ADVANCES IN AGRONOMY

Advisory Board
PAUL M. BERTSCH RONALD L. PHILLIPS
University of Kentucky University of Minnesota

KATE M. SCOW LARRY P. WILDING

University of California, Texas A&M University
Davis

Emeritus Advisory Board Members

JOHN S. BOYER KENNETH J. FREY
University of Delaware Iowa State University

EUGENE J. KAMPRATH MARTIN ALEXANDER

North Carolina State Cornell University
University

Prepared in cooperation with the

American Society of Agronomy, Crop Science Society of America, and Soil
Science Society of America Book and Multimedia Publishing Committee

DAVID D. BALTENSPERGER, CHAIR

LISA K. AL-AMOODI CRAIG A. ROBERTS
WARREN A. DICK MARY C. SAVIN
HARI B. KRISHNAN APRIL L. ULERY
SALLY D. LOGSDON
Academic Press is an imprint of Elsevier
525 B Street, Suite 1900, San Diego, CA 92101-4495, USA
225 Wyman Street, Waltham, MA 02451, USA
32 Jamestown Road, London, NW1 7BY, UK
Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands

First edition 2012

Copyright # 2012 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted

in any form or by any means electronic, mechanical, photocopying, recording or
otherwise without the prior written permission of the publisher

Permissions may be sought directly from Elsevier’s Science & Technology Rights
Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333;
email: [email protected]. Alternatively you can submit your request online by
visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting
Obtaining permission to use Elsevier material

Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons or
property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions or ideas contained in the material
herein. Because of rapid advances in the medical sciences, in particular, independent
verification of diagnoses and drug dosages should be made

ISBN: 978-0-12-394277-7
ISSN: 0065-2113 (series)

For information on all Academic Press publications

visit our website at elsevierdirect.com

Printed and bound in USA

12 13 14 15 10 9 8 7 6 5 4 3 2 1
CONTRIBUTORS

Numbers in Parentheses indicate the pages on which the authors’ contributions begin.

K. J. Boote (41)
Agronomy Department, University of Florida, Gainesville, Florida, USA
Jean-Pierre Caliman (71)
PT SMART Research Institute (SMARTRI), Pekanbaru, Riau, Indonesia
Qing Chen (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Xinping Chen (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
François Colin (71)
Montpellier-SupAgro, UMR-LISAH (Laboratory on Interactions between Soil,
Agrosystem and Hydrosystem), Montpellier cedex, France
Irina Comte (71)
Department of Natural Resource Sciences, Macdonald Campus of McGill Univer-
sity, Ste-Anne-de-Bellevue, Quebec, Canada, and CIRAD (International Coopera-
tion Centre in Agronomic Research for Development), Montpellier cedex, France
Zhenling Cui (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Mingsheng Fan (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Steven J. Fonte (123)
International Center for Tropical Agriculture (CIAT), Cali, Colombia
Olivier Grünberger (71)
IRD (Institut de Recherche pour le Développement), UMR-LISAH, Montpellier
cedex, France
Rongfeng Jiang (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Xiaotang Ju (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China

ix
x Contributors

Uttam Kumar (41)

International Crops Research Institute for the Semi-Arid Tropics (ICRISAT),
Patancheru, Andhra Pradesh, India
Patrick Lavelle (125)
International Center for Tropical Agriculture (CIAT), Cali, Colombia, and Institut
de Recherche sur le Développement (IRD)/Université Pierre et Marie Curie
(UPMC), Paris, France
Xin Li (219)
Department of Agronomy, Kansas State University, Manhattan, Kansas, USA
Xuejun Liu (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Guohua Mi (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Pedro Oyarzun (125)
EkoRural, Quito, Ecuador
Soroush Parsa (125)
International Center for Tropical Agriculture (CIAT), Cali, Colombia
D. Carolina Quintero (125)
International Center for Tropical Agriculture (CIAT), Cali, Colombia
Idupulapati M. Rao (125)
International Center for Tropical Agriculture (CIAT), Cali, Colombia
Terry J. Rose (185)
Southern Cross Plant Science, Southern Cross University, Lismore, NSW,
Australia
Jianbo Shen (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Piara Singh (41)
International Crops Research Institute for the Semi-Arid Tropics (ICRISAT),
Patancheru, Andhra Pradesh, India
Steven J. Vanek (125)
Department of Crop and Soil Science, Cornell University, Ithaca, New York,
USA
Jiankang Wang (219)
Institute of Crop Science and CIMMYT China, Chinese Academy of Agricultural
Sciences, Beijing, China
Contributors xi

Joann K. Whalen (71)

Department of Natural Resource Sciences, Macdonald Campus of McGill Univer-
sity, Ste-Anne-de-Bellevue, Quebec, Canada
Matthias Wissuwa (185)
Japan International Research Center for Agricultural Sciences (JIRCAS), Crop
Production and Environment Division, Ohwashi, Tsukuba, Ibaraki, Japan
Jianming Yu (219)
Department of Agronomy, Kansas State University, Manhattan, Kansas, USA
Fusuo Zhang (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Weifeng Zhang (1)
Department of Plant Nutrition, China Agricultural University, Beijing, PR China
Chengsong Zhu (217)
Department of Agronomy, Kansas State University, Manhattan, Kansas, USA
PREFACE

Volume 116 contains six excellent reviews dealing with environmental

sustainability and food security. Chapter 1 is an enlightening review on an
integrated nutrient management (INM) approach, developed on more than
20 years of research, to address serious environmental quality challenges,
related to excess use of nutrients, in China. The INM approach has led to
increased nutrient use efficiency and decreased inputs of fertilizers. Chapter
2 deals with the effect of climate change factors on crop growth, develop-
ment, and yield of groundnut. Chapter 3 is a comprehensive review on
practices used in oil palm plantations and impacts on hydrological changes,
nutrient fluxes, and water quality in Indonesia. Chapter 4 is an enlightening
overview of soil fertility decline in the high Andes of Bolivia, Ecuador, and
Peru. Approaches are presented to enhance nutrient cycling, crop nutrient
uptake, and overall increased productivity. Chapter 5 addresses an impor-
tant global factor affecting future food security, phosphorus utilization
efficiency (PUE) by plants. The review focuses on grain crops and covers
past attempts to improve PUE via plant breeding, and new approaches for
improving PUE. Chapter 6 is a stimulating review on the importance of
computer simulation in plant breeding.
I am grateful to the authors for their outstanding reviews.

DONALD L. SPARKS
Newark, Delaware, USA

xiii
 C H A P T E R O N E

Integrated Nutrient Management for

Food Security and Environmental
Quality in China
Fusuo Zhang, Zhenling Cui, Xinping Chen, Xiaotang Ju,
Jianbo Shen, Qing Chen, Xuejun Liu, Weifeng Zhang,
Guohua Mi, Mingsheng Fan, and Rongfeng Jiang

Contents
1. Introduction 2
2. Principles of INM 6
2.1. Optimizing nutrient inputs and taking all possible sources of
nutrients into consideration 9
2.2. Dynamically matching soil nutrient supply with crop
requirement spatially and temporally 11
2.3. Effectively reducing N losses in intensive managed Chinese
cropping systems 12
2.4. Taking all possible yield increase measures into consideration 15
3. Technology and Demonstration of INM in Different Cropping
Systems 16
3.1. INM for intensive wheat and maize system 18
3.2. INM for paddy rice 21
3.3. INM for vegetable systems 23
3.4. INM for orchards 26
4. Large-Scale Dissemination of INM 29
5. Summary and Conclusions 31
Acknowledgments 32
References 32

Abstract
While the concept of sustainability as a goal has become widely accepted, the
dominant agricultural paradigm still considers high yield and reduced environ-
mental impact being in conflict with one another. During the past 49years
(1961–2009), the 3.4-fold increase in Chinese agricultural food production can

Department of Plant Nutrition, China Agricultural University, Beijing, PR China

Advances in Agronomy, Volume 116 # 2012 Elsevier Inc.

ISSN 0065-2113, DOI: 10.1016/B978-0-12-394277-7.00001-4 All rights reserved.

1
2 Fusuo Zhang et al.

be partly attributed to a 37-fold increase in N fertilization and a 91-fold increase

in P fertilization, but the environment costs have been very high. New advances
for sustainability of agriculture and ecosystem services will be needed during
the coming 50years to improve nutrient use efficiency (NUE) while increasing
crop productivity and reducing environmental risk. Here, we advocate and
develop integrated nutrient management (INM) based on more than 20years
of studies. In this INM approach, the key components comprise (1) optimizing
nutrient inputs by taking all possible nutrient sources into consideration, (2)
matching nutrient supply in root zone with crop requirements spatially and
temporally, (3) reducing N losses in intensively managed cropping systems, and
(4) taking all possible yield-increasing measures into consideration. Recent
large-scale application of INM for cereal, vegetable, and fruit cropping systems
has shed light on how INM can lead to significantly improved NUE, while
increasing crop yields and reducing environmental risk. The INM has already
influenced Chinese agricultural policy and national actions, and resulted in
increasing food production with decreased climb of chemical fertilizer consump-
tion at a national scale over recent years. The INM can thus be considered an
effective agricultural paradigm to ensure food security and improve environmen-
tal quality worldwide, especially in countries with rapidly developing economies.

Abbreviations

AEN agronomy N efficiency

FNP farming practice
INM integrated nutrient management
NCP North China Plain
NUE nutrient use efficiency
ONR optimum N fertilizer rate
PFPN nitrogen partial factor productivity
REN recovery N efficiency

1. Introduction
The Green Revolution helped to create the world’s “Miracle in China,”
with 9% of the world’s arable land feeding 22% of the world population. In the
past 49years (1961–2009), cereal grain yields have increased 3.5-fold from 1.2
to 5.4tha
1, while total grain production has increased 3.4-fold from 110 to
483 million ton (MT) (FAO, 2011). In 1998, grain, meat, and egg production
per capita in China exceeded the world average. The increased demand in
Chinese grain production has affected the global food supply and the natural
Nutrient Management in China 3

resource bases required for nutrient production (fossil fuels, mineral sources of
P and K) and has attained world recognition.
However, this 3.4-fold increase in Chinese agricultural food production
during the past 49years can be partly attributed to a 48-fold increase in
chemical fertilizers from 1 to 49MT, including a 37-fold increase in N
fertilizer application and a 91-fold increase in P fertilizer use, and a 442-fold
increase in the area of irrigated croplands (Fig. 1). Total consumption of
chemical fertilizers worldwide increased by 3.9-fold from 32 to 164MT,
indicating that 36% of the global increase (132MT) came from China
during the past 49years. In the past 10years (2000–2009), 54% of the global
increase in chemical fertilizer consumption (27MT) was contributed
by China, including 11MT fertilizer N (54% of the global increase), 2.5
MT fertilizer P (52% of the global increase), and 1.1MT fertilizer K (58% of
the global increase) (Figs. 1 and 2A,B).
Cereal yields in the past 10years have continued to increase with no
proportional increases in fertilizer use in many developed countries or
regions such as Western Europe (rainfed cereal systems), North America
(rainfed and irrigated corn), and Japan and South Korea (irrigated rice)
(Dobermann and Cassman, 2005). For example, in the past 10years, chem-
ical fertilizer consumption in the United States increased by only 0.04MT
with 0.23% of total fertilizer consumption in 2009 and decreased by 0.32
MT in Western Europe (Fig. 2A). By contrast, the application rate of

600 80
Grain production
500 Total fertilizer Fertilizer consumption (MT)
N fertilizer
Grain production (MT)

60
400 P fertilizer
K fertilizer
300 40

200
20
100

0 0
1960 1970 1980 1990 2000 2010
Year

Figure 1 The trend of grain production and chemical fertilizer inputs (N, P, and K
fertilizers) in China from 1961 to 2009. The P and K fertilizers are calculated by P2O5
and K2O, respectively. Fertilizer consumption is defined as the difference between
fertilizer production and exports. Source: FAO (2011) and IFA (2011).
4 Fusuo Zhang et al.

200 80
A
160

Regional fertilizer (MT)

Global fertilizer (MT) 60

120
40
80

20
40

0 0
1960 1970 1980 1990 2000 2010
Year

600
B
500
Fertilizer rate (kg ha )
-1

400

300

200

100

0
1960 1970 1980 1990 2000 2010
Year

Global fertilizer China

United States Western Europe

Figure 2 Trend of total chemical fertilizer consumption (A) and fertilizer rate per
hectare (B) for global scale, China, United States, and Western European. Source: IFA
(2011).

chemical fertilizers in China was continually increasing and reached 448kg

ha
1 in 2009, which is 2.8, 2.9, and 1.4 times the world average and rates in
the United States and Western Europe, respectively (Fig. 2B).
On the other hand, Chinese cereal crop production has stagnated at
approximately 450MT since 1998. From 1998 to 2009, grain yields increased
Nutrient Management in China 5

by only 10%, while the consumption of chemical fertilizers increased by

nearly 49%, 19%, and 33% for N, P, and K, respectively (Fig. 1). That
means that the large increase in fertilizer nutrient inputs did not result in a
corresponding yield increase in the past decade in China. For example, the
REN (the percentage of N fertilizer recovered in the aboveground plant parts
at maturity) in Chinese cereal grain production decreased from about 35% in
the 1980s (Zhu, 1998) to 28% in the 2000s (Zhang et al., 2008a), lower than
the world average of 33% (Raun and Johnson, 1999). Often twice as much
fertilizer N or P is applied compared with the removed nutrients by crops,
and this nutrient imbalance in turn drives severe environmental problems,
such as eutrophication of surface waters (Le et al., 2010), soil acidification
(Guo et al., 2010), greenhouse gas emissions (Zheng et al., 2004), and other
forms of air pollution (Liu et al., 2011). For example, about 60% of inland
lakes in China show eutrophication, and 57% of N inputs and 67% of P inputs
are derived from agriculture (Chinese Ministry of Environmental Protection,
2010). Soil pH declined significantly (P<0.001) from the 1980s to the 2000s
in the major Chinese crop lands due to overuse of N fertilizer (Guo et al.,
2010). On the North China Plain (NCP), total wet and dry deposition of N
averaged 80–90kgNha
1 yr
1 in the 2000s (Liu et al., 2006b; Shen et al., 2009;
Zhang et al., 2008b), a value nearly 10 times that at Rothamsted, Harpenden,
UK (Goulding et al., 1998) or in central New York in the USA (Fahey et al.,
1999). These problems are meaningful on a global scale.
To meet the demand for grain and to feed a growing population on the
remaining arable land by 2030, crop production must reach 5.8MT (an
increase of >40%) and yields have to increase by 2% annually (Zhang et al.,
2011). Due to environmental and economic (e.g., rising cost of fossil fuels)
constraints, further increases in food supplies projected for the coming 50
years must be attained through improved resource use efficiency rather than
more agricultural inputs, especially N and P fertilizer applications (Cassman,
1999; Matson et al., 1997; Tilman et al., 2002). Toward this end, sound
agronomic and environmentally acceptable integrated nutrient management
(INM) is an essential approach for the achievement of a reduction in fertilizer-
derived environmental risk while also increasing crop productivity and NUE.
In most intensive agricultural areas, however, current nutrient manage-
ment strategies are focused on delivering soluble inorganic N and P from
fertilizers directly to crops and have uncoupled soil and environmental N
and P cycles spatially and temporally. As a result, agricultural ecosystems are
maintained in a state of N saturation and are inherently leaky because chronic
surplus additions of N and P are required to meet the goal of maximum yields
(Drinkwater and Snapp, 2007). For example, the N and P surpluses in
intensive wheat–maize systems on the NCP were recently estimated to be as
high as 227 and 53kgha
1 yr
1 (Vitousek et al., 2009). Therefore, all these
approaches have been successful in terms of maintaining grain yields; however,
attempts to reduce nutrient losses and improve NUE have met with limited
6 Fusuo Zhang et al.

success in intensive agricultural areas (Cassman et al., 2002; Drinkwater and

Snapp, 2007).
In INM, crop yields can be increased while minimizing nutrient losses to
the environment by managing nutrient supply in the root zone within a
reasonable range, which realizes the biological potential of crops, matches
high-yielding crop N requirement, and controls minimal nutrient losses.
Nutrient supply and nutrient requirements in high-yielding cropping sys-
tems must be matched in quantity and synchronized in time and space
(Chen et al., 2010; Cui et al., 2010a). To realize this goal, some improve-
ments must be made: using a variety of N sources from fertilizers, the
environment, and the soil to meet crop demand; calculating the nutrient
balance between the inputs and outputs to manage a variety of intrinsic
ecosystem processes at multiple scales to recouple elemental cycles; and
considering the biological potential of the root system and matching crop
requirements by supplying sufficient N only when plant demand exists (Cui
et al., 2010a; Fig. 3). In this chapter, we discuss the principles of INM and
the development of INM technology on a large scale with dissemination of
INM in different cropping systems up to national scale.

2. Principles of INM
The overall principle of INM is to maximize biological potential for
improving crop productivity and resources use efficiency through root
zone/rhizosphere management. Plant roots take up nutrients from soils

Nutrient supplies in root-zone Nutrient demand for high-yield crop

Nutrient from Bioavailable Characteristics Total and Root
environment nutrient of NPK periodic demand response

Nutrient management strategy

Quantify root- In-season Building-up and Correction when
zone nutrient management maintenance for deficient for
supplies for N P and K trace elements

High-yielding crop Optimal water

management management

Integrated Nutrient Management

Figure 3 Conceptual model illustrating the principles of Integrated Nutrient Manage-

ment (INM).
Nutrient Management in China 7

via the rhizosphere, a narrow zone of the soil that is directly influenced by
root growth, root secretions, and associated soil microorganisms. In crop-
ping systems, a rhizosphere continuum in the root zone can be formed due
to root/rhizosphere interactions among individual plants. The rhizosphere
is the important interface where interactions among plants, soils, and micro-
organisms occur and is a “bottleneck” controlling nutrient transformations,
availability, and flow from soils to plants. Therefore, the chemical and
biological processes occurring in the rhizosphere determine the mobiliza-
tion and acquisition of soil nutrients together with microbial dynamics, and
also control NUE by crops, and thus profoundly influence cropping system
productivity and sustainability (Zhang et al., 2004, 2010).
As plant growth proceeds, the roots can respond to and/or sense changes
in soil nutrient availability including nutrient supply intensity and composi-
tion. These responses involve a series of adaptive alterations in root mor-
phology and root physiology. P-deficient plants can commonly increase
their root/shoot ratio, root branching, root elongation, root topsoil forag-
ing, and formation of cluster roots and root hairs (Lynch and Brown, 2008;
Shen et al., 2011b; Vance, 2008). Mycorrhizal associations can also enhance
the spatial availability of P, extending the nutrient absorptive surface by
formation of mycorrhizal hyphae (Marschner, 1995). On the other hand,
root-induced chemical and biological changes in the rhizosphere affect the
bioavailability of soil P, mainly involving rhizosphere acidification, carboxy-
late exudation, secretion of phosphatases or phytases, and Pi transporter
expression (Neumann and Römheld, 2002; Zhang et al., 2010). It has been
reported that P deficiency increases the formation of cluster roots by white
lupin (Lupinus albus L.; Shen et al., 2005; Wang et al., 2007), axial root length
and total root length, and larger amounts of lateral roots and more root hair
formation in maize (Zea mays L.) or Arabidopsis (Bates and Lynch, 1996;
Linkohr et al., 2002; Liu et al., 2004b; Schachtman et al., 1998; Schenk and
Barber, 1979). In crop species, Liu et al. (2004b) found that efficient use of
P in calcareous soil by maize is related to its large root system, with a greater
ability to acidify the rhizosphere, and a positive response of acid phosphatase
production and excretion in low P conditions. High P acquisition efficiency
by modifying root morphology and root physiology in terms of rhizosphere
biological and chemical processes is important for achieving high crop
yields with savings in nutrient inputs. The nutrient supply intensity or
concentrations in the rhizosphere/root zone in cropping systems can be
optimized to a critical level through nutrient management to maximize the
biological potential for efficient use of soil P by plants.
Nitrogen fertilization is the most common practice for the regulation of
root growth in field conditions. Maize roots respond to N supply in two
ways. First, in uniform N supply systems, N deficiency increases maize root
length, resulting in longer axial roots (primary, seminal, and nodal roots;
Tian et al., 2006; Wang et al., 2003). This helps the roots to explore a larger
8 Fusuo Zhang et al.

soil volume and thus increases spatial N availability. However, root elonga-
tion can be inhibited if the N supply is too high. In maize, for example, the
optimum nitrate level for root length seems to be around 5mmolL
1 (Tian
et al., 2008). Second, root growth can be stimulated when plant roots
experience nutrient-rich patches, particularly when the patches are rich in
N and P (Drew, 1975; Hodge, 2004). When a maize plant is suffering from
N deficiency and part of the root mass is supplied with nitrate locally, the
growth of lateral roots in the supplied area is enhanced (Granato and Raper,
1989; Guo et al., 2005; Sattelmacher and Thoms, 1995). This helps plants to
compete with other plant species and/or microbes for limited N resources
(Hodge, 2004). It is suggested that NO
3 plays a key role as a nutrient signal
in regulating root proliferation (Zhang and Forde, 1998). Localized P
application effectively enhances crop growth and P use efficiency. More-
over, manipulating and managing nutrient supply intensity and composition
in the local fertilization zone can greatly strengthen root growth and
nutrient uptake through modifying rhizosphere processes and enlarging
the root absorbing surface. A field experiment showed that localized appli-
cation of P with addition of ammonium significantly enhanced P uptake
and crop growth through stimulating root proliferation and rhizosphere
acidification ( Jing et al., 2010). The leaf expansion rate was 20–50% higher,
the total root length 23–30% greater, and the plant growth rate 18–77%
greater with a localized supply of P plus ammonium compared with broad-
casting of these nutrients. Localized application of P combined with addi-
tion of ammonium significantly decreased rhizosphere pH in the fertilized
zone compared with the bulk soil ( Jing et al., 2010). The results suggest that
modifying rhizosphere processes in the field may be an effective manage-
ment strategy for increasing NUE and plant growth.
Rhizosphere management emphasizes maximizing the efficiency of
root/rhizosphere processes in nutrient acquisition and use by crops rather
than simply depending on excessive fertilizer inputs, which involves reg-
ulating the root system, rhizosphere acidification, carboxylate exudation,
microbial associations with plants, rhizosphere interactions in terms of
intercropping and rotation (Li et al., 2007), localized application of nutri-
ents, use of efficient crop genotypes and synchronizing rhizosphere nutri-
ent supply with crop demand. Rhizosphere management has been shown
to be an effective approach for increasing NUE and crop productivity
through “small causes with big effects” for sustainable agricultural produc-
tion (Zhang et al., 2010). Based on a better understanding of rhizosphere
processes, the key steps of INM are (1) optimizing nutrient inputs and
taking all possible sources of nutrients into consideration, (2) dynamically
matching soil nutrient supply with crop requirement spatially and tempo-
rally, (3) effectively reducing N losses in intensively managed Chinese
cropping systems, and (4) taking all possible yield increase measures into
consideration (Fig. 4).
Nutrient Management in China 9

Figure 4 Rhizosphere/root-zone nutrient management is a key component of INM

for achieving high grain yield and high NUE at the same time.

2.1. Optimizing nutrient inputs and taking all possible

sources of nutrients into consideration
Since the 1990s, excessive chemical N fertilization has often been considered
as the main practical strategy to pursue high yields in China. The average N
fertilizer application rate has far exceeded crop requirements for maximum
grain yield, up to double the crop N demand in some areas (Cui et al., 2010a).
Clearly, applying large amounts of N fertilizer does affect grain yield and N
uptake but also increases the potential for N losses to the environment. For
example, N fertilizer could be cut from 588 to 286kgNha
1 yr
1 without
a loss in yield or grain quality and, in the process, reduce N losses by <50%
( Ju et al., 2009). Therefore, we believe that “Controlling the total fertilizer N
application rate” should be a top priority policy and practice to reduce
overuse of N in China under current conditions.
Increasing N fertilizer application rates and associated increases in envi-
ronmental pollution have led to N derived from the soil and the environment
becoming an important N source for crop plants in China. Across numerous
on-farm experiments (n¼269), indigenous N supply (average N uptake in
control) typically provides around 274kgNha
1 yr
1 and accounts for 76% of
crop N uptake in intensive wheat–maize systems (Cui et al., 2010a). As a
10 Fusuo Zhang et al.

result, high crop yields can be readily obtained in arable soils without
application of N fertilizers (Tong et al., 2004). This large indigenous N supply
aggravates N surpluses and increases the potential for N losses from agro-
ecosystems unless it is considered to be a component part of the plant-
available N when constructing an integrated N management plan.
The large indigenous N supply is attributed to high soil nitrate-N accu-
mulation and environmental N supply. Soil nitrate-N (NO
3 -N) accumula-
tion in the top 90 or 100cm of the soil was above 200kgNha
1 under
conventional N practice in intensive wheat–maize systems (Cui et al.,
2008a,d; Liu et al., 2003); this residual NO
3 -N supply can reach 1173 and
613kgNha
1, respectively, in greenhouse vegetable and orchard systems in
North China ( Ju et al., 2004, 2006). Ju et al. (2006) observed that residual soil
nitrate-N after winter wheat harvest was 275kgNha–1 in the top 90cm of the
soil profile and 213kgNha–1 in the 90–180cm soil depth increment. Liu et al.
(2004a) reported an average residual soil nitrate-N content of 314kgNha–1 in
the top 2m of the soil profile and 145kgNha–1 at 2–4m soil depth based on
on-farm soil tests in annual winter wheat production systems in Beijing
suburbs on the surroundings of NCP (n¼93). Wheat grain yields on the
NCP showed no response to applied N when initial nitrate-N before
sowing in the top 90cm soil layer exceeded 200kgNha–1, but residual
nitrate-N content after harvest and N losses significantly increased (Cui
et al., 2008a). High nitrate-N accumulation in the soil profile is like a “time
bomb” that could explode at any moment and will finally be lost to the
environment through either denitrification or leaching under high N appli-
cation rates ( Ju et al., 2009; Zhao et al., 2006). Ju et al. (2002) observed
residual nitrate-N from the first growing season (wheat) to leach from the
0 to 1m soil profile during the second growing season (maize) on the NCP
due to high rainfall in summer.
The total amount of N available from the environment in China in the
2000s has more than doubled compared with the 1980s because of the rapid
continuing increases in both oxidized and reduced N emissions (Liu and
Zhang, 2009; Liu et al., 2010, 2011). According to Fig. 5, N inputs from
atmospheric N deposition and irrigation water in China were up to 33 and
12kgNha
1 yr
1 in the 2000s but only 14 and 4kgNha
1 yr
1 in the 1980s.
Similar rapid increases in environmental sources of N from atmospheric
deposition and irrigation water were reported on the NCP and in the Taihu
Lake region ( Ju et al., 2009). Nitrogen input from deposition plus irrigation
was only 30kgNha
1 yr
1 on the NCP and in Taihu Lake region in the
1980s, and this value has increased to 99 and 89kgNha
1 yr
1, respectively,
in these two intensive agricultural regions in the 2000s ( Ju et al., 2009).
Evidence from the NCP indicates that dry N deposition (50–60kgNha
1
yr
1) (Shen et al., 2009, 2011a) is likely to be a major contributor to the
environmental N in this semihumid/semiarid region compared with wet
N deposition (30kgNha
1 yr
1) (Liu et al., 2006b; Zhang et al., 2008b).
Nutrient Management in China 11

35
1980s
30
2000s
N input (kg N ha-1 yr -1) 25

20

15

10

0
N deposition N from irrigation

Figure 5 Changes of N input from atmospheric deposition and irrigation in Chinese

croplands during 1980s and 2000s. Modified from Liu et al. (2011).

Further studies confirm that about 50% of this air-borne N input (52kg
Nha
1 yr
1) can be utilized by current crops, according to the ITNI-system
which is based on a 15N dilution approach (He et al., 2007a, 2010).

2.2. Dynamically matching soil nutrient supply with crop

requirement spatially and temporally
Nutrient applications that meet, but do not exceed, crop nutrient require-
ments are essential for achieving maximum yields and minimizing environ-
mental risk. Recent research on improving NUE in crop production systems
has emphasized the need for greater synchrony between crop nutrient
demand and nutrient supply from all sources throughout the growing season
(Cassman et al., 2002; Cui et al., 2010a; Fan et al., 2008). Nitrogen fertilizers
applied during periods of the greatest crop demand, at or near the plant roots,
and in small quantities and frequent applications, can potentially reduce losses
while maintaining or increasing crop yield and quality (Matson et al., 1998;
Tilman et al., 2002).
In INM, we attempt to manage soil nutrient supply in the root zone
within a reasonable range that matches the quantity required by the crop, is
synchronized in terms of time and crop growth, and is coupled in space to
nutrient supply and crop nutrient requirements, especially in N manage-
ment (Chen et al., 2010, 2011; Cui et al., 2010b). Dry matter production
and thus N requirement is relatively low before the rapid growth stages of
crops, that is, stem elongation of wheat and the expanded leaf stage of maize.
In most cases, a small amount of N fertilizer and indigenous N supply can
12 Fusuo Zhang et al.

meet the crop N requirement, establish growth, and promote the develop-
ment of healthy roots during the early periods of crop growth (Cui et al.,
2008a,d). Nitrogen application in excess of crop N demand during this
period increases the potential for nitrate-N leaching and results in excessive
crop growth that is more susceptible to disease and lodging. Therefore,
more N fertilizer (around 60–70% of the total N fertilizer input) should be
applied during the most rapid growth stages of the crop to achieve synchro-
nization between the N supply and crop demand (Cui et al., 2008b,c).
Restricted by old knowledge and habits, farmers often apply large
amounts of N fertilizers before planting or at the early growth stages as a
conventional management practice in crop production. For example, the N
supply before planting is usually about 50% of the total amount given (Chen,
2003; Cui et al., 2008c,d). Many rice farmers in China apply 50–90% of the N
as a basal dressing and an early top-dressing within the first 10 days after
transplanting to reduced transplanting shock and stimulate early tillering
(Zhang et al., 2011). Clearly, this large amount of N fertilizer at the early
growth stages has resulted in poor synchronization between soil N supply and
crop demand, leading to high soil inorganic N concentrations before the
occurrence of rapid crop N uptake (Chen et al., 2006; Tilman et al., 2002).
Although the average N application rate in China is excessive for the
maximum crop N requirement, some low-income farmers or those in
remote areas apply N inadequately. If we simply use 150–200kgNha–1 as
a reasonable amount of N fertilizer for the main wheat and maize growing
regions of China (n¼10,000), only one-third of farmers would be in this
range, one-third would be applying too much (N rate>200kgNha
1), and
one-third would be applying too little (N rate<150kgNha
1) (Wang,
2007). On a regional scale, higher crop yields are likely to be achieved
through a combination of increased N application in regions with a low N
input and improved NUE in regions where N fertilizer use is already high.

2.3. Effectively reducing N losses in intensive managed

Chinese cropping systems
The fate of fertilizer N in cropping systems is an integrated result of crop N
uptake, immobilization, and residues in the soil, and losses to the environ-
ment, including ammonia volatilization, denitrification, N leaching and
runoff. There are close relationships among these three major fates of applied
fertilizer N, which are influenced by many factors such as crop characteristics,
soil properties, climatic conditions, and management practices. According
to an integrated study, Ju et al. (2009) found that the mechanisms of N loss
were very different in two major intensively managed crop rotation systems,
that is, wheat–maize and rice–wheat systems. Understanding N processes will
help us control the N losses and their harmful environmental effects while
maintaining high crop productivity.
Nutrient Management in China 13

Under current farming practice (about 550kgNha
1 yr
1 in both crop-

ping systems and usually without deep placement of urea or ammonium
bicarbonate), ammonia volatilization in the wheat–maize is about 20–25%
of applied N and this is the main N loss pathway due to high pH (around
8.0) in calcareous soils together with surface broadcasting of urea before
irrigation or rainfall under most farmers’ practices (Fig. 6). In paddy soils,
ammonia volatilization is about 12% of applied fertilizer N in the rice
growing phase and is very low and about 2% in the wheat phase (Fig. 6).
Another important N loss pathway in paddy soils is denitrification, which
was found to be 36–44% of applied N (Fig. 6). Because of low levels of
groundwater in these soils and copious rainfall in the wheat season, drying
and wetting cycles frequently occur and these are favorable for denitrifica-
tion (Zhao and Xing, 2009), with N2 and N2O representing the main
products of denitrification. Within wheat/maize systems under conven-
tional agricultural N management practices, nitrate-N leaching below the
root zone was found to be the important N loss pathway (Liu et al., 2003)
due to concentrated rainfall in summer and flooding irrigation, which
contradicts the conventional wisdom that leaching losses are not an impor-
tant N loss pathway in semihumid upland agricultural systems on the NCP
( Ju et al., 2004).

50
45 NH3 volatilization
40 Leaching
Denitrification
35
N losses (%)

30
25
20
15
10
5
0
Rice Wheat-South Wheat-North Maize
Taihu region North China Plain
Crops and regions

Figure 6 N loss pathways expressed as a percentage of N application rates in farmers’ N

practices for the Taihu region and North China Plain. N fertilizer rates for rice and wheat
are 300 and 250kgNha
1 in the Taihu region and are 325 and 263kgNha
1 for wheat
and maize in North China Plain, respectively. Modified from Ju et al. (2009).
14 Fusuo Zhang et al.

Total fertilizer N losses in wheat–maize rotation averaged 31%, lower

than the 48% in rice–wheat rotation (Fig. 6); correspondingly, the residual
N in the soil is higher than that in the latter system, so that the capacity to
retain synthetic N is higher than that in rice–wheat. However, released
reactive N in wheat–maize rotation systems is higher than in rice–wheat
rotations. Upland–upland crop rotations on calcareous soils lead to substan-
tial ammonia volatilization and nitrate-N accumulation or leaching, which
are the main loss pathways of fertilizer N in North China, and the frequent
flooding and drying cycles lead to N losses via denitrification and ammonia
volatilization in paddy-upland crop rotation systems ( Ju et al., 2009).
In practice, the amount of NH3 volatilization is largely influenced by
fertilization method and N fertilizer type. Deep placement of urea or ammo-
nium bicarbonate increases N use efficiency. For example, N loss through
NH3 volatilization during the maize growing season can be reduced by
11–48% of applied urea-N fertilizer with deep placement compared with
surface broadcasting (Li et al., 2002; Zhang et al., 1992). In China, N is
generally recommended to be applied to wheat before irrigation because of a
very low recovery of broadcast urea when precipitation is low and the fertilizer
cannot reach the rooting zone. In contrast, when urea is applied before
irrigation or plowing, NH3 volatilization is significantly reduced, indicating
the necessity for irrigation or rainfall to leach N fertilizer into the rooting zone.
The strategy to reduce nitrate leaching can be achieved simply by
decreasing the nitrate-N content in the soil profile. For environmentally
sound crop production, the residual nitrate-N content in soil should be
minimized, especially at the end of the growing season, because nitrate-N
leaching is directly related to the mineral N content in the root zone
(Dinnes, et al., 2002; Roth and Fox, 1990; Sogbedji, et al., 2000). However,
achieving high maize yields is impossible if nitrate-N in the root zone is too
low. It then becomes important to consider “the suitable soil nitrate-N level
at which the lower limit does not restrict grain yield and the upper limit
does not lead to unacceptable N losses to the environment.” According to
European experience and our own results (Cui et al., 2008b,c; Hofman,
1999; Schleef and Kleihanss, 1994), residual nitrate-N content after harvest
in the top 90cm soil layer should be maintained around 90kgNha
1. When
inorganic N exceeds 190kg nitrate-N ha
1 in the top 90cm of soil profiles
before planting, an increase in grain yield in response to added N is unlikely
in Chinese intensive wheat–maize systems (Cui et al., 2008b,d).
Application of nitrification inhibitors such as DCD, DMPP, and nitra-
pyrin together with NHþ 4 -based fertilizer can reduce N2O emission by 77%
on the North China because N2O emissions occur mostly during the
nitrification processes after fertilizer N application and irrigation ( Ju et al.,
2011). Recently, we estimated that N2O emission was 33.1GgNyr
1 from
paddy fields and 255.3GgNyr
1 from upland soils in 2007 (Gao et al.,
2011). If 50% of the total area is taken into account and 77% reduction by
Nutrient Management in China 15

using nitrification inhibitors in upland crops, the total emission in upland

soil will be 169.7GgNyr
1. The total N2O emission from Chinese crop-
lands would be reduced by 30% from the current amount with careful use of
nitrification inhibitors.

2.4. Taking all possible yield increase measures into

consideration
Many recently developed approaches and tools for fine-tuning N management
have increased the NUE by decreasing the N fertilizer rate, but substantial
and consistent yield increases have been demonstrated in only a few studies
(Dobermann and Cassman, 2005). The major challenge of current nutrient
management in China is how to increase crop yields to meet food demand
while also increasing NUE to protect the environment. Large numbers of
experiments indicate that the grain yield potential of currently grown cereal
varieties in China far exceeds actual yields obtained. For example, the average
maize yields in farmers’ fields are 5.3Mgha
1 in northeast China, 5.1tha
1
on the NCP, and 4.0tha
1 in hilly areas in south China (ECCAY, 2006).
However, maize yields in new variety trials in these regions typically average
8.5, 7.3, and 6.7tha
1, 60%, 45%, and 68% above the average farming yields
(Fan et al., 2010b). The highest attainable maize yields recorded, achieved with
high inputs of nutrients, water, and labor, were 16.8tha
1 in northeast China,
18.0tha
1 in the NCP, and 14.7tha
1 in south China (Fan et al., 2010a,b).
Similar results have been obtained for wheat and rice over the whole country.
This implies that, when an integrated management approach to crop produc-
tion is used, there is great potential to increase cereal grain yields above current
farming yields with significant enhancement of the ability of Chinese agricul-
ture to meet food demands in the coming decades.
Increased yields can be attributed to greater NUE from both indigenous
and applied nutrient sources, especially N sources, because rapidly growing
plants have larger root systems that more effectively exploit available soil
resources (Cassman et al., 2002). Crop health, insect and weed management,
moisture and temperature regimes, supplies of nutrients, and use of the best
adapted cultivars or hybrids all contribute to efficient uptake of available
nutrients and conversion of plant nutrients to grain yields. Key contributory
factors include (a) increased yields and more vigorous crop growth asso-
ciated with greater stress tolerance of modern hybrids, (b) improved man-
agement of production factors other than N (conservation tillage, seed
quality, and higher plant densities), and (c) improved N fertilizer manage-
ment. For example, the nitrogen partial factor productivity (PFPN) in North
China increased to 50 and 47kgkg
1 using integrated N management for
wheat and maize production compared with 15 and 25kgka
1 over those
with farmers’ N management practice. The PFPN could further to 60 and
59kgkg
1 when N-efficient varieties were used (Cui et al., 2009, 2011).
16 Fusuo Zhang et al.

Additional improvements in soil quality can occur when the benefits of C

sequestration are coupled with increases in crop yields from adoption of
cultivation practices that reduce yield losses from abiotic and biotic stresses,
such as returning straw back to the soil, increasing applications of organic
manures, and using reduced tillage. Over the past two decades, the soil organic
matter content of soils in north China has significantly increased because of
increasing incorporation of crop residues and organic manure and the devel-
opment of no-tillage and reduced-tillage practices (Huang and Sun, 2006).
Crop yields increased in this region at the same time (ECCAY, 2006).

3. Technology and Demonstration of INM in

Different Cropping Systems
Nitrogen is unique among the major nutrients since it is synthesized
from the atmosphere using the Haber–Bosch process and its transformation
and transport in the pedosphere and hydrosphere are mediated almost
entirely by biological processes (Galloway and Cowling, 2002). As a result,
N is mobile, hard to contain, and even N that is efficiently conserved and
taken away in crop harvest eventually makes its way back into the environ-
ment (Robertson and Vitousek, 2009). Hence, efficient use of N in crop
production is essential for increasing crop yields, environmental safety, and
the consequent economic considerations (Campbell et al., 1995).
Recent literature on improving N use efficiency in crop production has
emphasized the need for greater synchronization between crop N demand
and N supply from all sources throughout the growing season (Cassman et al.,
2002; Chen et al., 2010; Cui et al., 2009). Considering site-specific soil N
supply and crop demand, current research has demonstrated that in-season
applied N results in a high grain yield and high N use efficiency (Table 1; Cui
et al., 2009; Chen et al., 2010). As a net result of soil N transformation,
transport, and previous fertilizer applications, soil N supply can significantly
influence crop N uptake. It is very important to quantify the total N balance
including initial soil Nmin (NO
þ
3 þNH4 ), N mineralization, environmental
N supply, crop N uptake, N losses, and N immobilization during the growing
season. However, it is very difficult to measure accurately N balance with all
components, especially under field conditions (Blankenau et al., 2000; Engels
and Kuhlmann, 1993). The amount of soil Nmin at rooting depth may be a
good index to estimate the N balance situation. Substantial attempts should be
made to manage the N supply in the root zone in order to match the total
quantity of N required by the crop in both space and time.
Unlike N, management of fertilizer P and K focuses on maintenance of
adequate soil available P or K levels to ensure that neither P nor K supply
limits crop growth or becomes excessive due to overfertilization (Table 1).
Nutrient Management in China 17

Table 1 Resource characteristics of nutrient and the technologies of INM for different
nutrients

Nutrients Resource characteristics Management strategy

N Diverse sources In-season root-zone
Multidirectional losses management
Serious environmental harm
Crop sensitivity
P and K Limited resources Building-up and
Large soil pools but low maintenance approach
effectiveness
Long-term effects
Crop sensitivity
Micronutrients or Yield failure with deficiency Correction when
trace elementsa Too much causing toxicity deficiency
Just meeting crop demand
a
Trace elements include boron (B), chlorine (Cl), copper (Cu), iron (Fe), manganese (Mn), molybde-
num (Mo), nickel (Ni), and zinc (Zn), respectively.

In China, maintenance of fertilizer P or K rates is recommended on the basis

of constant monitoring of soil nutrient supplies and nutrient holding capa-
cities (Li et al., 2011; Wang et al., 1995). In soils with low P status and/or
high fixation capacity, capital investment is required to build up soil nutri-
ents until the system becomes profitable and sustainable. On soils with
moderate P and K levels and little fixation, management must focus on
balancing inputs and outputs at field and farm scales to maximize profit,
avoid excessive accumulation, and minimize risk of P losses (Dobermann,
2007). Therefore, managing nutrients to achieve synchrony between nutri-
ent supply and crop demand is crucial to increase NUE while maintaining
agricultural productivity and improving technical operability.
The management strategy of trace elements such as boron (B), chlorine
(Cl), copper (Cu), iron (Fe), manganese (Mn), molybdenum (Mo), nickel
(Ni), and zinc (Zn) is based on their plant availability in both soil and plant.
Their available contents in soil and (or) critical concentrations in plants are
determined and then the corresponding micronutrient fertilizers are applied
so that trace elements do not become a yield limiting factor (Table 1). This
strategy follows the law of the minimum and uses dose–response curves.
This type of curves for all the essential micronutrients show that the yields
can be affected by deficiencies and can also be reduced by toxicity due to
excessive concentrations. Not all micronutrients should therefore be applied
in the field. It is therefore important to monitor soils and/or crops to ensure
that the available micronutrient concentrations in soils are within the
optimum range and neither too low nor too high (Alloway, 2008).
18 Fusuo Zhang et al.

3.1. INM for intensive wheat and maize system

From 1961 to 2009, wheat production in China increased sevenfold from
14 to 115MT and maize production also increased sevenfold from 18 to 164
MT. In 2009, wheat production represented 24% of the Chinese cereal
output and 17% of global wheat output (FAO, 2011). Similarly, maize
production represented 34% of Chinese cereal output and 20% of global
corn output (FAO, 2011). The rapid development of wheat and maize
production has been attributed to increasing grain yield. During the same
period (1961–2009), wheat grain yield increased seven times from 0.56 to
4.74tha
1 and maize grain yield increased three times from 1.18 to 5.26tha
1
(FAO, 2011).
With simultaneous increases in fertilizer application rates and grain yields
before the mid 1990s, especially N and P fertilizers, most extension staff and
farmers still believed that the more fertilizer they use and the higher the
grain yields that would be obtained. Based on extensive on-farm investiga-
tions from 1997 to 2007 (Cui et al., 2010a; Ma et al., 1999), the typical
annual N rate for farmers was >500kgNha–1 for the intensive wheat–maize
rotation system in the NCP and approached 600kgNha–1 in some regions.
Average grain yields were around 11.5tha–1 yr–1 (around 5.5 and 6.0tha–1
for wheat and maize, respectively), and the estimated N uptake was only
300kgNha–1 yr–1 (around 160 and 140kgNha–1 for wheat and maize,
respectively). Results from region-wide experiments have demonstrated
that the economically optimal N rate is 130–160kgNha–1 crop–1 for the
intensive wheat–maize system (Cui et al., 2008b,c). Similar results showing
excess fertilizer application also observed for P. As a result, nutrient sur-
pluses in the main cropping systems, and hence environmental losses, are
very high (Vitousek et al., 2009). In addition, wheat and maize production
in China has stagnated since 1996. From 1996 to 2009, wheat and maize
grain yields increased by only 1% and 27%, with annual growth rates of
<0.1% and <2% (FAO, 2011).
Soil N tests (NO3-NþNH4-N) for upland soils are an important tool for
assessing soil N supply capacity (Wehrmann and Scharpf, 1979). However,
the original soil Nmin method that is based on a single soil N test before
planting and that disregards variations in the soil N cycle (N mineralization,
N immobilization, and losses) may result in underuse or overuse of N
fertilizer. Several versions of the presidedress soil nitrate-N test only estab-
lish the critical threshold above which there is a low probability of a yield
response to sidedress N application, but do not provide an accurate estimate
of the optimal N rate below a critical value (Magdoff et al., 1984; Meisinger
et al., 1992; Schmitt and Randall, 1994). Uncertainties in the prediction of
the seasonal crop N demand and soil N movement require the use of N
status indicators for fine-tuning of N rate and timing of N applications. This
requires monitoring soil N concentrations in the root zone at different
Nutrient Management in China 19

growth stages of crops to achieve the synchronization of crop N nutrient

uptake and N supply from wastes, indigenous soil resources, and environ-
mental sources. For optimum management of soil N supply in the root
zone, we have developed an in-season root-zone N management strategy
for intensive wheat–maize systems on the NCP (Chen et al., 2006; Cui et al.,
2008b,c; Zhao et al., 2006). According to this strategy, the total amount of
N fertilizer is divided into two or three applications over the course of the
growing season, with the optimum N fertilizer rate (ONR) for each applica-
tion being determined from soil nitrate-N tests in the root zone and a target
N value for the corresponding growth period of the crop (Fig. 7).
To confirm whether ONR using in-season root-zone N management
was the economically optimum N rate, we conducted on-farm experiments
(wheat, n¼121; maize, n¼148) with three N levels during 2003–2007 to
evaluate this N management in terms of agronomic performance and
environmental impact. Compared with farming practice (FNP), the in-
season root-zone N management strategy reduced N fertilizer by 60% and
40% for wheat and maize, and simultaneously increased grain yield by 4%

Soil nitrate-N testing in root-zone

Planting 3-leaf stage 10-leaf stage

N demand
N target value

Fertilizer+ N
from
environment,
N supply

etc.
Nutrient bioavailability
Planting
Nutrient spatial
availability
Nutrient temporal and spatial variability

Nitrogen target value for different grain yield for Chinese maize production

Target yield (t ha-1) 8.0–8.5 9.5–10.0 7.0–7.5 5.5–6.0 6.5–7.0

-1
Before sowing (kg N ha ) 20 30 20 20 25
Three leaf stage (kg N ha-1) 95 160 85 60 75
Ten leaf stage (kg N ha-1) 190 220 175 140 160

Figure 7 Model of in-season root-zone N management and N target value for differ-
ent grain yields for Chinese maize production. Date source: Cui et al., 2008b.
20 Fusuo Zhang et al.

and 5%, respectively. As expected, the ONR treatment had higher N use
efficiency and reduced apparent N loss compared with FNP (P<0.05).
Compared to FNP, apparent N losses were reduced by 73% from 159 to
43kgNha
1 and 43% from 151 to 86kgNha
1 in the wheat and maize
growing seasons, respectively (Fig. 8).
Clearly, in this study ONR using in-season root-zone N management
strategy was close to optimal, sustained high yields, and increased profits,
while minimizing the potential environmental impacts of N fertilization.
This superiority of in-season root-zone N management strategy can be
attributed to three major factors: (1) efficient utilization of all N sources
from the soil, environment, and fertilizer; (2) achievement of synchroniza-
tion between N supply and crop demand; and (3) addressing site-specific N
management. In China, small-scale farming has intensified variations in soil
N supply. In the research region, indigenous N supplies (average N uptake
in 0 N control) varied from 69 (site 14) to 191kgNha
1 (site 5), with a
coefficient of variation of 24%. The large disparity among experimental sites
reflects the net variation in soil N transformations (mineralization, immobi-
lization, denitrification, and volatilization), possible N transport (leaching
and runoff), and crop uptake. Clearly, one rate does not fit all circumstances.

Figure 8 Performance of INM for wheat and maize in the North China Plain. (A)
INM can reduce N fertilizer by 30% while increasing grain yield of 16% compared with
farmers’ practice in Binzhou, Shandong province. (B) Maize lodging for FNP in
Xiaoxian, Anhui Province. (C) Mean N fertilizer application rate, apparent N losses,
AEN and PFPN of INM compared with FNP. Modified from Cui et al. (2008b,c).
Nutrient Management in China 21

The in-season root-zone N management strategy can address variations due

to site-specific conditions using 3-tiered soil nitrate testing, as supported by
a close correlation between ONR and initial soil nitrate-N content. Our
results are consistent with earlier studies in demonstrating that yield
response to applied N is strongly affected by soil N supply or initial soil
nitrate-N content before planting (Binford et al., 1992). In addition, the
in-season root-zone N management strategy considers variation in yield
response to N application rate between fields, and increased fertilizer N rate
in 4% (wheat) or 18% (maize) of sites compared to farming practice. As a
result, higher crop yields are likely to be achieved through a combination of
increased N application in fields with low N fertilizer use and improved
NUE in fields where N fertilizer use is already high.

3.2. INM for paddy rice

Rice is a staple crop in China, accounting for 34% and 47% of the national
total cultivated area and cereal production in 2009, respectively (FAO,
2011). China was responsible for around 29% of global rice production
with 19% of the production area in 2009 (FAO, 2011). Intensification of
rice production over the past 50years has been achieved by using modern
high-yielding varieties and larger inputs of chemical fertilizers and weed and
pest control, and has led to higher yields. From 1961 to 2009, there was a
3.2-fold increase in the productivity of rice (from 2.0 to 6.6tha
1).
Unfortunately, since about 1990 the increase in rice grain production
was associated with a major decline in NUE, especially N, and with
widespread environmental damage. The REN for Chinese rice production
declined from 30–35% (Zhu, 1998) to 28.3% (Zhang et al., 2008a), which is
lower than the world values (40–60%).The REN can be lower in some rice
production areas, for example, the provincial average REN of rice was only
19.9% in Jiangsu (Li, 2000). Yoshida (1981) estimated AEN (the agronomy
N efficiency) to be 15–25kg rough rice per kg applied N in the tropics.
Cassman et al. (1996) reported that AEN was 15–18kgkg
1 N in the dry
season in agricultural fields in the Philippines. In China, AEN was 15–20kg
kg
1 N from 1958 to 1963 and declined to only 10.4kgkg
1 N in 2000s
(Zhang et al., 2008a).
The low NUE may be attributed to fertilizer overuse and high nutrient
loss resulting from inappropriate timing and methods of fertilizer applica-
tion. For example, the average fertilizer N application rate for rice of 150
kgNha
1 is higher than in most countries and up to 67% above the global
average, but current application rates of 150–250kgNha
1 are common
and can reach 300kgNha
1 in some regions (Peng et al., 2010; Roelcke
et al., 2004). Based on a national farm survey, Li et al. (2010) found that
fertilizer N rates for rice showed an increasing trend from 217kgNha
1 in
2000 to 231kgNha
1 in 2007.
22 Fusuo Zhang et al.

Fertilizer application is often not based on real-time nutrient require-

ments of the crop and/or site-specific knowledge of soil nutrient status. For
example, in rice production systems, most farmers apply N in two split
dressings (basal and top-dressing) within the first 10 days of the rice growing
season (Fan et al., 2007). This large amount of basal fertilizer N is prone to
loss over an extended period because the plants require time to develop
their root systems and a substantial demand for N. Overuse fertilizer with
high nutrient losses has directly or indirectly led to a series of environmental
problems such as groundwater pollution and eutrophication of surface
waters. It is therefore becoming a major challenge for rice production to
increase fertilizer N use efficiency by crops through improved rice nutrient
management (Cassman et al., 1996). This is a very important step toward
sustainable development of agricultural production and environmental
protection.
The INM for paddy rice has been developed and implemented through-
out China since 2002. The overall nutrient management strategy has focused
on optimizing fertilizer N application rates and timing in combination with
maintenance of P and K supply. The total fertilizer N application rate for a
rice field can be estimated by calculating the difference in N budget between
N requirement and N supply from the soil (Dobermann and Fairhurst,
2000a,b). On average, irrigated rice needs to absorb 17.5, 3.0, and 17.0kg
N, P, and K, respectively, to produce each ton of grain yield. The crop
requirement can be calculated from a yield target selected as 75–80% of the
climate yield potential. Soil N supply capacity is estimated using the yield of
nutrient omission plots including nutrient inputs from atmospheric deposi-
tion (rainfall and dust), irrigation, floodwater, and sediments (dissolved and
suspended nutrients) and biological N2-fixation. Based on soil nitrogen
supply capacity, the basal fertilizer N is recommended. Top-dressings are
applied through SPAD fine-tuning. Therefore, the split pattern is in accor-
dance with soil N supply capacity, crop growth stage, cropping season, the
variety used, and the crop establishment method. Critical levels of SPAD and
the corresponding fertilizer application rates have been developed for fertil-
izer N distribution at different growth stages in China. If the SPAD tests are
not available to farmers, a simplified distribution method for fertilizer N can
be used according to the nutrient uptake pattern by the rice crop at a specific
site with fixed distribution rates based on previous investigations on nutrient
management (e.g., 35–40% for basal fertilization, 20–25% for early tillering,
25–30% for panicle initiation, and 0–10% for heading).
Nearly 560 demonstration sites have been established in major rice-
producing areas since 2001 (Fig. 9). On average, rice yields using INM
have increased by 11% compared with FNP with fertilizer N use reduced by
24%. Across all sites, Heilongjiang province increased yields by 15% and
Zhejiang province saved fertilizer N by 35%. The AEN of INM increased
by 96% from 9 to 19kgkg
1 compared with FNP.
Nutrient Management in China 23

Figure 9 Performance of INM for rice in China. (A) Experimental bases (red dots) for
INM demonstration of rice in China; (B) INM reduced N fertilizer by 38% while
increasing grain yield of 11% compared with FNP; (C) rice yield, N rate and AEN for
INM and FNP for paddy rice in Chinese main rice production since 2001.

3.3. INM for vegetable systems

China has become the world’s largest center for vegetable production. The
vegetable industry has developed rapidly since the mid-1980s and is an
important land use type with a total area of 1.84million ha, accounting for
11.3% of the total cropping area. Production reached 62MT in 2009, 7.7
times that in 1980, respectively (China Statistical Yearbook, 2011). The
cropping area of greenhouse vegetables has increased 500 times from 7.00
103 ha in the 1980s to 3.35106 ha in 2008 to provide sufficient vegetables
in winter and accounts for 18.7% and 25%, respectively, of total vegetable
cropping area and production (Wang, 2011).
High nutrient demand and high variation in rooting depth (especially in
greenhouse vegetables with shallow root systems) are important character-
istics of vegetable crops. For example, rooting depth can range from 25cm
(onion) to >100cm (long season cabbage). Substantial quantities of fertilizer
are applied because of limited information or technology and strong incen-
tive to produce maximum yields and profits. Frequent furrow irrigation
24 Fusuo Zhang et al.

with flushing of water-soluble fertilizer is typical in vegetable production

using heavy basal applications of fertilizer and manure. Based on a literature
survey, mean N, P, and K application rates for open field vegetables (n¼
320) were 443kgN, 232kgP2O5, and 302kgK2Oha
1, respectively, com-
pared with 1817kgN, 1144kgP2O5, and 1465kgK2Oha
1 for greenhouse
vegetables (n¼2478). Conventional fertilizer inputs are more than two to
eight times crop nutrient uptake (Cong et al., 2011; Fan et al., 2010a; Ge,
2009; Jiao et al., 2010; Ma, 2010; Ye et al., 2008).
Excessive use of fertilizers is a typical practice to maintain relatively high
nutrient intensity in the root zone. It was observed that crop productivity
decreased distinctly over time (Du et al., 2006) and secondary salinization is
likely to be a major obstacle, particularly in greenhouses because of the
excessive use of N fertilizers (Cao et al., 2004; Chen et al., 2004; Shi et al.,
2008). Moreover, overirrigation at each event far exceeds the water-holding
capacity of the soil and leads to ready nutrient loss from the root zone. In the
case of N, NO3-N will accumulate in soils and crops, and NO3-N may
reduce crop quality if taken up in very large amounts. Ecological problems
comprise gaseous N losses to the atmosphere, leaching of NO3-N, eutrophi-
cation of water bodies and soil salinization, particularly when vegetable crops
are grown in greenhouses.
A failed example with chili pepper has indicated that the ONR (Fig. 10A)
cannot reach the same high yield as in FNP with a total volume irrigation
of 600mm applied over five occasions. High fertilizer rates for cucumber
can be greatly reduced when irrigation is optimized (Fig. 10B). High fertilizer
N inputs associated with high irrigation have resulted in the agronomic
problems, that is, low yields and low N use efficiency. The REN in Chinese
vegetable production was only 8–12% (Liang, 2011). On the other hand, in
32–71% of shallow groundwater samples in Beijing and Shandong from

Figure 10 Two examples of INM for vegetables in Shouguan, Shandong Province.

(A) Chili pepper yield reduced when reduced N rate from 1600 of FNP to 300kgha
1
per season of ONR without optimized irrigation water; (B) cucumber can maintain
yield when reducing N rate from 2800 of FNP to 470kg ha
1 of ONR and optimizing
irrigation water.
Nutrient Management in China 25

intensive greenhouse production areas, the nitrate concentrations exceeded

the drinking water standards (10.8mgL
1 NO3-N) (Song et al., 2009; Zhu
et al., 2005).
The INM method was developed for sustainable vegetable cropping
systems on the basis of understanding the pattern of crop nutrient demand
and the characteristics of the soil nutrient supply. Real-time N management
tools can be used to balance the N input and output in the root zone.
Nitrogen target values, which vary among different crops and different yield
levels, include the components N uptake, critical Nmin supply in the root
zone, and avoidable N loss. On the basis of N balance strategies, various N
sources such as fertilizer N, Nmin in the root zone before N recommenda-
tions, and nitrate from irrigation water can be taken into consideration. N
recommendations are therefore calculated taking account of the N target
value in the root zone and different N sources (Fig. 11).
N fertilization¼target N value
soil Nmin in the root-zone
(NO
3 -N
from irrigation).
The N target values for different cropping systems were determined
with crop N uptake curves based on the analysis of metadata or long-term
experiments (Ren et al., 2010). Soil Nmin in the root zone and the nitrate
content in irrigation water can be measured before N recommendations
are made. Compared with FNP, INM can reduce chemical N fertilizer

N rate (kg ha-1) 98 100 30 – 92 75

N uptake in growth stage (kg N ha-1)

Total N uptake (kg N ha-1)

300 Accumulation of N uptake 140

250 120
100
200
80
150
60
100
40
50 20
N target (kg N ha-1)

0 0
100
200
300
I II III

Figure 11 In-season root-zone N management of INM for winter–spring season

cucumber in greenhouse. (I, II, III) Mean seedling establishment period, the former
fruiting period and the later fruiting period, respectively. The target N supply in these
three periods is 150, 250, and 200kgNha
1, respectively.
26 Fusuo Zhang et al.

inputs by 72% on average, N losses by 54%, and N2O emissions by 38% of

in year-round greenhouse tomato cropping systems (He et al., 2006, 2007b,
2009; Ren et al., 2010; Tang et al., 2005). In greenhouse cucumber crop-
ping systems, 55% of chemical N fertilizer inputs and 40% of N losses on
average were reduced using the INM methods compared with FNP (Guo,
2007; Guo et al., 2008; Peng, 2006).
A P management strategy with a buildup and maintenance approach has
been established for major vegetable species in China. In this strategy, the
soil available P content in the root zone can be tested at intervals of 3–5
years. P fertilization was controlled as “buildup,” “maintenance,” or “draw-
down” strategies in response to “low,” “moderate,” and “high” soil avail-
able P levels, respectively, to maintain soil available P at an optimum level in
the long term. Considering the agronomic threshold to obtain high yields
and environmental thresholds, the optimum soil Olsen-P level has been
developed for several vegetable crops based on plant response to P fertiliza-
tion. For example, in open fields the optimum Olsen-P level in the root
zone for tomato production in North China was found to be 50mgPkg
1
(Zhang et al., 2007) and was 43.7mgPkg
1 for leafy vegetables in South
China (Zhang et al., 2007).
Fertigation techniques with INM, combining irrigation with fertiliza-
tion to achieve integrated effects, was also developed in our INM to control
nutrients, especially N and K concentrations in the root zone in order to
further improve NUE and water use efficiency. Since 2005, fertigation
techniques have been developed in different vegetable systems and demon-
strated in different Chinese regions. On average (n¼46), fertigation can
increase vegetable yields by 5–20% while reducing inputs of chemical
fertilizers and water by 20–60% and 20–50%, respectively. Fertigation can
also reduce the environmental risk. For example, N losses declined by over
40% in greenhouse experiments on tomato compared with farming practice
(Gao et al., 2009; Zhang, 2010). Catch crops such as sweet corn during
summer fallow periods were considered to reduce N and P losses and retain
soil N deep soil profile in our INM. In greenhouse cucumber and tomato
systems, the accumulated Nmin in the soil (0–180cm) was reduced by
303–343 and 154–403kgNha
1 after introducing sweet corn as a catch
crop during the summer fallow period (Guo, 2007).

3.4. INM for orchards

Fruit production in China has developed rapidly since 1992 and includes
the most important cash crops with the highest production and acreage of
fruit trees in the world, with an area of 11.1million ha, accounting for
7.02% of the total area of crops, and output reached 203.95MT in China in
2009 (FAO, 2011). Among deciduous fruit crops, apples are far ahead in
acreage and production and pears, peach, and grapes are also leading crops.
Nutrient Management in China 27

Apricot, plum, walnut, chestnut, and kiwifruit are next in importance.

Citrus is the commonest of the tropical and subtropical fruit species, and
banana, pineapple, and lychee are also major tropical fruits in China.
Although arable land may be scarce, labor is available for labor-intensive
fruit production. Subsequently, China has substantially raised its profile in
the global fruit market and increased farmer’s income.
Much higher incomes can be obtained from orchard than cereal crops
in mountain and hilly areas. In many regions, the fertile lowlands have
switched from cereals to fruit and form the major fruit production areas.
Some soils may be too shallow for good root development. However, deep
tillage has been used to improve soil conditions in hilly areas for establish-
ment of orchards. Fruit and vegetable production relies on high inputs of
manures and fertilizers and this can adversely affect the safety and quality of
the produce. Nutrient inputs for orchards are often higher than the nutrient
recommendations in the effort to obtain higher yields and net returns ( Ju
et al., 2006; Liu et al., 2006a).
Compared to other organic fertilizers, animal manures are most exten-
sively used in orchards in conventional practice. Based on a survey of 916
orchards in North China, excessive N and P applications were common,
with average input rates of 588.4kgN and 156.7kgPha
1, respectively,
which were 2.5–3.0-fold higher than the fruit N and P requirements (Lu
et al., 2008). However, nutrient deficiency was observed in some orchards,
especially in hilly areas. For example, in Shaanxi province about 20–80% of
orchards received no fertilizers and about 80% of orchards received no
manures (Liu et al., 2002). Most farmers are accustomed to applying basal
fertilizers once a year after autumn leaf fall or before the spring bud burst and
additional fertilizer dressings are applied at the beginning of the phenologi-
cal period (Chai et al., 2009; Zhou, 2009).
Misuse of fertilizers is associated with widespread problems of fruit
production and environments. Unstable yields and poor fruit quality includ-
ing serious physiological diseases of fruit trees (i.e., bitter pit, shrinking fruit,
yellow leaves, internal bark necrosis) occur widely due to calcium and boron
deficiencies (Shorrocks and Nicholson, 1980), iron deficiency, and manga-
nese toxicity (Berg, 1973; Gao et al., 2006; Xu et al., 2008) with serious effects
on fruit exports and marketable values (Ferguson and Watkins, 1989; Wang
et al., 2001, 2010b; Zhao et al., 2007).
High proportions of surplus N and P and low NUE have been found
in vineyards and in apple, pear, and peach orchards in the plains regions
with high economic returns. N surplus reduces soil C/N ratio on the plains,
plateaus, and mountain regions. Controlling N and P fertilization is the key
to maintaining sustainable fruit production in North China (Lu et al., 2008,
Ren, 2007). Soil acidification is a growing problem in most orchards. For
example, in the Bohai Gulf area, high soil pH (>6.5) accounts for only 7.5%
of 1338 orchards in Jiaodong peninsula (Li et al., 2009) and the pH value has
28 Fusuo Zhang et al.

declined from 6.8–7.6 in 1982 to 4.0–5.3 in 2008 in Jiaodong peninsula

(Wang et al., 2010a).
Nutrient management is an important component in orchard man-
agement for high efficiency and high fruit quality. Fruit trees are
perennial woody plants which can store nutrients and then release
them for growth. Therefore, the nutrient uptake of fruit trees includes
the total nutrients of the fruits, leaves, new branches, and time incre-
ments of storage. Nutrient uptake is different under different yield
levels. For example, the N uptake of apple trees was 100–120, 110–
130,120–140, and 130–150kgha
1 under yield levels of 30, 45, 60, and
75tha
1, respectively. Nitrogen fertilizer strategies have been studied in
detail by Zhang et al. (2009).
Basal fertilization is considered to be important for high-quality fruit
production by providing nutrients for tree growth and also improving
soil structure which promotes better root development. It is usually
recommended and immediately applied during the dormant season up to
autumn, after fruit harvest. This application coincides with the third peak
of annual root growth. The fertilizers not only help trees to recover from
the nutrient depletion caused by a heavy fruit load but also enhance
development of quality fruit buds. In our studies, 60% of N fertilizer was
applied in autumn after fruit picking to add N storage and restore tree
vigor. Top dressing applications are usually chemical fertilizers which are
applied two to four times before bud break or prior to or after blooming
and at the fruit enlargement stage, depending on the requirements of
each type of fruit tree. For apple trees, 20% or so of N fertilizer is applied
from the period of blooming to speed growth of new branches, a further
20% of N applied to speed growth of new branches until before fruit
picking, and about 60% of N fertilizer is applied from fruit picking to leaf
fall ( Jiang et al., 2007). However, over 50% of N fertilizer is commonly
applied after flower bud breaking or fruit forming in farming practice.
This might cause the phenomenon of alternate bearing due to inadequate
storage of nutrients.
Using this INM method, the total nutrient demand is calculated accord-
ing to target fruit yield level and NPK fertilizers are recommended based on
soil fertility index (High, Moderate, and Low soil fertility) and nutrient
demand with different stages. For moderate apply trees, most of N fertilizer
is applied in autumn after picking to add N storage and renew tree vigor,
and most of the P fertilizer is applied in spring with flowing stage to improve
flower bud quality, and most of the K fertilizer is applied in summer fruiting
period to improve fruit quality. Management of high-yielding fruits using
suitable pruning, and irrigation, are combined in this INM.
Since 2005, INM has been developed in different orchard and vineyard
systems, such as apple, peach, jujube, grape, cherry, mandarin orange,
banana, lychee, and citrus, and demonstrated for different regions of the
Nutrient Management in China 29

country (total demonstration area of 7460ha). On average, compared to

farming practice, INM increased fruit yields by 11.1–18.3% while reducing
chemical fertilizers (N, P2O5, and K2O) by 23.1–36.2%, 10.4–17.4%, and
10.4–16.6%, respectively.

4. Large-Scale Dissemination of INM

From 2003 to 2010, a large number of on-farm demonstration trials
(n¼5147) were conducted with 158 experimental bases by >30 research
units including universities and academies of agricultural sciences. Across
all 5147 sites, on average, reduce N fertilizer inputs by 24%, increase yield
by 12%, and increased net farming income by $132 per ha (Fig. 12). Four
factors contributed to this improvement: (1) elite varieties capable of
producing well at high planting densities and also with high yield poten-
tial; (2) integrated nutrient and water management, especially N manage-
ment; (3) better crop management including plowing, sowing, density,
and pest management; and (4) improved soil quality, largely due to more
widespread adoption of the practice of returning straw to soils as opposed
to burning and the growth of conservation tillage. Improved N manage-
ment includes significant reduction in fertilizer N application rate by
efficient utilization of all N sources in the soil, N entering soils from the
environment, and fertilizers; greater use of split fertilizer N applications to
match crop demands; and addressing field-to-field variation by predicting

Figure 12 Performance of INM in China. (A) 158 experimental bases (red dots) for
INM dissemination in China; (B) increased yield, reduced N fertilizer rate, and
increased net income of INM for different crops including wheat, maize, rice, vegeta-
ble, fruit, rape, and cotton, compared farmers’ practice. D yield, N fertilizer, and net
income mean the different yields, N fertilizer rate, and net income between INM and
farmer’s practice.
30 Fusuo Zhang et al.

plant-available soil N through soil mineral N or soil nitrate-N testing

during the crop growing season (Cui et al., 2008b,c). Poor crop manage-
ment by farmers may lead to lower exploitation of yield potential in their
fields than in the regional variety test experiments. For example, our study
in southwest China showed that combining a triangular transplanting
pattern with split N fertilizer applications increased rice yields by 20%,
saved fertilizer N inputs by 18%, and reduced N losses by 44% compared
with traditional farming practice (Fan et al., 2009).
The Chinese government regards agriculture as the primary field of
development of the national economy in the twenty-first century. Our
INM, as a major agricultural development technology, has been advocated
in Chinese policy and national actions. For example, since 2005 INM has
been practiced as most important nutrient management strategy for soil
testing and fertilizer recommendation project, covering all agricultural
counties with total funding of nearly 6 billion Yuan until 2010 ($923
million). As a result, fertilizer application rates in the project area have
decreased and NUE has also improved. As shown in Fig. 13, fertilizer
consumption has decreased steadily since 2004, coupled with a trend of
increasing partial factor productivity for chemical fertilizers from 17kgkg
1
in 2004 to about 21kgkg
1 in 2008.
The need for increased global food production while also protecting
environmental quality and conserving natural resources raises the issue

Figure 13 Trends in fertilizer consumption for cereal crop and partial factor produc-
tivity for fertilizer (PFP) in China from 1981 to 2008. The PFP was defined the ratio of
crop yield per unit of applied chemical fertilizer. Modified from Zhang et al. (2011).
Nutrient Management in China 31

on how to achieve high yields without excessive nutrient losses to the

environment worldwide (Cassman, 1999; Matson et al., 1997; Tilman,
1999). This challenge is particularly daunting in rapidly developing
countries such as China, India, Brazil, Mexico, Indonesia, Vietnam, Paki-
stan, and Sri Lanka. First, rapidly growing populations need further increases
in grain production while rates of gain in cereal yields have slowed markedly
in the past 10–20years (FAO, 2011), even though agricultural inputs such as
N and P have continued to increase. Second, rapidly developing countries
account for 89% of the total global increase in N fertilizer use (25MT)
since 2000, especially in China (12MT) and India (4MT). By 2050, 59% of
all fertilizer N will be applied in developing regions (Beman et al., 2005).
Environmental pollution is becoming an issue of serious concern due to
excessive use of fertilizer N in crop production (Guo et al., 2010; Le et al.,
2010; Zheng et al., 2004). Third, some developed nutrient management
strategies using sophisticated decision-support tools for large-scale enter-
prises are not available for hundreds of millions of holder farmers on small
parcels of land such as site-specific management based on GPS, GIS, and
remote sensing (Chen et al., 2011; Raun et al., 2002). The challenge in
rapidly developing economies is to increase global food production while
also protecting environmental quality and conserving natural resources
as in Chinese agricultural production. The success of INM in China as
indicated by our INM in this study shows that advanced methodologies
can be employed, to ensure food security and protect environmental
quality in rapidly developing countries.

5. Summary and Conclusions

Demonstration in different cropping systems and large-scale dissem-
ination of INM has indicated the potential of significant increase in crop
yields and NUE. However, there is still a long way to go to realize the aims
of the “Four WINs,” namely, increasing crop yield and NUE while
simultaneously improving soil productivity and environmental quality.
Achieving these goals will require continued and expanded efforts nation-
wide to develop new technologies by integration of different disciplines
such as plant breeding, agronomy, soil science, plant nutrition, plant
protection, and agricultural engineering, and to extend these technologies
to millions of small-holder farmers. Fortunately, the Chinese Government
is aware of the importance of INM in the sustainable development of
Chinese agriculture and supports the adoption of these technologies on a
national scale. INM can also be used as part of the global strategy to ensure
food security and protect the environment now that there are over 7
billion human beings on the planet.
32 Fusuo Zhang et al.

ACKNOWLEDGMENTS
We thank the National Basic Research Program of China (973 Program: 2009CB118606),
the Special Fund for the Agriculture Profession (201103003). This research was supported
by the National Natural Science Foundation of China (30890130) and the Innovative
Group Grant of the National Science Foundation of China (31121062) for financial
support. We will give special thanks to Prof. P. Christie in Queen’s University Belfast,
UK, and Prof. C. Tang in La Trobe University, Bundoora, Australia, for their comments
and linguistic revisions.

REFERENCES
Alloway, B. J. (2008). Micronutrients and crop production: An introduction.
In “Micronutrient Deficiencies in Global Crop Production” (B. J. Alloway, et al., Eds.),
pp. 1–39. Springer, The Netherlands.
Bates, T. R., and Lynch, J. P. (1996). Stimulation of root hair elongation in Arabidopsis
thaliana by low phosphorus availability. Plant Cell Environ. 19, 529–538.
Beman, J. M., Arrigo, K. R., and Matson, P. A. (2005). Agricultural runoff fuels large
phytoplankton blooms in vulnerable areas of the ocean. Nature 434, 211–214.
Berg, A. (1973). The relation of manganese to internal bark necrosis of apple. Science 74,
485–490.
Binford, G. D., Blackmer, A. M., and Cerrato, M. E. (1992). Relationships between corn
yields and soil nitrate in late spring. Agron. J. 84, 53–59.
Blankenau, K., Olfs, H. W., and Kuhlmann, H. (2000). Effect of microbial nitrogen
immobilization during the growth period on the availability of nitrogen fertilizer for
winter cereals. Biol. Fertil. Soils 32, 157–165.
Campbell, C. A., Myers, R. J. K., and Curtin, D. (1995). Managing nitrogen for sustainable
crop production. Fertil. Res. 42, 277–296.
Cao, Z. H., Huang, J. F., Zhang, C. S., and Li, A. F. (2004). Soil quality evolution after land
use change from paddy soil to vegetable land. Environ. Geochem. Health 26, 97–103.
Cassman, K. G. (1999). Ecological intensification of cereal production systems:
Yield potential, soil quality, and precision agriculture. Proc. Natl. Acad. Sci. USA 96,
5952–5959.
Cassman, K. G., Gines, G. C., Dizon, M. A., Samson, M. I., and Alcantara, J. M. (1996).
Nitrogen-use efficiency in tropical lowland rice systems: Contributions from indigenous
and applied nitrogen. Field Crop Res. 47, 1–12.
Cassman, K. G., Dobermann, A., and Walters, D. T. (2002). Agroecosystems, nitrogen use
efficiency, and nitrogen management. Ambio 31, 132–140.
Chai, Z. P., Jiang, P. A., and Wang, X. M. (2009). The investigation research of fertilization
about several main characteristic fruit trees in Xinjiang. Chin. Agric. Sci. Bull. 24, 231–
234 (in Chinese with English abstract).
Chen, X. P. (2003). Optimization of the N Fertilizer Management of a Winter Wheat/
Summer Maize Rotation System in the North China Plain.PhD Thesis, University of
Hohenheim, Stuttgart, Germany.
Chen, Q., Zhang, X., Zhang, H., Christie, P., Li, X., Horlacher, D., and Liebig, H. (2004).
Evaluation of current fertilizer practice and soil fertilizer in vegetable production in the
Beijing region. Nutr. Cycl. Agroecosyst. 69, 51–58.
Chen, X. P., Zhang, F. S., Römheld, V., Horlacher, D., Schulz, R., Böning-Zilkens, M.,
Wang, P., and Claupein, W. (2006). Synchronizing N supply from soil and fertilizer
and N demand of winter wheat by an improved Nmin method. Nutr. Cycl. Agroecosyst. 74,
91–98.
Nutrient Management in China 33

Chen, X. P., Zhang, F. S., Cui, Z. L., Li, F., and Li, J. L. (2010). Optimizing soil nitrogen
supply in the root zone to improve maize nitrogen management. Soil Sci. Soc. Am. J. 74,
1367–1373.
Chen, X. P., Cui, Z. L., Vitousek, P. M., Cassman, K. G., Matson, P. A., Bai, J. S.,
Meng, Q. F., Hou, P., Yue, S. C., Römheld, V., and Zhang, F. S. (2011). Integrated
soil-crop system management for food security. Proc. Natl. Acad. Sci. USA 108,
6399–6404.
China Agricultural Yearbook. (2011). Editorial committee of China Agricultural Yearbook,
Chinese Statistical Publishing House, Beijing.(in Chinese).
Chinese Ministry of Environmental Protection (2010). http://www.gov.cn/jrzg/2010-02/
10/content_1532174.htm..
Cong, L. M., Jiao, X. Y., Wang, L. G., Wang, J. S., and Dong, E. W. (2011). Present
situation of fertilization in solar-greenhouse in Quwo county and suggestions. J. Shanxi
Agric. Sci. 39(4), 342–345. (in Chinese with English abstract).
Cui, Z. L., Chen, X. P., Miao, Y. X., Li, F., Zhang, F. S., Li, J. L., Ye, Y. L., Yang, Z. P.,
Zhang, Q., and Liu, C. S. (2008a). On-farm evaluation of winter wheat yield response to
residual soil nitrate-N in North China Plain. Agron. J. 100, 1527–1534.
Cui, Z. L., Chen, X. P., Miao, Y. X., Zhang, F. S., Sun, Q. P., and Schroder, J. (2008b).
On-farm evaluation of the improved soil Nmin-based nitrogen management for summer
maize in North China Plain. Agron. J. 100, 517–525.
Cui, Z. L., Zhang, F. S., Chen, X. P., Miao, Y. X., Li, J. L., Shi, L. W., Xu, J. F., Ye, Y. L.,
Liu, C. S., Yang, Z. P., Zhang, Q., Huang, S. M., et al. (2008c). On-farm evaluation of an
in-season nitrogen management strategy based on soil Nmin test. Field Crop Res. 105, 48–55.
Cui, Z. L., Zhang, F. S., Miao, Y. X., Sun, Q. P., Li, F., Chen, X. P., Li, J. L., Ye, Y. L.,
Yang, Z. P., Zhang, Q., and Liu, C. S. (2008d). Soil nitrate-N levels required for high
yield maize production in the North China Plain. Nutr. Cycl. Agroecosyst. 82, 187–196.
Cui, Z. L., Zhanf, F. S., Mi, G. H., Chen, F. J., Li, F., Chen, X. P., Li, J. L., and Shi, L. F.
(2009). Interaction between genotypic difference and nitrogen management strategy in
determining nitrogen use efficiency of summer maize. Plant Soil 317, 267–276.
Cui, Z. L., Chen, X. P., and Zhang, F. S. (2010a). Current nitrogen management status and
measures to improve the intensive wheat–maize system in China. Ambio 39, 376–384.
Cui, Z. L., Zhang, F. S., Chen, X. P., Dou, Z. X., and Li, J. L. (2010b). In-season nitrogen
management strategy for winter wheat: Maximizing yields, minimizing environmental
impact in an over-fertilization context. Field Crop Res. 116, 140–146.
Cui, Z. L., Zhang, F. S., Chen, X. P., Li, F., and Tong, Y. P. (2011). Using in-season
nitrogen management and wheat cultivars to improve nitrogen use efficiency. Soil Sci.
Soc. Am. J. 75, 967–983.
Dinnes, D. L., Karlen, D. L., Jaynes, D. B., Kaspar, T. C., Hatfield, J. L., Colvin, T. S., and
Cambardella, C. A. (2002). Nitrogen management strategies to reduce nitrate leaching
in tile-drained midwestern soils. Agron. J. 94, 153–171.
Dobermann, A. (2007). Nutrient use efficiency-measurement and management. In “IFA
International Workshop on Fertilizer Best Management Practices,” International Fertil-
izer Industry Association, Paris, 7–9 March 2007, Brussels, Belgium.
Dobermann, A., and Cassman, K. G. (2005). Cereal area and nitrogen use efficiency are
drivers of future nitrogen fertilizer consumption. Sci. China C Life Sci. 48, 745–758.
Dobermann, A., and Fairhurst, T. (2000). Rice: Nutrient Disorders & Nutrient Manage-
ment. PPI, PPIC and IRRI, Singapore.
Drew, M. C. (1975). Comparison of the effects of a localized supply of phosphate, nitrate,
ammonium and potassium on the growth of the seminal root system, and the shoot, in
barley. New Phytol. 75, 479–490.
Drinkwater, L. E., and Snapp, S. S. (2007). Nutrients in agroecosystems: Rethinking the
management paradigm. Adv. Agron. 92, 163–186.
34 Fusuo Zhang et al.

Du, L. F., Zhang, W. L., and Li, Z. H. (2006). Soil quality with various planting patterns
in Yangtze River Delta Area. J. Agro Environ. Sci. 25, 95–99. (in Chinese with English
abstract).
Editorial Committee of China Agriculture Yearbook (ECCAY). (2006). China Agriculture
Yearbook. China Agriculture Press, Beijing.
Engels, T., and Kuhlmann, H. (1993). Effect of the rate of N fertilizer on apparent net
mineralisation of N during and after cultivation of cereal and sugar beet crops.
Z. Pflanzenernähr. Bodenkd. 156, 149–154.
Fahey, T. J., Williams, C. J., and Rooney-Varga, J. N. (1999). Nitrogen deposition in and
around an intensive agricultural district in central New York. J. Environ. Qual. 28,
1585–1600.
Fan, M., Lu, S., Jiang, R., Liu, X., Zeng, X., Goulding, K., and Zhang, F. (2007). Nitrogen
input, 15 N balance and mineral N dynamics in a rice-wheat rotation in southwest China.
Nutr. Cycl. Agroecosyst. 79, 255–265.
Fan, M. S., Cui, Z. L., Chen, X. P., Jiang, R. F., and Zhang, F. S. (2008). Integrated nutrient
management for improving crop yields and nutrient utilization efficiencies in China.
J. Soil Water Conserv. 63, 126–128.
Fan, M. S., Lu, S. H., Jiang, R. F., Liu, X. J., and Zhang, F. S. (2009). Triangular
transplanting pattern and split nitrogen fertilizer application increase rice yield and
nitrogen fertilizer recovery. Agron. J. 101, 1421–1425.
Fan, H. Z., Chen, Y. X., Lv, S. H., and Feng, W. Q. (2010a). Current status and evaluation
of nutrient management of turnip in Sichuan province. Soils 42(5), 828–832. (in Chinese
with English abstract).
Fan, M. S., Christie, P., Zhang, W., and Zhang, F. S. (2010b). Crop production, fertilizer
use and soil quality in China. In “Advances in Soil Science Food Security and Soil
Quality” (R. Lal, B. A. Stewart, et al., Eds.), pp. 87–108. CRC Press, Taylor & Francis
Group, Boca Raton, FL.
FAO. (2011). FAOSTAT Database—Agriculture Production. Food and Agriculture Orga-
nization of the United Nations, Rome.
Ferguson, I. B., and Watkins, C. B. (1989). Bitter pit in apple fruit. Hortic. Rev. 11, 289–355.
Galloway, J. N., and Cowling, E. B. (2002). Reactive nitrogen and the world: 200 years of
change. Ambio 31, 64–71.
Gao, Y. M., Xu, J., and Gao, S. Q. (2006). Effects of silicon application on apple internal
bark necrosis induced by high content of manganese. Plant Nutr. Fertil. Sci. 12(4), 571–
577. (in Chinese with English abstract).
Gao, B., Li, J. L., Chen, Q., Liu, Q. H., and Wang, J. (2009). Study on nitrogen
management and irrigation methods of greenhouse tomato. Sci. Agric. Sin. 42(6),
2034–2042. (in Chinese with English abstract).
Gao, B., Ju, X. T., Zhang, Q., Christie, P., and Zhang, F. S. (2011). New estimates of direct
N2O emissions from Chinese croplands from 1980 to 2007 using localized emission
factors. Biogeosci. Discuss. 8, 6971–7006.
Ge, X. Y. (2009). Assessment of NPK Fertilizer Consumption and Demand in Vegetable
System in China. Master Thesis, China Agricultural University, Beijing, China (in
Chinese with English abstract).
Goulding, K. W. T., Bailey, N. J., Bradbury, N. J., Hargreaves, P., Howe, M.,
Murphy, D. V., Poulton, P. R., and Willison, T. (1998). Nitrogen deposition and its
contribution to nitrogen cycling and associated soil processes. New Phytol. 139, 49–58.
Granato, T. C., and Raper, C. D. (1989). Proliferation of maize (Zea mays L.) roots in
response to localized supply of nitrate. J. Exp. Bot. 40, 263–275.
Guo, R. Y. (2007). Studies on Nitrogen Control in Root-Zone and Summer Catch Crop
Planting for Reducing N Loss in Greenhouse Cucumber Cropping System. PhD Thesis,
China Agricultural University, Beijing, China. (in Chinese with English abstract).
Nutrient Management in China 35

Guo, Y. F., Mi, G. H., Chen, F. J., and Zhang, F. (2005). Effect of NO
3 supply on lateral
root growth in maize plants. J. Plant Physiol. Mol. Biol. 31, 90–96. (in Chinese with
English abstract).
Guo, R. Y., Li, X. L., Christie, P., Chen, Q., and Zhang, F. S. (2008). Seasonal temperatures
have more influence than nitrogen fertilizer rates on cucumber yield and nitrogen uptake
in a double cropping system. Environ. Pollut. 151, 433–451.
Guo, J. H., Liu, X. J., Zhang, Y., Shen, J. L., Han, W. X., Zhang, W. F., Christie, P.,
Goulding, K. W. T., Vitousek, P. M., and Zhang, F. S. (2010). Significant acidification
in major Chinese croplands. Science 327, 1008–1010.
He, F. F., Xiao, W. L., Li, J. L., Chen, Q., Jiang, R. F., and Zhang, F. S. (2006). Integrated
nitrogen management in greenhouse tomato production. Plant Nutr. Fertil. Sci. 12(3),
394–399. (in Chinese with English abstract).
He, C. E., Liu, X. J., Fangmeier, A., and Zhang, F. S. (2007a). Quantifying the total airborne
nitrogen input into agroecosystems in the North China Plain. Agric. Ecosyst. Environ. 121,
395–400.
He, F. F., Chen, Q., Jiang, R. F., Chen, X. P., and Zhang, F. S. (2007b). Yield and nitrogen
balance of greenhouse tomato (Lycopersicum esculentum Mill.) with conventional and
site-specific management practice in Northern China. Nutr. Cycl. Agroecosyst. 77, 1–14.
He, F. F., Jiang, R. F., Chen, Q., Zhang, F. S., and Su, F. (2009). Nitrous oxide emissions
from an intensively managed greenhouse vegetable cropping system in Northern China.
Environ. Pollut. 157, 1666–1672.
He, C. E., Wang, X., Liu, X. J., Fangmeier, A., Christie, P., and Zhang, F. S. (2010).
Nitrogen deposition and its contribution to nutrient inputs to intensively managed
agricultural ecosystems. Ecol. Appl. 20, 80–90.
Hodge, A. (2004). The plastic plant: Root responses to heterogeneous supplies of nutrients.
New Phytol. 162, 9–24.
Hofman, G. (1999). Nutrient management legislation in European countries. NUMALEC
Report. Concerted action, Fair6-CT98-4215.
Huang, Y., and Sun, W. J. (2006). Changes in topsoil organic carbon of croplands in
mainland China over the last two decades. Chin. Sci. Bull. 51, 1785–1803.
IFA (2011). IFADATA—International Fertilizer Industry Association..
Jiang, Y. M., Zhang, H. Y., and Zhang, F. S. (2007). The Theory and Practice of Integrated
Nutrient Management for Fruit Trees in North China. China Agricultural University Press,
Beijing. (in Chinese).
Jiao, X. Y., Wang, L. G., Zhang, D. L., Zhang, J. S., and Dong, E. W. (2010). Present
situation of fertilizer application, its problems and suggestions concerning vegetable
production under conditions of solar-greenhouse. J. Shanxi Agric. Sci. 38(4), 37–41. (in
Chinese with English abstract).
Jing, J. Y., Rui, Y. K., Zhang, F. S., Rengel, Z., and Shen, J. B. (2010). Localized application
of phosphorus and ammonium improves growth of maize seedlings by stimulating root
proliferation and rhizosphere acidification. Field Crop Res. 119, 355–364.
Ju, X. T., Liu, X. J., Zou, G. Y., Wang, Z. H., and Zhang, F. S. (2002). Evaluation of
pathway of nitrogen loss in winter wheat and summer maize rotation system. Agric. Sci.
Chin. 1, 1224–12316. (in Chinese with English abstract).
Ju, X. T., Liu, X. J., Zhang, F. S., and Roelcke, M. (2004). Nitrogen fertilization, soil nitrate
accumulation, and policy recommendations in several agricultural regions of China.
Ambio 33, 300–305.
Ju, X. T., Kou, C. L., Zhang, F. S., and Christie, P. (2006). Nitrogen balance and
groundwater nitrate contamination: Comparison among three intensive cropping sys-
tems on the North China Plain. Environ. Pollut. 143, 117–125.
Ju, X., Xing, G., Chen, X., Zhang, S., Zhang, L., Liu, X., Cui, Z., Yin, B., Christie, P.,
Zhu, Z., and Zhang, F. (2009). Reducing environmental risk by improving N manage-
ment in intensive Chinese agricultural systems. Proc. Natl. Acad. Sci. USA 106, 3041–3046.
36 Fusuo Zhang et al.

Ju, X., Lu, X., Gao, Z., Chen, X., Su, F., Kogge, M., Römheld, V., Christie, P., and
Zhang, F. (2011). Processes and factors controlling N2O production in an intensively
managed low carbon calcareous soil under sub-humid monsoon conditions. Environ.
Pollut. 159, 1007–1016.
Le, C., Zha, Y., Li, Y., Sun, D., Lu, H., and Yin, B. (2010). Eutrophication of lake waters in
China: Cost, causes, and control. Environ. Manage. 45, 662–668.
Li, R. (2000). Efficiency and Regulation of Fertilizer Nitrogen in High-Yield Farmland: A
Case Study on Rice and Wheat Double Maturing System Agriculture Area of Tai Lake
for Deducing to Jiangsu Province. PhD Thesis, China Agricultural University, Beijing,
China. (in Chinese with English abstract).
Li, G. T., Li, B. G., and Chen, D. L. (2002). Method for measurement of ammonia
volatilization from large area field by Bowen Ratio System. J. Chin. Agric. Univ. 6,
56–62. (in Chinese with English abstract).
Li, L., Li, S. M., Sun, J. H., Zhou, L. L., Bao, X. G., Zhang, H. G., and Zhang, F. S. (2007).
Diversity enhances agricultural productivity via rhizosphere phosphorus facilitation on
phosphorus-deficient soils. Proc. Natl. Acad. Sci. USA 104, 11192–11196.
Li, Z. D., Wang, Y. H., and Li, L. Y. (2009). Orchard soil acidification status in Jiaodong
peninsula of Shandong province and control techniques. Phosphate Compound Fertil. 24
(6), 80–81. (in Chinese with English abstract).
Li, H., Zhang, W., Zhang, F., Du, F., and Li, L. (2010). Chemical fertilizer use and
efficiency change of main grain crops in China. Plant Nutr. Fertil. Sci. 16(5),
1136–1143. (in Chinese with English abstract).
Li, H., Huang, G., Meng, Q., Ma, L., Yuan, L., Wang, F., Zhang, W., Cui, Z., Shen, J.,
Chen, X., Jiang, R., and Zhang, F. (2011). Integrated soil and plant phosphorus
management for crop and environment in China. A review. Plant Soil 349, 157–167.
10.1007/s11104-011-0909-5.
Liang, J. (2011). Study on the Situation of Nitrogen Input and Output and the Potential of
Fertilizer Saving in Vegetable System in China. Master Thesis, China Agricultural
University, Beijing, China. (in Chinese with English abstract).
Linkohr, B. I., Williamson, L. C., Fitter, A. H., and Leyser, H. M. O. (2002). Nitrate and
phosphate availability and distribution have different effects on root system architecture
of Arabidopsis. Plant J. 29, 751–760.
Liu, X. J., and Zhang, F. S. (2009). Nutrient from environment and its effects in nutrient
management of ecosystems—A case study on atmospheric nitrogen deposition. Arid Zone
Res. 3, 306–311. (in Chinese with English abstract).
Liu, H. J., Ju, X. T., and Tong, Y. A. (2002). The status and problems of fertilization of main
fruit trees in Shaanxi Province. Agric. Res. Arid Areas 20(1), 38–44. (in Chinese with
English abstract).
Liu, X., Ju, X., Zhang, F., Pan, J., and Christie, P. (2003). Nitrogen dynamics and budgets in
a winter wheat–maize cropping system in the North China Plain. Field Crop Res. 83,
111–124.
Liu, H. B., Li, Z. H., and Zhang, Y. G. (2004a). Characteristics of nitrate distribution and
accumulation in soil profiles under main agro-land use types in Beijing. Sci. Agric. Sin. 37,
692–698. (in Chinese with English abstract).
Liu, Y., Mi, G., Chen, F., Zhang, J., and Zhang, F. (2004b). Rhizosphere effect and root
growth of two maize (Zea mays L.) genotypes with contrasting P efficiency at low P
availability. Plant Sci. 167, 217–223.
Liu, J. L., Liao, W. H., Zhang, Z. H., and Sun, J. S. (2006a). The change of soil nutrition and
the status of distribution in the apple orchard in the south and central part of Hebei
province. Acta Hortic. Sin. 33, 705–708. (in Chinese with English abstract).
Liu, X. J., Ju, X. T., Zhang, Y., Kopsch, J., and Zhang, F. S. (2006b). Nitrogen deposition in
agroecosystems in the Beijing area. Agric. Ecosyst. Environ. 113, 370–377.
Nutrient Management in China 37

Liu, X. J., Song, L., He, C. E., and Zhang, F. S. (2010). Nitrogen deposition as an important
nutrient from the environment and its impact on ecosystems in China. J. Arid Land 2,
137–143. (in Chinese with English abstract).
Liu, X. J., Duan, L., Mo, J. M., Du, E. Z., Shen, J. L., Lu, X. K., Zhang, Y., Zhou, X. B.,
He, C. E., and Zhang, F. S. (2011). Nitrogen deposition and its ecological impact in
China: An overview. Environ. Pollut. 159, 2251–2264.
Lu, S. C., Chen, Q., and Zhang, F. S. (2008). Analysis of nitrogen input and soil nitrogen
load in orchards of Hebei province. Plant Nutr. Fertil. Sci. 14(5), 858–865. (in Chinese
with English abstract).
Lynch, J. P., and Brown, K. M. (2008). Root strategies for phosphorus acquisition. In “The
Ecophysiology of Plant-Phosphorus Interactions” (P. J. White and J. P. Hammond,
Eds.), pp. 83–116. Springer, Dordrecht, The Netherlands.
Ma, W. J. (2010). The Growth, Nutrient Uptake and Accumulation in Different Cash
Crops. PhD Thesis, Northwest A & F University, Yangling, Shaanxi, China (in Chinese
with English abstract)..
Ma, W. Q., Mao, D. R., and Zhang, F. S. (1999). Current status and evaluation of crop
fertilization in Shandong province. Chin. J. Soil Sci. 30, 217–220. (in Chinese with
English abstract).
Magdoff, F. R., Ross, D., and Amadon, J. (1984). A soil test for nitrogen availability to corn.
Soil Sci. Soc. Am. J. 48, 1301–1304.
Marschner, H. (1995). Mineral Nutrition in Higher Plants. Academic Press, London.
Matson, P. A., Parton, W. J., Power, A. G., and Swift, M. J. (1997). Agricultural intensifi-
cation and ecosystem properties. Science 277, 504–509.
Matson, P. A., Naylor, R., and Ortiz-Monasterio, I. (1998). Integration of environmental,
agronomic, and economic aspects of fertilizer management. Science 280, 112–115.
Meisinger, J. J., Bandel, V. W., Angle, J. S., O’Keefe, B. E., and Reynolds, C. M. (1992).
Presidedress soil nitrate test evaluation in Maryland. Soil Sci. Soc. Am. J. 56, 1527–1532.
Neumann, G., and Römheld, V. (2002). Root-induced changes in the availability of
nutrients in the rhizosphere. In “Plant Roots, The Hidden Half” (Y. Waisel, A. Eshel,
U. Kafkafi, et al., Eds.), pp. 617–649. Marcel Dekker, Inc., New York.
Peng, L. H. (2006). Application of Real-Time Testing of Nitrate Technique in Nitrogen
Management of Cucumber in Greenhouse. PhD Thesis, China Agricultural University,
Beijing, China. (in Chinese with English abstract).
Peng, S., Buresh, R., Huang, J., Zhong, X., Zou, Y., Yang, J., Wang, G., Liu, Y., Hu, R.,
and Tang, Q. (2010). Improving nitrogen fertilization in rice by site-specific N manage-
ment. A review. Agron. Sustain. Dev. 30, 649–656.
Raun, W. R., and Johnson, G. V. (1999). Improving nitrogen use efficiency for cereal
production. Agron. J. 91, 357–363.
Raun, W. R., Solie, J. B., Johnson, G. V., Stone, M. L., Mullen, R. W., Freeman, K. W.,
Thomason, W. E., and Lukina, E. V. (2002). Improving nitrogen use efficiency in cereal
grain production with optical sensing and variable rate application. Agron. J. 94, 815–820.
Ren, H. Q. (2007). Characteristics of Nitrogen Input and in Peach Orchards and Effects on
Nitrate Content of Groudwater in the Farms of Pinggu Arable Land. Master Thesis,
China Agricultural University, Beijing, China (In Chinese with English abstract).
Ren, T., Christie, P., Wang, J. G., Chen, Q., and Zhang, F. S. (2010). Root zone soil
nitrogen management to maintain high tomato yields and minimum nitrogen losses to
the environment. Sci. Hortic. 125, 25–33.
Robertson, G. P., and Vitousek, P. M. (2009). Nitrogen in agriculture: Balancing the cost of
an essential resource. Annu. Rev. Environ. Resour. 34, 97–125.
Roelcke, M., Han, Y., Schleefk, K. H., Zhu, J. G., Liu, G., Cai, Z. C., and Richter, J.
(2004). Recent trends and recommendations for nitrogen fertilization in intensive
agriculture in eastern China. Pedosphere 14, 449–460.
38 Fusuo Zhang et al.

Roth, G. W., and Fox, R. H. (1990). Soil nitrate accumulation following nitrogen fertilized
corn in Pennsylvania. J. Environ. Qual. 9, 243–248.
Sattelmacher, B., and Thoms, K. (1995). Morphology and physiology of the seminal root
system of young maize (Zea mays L.) plants as influenced by a locally restricted nitrate
supply. Z. Pflanz. Bodenkunde. 158, 493–497.
Schachtman, D. P., Reid, R. J., and Ayling, S. M. (1998). Phosphorus uptake by plants:
From soil to cell. Plant Physiol. 116, 447–453.
Schenk, M. K., and Barber, S. A. (1979). Root characteristics of corn genotypes as related to
P uptake. Agron. J. 71, 921–924.
Schleef, K. H., and Kleihanss, W. (1994). Mineral Balance in Agriculture in EU. Institute of
Farm Economics, Federal Agricultural Research Centre, Brounchweig, Germany.
Schmitt, M. A., and Randall, G. W. (1994). Developing a soil nitrogen test for improved
recommendations for corn. J. Prod. Agric. 7, 328–334.
Shen, J., Li, H., Neumann, G., and Zhang, F. (2005). Nutrient uptake, cluster root
formation and exudation of protons and citrate in Lupinus albus as affected by localized
supply of phosphorus in a split-root system. Plant Sci. 168, 837–845.
Shen, J. L., Tang, A. H., Liu, X. J., Fangmeier, A., Goulding, K. W. T., and Zhang, F. S.
(2009). High concentrations and dry deposition of reactive N species at two sites in the
North China Plain. Environ. Pollut. 157, 3106–3113.
Shen, J., Liu, X., Fangmeier, A., Goulding, K., and Zhang, F. (2011a). Atmospheric
ammonia and particulate ammonium from agricultural sources in the North China
Plain. Atmos. Environ. 45, 5033–5041.
Shen, J., Yuan, L., Zhang, J., Li, H., Bai, Z., Chen, X., Zhang, W., and Zhang, F. (2011b).
Phosphorus dynamics: From soil to plant. Plant Physiol. 156, 997–1005.
Shi, W. M., Yao, J., and Yan, F. (2008). Vegetable cultivation under greenhouse conditions
leads to rapid accumulation of nutrients, acidification and salinity of soils and ground-
water contamination in South-Eastern China. Nutr. Cycl. Agroecosyst. 83, 73–84.
Shorrocks, V. M., and Nicholson, D. D. (1980). The influence of boron deficiency on fruit
quality. Acta Hortic. (ISHS) 92, 103–110.
Sogbedji, J. M., Van Es, H. M., and Yang, C. L. (2000). Nitrate leaching and nitrogen
budget as affected by maize nitrogen rate and soil type. J. Environ. Qual. 29, 1813–1820.
Song, X. Z., Zhao, C. X., Wang, X. L., and Li, J. (2009). Study of nitrate leaching nitrogen fate
under intensive vegetable production pattern in northern China. C. R. Biol. 332, 385–392.
Tang, L. L., Chen, Q., Li, X. L., Chen, Y. Z., and Ding, G. G. (2005). Studies on the target
value of nitrogen supply for greenhouse tomato growth during autumn-winter season.
Plant Nutr. Fertil. Sci. 11(2), 230–235. (in Chinese with English abstract).
Tian, Q. Y., Chen, F. J., Zhang, F. S., and Mi, G. H. (2006). Genotypic difference nitrogen
acquisition ability in maize plants is related to the coordination leaf and root growth.
J. Plant Nutr. 29, 317–330.
Tian, Q. Y., Chen, F. J., Liu, J. X., Zhang, F. S., and Mi, G. H. (2008). Inhibition of maize
root growth by high nitrate supply is correlated to reduced IAA levels in roots. J. Plant
Physiol. 165, 942–951.
Tilman, D. (1999). Global environmental impacts of agricultural expansion: The need for
sustainable and efficient practices. Proc. Natl. Acad. Sci. USA 96, 5995–6000.
Tilman, D., Cassman, K. G., Matson, P. A., Naylor, R., and Polasky, S. (2002). Agricultural
sustainability and intensive production practices. Nature 418, 671–678.
Tong, T. A., Emteryd, O., and Zhang, S. L. (2004). Evaluation of over-application of
nitrogen fertilizer in China’s Shanxi province. Sci. Agric. Sin. 37, 1239–1244. (in Chinese
with English abstract).
Vance, C. P. (2008). Plants without arbuscular mycorrhizae. In “The Ecophysiology of
Plant-Phosphorus Interactions” (P. J. White and J. P. Hammond, Eds.), pp. 117–142.
Springer, Dordrecht, The Netherlands.
Nutrient Management in China 39

Vitousek, P. M., Naylor, R., Crews, T., David, M. B., Drinkwater, L. E., Holland, E.,
Johnes, P. J., Katzenberger, J., Martinelli, L. A., Matson, P. A., Nziguheba, G.,
Ojima, D., et al. (2009). Nutrient imbalances in agricultural development. Science 324,
1519–1520.
Wang, J. Q. (2007). Analysis and Evaluation of Yield Increase of Fertilization and Nutrient
Utilization Efficiency for Major Cereal Crops in China. PhD Thesis, China Agricultural
University, Beijing, China (in Chinese with English abstract).
Wang, J. G. (2011). Management of Degraded Vegetable Soils in Greenhouse. China
Agricultural University Press, Beijing, China (in Chinese).
Wang, X., Cao, Y., Zhang, F., and Chen, X. (1995). Application of building-up and
maintenance approach in agriculture. Plant Nutr. Fertil. Sci. 1, 59–63. (in Chinese with
English abstract).
Wang, L. J., Jiang, W. B., He, Q. F., and Fan, H. B. (2001). Studies on the relationship of
development of bitter pit in apple fruits with the contents of calcium and magnesium and
the activities of antioxidant enzymes. Acta Hortic. Sin. 28(3), 200–205. (in Chinese with
English abstract).
Wang, Y., Mi, G. H., Chen, F. J., and Zhang, F. (2003). Genotypic differences uptake by
maize inbred lines its relation to root morphology. Acta. Ecol. Sin. 23, 297–302.
Wang, B. L., Shen, J., Zhang, W. H., Zhang, F. S., and Neumann, G. (2007). Citrate
exudation from white lupin induced by phosphorus deficiency differs from that induced
by aluminum. New Phytol. 176, 581–589.
Wang, J. Y., Liu, Q. H., and Liu, J. L. (2010a). Analysis on the characteristic and cause of
orchard soil acidification in the area of Shandong peninsula. Chin. Agric. Sci. Bull. 26(16),
164–169. (in Chinese with English abstract).
Wang, X. Y., Hang, B., and Liu, C. L. (2010b). Distribution of calcium in bagged apple fruit
and relationship between antioxidant enzyme activity and bitter pit. Agric. Sci. Technol. 11
(1), 82–85. (in Chinese with English abstract).
Wehrmann, J. V., and Scharpf, H. C. (1979). Mineral nitrogen in soil as an indicator for
nitrogen fertilizer requirements (Nmin-method). Plant Soil 52(1), 109–126.
Xu, S. Y., Zhang, F. S., and Wang, H. (2008). Effects of environmental factors on internal
bark necrosis of apple trees. J. Fruit Sci. 25(1), 73–77. (in Chinese with English abstract).
Ye, Y. L., Yang, S. Q., Liu, S. L., and Wang, W. L. (2008). Study on vegetable production,
fertilizer application, soil chemical and physical property variance in suburb of Zhengz-
hou city. Henan Sci. 26(1), 51–55. (in Chinese with English abstract).
Yoshida, S. (1981). Fundamentals of Rice Crop Science. International Rice Research
Institute, Los Baños, Philippines. 269 pp.
Zhang, X. M. (2010). Effect of Fertilization on the Yield and the Apparent N Balance for
Greenhouse Tomato in Shouguang. Master Thesis, China Agricultural University, Beij-
ing, China. (in Chinese with English abstract).
Zhang, H. M., and Forde, B. G. (1998). An Arabidopsis MADS box gene that controls
nutrient-induced changes in root architecture. Science 279, 407–409.
Zhang, S. L., Cai, G. X., Wang, X. Z., Xu, Y. H., Zhu, Z. L., and Freney, J. R. (1992). Loss
of urea-nitrogen applied to maize grown on a calcareous fluvo-aquic soil in North China
Plain. Pedosphere 2, 171–178.
Zhang, F. S., Shen, J. B., Li, L., and Liu, X. (2004). An overview of rhizosphere processes
related with plant nutrition in major cropping systems in China. Plant Soil 260, 89–99.
Zhang, X. S., Liao, H., Chen, Q., Christie, P., Li, X. L., and Zhang, F. S. (2007). Response
of tomato on calcareous soils to different seedbed phosphorus application rates. Pedosphere
17, 70–76.
Zhang, F. S., Wang, J. Q., Zhang, W. F., Cui, Z. L., Ma, W. Q., Chen, X. P., and Jiang, R. F.
(2008a). Nutrient use efficiencies of major cereal crops in China and measures for
improvement. Acta Pedolog. Sin. 45, 915–924. (in Chinese with English abstract).
40 Fusuo Zhang et al.

Zhang, Y., Liu, X. J., Fangmeier, A., Goulding, K. T. W., and Zhang, F. S. (2008b).
Nitrogen inputs and isotopes in precipitation in the North China Plain. Atmos. Environ.
42, 1436–1448.
Zhang, F. S., Chen, X. P., and Chen, Q. (2009). The Fertilization Guideline of Major Crop
in China. China Agricultural University Press, Beijing. (in Chinese).
Zhang, F., Shen, J., Zhang, J., Zuo, Y., Li, L., and Chen, X. (2010). Rhizosphere processes
and management for improving nutrient use efficiency and crop productivity: Implica-
tions for China. Adv. Agron. 107, 1–32.
Zhang, F. S., Cui, Z. L., Fan, M. S., Zhang, W. F., Chen, X. P., and Jiang, Q. F. (2011).
Integrated soil-crop system management: Reducing environmental risk while increasing
crop productivity and improving nutrient use efficiency in China. J. Environ. Qual. 40,
1–7.
Zhao, X., and Xing, G. (2009). Variation in the relationship between nitrification and
acidification of subtropical soils as affected by the addition of urea or ammonium sulfate.
Soil Biol. Biochem. 41, 2584–2587.
Zhao, R., Chen, X., Zhang, F., Zhang, H., Schroder, J., and Römheld, V. (2006).
Fertilization and nitrogen balance in a wheat–maize rotation system in North China.
Agron. J. 98, 938–945.
Zhao, T. S., Yu, L. C., and Jiao, R. (2007). Study on the relationship between calcium
nutrition and bitter pit in bagged apples. J. Fruit Sci. 24(5), 649–652. (in Chinese with
English abstract).
Zheng, X. H., Han, S. H., Huang, Y., Wang, Y. S., and Wang, M. X. (2004). Re-
quantifying the emission factors based on field measurements and estimating the direct
N2O emission from Chinese croplands. Glob. Biogeochem. Cycl. 18, 1–19. GB2018.
Zhou, J. P. (2009). The problems of manure fertilization of orchards and correct methods.
Bull. Agric. Sci. Technol. 3, 151–152. (in Chinese with English abstract).
Zhu, Z. L. (1998). The status, problems and countermeasures of nitrogen fertilizer applica-
tion in China (in Chinese). In “Fertilizer Issues of Sustainable Agriculture Development
in China” (Q. K. Li, et al., Eds.), pp. 28–51. Jiangsu Science and Technology Publishing,
Nanjing, Jiangsu.
Zhu, J. H., Li, X. L., Christie, P., and Li, J. L. (2005). Environmental implications of low
nitrogen use efficiency in excessively fertilized hot pepper (Capsicum frutescens L.)
cropping systems. Agr Ecosyst Environ 111, 70–80.
 C H A P T E R T W O

Effect of Climate Change Factors on

Processes of Crop Growth and
Development and Yield of Groundnut
(Arachis hypogaea L.)
Uttam Kumar,* Piara Singh,* and K. J. Boote†

Contents
1. Introduction 42
2. Vegetative Development 43
2.1. Germination and emergence 43
2.2. Leaf appearance and leaf number 44
3. Canopy Expansion and Growth Processes 45
3.1. Leaf thickness 45
3.2. Leaf area and stem elongation 46
3.3. Leaf senescence 47
3.4. Stomatal conductance and transpiration 47
3.5. Photosynthesis 48
3.6. Net assimilation and growth rates 49
4. Reproductive Development and Growth 50
4.1. Appearance of flowers, pegs, and pods 50
4.2. Rate of flower production 51
4.3. Pollen production and viability and fruit-set 53
4.4. Number of pegs, pods, and seeds 54
4.5. Pod and seed growth rates and their size 55
5. Total Dry Matter, Pod, and Seed Yield 56
6. Harvest Index and Shelling Percentage 58
6.1. Harvest index 58
6.2. Shelling percentage 59
7. Root Growth and Root-to-Shoot Ratio 59
7.1. Root growth 59
7.2. Root-to-shoot ratio 60

* International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Andhra Pradesh,
India
{
Agronomy Department, University of Florida, Gainesville, Florida, USA

Advances in Agronomy, Volume 116 # 2012 Elsevier Inc.

ISSN 0065-2113, DOI: 10.1016/B978-0-12-394277-7.00002-6 All rights reserved.

41
42 Uttam Kumar et al.

8. Synthesis of the Review for Improving the CROPGRO or Other

Models for Groundnut 61
8.1. Vegetative development 61
8.2. Reproductive progression 61
8.3. Vegetative expansion and photosynthesis processes 62
8.4. Pod addition, seed growth, and partitioning intensity 63
8.5. Climatic effects on root growth 64
9. Concluding Comments 64
Acknowledgment 65
References 65

Abstract
Global warming is changing climate in terms of increased frequency of extreme
weather events as well as increased air temperature and vapor pressure deficit
of air and spatial and temporal change in rainfall. In spite of the beneficial effect
of increased atmospheric CO2 concentration, climate change will adversely
impact the production and productivity of groundnut grown in subtropical and
tropical regions of the world. This chapter reviews the current state of knowl-
edge on effects of climate change factors on the growth and development of
groundnut. This review identifies research gaps and suggests upgrades to
groundnut models, such as the CROPGRO-Groundnut model, which is being
used as a tool to assess impacts of climate change on groundnut crop. This
review revealed that the direct and indirect effects of most climate change
factors on plant growth and development processes are well understood and
already incorporated in the CROPGRO-Groundnut model. Extreme events asso-
ciated with climate change may sometimes cause water-logging, extreme soil
water deficiency, or extreme humidity conditions, and these effects could be
better addressed in the models.

1. Introduction
The Fourth Assessment report of the Inter-Governmental Panel on
Climate Change (IPCC, 2007) has reconfirmed that the atmospheric con-
centrations of carbon dioxide, methane, and nitrous oxide greenhouse gases
(GHGs) have increased markedly since 1750. The global increases in CO2
concentrations are due primarily to fossil-fuel use and land-use change,
while those of methane and nitrous oxide are primarily due to agriculture.
The IPCC has also shown that these increases in GHGs have resulted in
warming of the climate system by 0.74
C over the past 100years, and the
projected increase in temperature by 2100 is about 1.8–4.0
C. For the
South Asia region, the IPCC has projected 0.5–1.2
C rise in temperature
by 2020, 0.88–3.16
C by 2050, and 1.56–5.44
C by 2080 depending upon
the scenario of future development. Overall, the temperature increases are
Effect of Climate Change Factors on Processes of Crop Growth 43

likely to be much higher in winter season than in rainy season. With climate
change, more frequent hot days, heat waves, and warm spells are expected
to increase. These increases in the temperatures are likely to result in both
spatial and temporal variations in rainfall. Overall, there will be an increase
in rainfall especially in the tropical regions. The pattern of precipitation is
already changing and will become more erratic and intense with warming of
the globe. Because of the increase in temperatures, vapor pressure deficit of
the air will increase in spite of the increase in humidity with the increase in
rainfall. For the A1B SRES scenario, the expected increase in CO2 concen-
tration will be 420ppm by 2020, 530ppm by 2050, and 650ppm by 2080 as
estimated by the SPAM model (IPCC, 2001).
These changes in climatic factors (CO2, temperature, vapor pressure defi-
cit, and rainfall) will alter plant growth and development processes and most
likely have negative impact on crop productivity, especially in the semiarid
tropical regions, where the current temperatures are already high and close to
the upper limits beyond which the plant processes will be adversely affected.
Therefore, in spite of some expected benefits of the increased CO2 concentra-
tion on some crops, global warming poses a potential threat to agricultural
production and productivity throughout the world. Increased incidence of
weeds, pests, and plant diseases with climate change may cause even greater
economic losses to agricultural production. It is projected that even a small rise
in temperature (1–2
C) at lower latitudes, especially in the seasonally dry
tropical regions (IPCC, 2007), would decrease crop productivity.
Groundnut (Arachis hypogaea L.) is one of the major oilseed and food
crops grown in subtropical and tropical regions of the world. It is grown in
different rainfall and temperature regimes on a variety of soils. Being a C3
crop, higher temperatures and other climatic factors may affect its produc-
tivity and to some extent its distribution. This chapter attempts to review
the current state of knowledge of climate factor effects on growth and
development response of groundnut and revisits the need to fine-tune the
CROPGRO and other groundnut models to determine the impacts and
adaptation of groundnut to the climate change in future.

2. Vegetative Development
2.1. Germination and emergence
After groundnut seeds are sown, germination and emergence are primarily
determined by the temperature and soil moisture in the seeding zone.
The processes of germination and emergence have a minimum threshold
value, optimum range, and maximum threshold value for both temperature
and soil moisture contents. At minimum threshold values of temperature
(base temperature) and soil moisture content, the processes of germination are
44 Uttam Kumar et al.

not initiated. At the optimum range of temperature and soil moisture both,
germination and emergence takes place at a maximum rate. Between their
minimum threshold and lower optimum values, the rates of germination and
emergence increase with the increase in temperature and soil moisture.
Above their optimum range, these processes are progressively slowed down
until they completely stop at their respective maximum threshold values
(damaging thresholds). For example, Awal and Ikeda (2002) and Prasad
et al. (2006) reported that base temperature for germination of groundnut is
approximately 10
C and the optimum temperature (OT) for emergence is
between 25 and 30
C. Mohamed et al. (1988) and Angus et al. (1981)
reported base temperatures ranging from 8 to 13
C for groundnut seed
germination. These differences in base temperature suggest genotypic differ-
ence among cultivars studied. In terms of soil temperature, the optimum
mean soil temperature for seed germination is between 29 and 30
C
(Mohamed et al., 1988) and for root growth it is close to 30
C (Suzuki,
1966). Leong and Ong (1983) also reported that in two cooler (wet) soil
temperatures (19 and 22
C) less than 50% emergence of groundnut seedling
took place, while at warmer temperatures (25, 28, and 31
C) the percentage
of emergence varied from 70% to 80%. Seedling emergence started within
5 days after sowing (DAS) in warm temperatures but in 10 DAS at 19
C.

2.2. Leaf appearance and leaf number

Like germination and emergence, vegetative development of groundnut
crop is also determined by temperature and soil moisture availability. As soil
moisture availability decreases, turgor pressure in leaves decreases and slows
leaf appearance and expansion. There may also be limited variation among
genotypes (ecotypes) in response to temperature and soil moisture. Leong
and Ong (1983) reported that base temperature, below which there is no
development, varied between 8 and 11
C among several genotypes. They
also reported decrease in leaf appearance rate under water deficit conditions.
Bagnall and King (1991a) estimated that Spanish varieties have a phenologi-
cal base temperature of 13.6
C, whereas Valencia and Virginia varieties
have a base temperature of 12.6 and 11.4
C, respectively. As far as soil
temperature is concerned, the rate of leaf appearance showed positive linear
functions with soil temperatures (Awal and Ikeda, 2002). The plants grown
in comparatively warmer soil produced more leaves on their branches than
on the main axis. This phenomenon of increasing leaf number on branches
in warmer soil gives plants the initial vigor for establishment by capturing
more light and CO2. The impact of soil temperature is less at later stages as
plants become more dependent on air temperature rather than soil temper-
ature for their development. Studies on day and night air temperatures
showed that OTs for vegetative development in groundnut range from
25/25 (Wood, 1968) to 30/26
C (Cox, 1979). Marshall et al. (1992)
Effect of Climate Change Factors on Processes of Crop Growth 45

recorded maximum rate of foliage development for groundnut (cv. Robut

33-1) in the temperature range of 28–30
C. More recently, Williams and
Boote (1995) and Weiss (2000) reported the OT range from 25 to 30
C for
vegetative development of groundnut.
Rao (1999) studied the interactions of CO2 and temperature on ground-
nut (cv. TMV 2) growth and development using open top chambers. Plants
were grown in ambient conditions for 30days in pots and then transferred to
open top chambers maintained at combinations of two levels of temperature
(35 and 40
C) and two levels of CO2 (330 and 660mmolmol1). At all
temperature and CO2 levels, the total number of leaves per plant ranged
from 33 to 36 per plant at 60days of plant age. Elevated CO2 did not
significantly change the total leaf numbers; however, leaf area and leaf
weights were higher at elevated CO2 than at ambient CO2. There was no
interaction between CO2 and temperature for leaf numbers per plant.

3. Canopy Expansion and Growth Processes

3.1. Leaf thickness
Specific leaf area (SLA) influences canopy expansion and growth through its
effect on total leaf area per plant affecting light interception and light use
efficiency (LUE). Temperature is the major factor affecting SLA of ground-
nut. Ketring (1984) studied the effect of temperatures ranging from 30/22
to 35/22
C on the growth and development of two groundnut cultivars
(Tamnut 74 and Starr). Observations made at 63 and 91days after planting
(DAP) showed that SLA of both the cultivars was unaffected over time
in growth chambers maintained at 30/22
C, whereas at 35/22
C, the SLA
of both the cultivars increased much faster during the same period, cultivar
Tamnut 74 being less sensitive than Starr. However, Talwar et al. (1999) did
not observe any significant effect of temperature increase from 25/25 to
35/25
C on the SLA of three cultivars studied. Pilumwong et al. (2007)
studied the growth and development responses of groundnut cultivar Tai-
nan 9 to the combination of two temperatures (25/15 and 35/25
C) and
three CO2 concentrations (400, 600, and 800mmolmol1). Observation
made at 112 DAP showed that SLA of plants was 22% less at low tempera-
ture than at high temperature (HT). Elevated CO2 did not affect SLA. In an
open top chamber study, Rao (1999) did not observe any significant effect
of temperature increase from 35 to 40
C on SLA of TMV 2 variety.
Increase in CO2 concentration from 330 to 660mmolmol1 did not affect
SLA. In both the studies, the interaction between CO2 and temperature
for SLA was nonsignificant. From these studies, it is clear that SLA of
groundnut increases with the increase in temperature. However, different
results were obtained in different studies.
46 Uttam Kumar et al.

3.2. Leaf area and stem elongation

In a growth chamber study, Ketring (1984) showed that when groundnut
plants were transferred from 30/251
C to experimental temperatures
(30/22, 32/22, and 35/22
C) the leaf area of two cultivars (Tamnut 74
and Starr) progressively decreased with the increase in temperature when
observed at 63 and 91 DAP. At harvest (91 DAP), the decrease in leaf area
per plant was about 49% for Tamnut 74 and about 80% for Starr at 35/22
C
as compared to leaf area of respective cultivars at 30/22
C. Stem elongation
was significantly inhibited by both 32/22 and 35/22
C for Tamnut 74 and
by 35/22
C for Starr. Contrary to the Ketring’s results, Talwar et al. (1999)
in a glasshouse study observed that all vegetative growth parameters (such as
leaf area, stem elongation, etc.) of three genotypes (ICG 1236, ICGS 44, and
Chico) increased at 35/25
C as compared to those observed at 25/25
C.
These contradicting results between the two studies may be caused by lower
light intensity in growth chamber studies.
In the Rao (1999) study, both HTs (40 vs. 35
C) and high CO2 (660 vs.
330mmolmol1) increased leaf area per plant. Leaf area per plant was
maximum in elevated CO2 at 40
C and minimum in ambient CO2 at
35
C. Length of the longest stem in all treatments was not significantly
affected by temperature or enrichment of CO2. Pilumwong et al. (2007) in a
growth chamber study observed that at 112 DAP the total plant leaf area
decreased with increasing temperature from 25/15 to 35/25
C at all levels of
CO2 concentrations. Leaf area per plant averaged over two temperatures was
greatest in 600mmolmol1 CO2, followed by 800mmolmol1 CO2, and
400mmolmol1 CO2. The interaction between temperature and CO2 was
not significant for leaf area per plant. At 25/15
C, main stem length was 24%
and 44% longer in 600 and 800mmolmol1 CO2, respectively, in comparison
to plants grown at 400mmolmol1 CO2, while at 35/25
C, the main stem
lengths were similar across CO2 concentrations. These responses of increase
in stem length with increasing CO2 concentration at 25/15
C and no
significant change at 35/25
C might be because of detrimental effect of
HT in combination with low light on synthesis and translocation of assim-
ilates to plant parts (Pilumwong et al., 2007). The differences in results
between the two studies for leaf area and main stem lengths may be due to
different experimental setups for the two studies. Rao (1999) conducted the
experiment in an open top chamber, while Pilumwong et al. (2007) con-
ducted in controlled growth chamber. However, these studies give an indi-
cation that the leaf area per plant and stem elongation may increase up to 35
C
with the increase in temperature.
Clifford et al. (1993) studied the growth and yield of groundnut variety
Kadiri 3 grown in controlled-environment glasshouses at 28
C (5
C)
under two levels of atmospheric CO2 (350 or 700ppm) and two levels of
soil moisture (irrigated weekly or no water after 35 DAS). In the irrigated
Effect of Climate Change Factors on Processes of Crop Growth 47

treatment, the maximum leaf area index (LAI) reached 7.5 in ambient CO2
and 8.0 in elevated CO2 at the end of the season. Under drought conditions,
elevated CO2 had a highly significant effect on canopy development. Plants
achieved a maximum LAI of 3 in ambient CO2 and 4.3 in elevated CO2.
Later when the drought conditions intensified, LAI declined to 1.9 in the
ambient CO2 and 3.0 in the elevated CO2. Groundnut plants grown under
elevated CO2 in drought conditions maintained less negative leaf water
potential than the plants grown in ambient CO2, which helped in main-
taining the turgor potential for growth and expansion of leaves. These
results showed that elevated CO2 benefits the crop growth under both
water limiting and nonlimiting conditions; however, the relative benefits
are more under water limiting conditions (something that model simula-
tions also show).

3.3. Leaf senescence

Hardy and Havelka (1977) reported that CO2-enriched treatment acceler-
ated the leaf senescence in groundnut plants. In contrast, Chen and Sung
(1990) found that groundnut plants grown at two concentrations of CO2
(1000mLL1 and ambient 340mLL1) had similar timing of start of leaf
senescence. The study of Hardy and Havelka (1977) might have had
confounding effect of ethylene contamination of CO2.

3.4. Stomatal conductance and transpiration

In a controlled growth chamber study, Prasad et al. (2003) reported that
stomatal conductance and transpiration rates significantly increased with the
increase in temperature and decreased with the increase in CO2 concentra-
tion. In the temperature range of 32/22–44/34
C, the stomatal conduc-
tance increased linearly by 0.12 and 0.04molm2 s1 and transpiration by
1.4 and 0.8mmolm2 s1 with every degree Celsius rise in temperature
under both ambient and elevated CO2, respectively. The interaction
between temperature and CO2 was also significant (P¼0.08) for these
processes (Prasad et al., 2003).
Clifford et al. (1995) did not observe any significant effect of CO2
enrichment (700 vs. 375ppm) on stomatal conductance during early season
(up to 28 DAS) when plants were well supplied with water; however, later
in the life cycle, conductance was less for CO2-enriched compared to
ambient plants under full irrigation. At 114 DAS under drought, the
conductance of droughted plants had fallen to zero under ambient CO2,
whereas measurable conductance was still recorded for the adaxial leaf
surface of plants grown at elevated CO2, which indicates soil water conser-
vation. Elevated CO2 as compared to the ambient CO2 decreased stomatal
48 Uttam Kumar et al.

frequency on both the surfaces of leaves up to 16% in the irrigated treatment

and by 8% in the droughted plants on the adaxial surface only. However,
elevated atmospheric CO2 promoted larger reduction in leaf conductance
than changes in stomatal frequency, indicating partial stomatal closure.
These results suggest that the effects of future increase in atmospheric
CO2 concentration on stomatal frequency in groundnut are likely to be
small, especially under conditions of water stress, but that combination of
associated reductions in leaf conductance at elevated CO2 will be important
in the semiarid tropics.
Stronach et al. (1994) conducted a study on stands of groundnut
(cv. Kadiri 3) in controlled-environment glasshouses at two mean air tem-
peratures (28 and 32
C), two atmospheric CO2 concentrations (375 and
700ppm) and two soil moisture regimes (irrigated weekly to field capacity
or allowed to dry from 22 DAS). Transpiration equivalent (product of
accumulated biomass/transpiration and saturation deficit of air, gkPakg1)
was calculated using total above and below ground plant biomass. Neither
temperature nor soil moisture treatments had any effect on transpiration
equivalent. Increase in CO2 concentration raised transpiration equivalent
value from 6.210.30 to 7.670.29gkPakg1 in the dry treatment. This
increase of 24% is on the order of the change in the water use efficiency as
predicted by Morison (1985) for the whole plants, which is of significant
importance for crops grown with limited soil water availability.

3.5. Photosynthesis
Talwar et al. (1999) recorded higher net photosynthetic rate in three
groundnut genotypes grown at 35/30
C as compared to those grown at
25/25
C at 30 and 60 DAS. They also observed genotypic differences in net
photosynthesis at both temperatures. In crops like groundnut (C3 crops),
Rubisco is not saturated by the current concentration of CO2 in the
atmosphere. So an increase in CO2 concentration will improve the balance
of CO2 and O2 at Rubisco site, thus improving the CO2-exchange rate
(CER) of the plant by providing more substrate for photosynthesis. Prasad
et al. (2003) reported that doubling of ambient CO2 concentration (350 vs.
700mmolmol1) enhanced leaf photosynthesis of groundnut by 27% across
a range of daytime temperatures (32–44
C), but they found no CO2 by
temperature interaction on leaf photosynthesis. On the other hand, some
researchers have suggested that optimum growth temperature for several
plants may rise significantly with increasing concentration of atmospheric
CO2 (Berry and Bjorkman, 1980; McMurtrie and Wang, 1993; McMurtrie
et al., 1992; Stuhlfauth and Fock, 1990). Long (1991) calculated from well-
established plant physiological principles that most C3 plants should increase
their OT for growth by approximately 5
C with 300ppm increase in
CO2 concentration. Thus, photosynthetic rates are expected to rise with
Effect of Climate Change Factors on Processes of Crop Growth 49

simultaneous increases in both the CO2 concentration and canopy temper-

ature as suggested by Idso and Idso (1994).
Clifford et al. (1993) reported that, under irrigated condition, the maxi-
mum rate of net photosynthesis of groundnut increased up to 40% by
elevated CO2 (700ppm) compared to ambient CO2. This was also accom-
panied by increase in LUE for biomass production by 30%, from 1.66 to
2.16gMJ1 in elevated CO2. Where no irrigation was given after 35 DAS,
the increase in LUE was 94%, from 0.64 to 1.24gMJ1 in elevated CO2.
Such differences in photosynthetic efficiency were also observed in another
study by Clifford et al. (1995), where under gradual imposition of severe
drought, the net photosynthesis increased under enriched CO2, while it was
negative under ambient CO2 at 114 DAS of groundnut crop. At elevated
CO2, plants maintained less negative and higher leaf water potential which
enables them to remain active for longer period of time in dry soil condi-
tions (Clifford et al., 1993).
Chen and Sung (1990) reported that leaf CO2 exchange rate increased
with increasing photosynthetic photon flux density (PPFD) in plants grown
at 340 and 1000mLCO2 L1. Plants grown in 1000mLCO2 L1 had greater
leaf CER at all PPFD levels. The apparent maximum quantum yield
estimated from the initial slope of the light response curve of high CO2-
grown plants (0.06mmolCO2 mmol1 quanta) was much higher than that of
ambient CO2-grown plants (0.026mmolCO2 permmol1 quanta), indicat-
ing better efficiency of light utilization by photosynthesis in high CO2-
grown plants. Leaf CER responded to intercellular partial pressure of CO2
(Ci) in a curvilinear manner with increasing Ci level. Plants grown at 1000mL
CO2 L1 consistently exhibited a higher leaf CER than the plants grown at
340mLCO2 L1.

3.6. Net assimilation and growth rates

Rao (1999) in his study reported that both HTs (40 vs. 35
C) and CO2
(660 vs. 330ppm) significantly increased the net assimilation rate (NAR) of
groundnut. At 330ppm CO2, NAR increased from 4.092 to 4.328gm2
day1 with the increase in temperature from 35 to 40
C. At 660ppm CO2
level, it increased from 4.660 to 4.890gm2 day1 with the same increase in
temperature. Relative growth rate (RGR) showed a similar trend as NAR
in response to temperature and CO2. The interaction between CO2 and
temperature for both NAR and RGR was significant. Greater NAR and
RGR in elevated CO2 are linked to the increase in rate of photosynthesis
(Hertog et al., 1993; Lenssen and Rozema, 1990).
Nigam et al. (1994) studied the effect of temperature and photoperiod on
growth and development of three genotypes of groundnut (TMV 2, NC Ac
17090, and VA 81B). Mean plant growth rate of three genotypes decreased
from 87.5 to 52.4mg pl1 Cd–1 with the increase in temperature from 22/18
50 Uttam Kumar et al.

to 30/26
C. These results are in contrast to the results obtained by Rao

(1999) in an open top chamber study. Mean plant growth rate of genotypes
was significantly higher in long-day (12h) photoperiod (84.8mg pl1oCd1)
than those in short-day (9h) photoperiod (53.8mg pl1 Cd1). There was no
interaction between photoperiod and temperature for plant growth rate.

4. Reproductive Development and Growth

4.1. Appearance of flowers, pegs, and pods
Leong and Ong (1983) reported that flowering at 19, 22, 25, 28, and 31
C
occurred at 61, 49, 40, 32, and 31 DAS, respectively, in the wet treatment.
In the dry treatment, flowering occurred at 56, 43, 37, 31, and 28 DAS in
the same order of increasing temperatures. The calculated base tempera-
ture for the appearance of flowering was 10.8
C. Bagnall and King
(1991a) studied the effect of four temperature regimes (24/19, 27/22,
30/25, and 33/28
C) on flowering, fruiting, and growth of cv. Early
Bunch. The lowest temperature regime (24/19
C) considerably slowed
the appearance of first flower, and subsequent flower and peg production
rates were also strongly depressed by low temperature. In the Talwar et al.,
(1999) study, when the temperatures were increased from 25/25 to
35/30
C, the days to first flower appearance decreased from 37 to 31 for
ICG 1236, 38 to 33 for ICGS 44, and 33 to 27days for Chico. Earlier
studies (Bolhuis and de Groot, 1959; Fortanier, 1957) showed that OT for
time to flowering and vegetative growth for different groundnut varieties
is in the range of 28–30
C. Marshall et al. (1992) also reported that the rate of
foliage development increased to maximum in this range of temperatures
for cv. Robut 33-1.
Pilumwong et al. (2007) reported that the duration from planting to first
flower was 22 and 34days at 35/25 and 25/15
C, respectively, for both
ambient and elevated CO2. Prasad et al. (2003) observed that the duration of
groundnut from sowing to flowering at temperatures 32/22, 36/26, 40/30,
and 44/34
C was 30, 31, 26, and 28days, respectively, under both ambient
(350mmolmol1) and elevated CO2 (700mmolmol1). Thus the OT for
flower appearance was 40/30
C (35
C). HT (40/30
C and higher) delayed
pegging and podding in groundnut, indicating greater sensitivity of pegging
and podding than flowering to HTs. Duration from flowering to pegging at
both 32/22 and 36/26
C was about 8 days, while at 40/30
C it took about
10days. The time from flowering to podding was about 16days at 32/22
and 36/26
C, while at 40/30
C, it was 19days. Prasad et al. (2003) did not
observe any affect of enhanced CO2 on the phenology of groundnut.
Bagnall and King (1991a) reported that at 30/25
C, six photoperiod
treatments ranging from 10 to 14h, had little effect on days to first flower
Effect of Climate Change Factors on Processes of Crop Growth 51

appearance in four groundnut cultivars (two Spanish and two Virginia

types). However, flower production was enhanced significantly in short-
day photoperiods. To observe the interaction of photoperiod and temper-
ature for flower appearance, two temperature (24/19 and 30/25
C) and
five photoperiod treatments (11–14h) were studied on 12 cultivars (four
Spanish, three Valencia, and five Virginia types). Average daily irradiance
at canopy level during this experiment was 13.7MJm2. Bagnall and King
(1991a) found no effect of photoperiod or interaction between tempera-
ture and photoperiod on the time to flower. They also subjected a similar
range of groundnut varieties to two photoperiods (12 and 14h) and three
temperatures regimes (33/28, 27/22, and 21/16
C) in winter with an
irradiance level of 7.0MJm2 d1. Most of the varieties examined showed
a short-day photoperiodic response; they flowered faster under short-day
at higher temperatures (33/22 or 27/22
C). At low temperature (21/16
C),
the time to first flower was similar under both short and long days in
all varieties. Bagnall and King (1991a) also reported that photon flux density
(Q) below 500mmolm2 s1 considerably slowed down the progress toward
flowering at a constant temperature of 30
C. At photon flux density (Q) of
500mmolm2 s1 and higher, different varieties flowered at a particular dry
weight (leaf and stem), whereas at low Q, plant dry weights were much
reduced at the time of flowering. Thus, delay in flowering associated with
low Q is correlated with slowing of dry matter production. Under low Q,
there was evidence of Qphotoperiod interaction for days to first flower.
These studies by Bagnall and King (1991a) indicated that while temperature
has a major role in flowering of groundnut, some modulation by photope-
riod and irradiance may be needed under certain climatic conditions.

4.2. Rate of flower production

Bagnall and King (1991b) studied the reproductive development of ground-
nut in the temperature range of 24/19–33/28
C. Average rate of flower
production (per plant) from the first flower appearance to peak flower
production was 11 flowers week1 at 33/28
C, 7.4 flowers week1
at 30/25
C, 6.6 flowers week1 at 27/22
C, and 1.8 flowers week1 at
24/19
C. They observed that total flower and total peg numbers were
strongly correlated with vegetative growth, particularly main stem leaf
number, at 70days of sowing. Disregarding the initial vegetative phase to
about 12.5 leaves, on an average in all the temperature regimes, 14.7 flowers
were formed for every new leaf on the main stem. Similarly, Talwar et al.
(1999) also reported that flower number per plant increased at HT (35/30

C) in three genotypes (ICG 1236, ICGS 44, and Chico) compared to 25/
25
C. Total flower numbers were also correlated with plant dry weight and
number of leaves per plant.
52 Uttam Kumar et al.

Prasad et al. (1999a) studied the effect of HT on two groundnut cultivars,

ICGV 86015 and ICGV 87282. Initially, both cultivars were grown at OT
(28/22
C), and after first appearance of flower bud (21DAP), half the plants
were transferred to HT (38/22
C). Thereafter, the plants were transferred
at 3-day intervals from OT to HT and from HT to OT, up to 46DAP,
giving a total of nine transfer treatments. Plants remained in the new
temperature regime for 6days before being returned to their original
regime, where they remained until harvest at 67DAP. HT had a significant
effect (P<0.001) on the total flower number in the controls and in the
reciprocal transfer treatments. HT increased flower production in the HT-
control and OT to HT transfer treatments and vice versa in the HT to OT
transfer treatments. However, these changes in flower production only
occurred 6 days following transfer to HT or OT (P<0.01). During the
6-day OT or HT stress period, temperature had no significant (P<0.35)
effect on flower production. These results show that HT had no deleterious
effect on flower production, but these results did not address the effect of
HT on fruit-set. The effect of temperature treatments was similar for both
cultivars, and there was no temperaturecultivar interaction.
Bagnall and King (1991b) examined two groundnut cultivars (Robut
33-1 and Early Bunch) in long- (16h) and short-day (12h) treatments and
found that short days promoted greater flowering numbers in both ground-
nut cultivars as compared to long-day treatment. Cumulative flower num-
bers were greater in short-day treatment than in long-day by 70% for Robut
33-1 and 88% for Early Bunch at 30
C at 24days after beginning of
flowering. In the same study, they also reported that flower numbers in
groundnut variety White Spanish were also influenced by photon flux
density (Q) imposed after first flower appearance. In four treatments of
photon flux density, viz. 400, 550, 700, 1000mmolm2 s1 from first week
to next 17days of flowering, flower number at 1000mmolm2 s1 was
double that of those plants at 400mmolm2 s1. Plants grown at high Q
had more plant dry weight than the plants grown at low Q. The ratio of
flower number to dry weight suggested that at higher Q there were
proportionally more flowers (35%) than at lowest Q.
Lee et al. (1972) grew groundnut plants (cv. Starr) in a greenhouse at 30
C
until beginning of flowering (30–35days of age). At this time, one group of
plants was moved to growth room at 95% relative humidity. At 50days
of age, the relative humidity of the growth room was lowered to 50%.
A second group of plants at beginning of flowering was placed into a growth
room at 50% relative humidity and at 50days the humidity was raised to
95%. Flowering was stimulated by transfer from low to high humidity and
these plants set the largest percentage of pegs. The plants in the high to low
humidity transfer had least number of flowers and formed the lowest
percentage of pegs. These results indicate that when plants are exposed to
high humidity the flower production is increased.
Effect of Climate Change Factors on Processes of Crop Growth 53

4.3. Pollen production and viability and fruit-set

Prasad et al. (1999b) studied the effects of short episodes of heat stress on
pollen production and viability and fruit yield. Plants of cultivar ICGV
86015 were grown at a day/night temperature of 28/22
C from sowing
until 9 days after flowering (DAF). Cohorts of plants were then exposed to a
factorial combination of 4-day temperatures (28, 34, 42, and 48
C) and two
night temperatures (22 and 28
C) for 6days. Thereafter, all plants were
maintained at 28/22
C until final harvest 9days later. Both hot days and
warm nights had prominent effect on groundnut pollen production and its
viability. As the day temperatures increased from 28 to 48
C, pollen
production and pollen viability reduced by 390 per flower
C1 and
1.9%
C1, respectively. Warmer nights (28 vs. 22
C) reduced mean pollen
number from 4389 to 2800 per flower and mean pollen viability from 49%
to 40%. Reduced fruit-set was a consequence of fewer pollen grains and
reduced pollen viability. The threshold temperature for pollen production
and viability was 34
C and there was strong negative linear relationship
between both pollen production and viability and accumulated temperature
above 34
C. Prasad et al. (2000a) exposed the groundnut plants for 6-day
periods starting 9 DAF to the day temperature range of 28–48
C either for
whole day (08:00–20:00h) or for 6h during AM or PM of the day. Along
with air temperatures of growth cabinets, floral bud temperatures were
continuously measured over a 6-day period. Variation in flower number
was quantitatively related to floral bud temperature during the day over the
range 28–43
C. In contrast, floral bud temperatures above 36
C during
AM and whole day significantly reduced fruit-set (number of pegs and
pods), whereas high PM temperature had no effect on fruit-set. They
recommended that number of pegs and pods per plant can be modeled by
combining the response of flower numbers and fruit-set to temperature.
Talwar (1997) showed that flower buds of groundnut are sensitive to
temperature stress at a phase 3–5 days before anthesis, which coincides with
microsporogenesis (Martin et al., 1974; Xi 1991). HT during microsporo-
genesis causes low pollen viability, poor anther dehiscence, and hence male
sterility. This pollen sterility at HT may be associated with early degenera-
tion of tapetal layer (Ahmed et al., 1992; Suzuki et al., 2001) and reduction
in carbohydrates in developing pollen (Pressman et al., 2002).
Prasad et al. (2003) also studied the season-long effect of super-optimal
temperatures (32/22–44/34
C) and elevated CO2 (350 vs. 700mmol
mol1) on reproductive processes of groundnut. Pollen viability decreased
with increasing temperature under both ambient (350mmolmol1) and
elevated CO2 (700mmolmol1) treatments. Pollen viability of the tagged
flowers was about 90–95% at 32/22 and 36/26
C, but decreased to 68% at
40/30
C and zero at 44/34
C. Seed set was 70–80% at 32/22 and 36/26
C,
50% at 40/30
C, and 0% at 44/34
C under both ambient and elevated
54 Uttam Kumar et al.

CO2. There was no effect of CO2 or interaction between temperature and

CO2 on pollen viability.

4.4. Number of pegs, pods, and seeds

Bolhuis and De Groot (1959) in their study recorded highest number of
pegs at 27 or 30
C. Bagnall and King (1991b) reported increase in peg
numbers when the temperature was increased from 24/19 to 33/28
C.
Similarly, Talwar et al. (1999) reported increase in peg numbers of ground-
nut cultivars when temperature was increased from 25/25 to 35/30
C, but
the pod numbers decreased with the increase in temperature. These results
indicate that peg formation is not adversely affected by temperatures up to
the range of 33/28–35/30
C. However, Ketring (1984) in his range of
temperatures (30/22, 32/22, and 35/22
C) reported a 33% decrease in
number of pegs with increasing temperature from 30/22 to 35/22
C, but
this was in low-light chambers.
In the Prasad et al. (2003) study, both pegging and podding were delayed
above the 32/22–36/26
C temperature range. As the temperatures increased
from 32/22 to 44/34
C pod number decreased from 353 to 74m2 under
ambient CO2 (350mmolmol1) and from 407 to 116m2 under elevated CO2
(700mmolmol1). Similarly, with the same temperature increase, seed number
decreased from 587 to 43m2 at ambient CO2 and 709 to 132m2 at elevated
CO2. Across all temperatures, elevated CO2 compared with ambient CO2
increased pod number by 40% and seed number by 31%. The interaction
between temperature and CO2 for pod and seed number was not significant.
Air and soil temperature both are important factors to determine the
yield of groundnut as groundnut flowers develop aerially and pods in the
soil. The optimum soil temperature range for pod formation and develop-
ment is between 31 and 33
C, and soil temperatures above 33
C signifi-
cantly reduce the number of mature pods and seed yields (Dreyer et al.,
1981; Ono, 1979; Ono et al., 1974). However, Golombek and Johansen
(1997) found that the greatest number of pods were produced at slightly low
range of mean soil temperatures, that is, between 23 and 29
C, while
temperatures of 17 and 35
C were sub- and supraoptimal, respectively.
Prasad et al. (2000b) studied the individual as well as combined response
of air and soil temperature on yield and yield components of groundnut.
The effects of high air (38/22 vs. 28/22
C) and high soil temperatures (38/30
vs. 26/24
C) were imposed from flowering or podding. High air tempera-
ture had no significant effect on total flower production but significantly
reduced the proportion of flowers setting pegs (fruit-set) and hence the fruit
numbers. In contrast, high soil temperature significantly reduced flower
production, the proportion of pegs forming pods, and 100-seed weight.
The combined treatment of high soil and air temperatures reduced fruit-set
and pod weight by 58% and 57% at podding and 49% and 52% at flowering,
Effect of Climate Change Factors on Processes of Crop Growth 55

respectively, indicating high sensitivity to temperatures at podding stage.

The effects of high air and soil temperature were mostly additive and
without any interaction.
Bell et al. (1991) studied the effect of temperature and photoperiod on
Spanish, Virginia, and Valencia types of groundnut and reported strong
photoperiodtemperature interaction for number of pegs and pods pro-
duced. Photoperiod did not affect time to first flower, but the number of
pegs and pods and total pod weight per plant decreased in long (16 or 17h)
photoperiods. For example, pod numbers of two cultivars, that is, White
Spanish and NC 17090, decreased with increasing photoperiod (17 vs. 11.9–
13.5h) at two temperatures (33/17 and 33/23
C). Similarly, Bagnall and
King (1991b) studied the response of groundnut to temperature, photope-
riod, and irradiance on flowering and development of pegs and pods. Flower
and peg number at 60–70 days from emergence were approximately doubled
by 12-h days (SD) compared with plants with 16-h days (LD). Peg numbers
were highly correlated to flower numbers and their ratio was independent of
differing photoperiod treatments, suggesting that there was no major effect
of day length on flower abortion. However, the pod number, and therefore,
yield, was more influenced by photoperiod than was flower or peg formation.
Photoperiod-induced changes in flower and fruit numbers were independent
of growth and plant dry weight. Conversely, temperature and light intensity
affected flower numbers and these changes were correlated with growth-
related changes in leaf number and plant dry weight.
Leong and Ong (1983) reported that rate of peg and pod formation,
mainly controlled by temperature, was not significantly affected by dry or
wet soil treatments. However, Rao et al. (1985) observed significant yield
reductions when water stress was imposed from start of flowering to start
of seed growth. They attributed yield reductions due to water deficits in
the top 4–5cm of soil that prevented peg and pod development in the dry
and hard soil. Similar results have also been obtained in other studies (Boote
et al., 1976; Matlock et al., 1961; Ono et al., 1974; Pallas et al., 1979;
Underwood et al., 1971).

4.5. Pod and seed growth rates and their size

Optimum air temperature for pod growth as suggested by various research-
ers appears to lie between 20 and 24
C (Cox, 1979; Williams et al., 1975).
Cox (1979) observed that the individual and total pod weights and the rate
of increase in pod weight were greatest at the mean temperature of 23.5
C.
So partitioning of dry matter to pods would, therefore, be expected to
decrease as temperature increases above 24
C (Ong, 1984). Pilumwong
et al. (2007) found that as temperature increases from 25/15 to 35/25
C,
pod dry weight reduced by 50%. Pod weight reduction by HT (35/30 vs.
25/25
C) was also reported by Talwar et al. (1999) for three genotypes.
56 Uttam Kumar et al.

Nigam et al. (1994) reported that temperature had a significant effect

(P<0.01) on pod growth rate but there was no overall effect of photope-
riod. In the tested genotypes, highest pod growth rate was observed at
26/22
C compared to 22/18 and 30/26
C. Photoperiod effects on pod
growth rate for cvs. TMV 2 and Nc Ac 17090 were not significant in any
temperature regimes. On the other hand, significantly greater pod growth
rate for VA 81B occurred in long day than in short day 26/22
C. The study
may provide evidence of genotypic variability for photoperiodtempera-
ture interaction which could influence adaptation for groundnut genotypes
to new environments.

5. Total Dry Matter, Pod, and Seed Yield

Cox (1979) observed that accumulation of top dry weight in early
growth was optimum at a weighted mean temperature of 27.5
C and no
shoot growth was observed at 15.5
C indicating positive linear function of
growth above 15.5
C. But further increase in temperature above optimum
range may decrease dry matter production. Craufurd et al. (2002) observed
that HT (38/22
C) significantly (P0.001) reduced total dry weight of four
groundnut cultivars (ICGV 86015, 796, ICGV 87282, and 47-16) by
20–35% as compared to the 28/22
C treatment. Similar results were
obtained by Prasad et al. (2000b) in a polytunnel study where the groundnut
plants exposed to high air (38/22
C) and/or high soil temperature (38/30

C) significantly reduced total dry matter production, its partitioning to pods
and pod yields of groundnut. Cox (1979) reported that temperatures above
26/22
C (24
C mean temperature) reduced the pod weight per plant.
Ong (1984) observed significant reduction in number of subterranean pegs
and pods, seed size, and seed yield by 30–50% at temperature above 25
C.
Using semiclosed chambers, Chen and Sung (1990) exposed peanut plants
(cv. Li-Chih-Taze) to enriched CO2 atmosphere (1000mLCO2 L1) during
two different growth periods, that is, from pod formation (R3 stage) to final
harvest (R8 stage) or seed filling (R5 stage) to final harvest. Groundnut plants
produced more dry matter accumulation and higher pod yield in the enriched
treatment (1000mmolmol1 CO2) as compared to the ambient treatment
(340mmolmol1 CO2). The enrichment-stage effect on these parameters
was not significant.
Pilumwong et al. (2007) reported that above-ground biomass of ground-
nut was increased by elevated CO2 (800 vs. 400mmolmol1) in both the
low (25/15
C) and the HT (35/25
C) treatments. Pod dry weight
increased with increasing CO2 at 25/15
C but was not different among
CO2 levels at 35/25
C. At 25/15
C, pod dry weight was 50% higher than
at 35/25
C. Highest above-ground biomass production at 35/25
C, under
Effect of Climate Change Factors on Processes of Crop Growth 57

800mmolmol1 CO2, indicates that the HT regime chosen in this study was
still in the OT range for biomass production of groundnut. Rao (1999)
reported increased dry weight of shoot in elevated CO2 (660 vs. 300ppm)
even at 40
C.
Prasad et al. (2003) reported increase in total dry matter production of
groundnut with increase in CO2 between temperatures of 32/22 and
40/30
C. Further increase in temperature to 44/34
C decreased total dry
matter under both ambient (350mmolmol1) and elevated CO2 (700mmol
mol1). As the temperature increased from 32/22 to 44/34
C, pod yield
decreased by 89% and 87% under ambient and elevated CO2, respectively.
With the same increase in temperature, the seed yield decreased by 90% and
88% under ambient and elevated CO2, respectively. Temperature and CO2
effect on total dry matter, pod, and seed yields was statistically significant;
however, the interaction between temperature and CO2 for all yields was
not significant. On average, total dry matter yield increased by 36%, and
both pod and seed yields increased by 30% under elevated CO2 across all the
temperature regimes. The study showed that when the groundnut crop is
exposed to HTs throughout the full season, total dry matter production is
reduced at temperatures above 40/30
C (35
C), whereas the pod and seed
yields are adversely affected above temperatures of 32/22
C (27
C). These
results differ from the Cox (1979) study results that OT for dry matter
production ranges from 25 to 30
C with a mean of 27.5
C, whereas the
pod and seed yields start declining above 24
C. The study of Cox (1979)
used pot-grown plants at lower light intensity.
Clifford et al. (1993) reported that, in well-irrigated conditions,
elevated CO2 (700ppm) increased above-ground dry matter accumulation
by an average of 16% over the ambient CO2 concentration (350ppm).
Droughted plants grown at elevated CO2 produced more than double the
dry matter of plants grown at ambient CO2. Average increase in pod yield
with elevated CO2 was 25%, from 2.73 to 3.42tha1 in well-irrigated
plots, with a sixfold increase from 0.22 to 1.34tha1 in the droughted
treatment. The reason for such differential response to CO2 in two
moisture regimes was discussed earlier as a result of CO2-induced water
conservation in the section on stomatal conductance and photosynthesis.
Timing and intensity of water stress can enhance or reduce yield of
groundnut. Rao et al. (1985) reported that when groundnut plants received
12–15% less water during vegetative growth (or up to start of pegging) pod
yields increased by 12–19% compared to the fully irrigated control. Earlier
work at ICRISAT (ICRISAT Annual Report, 1981); Ong (1984) showed
similar increase in pod yield under mild water stress during vegetative phase
of groundnut. In the Rao et al. (1985) study when plants were stressed from
start of flowering to start of seed growth, total biomass and pod yield were
reduced as much as 50% and 77%, respectively. Greatest reduction in kernel
yield occurred when stress was imposed during the seed-filling phase.
58 Uttam Kumar et al.

As fruit initiation continues even after the start of kernel growth, soil water
deficits during pod-filling stage reduce both the initiation and the develop-
ment of pods (Boote et al., 1976; Matlock et al., 1961; Ono et al., 1974;
Pallas et al., 1979; Underwood et al., 1971).

6. Harvest Index and Shelling Percentage

6.1. Harvest index
Prasad et al. (2003) found that pod and seed harvest indices (HIs) at harvest
maturity were significantly affected by temperature, but not by CO2.
As temperatures increased from 32/22 to 44/34
C, pod HI decreased
from 0.50 to 0.07 and seed HI from 0.41 to 0.05, respectively, under both
ambient and elevated CO2. Talwar et al. (1999) reported that HI decreased
significantly at HT (35/30
C) compared to OT (25/25
C) and the decrease
was more than 59% in all the tested genotypes. Craufurd et al. (2002)
also reported similar reduction in seed HI ranging from 0% to 65% at HT
(38/22
C) for the four cultivars. Temperature had similar effect of reducing
the dHI/dt (rate of change in HI) for pod and seeds in all genotypes. HT had
no effect on dHI/dt of moderately heat tolerant genotypes, that is, 796 and
47-16. But in susceptible genotypes, ICGV 86016 and ICGV 87282, the
start of pod and seed filling was delayed by 5–9 days and dHI/dt was reduced
by 20–65% at 38/22
C. Craufurd et al. (2002) concluded that crop models
need to account for genotypic differences in HT effect on timing and rate
of dHI/dt to successfully simulate yields in warmer climates.
Bell et al. (1991) observed that the HI of cvs. White Spanish and
NC 17090 decreased under long day (17h) as compared to the short days
(11.9–13.5h) at both the temperatures; however, the decrease was more at
higher temperature (33/23
C) than at lower temperature (33/17
C).
Nigam et al. (1994) also reported decrease in partitioning coefficient (pod
growth rate/plant growth rate) of three selected genotypes with HT and
long photoperiod. Flohr et al. (1990) suggested that long days increase the
thermal time for initiation of pegs and pods, thus resulting in less partition-
ing of dry matter to these reproductive organs.
Ong (1984) reported that partitioning of dry matter to pods [expressed as
pod weight ratio (PWR)] was 0.178 and 0.042 at 25 and 31
C, respectively,
in an irrigated treatment. In the water limited treatment, PWR decreased
with increasing water deficit. At 27
C, PWR was 0.104 and 0.067 in the
wet treatment having saturation vapor pressure deficit (SVPD) of 1.0 and
dry treatment with SVPD of 3.0, respectively. Clifford et al. (1993) did not
observe any marked difference in seed HI of groundnut in two CO2
treatments (350 and 700ppm) in irrigated condition, which was 0.20
under ambient CO2 (350ppm) and 0.21 under elevated CO2 (700ppm).
Effect of Climate Change Factors on Processes of Crop Growth 59

In the drought treatment, HI was 0.05 in ambient CO2, which increased to

0.15 in elevated CO2. Similar results were obtained by Stronach et al. (1994)
on fraction of biomass partitioning to pods in ambient and elevated CO2
(375 vs. 700ppm) in irrigated and drought conditions.

6.2. Shelling percentage

In Prasad et al. (2003) study, shelling percentage decreased from 82% to 74%
(by 0.7units
C1) as temperature increased from 32/22 to 44/34
C under
both ambient and elevated CO2. HT decreases the partitioning of dry
matter to seeds which results in low shelling percentage (Craufurd et al.,
2002). Ketring (1984) reported a 25 and 20% reduction in mature seed
weight at 35
C compared to 30
C for Tamnut 74 and Starr cultivars,
respectively. Similarly, Talwar et al. (1999) reported that seed setting and
seed weight of three tested genotypes (ICG 1236, ICGS 44, and Chico)
were significantly reduced under HT 35/30
C compared to 25/25
C.
Shelling percentage was 60–76% at 25/25
C and 41–62% at 35/30
C for
three genotypes, viz. ICG 1236, ICGS 44, and Chico. Rao et al. (1985)
reported decrease in shelling percentage when water stress was imposed
during pod-filling stage.

7. Root Growth and Root-to-Shoot Ratio

7.1. Root growth
In a phytotron experiment, Wood (1968) reported that root dry weights of
groundnut plants decreased with increasing day temperatures from 20 to
35
C keeping night temperature the same (25
C). At 35/25
C, root dry
weight was only 35% of the weight at 20/25
C and the difference was
highly significant. In a short-term rhizotron study, Pilumwong et al. (2007)
reported that total root length and number of roots at 17 DAP were
significantly greater in the plants grown at low temperature (25/15
C)
than those at HT (35/25
C) in all CO2 concentrations. However, in the
long-term rhizotron study, plants grown at HT (35/25
C) had significantly
greater root number, greater root length, and greater root length density at
99 DAP than those at 25/15
C. This shows that short-term study in this
case does not represent long-term study in terms of HT impacts on root
growth. In terms of soil temperature, Suzuki (1966) reported OT close to
30
C for root growth.
Chen and Sung (1990), using semiclosed CO2 enrichment chamber,
studied the effect of CO2 enrichment on the growth of Virginia type
groundnut. In the 340mLCO2 L1 treatment, root dry weight was 2.01–
2.33gplant1. In the enriched treatment (1000mLCO2 L1), root dry
60 Uttam Kumar et al.

weight was 3.28–3.67gplant1 when applied from pod to harvest and 2.79–
3.41gplant1 when applied from seed filling to harvest stage. Similarly, Rao
(1999) using open top chamber observed increase in dry weight of root with
CO2 enrichment from 330 to 660ppm at 35 to 40
C. Pilumwong et al.
(2007) in a rhizotron study observed that when CO2 concentration was
increased from 400 to 800mmolmol1 the fibrous root dry weight of ground-
nut plants increased at 25/15
C but decreased at 35/25
C. Clifford et al.
(1993) using closed environment glasshouse observed that under ambient
(350ppm) and elevated (700ppm) CO2 the dry root weights were 180.2 and
177.3gm2 in the irrigated treatment and 274.0 and 274.7gm2 in the
drought treatment in the respective CO2 concentrations. This indicates that
root dry weight was unaffected by CO2 at a given moisture regime but was
increased by drought. These differences in root weight response to CO2 in
different studies may be attributed to the differences in the crop growth
facility used for experimentation.

7.2. Root-to-shoot ratio

Prasad et al. (2000b) reported that partitioning of dry matter to root
increased when the plants were exposed to high air temperature (38/22
C)
at the beginning of flowering. But when the treatment was applied at
beginning pod, partitioning of dry matter to root reduced significantly
and no change in total dry matter was observed. This difference in dry
matter partitioning to root under HT at these two stages could be caused by
preferential partitioning of dry matter to reproductive organs when stressed
at pod formation stage (Yamagata et al., 1987). Prasad et al. (2000b) also
observed that partitioning of dry matter to roots was greater when plants
were grown at high soil temperature (38/30 vs. 26/24
C) than when
grown in high air temperature (38/22 vs. 28/22
C). Both high air and
soil temperatures above 30
C increase dry matter partitioning to roots.
Craufurd et al. (2002) in their study on four groundnut cultivars (two
Spanish and two Virginia genotypes) found that root-to-shoot ratio of
different genotypes was significantly reduced (P¼0.01) by 20–35% at
38/22
C as compared to 28/22
C. Rao (1999) did not find any effect of
temperature or CO2 concentration on the root-to-shoot ratio. Root-to-
shoot ratio under ambient CO2 (330mmolmol1) was 0.039 at 35
C and
0.037 at 40
C. Under elevated CO2 (660mmolmol1), it was 0.038 at both
the temperatures.
Root-to-shoot ratio considerably decreased under elevated CO2
(700ppm) in both irrigated and drought treatments (Clifford et al., 1993).
In irrigated treatment, root-to-shoot was decreased from 0.19 to 0.12 when
CO2 concentration increased from 330 to 700ppm. In the drought treat-
ment, it decreased from 0.70 to 0.33 in respective concentrations of CO2.
Effect of Climate Change Factors on Processes of Crop Growth 61

Overall, in drought treatment, root-to-shoot ratio was greater than irrigated

treatments (Clifford et al., 1993).

8. Synthesis of the Review for Improving the

CROPGRO or Other Models for Groundnut
8.1. Vegetative development
Base temperature for germination of groundnut seeds is 10
C and the OT
for emergence ranges from 25 to 30
C (Awal and Ikeda, 2002; Prasad et al.,
2006). However, different genotypes may have different base temperature
ranging from 8 to 13
C (Leong and Ong, 1983; Mohamed et al., 1988).
Optimum soil temperature for germination is 29–30
C (Mohamed et al.,
1988). Base temperature for vegetative development of groundnut geno-
types ranges from 8 to 11
C (Leong and Ong, 1983) and the OT is between
25 and 30
C (Weiss, 2000; Williams and Boote, 1995). Elevated CO2 does
not affect vegetative progression of groundnut (Rao, 1999).
Currently in the groundnut model (Boote et al., 1986, 1991, 1998; Singh
et al., 1994a,b), the base temperature is 11
C and the OTs for vegetative
development range from 28 to 30
C, and the damaging threshold tempera-
ture is taken as 55
C. There is little information in the literature on how
vegetative development is affected by temperatures above 30
C. Soil tem-
perature and soil water status are considered in the model for germination and
emergence, but only air temperature (not soil) is used for subsequent vegeta-
tive development. Less is known about how soil moisture stress, especially
excess soil water, affects the groundnut crop and what is the optimum range
or threshold values affecting germination or vegetative development of
groundnut. Extreme events associated with climate change may cause
water-logging or extreme soil water deficiency, and these effects, if suffi-
ciently understood, need to be incorporated in the model.

8.2. Reproductive progression

Base temperature for first flower appearance is 10.8
C (Leong and Ong,
1983) and the OT is in the range of 28–30
C (Bolhuis and de Groot,
1959; Fortanier, 1957). On the other hand, Prasad et al. (2003) reported
that appearance of flowers was hastened with the increasing temperatures
up to 40/30
C (35
C) but slowed down beyond this temperature.
Temperatures above 36/26
C (31.5
C) delayed pegging and podding
in groundnut. Thus, HTs increase rate of flowering and flower produc-
tion but have deleterious effect on fruit-set. At high irradiance level, day
length has no effect on days to flower. At low irradiance level, short days
enhance time to first flower at HTs but not at low temperatures. Low
62 Uttam Kumar et al.

photon flux density (Q<500mmolm2 s1) slows the progress toward

flowering and the interaction between Q and photoperiod was signifi-
cant for days to first flower (Bagnall and King, 1991a). Soil moisture
regime or CO2 concentration does not influence the appearance of
flowers in groundnut.
Currently in the groundnut model (Boote et al., 1998), the base temper-
ature for progression to flowering is 11
C and the OT range is 28–30
C,
with progressively slower progress above 30
C, reaching zero progress
(damaging threshold) at 55
C. In the model, after the beginning seed
stage (R5 stage), the base temperature for development is reduced to 5
C
and the optimum to 26
C. There is no short-day photoperiod effect for any
cultivar currently used in the groundnut model, but the code is pro-
grammed to accept a short-day sensitivity if sufficient evidence is provided.
So far, none of the 30 or so commonly grown cultivars exhibit any short-
day acceleration of time to flower (we think NC 17090 is not typical of
current cultivars). Low Q effect on time to flower is not incorporated in the
model, although it could be important for low-light growth cabinets.

8.3. Vegetative expansion and photosynthesis processes

Leaf area expansion and stem elongation increase with the increase in
temperature up to 35/25
C (Talwar et al., 1999). Drought reduces leaf
extension rates. Elevated CO2 benefits the crop growth under both water
limiting and nonlimiting conditions; however, the relative benefits are more
under water limiting conditions (Clifford et al., 1993). Threshold tempera-
ture up to which SLA increases appears to be 30
C. Elevated CO2 does not
influence SLA of groundnut (Ketring, 1984; Pilumwong et al., 2007).
Stomatal conductance and transpiration rates increase with temperature,
whereas elevated CO2 reduces these processes. Elevated CO2 enhances
CER, photosynthesis, LUE, and transpiration efficiency of groundnut
(Chen and Sung, 1990; Clifford et al., 1993; Prasad et al., 2003). Talwar
et al. (1999) observed increase in crop growth and net photosynthesis when
temperature increased from 25/25 to 35/25
C, whereas Bell et al. (1991)
observed increase in crop growth rates up to 33/23
C. Rao (1999) reported
that increase in temperature (35–40
C) and elevated CO2 (330–660mmol
mol1) had positive effect on RGR.
In the CROPGRO-Groundnut model (Boote et al., 1998), the expan-
sion processes for plant height and width are decreased at temperatures
below 26
C. The model reduces leaf expansive processes, for example,
SLA, when temperature falls below the 27
C optimum, being reduced to
20% of optimum at 14
C. Thus these expansive processes are sufficiently
represented in the model. Exact cardinal temperatures for crop growth rate
and biomass increase are more difficult to interpret because leaf appearance
rate, leaf area expansion, as well as leaf photosynthesis have separate effects,
Effect of Climate Change Factors on Processes of Crop Growth 63

and maintenance respiration increases with rising temperature (in the

model). Leaf photosynthesis in the model has an electron-transport rate
that has a linear response from zero rate at 8
C up to optimum at 40
C, but
the rubisco competition for CO2 versus O2 is programmed in the code and
causes quantum efficiency to be reduced as temperature rises; thus single leaf
photosynthesis is practically at its maximum between 30 and 40
C (Boote
and Pickering, 1994). There is also a minimum night temperature effect that
reduces light-saturated rate if the minimum temperature is less than 22
C.
All the processes of CO2 and temperature sensitivity of photosynthesis
are represented in the model directly or indirectly (see method in Boote
and Pickering, 1994) and have been tested and shown to work well (Boote
et al., 2010).

8.4. Pod addition, seed growth, and partitioning intensity

Increase in temperature, short days, light intensity, high Q, and high
humidity promote flower numbers in groundnut (Bagnall and King,
1991a,b; Lee et al., 1972; Prasad et al., 1999b; Talwar et al., 1999). Threshold
temperature for pollen production and viability is 34
C, above which both
pollen production and viability decrease linearly with the increase in
temperature (Prasad et al., 1999b). Floral bud temperatures above 36
C
during AM and whole day significantly reduce fruit-set (Prasad et al.,
2000a). Elevated CO2 does not affect pollen viability (Prasad et al., 2003).
Peg formation is not affected up to the air temperature range of 33/28 to
35/30
C (30.5–32.5
C), but the pod and seed numbers are decreased.
Optimum air temperature for podding is around 36/26
C (31
C) (Prasad
et al., 2003). Optimum air temperature for pod growth as suggested by
many researchers appears to lie between 20 and 24
C, whereas optimum
soil temperature for pod formation and development is between 29 and 33

C (Dreyer et al., 1981; Golombek and Johansen, 1997; Ono, 1979; Ono
et al., 1974). Both air and soil temperatures have additive effect on repro-
ductive growth (Prasad et al., 2003). Elevated CO2 increases pod and seed
numbers. Long photoperiod decreases number of flowers, pegs, and pods
and pod weight. Pod numbers are more sensitive to photoperiod than
number of flowers and pegs (Bagnall and King, 1991a,b; Bell et al., 1991).
Soil water deficits prevent peg and pod development (Rao et al., 1985). HT
and water stress decreases HI, except when mild water stress occurs prior to
flowering (Craufurd et al., 2002; Rao et al., 1985). Long days decrease HI
and the temperaturephotoperiod interaction was significant for HI.
Enhanced CO2 increases HI. HT decreases shelling percentage. Drought
reduces shelling percentage when it occurs during pod-filling period.
The CROPGRO-Groundnut model (Boote et al., 1998) has a parabolic
temperature function for relative rate of flower and pod formation per day
that has a base temperature of 15
C, with an optimum between 20 and
64 Uttam Kumar et al.

26.5
C, declining to zero addition at 40
C. The individual pod and

individual seed growth rates function (per shell or per seed) in the model
depend on a similar parabolic function, with a base temperature of 6
C,
with an optimum between 21 and 23.5
C, declining to zero growth rate
at 41
C (this is strongly supported by Cox, 1979). In addition, there is a
function that reduces partitioning to pods and seeds as maximum tempera-
ture exceeds 33
C, going to a 0.40 relative value at 46
C (but of course, no
flowers or pods would be added above 40
C). These three functions were
found by Boote et al. (2010, see their Fig. 4) to mimic well the data of Prasad
et al. (2003), showing that optimum pod yield was at 24
C and progressively
declined to zero yield at a mean temperature of 39–40
C. The model also
well reproduced data of Cox (1979) showing OT for pod and seed growth
to be about 24
C. With coding, these functions could be replaced by
explicit temperature effects on transitions from individual flowers to suc-
cessful pegs and pods using information similar to Prasad et al. (1999a,b).
The CROPGRO model does allow mild photoperiod effects on seed
growth rate of soybean based on reliable data, but data for same effect on
groundnut are too tenuous to turn this effect on at present.

8.5. Climatic effects on root growth

The change in root growth or root-to-shoot ratio at HT depends upon the
timing, duration, and intensity of temperature stress in relation to crop
growth stage. OT for root growth is close to 30
C (Suzuki, 1966). Gener-
ally, both high air and soil temperatures above 30
C decrease dry matter
partitioning to roots. Soil water deficit and enhanced CO2 increase root
growth. HT and enhanced CO2 decrease root-to-shoot ratio, while water
stress increases root-to-shoot ratio.
The effects of CO2 and water stress on root growth are indirectly taken
care of in the model via their effect on plant water deficit and partitioning
to roots. Presently, the CROPGRO-Groundnut model does mimic
increased root growth under CO2 enrichment, as well as enhanced parti-
tioning to root as a function of water deficit. However, the direct effects of
HT on root-to-shoot ratio are not modeled, unless that operates via
enhanced water deficit.

9. Concluding Comments
Groundnut (A. hypogaea L.) is one of the major oilseed and food crops of
the subtropical and tropical regions of the world. It is grown in different rainfall
and temperature regimes on a variety of soils. Depending upon the location on
the globe, climate change may benefit or adversely affect the productivity of
Effect of Climate Change Factors on Processes of Crop Growth 65

this crop. This chapter has reviewed the current state of knowledge on effects
of climate change factors, such as extremes of air and soil temperatures, relative
humidity, water availability and their interactions with photoperiod, light
intensity, and increased atmospheric CO2 concentration, on the growth and
development of groundnut. This review identified research gaps and needs to
generate information to upgrade the CROPGRO-Groundnut model. This
review revealed that the direct and indirect effects of most climate change
factors on plant growth and development processes are well understood and
already incorporated in the model. Extreme events associated with climate
change such as water-logging, extreme soil water deficiency or extreme
humidity conditions will affect the productivity of the crop. Low-light inten-
sity affects flowering and high air and soil temperatures affect root growth and
root-to-shoot ratio. The effects of these factors on groundnut crop growth and
development need to be sufficiently understood before these are suitably
incorporated in the model to enhance its capability for better assessment
of climate change impacts and to develop adaptation strategies to cope up
with climate change in different agro-climates. Direct comparison of model
simulations against experimental data reported in some studies listed in this
review would be useful.

ACKNOWLEDGMENT
We are grateful to ICRISAT for providing financial support through the USAID
linkage fund.

REFERENCES
Ahmed, F. E., Hall, A. E., and DeMason, D. A. (1992). Heat injury during f loral develop-
ment in cowpea (Vigna unguiculata, Fabaceae). Am. J. Bot. 79, 784–791.
Angus, J. F., Cunningham, R. B., Moncur, M. W., and MacKenzeie, D. H. (1981). Phasic
development in field crops. I. Thermal response in the seedling phase. Field Crop Res. 3,
365–378.
Awal, M. A., and Ikeda, T. (2002). Effects of changes in soil temperature on seedling
emergence and phenological development in field-grown stands of peanut (Arachis
hypogaea). Environ. Exp. Bot. 47, 101–113.
Bagnall, D. J., and King, R. W. (1991a). Response of peanut (Arachis hypogaea) to tempera-
ture, photoperiod and irradiance 1. Effect on flowering. Field Crop Res. 26, 263–277.
Bagnall, D. J., and King, R. W. (1991b). Response of peanut (Arachis hypogaea) to tempera-
ture, photoperiod, and irradiance. 2. Effect on peg and pod development. Field Crop Res.
26, 279–293.
Bell, M. J., Bagnall, D. J., and Harch, G. (1991). Effect of photoperiod on reproductive
development of peanut (Arachis hypogaea L.) in a cool subtropical environment. II.
Temperature interactions. Aust. J. Agric. Res. 42, 1151–1161.
Berry, J., and Bjorkman, O. (1980). Photosynthetic response and adaptation to temperature
in higher plants. Annu. Rev. Plant Physiol. 31, 491–543.
66 Uttam Kumar et al.

Bolhuis, G. G., and De Groot, W. (1959). Observations on the effect of varying temperatures
on the flowering and fruit set in three varieties of groundnut. Neth. J. Agric. Sci. 7, 317–326.
Boote, K. J., and Pickering, N. B. (1994). Modeling photosynthesis of row crop canopies.
HortScience 29, 1423–1434.
Boote, K. J., Varnell, R. J., and Duncan, W. G. (1976). Relationships of size, osmotic
concentration, and sugar concentration of peanut pods to soil water. Soil Crop Sci. Soc.
Fla. Proc. 35, 47–50.
Boote, K. J., Jones, J. W., Mishoe, J. W., and Wilkerson, G. G. (1986). Modeling growth
and yield of groundnut. In “Agrometeorology of Groundnut: Proceedings of an Interna-
tional Symposium, 21–26 Aug 1985, ICRISAT Sahelian Center, Niamey, Niger,”
pp. 243–254. ICRISAT, Patancheru, A.P., India.
Boote, K. J., Jones, J. W., and Singh, P. (1991). Modeling growth and yield of groundnut—
State of the art. In “Groundnut—A Global Perspective: Proceedings of an International
Workshop, 25–29 Nov 1991,” pp. 331–343. ICRISAT Center, India.
Boote, K. J., Jones, J. W., Hoogenboom, G., and Pickering, N. B. (1998). The CROPGRO
model for grain legumes. In “Understanding Options for Agricultural Production”
(G. Y. Tsuji, G. Hoogenboom, and P. K. Thornton, Eds.), pp. 99–128. Kluwer
Academic Publishers, Dordrecht.
Boote, K. J., Allen, L. H. Jr., Vara Prasad, P. V., and Jones, J. W. (2010). Testing effects of
climate change in crop models. In “Handbook of Climate Change and Agroecosystems”
(D. Hillel and C. Rosenzweig, Eds.), pp. 109–129. Imperial College Press, London UK.
Chen, J. J., and Sung, J. M. (1990). Gas exchange rate and yield response of Virginia-type
peanut to carbon dioxide enrichment. Crop Sci. 30, 1085–1089.
Clifford, S. C., Stronach, I. M., Mohamed, A. D., Azam-Ali, S. N., and Crout, N. M. J. (1993).
The effects of elevated atmospheric carbon dioxide and water stress on light interception,
dry matter production and yield in stands of groundnut (Arachis hypogaea L.). J. Exp. Bot. 44,
1763–1770.
Clifford, S. C., Black, C. R., Roberts, J. A., Stronach, I. M., Singleton-Jones, P. R.,
Mohamed, A. D., and Azam-Ali, S. N. (1995). The effect of elevated atmospheric
CO2 and drought on stomatal frequency in groundnut (Arachis hypogaea L.). J. Exp.
Bot. 46, 847–852.
Cox, F. R. (1979). Effect of temperature treatment on peanut vegetative and fruit growth.
Peanut Sci. 6, 114–117.
Craufurd, P. Q., Prasad, P. V. V., and Summerfield, R. J. (2002). Dry matter production and
rate of change of harvest index at high temperature in peanut. Crop Sci. 42, 146–151.
Dreyer, J., Duncan, W. G., and McClaud, D. E. (1981). Fruit temperature growth and yield
of peanut. Crop Sci. 21, 686–688.
Flohr, Marie-Luise, Williams, J. H., and Lenz, F. (1990). The effect of photoperiod on
the reproductive development of a photoperiod sensitive groundnut (Arachis hypogea L.)
cv. NC Ac 17090. Exp. Agric. 26, 397–406.
Fortanier, E. J. (1957). Control of flowering in Arachis hypogaea L. PhD. Thesis. Mededelin-
gen van de Landouwhoogexhool te Wageningen, The Netherlands.
Golombek, S. D., and Johansen, C. (1997). Effect of soil temperature on vegetative and
reproductive growth and development in three Spanish genotype of peanut (Arachis
hypogaea L.). Peanut Sci. 24, 67–72.
Hardy, R. W. F., and Havelka, U. D. (1977). Possible routes to increase the conversion of
solar energy to food and feed by grain legumes and cereal grains (crop production): CO2
and N fixation, foliar fertilization, and assimilate partitioning. In “Biological Solar Energy
Conversion” (A. Mitsui, et al., Eds.), pp. 299–322. Academic Press, New York.
Hertog, L. D., Stulen, I., and Lambers, H. (1993). Assimilation, respiration and allocation of
carbon in Plantago major as affected by atmospheric CO2 Levels—A case study. In “CO2
and Biosphere” ( J. Rozema, et al., Eds.), pp. 369–378. Kluwer Acadamic Publishers,
Belgium.
Effect of Climate Change Factors on Processes of Crop Growth 67

ICRISAT Annual Report (1981). Published 1982, p. 190.

Idso, K. E., and Idso, S. B. (1994). Plant responses to atmospheric CO2 enrichment in the
face of environmental constraints: A review of the past 10 years’ research. Agric. Forest
Meteorol. 69, 153–203.
IPCC (Intergovernmental Panel on Climate Change). (2001). Climate change 2001: The
scientific basis.( J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden,
X. Dai, K. Maskell, and C. A. Johnson, Eds.), In “Contribution of Working Group I to
the Third Assessment Report of the Intergovernmental Panel on Climate Change”.
p. 881. Cambridge University Press, Cambridge.
IPCC (Intergovernmental Panel on Climate Change). (2007). Climate change 2007: The
physical science basis.(S. Solomon, et al., Eds.), In “Contribution of Working Group I to
the Fourth Assessment Report of the Intergovernmental Panel on Climate Change”.
p. 996. Cambridge University Press, Cambridge, UK.
Ketring, D. L. (1984). Temperature effects on vegetative and reproductive development
of peanut. Crop Sci. 24, 877–882.
Lee, T. A. Jr., Ketring, D. L., and Powell, T. D. (1972). Flowering and growth response of
peanut plants (Arachis hypogaea L. var. starr) at two levels of relative humidity. Plant
Physiol. 49, 190–193.
Lenssen, G. M., and Rozema, J. (1990). The effect of atmospheric CO2 enrichment and
salinity on growth, photosynthesis and water relation of salt marsh species. In “The
Greenhouse Effects and Primary Productivity in European Agro-Ecosystems”
( J. Gouriaan, H. Van Keullen, and H. H. Va Laar, Eds.), pp. 64–67. Prodoc,
Wageningen.
Leong, S. K., and Ong, C. K. (1983). The influence of temperature and soil water deficit on
the development and morphology of groundnut (Arachis hypogaea L.). J. Exp. Bot. 34(1),
1551–1561.
Long, S. P. (1991). Modification of the response of photosynthetic productivity to rising
temperature by atmospheric CO2 concentrations: Has its importance been underesti-
mated? Plant Cell Environ. 14, 729–739.
Marshall, B., Squire, G. R., and Terry, A. C. (1992). Effect of temperature on interception
and conversion of solar radiation by stands of groundnut. J. Exp. Bot. 43(246), 95–101.
Martin, J. P., Cas, S., and Rabechault, H. (1974). Cultures in vitro d etamines arachide
(Arachis hypogea L.). 1. Stades du development des boutons floraux et microsporogenesis.
Oleagineux 29, 145–149.
Matlock, R. S., Garton, J. E., and Stone, J. F. (1961). Peanut irrigation studies in Oklahoma,
1956–1959. “Oklahoma Agricultural Experiment Station Bulletin”, p. 19. Oklahoma
State University, Stillwater, Oklahoma, No. B-580.
McMurtrie, R. E., and Wang, Y. P. (1993). Mathematical models of the photosynthetic
response of tree stands to rising CO2 concentrations and temperatures. Plant Cell Environ.
16, 1–13.
McMurtrie, R. E., Comins, H. N., Kirschbaum, M. U. F., and Wang, Y. P. (1992).
Modifying existing forest growth models to take account of effects of elevated CO2.
Aust. J. Bot. 40, 657–677.
Mohamed, H. A., Clark, J. A., and Ong, C. K. (1988). Genotypic differences in the
temperature responses of tropical crops I. Germination characteristics of groundnut
(Arachis hypogaea L.) and pearl millet (Pennisetum typhoides S & H). J. Exp. Bot. 39,
1121–1128.
Morison, J. I. L. (1985). Sensitivity of stomata and water use efficiency to high CO2. Plant
Cell Environ. 8, 467–474.
Nigam, S. N., Rao, R. C. N., Wynne, J. C., Williams, J. H., Fitzner, M., and
Nagabhushanam, G. V. S. (1994). Effect and interaction of temperature and photoperiod
on growth and partitioning in three groundnut (Arachis hypogaea L.) genotypes. Ann.
Appl. Biol. 125, 541–552.
68 Uttam Kumar et al.

Ong, C. K. (1984). The influence of temperature and water deficits on the partitioning of
dry matter in groundnut (Arachis hypogaea L.). J. Exp. Bot. 35, 746–755.
Ono, Y. (1979). Flowering and fruiting of peanut plants. Jpn. Agric. Res. Q. 13, 226–229.
Ono, Y., Nakayama, K., and Kubota, M. (1974). Effects of soil temperature and soil
moisture in podding zone on pod development of peanut plants. Proc. Crop Sci. Soc.
Jpn. 43, 247–251.
Pallas, J. E. Jr., Stansell, J. R., and Koske, T. J. (1979). Effect of drought on florunner
peanuts. Agron. J. 71, 853–858.
Pilumwong, J., Senthonga, C., Srichuwongb, S., and Ingram, K. T. (2007). Effects of
temperature and elevated CO2 on shoot and root growth of peanut (Arachis hypogaea L.)
grown in controlled environment chambers. Sci. Asia 33, 79–87.
Prasad, P. V. V., Craufurd, P. Q., and Summerfield, R. J. (1999a). Sensitivity of peanut to
timing of heat stress during reproductive development. Crop Sci. 39, 1352–1357.
Prasad, P. V. V., Craufurd, P. Q., and Summerfield, R. J. (1999b). Fruit number in relation
to pollen production and viability in groundnut exposed to short episodes of heat stress.
Ann. Bot. 84, 381–386.
Prasad, P. V. V., Craufurd, P. Q., Summerfield, R. J., and Wheeler, T. R. (2000a). Effects of
short episodes of heat stress on flower production and fruit-set of groundnut (Arachis
hypogaea L.). J. Exp. Bot. 51, 777–784.
Prasad, P. V. V., Craufurd, P. Q., and Summerfield, R. J. (2000b). Effect of high air and soil
temperature on dry matter production, pod yield and yield components of groundnut.
Plant Soil 222, 231–239.
Prasad, P. V. V., Boote, K. J., Allen, L. H. Jr., and Thomas, J. M. G. (2003). Super-optimal
temperatures are detrimental to peanut (Arachis hypogaea L.) reproductive processes and
yield at both ambient and elevated carbon dioxide. Glob. Chang. Biol. 9, 1775–1787.
Prasad, P. V. V., Boote, K. J., Thomas, J. M. G., Allen, L. H. Jr., and Gorbet, D. W. (2006).
Influence of soil temperature on seedling emergence and early growth of peanut cultivars
in field conditions. J. Agron. Crop Sci. 192, 168–177.
Pressman, E., Peet, M. M., and Pharr, M. (2002). The effect of heat stress on tomato pollen
characteristic is associated with changes in carbohydrate concentration in the developing
anthers. Ann. Bot. 90, 631–636.
Rao, K. V. (1999). The combined effect of elevated CO2 levels and temperature on growth
characteristics of groundnut (Arachis hypogaea L.). Indian J. Plant Physiol. 4, 297–301.
Rao, R. C. N., Singh, S., Sivakumar, M. V. K., Srivastava, K. L., and Williams, J. H. (1985).
Effect of water deficit at different growth of peanut. I. yield responses. Agron. J. 77, 782–786.
Singh, P., Boote, K. J., and Virmani, S. M. (1994a). Evaluation of the groundnut model
PNUTGRO for crop response to plant population and row-spacing. Field Crop Res. 39,
163–170.
Singh, P., Boote, K. J., Yogeswara Rao, A., Iruthayaraj, M. R., Sheikh, A. M.,
Hundal, S. S., Narang, R. S., and Singh, Phool (1994b). Evaluation of the groundnut
model PNUTGRO for crop response to water availability, sowing dates and seasons.
Field Crop Res. 39, 147–162.
Stronach, I. M., Clifford, S. C., Mohamed, A. D., Singleton-Jones, P. R., Azam-Ali, S. N., and
Crout, N. M. J. (1994). The effect of elevated carbon dioxide, temperature and soil moisture
on the water use of stands of groundnut (Arachis hypogeae L.). J. Exp. Bot. 45, 1633–1638.
Stuhlfauth, T., and Fock, H. P. (1990). Effect of whole season CO2 enrichment on the
cultivation of a medicinal plant, Digitalis lanata. J. Agron. Crop Sci. 164, 168–173.
Suzuki, M. (1966). Studies on thermoperiodicity of crops. II. The effects of soil temperature
on fructification of peanuts. Chiba Univ. Tech. Bull. 13, 95–101.
Suzuki, K., Takeda, H., Tsukaguchi, T., and Egawa, Y. (2001). Ultrastructural study of
degeneration of tapetum in anther of snap bean (Phaseolus vulgaris L.) under heat-stress.
Sex. Plant Reprod. 13, 293–299.
Effect of Climate Change Factors on Processes of Crop Growth 69

Talwar, H. S. (1997). Physiological Basis for Heat Tolerance During Flowering and Pod
Setting Stages in Groundnut (Arachis hypogaea L.). JIRCAS Visiting Fellowship Report
1996–97 JIRCAS, Okinawa.
Talwar, H. S., Takeda, H., Yashima, S., and Senboku, T. (1999). Growth and photosyn-
thetic responses of groundnut genotypes to high temperature. Crop Sci. 39(2), 460–466.
Underwood, C. V., Taylor, H. M., and Hoveland, C. S. (1971). Soil physical factors
affecting peanut pod development. Agron. J. 63, 953–954.
Weiss, E. A. (2000). Oilseed Crops. Blackwell Science, London.
Williams, J. H., and Boote, K. J. (1995). Physiology and modelling—Predicting the unpre-
dictable legume. In “Advances in Peanut Science” (H. E. Pattee and H. T. Stalker, Eds.),
pp. 301–335. APRES, Stillwater, Oklahoma.
Williams, J. H., Wilson, J. H., and Bate, G. C. (1975). The growth of groundnuts (Arachis
hypogaea L. cv. Makulu Red) at three altitudes. Rhodosian J. Agric. Res. 13, 33–43.
Wood, I. M. W. (1968). The effect of temperature at early flowering on the growth and
development of peanuts (Arachis hypogaea). Aust. J. Agric. Res. 19, 241–251.
Xi, X. Y. (1991). Development and structure of pollen and embryo sac in peanut (Arachis
hypogaea L.). Bot. Gaz. 152, 164–172.
Yamagata, M., Kouchi, H., and Yoneyama, T. (1987). Partitioning and utilization of photo-
synthate produced at different growth stages after anthesis in soybean (Glycine max L.
Merr.): Analysis by long term 13C-labelling experiments. J. Exp. Bot. 38, 1247–1259.
 C H A P T E R T H R E E

Agricultural Practices in Oil Palm

Plantations and Their Impact on
Hydrological Changes, Nutrient
Fluxes and Water Quality in
Indonesia: A Review
Irina Comte,*,† François Colin,‡ Joann K. Whalen,*
Olivier Grünberger,§ and Jean-Pierre Caliman}

Contents
1. Introduction 72
2. Expansion of Oil Palm Cultivation in Indonesia and
Environmental Stakes 74
2.1. Expansion of oil palm cultivation 74
2.2. Environmental stakes 75
3. Oil Palm Cultivation 81
3.1. Climate and soil conditions 81
3.2. Production systems: Industrial versus smallholder plantations 82
3.3. Land clearing and site preparation 82
3.4. Water and soil management 83
3.5. Nutrient-demand assessment 85
3.6. Fertilizer management 90
3.7. Synthesis 92
4. Hydrological Processes and Associated Nutrient Transfers in Oil
Palm Plantations 93
4.1. Precipitation in Indonesia 94
4.2. Interception 94
4.3. Evapotranspiration 95

* Department of Natural Resource Sciences, Macdonald Campus of McGill University, Ste-Anne-de-Bellevue,

Quebec, Canada
{
CIRAD (International Cooperation Centre in Agronomic Research for Development), Montpellier cedex,
France
{
Montpellier-SupAgro, UMR-LISAH (Laboratory on Interactions between Soil, Agrosystem and Hydro-
system), Montpellier cedex, France
}
IRD (Institut de Recherche pour le Développement), UMR-LISAH, Montpellier cedex, France
}
PT SMART Research Institute (SMARTRI), Pekanbaru, Riau, Indonesia

Advances in Agronomy, Volume 116 # 2012 Elsevier Inc.

ISSN 0065-2113, DOI: 10.1016/B978-0-12-394277-7.00003-8 All rights reserved.

71
72 Irina Comte et al.

4.4. Soil infiltration, leaching, and groundwater quality 96

4.5. Surface runoff and erosion 103
4.6. Stream flow and stream water quality 108
4.7. Synthesis 111
5. Conclusion 113
Acknowledgments 114
References 114

Abstract
Rapid expansion of oil palm (Elaeis guineensis Jacq.) cultivation in Southeast Asia
raises environmental concerns about deforestation and greenhouse gas emis-
sions. However, less attention was paid to the possible perturbation of hydrolog-
ical functions and water quality degradation. This work aimed to review (i) the
agricultural practices commonly used in oil palm plantations, which potentially
impact hydrological processes and water quality and (ii) the hydrological changes
and associated nutrient fluxes from plantations. Although many experimental
trials provide clear recommendations for water and fertilizer management, we
found that few studies investigated the agricultural practices actually followed by
planters. Our review of hydrological studies in oil palm plantations showed that
the main hydrological changes occurred during the first years after land clearing
and seemed to dissipate with plant growth, as low nutrient losses were generally
reported from plantations. However, most of those studies were carried out at the
plot scale and often focus on one hydrological process at a single plantation age.
So, there is insufficient information to evaluate the spatiotemporal fluctuations in
nutrient losses throughout the entire lifespan of a plantation. Furthermore, few
studies provided an integrated view at the watershed scale of the agricultural
practices and hydrological processes that contribute to nutrient losses from oil
palm plantations and the consequences for surface and groundwater quality.
Future research efforts need to understand and assess the potential of oil palm
plantations to change hydrological functions and related nutrient fluxes, consid-
ering agricultural practices and assessing water quality at the watershed scale.

1. Introduction
Oil palm (Elaeis guineensis) is one of the most rapidly expanding crops
in the tropics. Since the early 1980s, the global land area under oil palm
production has more than tripled, reaching almost 15 million ha in 2009 and
accounting for almost 10% of the world’s permanent crop land (FAOSTATS,
2011; Sheil et al., 2009) Most of this increase has taken place in Southeast Asia.
Together, Malaysia and Indonesia account for almost 85% of the 46.5 million
tons of crude oil palm produced in the world, Indonesia being the top
producer since 2007 (Oil World, 2011; USDA, 2007). The area covered by
smallholder plantations in Indonesia increased nearly 1000-fold between 1979
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 73

and 2008, reaching almost 3 million ha, representing 39% of Indonesian oil
palm plantations currently, the remaining 4.5 million ha being large private
(52%) and government-owned (8%) plantations (IMA, 2010).
Although oil palm cultivation is a strong driver of economic development
in Indonesia, providing jobs and incomes to millions of people (USDA, 2007),
it is strongly denigrated for its environmental impacts. Many media and NGOs
accuse oil palm plantation development in Southeast Asia of triggering defor-
estation, loss of biodiversity, peatland degradation, and high greenhouse gas
(GHG) emissions (Greenpeace, 2011; WWF, 2011). In the scientific commu-
nity, there is controversy about the positive and negative aspects of the
expanding oil palm cultivation and potential environmental risks, which has
been discussed at length in the scientific literature (Basiron, 2007; Lamade and
Bouillet, 2005; Nantha and Tisdell, 2009; Sheil et al., 2009). The development
of oil palm plantations, which frequently cover tens of square kilometers in
Southeast Asia, involves land clearing, road and drainage network construc-
tion, and sometimes earthworks such as terracing on undulating areas. The use
of agrochemicals, such as fertilizers and pesticides, might represent a potential
risk for the sustainability of aquatic ecosystem and hydrological functions
when agricultural practices are not optimized. In particular, oil palm growers
usually apply large amounts of commercial fertilizer and thus are among the
largest consumers of mineral fertilizers in Southeast Asia (Härdter and
Fairhurst, 2003). However, hydrological processes within oil palm plantations
are still not fully understood, and few studies have examined the impacts of
agricultural practices on terrestrial hydrological functions and water quality in
nearby aquatic ecosystems (Ah Tung et al., 2009), although aspects that impact on
water quality are by far the largest component of an environmental risk register accounting
for nearly 50% of all entries in oil palm plantations (Lord and Clay, 2006).
This review aims to document the current state-of-knowledge of agricul-
tural practices in oil palm plantations that potentially impact hydrological
functions and water quality in surface waters, with a focus on nutrient loading
of surface waterways, and to highlight research gaps in the understanding of
these processes. This work focuses on the situation in Indonesia, with examples
from other oil palm producing countries in the humid tropics as appropriate.
First, the expansion of oil palm cultivation in Indonesia, relevant environmen-
tal issues, and polemics will be presented. Next, typical agricultural practices in
industrial and smallholder oil palm plantations will be discussed, focusing on
nutrient, soil, and water management. Finally, the last section gives the state-of-
the-art knowledge of hydrological changes and associated nutrient fluxes from
oil palm plantations compared to tropical rainforests, which were the dominant
natural ecosystem prior to oil palm plantation establishment. Relevant pro-
cesses in the hydrological cycle and their magnitude and relevance in oil palm
plantations will be explained in this section, but we do not provide an in-depth
discussion of hydrological processes in rainforests, as a number of reviews were
already published on this topic (Bruijnzeel, 1991, 2004; Elsenbeer, 2001).
74 Irina Comte et al.

2. Expansion of Oil Palm Cultivation in

Indonesia and Environmental Stakes
2.1. Expansion of oil palm cultivation
2.1.1. Palm oil utilization
Palm oil is derived from the plant’s fruit, which produces two types of oils:
crude palm oil (CPO), which comes from the mesocarp of the fruit, and palm
kernel oil, which comes from the seed in the fruit. Most CPO is used for food
products, while most palm kernel oil is used in nonedible products such as
detergents, cosmetics, plastics, as well as a broad range of other industrial and
agricultural chemicals (Wahid et al., 2005). The oil palm is the most produc-
tive oil crop in terms of oil yield per hectare and resource use efficiency due to
its high efficiency at transforming solar energy into vegetable oil. The average
yield of palm oil is approximately 4.2 tha
1 year
1, with yields exceeding 6.0t
ha
1 year
1 in the best-managed plantations, greatly exceeding vegetable oils
such as rapeseed and soybean that produce only 1.2 and 0.4tha
1 year
1,
respectively (Fairhurst and Mutert, 1999). In addition, little fossil fuel energy
is used, as most of the energy required by the oil palm mill for processing of the
fruits is provided by burning the palm by-products (shells and fibers). Conse-
quently, the energy balance, expressed as the ratio of outputs to inputs, is
higher for oil palm (9.6) than other commercially grown oil crops (e.g.,
rapeseed: 3.0; soybean: 2.5), making oil palm the most attractive candidate
for biofuel production (Fairhurst and Mutert, 1999).

2.1.2. Extent of oil palm cultivation in Indonesia: 1911 to present

The first commercial plantation was developed in Sumatra in 1911 and the
area planted in Indonesia increased from about 31,600ha by 1925 to 7.3
million ha by 2008 (Corley and Tinker, 2003; IMA, 2010). Since 2007,
Indonesia has been the world’s largest and most rapidly growing producer.
Its production rose from 168,000 tons in 1967 to 22 million tons by 2010
(IMA, 2010). CPO and kernel oil prices have been rising, encouraging
investors to develop plantations on the large areas of suitable land in the
islands of Sumatra, Indonesia (Fig. 1) and elsewhere, mainly on the island of
Borneo (USDA, 2007).

2.1.3. Expected future expansion of oil palm cultivation

Continued expansion of oil palm plantations is forecast due to growing
global demand for palm oil as a source of fats and oil for human consump-
tion, nonedible products, and biofuel to keep pace with human population
growth, expected to reach 8.9 billion in 2050 (Bangun, 2006; Tan et al.,
2009; UN, 2004). Present plans are to increase production up to 40 million
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 75

Figure 1 Map of Indonesian palm oil production in 2007/2008 (from USDA, 2009).

tons of CPO by 2020 (IMA, 2010; Rist et al., 2010). According to USDA
(2007), the availability of land in Indonesia, coupled with other factors—
high seed sales, record energy prices, and high vegetable oil prices—ensures
that Indonesia will continue to lead the world in palm oil production for
years to come. However, few developments generate as much controversy
as the rapid expansion of oil palm in developing countries such as Indonesia
(Koh and Wilcove, 2008; Nantha and Tisdell, 2009). Negative conse-
quences reported by environmental groups include deforestation, loss of
biodiversity, peatland degradation, GHG emissions, and water pollution.

2.2. Environmental stakes

2.2.1. Deforestation and loss of biodiversity
The most contentious environmental issue facing the oil palm industry is
deforestation as huge tracts of tropical rainforest are converted to plantations
(Germer and Sauerborn, 2008; Wakker, 1999). Indonesian tropical forested
area ranks third behind Brazil and the Democratic Republic of Congo, and
76 Irina Comte et al.

harbors numerous endemic or rare species (Koh and Ghazoul, 2008; WRI,
2002). Many sources are claiming that virgin tropical forests are being
cleared for oil palm plantations, leading to natural habitat loss for many
endangered species and biodiversity reduction. For instance, it was reported
that Sumatran orangutans (Pongo abelii) and Bornean orangutans (Pongo
pygmaeus) face extinction due to plantation expansion (Nantha and
Tisdell, 2009; Nellemann et al., 2007; Tan et al., 2009). Herds of elephants,
tigers, and rhinos are reported to be critically threatened due to this
expansion (Danielsen et al., 2008; WRI, 2002). Studies in oil palm frontier
areas on the island of Sumatra concluded that oil palm plantations result
in a significant reduction in biodiversity if plantations replace natural forests,
secondary forests, agroforests, or even degraded forests and scrubby
unplanted areas (Gillison and Liswanti, 1999; Sheil et al., 2009). However,
others mentioned that the expansion of oil palm plantations is only one of
the factors contributing to herd displacement and local extinction, as other
anthropogenic activities like illegal logging, forest fires, and illegal hunting
are also problematic for large mammals (Nellemann et al., 2007; Tan et al.,
2009). According to Koh and Ghazoul (2008), at least 56% of the oil palm
expansion in Indonesia during the period 1990–2005 occurred at the
expense of primary, secondary, or plantation forests, and 44% was on
cropland area. Deforestation in Southeast Asia cannot be attributed solely
to oil palm production. According to the World Rainforest Movement
(WRM, 2002), the immediate causes of rainforest destruction in Southeast
Asian countries are logging by commercial companies, shifting agriculture,
monoculture plantations (e.g., rubber in Thailand), cattle ranching, fuel-
wood harvesting, hydroelectric dams, mining and oil exploitation, and
colonization schemes.

2.2.2. Peatland degradation

2.2.2.1. Peatland formation Southeast Asia has an estimated 27.1 million
ha of peatlands, most of which are located in Indonesia (22.5million ha),
representing 12% of its land area (Hooijer et al., 2006). Peatlands develop
in depressions or wet coastal areas when the rate of biomass deposition
from adapted vegetation (i.e., mangroves, swamp forest) is greater than the
rate of decomposition. The accumulation of organic matter that degrades
very slowly, over a period of hundreds of years, makes peat soil. This is due
to the presence of a permanently high water table that prevents aerobic
microorganisms from decomposing the plants debris (Mutert et al., 1999;
Siegel and Glaser, 2006). A soil is considered to be peat when it includes an
organic layer thicker than 40–50cm (USDA, 2006). In Southeast Asia, all
low-lying peatlands are naturally forested with an average canopy height of
40m and emergent trees of up to 50m (WI, 2010). In the Eastern coast of
the island of Sumatra, Indonesia, peat deposits are usually at least 50cm thick
but can form a deep profile that extends up to 20m (CAAL, 2011).
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 77

2.2.2.2. Peatland ecological functions Peatlands regulate water flow by

capturing rainwater during the wet season and slowly releasing it, over a
period of months, during the dry season. Consequently, peatlands help to
prevent floods and droughts (Clark et al., 2002; Tan et al., 2009). In
addition, peatlands are an important carbon (C) sink in the global C cycle
because they cover nearly 3% (some 4 million km2) of the earth’s land area
and store about 528,000 million tons C, which is equivalent to one-third of
global soil C and 70 times annual global emissions from fossil fuel burning
(Hooijer et al., 2006; Tan et al., 2009). Peatland attributes also include
biological diversity, since tropical peatlands are important genetic reservoirs
of certain animals and plants. Tropical peatlands have long-provided goods
and services for local communities to fulfill their daily, basic requirements,
for example, hunting grounds and fishing areas, food, medicines, and
construction materials (Rieley, 2007).

2.2.2.3. Peatland conversion to oil palm Peat swamp forests have

remained relatively undisturbed until recently, as they were unattractive
for agriculture. But the increasing international demand for biofuel and
the current lack of available land on mineral soils has accelerated the
conversion of peatlands to oil palm plantations especially in Indonesia
(Kalimantan, Sumatra, and West Papua) where nearly 25% of all oil palm
plantations are located on peatlands (Sheil et al., 2009; Tan et al., 2009).
However, oil palms cannot survive in undrained waterlogged peatlands.
Drainage for oil palm growth in peatlands is installed between 40 and 80cm
depth, but the water table could recede below 80cm during an extended
drought (WI, 2010). Many authors reported that there is a direct relation-
ship between the depth of the water table and the rate of peat subsidence
and thus the sustainability of the peat (Strack and Waddington, 2007;
Wösten et al., 1997, 2008). Drainage results in rapid peat subsidence and
compaction, leading to various changes in its physical properties including
greater bulk density and less total porosity, oxygen diffusion, air capacity,
available water volume, and water infiltration rate (Rieley et al., 2007).
Drainage ultimately destroys the “sponge effect” of peat swamps and their
reservoir function (Andriesse, 1988). In addition, the exploitation or removal
of the overlying forest resource further reduces the ability of the ecosystem to
hold rainfall, and water is flushed more quickly into the rivers, increasing
flooding in the rainy season and drought in the dry season (Rieley, 2007;
Rieley and Page, 1997). Moreover, the drainage of C-rich peatlands leads to
aeration of the peat material and hence to the oxidation (or aerobic decom-
position) of peat material resulting in massive CO2 gas emissions to the
atmosphere (Hooijer et al., 2006; Schrevel, 2008). Although the exploitation
of peat swamp forests also provides employment, local income, new jobs, and
business opportunities, contributing to poverty alleviation of the country, it is
at the expense of the ecosystem and the environment (Noor et al., 2007).
78 Irina Comte et al.

2.2.3. GHG emissions and carbon storage

As mentioned in previous sections, establishment of oil palm plantations
requires deforestation or peatland conversion, which irreversibly alters the
GHG balance. Before 1998, most of the deforestation in Southeast Asia
involved burning, which caused numerous, large, and persistent fires and
consecutive GHG emissions (Tan et al., 2009). In 1997–1998, these fires
were particularly devastating due to a severe drought caused by the
El-Niño climatic phenomenon. In Indonesia, the total area damaged or
destroyed by the 1997–1998 fires was estimated at nearly 10million ha and
the overall economic cost of fire and haze in the region at $9 billion
(Glastra et al., 2002). Globally, peatlands emit 2000Mg CO2 equivalents
(CO2eq) year
1, equal to almost 8% of global emissions from fossil fuel
burning, and more than 90% of this annual emission originates from
Indonesia. Due to GHG emissions from forest burning and peatland
conversion, Indonesia is in fourth place in the global CO2 emission
ranking (Barnett, 2007).
Nevertheless, some authors suggested that oil palm plantations may act
as a C sink as they assimilate more CO2 and release more oxygen (O2) than
tropical forest (Lamade and Bouillet, 2005). Henson (1999) reported oil
palm CO2 uptake at 25.71tha
1 year
1 and O2 production at 18.70tha
1
year
1 compared to 9.62tha
1 year
1 and 7.00tha
1 year
1, respectively,
in tropical forest. According to several studies reported by Henson (1999),
the total dry matter production in oil palm stands was from 19.1 to 36.5t
ha
1 year
1, and the total dry matter production recorded in a Malaysian
forest reached 25.68tha
1 year
1. While a new oil palm plantation may
grow faster and sequester C at a higher annual rate than a naturally
regenerating forest, ultimately, the oil palm plantation will store less C
(50–90% less over 20years) than the original forest cover (Germer and
Sauerborn, 2008; Henson, 1999). This is due to the wide spacing between
oil palm trees and control of the understory vegetation to avoid competi-
tion with the trees, whereas a natural forest has a plant community that is
stratified horizontally and vertically to maximize primary production on a
given acreage. Germer and Sauerborn (2008) estimated that forest
conversion to oil palm causes a net release of approximately 650Mg
CO2eq ha
1 on mineral soils and more than 1300Mg CO2eq ha
1 on
peatlands, during the first 25-year cycle of oil palm growth (e.g., an
average net release of 26tha
1 year
1 and 52tha
1 year
1 for forest and
peatland conversion, respectively). However, Germer and Sauerborn
(2008) also reported that if tropical grassland (instead of forest or peatlands)
is converted to oil palm plantation, C fixation in plantation biomass and
soil organic matter not only neutralizes emissions caused by grassland
conversion but also results in the net removal of about 135Mg CO2eq
ha
1 from the atmosphere.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 79

2.2.4. Water pollution

Runoff and sedimentation; leaching of nutrients from fertilizer, pesticides,
and other agrochemicals; effluent discharge and sewage from the worker
population, all are potential factors that could affect water quality and can be
significant impacts of oil palm cultivation (ECD, 2000; Lord and Clay,
2006). The consequences of poor water quality will be borne by much of
the Indonesian population: at the beginning of the twenty-first century, at
least 80% of the Indonesian population (250million) had no access to piped
water, while in 2002, 66.2% of the population used river water for washing
and bathing and 22.5% relied on it for drinking water (WEPA, 2011).
Runoff water can transport eroded soil particles from fields to water
bodies. Suspended particles contribute significantly to water turbidity, which
reduces light penetration, impairs photosynthesis, alters oxygen levels, and
reduces the food supply for certain aquatic organisms (Bilotta and Brazier,
2008). Sediment clogs streams, reducing their water-holding capacity, and can
cover spawning beds, destroying fish populations (Kemp et al., 2011).
Mineral fertilizer application can lead to a marked increase in the
nutrient concentrations of water draining from the fertilized areas (ECD,
2000). Of greatest importance for water quality are N and P exported from
agroecosystems to waterways. Nitrogen is mainly applied to agroecosystems
as ammonium sulfate or urea. Both ammonium compounds and urea are
eventually converted into nitrate in the soil under well-drained conditions.
Nitrate in water promotes undesirable growth of aquatic microflora in
surface water bodies, and concentrations exceeding 10mg NO3 l
1 are not
recommended in drinkable water (WHO, 2008). Phosphorous in the form


of orthophosphate (PO3
2

4 , HPO4 , and H2PO4 ) has a similar eutrophica-
tion effect in surface water as nitrate, causing excessive growth of cyano-
bacteria and blocking sunlight and oxygen diffusion to aquatic life in deeper
water (Schindler et al., 1971; Turner and Rabalais, 1994).
Pesticides, including herbicides, are commonly used in oil palm planta-
tions, despite their adverse impacts on human beings and the environment
(Sheil et al., 2009). As rainfall can easily exceed 2500mmyear
1 in Indone-
sia, herbicides can be washed into streams and rivers that are the only water
source for all household needs—including drinking water—in villages
around the plantations and contaminating fishing grounds (DE, 2005).
Yet the risk of water pollution from pesticides originating from oil palm
plantations is probably low, as the Malaysian Environmental Conservation
Department (ECD, 2000) noted that biological control methods can be
quite effective. However, to our knowledge, there has been no study on the
impact of pesticide use on water quality within oil palm plantations.
Finally, palm oil production generates large amounts of waste that can
pollute local waterways when disposed incorrectly. For instance, in 2001,
Malaysia’s production of 7 million tons of CPO generated 9.9 million tons
80 Irina Comte et al.

of solid oil wastes, palm fiber, and shells. Moreover, 10 million tons of palm
oil mill effluent (POME), a polluted mix of crushed shells, water, and fat
residues, was produced and often returned without treatment to natural
watercourses downstream from the mill (Lord and Clay, 2006), leading
to the degradation of the aquatic ecosystems (Briggs et al., 2007; Sheil et al.,
2009). The POME is an acidic colloidal suspension characterized by high
concentration of suspended solids, with a biological oxygen demand (BOD)
of 25,000ppm and a chemical oxygen demand (COD) of 60,000ppm
( Jacquemard, 1995; Olaleye and Adedeji, 2005). According to Olaleye
and Adedeji (2005), riparian rivers and streams receiving untreated mill
effluent are expected to be heavily polluted. Fortunately, technologies
have been developed to treat and reclaim drinking water from the POME
(Ahmad et al., 2006; Rupani et al., 2010; Singh et al., 2010; Yi Jing et al.,
2010), and they need to be widely implemented to protect downstream
waters. Sewage from the worker population in the plantation is another
waste byproduct that is expected to elevate COD, BOD, and fecal coliform
levels in waterways.

2.2.5. Agricultural policies

The pressing environmental and social issues associated with the palm oil
industry were the impetus for dialog between palm oil stakeholders and
NGO representatives regarding methods to achieve sustainable palm oil
production and use, including better management practices (BMP) for
agronomists and planters working with this industry (Darussamin, 2004).
Thus, the Roundtable on Sustainable Palm Oil (RSPO), an international
organization of producers, distributors, environmental NGOs, and social
NGOs, was created and defines sustainable palm oil production as a legal,
economically viable, environmentally appropriate, and socially beneficial management
and operations (Tan et al., 2009). RSPO has recently established a set of
principles and criteria for the management of oil palm plantations and palm
oil mills and encourages the planters to follow BMP (Lord and Clay, 2006).
These practices are environment-friendly approaches like zero burning for
land clearing, conservation of wildlife and habitat, integrated pest manage-
ment (IPM), and waste minimization and recycling. For example, IPM in
plantations relies on barn owls or snakes to reduce rat populations instead of
pesticides (Hansen, 2007; Sheil et al., 2009). Planting leguminous cover crops
(LCC) to minimize soil erosion and maintain soil fertility and the recycling of
palm oil empty fruit bunches (EFB) and POME as fertilizer in the plantation
are also promoted. These environment-friendly practices could reduce the
use of mineral fertilizers (Tan et al., 2009) and improve water quality.
The Indonesian government also has recently started to encourage
planters to improve the sustainability of their plantations. The Indonesian
Sustainable Palm Oil (ISPO) was launched in 2010 by the government
of Indonesia and will be compulsory in the coming years for all oil palm
growers, while RSPO is still a voluntary process. In addition, Indonesia
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 81

government has recently edited a guideline on oil palm cultivation on

peatland “to manage the peatland in a sustainable way considering the
ecological function of peatland.” It stipulates, among other things, that the
land clearing will be done without burning (since 1998, zero burning is
compulsory in Indonesia) and an intensive and quick process of drying is not
allowed, to avoid irreversible shrinkage. The decree also lays down that oil
palm cultivation on peatland must be restricted to (a) areas in which the peat
extended less than 3m below the surface; (b) areas where the subsoil under
the peatland is not silica sand or acid sulfate soil to avoid the toxic effects
of the oxidation of the pyrite and accumulation of sulfuric acid, due to
the drop in water table level (Zaidel’man, 2008); (c) areas with mature
peat soils, classified as sapric (the most decomposed) or hemic (somewhat
decomposed); and (d) areas with eutrophic peatlands (Ministerial Decree on
Agriculture, 2009).

2.2.6. Implications for future research

To date, most of the attention and research related to the environmental
impact of oil palm cultivation has focused on deforestation and its conse-
quences (loss of biodiversity, GHG emissions from burning, etc.) during the
initial phase of oil palm plantation establishment. Much less attention has
been paid to the environmental impacts of established oil palm plantations,
particularly hydrological changes and water pollution. Understanding and
assessing the activities in oil palm plantations that impact hydrological
functions and associated nutrient fluxes requires a good knowledge of all
agricultural practices followed throughout the entire life cycle of the plan-
tation. In the next section, we describe the preferred growing conditions for
oil palm (climate, soils) and agronomic considerations (site preparation, tree
spacing), with emphasis on nutrient, soil, and water management in differ-
ent production systems (industrial and smallholders).

3. Oil Palm Cultivation

3.1. Climate and soil conditions
The African oil palm (Elaeis guineensis, family Arecaceae) is a tropical forest
palm native to West and Central African forests. Oil palm needs humid
equatorial conditions (1780–2280mm annual rainfall and a temperature
range of 24–30  C) to thrive, and so conditions in Southeast Asia are ideal
(Corley and Tinker, 2003). Palm productivity benefits from direct sunshine:
the lower incidence of cloud cover over much of Southeast Asia is thought
to be one reason why oil palm yields are higher there than in West Africa
(Dufrêne et al., 1990). Oil palm is tolerant of a wide range of soil types, as
long as it is well watered. Seasonal droughts at higher tropical latitudes
82 Irina Comte et al.

greatly reduce yields (Basiron, 2007). The oil palm is cultivated predomi-
nantly on tropical soils in the orders Ultisol, Oxisol, and Inceptisol. These
soils are highly acidic with low buffering capacities (Ng, 2002) as a conse-
quence of cation leaching (Caliman et al., 1987). Once the pH drops below
5.5, aluminum and manganese compounds start to dissolve, which may
cause root deterioration (Godbold et al., 1988). However, oil palm is
adapted to acidic conditions (Omoti et al., 1983), and with appropriate
management, oil palm plantations can also be productive on “problem soils”
such as acid sulfate soils, deep peat and acidic high aluminum soils, where
few other crops are successful (Corley and Tinker, 2003).

3.2. Production systems: Industrial versus smallholder

plantations
Oil palm-growing areas in Indonesia are distributed among three production
systems: government holdings, private companies, and smallholders. In 2010,
the Indonesian Ministry of Agriculture estimated that of almost 8million ha
under oil palm cultivation, private companies held 53%, smallholders had 39%,
and the remaining 8% belonged to government plantations. Private and gov-
ernmental estates typically range in size from 3000 to 20,000ha (Sheil et al.,
2009), while the smallholders are defined as family-based enterprises producing
palm oil from less than 50ha of land, often about 2ha (Vermeulen and Goad,
2006). When the price of palm oil in the international market was exceptionally
high (around US$ 700 per ton in 1974), efforts were made to increase
production. The government established a strategy called the nucleus estate
scheme (NES), where state-owned or private plantation companies (nucleus)
helped smallholder farmers to grow oil palm on 2
3ha of land in the surround-
ing area (plasma) (Bangun, 2006; IEG, 1993; Zen et al., 2005). Smallholders
working under contract to the plantation companies received seedlings of high-
yielding cultivars, technical assistance for land preparation and planting, agro-
chemical inputs (fertilizers and pesticides), management assistance, and loan
access (Bangun, 2006; Vermeulen and Goad, 2006). The estates benefited
through their fees for services and return from milling smallholder fruit into
CPO (Zen et al., 2005). In contrast to smallholders working under the NES,
independent smallholders cultivate oil palm without direct assistance from
government or private companies. They sell their crop to local mills directly
or through buyers (Vermeulen and Goad, 2006).

3.3. Land clearing and site preparation

Many authors provide recommendations for the establishment of an oil
palm plantation including land clearing, road and drainage network con-
struction, caring for prenursery and nursery stages, planting, and manage-
ment of immature and mature trees until replanting. Among the most
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 83

popular agronomic handbooks for oil palm cultivation are Corley and
Tinker (2003), Jacquemard (1995), Rankine and Fairhurst (1998a,b,c), and
Hartley (1988).
Once the site is selected, establishment of the oil palm plantation starts
with land clearing. At present, mechanical methods are used in all major oil
palm-growing countries with chainsaws, winches, and bulldozers. Then,
the felled vegetation is either burnt or allowed to rot. The issue of whether
or not to burn the felled vegetation has remained a subject of controversy
for years (Corley and Tinker, 2003) due to environmental impacts of
burning: smoke, haze, and large nutrient losses through volatilization and
ash carried away (Mackensen et al., 1996). Since the massive fires in
Kalimantan and Sumatra in 1997, the Indonesian government prohibited
burning (Corley and Tinker, 2003). However, according to some media,
laws to limit agricultural burning are poorly enforced and burning still
continues (Mongabay Editorial, 2006).
The general layout of a plantation is decided by the topography, the
drainage, the position of the mill, and the distance to transport fresh fruit
bunches (FFB) to the nearest road. Hartley (1988) recommends a gap of 320
m between roads, giving a density of 33m roads ha
1, that is, 3% of the land
area. In hilly areas, the road density should increase, with distances between
roads not exceeding 200m, due to the difficulty of transporting FFB across
platforms and terraces. With terraces, the distance should be 125m, the
density being 80m roads ha
1. A typical road layout in hilly estates has to be
arranged in relation to the drainage lines and streams. If the roads can run
parallel to streams, this reduces the number of bridges needed and helps to
avoid crossing over swampy areas (Corley and Tinker, 2003). Steps for the
establishment and exploitation of an oil palm plantation on a typical private
estate are summarized in Fig. 2.

3.4. Water and soil management

Water management is a crucial aspect of oil palm cultivation as deficit or
excess of water stresses oil palm trees and is highly detrimental for crop yields.
Water management mainly aims at minimizing impacts of drought and floods
and optimizing the use of rainwater and fresh water from streams by drainage,
irrigation, and soil moisture conservation practices. Corley and Tinker (2003)
provided detailed information on irrigation in oil palm plantations. In Indo-
nesia, rainfall is generally well distributed over the year, so irrigation is not
common in oil palm plantations, except in South Sumatra where yield-
limiting water deficits may occur during the dry season.
Drainage systems are common in Indonesian oil palm plantations, par-
ticularly in flat areas with a high water table where drainage is compulsory to
remove excess water and promote oil palm root proliferation in deeper soil
layers. Indeed, optimum depth of water table for oil palm growth is 50–75cm
84 Irina Comte et al.

Nursery establishment Site preparation

Access road
Site clearing, underbrushing, and clear felling

Biomass management and disposal

Earthworks, drainage, irrigation, and infrastructure

Planting and maintenance
Cover crop establishment
of seedlings

Field establishment
Field lining and holing
Final culling
Transplant seedlings

Maintenance and harvesting

Fertilizer application

Pest control
General field upkeep
Harvesting
Transportation of fresh fruit bunches to mills

Replanting Abandonment
Removal of old palms Evacuation of plantation staff and workers

Removal of equipment, machinery, and

structure
Site restoration/rehabilitation

Figure 2 Typical oil palm plantation development activities (adapted from ECD,
2000).

(APOC, 2011; Goh and Chew, 1995). To achieve this, a good outlet with
sufficient capacity to discharge excess water is needed (Corley and Tinker,
2003). Another possibility is to install controlled drainage systems that retain
subsurface water in the drains prior to dry periods. Generally, the drainage
system consists of an interconnected network of collection and main drains of
varying dimensions depending on the hydrological and rainfall characteristics
of the area (Othman et al., 2010). The slope, intensity, and dimension of
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 85

drains depend largely on the expected amount of water to be removed during

wet periods. To achieve the desired water level, the minimum drain intensity
is at least one drain for every eight rows of palm and the intensity could be
further increased to one in every four or even to one in every two rows of
palms. Higher drainage intensity is adopted on clayey soils than sandy soils,
which naturally have better drainage (Goh and Chew, 1995).
Water management is more critical on undulating, hilly, or inland soils
because growers need to maintain soil moisture and minimize soil erosion
and nutrient losses. The American Palm Oil Council (APOC) recommends
(1) digging of silt-pits and foothill drains to trap water sediments from
surface runoff; (2) stacking fronds across the slope to minimize the velocity
of water runoff downhill slopes and to conserve water through mulching;
and (3) planting LCC, which not only helps replenish soil organic matter
stock but also reduces the velocity of soil and water movement. The LCC
commonly established in oil palm estates include Mucuna bracteata, Pueraria
phasesloides, and Calopogonium cearuleum. In steep terraced areas, deep root-
ing Vetiver grass can be used to prevent soil erosion and regulate water
excess.
However, water management and soil conservation are not practiced by
all planters. Plasma smallholders may benefit of drainage network imple-
mented by large companies involved in the nucleus–plasma scheme,
whereas independent smallholders may not have the financial and technical
means to dig and maintain a drainage network within their plantations, nor
knowledge of soil moisture and soil conservation practices. Despite abun-
dant literature and recommendations for water management, studies inves-
tigating actual and current water management practices in oil palm
plantations, especially in smallholdings, are almost nonexistent.

3.5. Nutrient-demand assessment

Despite the absence of mineral elements in the palm oil, large quantities of
nutrients are required by the palm tree to support its vegetative growth and
fruit production, which cannot be provided by inland and upland soils in
Sumatra and Borneo due to their low fertility status (Goh and Härdter,
2003). Thus, mineral fertilizers are compulsory to supplement the low
indigenous soil nutrient supply and to ensure suitable yields. Fertilizers
account for 50–70% of field operational costs and about 25% of the total
cost of production (Caliman et al., 2007; Goh and Härdter, 2003). Under-
standing the factors that contribute to efficient fertilizer use is crucial to
maximize yields and enhance economic returns (Goh and Härdter, 2003).
Consequently, the Indonesian oil palm industry has invested millions of
dollars in research and development to improve fertilizer use. Many trials
have been conducted on a wide range of soil types, climate, and tree ages in
order (i) to determine the ecophysiological nutrient demand of oil palm for
86 Irina Comte et al.

targeted yields (Tarmizi and Mohd, 2006); (ii) to determine input levels to
achieve an economically optimum production (income vs. costs of produc-
tion) (Breure, 2003; Caliman, 2001; Caliman et al., 1994; Goh and Härdter,
2003) including recommended types, rates, and timing of fertilizer applica-
tions to minimize nutrient losses (Goh and Chew, 1995); and also (iii) to
develop agricultural practices aiming at minimizing chemical fertilizers
inputs such as the establishment of a LCC during the immature stage
(Agamuthu and Broughton, 1985), recycling of pruned fronds and male
inflorescences (Ng and Thamboo, 1967), mulching of EFB (Caliman et al.,
2001; Chiew and Rahman, 2002; Loong et al., 1987), and POME spread-
ing (Rupani et al., 2010; Wood et al., 1979). These studies led to an
abundant literature on recommendations for optimum nutrient manage-
ment in oil palm plantations (Corley and Tinker, 2003; Fairhust et al.,
2005; Kee and Goh, 2006).
Fertilizer management in oil palm plantations is based on nutrient
balance principle, which estimates the total demand of the palm and
matches it with the nutrient supply in the oil palm plantation and from
supplemental fertilizers (Goh et al., 1999). Nutrient demand can be divided
in two categories: nutrient uptake and nutrient losses from soil through
processes such as runoff, leaching, and gaseous emissions. Within nutrient
uptake, we distinguish nutrient stocks exported by harvesting (FFB) and
nutrients immobilized in the palm biomass (for growth) (Goh, 2004; Goh
and Härdter, 2003). Some authors also calculate nutrients recycled from
pruned fronds and male inflorescences because they are usually returned to
the soil (Tarmizi and Mohd, 2006). The nutrient requirements of oil palm
vary widely, depending on the target yield, genetic potential of the planting
material used, and numerous environmental factors such as tree spacing,
palm age, soil fertility, groundcover conditions, and climate (Fairhurst and
Mutert, 1999; Goh and Härdter, 2003; Tarmizi and Mohd, 2006). Table 1
compares the total nutrient stocks in standing biomass of oil palm plantation
and tropical forest, illustrating that 1ha of forest generally contains more
nutrients in plant biomass than a plantation. Table 2 shows nutrient uptake
and allocation for production, immobilization in palm biomass, and recy-
cling, based on data from trials in Southeast Asia. Generally, a larger
proportion of nutrient uptake is needed for FFB than for immobilization.
For example, Ng et al. (1999) reported, for a target yield of 25tha
1 year
1,
the annual nutrient uptake in FFB was 2.3-fold greater than nutrients
immobilized in new biomass.
The soil nutrient supply in an oil palm plantation comes from dissolved
nutrients in atmospheric deposition, including precipitation, nutrients
recycled from pruned fronds and male inflorescences when these are
returned to the soil, nutrients leached by rainfall from the leaf canopy (leaf
wash), nutrient returns from LCC, available nutrients present in the soil, and
fertilizer applications (Goh, 2004; Goh and Härdter, 2003; Goh et al., 1999;
Table 1 Standing stock biomass in oil palm plantation and topical forest (adapted from Henson, 1999)

Total stocks in standing biomass (kgha
1)

Vegetal cover Mean biomass N P K Mg Ca Source

1
Oil palm, 14years 94tha for 588 58 1112 151 173 Ng et al. (1968)
136 palms ha
1
Rainforest 385tha
1 1067–2980 33–315 340–3163 161–595 476–4168 Anderson and Spencer (1991)
(5 studies range)
Table 2 Nutrient uptake in different components of oil palm plantations

Nutrient contents in oil palm biomass

(kg ha
1 yr
1)
Target yield
Oil palm components (ton FFB ha
1 yr
1) N P K Mg Ca Source

Production
Harvested fruit bunches 24 72.5 12.1 93.2 20.7 – Ng and Thamboo (1967),
Ng et al. (1968)
25 73.2 11.6 93.4 20.8 19.5 Ng et al. (1968)
30 97.6 10.0 105.4 18.2 – Tarmizi and Mohd (2006)
30 99.1 15.6 129.3 33.3 – Ng et al. (1999)
Immobilized
Trunk 30 42.4 4.1 121.6 10.2 – Ng et al. (1999)
Roots – 16.6 1.1 2.8 0.42 – Corley et al. (1971)
Trunk & roots 30 18.5 2.4 61.9 3.8 – Tarmizi and Mohd (2006)
Immobilized in new biomass 25 40.0 3.1 55.7 11.5 13.8 Ng et al. (1968)
Recycled
Pruned fronds and male inflorescences 24 78.4 11.3 102.1 28.1 – Ng and Thamboo (1967),
Ng et al. (1968)
FFB, Fresh Fruit Bunch.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 89

FFB 160 kg Kernels 7 kg

Outer
Inner cycle
Replanting
cycle Shell, fibre
cycle Increment
120 kg EFB to land 39 kg
(25 years) Inflorescences,
63 kg
rainfall leaching,
litter fall, POME Loss on
Felled pruned fronds to land combustion
Trunk & 30 kg
palms & roots leaves 290 kg Pruned fronds
1900 kg 1800 kg & cover crop Boiler ash Fertilizer
80 kg 210 kg

Ground level

Flow to roots Exchangeable soil K

Roots 1150 kg
570 kg Leaching,
110 kg runoff
28 kg
Non exchangeable K
9500 kg

Figure 3 Total demands and sinks for potassium in an oil palm plantation with 30ton
FFB yield (from Corley and Tinker, 2003). EFB, empty fruit bunch; FFB, fresh fruit
bunch; POME, palm oil mill effluent.

Tarmizi and Mohd, 2006). As an example, the potassium fluxes in oil palm
plantation are illustrated in Fig. 3.
Some research groups developed complex models, based on the nutrient
balance principle, to assess nutrient requirements, taking account of the
commodity price and fertilizer cost. Corley and Tinker (2003) reported on
three of these models, which are summarized below:
1. The Applied Agriculture Research group in Malaysia developed a linked
group of models for oil palm management, which include a model to
assess site-specific yield potential (ASYP), a model for predicting the
fertilizer requirements, the integrated site-specific fertilizer recommen-
dation system (INFERS), and expert systems for determining best
month for fertilizer application, and the timing and allocation of differ-
ent fertilizers (Kee et al., 1994; Kok et al., 2000).
2. Leaf analysis is the most common diagnostic tool to determine the nutri-
tional status of oil palm and estimate the appropriate fertilizer rates because
of significant relationship between leaf nutrient concentration and FFB
yield (Foster and Chang, 1997; Goh, 2004). It is largely used by the
International Cooperation Centre in Agronomic Research for
90 Irina Comte et al.

Development (CIRAD), which carries out leaf analysis of palms located on

important soil types within the plantation and uses response curves of the
leaf analysis results to determine the critical level corresponding to the
economically optimal fertilizer rate (Caliman et al., 1994, 2001).
3. The Foster system (PORIM fertilizer recommendation system) involves
two basic approaches: (i) the use of site-specific characteristics to deter-
mine yield without fertilizer, fertilizer need, and the efficiency of
response to fertilizer (Foster, 1995; Foster et al., 1986) and (ii) leaf
analysis data (Foster and Chang, 1977; Foster et al., 1988).

3.6. Fertilizer management

3.6.1. Chemical fertilizer
The dominant fertilizers produced and used in Indonesia are urea (46% N);
triple superphosphate (TSP, 46% P2O5); rock phosphates (RP, 27–34%
P2O5); ammonium sulfate (AS, 21% N and 24% S); potassium chloride,
also called muriate of potash (KCl or MOP, 60% K2O); magnesium sulfate,
also called kieserite (17% Mg, 23% S); and blended NPK, NP, and PK
fertilizers (FAO, 2005). Table 3 shows some recommended fertilizer appli-
cation rates, which vary according to climatic conditions, soil type, age of
palms, and palm yield potential. The optimal frequency of fertilizer applica-
tion depends on crop requirements, tree age, ground conditions, types of
fertilizer available, and rainfall. For example, frequent application of fertilizer
at low rates is preferred for sandy or sloped land where the risk of nutrient
losses through runoff or drainage is great. In such areas, a single annual
application of water-insoluble rock phosphate is recommended, whereas
soluble fertilizers would be applied in low doses several times a year. More
frequent fertilizer application is also advised for immature trees (Goh and
Chew, 1995). Some authors recommend fertilizer applications close to the
tree base in the initial years, and to be gradually extended to the tree avenues
when the canopy overlaps and good root development is reached (Goh and
Chew, 1995; Goh et al., 2003). Moreover, the timing of fertilizer application
should account for the rainfall pattern to avoid substantial nutrient losses (Goh
and Chew, 1995; Goh et al., 2003). Thus, the general guideline is to avoid
fertilizer applications during period with high rainfall, such as during months
with more than 250mm month
1, months with high rainfall on more than
16days month
1, and months with high-intensity rainfall events of more than
25mm day
1 (Goh and Chew, 1995).

3.6.2. Organic fertilizer

The value of oil palm residues such as pruned fronds and other wastes from
processing mills for mulch and organic manure is well documented (Dolmat
et al., 1987; Khalid et al., 2000; Loong et al., 1987). According to Fairhurst
Table 3 Recommended fertilizer applications for oil palm in South East Asia

Fertilizer applications (kgha
1 year
1a)

Notes N Notes P Notes K Notes Mg Notes B Source

Immature 45 Immature 24 Immature 108 Immature 28 Immature 0.6 FAO (2005)

Mature 120 Mature 22 Mature 286 Mature 24 Mature 0.6
Immature 35–105 Coastal clay 42–56 Depending 42–420 Mg sufficiency 8.4–35 Immature 1.4 Goh et al. (2003)
soil on site
Mature 35–245 Malaysian 56–98 conditions Mg deficiency 42–105 Mature 2.1–4.9 Goh et al. (2003)
inland soils
Immature 50–120 Immature 22–48 Immature 54–216 Immature 7–24 Immature 1.2–3.7 von Uexküll
(2007)
Mature 120–200 Mature 30–87 Mature 183–581 Mature 0–36 Mature 2.5–5.6 von Uexküll
(2007)
a
Assuming 140 palms ha
1 when recommendations were expressed as kg palm
1.
92 Irina Comte et al.

(1996), field palms yielding 30t FFBha
1 in West Sumatra return to the
soil about 10tha
1 year
1 dry matter from pruned fronds, which contain
125kg N, 10kg P, 147kg K, and 15kg Mg. Also, both EFB and POME
contain substantial amounts of nutrients and organic matter that can replen-
ish the soil fertility and help meet the nutrient requirements for oil palm.
According to Taillez (1998), mulched EFB can reduce the need for chemi-
cal fertilizers by more than 50% in immature stands and by 5% in mature
stands. Application of 40–60t EFBha
1 year
1 or 750m3 POMEha
1 year
1
is recommended to add organic matter and improve soil fertility on poor
inland soils (Goh et al., 1999).

3.7. Synthesis
Industrial and smallholder planters do not have the same resources to ensure
optimum nutrient management. Specifically, independent smallholders do
not benefit from techniques such as leaf diagnosis and soil analysis (Fairhurst
and Mutert, 1999; Pushparajah, 1994) to assess the nutrient requirements of
their plantations. In many smallholdings, low-productivity palms are
planted unevenly without terracing, and fertilizer use is inadequate and
unbalanced (urea applied alone) (FAO, 2005; Webb et al., 2011; Zen
et al., 2005). Some smallholders observe and copy the industrial plantations
in recycling pruned fronds and male inflorescences, but not all. Moreover,
they do not benefit from organic inputs of EFB and POME, which are
available only in industrial or governmental estates where the mills are
located. Unfortunately, few data are available on actual fertilization practices
in most oil palm plantations in Indonesia. Some authors provide estimates of
average yields of palm oil from the different management groups, which are
generally higher for private plantations than smallholdings, although the
FAO estimated similar yields for these two groups (Table 4). However,
these estimates should be interpreted with caution as they do not account
for the high variability for cultural practices among planters. Although some

Table 4 Average yields of palm oil production for the different production systems in
Indonesia (tha
1)

Industrial
Smallholder Government Private Source

1.8 3.7 1.7 FAO (2005) (in 2002)

3.3 (5) 4.2 (7) 4.1 (7) USDA (2009) (in 2008)
2.34 3.04–5.52 Bangun (2006)
( ), potential yields.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 93

private producers have recorded high yields of 6.5–8.0tha
1 on individ-

ual plantations, the gap between yields on smallholdings and private
estates is inexplicably low, given the tremendous advantages that private
estates show in capital, land management abilities, fertilizer availability,
and access to high-yielding varieties (USDA, 2009). Two possibilities to
explain this discrepancy are (1) producers on private estates have seri-
ously underinvested in soil fertility and fertilizer use efficiency, as they
have the financial resources to manage nutrients for higher yields, or (2)
producers on private estates have not been motivated to manage their
trees intensively, as they already achieve sufficient profit margins (USDA,
2009).
Abundant recommendations exist for soil, water, and nutrient manage-
ment on oil palm plantations, especially for large plantations, but small-
holder farms are generally not large enough for trials to be implemented
(Webb et al., 2011). Moreover, there is scant or no data available from study
cases on the actual management practices and fertilizer use in industrial
plantations, let alone smallholdings. However, these practices have a high
potential to impact hydrological processes and associated nutrient fluxes
from oil palm plantations.

4. Hydrological Processes and Associated

Nutrient Transfers in Oil Palm Plantations
Replacing a natural forest with an oil palm plantation is expected to
drastically change the existing characteristics of the area and to modify the
hydrological cycle, shown in Fig. 4 (Henson, 1999). The activities related to
the oil palm plantation establishment and exploitation (e.g., complete
clearing of forested areas, construction of roads and drainage networks,
fertilizer and agrochemical use, wastewaters release from mill and worker
residences) are expected to cause problems related to water flow (e.g.,
flooding incidents downstream) and increase nutrient and sediment delivery
to streams, causing deterioration in water quality (ECD, 2000; Goh et al.,
2003). This section aims to (1) compare hydrological processes in naturally
forested watersheds with those under oil palm plantation, including the
hydrological changes occurring during oil palm development (immature vs.
mature plantations) and (2) highlight gaps in literature regarding the under-
standing of hydrological processes in oil palm plantations and the conse-
quences for water quality. We do not review hydrological processes in
tropical forests as a number of detailed reviews were already published
(Bruijnzeel, 1990, 2004; Douglas, 1999; Elsenbeer, 2001), but will refer
to those data to illustrate the magnitude of change in hydrological processes
occurring in oil palm plantations.
94 Irina Comte et al.

Evapotranspiration Rainfall

Evaporation Interception

Evaporation Transpiration Throughfall

Stemflow

Water in
biomass
Overland flow
Infiltration
Stormflow Streamflow

Soil water Lateral flow

Deep Capillary
drainage rise

Groundwater Baseflow

Figure 4 The hydrological cycle in an oil palm plantation. Boxes indicate storage
pools, arrows indicate fluxes.

4.1. Precipitation in Indonesia

Indonesia’s humid tropical climate is characterized by seasonal changes in
rainfall largely determined by monsoons winds, due to its location in the
equatorial zone, between the Asian and Australian landmasses. Typically, the
northwest monsoon brings rain from December through March, followed by
the southeast monsoon, which brings drier weather from June through
September (Galdikas, 2009). Rainfall patterns in Indonesia vary from one
region to another (Aldrian and Dwi Susanto, 2003). Annual rainfall in
lowland areas is between 1800 and 3200mm, increasing with altitude to an
average of 6100mm in some mountainous regions. In the lowlands of
Sumatra and Kalimantan, the annual rainfall averages 3000–3700mm.

4.2. Interception
Studies of hydrological processes in tropical rainforests indicate the impor-
tance of interception by the leaves and branches (canopy) of plants. In his
review, Bruijnzeel (1990) concluded that forest interception was between
4.5% and 22% of the rainfall incident on the canopy, with an average value
of 13%. Compared to natural forest vegetation, a lower proportion of
rainfall is expected to be intercepted by palms as a result of lower leaf area
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 95

index, even in mature oil palm plantations (Henson, 1999) due to clearing
of most understory vegetation. Dufrêne (1989) reported that less than
5% rainfall was intercepted following a 30-mm precipitation event and
lower values were recorded at higher rainfall intensity. However, Squire
(1984) found that interception was 17–22% of precipitation in oil palm
plantation, the amount varying with palm age, erectness of canopy, and
rainfall intensity.
In tropical forests, only a small amount of rain falls directly on the ground
or into water bodies, as most of the rainfall will reach the soil via throughfall
and stemflow, which are responsible for transferring nutrients from the
canopy to the soil. Some Malaysian studies found that throughfall trans-
ferred 70–78% of rainfall in mature oil palm plantations (Kee et al., 2000). In
Papua New Guinea where rainfall is higher, Banabas et al. (2008) found that
83% of the rain reached the ground as throughfall. These values are the same
order of magnitude as those reported by Bruijnzeel (1990) for lowland
rainforest: 77–93% of incident rainfall (on average 85%) was transported
by throughfall. We are not aware of published data on throughfall, stem-
flow, and rainfall interception by immature oil palm when the canopy has
not yet closed.

4.3. Evapotranspiration
A number of evapotranspiration (ET) studies were carried out in tropical
forests (Noguchi et al., 2004; Tanaka et al., 2008; Zulkifli et al., 1998). In
three catchments in Selangor, Malaysia with more than 94% forest cover,
Low and Goh (1972) found ET accounted for half of the 2162–2482mm
annual rainfall when estimated as the difference between measured rainfall
and water outflow from the catchment. In a study in the Sungai Tekam
Experimental Basin, Malaysia (DID, 1989) that received average annual
rainfall of 1878mm for the period 1977/1978 to 1985/1986, the actual ET
of a forested catchment averaged 1500mmyear
1 (range: 1374–1583mm
year
1), which represents 99% of the potential ET, on average 1515mm
year
1 (range: 1449–1567mmyear
1) based on the Penman & Thornwaite
method of estimation. In his review, Bruijnzeel (2004) reported ET typi-
cally ranged from 1000 to 1800mmyear
1 in lowland and hill dipterocarp
forests in Peninsular Malaysia.
There have been few ET studies in oil palm plantations (Yusop et al.,
2008). Micrometeorological measurements in mature oil palm on a Malaysian
coastal site showed that ET accounted for 83% of rainfall and was close to
potential ET calculated by the Penman equation (Henson, 1999). A study of
oil palm on volcanic soils in Papua New Guinea gave an estimated ET of
1334–1362mmyear
1; Banabas et al. (2008) noted that water deficit was
unlikely to limit transpiration in this area, which had average annual rainfall
between 2398 and 3657mm. Radersma and de Ridder (1996) estimated oil
96 Irina Comte et al.

palm ET of 1018–1051mmyear
1 from literature data. We know of one

published report on ET in immature oil palm plantations. Yusop et al. (2008)
estimated ET in three catchments where oil palm ages ranging from 2 to 9
years, using short-time period water budget and catchment water balance.
Surprisingly, they found higher values (1405 and 1365mmyear
1 for the two
methods, respectively) in the 2-year-old plantation than the 9-year-old
plantation (927 and 1098mmyear
1, respectively), as higher transpiration is
expected in mature plantation. They concluded that ET values in the older
plantation were underestimated. Overall, these literature reports suggest high
ET in mature oil palm plantations in Southeast Asia, ranging from about
1000–1300mmyear
1, which is similar to ET in tropical rainforests. How-
ever, few studies report ET in immature oil palm plantation, and the differ-
ence in ET of different age plantations remains to be fully investigated.

4.4. Soil infiltration, leaching, and groundwater quality

4.4.1. Soil infiltration
Water reaching the ground may either infiltrate into the soil or flow directly
to the stream through surface runoff. Partitioning between surface runoff
and infiltrated water essentially depends on soil infiltrability and rainfall
intensity. Once infiltrated, water can either percolate downward as vertical
flow or reach the stream through lateral (downslope) subsurface flow,
depending on soil hydraulic conductivity gradient. In rainforests, water
moves within the soil as matrix flows or more rapidly through bypass or
preferential flows (roots channels, macropores) in both vertical and lateral
directions (Noguchi et al., 1997a). It is well known that soil infiltration
capacity depends on soil texture (FAO, 1990) and on land use (Chorley
et al., 1984; Osuji et al., 2010). Many studies showed that tropical forest soils
have high infiltration capacities (Bruijnzeel, 1990), due to dense vegetation
(increases water uptake and ET from the system) and high soil organic
matter content that improves the soil structure and enhances its porosity.
For example, high infiltration rates of 200–250mmh
1 were found in two
tropical forest soils of western Nigeria under bush fallow (natural regrowth)
(Wilkinson and Aina, 1976). According to Bruijnzeel (2004), about
80–95% of incident rainfall infiltrates the soil in a mature tropical rainforest.
However, some studies showed that hydraulic conductivities decrease rap-
idly with soil depth in forested land, leading to shallow subsurface flow from
the nearly saturated topsoil and contributing to stormflow generation (Bidin
et al., 1993; Malmer, 1996; Noguchi et al., 1997b). This phenomenon was
also observed by Yusop et al. (2006) in forested catchments in Malaysia due
to positive hydraulic pressure at 10cm depth during storms, despite high
hydraulic conductivity (169–1485mmh
1).
Deforestation can affect infiltration capacity of soils in a number of ways.
According to DID (1989), removal of forest reduces transpiration and
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 97

increases soil moisture storage, which resulted in soils reaching field capacity
earlier during rainfall events. Infiltration rates will also be reduced immedi-
ately following deforestation when soils are compacted by heavy machinery.
However, DID (1989) showed that infiltration rates return to predistur-
bance levels following the establishment of crops due to improvement in
soil structure under vegetative ground cover.
According to Banabas (2007), soils under mature oil palm typically have
high infiltrability (80–8500mmh
1, depending on soil texture). However,
the infiltrability of the soil is highly variable due to the ordered structure of
vegetation in oil palm plantations. A study in West Sumatra on 10-year-old
oil palms demonstrated significant spatial variability when comparing water
infiltration rates in soil beneath the palm circle, harvest path, and frond piles.
Infiltration rate increased in the order path<circle<frond pile (Fairhurst,
1996). In Papua New Guinea, Banabas et al. (2008) recorded similar results
(Table 5). They attributed the highest values in the frond pile zones to the
macroporosity-enhancing effect of the organic matter present in this zone.
They ascribed the lower values found in the weeded circle and harvest-path
zones to topsoil compaction from falling bunches in the weeded circle,
wheel and foot traffic in the harvest paths, and sparse understory vegetation.
Thus, apart from its high dependence on soil texture, infiltrability will
vary temporally and spatially in oil palm plantation. After a strong decrease
due to soil compaction after land clearing, infiltrability in oil palm plantation
may partly recover due to plant growth and organic matter addition to the
soil. Yet, infiltrability will remain low along roads and harvest pathways and
in weeded circles.

4.4.2. Leaching and groundwater quality

Leaching losses are generally assumed to be higher in the humid tropics than
temperate areas due to the frequent and intense rain storms, higher temper-
ature, and high carbonic acid content in soil (Ah Tung et al., 2009; Johnson

Table 5 Soil infiltrability (mmhr
1) for major soil types and at specific locations in oil
palm plantations

Frond Between Weeded Harvest

Soil piles zones circles path Source
Typic Hapluland, sandy 1351 997 270 265 Banabas et al.
clay to clay loam (2008)
Typic Udivitrand, 7350 1230 340 60 Banabas et al.
sandy loam to sand (2008)
Typic Hapludult 2050

175 Maena et al.
(1979)
98 Irina Comte et al.

et al., 1975). Depending on the amount of water draining out of the rooting
zone, leached solutes may simply accumulate in a deeper layer of the soil
profile or may reach the underlying groundwater (Ah Tung et al., 2009).
Table 6 summarizes results from Bruijnzeel (1991), including estimates of
annual runoff and nutrient losses in drainage water from tropical rainforests.
The author also reported that forest clear-cutting often increased nutrient
losses to streams. Brouwer and Riezebos (1998) compared nutrient leaching
in closed and logged tropical rainforest in Guyana. Logging clearly increased
leaching, which they ascribed to an increased percolation of water through
the soil (2800mm after 22months logging compared to 1800mm in closed
forest) and increased solute concentrations in the percolating water. They
found that Ca, K, and Mg concentrations were 2–10 times greater in the gaps
after logging than in closed forest, while the NO3 concentration was 10–20
times greater than closed forest. The major pulse of leaching occurred during
the first year after logging, and most solute concentrations in percolating soil
water remained higher up to 15months after logging. However, vegetative
regrowth reduced leaching losses as plants absorbed soluble nutrients and
immobilized them in standing stock biomass.
Large nutrient losses are expected in oil palm plantations (Goh et al.,
2003). According to Ng et al. (2003), most of the oil palm root biomass is
found within 1m of the soil, but the distribution of oil palm active roots
favors the nutrient uptake in the upper 30cm, which may increase the
potential risk of nutrient leaching. Omoti et al. (1983) measured amounts
of nutrients leached under immature (4years) and mature (22years) oil
palms, distinguishing between the loss of native nutrients and the loss of
applied nutrients (fertilizers). The losses of native nutrients from closed and
logged tropical forest on one hand and in immature and mature oil palm
plantations on the other hand are summarized in Table 7. To our knowl-
edge, there is no chronological study comparing the amounts of nutrient
leached under natural forest and oil palm plantation established on the same
site neither after forest clearing nor between forested and oil palm catch-
ments with similar climatic and soil conditions. This makes it difficult to
quantify the impact of oil palm plantation establishment on native nutrient
leaching. The most complete study to assess the impact on hydrological
processes of forest conversion to tree crop and oil palm plantations (DID,
1989) did not evaluate leaching, leaving a considerable gap in our knowl-
edge of this process in plantations at the catchment scale.
Despite the paucity of large-scale data on leaching processes, many plot-
scale studies investigated the percentage of applied fertilizers lost through
leaching in oil palm plantations, some of them comparing young and mature
oil palm stands (Chang and Zakaria, 1986; Foong et al., 1983; Maena et al.,
1979). For example, a field lysimeter study conducted on Munchong series
soil in Malaysia found higher fertilizer losses when the palm was 1–4years old
(17% for N and 10% for K), which declined to 2.1% of applied fertilizer N
Table 6 Catchment studies of annual rainfall, annual runoff, and nutrient losses in drainage water from South East Asian tropical forests (modified
from Bruijnzeel, 1991)

Nutrient losses
Annual Annual
(kgha
1 year
1)
Catchment rainfall runoff Q/P Sources quoted by
Location Type of forest Soil area (ha) P (mm) Q (mm) (%) Ca Mg K P N Bruijnzeel (1991)
Ulu Gombak, Partly disturbed Oxisol 31 2500 750 30 2.1 1.5 11.2 – – Kenworthy (1971)
Malaysia dipterocarp forest
Bt. Berembun, Undisturbed Deep Ultisols (2/3) 29.6 2005 225 11 5.8 3.6 8 – – Abdul Rahim and
Malaysia dipterocarp forest and Oxisols (1/3), Zulkifli (1986),
sandy clay (loam) Zulkifli (1989),
and Zulkifli et al.
(1989)
Watubelah, Plantation forest of Andesitic tuffs 18.7 4670 3590 77 29 30.5 22 0.7 10.6 Bruijnzeel (1983a,c,
Indonesia Agathis dammara underlain by 1984)
andesitic breccias
Kinta Valley, Lowland rainforest Limestone – 2845 1605 56 795 90 76 – – Crowther (1987a,b)
Malaysia (karst terrain)
Ei Creek, Papua Colline rainforest Basaltic volcanic 16.25 2700 1480 55 24.8 51 14.9 – – Turvey (1974)
agglomerates
Gua Anak Lowland rainforest Limestone – 2440 1255 51 764 45 20 – – Crowther (1987a,b)
Takun,
Malaysia
Table 7 Nutrient leaching losses in undisturbed forest, highly disturbed forest, and oil palm plantations

Amount of element leached (kgha
1 year
1)

Ca K Mg Na Cl NH4-N NO3-N N-total SO4-S Source
a
Closed tropical rain forest, Guyana 2 5 1 24 25 1 3 4 Brouwer and Riezebos (2002)
Large gap sikkder zone, Guyanaa 14 23 16 65 56 1 90 91 Brouwer and Riezebos (2002)
Unfertilized young oil palm (4years), 165 3 32
53

32 53 Omoti et al. (1983)
Nigeria
Unfertilized adult oil palm (22years), 123 29 32
30

65 83 Omoti et al. (1983)
Nigeria
a
Annual average on 1030days.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 101

and 2.7% of fertilizer K when the palm was 5–14years old (Foong, 1993).
Higher nutrient losses through leaching from immature palm implies less
plant nutrient uptake, whereas older palms have more extensive root system
that can absorb applied and indigenous soil nutrients, a greater nutrient
demand and a higher transpiration rate that lowers water loss via leaching.
However, a field-scale study on Orlu and Algba series (Rhodic Paleudult)
soils in Nigeria showed no significant differences in nutrient leaching
from applied fertilizers between the immature (4years) and mature plantations
(22years) (Omoti et al., 1983). According to some authors, the adult stage
poses a high risk of nutrient losses because ground vegetation is sparse due to
poor light penetration through the closed oil palm canopy (Breure, 2003).
Moreover, the LCC dies off at canopy closure, releasing a large amount of
N from the decomposing legume biomass and increasing the risk of N loss via
leaching (Campiglia et al., 2010; Goh et al., 2003). According to Goh and
Chew (1995), leaching losses also depends on soil texture and greater losses
were recorded in sandier soils, as summarized in Table 8. In general, leached
P losses are low due to the relative immobility of P in acidic, weathered
tropical soils (Goh et al., 2003; Omoti et al., 1983).
A plot-scale study by Ah Tung et al. (2009) is the only one to our
knowledge that investigated leaching losses of inorganic N and K and
measured their concentrations in groundwater. They found leaching losses
of inorganic N represented between 1.0% and 1.6% of applied N fertilizer
and the K losses were between 2.4% and 5.3%, depending on fertilizer
application rates. The concentration of N and K in the soil solution
decreased with soil depth, which they explained by nutrient removal and
uptake by palm roots, resulting in lower nutrient concentrations in the soil
solution with depth. However, another explanation they did not mention
is that fertilizers are applied near the soil surface, so the topsoil layers have a
higher concentration of nutrients in soil solution; this naturally declines
with depth because there was no fertilizer injection deeper in the soil
profile. The measured concentrations of NH4-N, NO3-N, and K in
groundwater ranged from 0.23 to 2.7, 0.07 to 0.25, and 0.63 to 9.54mgl
1,
respectively, which did not exceed the water quality standards set by the
World Health Organization (WHO, 2008). The authors did not specify
the distance of groundwater sampling wells from the palm circles where
the fertilizers were applied. However, they mentioned the possibility of
groundwater pollution when excessive N fertilizer was applied or if NO3-N
leached from the soil profile into groundwater in the intertree spaces between
palms. This supposition is supported by Schroth et al. (2000), who reported
pronounced spatial pattern of NO3-N concentrations within the plantation.
Low NO3-N concentrations were measured in the soil profile within 1m of
palm trees, indicating efficient absorption of mineral N by the palms, whereas
soil NO3-N concentrations increased with increasing distance from palm
trees. At 4m away from the trees, the vertical NO3-N concentration gave
Table 8 Nutrients leached as percentages of applied fertilizers in immature (1–4years) and mature (>4years) oil palm plantations

Annual Age of Nutrient losses of applied fertilizer (%)

rainfall Palms palm
Location Soil (mm) per ha (year) N P K Mg Ca S Cl Notes Source
Nigeria Rhodic Paleudult 1923 150 4 and 22 34 – 18 172 60 14 141 Omoti et al. (1983)
Sabah, Typic Hapludults >2500 26 1–1.6 – 2.4–5.3 – – – – Ah Tung et al. (2009)
Malaysia
Malaysia Typic Hapludox 1909 145 1 26.5 Trace 19.5 169.4 – – – Rainfed Foong et al. (1983)
(Munchong series) 1495 2 10.9 Trace 3.4 8.4 – – – Rainfed
2729 3 12.2 1.4 10.4 53.6 – – – Irrigated
2787 4 16.8 5.8 5.6 47.6 – – – Irrigated
2391 5 2.7 1.7 1.9 5.4 – – – Irrigated
2193 6 4.8 1.4 3.3 6.6 – – – Irrigated
Malaysia Typic Paleudult 2352 – – 10.4 – 5.1 – – – – Chang and Zakaria
(Serdang series) (1986)
Malaysia Typic Hapludox 1–4 16.6 1.8 9.7 69.8 – – – Foong (1993)
(Munchong series) – – 5–8 1.2 1.6 2.5 11.5 – – –
9–14 3.0 1.5 2.9 15.5 – – –
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 103

clear evidence of NO3-N leaching into the deeper subsoil as concentrations

almost doubled from surface soil to 175cm depth (6–11.5mgg
1, respectively).
They concluded that much of the soil volume in the plantation was apparently
not accessed by palm roots, leaving surplus N at risk of leaching loss (Schroth
et al., 2000).
Thus, nutrient leaching in oil palm plantations may be impacted by soil
type and rainfall intensities (that both influence infiltration rates), oil palm
age, agricultural practices such as LCC establishment, water management
that impacts the water table level and subsequently increases the possibility
of leaching, fertilizer types, and rate applications, etc. In oil palm plantations
receiving chemical fertilizer, low-nutrient losses via leaching and nutrient
concentrations in groundwater quality were generally reported. Higher
nutrient losses (as % of fertilizer application) are expected in immature
plantations due to lower nutrient uptake by palm roots. However, higher
fertilizer applications are recommended in mature plantations, which may
lead to higher losses in absolute terms. The link between nutrient leaching
and groundwater quality is clearly not fully investigated or understood. A
shortcoming of the current literature is the reliance on plot-scale studies
with lysimeters installed close to the trees, despite high spatial heterogeneity
in the structure of the oil palm plantation that affects soil infiltration rates,
distribution of vegetation, and root growth in the soil profile. Although
organic fertilizer applications (POME and EFB) also supply nutrients that
could potentially leach, impacting groundwater quality, work on oil palm
plantations amended with these residuals has focused on soil properties
rather than groundwater quality (Okwute and Isu, 2007).

4.5. Surface runoff and erosion

In rainforest watersheds, surface runoff may occur as infiltration excess
overland flow and/or saturated overland flow. Malmer (1996) reported that
infiltration excess overland flow was more likely than saturated overland flow
across the sloping uplands of a tropical forested catchment, whereas areas close
to the stream are susceptible to saturated overland flow or return flow.
However, infiltration excess overland flow in forests is regarded as a rare
phenomenon, as vegetation plays an important role in holding and absorbing
rainfall (Bonnell, 2005; Bruijnzeel, 1990; Zhang et al., 2007). Noguchi et al.
(1997a) concluded that neither saturation overland flow nor infiltration excess
overland flow is likely to occur at Bukit Tarek Experimental Watershed, a
forested watershed in Peninsular Malaysia. Data on annual runoff soil and
nutrient losses in tropical rainforests, reviewed by Bruijnzeel (1990), are
summarized in Table 9. Although surface runoff is rare in forests, it is likely
in oil palm plantations during short-term high-intensity rainfall events
because of the high intrinsic variability in soil infiltrability (Banabas et al.,
2008). When runoff does occur, it would be initiated mainly from the
Table 9 Runoff, soil erosion, dissolved solutes, and sediment losses via runoff in Southeast Asian rainforest

Annual
rainfall (P) Runoff Q/P Dissolved loss TSS outflow
Soil Study scale mm (Q) mm (%) Soil erosion (tha
1 year
1) (tha
1 year
1) Notes
Red-yellow Plot 210m 2862–3563 2.50 0.5 0.41gm
2 – – Period 27/3 to 6/6/98;
podsol rainfall 495mm
2.50 0.8 1.35gm
2 – – Period 28/12/97 to
21/2/98; rainfall 316mm
Gleyic Runoff plot 50–200m2; 1950 56.55 2.9 38kgha
1 0.16 0.3 Excluding valley bottom
podsol catchment scale year
1
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 105

weeded circle (where stem flow causes the highest local water inflows and the
infiltrability is quite low) and the harvest-path zone (where infiltrability is
lowest due to soil compaction) (Table 10).
While erosion is never excessive (i.e., greater than the rate of soil
formation) in forests, soil loss can be pronounced at particular stages in an
oil palm plantation. Many studies reported highest erosion rates immedi-
ately after land clearing, resulting from increased exposure of the soil surface
to erosion and surface runoff losses (Bruijnzeel, 1990, 2004; Douglas, 1999;
Goh et al., 2003). According to Clay (2004), bare soil resulting from road
construction and other infrastructure such as bridges, culverts, and drains
increases soil erosion in oil palm plantations. DID (1989) showed that
deforestation activities such as timber harvesting, construction of roads,
and preparation of land for crop planting account for as much as 91% of
all the sediment exported from the catchments. Results from erosion plots
on two soil types revealed five to seven times more erosion from deforested
land than forested lands during the first year after planting the LCC.
However, once the ground cover was established, erosion was greatly
reduced but not eliminated (DID, 1989). In mature plantations, erosion
still occurs from harvest paths, roads, and localized areas of steep elevation.
Clay (2004) reported that in Papua New Guinea, every 100m of road has
the potential to produce as much sediment as each hectare of oil palm, but
this is not unrelated as there are 50 linear meters of road for every hectare of
oil palm planted.
Some authors estimated contribution of rainfall to runoff and associated
nutrient losses, usually expressed as percentage of applied fertilizers, carrying
out plot-scale studies in oil palm plantations. Some of them computed
nutrient losses, via runoff and/or sediment transport to be large, accounting
for up to 10% of applied fertilizers (Maena et al., 1979). They observed
greater losses from surface runoff in the uncovered soil in the harvest path,
compared to the interrows, where pruned fronds provide soil cover
(Fairhurst, 1996; Goh et al., 2003; Maena et al., 1979). Others reported
low losses of nutrient via runoff in oil palm plantation (Banabas et al., 2008).
Results from key papers are summarized in Table 11. Moreover, runoff
losses of applied fertilizers also depend on the lag time between the applica-
tion and the subsequent rainfall. While Chew et al. (1995) showed that high
rainfall prior to fertilizer application resulted in substantial nutrient loss, Kee
and Chew (1996) found that the first rain event following fertilizer applica-
tion in a wet month gave N concentrations in runoff water of 89 and 135
mgkg
1 for 65 and 130kgN ha
1 rates, respectively, compared to 4mgN
kg
1 in the unfertilized control plot. Thus, the amount of fertilizer nutrients
lost through runoff and sediment transport depends on the soil texture, the
age of the oil palms, the local topography and infiltrability, and the lag time
between fertilizer application and rainfall (Banabas et al., 2008). The con-
tinual compaction of harvest pathways and roads and the disappearance of
Table 10 Soil erosion and nutrient losses in surface runoff water from spatial components of an oil palm plantation on a Typic Hapludult in
Malaysia (after Goh et al., 1999; Maena et al., 1979)

Nutrient losses (% of applied fertilizer)

Average annual runoff Soil erosion
Fertilizer placement (% of rainfall) (tha
1 year
1) N P K Mg Ca B
Oil palm row 20.2 7.47 13.3 3.5 6.0 7.5 6.8 22.9
Harvest path 30.6 14.92 15.6 3.4 7.3 4.5 6.2 33.8
Frond pile 2.8 1.1 2.0 0.6 0.8 2.7 0.8 3.3
Frond pile/harvest path

6.6 1.4 3.5 2.2 3.4 12.5
Average for the field

11.1 2.8 5.0 5.6 5.2 20.7
Fertilizer nutrients applied (kgha
1) 90.2 52.0 205.9 32.8 78.9 2.4
Table 11 Nutrient losses through runoff and eroded sediment in oil palm plantations (plot-scale studies)

Age of oil Annual Annual Annual nutrient losses, kgha
1 year
1 (% of applied fertilizers)
palm rainfall runoff
Location Soil plantation (mm) (% of rainfall) N P K Mg Ca Transport Source
Malaysia Orthoxic 11 1426 2.8–30.6% 9.93 (11.1%) 1.43 (2.8%) 10.40 (5.0%) 1.82 (5.6%) 4.04 (5.2%) In runoff Maena et al.
Tropudult 5.57 (6.2%) 3.63 (7.0%) 8.79 (0.0%) 21.10 (64.3%) 7.40 (9.4%) In eroded (1979)
sediment
Malaysia Typic 4.5–7.2 0.7–1.1 20.8–33.0 3.6–6.8 – In runoff Kee and
Paleudult (4.4–7.2%) (0.5–0.8%) (9.7–15.4%) (4.0–7.6%) – Chew
0.5–0.8 0.5–1.3 Trace 0.1 (0.1%) – In eroded (1996)
(0.5–0.9%) (0.3–0.9%) sediment
PNG Typic Mature 2398 6 (0–44 for 2.2 – – – – In runoff Banabas
Hapluland (135 374 individual et al.
palms events) (2008)
ha
1)
PNG Typic Mature 3657 1.4 (0–8 for 0.3 – – – – In runoff
Udivitrand (135 682 individual
palms events)
ha
1)

PNG, Papua New Guinea.

108 Irina Comte et al.

understory cover (including LCC) due to canopy closure may also contrib-
ute to soil and nutrient losses via runoff and erosion within a mature oil palm
plantation (Table 11).

4.6. Stream flow and stream water quality

4.6.1. Stream flow
Hydrological changes observed at the plot scale will have consequences on
stream flow and water quality. It is well known that total stream flow
increases following clearance of forest cover and conversion to other types
of land use (Bruijnzeel, 1990). The absence of vegetation allows a greater
proportion of direct rainfall to reach the forest floor, and reduced ET rates
(due to absence of plants) translate into a greater volume of water leaving the
catchment. When land-use change increases the amount of disturbed and
compacted surfaces, there will be an associated increase in surface runoff
and stream flow. This increase may be permanent when converting natural
forest to grassland or shallow-rooted agricultural crops, or temporary in the
case of conversion to tree plantations (Abdul Rahim and Harding, 1992;
Bruijnzeel, 1990).
Hydrological studies carried out in oil palm plantation at a watershed
scale are scarce, except those of Sungai Tekam Experimental Basin in
Pahang by DID (1989). They observed an increase in total flow immedi-
ately in response to deforestation (þ85 to 157% for the first 3years after
deforestation), which declined gradually with the planting of LCC and tree
crops. Total flow increase was due to greater baseflow rather than more
runoff. The authors ascribed the large increase in baseflow to rising water
table level due to reduced ET and ponding effects immediately after defor-
estation, as felled logs and debris were left in the stream channels for long
period acting as debris dams. However, in Bruijnzeel’s review (2004), other
authors reported a decrease in baseflow following deforestation, especially
during the dry season because more surface runoff from compacted soils
decreased the groundwater recharge and the subsequent release of baseflow.
Soil compaction following land clearing triggers a shift from dominant
subsurface flow to overland flow, increasing peak flow during storm events
(Bruijnzeel, 2004). DID (1989) observed also that peak discharge increased
after deforestation and that time-to-peak decreased significantly from 3 to 1
h immediately after deforestation. Although this study provides insight into
short-term hydrological changes when oil palm plantations are established,
it was stopped before the oil palm plantation reached maturity (Hui, 2008).
Further study of the mature plantation would be helpful to determine
whether older oil palm plantations continue to experience more surface
runoff and higher peak flows during storms than undisturbed forest, leading
to higher stream flow in the watershed. A study in a small oil palm
catchment (8.2ha) on a coarse sandy clay loam Ultisol in the upstream of
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 109

Skudai River in Johor, Malaysia, showed a high proportion of baseflow,

approximately 54% of the total runoff and rapid responses to rainfall with a
short time (6–48min) to peak flow (Yusop et al., 2007). However, baseflow
can be higher in forested catchments and reach as much as 70% of the total
annual flow (Abdul Rahim and Harding, 1992; Yusop et al., 2007).

4.6.2. Stream water quality

It is clear that oil palm plantations have different hydrological characteristics
from natural forests at the plot scale, which may impact the quality of
receiving waters at a watershed scale. The increase of surface runoff loaded
with eroded soil particles, the use of agrochemical (fertilizers and pesticides),
and the release of POME in the streams are expected to affect the aquatic life
and drinkable water quality of the receiving water bodies (ECD, 2000; Sheil
et al., 2009). However, catchment-scale studies on water quality and nutrient
losses in tropical areas have focused primarily on forested areas and the impacts
of rainforest disturbances (Malmer, 1996; Malmer and Grip, 1994). In
Malaysia, some researchers reported slightly acidic stream water, low electrical
conductivity, and low solute concentrations from forested catchments (DID,
1989; Yusop et al., 2006) and in a catchment with diversified land uses,
including oil palm plantation (15%), forest (50%), mining, and urbanized
area (Gasim et al., 2006) (Table 12). As expected, deforestation greatly
increased outflow of sediment loads and nutrients after clearing (e.g., EC
(þ16%), Ca (þ26%), and Mg (þ37%) by DID, 1989; turbidity (9) and
suspended solids (12), Zulkifli et al., 1987). Temporal variations of stream
water quality at the storm event scale were noted by Yusop et al. (2006), in
particular higher export of nitrates (3) and suspended solids in stormflow
than in baseflow but greater export of SiO2 during baseflow, suggesting that
low flow removed solutes associated with soil weathering processes. At the
seasonal scale, Gasim et al. (2006) observed higher values for most water
quality parameters in the wet season than the dry season, while DID (1989)
observed higher values for turbidity, suspended solids, and iron in wet season
and higher values of conductivity, pH, Mg, and Ca during dry months.
In large oil palm plantations, POME is released directly to streams,
sometimes without treatment, which is expected to cause water pollution
(cf. Section 2.2.4). To our knowledge, the study by Olaleye and Adedeji
(2005) is the only one published in the peer-reviewed literature to assess the
water quality of a river impacted by POME release from oil palm planta-
tions. They ascribed the near absence of zooplankton in a large Nigerian
river to the deleterious effect of POME discharge in the stream. Pesticides
originating from oil palm plantations are expected to have a strong impact
on water quality according to NGOs, while oil palm managers expect low
impact due to low application rates. The absence of data on this topic in the
peer-reviewed literature is a major knowledge gap.
Table 12 Water quality in streams as impacted by oil palm plantation, managed and natural forest catchments in Southeast Asia (Malaysia)

Annual
rainfall EC (ms Turbidity DO TDS TSS K Ca Mg Na NH3 NO3 PO4 Cl SO4 SiO2
Land use Soil (mm) Notes pH cm
1) (NTU) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) (mgl
1) Source

Variously
2235 3.2– 14.3– 4.7–28.7 0.3–6.4 22.7– 1.2– 0.007– 0.7–2.9 0.0– 0.0–2.0 Gasim
vegetated 6.3 85.7 184 79.1 0.57 0.50 et al.
with 50% (2006)
forest and
15% oil
palm
Two- Araceneous 2348– Low 5.6 7.3– 0.34– 0.17– 0.32– 0.25– 0.03 0.08– 0.1 0.4–0.5 0.005 9.23– Yusop
forested series 3169 flow 7.5 0.38 0.19 0.35 0.28 0.1 9.24 et al.
catchments (derived Storm 5.1– 10.2– 0.83– 0.2– 0.37– 0.2– 0.04 0.23– 0.1 0.6 0.005 5.15– (2006)
(30ha) from flow 5.2 11.6 0.92 0.24 0.45 0.21 0.32 6.34
sandstone)
Forested Tropeptic 1878 6years 6.9 55.7 48.6 30.1 1.48 6.81 2.48 3.36 1.29 26.50 DID
control Harplothox average (1989)
catchment
(56ha)
Cleared Tropeptic 1878 6years 7.0 83.6 43.2 47.8 2.56 9.04 3.83 2.94 1.55 20.49 DID
catchment Harplothox average (1989)
for oil palm
(97ha)
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 111

Despite the potential risk of water pollution expected from oil palm
plantation activities, there have been very few studies at the watershed scale
to assess water quality in streams within a plantation at different develop-
ment stages (i.e., immature vs. mature palms).

4.7. Synthesis
There is an abundance of literature on hydrological processes in tropical
rainforest ecosystems, immediate and short-term impacts of rainforest distur-
bance (logging, clearing), as depicted in Fig. 5. It is generally accepted that
natural regrowth in tropical forests leads to a relatively fast return to previous
levels of soil infiltrability, streamflow, water budget, and soil nutrient stocks.
However, the impact of oil palm establishment on changes to hydrological
processes and associated nutrient losses and their evolution during oil palm
growth are much less investigated, documented, and understood.
Table 13 summarizes the evolution of hydrological processes and asso-
ciated nutrient losses occurring from forested land to mature oil palm planta-
tion stage, based on observed or expected outcomes and highlights research
gaps in the understanding of these processes. We compare consecutive stages
(cleared land vs. tropical forest; immature plantation vs. cleared land; mature
vs. immature plantation) and also compare both immature and mature stages
to forest, the original land cover. Expected trends for each stage are described
qualitatively because observations were often made for a specific plantation
age (without long-term monitoring to cover all stages) across areas with a
broad range of climatic and soil conditions. The impacts of forest clearing rely
on many studies reviewed by Bruijnzeel (1990, 1991) that generally reported
strong impacts of complete forest clearing. Information regarding the imma-
ture oil palm stage is based on the study by DID (1989), that represents, to our
knowledge, the only chronological study focused on the evolution of hydro-
logical process dynamics from forest to oil palm plantation establishment, at
both plot scale and watershed scale. Unfortunately, it was stopped before oil
palms reached maturity and did not investigate leaching process at the plot
scale. It concluded that after clearing, the growth of oil palm (with LCC)
tends to counteract the negative impacts of clearance, without always return-
ing to predisturbance levels. Runoff and erosion remain high in compacted
areas such as roads, harvest paths, and weeded circles. Due to the high
nutrient-uptake rate and large evaporative demand of the palms, low-nutrient
losses via leaching were generally reported in oil palm plantations in Southeast
Asia despite high-rainfall intensities.
Few studies compared the water and nutrient budgets between young
and mature oil palm stands, although leaching losses at the plot scale were
examined by Foong et al. (1983) and Omoti et al. (1983) and ET was
measured by Yusop et al. (2008). In mature plantations, data were available
from a number of plot scale focusing on single hydrological processes, such
112 Irina Comte et al.

Forest clearing

Loss of vegetal cover Soil compaction

and litter layer from machinery

Reduced Reduced soil

transpiration infiltration
demand

Reduced sub Increased

surface flow surface runoff

Increased soil
moisture Increased soil Increased
erosion nutrient
losses

Increased
sediment
loads

Reduced Increased
groundwater stormflow
recharge

Increased Reduced dry Increased

baseflow season baseflow localized
floods

Figure 5 Hydrological impacts of forest clearing (modified from Henson, 1999).

as runoff or leaching, with the exception of Banabas et al. (2008). Investi-

gating water budget in a mature plantation, including ET, soil water storage,
runoff, and leaching losses, authors demonstrated that nutrient losses
occurred primarily from leaching rather than from runoff. Few studies
have taken in account the spatial heterogeneity of the plantation. Finally,
hydrological studies carried out at the watershed scale in mature oil palm
plantation are almost nonexistent, with the exception of Yusop et al. (2008)
who quantified runoff processes on a small oil palm watershed (8.2ha).
The only research on stream water quality within oil palm plantations
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 113

Table 13 Qualitative description of the expected change in hydrological processes

and associated nutrient losses from forest clearing to mature oil palm plantation,
compared to undisturbed tropical forest

Clearance Immature oil palm Mature oil palm

(bare soil) plantation plantation
Versus Versus Versus Versus Versus
Hydrological process forest clearance forest immature forest
Plot scale
Infiltration rate &a,b %b "b
lc,d
Leaching %a &ea "ee &e or¼ f "g,h
Actual evapotranspiration &a %eb #ea,i
¼a,c
Runoff and erosion %a,b &b "b
"eg
Catchment scale
Water yields %b &b "b


Dry season flow &a




Stormflow %a &a,b "ea
"eg
Peakflow %b &b "b
"eg
Time to peak &b %b



Nutrient outflow %a,b &b "b
"eg
Sediment loads in stream %b &b "b &ea,b "eg
&, decrease; %, increase; ", higher; #, lower; l, variable; e, expected; ¼, similar.
a
Bruijnzeel (1990).
b
DID (1989).
c
Banabas et al. (2008).
d
Maena et al. (1979).
e
Foong (1993).
f
Omoti et al. (1983).
g
ECD (2000).
h
Schroth et al. (2000).
i
Ling (1979) (quoted by Bruijnzeel, 1990).

agroecosystems comes from DID (1989) for immature plantations. Thus,

hydrological process dynamics and magnitude (e.g., total water yields, dry
season baseflow, stormflow dynamics) and nutrient outflows from oil palm
plantation are far from being fully assessed and understood.

5. Conclusion
Since the 1960s, research effort focused on plot-scale trials in South-
east Asia to provide agronomic recommendations for plantation managers
that would increase productivity and economic returns for the palm oil
industry. Growing awareness of environmental impacts from the rapidly
114 Irina Comte et al.

expanding oil palm sector, driven by media and socioenvironmental

NGOs, led to the creation of RSPO to promote a sustainable palm oil
production. This organization encourages planters to assess the environ-
mental impacts of oil palm cultivation and develop eco-friendly agricul-
tural practices. Although RSPO encourages an evaluation of oil palm
plantation activities impacting water quality and hydrological processes,
this review demonstrated that the topic remains largely underinvestigated.
First of all, the actual agricultural practices for nutrient and water manage-
ment currently used in Southeast Asian oil palm plantations are poorly
described, especially in smallholdings. Assessing actual agricultural prac-
tices is challenging as high variability likely occurs not only between large
companies and smallholders but also within both production systems, due
to variable access to knowledge, technical, and financial means. Another
constraint is that palm oil is produced in developing countries, which may
lack the resources to monitor the impact of oil palm plantation on
hydrological functions at different stages throughout its long lifespan
(about 25years). Indeed, most of hydrological studies in oil palm planta-
tions were carried out at the plot scale (i.e., a few hectares), whereas oil
palm plantations can reach thousands contiguous hectare across several
watersheds. Few studies provided an integrated view of hydrological
processes or have taken account of the intrinsic spatial variability of an
oil palm plantation. Spatiotemporal variation in surface water quality and
groundwater quality within oil palm plantations has been very poorly
investigated, and the link to agricultural practices remains tenuous. There-
fore, study cases that include a survey of actual agricultural practices, water
quality assessment, and hydrological processes investigation at the water-
shed scale are needed to better understand and assess the potential risk to
waterways of oil palm plantations. In the end, this information will help
planters to manage their oil palm plantations more sustainably.

ACKNOWLEDGMENTS
This study was partly funded by the International Cooperation Centre in Agronomic
Research for Development (CIRAD) and the Natural Sciences and Engineering Research
Council of Canada (NSERC).

REFERENCES
Abdul Rahim, N., and Zulkifli, Y. (1987). Stream water quality of undisturbed forest catche-
ments in Peninsular Malaysia. In “Workshop on Impacts of man’s activities on tropical
upland forest ecosystems” (H. Yusuf, et al., Eds.), pp. 289–308. UPM, Serdang.
Abdul Rahim, N., and Harding, D. (1992). Effects of selective logging methods on water
yield and streamflow parameters in Peninsular Malaysia. J. Trop. For. Sci. 5, 130–154.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 115

Agamuthu, P., and Broughton, W. J. (1985). Nutrient cycling within the developing oil
palm-legume ecosystem. Agric. Ecosyst. Environ. 13, 111–123.
Ah Tung, P. G., Yusoff, M. K., Majid, N. M., Joo, G. K., and Huang, G. H. (2009). Effect
of N and K fertilizers on nutrient leaching and groundwater quality under mature oil
palm in Sabah during the monsoon period. Am. J. Appl. Sci. 6, 1788–1799.
Ahmad, A. L., Chong, M. F., Bhatia, S., and Ismail, S. (2006). Drinking water reclama-
tion from palm oil mill effluent (POME) using membrane technology. Desalination
191, 35–44.
Aldrian, E., and Dwi Susanto, R. (2003). Identification of three dominant rainfall regions
within Indonesia and their relationship to sea surface temperature. Int. J. Climatol. 23,
1435–1452.
Anderson, J. M., and Spencer, T. (1991). Carbon, nutrient and water balances of tropical rain
forest ecosystems subject to disturbance. In “MAB digest 7,” p. 95. UNESCO, Paris.
Andriesse, J. P. (1988). Nature and management of tropical peat soils. FAO Soils Bulletin 59,
Rome.
APOC (American Palm Oil Council). Available at: http://www.americanpalmoil.com/
(consulted on March 7th, 2011).
Banabas, M. (2007). Study of nitrogen loss pathway in oil palm (Elaeis guineensis Jacq.)
growing agro-syatems on volcanic ash soils in Papua New Guinea. PhD thesis, Massey
University, Palmerston North.
Banabas, M., Turner, M. A., Scotter, D. R., and Nelson, P. N. (2008). Losses of nitrogen
fertiliser under oil palm in Papua New Guinea: 1. Water balance, and nitrogen in soil
solution and runoff. Aust. J. Soil Res. 46, 332–339.
Bangun, D. (2006). In “Indonesian palm oil industry”. Paper Presented at the National
Institute of Oilseed Products Annual Convention, Phoenix, Arizona.
Barnett, R. (2007). Top 10 Outcomes 2007. 2011 World Resources Institute, Washington,
DC, USA.
Basiron, Y. (2007). Palm oil production through sustainable plantations. Eur. J. Lipid Sci.
Tech. 109, 289–295.
Bidin, K., Douglas, I., and Greer, T. (1993). Dynamic response of subsurface water levels in a
zero-order tropical rainforest basin, Sabah, Malaysia. In “Hydrology of Warm Humid
Regions” ( J. S. Gladwell, Ed.), Vol. 216, pp. 491–496. IAHS Pub., Wallingford.
Bilotta, G. S., and Brazier, R. E. (2008). Understanding the influence of suspended solids on
water quality and aquatic biota. Water Res. 42, 2849–2861.
Bonnell, M. (2005). Runoff geenration in tropical forests. In “Forest-Water-People in the
Humid Tropics: Pas, Present and Future Hydrological Research for Integrated Land and
Water Management” (M. Bonnell and L. A. Bruinjnzeel, Eds.), pp. 314–406. Cambridge
University Press, Cambridge.
Breure, K. (2003). The search for yield in oil palm: Basic principles. In “Oil Palm:
Management for Large and Sustainable Yields” (T. Fairhurst and R. Hardter, Eds.),
pp. 59–98. Potash & Phosphate Institute/Potash Institute of Canada and International
Potash Institute, Singapore.
Briggs, A. O., Stanley, H. O., Adiukwu, P. U., and Ideriah, T. J. K. (2007). Impact of palm
oil (Elaeis guineensis Jacq; Banga) mill effluent on water quality of receiving Oloya lake in
Niger Delta, Nigeria. Res. J. Appl. Sci. 2, 842–845.
Brouwer, L. C., and Riezebos, H. T. (1998). Nutrient dynamics in intact and logged tropical
rain forest in Guyana. In “Soils of Tropical Forest Ecosystems. Characteristics, Ecology
and Management” (A. Schulte and D. Ruhiyat, Eds.), pp. 73–86. Springer-Verlag Berlin
Heidelberg, New York.
Bruijnzeel, L. A. (1983a). Hydrological and biogeochemical aspects of man-made forests in
South-central Java, Indonesia. PhD thesis, Free University, Amsterdam. p. 256.
116 Irina Comte et al.

Bruijnzeel, L. A. (1983b). The chemical balance of a small basin in a wet monsoonal

environment and the effet of fast-growing plantation forest. Int. Assoc. Hydrolog. Sci.
Publ. 141, 229–239.
Bruijnzeel, L. A. (1984). Immobilization of nutrients in plantation forests of Pinus merkusii
and Agathis dammara growing on volcanic soils in central Java, Indonesia. In “Soils and
Nutrition of Perennial Crops” (A. Thajib and E. Pushparadjah, Eds.), pp. 19–29.
Malaysian Soil Science Society, Kuala Lampur.
Bruijnzeel, L. A. (1990). Hydrology of Moist Tropical Forests and Effects of Conversion: A
State of Knowledge Review. Humide Tropics Programme, UNESCO International
Hydrological Programme, UNESCO, Paris.
Bruijnzeel, L. A. (1991). Nutrient input-output budgets of tropical forest ecosystems: A
review. J. Trop. Ecol. 7, 1–24.
Bruijnzeel, L. A. (2004). Hydrological functions of tropical forests: Not seeing the soil for the
trees? Agric. Ecosyst. Environ. 104, 185–228.
CAAL (Carbon Association of Australasia Limited). (2011). Sumatran Peat Swamp Forests.
Available at: http://caaltd.org/Rainforest/Indonesia/PeatSwampForests.aspx. (consulted
on Feb 22nd, 2011).
Caliman, J. P. (2001). Strategy for fertilizer management during low commodity price.
In “PIPOC International Palm Oil Congress—Agriculture”, pp. 295–312. Malaysian oil
palm board (MOPB), Kuala Lampur.
Caliman, J. P., Olivin, J., and Dufour, O. (1987). Dégradation des sols ferrallitique sableux
en culture de palmiers à huile par acidification et compaction. Oléagineux Corps gras
Lipides 42, 393–398.
Caliman, J. P., Daniel, C., and Tailliez, B. (1994). Oil palm nutrition. Plant. Rech. Dev. 1,
36–54.
Caliman, J. P., Budi, M., and Saletes, S. (2001). Dynamics of nutrient release from empty
fruit bunches in field conditions and soil characteristics changes. In “MPOB International
Palm Oil Congress (PIPOC)”, Kuala Lampur.
Caliman, J. P., Carcasses, R., Perel, N., Wohlfahrt, J., Girardin, P., Wahyu, A., Pujianto,
Dubos, B., and Verwilghen, A. (2007). Agri-environmental indicators for sustainable
palm oil production. Palmas 28, 434–445.
Campiglia, E., Mancinelli, R., Radicetti, E., and Marinari, S. (2010). Legume cover crops
and mulches: Effects on nitrate leaching and nitrogen input in a pepper crop (Capsicum
annuum L.). Nutr. Cycl. Agroecosyst. 89, 399–412.
Chang, K. C., and Zakaria, A. (1986). Leaching losses of N and K fertilisers from mature fields
of perennial crop in Malaysia—A review of local work. The Planter 62, 468–487.
Chew, P. S., and Pushparajah, E. (1995). Nitrogen management and fertilization of tropical
plantation tree crops. In “Nitrogen Fertilization in the Environment” (P. E. Bacon, Ed.),
pp. 225–293. Marcel Drekker Inc., New York.
Chiew, L. K., and Rahman, Z. A. (2002). The effects of oil palm empty fruit bunches on oil
palm nutrition and yield, and soil chemical properties. J. Oil Palm Res. 14, 1–9.
Chorley, R. J., Schumm, S. A., and Sugden, D. E. (1984). Geomorphology Methuen & Co.
Ltd, Cambridge.
Clark, D. B., Clark, D. A., Brown, S., Oberbauer, S. F., and Veldkamp, E. (2002). Stocks
and flows of coarse woody debris across a tropical rain forest nutrient and topography
gradient. For. Ecol. Manage. 164, 237–248.
Clay, J. (2004). Palm oil. In “World Agricultural and the Environment: A Commodity by
Commodity Guide to Impacts and Practices” (Jason Clay, Ed.), pp. 203–235. Island
Press, Wahsington, DC, USA.
Corley, R. H. V., and Tinker, P. B. (2003). “The Oil Palm”, 4th edn., Blackwell publishing,
Oxford.
Corley, R. H. V., Gray, B. S., and Kee, N. S. (1971). Productivity of the oil palm (Elaeis
guineensis Jacq.) in Malaysia. Exp. Agric. 7, 129–136. Cambridge Journals Online.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 117

Crowther, J. (1987a). Ecological observations in tropical karst terrain, West Malaysia. II.
Rainfall interception, litterfall and nutrient cycling. J. Biogeogr. 14, 145–155.
Crowther, J. (1987b). Ecological observations in tropical karst terrain. West Malaysia. III.
Dynamics of the vegetation-soil-bedrock system. J. Biogeogr. 14, 157–164.
Danielsen, F., Beukema, H., Burgess, N. D., Parish, F., Brühl, C. A., Donald, P. F.,
Mudiyarso, D., Phalan, B., Reijnders, L., Struebig, M., and Fitzherbert, E. B. (2008).
Biofuel plantations on forested lands: Double jeopardy for biodiversity and climate.
Conserv. Biol. 23, 348–358.
Darussamin, A. (2004). Impact assessment on oil palm development. In “The 2nd Routd-
table Meeting on Sustainable Palm Oil”, Jakarta.
DE (Down to Earth). (2005). Pesticide Use in Oil Palm Plantations. N 66. Available at:
http://dte.gn.apc.org/66pes.htm. (consulted on March 8th, 2011).
DID (Department of Irrigation and Drainage). (1989). In “Sungai Tekam Experimental
Basin Final Report, July 1977 to June 1986. Malaysian Ministry of Agriculture, Depart-
ment of Irrigation and Drainage, Kuala Lampur”.
Dolmat, M. T., Lim, K. H., Zakaria, Z. Z., and Hassan, A. H. (1987). Recent studies on the
effects of Land Application of Palm Oil Mill Effluent on oil palm and the environment.
In “Proceedings of the International Oil Palm/Palm Oil Conferences” (H. A. H. Halim,
P. S. Chew, B. J. Wood, and E. Pushparaja, Eds.), Vol. 2, pp. 596–604. PORIM and
Incorporated Society of Planters, Kuala Lampur.
Douglas, I. (1999). Hydrological investigations of forest disturbance and land cover impacts
in South-East Asia: A review. Philos. Trans. R. Soc. Lond. B Biol. Sci. 354, 1725–1738.
Dufrêne, E. (1989). Photosynthèse, consommation en eau et modélisation de la production
chez le palmier à huile (Elaeis guineensis Jacq.). PhD Thesis, Université de Paris XI, Orsay,
Paris, pp. 156.
Dufrêne, E., Ochs, R., and Saugier, B. (1990). Oil palm photosynthesis and productivity
linked to climatic factors. Oleagineux 45, 345–353.
ECD (Environment Conservation Department). (2000). Environmental Impact Assessment
(EIA) Guidelines on Oil Palm Plantation Development. Environmental Conservation
Department, Sabah, Malaysia. Available at: http://www.sabah.gov.my/jpas/programs/
ecd-cab/technical/OP211100.pdf. (consulted on Apr 13th, 2011).
Elsenbeer, H. (2001). Hydrologic flowpaths in tropical rainforest soilscapes: A review.
Hydrol. Processes 15, 1751–1759.
Fairhurst, T. (1996). Management of Nutrients for Efficient Use in Smallholder Oil Palm
Plantations. Wye College, London.
Fairhurst, T., and Mutert, E. (1999). Introduction to oil palm production. Better Crops Int.
13, 3–6.
Fairhust, T., Caliman, J.-P., Härdter, R., and Witt, C. (2005). Oil Palm: Nutrient Disorders
and Nutrient Managemen: diagnosis, causes, prevention, treatmentt. Potash & Phosphate
Institute/Potash & Phosphate Institute of Canada and International Potash Institute (PPI/
PPIC and IPI), Singapore.
FAO (Food and Agriculture Organization of the United Nations). (1990). Irrigation Water
Management: Irrigation methods, by C. Brouwer, K. Prins , M. Kay & M. Heibloem.
Irrigation water management Training manuals No. 5. Rome.
FAO (Food and Agriculture Organization of the United Nations). (2005). Fertilizer Use by
Crop in Indonesia, FAO, Natural Resources Management and Environment. Available
at: http://www.fao.org/docrep/008/y7063e/y7063e00.htm. (consulted on Feb 14th,
2011).
FAOSTATS, 2011. Available at: http://faostats.fao.org. (consulted on Feb 12th, 2011).
Foong, S. F. (1993). Potentail evaporation, potential yields and leaching losses of oil palm.
In “Proceeding of the 1991 PORIM International Palm Oil Conference” (Y. Basiron,
S. Jalani, K. C. Chang, S. C. Cheah, I. E. Henson, N. Kamarudin, K. Paranjothy,
118 Irina Comte et al.

N. Rajanaidu, and D. Tayeb, Eds.), pp. 105–119. Agriculture Palm Oil Research
Institute, Kuala Lampur.
Foong, S. F., Syed Sofi, S. O., and Tan, P. Y. (1983). A lysimetric simulation of leaching
losses from an oil palm field. In “Proceedings of the Seminar on Fertilizers in Malaysian
Agriculture” (C. P. Soon, K. B. Lian, W. S. W. Harun, and Z. Z. Zakaria, Eds.),
pp. 45–68. Malaysian Society of Soil Science, Kuala Lampur.
Foster, H. L. (1995). Experience with fertilizer recommendation systems for oil palm.
In “Proceedings of the 1993 PORIM International Palm Oil Congress—Agriculture”
(B. S. Jalani, et al., Eds.), pp. 313–328. Palm Oil Research Institute, Kuala Lampur.
Foster, H. L., and Chang, K. C. (1977). The diagnosis of the nutrient status of oil palms in
West Malaysia. In “International Development in Oil Palms” (D. A. Earp and
W. Newall, Eds.), pp. 290–312. Incorporated Society of Planters, Kuala Lampur.
Foster, H. L., Tarmizi Mohammed, A., and Zin, Z. Z. (1986). Fertilizer recommendations
for oil palm in Peninsular Malaysia. PORIM Techonol. 13, 1–25.
Foster, H. L., Tarmizi Mohammed, A., and Zakaria, Z. Z. (1988). Foliar diagnosis of
oil palm in Peninsular Malaysia. In “Proceedings of the 1987 International Oil Palm/
Palm Oil Conference—Agriculture” (H. A. H. Halim, P. S. Chew, B. J. Wood, and
E. Pushparajah, Eds.), pp. 244–261. PORIM and Incorporated Society of Planters, Kuala
Lampur.
Galdikas, B. M. (2009). Indonesian monsoons. Climate, Forest Ecology and Orangutans.
Available at: http://www.orangutan.org/archives/550. (consulted on Feb 14th, 2011).
Gasim, M. B., Toriman, M. E., Rahim, S. A., Islam, M. S., Chek, T. C., and Juahir, H.
(2006). Hydrology and water quality and land-use assessment of Tasik Chini’s feeder
rivers, Malaysia. Geografia 3, 1–16.
Germer, J., and Sauerborn, J. (2008). Estimation of the impact of oil palm plantation
establishment on greenhouse gas balance. Environ. Dev. Sustainability 10, 697–716.
Gillison, A., and Liswanti, N. (1999). Impact of Oil Palm Plantations on Biodiversity in
Jambi, Central Sumatra, Indonesia. CIFOR. Available at: http://www.asb.cgiar.org/
data/dataset/IDA1AMZB.htm. (consulted on March 9th, 2011).
Glastra, R., Wakker, E., and Richert, W. (2002). Oil Palm Plantations and Deforestation in
Indonesia. What Role Do Europe and Germany Play? WWF, Germany, Amsterdam,
Netherlands.
Godbold, D. L., Fritz, E., and Hüttermann, A. (1988). Aluminum toxicity and forest decline.
Proc. Natl. Acad. Sci. USA 85, 3888–3892.
Goh, K. J. (2004). Fertilizer recommendation systems for oil palm: Estimating the fertilizer
rates. In “Proceedings of MOSTA Best Practices Workshops: Agronomy and Crop
Management” (C. P. Soon and T. Y. Pau, Eds.), pp. 235–268. Malaysian Oil Scientists
and Technologies Association, Kuala Lampur.
Goh, K. J., and Chew, P. S. (1995). Managing soils for plantation tree crops. 1. General soil
management. In “Course on Soil Survey and Managing Tropical Soils” (S. Paramanathan,
Ed.), pp. 228–245. MSSS and PASS, Kuala Lampur.
Goh, K. J., and Härdter, R. (2003). General oil palm nutrition. In “Oil Palm: Management
for Large and Sustainable Yields” (T. Fairhurst and R. Hardter, Eds.), pp. 191–230. PPI/
PPIC and IPI, Singapore.
Goh, K. J., Teo, C. B., Chew, P. S., and Chiu, S. B. (1999). Fertiliser management in oil
palm: Agronomic principles and field practices. In “Fertiliser Management for Oil Palm
Plantations”, p. 44. ISP North-east Branch, Sandakan. Available at: http://www.aarsb.
com.my/AgroMgmt/OilPalm/FertMgmt/Principle/Fertiliser%20management%20in%
20oil%20palm–%20Agronomic%20Principles%20an%E2%80%A6.pdf (downloaded on
Jan 12 th 2011).
Goh, K. J., Härdter, R., and Fairhurst, T. (2003). Fertilizing for maximum return. In “Oil
Palm: Management for Large and Sustainable Yields” (T. Fairhurst and R. Hardter, Eds.),
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 119

pp. 279–306. Potash & Phosphate Institute/Potash & Phosphate Institute of Canada and
International Potash Institute (PPI/PPIC and IPI), Singapore.
Greenpeace. (2011). Deforestation for Palm Oil. Available at:, http://www.greenpeace.org/
usa/en/campaigns/forests/forests-worldwide/paradise-forests/our-work-in-paradise/
(consulted on April 4th, 2011).
Hansen, S. (2007). Feasibility study of performing an life cycle assessment on crude palm oil
production in Malaysia (9 pp). Int. J. LCA 12, 50–58.
Härdter, R., and Fairhurst, T. (2003). Introduction: Oil palm: Management for large and
sustainable yields. In “Oil Palm: Management for Large and Sustainable Yields”
(T. Fairhurst and R. Härdter, Eds.), pp. 1–12. PPI/PPIC IPI, Singapore.
Hartley, C. W. S. (1988). The Oil Palm Longman, London.
Henson, I. E. (1999). Comparative ecophysiology of oil palm and tropical rain forest. In “Oil
Palm and the Environment—A Malaysian Perspective” (G. Singh, L. Kim Huan,
T. Leng, and D. Lee Kow, Eds.), pp. 9–39. Malaysian oil palm grower’s concil, Kuala
Lampur.
Hooijer, A., Silvius, M., Wosten, H., and Page, S. (2006). Peat-CO2: Assessment of CO2
emissions from drained peatlands in South East Asia. Wetlands International. In “Delft
Hydraulics 36,” Delft, The Netherlands.
Hui, C. M. (2008). “Comparison of Rainfall Runoff Characteristics and Evapotranspiration
in Oil Palm Catchments”, p. 204. Universiti Teknologi Malaysia, Kuala Lampur,
Malaysia.
IEG (Independent Evaluation Group). (1993). Nucleus Estates and Smallholders Projects in
Indonesia the World Bank Group. Available at: http://lnweb90.worldbank.org/oed/
oeddoclib.nsf/DocUNIDViewForJavaSearch/95d104dd2107d21d852567f5005d8461?
OpenDocument&Click¼ (consulted on Jan 10th, 2011).
IMA (Indonesian Ministry of Agriculture). (2010). Area and Production by Category of
Producers: Palm Oil, Direktorat Jenderal Perkebunan. Kementerian Pertanian. Available
at: http://ditjenbun.deptan.go.id/index.php/direktori/3-isi/4-kelapa-sawit.html. (con-
sulted on April 13th, 2011).
Jacquemard, J. C. (1995). Le palmier à huile. Maisonneuve et Larose, Paris.
Johnson, D., Cole, D. W., and Gessel, S. P. (1975). Processes of nutrient transfer in a tropical
rainforest. Biotropica 7, 208–215.
Kee, K. K., and Chew, P. S. (1996). Nutrient losses through surface runoff and soil
erosion—Implications for improved fertilizer efficiency in mature oil palms.
In “Proceedings of the PORIM Internation Palm Oil Congress” (A. Ariffin,
M. B. Wahid, N. Rajanaidu, D. Tayeb, K. Paranjothy, S. C. Cheah, K. C. Chang,
and S. Ravigadevi, Eds.), pp. 153–169. Palm Oil Research Institute of Malaysia, Kuala
Lampur.
Kee, K. K., and Goh, K. J. (2006). Efficient fertiliser management for higher productivity
and sustainability in oil palm production. In “Proceedings of the International Planters
Conference on Higher Productivity and Efficient Practices for Sustainable Plantation
Agriculture”, Technical Papers, Vol. 1, pp. 157–182. Incorporated Society of Planters,
Kuala Lampur.
Kee, K. K., Goh, K. J., Chew, P. S., and Tey, S. H. (1994). An integrated site specific
fertilizer recommendation system (INFERS) for high productivity in mature oil palms.
In “Management for Enhanced Profitability in Plantations” (K. H. Chee, Ed.),
pp. 83–100. The Incorporated society of Planters, Kuala Lampur.
Kee, K. K., Goh, K. J., and Chew, P. S. (2000). Water cycling and balance in a mature oil
palm agroecosystem in Malaysia. In “Proceedings of the International Planters Confer-
ence” (E. Pushparajah, Ed.), pp. 153–169. The Incorporated society of Planters, Kuala
Lampur.
120 Irina Comte et al.

Kenworthy, J. B. (1971). Water and nutrient cycling in a tropical rain forest. In “The Water
Relations of Malesian Forests” (J. R. Flenley, Ed.), Miscellaneous Series No. 11, p. 11.
University of Hull.
Kemp, P., Sear, D., Collins, A., Naden, P., and Jones, I. (2011). The impacts of fine
sediment on riverine fish. Hydrol. Processes 25, 1800–1821. 10.1002/hyp.7940.
Khalid, H., Zin, Z. Z., and Anderson, J. M. (2000). Nutrient cycling in an oil palm
plantation: The effects of residue management practices during replanting on dry matter
and nutrient uptake of young palms. J. Oil Palm Res. 12, 29–37.
Koh, L. P., and Ghazoul, J. (2008). Biofuels, biodiversity, and people: Understanding the
conflicts and finding opportunities. Biol. Conserv. 141, 2450–2460.
Koh, L. P., and Wilcove, D. S. (2008). Is oil palm agriculture really destroying tropical
biodiversity?. Conserv. Lett. 2, 1–5.
Kok, T. F., Goh, K. J., Chew, P. S., Gan, H. H., Heng, Y. C., Tey, S. H., and Kee, K. K.
(2000). Advances in oil palm agronomic recommendations. In “Plantation Tree Crop in
the New Millennium: The Way Ahead” (E. Pushparajah, Ed.), pp. 215–232. The
Incorporated society of Planters, Kuala Lampur.
Lamade, E., and Bouillet, J.-P. (2005). Carbon storage and global change: The role of oil
palm. Oléaginaux Corps Gras Lipides 12, 154–160.
Ling, A. H. (1979). Some Lysimetric measurements of evapotranspiration of oil palm in
Central Peninsular Malaysia. In “Proceedings Symposium on Water in Malaysian agri-
culture”, pp. 89–99. Malaysian Soil Science Society, Kuala Lampur.
Loong, S. G., Mazeeb, M., and Letchumanan, A. (1987). Optimising the use of EFB mulch
on oil palms on two different soils, p. 24. PORIM International Palm Oil Development
Conference, Kuala Lampur.
Lord, S., and Clay, J. (2006). Environmental impacts of oil palm—Practical considerations in
defining sustainability for impacts on air, land and water. In “International Planters
Conference on Higher Productivity and Efficient Practices for Sustainable Agriculture,”
The Incorporated Society of Planters, Putrajaya.
Low, K. S., and Goh, G. C. (1972). The water balance of five catchments in Selangor.
J. Trop. Geogr. 35, 60–66.
Mackensen, J., Hölscher, D., Klinge, R., and Fölster, H. (1996). Nutrient transfer
to the atmosphere by burning of debris in eastern Amazonia. For. Ecol. Manage. 86,
121–128.
Maena, L. M., Thong, K. C., Ong, T. S., and Mokhtaruddin, A. M. (1979). Surface wash
under mature oil palm. In “Proceedings of the Symposium on Water Agriculture in
Malaysia” (E. Pushparajah, Ed.), pp. 203–216. Malaysian Society of Soil Science, Kuala
Lampur.
Malmer, A. (1996). Hydrological effects and nutrient losses of forest plantation establishment
on tropical rainforest land in Sabah, Malaysia. J. Hydrol. 174, 129–148.
Malmer, A., and Grip, H. (1994). Converting tropical rainforest to forest plantation in Sabah,
Malaysia. Part II. Effects on nutrient dynamics and net losses in streamwater. Hydrol.
Processes 8, 195–209.
Ministerial Decree on Agriculture. (2009). Peraturan Menteri Pertanian nomor: 14/Per-
mentan/PL.110/2/2009.
Mongabay Editorial. (2006). Forest Fires Result from Government Failure in Indonesia.
Available at: http://news.mongabay.com/2006/1015-indonesia.html. (consulted on
March 4th, 2011).
Mutert, E., Fairhurst, T. H., and von Uexküll, H. R. (1999). Agronomic management of oil
palms on deep peat. Better Crops Int. 13, 22–27.
Nantha, H. S., and Tisdell, C. (2009). The orangutan–oil palm conflict: Economic con-
straints and opportunities for conservation. Biodivers. Conserv. 18, 487–502.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 121

Nellemann, C., Miles, L., Kaltenborn, B. P., Virtue, M., and Ahlenius, H. (2007). The Last
Stand of the Orangutan—State of Emergency: Illegal Logging, Fire and Palm Oil in
Indonesia’s National Parks. UNEP GRID-Arendal, UNEP-WCMC, UNESCO, Vol.
2011.
Ng, S. K. (2002). Nutrition and nutrient management of the oil palm. New thrust for the future
perspective. In “Potassium for Sustainable Crop Production. International Symposium on
Role of Potassium in India” (N. S. Pasricha and S. K. Bansal, Eds.), pp. 415–429. Potash
Research Institute of India, and International Potash Institute, New Delhi.
Ng, S. K., and Thamboo, S. (1967). Nutrient contents of oil palms in Malaysia. I. Nutrients
in reproductive tissue fruit bunches and male inflorescence. Malays. Agric. J. 46, 3–15.
Ng, S. K., Thamboo, S., and de Souza, P. (1968). Nutrient contents of oil palms in Malaysia.
II. Nutrients in reproductive tissue, fruit bunches and male inflorescence. Malays. Agric. J.
46, 332–391.
Ng, P. H. C., Chew, P. S., Goh, K. J., and Kee, K. K. (1999). Nutrient requirements and
sustainability in mature oil palms—An assessment. The Planter 75, 331–345.
Ng, S. K., von Uexküll, H., and Härdter, R. (2003). Botanical aspects of the oil palm
relevant to crop management. In “Oil Palm: Management for Large and Sustainable
Yields” (T. Fairhurst and R. Härdter, Eds.), pp. 13–26. Potash & Phosphate Institute/
Potash Institute of Canada and International Potash Institute, Singapore.
Noguchi, S., Nik, A., Yusop, Z., Tani, M., and Sammori, T. (1997a). Rainfall-runoff
responses and roles of soil moisture variations to the response in tropical rain forest,
bukit tarek, peninsular Malaysia. J. Forest Res. 2, 125–132.
Noguchi, S., Rahim Nik, A., Kasran, B., Tani, M., Sammori, T., and Morisada, K. (1997b).
Soil physical properties and preferential flow pathways in tropical rain forest, Bukit
Tarek, Peninsular Malaysia. J. Forest. Res. 2, 115–120.
Noguchi, S., Abdul Rahim, N., Shamsuddin, S. A., Tani, M., and Sammori, T. (2004).
Evapotranspiration estimates of a tropical rain forest, Bukit Tarek Experimental Water-
shed in Peninsular Malaysia, using short-time period water-budget method. J. Jpn. Soc.
Hydrol. Water Resour. 17, 427–444.
Noor, Y. R., Cahyo Wibisono, I. T., and Suryadiputra, I. N. (2007). Poverty alleviation
projects in peatland in Indonesia. In “Proceedings of the International Symposium and
Workshop on Tropical Peatland” (J. O. Rieley, C. J. Banks, and B. Ragkagukguk, Eds.),
pp. 177–186. Gadjah Mada University, Indonesia and University of Leicester, United
Kingdom, Yogyakarta, Indonesia, EU CARBOPEAT and RESTORPEAT Partnership.
Oil World, 2011. Available at: http://www.oilworld.biz/app.php?fid¼310&fpar¼YToy
OntzOjI6IklkIjtzOjQ6IjYzMTQiO3M6NDoicGNpZCI7czoyOiIxMiI7fQ%3D%3D
&isSSL¼0&aps¼0&blub¼26bc92353fe185973450e4160e44224d&ista¼apdxujfvn (con-
sulted on May 6th, 2011).
Okwute, L. O., and Isu, N. R. (2007). The environmental impact of palm oil mill effluent
(pome) on some physico-chemical parameters and total aerobic bioload of soil at a dump
site in Anyigba, Kogi State, Nigeria. Afr. J. Agric. Res. 2, 656–662.
Olaleye, V. F., and Adedeji, A. A. (2005). Planktonic and water quality of palm oil effluent
impacted river in Ondo state, Nigeria. Int. J. Zool. Res. 1, 15–20.
Omoti, U., Ataga, D., and Isenmila, A. (1983). Leaching losses of nutrients in oil palm
plantations determined by tension lysimeters. Plant Soil 73, 365–376.
Osuji, G. E., Okon, M. A., Chukwuma, M. C., and Nwarie, I. I. (2010). Infiltration
characteristics of soils under selected land use practices in Owerri, Southeastern Nigeria.
World J. Agric. Sci. 6, 322–326.
Othman, A., Mohammed, A. T., Harun, M. H., Darus, F. M., and Mos, H. (2010). Best
management practices for oil palm planting on peat: Optimum groundwater table.
Malaysian Palm Oil Board (MPOB) Information series 472.
122 Irina Comte et al.

Pushparajah, E. (1994). Leaf Analysis and Soil Testing for Plantation Tree Crops. Interna-
tional Board for Soil Research and Management (IBSRAM), Thailand. Available at:
http://www.agnet.org/library/eb/398/. (consulted on Jan 13th 2011).
Radersma, S., and deRidder, N. (1996). Computed evapotranspiration of annual and
perennial crops at different temporal and spatial scales using published parameter values.
Agric. Water Manag. 31, 17–34.
Rankine, I. R., and Fairhurst, T. (1998a). Field Handbook—Oil Palm Series. Nursery.
Vol. 1, Potash and Phosphate Institute, Singapore.
Rankine, I. R., and Fairhurst, T. (1998b). Field Handbook—Oil palm. Immature. Vol. 2,
Potash and Phosphate Institute, Singapore.
Rankine, I. R., and Fairhurst, T. (1998c). Field Handbook—Oil palm. Mature. Vol. 3,
Potash and Phosphate Institute, Singapore.
Rieley, J. O. (2007). Tropical peatland - The amazing dual ecosystem: coexistence and
mutual benefit. In “Proceedings of the international symposium and workshop on
tropical peatland” (J. O. Rieley, C. J. Banks, and B. Ragkagukguk, Eds.), pp. 1–14.
Gadjah Mada University, Indonesia and University of Leicester, United Kingdom,
Yogyakarta, Indonesia, EU CARBOPEAT and RESTORPEAT Partnership.
Rieley, J. O., and Page, S. O. (1997). Biodiversity and sustainibility of tropical peatlands.
In “Proceedings of the International Symposium on Biodiversity, Environmental Impor-
tance and Sustainability of Tropical peat and Peatlands. Palangka Raya, Central
Kalimantan, 4-8 September, 1995,” Samara Publishing Limited, Cardigan.
Rist, L., Feintrenie, L., and Levang, P. (2010). The livelihood impacts of oil palm: Small-
holders in Indonesia. Biodivers. Conserv. 19, 1009–1024.
Rupani, P. F., Singh, R. P., Ibrahim, M. H., and Esa, N. (2010). Review of current palm oil
mill effluent (POME) treatment methods: Vermicomposting as a sustainable practice.
World Appl. Sci. J. 11, 70–81.
Schindler, D. W., Armstrong, F. A. J., Holmgren, S. K., and Brunskil, G. J. (1971).
Eutrophication of lake 227, experimental lakes area, Northwester Ontario, by addition
of phosphate and nitrate. J. Fish. Res. Board Can. 28, 1763–1782.
Schrevel, A. (2008). Oil-Palm Estate Development in Southeast Asia: Consequences for Peat
Swamp Forests and Livelihoods in Indonesia. Available at: ftp://ftp.fao.org/docrep/fao/
011/i0314e/i0314e06.pdf. (consulted 10th Jan 2011).
Schroth, G., Rodrigues, M. R. L., and D’Angelo, S. A. (2000). Spatial patterns of nitrogen
mineralization, fertilizer distribution and roots explain nitrate leaching from mature
Amazonian oil palm plantation. Soil Use Manag. 16, 222–229.
Sheil, D., Casson, A., Maijaard, E., van Noordwijk, M., Gaskell, J., Sunderland-groves, J.,
Wertz, K., and Kanninen, M. (2009). The Impacts and Opportunities of Oil Palm in
Southeast Asia Center for International Forestry Research (CIFOR), Bogor.
Siegel, D. I., and Glaser, P. (2006). The hydrology of peatlands. In “Boreal Peatland
Ecosystems” (R. K. Wieder and D. H. Vitt, Eds.), pp. 289–311. Springer, Berlin
Heidelberg.
Singh, R. P., Ibrahim, M. H., Esa, N., and Iliyana, M. S. (2010). Addressing the threats
to biodiversity from oil-palm agriculture. Composting of waste from palm oil mill:
A sustainable waste management practice. Rev. Environ. Sci. Biotechnol. 9, 311–344.
Squire, G. R. (1984). Techniques in environmental physiology of oil palm. 2. Partitioning of
rainfall above ground. PORIM Bull. 9, 1–9.
Strack, M., and Waddington, J. M. (2007). Response of peatland carbon dioxide and
methane fluxes to a water table drawdown experiment. Global Biogeochem. Cycles 21,
GB1007.
Taillez, B. (1998). Oil Palm : A new crop. . . for what future? Oléagineux Corps gras Lipides 5,
106–109.
Tan, K. T., Lee, K. T., Mohamed, A. R., and Bhatia, S. (2009). Palm oil: Addressing issues
and towards sustainable development. Renew. Sustain. Energy Rev. 13, 420–427.
Agricultural Practices, Nutrient Fluxes and Water Quality in Oil Palm Plantations 123

Tanaka, N., Kume, T., Yoshifuji, N., Tanaka, K., Takizawa, H., Shiraki, K., Tantasirin, C.,
Tangtham, N., and Suzuki, M. (2008). A review of evapotranspiration estimates from
tropical forests in Thailand and adjacent regions. Agric. Forest. Meteorol. 148, 807–819.
Tarmizi, M. A., and Mohd, T. D. (2006). Nutrient demands of Tenera oil palm planted on
inlands soils of Malaysia. J. Oil Palm Res. 18, 204–209.
Turner, R. E., and Rabalais, N. N. (1994). Coastal eutrophication near the Mississippi river
delta. Nature 368, 619–621.
Turvey, N. D. (1974). Nutrient cycling under tropical rain forest in central Papua. Occa-
sional paper, 10. Department of Geography. University of Papua New Guinea, Port
Moresby, p. 96.
UN (United Nations). (2004). World Population to 2300. United Nations, Department
of Economic and Social Affairs, Population Division. Available at: http://www.un.org/
esa/population/publications/longrange2/WorldPop2300final.pdf. (consulted on March
9th, 2011).
USDA (United States Department of Agriculture). (2006). Keys to Soil Taxonomy.10th
edn.. United States Department of Agriculture and Natural Resources Conservation
Service, Washington, DC. Available at: http://terra-geog.lemig2.umontreal.ca/donnees/
geo2142/Keys%20taxonomy.pdf. (consulted on May 6th, 2011).
USDA (United States Department of Agriculture). (2007). Indonesian Palm Oil Production.
Available at: http://www.pecad.fas.usda.gov/highlights/2007/12/Indonesia_palmoil/.
(consulted on Jan 7th, 2011).
USDA (United States Department of Agriculture). (2009). Indonesia: Palm Oil Production
Growth to Continue. Available at: http://www.pecad.fas.usda.gov/highlights/2009/03/
Indonesia/. (consulted on March, 17th, 2011).
Vermeulen, S., and Goad, N. (2006). Towards better practices in smallholder palm oil
production. In “Natural Resource Issues Series”, 5, International Institute for Environ-
ment and Development (IIED), London.
von Uexküll, H. R. (2007). Oil Palm (Elaeis guineensis Jacq.). E. & S.E. Asia Program for the
Potash & Phosphate Institute/International Potash Institute, Singapore.
Wahid, M. B., Abdullah, S. N. A., and Henson, I. E. (2005). Oil palm—Achievement and
potential. Plant Prod. Sci. 8, 288–297.
Wakker, E. (1999). Forest Fires and the Expansion of Indonesia’s Oil-Palm Plantations.
WWF-Indonesia, Jakarta.
Webb, M. J., Nelson, P. N., Rogers, L. G., and Curry, G. N. (2011). Site-specific fertilizer
recommendations for oil palm smallholders using information from large plantations.
J. Plant Nutr. Soil Sci. 174, 311–320.
WEPA (Water Environment Partnership in Asia). (2011). State of Water Environmental
Issues. Available at: http://www.wepa-db.net/policies/state/indonesia/indonesia.htm.
(consulted on Feb 25th, 2011).
WHO (World Health Organization). (2008).Guidelines for drinking-water quality: Incor-
porating the 1st and 2nd addenda. 3rd edn. , Recommendations. 1, World Health
Organization, Geneva.
WI (Wetlands International). (2010). Input to World Bank Group Palm Oil Strategy
Consultations, World Bank Group Framework & IFC Strategy in the Palm Oil Sector.
Available at: http://www.ifc.org/ifcext/agriconsultation.nsf/Content/2E1FC2F0E9F1
D8FD8525772F0073CFB6?OpenDocument. (consulted on March 1st, 2011).
Wilkinson, G. E., and Aina, P. O. (1976). Infiltration of water into two Nigerian soils under
secondary forest and subsequent arable cropping. Geoderma 15, 51–59.
Wood, B. J., Pillai, K. R., and Rajaratnam, J. A. (1979). Palm oil mill effluent disposal on
land. Agric. Wastes 1, 103–127.
Wösten, J. H. M., Ismail, A. B., and van Wijk, A. L. M. (1997). Peat subsidence and its
practical implications: A case study in Malaysia. Geoderma 78, 25–36.
124 Irina Comte et al.

Wösten, J. H. M., Clymans, E., Page, S. E., Rieley, J. O., and Limin, S. H. (2008). Peat-
water interrelationships in a tropical peatland ecosystem in Southeast Asia. Catena 73,
212–224.
WRI (World Resource Institute). (2002). The State of the Forest: Indonesia. Available at:
http://www.wri.org/publication/state-of-the-forest-indonesia. (consulted on March
9th, 2011).
WRM (World Rainforest Movement. (2002). The Direct and Underlying Causes of Forest
Loss, WRM Briefings. Available at: http://www.wrm.org.uy/. (consulted on April 7th,
2011).
WWF (World Wild Fund). (2011). Orangutans Under Threat. How Unsustainable Palm Oil
Is Destroying the Orangutan’s Home. Available at: http://www.wwf.org.uk. (consulted
on March 9th, 2011).
Yi Jing, C., Mei Fong, C., and Chung Lim, L. A. W. (2010). Biological treatment of
anaerobically digested palm oil mill effluent (POME) using a Lab-Scale Sequencing Batch
Reactor (SBR). J. Environ. Manage. 91, 1738–1746.
Yusop, Z., Douglas, I., and Nik, A. R. (2006). Export of dissolved and undissolved nutrients
from forested catchments in Peninsular Malaysia. For. Ecol. Manage. 224, 26–44.
Yusop, Z., Chan, C. H., and Katimon, A. (2007). Runoff characteristics and application of
HEC-HMS for modelling stormflow hydrograph in an oil palm catchment. Water Sci.
Technol. 56, 41–48.
Yusop, Z., Hui, C. M., Garusu, G. J., and Katimon, A. (2008). Estimation of evapotranspi-
ration in oil palm catchment by short-time period water-budget method. Malays. J. Civil
Eng. 20, 160–174.
Zaidel’man, F. (2008). Protecting soils against degradation. Herald Russ. Acad. Sci. 78,
362–369.
Zen, Z., Barlow, C., and Gondowarsito, R. (2005). Oil Palm in Indonesian Socio-
Economic Improvement—A Review of Options. Department of Economics, Research
School of Pacific and Asian Studies, Australian National University, Canberra.
Zhang, X., Yu, X., Wu, S., and Liu, H. (2007). Effects of forest vegetation on runoff
and sediment transport of watershed in Loess area, west China. Front. Forest. China 2,
163–168.
Zulkifli, Y., Anhar, S., and Mohd Fauzi, Z. (1987). Effects of selective logging on physical
stream water quality in hill tropical rain forests. In “Paper Presented at Workshop on
Impact of Operations in Natural and Plantation Forests on Conservation of Soil and
Water Resources,” Universiti Pertanian Malaysia, Serdang, Malaysia. (23–26th June).
Zulkifli, Y., Abdul Rahim, N., Anhar, S., and Fauzi Zakaria, M. (1989). Rainfall chemistry
and nutrient loading in a Peninsular Malaysia forest site. J. Trop. For. Sci. 1, 201–214.
Zulkifli, Y., Abdul Rahim, N., and Baharuddin, K. (1998). Evapotranspiration loss from a
second growth forest of Peninsular Malaysia. In “International Conference of Hydrology
and Water resources of Humid Tropics,” Pangkor, Malaysia. (24–26th Nov).
 C H A P T E R F O U R

Pathways to Agroecological
Intensification of Soil Fertility
Management by Smallholder Farmers
in the Andean Highlands
Steven J. Fonte,* Steven J. Vanek,† Pedro Oyarzun,‡
Soroush Parsa,* D. Carolina Quintero,* Idupulapati M. Rao,* and
Patrick Lavelle*,§

Contents
1. Introduction: Agricultural and Soil Fertility Issues in the High Andes 126
1.1. Cropping systems of the Andes 126
1.2. Biophysical limitations and risks 127
1.3. Socioeconomic and cultural setting of Andean agriculture 129
1.4. Current challenges and emerging threats 131
1.5. Ecologically based intensification in the Andean context 132
2. Examining Soil Fertility and Management Strategies in
Smallholder Systems 133
2.1. General concept of soil fertility 133
2.2. Approaches to examining soil fertility 133
3. Additional Considerations for Soil Fertility Interventions 162
3.1. Need to incorporate co-limiting crop growth factors 162
3.2. Integrating local needs and knowledge into soil
fertility research 162
4. Conclusions and Recommendations 163
4.1. Outlook for agroecological intensification in the
Andean context 163
4.2. Recommendations for future research and interventions 165
Acknowledgments 166
References 166

* International Center for Tropical Agriculture (CIAT), Cali, Colombia

{
Department of Crop and Soil Science, Cornell University, Ithaca, New York, USA
{
EkoRural, Quito, Ecuador
}
Institut de Recherche sur le Développement (IRD)/Université Pierre et Marie Curie (UPMC), Paris, France

Advances in Agronomy, Volume 116 # 2012 Elsevier Inc.

ISSN 0065-2113, DOI: 10.1016/B978-0-12-394277-7.00004-X All rights reserved.

125
126 Steven J. Fonte et al.

Abstract
Small farmers in the high Andes (>2500m) of Bolivia, Ecuador, and Peru face
increasing threats to their livelihoods due to land degradation, climate change,
and overall decreases in agricultural productivity. The fragile nature of these
agroecosystems and limited capacity of resource-poor farmers in the region to
adopt the large-scale use of conventional fertilizer and pest control technolo-
gies suggest the need for agroecological intensification to restore soil function-
ing and ensure long-term sustainability in these systems. This review addresses
soil fertility decline from a management perspective and considers six basic
approaches to enhance nutrient cycling, crop nutrient acquisition, and long-
term productivity. A mass balance approach first defines basic boundaries for
nutrient cycling and suggests that erosion control and identification of alterna-
tive nutrient sources (e.g., peri-urban wastes, rock phosphate) are critical for
reversing negative nutrient budgets. Meanwhile, short-term nutrient dynamics
could benefit greatly from improved management of organic residues in combi-
nation with low-level inorganic fertilizer applications. There is also a need for
greater understanding of soil physiochemical properties throughout much of the
Andes and the impacts of management. Similarly, soil biological functioning is
critical for successful agroecological intensification and there is great potential
for both inoculative and management strategies to promote beneficial soil
communities. Crop breeding for smallholder environments should complement
strategies of agroecological intensification, taking advantage of high regional
agrobiodiversity and experience from stress breeding programs in other regions.
Finally, we suggest several means by which the spatial and temporal organization
of farms may be improved to enhance overall agroecosystem function.

1. Introduction: Agricultural and Soil Fertility

Issues in the High Andes
1.1. Cropping systems of the Andes
Agriculture in the tropical Andean highlands (>2500m) represents a diverse
mix of cultures and cropping systems dating back several millennia (Sandor
and Eash, 1995; Stanish, 2007). Today, these agroecosystems are largely
managed by smallholder farmers in rural communities growing a mix of old
and new world crops across a range of socioeconomic and environmental
settings. Although potato (Solanum spp.) has long dominated food produc-
tion in the region, other important crops include grains (e.g., maize, quinoa,
barley, oats), legumes (e.g., fava beans, peas, Andean lupine—tarwi), other
native tubers (e.g., oca, mashwa, and ullucu), and a wide array of vegetables
(e.g., carrots, onions, cabbage). Livestock remains an integral part of these
agroecosystems and includes native camelids (e.g., llama and alpaca), cattle,
sheep, chickens, and guinea pigs (cuyes). Mixed crop-livestock farming
Andean Soil Fertility 127

generally dominates areas below 3800m; livestock becomes more important

with altitude, as frosts become increasingly limiting to crop production.
Generally speaking, there exists an altitudinal gradient of agricultural inten-
sity, with more intensive management and the production of more diverse
crops possible at lower altitudes and more extensive systems at higher eleva-
tions. Agroecosystems throughout the high Andes employ diverse cropping
strategies and a range of farming intensities, from low-input sectoral fallow
systems, which typically involve several years of cropping followed by 3–15
years of grazed fallow (Orlove and Godoy, 1986; Pestalozzi, 2000), to more
intensively managed permanent structures such as terraces (andenes) and raised
beds (waru-waru or suka) (de la Torres and Burga, 1986). More recently,
industrial agriculture has introduced mechanized tillage and agrochemical
inputs to enable continuous cropping across large expanses, particularly at
lower altitudes. The distribution of these management practices varies con-
siderably depending on cultural, geographic, and demographic factors. For
example, higher population densities and relatively short distance to markets
in peri-urban agricultural areas have led to more intensively managed, high-
input commercial agriculture, while more geographically isolated regions
in the Andes show greater emphasis on low-input, subsistence farming
(Caycho-Ronco et al., 2009). Environmental constraints also play a critical
role, as precipitation, temperature, and soil type are highly variable through-
out the region and largely determine what crops can be grown.

1.2. Biophysical limitations and risks

Despite the diversity of agroecosystems in the tropical Andean highlands,
agriculture is limited by a common set of environmental constraints through-
out the region (Stadel, 1991). Climate plays perhaps the most important role
(Fig. 1A and B). With little access to irrigation, most agriculture in the region
is rainfed, leading to yield depressions when rainfall frequency is low and crop
failure in drought years, with severe implications for overall yield stability.
This threat is perhaps more critical for the relatively arid Altiplano, as well
as inter-Andean valleys of Bolivia and the western slope of Peru, than in
the northern Andes where annual precipitation is generally higher (Fig. 1A;
Bottner et al., 2006; Geerts et al., 2006). Cool temperatures, especially at high
elevations, limit growth and increase the risk of damage incurred by frost and
hail. Meanwhile, the mountainous nature of the region results in many farms
being dominated by sloping, rocky terrain with high inherent spatial hetero-
geneity in soils and microclimate (Buytaert et al., 2007; Zehetner and Miller,
2006). With a few exceptions, soils in the high Andes are generally thin,
fragile, and highly nutrient limited. Although the volcanic soils (Andosols) of
Ecuador and northern Peru (Fig. 2) can be more fertile due to their high soil
organic matter (SOM) content and more favorable soil structure, this depends
greatly upon weathering conditions and the parent materials on which they
128 Steven J. Fonte et al.

A N Annual
B N Annual mean
W E
Precipitation (mm) W E
temperature (°C)
S S
0–261 –6–7
262–610 8–12
611–792 13–18
793–1010 19–22
1011–1357 23
1358–1627 24
1628–1875 25
1876–2276 26
2277–2786 27
2787–6081

0 125 250 500 750 0 125 250 500 750

Kilometers Kilometers

Figure 1 Average annual precipitation (A) and temperature (B) for the Andean focus
region (Ecuador, Peru, and Bolivia). The maps were generated using the WorldClim
database (http://www.worldclim.org) following methods of Hijmans et al. (2005).

formed. For example, allophanic Andosols represent some of the most fertile
soils in the world, while non-allophanic Andosols, which dominate the
Ecuadorean highlands (Poulenard et al., 2001), can have severe problems of
phosphorus deficiency (due to high P-fixation), acidity, and aluminum tox-
icity (Dahlgren et al., 2004). Meanwhile, soils of the central Andes (southern
Peru and Bolivia; Fig. 2) are generally low in SOM and nutrient content,
coarser in texture, highly susceptible to erosion, and on average have a more
neutral pH (Bottner et al., 2006; Cárdenas et al., 2008; Valente and Oliver,
1993). Despite these generalities, it should be noted that targeted manage-
ment with manure and fallows on less vulnerable parts of farmed landscapes
can lead to localized areas of high fertility that support impressive productivity
(Garcı́a, 2011). Also, in contrast to the highly weathered lowland soils
commonly associated with tropical latitudes, many soils in the central Andes
are similar to Ustic soils found in temperate regions (e.g., southern Europe,
parts of North America), thus permitting some degree of knowledge transfer
on management from these better studied regions.
It is generally accepted that crops face the greatest threat from specialist
herbivores, both pests and pathogens, in their centers of origin ( Jennings
and Cock, 1977). The Andean region is no exception. For example,
potatoes in the Andes face the dual threats of potato late blight and several
species of potato tuber moths, weevils, and nematodes, each with the
Andean Soil Fertility 129

Andosols- N. Andes

Hillslope soils, sedimentary

Hillslope soils, mixed

Altiplano soils

Figure 2 Broad characterization of soil environments in the high Andes, divided into
four key regions: Andosols, hillslope soils with sedimentary or mixed parent materials,
and soils associated with the Alitplano. Boundaries are approximate and considerable
intergrading occurs among zones.

potential to cause total crop losses. In recent years, the severity of pest
problems in the Andes has increased significantly, presumably due to the
influences of agricultural intensification and climate change, among other
factors (Parsa, 2010). Despite abundant research, solutions to these problems
are lagging, particularly for neglected crops such as oca (Oxalis tuberosa),
which are facing increasing farmer abandonment due to heavy pest infesta-
tions (Hersh, 2000).

1.3. Socioeconomic and cultural setting of Andean agriculture

In addition to constraints imposed by climate, soil type, and pests, many
farmers in this region have limited access to a number of basic agricultural
inputs such as fertilizer, pesticides, improved crop varieties, mechanized
tillage, and irrigation. Although traditional agriculture may have been largely
sustainable for centuries in the absence of these inputs (Dick et al., 1994),
recent pressures (e.g., population growth, market access, local attitudes,
130 Steven J. Fonte et al.

climate change) have created more demand for their use (Sarmiento et al.,
1993). Locally available inputs of organic matter have also become more
limiting, due to increasing demand for fuel and fodder, as well as lower
biomass production—driven by declining soil fertility and competing land
uses (Orsag, 2009; Swinton and Quiroz, 2003). A long tradition of agriculture
in the region, combined with the preservation of indigenous cultures and
knowledge, has endowed many Andean communities with a sophisticated
level of agroecological management (Sandor and Furbee, 1996; Winklerprins,
1999). However, due to rapidly shifting agricultural, climatic, and socioeco-
nomic contexts, new information and increased knowledge exchange are
needed to help farmers adapt to emerging challenges.
Access to agricultural inputs and external knowledge is especially
limited for rural communities located far from population centers. Due
to their isolation, imported products have a higher cost and technology
transfer (both externally and locally) is often slow, while the distance to
markets is so great that farmers have little incentive to invest in their land
beyond what is needed for home consumption and local trade (Swinton
and Quiroz, 2003). Poverty also represents a significant obstacle, as poor
farmers have neither the monetary resources to invest in new technologies
nor the economic luxury to assume the risk that is associated with these
activities. These issues are particularly relevant for the Altiplano, where
poverty levels greatly exceed the national averages for Bolivia and Peru,
and over 60% of the population lives in rural areas (Quiroz et al., 2003).
Labor shortages can also limit farmer investment in rural areas, as more
farming families seek to diversify their incomes through off-farm employ-
ment (Zimmerer, 1993).
In addition to resource access, other factors influence farm productivity
and farmers’ ability to adopt new technologies. For example, land tenure
can greatly influence farmer decisions, as farmers are generally reluctant to
invest in long-term agriculture improvements (i.e., soil conservation struc-
tures, proper management of SOM) on land that they do not own (Tenge
et al., 2004). Ongoing shifts in land management from communal to private
control represent another unique facet of Andean agriculture that effectively
results in smaller management units and intensification of these fields
(Cárdenas et al., 2008; Mayer, 1979). Despite the potential for more rapid
innovation to occur under individualized management, growing evidence
suggests that a lack of coordinated community decision-making about
agricultural practices can exacerbate problems of soil fertility, livestock
management, and pest regulation (Mayer, 1979; Orsag, 2009; Parsa,
2010). Local and cultural food preferences also affect farmers’ use of agri-
cultural inputs and practices. For example, many farmers choose not to
grow improved potato varieties or use agrochemicals for crops consumed at
home or sold locally, as they feel that traditional practices and varieties yield
superior taste and quality (Caycho-Ronco et al., 2009).
Andean Soil Fertility 131

1.4. Current challenges and emerging threats

Despite a long agricultural history in the region, a number of issues in recent
decades have emerged that threaten the long-term capacity of these agroe-
cosystems to provide food and key ecosystem services. Population growth
and the overall increase in food demand clearly play a role in agricultural
intensification (Orsag, 2009; Winters et al., 1998); however, a number of
technological and societal transformations have also contributed to degra-
dation and the loss of soil fertility in the Andes (Córdoba and Novoa, 1997;
Winters et al., 1998). The introduction of new technologies, such as
mechanized tillage and chemical fertilizers, have aggravated SOM loss and
erosion (Poulenard et al., 2001; Quintero, 2009; Sarmiento et al., 1993),
while other forms of intensification (e.g., reduced time under fallow) can
have similar impacts across a range of farming conditions (Pestalozzi, 2000;
Sarmiento and Bottner, 2002). At the same time, local labor shortages and
high emigration (more typical of rural areas) can lead to de-intensification
and resultant increases in erosion, due to the associated deterioration of
soil conservation structures or continued soil loss from fields with degraded
vegetation and poor soil cover (Harden, 1996; Wiegers et al., 1999;
Zimmerer, 1993). Soil degradation is often a self-perpetuating process,
where declining crop productivity drives farmers to bring new and often
more marginal lands under production (Kessler and Stroosnijder, 2006).
These newly cultivated lands are often at higher elevations, on steeper slopes
and/or on inappropriate soil types, and thus more susceptible to erosion and
soil degradation (Córdoba and Novoa, 1997; Hofstede et al., 2002;
Poulenard et al., 2001), and also possess lower yield potentials. Apart from
issues of soil degradation and sedimentation, the colonization of natural
areas threatens key watershed services (Buytaert et al., 2002, 2006; Hofstede,
1995) and regional biodiversity ( Jaimes and Sarmiento, 2002; Lawler et al.,
2009).
In addition to the direct consequences of human intervention in Andean
landscapes, climate change presents a grave threat to highland farmers in the
region (Perez et al., 2010; Valdivia and Quiroz, 2003; Valdivia et al., 2010).
Although an increase in temperature and elevated CO2 may increase
potential crop growth in some areas (Buytaert et al., 2010a), higher varia-
bility in climate (particularly more erratic precipitation) along with reduced
glacial water supply (Bradley et al., 2006) are causes for great concern.
Additionally, warmer temperatures combined with stronger winds will
likely increase evapotranspiration, thus further exacerbating issues of crop
water stress. While overall changes to total precipitation are unclear, rains
may start later in the year and are expected to become less frequent with
higher intensity (Thibeault et al., 2010; Valdivia et al., 2010). Impacts of
climate change on soils are perhaps less clear. While greater plant produc-
tivity in some areas could contribute to SOM, this may also lead to increased
132 Steven J. Fonte et al.

soil nutrient depletion (van Groenigen et al., 2006). Decreased plant cover
and increased variability in precipitation in some areas will likely result in
greater erosion and subsequent nutrient loss. At the same time, higher
temperatures will likely accelerate losses of soil C due to decomposition of
SOM (Buytaert et al., 2010a; Davidson and Janssens, 2006) and can nega-
tively impact soil decomposer communities (Briones et al., 2009).
High microclimatic heterogeneity, complex topography, and low reso-
lution of global climate models generate substantial uncertainty regarding
future climate predictions in the Andes and limit the ability of farmers and
regional decision makers to anticipate and adapt to these changes (Buytaert
et al., 2010b; Valdivia et al., 2010). Alterations to climate along with
agricultural intensification may lead to increased pest pressure, as warmer
weather can favor pest reproduction and weaken plant resistance due to
increased water stress (Bale et al., 2002; Garret et al., 2006). Farmer surveys
and preliminary evidence suggest that climate change is already yielding
negative impacts on Andean croplands via enhanced water deficits, erosion,
and SOM loss (Aguilera, 2010) and emphasizes the need to develop new
agricultural strategies that improve ecological resilience and flexibility of
agroecosystems.

1.5. Ecologically based intensification in the Andean context

The great challenges that lie ahead for the Andean region indicate the need
for drastic intervention to counteract widespread degradation and help farm-
ers adapt to rapidly changing conditions. Agricultural intensification based
solely upon ‘green revolution’ technologies may not be adequate or desirable,
given the complex socioeconomic setting and relative environmental fragility
of the region. Great heterogeneity in climate, topography, soil type, and
culture further hinders the implementation of broadly prescribed solutions
(e.g., improved cultivars) and emphasizes the need for locally adapted tech-
nologies that address the specific constraints of each region. The unique
cultural needs and ecological complexities of the Andes underscore the
need for ecologically based approaches that seek to maximize long-term
stability rather than short-term economic return. In the context of soil
fertility, such agroecological intensification needs to focus on utilizing and
augmenting the supply of locally produced and renewable soil fertility
resources (e.g., manure, cover crops, compost) and optimizing nutrient use
efficiency throughout the farm. Soil fertility and general management prac-
tices should aim to promote beneficial soil organisms, biological transforma-
tions that increase nutrient availability, and a greater coupling of nutrient
cycles (especially N and P) with C turnover (Drinkwater and Snapp, 2007).
The Andean region offers a valuable test case for such agroecological
intensification due to a number of key attributes: (1) a high diversity of
crops, genotypes, management strategies, and growing environments
Andean Soil Fertility 133

(e.g., NRC, 1989); (2) a burgeoning environmental consciousness and

societal support for indigenous crop varieties and practices (Bebbington
et al., 1993; Vargas, 2009); (3) a valuable history of local knowledge and
agricultural extension; and (4) a relatively low population density that allows
for additional flexibility in exploring alternative technology options. Taking
into consideration the above ecological principles and unique assets of the
region, this chapter seeks to evaluate past research on soil fertility in the
tropical Andean highlands and explore the most promising options and
potential synergies for agroecological intensification in the region. We
consider promising soil fertility technologies and research from other parts
of the world, with key economic and/or environmental similarities to the
Andes, as potential sources of innovation. The review was conducted to
generate recommendations for strategic on-farm interventions as well as
promising areas of future research in soil fertility to advance sustainable
agricultural development in the region.

2. Examining Soil Fertility and Management

Strategies in Smallholder Systems
2.1. General concept of soil fertility
In this review, we employ an integrative definition of soil fertility that
considers nutrient availability as well as the suite of physical, chemical,
and biological aspects that characterize the soil environments and the
ecosystem services they provide. Additionally, we consider co-limiting
factors to crop growth as they have important implications for nutrient
uptake and crop productivity. This section attempts to address soil fertility
from a management perspective, considering six basic approaches for
improving overall nutrient cycling, crop nutrient acquisition, and long-
term productivity. These are: (1) a mass balance consideration of agroeco-
system nutrient flows; (2) short-term nutrient dynamics; (3) physiochemical
environment of soils; (4) the biological functioning of soils; (5) plant
breeding for agroecological intensification; and (6) the spatial and temporal
organization of farms. Alternatively, these approaches may be viewed as
responses to common soil fertility problems in smallholder farms.

2.2. Approaches to examining soil fertility

2.2.1. Mass balance
One of the simplest, yet most valuable approaches for evaluating long-term
soil fertility dynamics of agroecosystems is to consider an overall budget of
nutrient input versus output (Cobo et al., 2010; Smaling and Fresco, 1993).
This approach considers both intentionally managed imports (e.g., fertilizer,
134 Steven J. Fonte et al.

manure) and exports (e.g., nutrients in crops or livestock produced for

market), as well as unintended nutrient transfers (e.g., such as nutrient inputs
in irrigation water and atmospheric deposition; and export due to erosion,
burning, and nutrient loss via leaching or gaseous emissions). According to
this perspective of fertility management, the nutrients entering a system
must match or exceed nutrient outflow in order to maintain long-term yields
(Fig. 3). Although N and P are the most limiting nutrients to agriculture in the
Andes (Bossio and Cassman, 1991; Devaux et al., 1997), other nutrients
such as K, S, and Ca may become limiting as well under a variety circum-
stances (de Koning et al., 1997; van de Kop, 1996; Vanek, 2010). Improved

Community/regional/national scales
Market
Inorganic
export
fertilizers

Organic Farm/household scale

Human + Fertilizers
Biological Inputs
animals
N Fixation

Agricultural field

Organic N Internal
(labile and Inorganic N cycling
recalcitrant pools)

Crop N
removal
Losses

Gaseous Leaching
Erosion
N losses

Figure 3 Generalized nitrogen flow diagram for a typical smallholder field, incorpor-
ating spatial scales of nutrient sources. Rectangles represent the boundaries of the field
(inner solid rectangle), household or farm management unit (outer rectangle), and
boundary between the community scale and urban markets that supply inorganic
fertilizer and other external fertility materials and also are a destination of marketed
agricultural products and N export. Drawing shows that some N is cycled internal to
the field and most crops and residue nutrients are cycled internally within farms or lost
to erosion. Rangeland can represent an important fertility resource accessed via live-
stock grazing in extensive systems in the central Andes. Arrows represent approximate
size of loss pathways.
Andean Soil Fertility 135

management of soil fertility at this level needs to focus on improved nutrient

recycling within agroecosystems and requires careful understanding of system
inputs and losses. We recognize the mass balance approach as useful starting
point that forces a consideration of system boundaries and long-term sustain-
ability. However, we emphasize that whole system nutrient budgets often
embody large uncertainties and neglect critical details of internal cycling and
nutrient inputs via weathering.
Nutrient removal from agroecosystems via crop (or livestock) export
often represents the most important mode of nutrient loss in cropping
systems (Smaling et al., 1993; Vitousek et al., 2009) and understanding
these losses is fundamental to ensuring long-term productivity. For exam-
ple, crop export of K has been estimated to exceed other mechanisms of K
removal (i.e., erosion, leaching) by an order of magnitude in some Andean
cropping systems (de Koning et al., 1997; Vanek, 2010). Removal of N and
P in crops can also be quite large, but not always the dominant mechanism
of loss in the Andes. Given that the export of nutrients via harvest is
ultimately desirable, as it relates to yield and the nutritional value of a
crop, efforts to balance nutrient budgets are best focused on a variety of
other, unintended mechanisms of loss.
Due to the steep terrain and general vulnerability of soils in the Andes,
rates of erosion can be high (Alegre et al., 1990; Harden, 1988, 1993) and
have garnered considerable attention in past years. Erosion represents a
critical mechanism of nutrient loss for agroecosystems in the region. For
example, rates of soil loss for agricultural fields in the Peruvian Andes have
been estimated on the order of 10–100Mgha
1 yr
1 (Felipe-Morales, 2002;
Romero-León, 2005). Applying a conservative estimate for soil N content
of 1gNkg
1 soil in the surface layer (Sandor and Eash, 1995) and a loss rate
of 50Mgha
1 yr
1, this translates into an annual loss of roughly 50kgN
ha
1. This value is in accordance with losses reported elsewhere in the
Andes and generally exceeds N losses associated with harvest of 10–30kgN
ha
1 yr
1 (de Koning et al., 1997). Nutrient losses due to erosion, however,
may be considerably higher under some circumstances. In the Ecuadorian
Andes, which tend to have higher SOM and nutrient content (Buytaert
et al., 2007; Tonneijck et al., 2010), Harden (1988) estimated erosion rates
as high as 800Mgha
1 yr
1 on steep, highly susceptible agricultural land.
Although this example represents an extreme case, it highlights the potential
for erosion to rapidly offset nutrient balances in Andean agroecosystems.
We also note that erosion losses for P may be even more important than for
N, as they often exceed the limited ability of farmers to replenish P with
manure or fertilizers. Perhaps a more critical consequence of high erosion
rates is the loss of suitable substrate for crop growth. While nutrient stocks
can be restored by fertilization, soil formation occurs slowly, with the global
average estimated to be as low as 1Mg soil formed ha
1 yr
1 (Pimentel,
2006). Thus, the implementation of viable soil conservation strategies is
136 Steven J. Fonte et al.

critical throughout the Andes from both a nutrient balance perspective and
more general goals of agroecosystem sustainability.
The susceptibility of soils to erosion depends greatly not only on topog-
raphy, slope, climate, and soil type, but also to a large extent on management
factors such as vegetation cover, cropping system, tillage, and livestock
intensity (Coppus et al., 2003; Inbar and Llerena, 2000). Given the widely
acknowledged role of agricultural disturbance in exacerbating erosion
(Montgomery, 2007), a number of technologies have been put forth to
conserve soils in the Andes and elsewhere. For steeper hillside farms, stone
and earthen terraces represent one of the oldest approaches to soil conser-
vation in the region and have proven largely successful in combating erosion
(Goodman-Elgar, 2008; Sandor and Eash, 1995), yet these structures may
no longer represent a viable option due to high labor requirements and
altered political and socioeconomic conditions (Dehn, 1995; Posthumus
and De Graaff, 2005). Terraces formed by live barriers are attractive as they
require considerably lower initial investments, offer key byproducts (e.g.,
fodder, fuel, organic matter inputs), and can effectively control erosion
(Craswell et al., 1998; Sims et al., 1999). Sims et al. (1999) working in
Bolivia suggested that grasses (Phalaris sp.) offer the best erosion control in
such systems, as they grow rapidly and provide fodder for livestock. How-
ever, competition for nutrients and water, along with high spatial variability
within live barrier terraces, indicates that such technologies require addi-
tional study and/or modification (Dercon et al., 2006). Cover cropping
offers another viable means to control erosion on hillside soils, by protecting
the soil when and where crops are absent (Bunch, 2004; Sims et al., 1999).
The development of adapted cover crop technologies for erosion control in
the Andes is highly desirable, as cover crops can also contribute significantly
to soil nutrient and organic matter stores (Snapp and Silim, 2002; Wheeler
et al., 1999). Despite the great potential for plant-based technologies to
control erosion, maintenance costs (e.g., labor, seed) and knowledge
requirements can limit their adoption by small farmers (Bunch, 2004;
Posthumus et al., 2010; Snapp et al., 1998). Other erosion control strategies
focus on reduced soil disturbance (tillage) in order to decrease labor require-
ments and conserve soils. For example, the Wacho rozado system from
Ecuador (Sherwood et al., 1999) as well as the chiwa and chacmeo systems
from the Central Andes (Oswald et al., 2009) all represent unique indige-
nous potatoes planting systems that seek to minimize soil movement on
smallholder farms. Reduced tillage options have also been developed for
larger farms that employ animal traction or mechanized tillage (Mamani
et al., 2001; Quintero, 2009; Wall, 1999) and may be especially relevant for
hillside farms in the northern Andes where mechanical tillage is more
common and poses a greater threat. However, significant investment is
still needed to work with farmers in developing locally adapted technologies
and implements that minimize soil disturbance and provide viable
Andean Soil Fertility 137

alternatives for small- and medium-sized farms in the region. At a watershed

scale, erosion has been mitigated in some valley areas by recapturing eroded
soil (and nutrients) as siltation of fields on river margins. This was a key
feature of ancient spate irrigation practices in the Andes and elsewhere
(Mehari et al., 2011; Zimmerer, 2011), and the same benefit is sought
from gabion walls for soil capture along rivers promoted by some NGOs
in the region (Pacheco et al., 1992). Despite advances in soil conservation
technologies, the potential of reduced tillage and other strategies in the
Andes needs to be further evaluated.
Other forms of nutrient loss can occur via leaching and gaseous emis-
sions, particularly in the case of N. These losses not only threaten long-term
yields but also can have highly deleterious impacts on regional water quality,
downstream aquatic ecosystems, and greenhouse gas emissions (Matson
et al., 1998). Such forms of nutrient export may be considerable in some
agroecosystems are often the least understood by farmers and can be difficult
to quantify. Management strategies to control them can involve reducing
the availability of nutrients during times of low plant demand (see Fig. 4 and
Section 2.2.2) as well as increasing SOM and the nutrient storage capacity of
soils (Craswell and Lefroy, 2001).
Given that most small farmers are highly resource limited, improved
nutrient recycling within farms offers an attractive means to close the gap, as
many inputs (e.g., crop residues, manure, compost) are free and only require
a limited amount of additional labor. For example, Osman (1999) suggested
improved management of livestock manure and urine (e.g., composting, rapid
incorporation to soil) as a means to reduce gaseous N losses and significantly
improve the farm N budgets for small farmers in Peru. Additionally, for a

IF
Nutrient availability/uptake

Combined OR +IF

OR
Plant demand

Time (growing season)

Figure 4 Theoretical curves for plant nutrient uptake an availability under different
fertility strategies (IF, Inorganic fertilizer; OR, Organic resources). Curves for plant
demand and organic matter decay represent what might be seen during a typical
growing season for a crop such as maize or potato. Twin peaks for IF represent two
fertilizations (at planting and at peak crop demand).
138 Steven J. Fonte et al.

number of reasons (convenience in threshing of grain, use of crop residues for

forage and fuel) residues of quinoa, maize, lupine, and other grains are often
removed off-field at harvest. On-field crop residue retention (or return) could
contribute substantially to restoring nutrient balances and combating SOM
depletion (Fuentes et al., 2009). Despite the promise of such simple interven-
tions, many loss mechanisms are more difficult to control—particularly those
associated with human urine and fecal waste. The safe recovery and applica-
tion of human waste, which can comprise a significant portion of small
farm nutrient budgets (Kanmegne et al., 2006), often faces technological and
cultural obstacles (Cofie et al., 2005; Karak and Bhattacharyya, 2011), resulting
in persistent losses of key elements. Even under ideal management practices,
losses are inevitable and alternative nutrient input sources are ultimately
necessary to restore nutrient equilibrium on smallholder farms.
While organic inputs (i.e., manure, plant residues) and long fallows were
traditionally used to restore soil fertility (Hervé, 1994; Sarmiento and
Bottner, 2002), intensification of cropping systems (i.e., reduced fallow
periods) has led to a growing reliance on synthetic fertilizers (when available
and affordable) to make up the difference (Claverı́as, 1994; Wiegers et al.,
1999). However, negative nutrient balances, reported for various highland
Andean agroecosystems and driven strongly by erosion (de Koning et al.,
1997; Osman, 1999; Vanek, 2010), suggest that the levels of fertilizer
applied by farmers are often insufficient, likely due to inaccessibility (i.e.,
cost, availability) of these inputs to small farmers. Increased reliance on
legumes and biological N-fixation offers perhaps the most obvious means
to address long-term N deficits, as legumes already comprise a fundamental
component of agricultural systems worldwide and their benefits are well
understood by farmers. However, current N inputs from legumes are
generally insufficient to meet crop demand and replenish soil N stores,
and further integration of legumes into cropping systems (on- and off-
field) is necessary to better meet typical farm N requirements (Snapp et al.,
2005). Limitation of other nutrients, particularly P, is more complicated and
may also have consequences for N-fixation (Chalk, 2000; Reed et al., 2007).
Addition of P, K, Ca, and other potentially limiting elements to crop
production generally depends upon the mining, refinement, and import
of these nutrients from great distances, and thus costs are high. Improve-
ment of rural markets and distribution networks for these fertilizers could
thus help alleviate some issues of nutrient depletion, but the dependency of
these inputs on oil prices and market fluctuations does not ensure that they
are sustainable in the long-term. Rock phosphate represents a low quality,
unrefined source of P that is relatively available in various parts of the Andes
and offers promise for restoring P budgets in the region (Lorion, 2004).
Other potential sources of nutrient inputs include peri-urban wastes from
residential and commercial sources, but this largely depends on the proxim-
ity of farm to population centers and requires necessary infrastructure for
Andean Soil Fertility 139

collection and transport of these materials (Harris et al., 2001). Longer-range

transport of these resources may also be economically viable and highly
advantageous under some circumstances for correcting nutrient imbalances
and thus merits additional consideration.

2.2.2. Short-term nutrient dynamics and synchronization

The nutrient balance approach discussed above is a critical starting point for
agroecological intensification and the long-term management of soil fertility.
However, in addition to maintaining sufficient quantities, nutrients must also
be present in plant available forms and at a time when plants need them. For
example, in the case of Andosols in the northern Andes and other P-fixing
soils, sorption of P limits its availability even when the P balances in a given
year are highly positive (Dahlgren et al., 2004; Espinosa, 1991). Sorption may
be irreversible unless management (organic residues, root/phosphate prox-
imity, and timing of P application) addresses the low availability of P
(Nziguheba et al., 1998). Synchronization of nutrient availability with plant
demand is thus critical for maximizing crop productivity as well as for
minimizing nutrient losses and associated environmental consequences
(Cassman et al., 2002; Woomer and Swift, 1994). Inorganic fertilizers offer
the distinct advantage of providing nutrients that are immediately available to
plants and, in theory, allow for relatively simple regulation of soil nutrient
availability to support crop growth. However, various economic and man-
agement constraints (i.e., labor, fluctuating costs, risk of crop damage asso-
ciated with multiple field entries) typically result in suboptimal application of
fertilizers from a soil fertility standpoint. Throughout the Andes, considerable
research has examined the potential of synthetic fertilizers to increase pro-
ductivity in various cropping systems. While high synthetic fertilizer applica-
tions may be representative of commercial fields in the region, it is much less
common in smallholder agriculture that dominates the Andean highlands
(Caycho-Ronco et al., 2009; Terrazas et al., 1998). Despite considerable
emphasis on inorganic fertilizer research at regional universities, few studies
have examined the issues of nutrient leaching and gaseous losses associated
with inefficient fertilizer application (e.g., Machado et al., 2010). Evidence
suggests that fertilizer application may greatly exceed crop demand in some
circumstances, for example, at higher altitudes where potential yields are low
and synthetic fertilizers represent a more recent introduction (NRC, 1989;
Oswald, 2010). Thus, environmental and economic costs of fertilizer appli-
cation may be substantial, but remain largely unknown for the high Andes.
Organic nutrient sources in the Andean highlands encompass both
traditional manure and plant residue inputs and a number of innovations
such as off-farm residues and manures, compost, and bioles (fermented
liquid nutrient amendments). Considerable research has been devoted to
these organic fertility inputs, and simple yield comparisons to synthetic
fertilizer have been a common theme (e.g., Aguilera, 2010; Garcı́a, 2011).
140 Steven J. Fonte et al.

Organic inputs have some challenging aspects for soil fertility management.
For example, compared to inorganic fertilizers, more must be transported
and applied due to their lower nutrient contents. Additionally, the mineral-
ization of plant nutrients from organic matter for plant uptake is often
harder to predict, as amount and timing of nutrient release depend on a
number of factors including organic matter quality, moisture, temperature,
and the soil decomposer community receiving organic additions (Lavelle
et al., 1993; Palm et al., 2001). The quality of organic resources (nutrient
content and ease of decomposition) also tends to be more heterogeneous
and can depend greatly on the source of material and timing of collection
or application. Thus, additional labor and knowledge may be required
to effectively manage organic resources, such that nutrient release is syn-
chronized with crop growth. Despite these potential drawbacks, organic
resources are the most important form of nutrient inputs for crop produc-
tion in the Andean highlands (Caycho-Ronco et al., 2009; Terrazas et al.,
1998) and offer several key benefits to farmers. First, organic resources are
typically less expensive and more readily available to farmers than inorganic
nutrient sources, especially in rural areas. Additionally, Andean smallholders
have developed complex local knowledge that is well adapted to the use of
manures and fallow residues and that may be readily transferable to innova-
tions in organic inputs. Organic resource application also contributes to
the maintenance of SOM in cropping systems (Fernandes et al., 1997), with
vital implications for soil structure, water storage and movement, nutrient
supply and retention, and the promotion of healthy soil biological commu-
nities (Craswell and Lefroy, 2001). Finally, they generally release nutrients
more slowly and over time, potentially decreasing the susceptibility of
nutrients to loss, as compared to synthetic fertilizers (Kramer et al., 2006).
Recognizing the inherent complexities of organic nutrient sources,
researchers have invested considerable effort toward classifying and processing
these materials in order to simplify management, provide a more predictable
release of nutrients, and to enrich materials for reducing bulk and accelerating
nutrient mineralization. For example, Palm et al. (2001) proposed a classifica-
tion system for plant residues and organic materials commonly found in
the humid tropics that is based on residue N content and the presence of
recalcitrant compounds (polyphenols and lignin). This system identifies four
organic matter quality classes with distinct recommendations of how best to
apply the residues for optimal nutrient release. Although a few studies have
examined the quality of organic resources available in the Andes (Coûteaux
et al., 2008; Machado et al., 2010; Mahboubi et al., 1997), this information is
limited and has not been adequately translated into targeted management
recommendations for the Andean agroecosystems.
Efforts for processing organic materials prior to application have become
commonplace throughout the world, as well as in the Andean highlands
(Felipe-Morales, 2002; Herbas, 2000). Composting encompasses a broad
Andean Soil Fertility 141

array of practices that involve mixing organic materials of different qualities

(e.g., manure, plant residues, kitchen wastes, ash, sawdust) and a managed
decomposition of residues to concentrate nutrients, remove pests and patho-
gens, homogenize materials, and form a more stable, yet nutrient rich sub-
strate to apply to soil and stimulate biological activity (Litterick et al., 2004;
Misra et al., 2003). Proper composting can also help prevent nutrient losses
from high quality organic resources (e.g., urine and fresh manure), especially
when materials are produced at a time that is inappropriate for field applica-
tion. Other forms of composting, such as vermicomposting, rely on epigeic
earthworms to breakdown organic materials, while concentrating and stabi-
lizing nutrients in their casts. Although potentially more management inten-
sive, this process yields a high quality soil amendment in a relatively short time
and has received increasing attention in recent years, in part due to the
beneficial byproducts associated with earthworm activity (Blouin et al.,
2005; Cardoza, 2011). Anerobic methods of composting also exist in the
Andes for the production of bioles, plant- or manure-based liquid fertilizers
(Caycho-Ronco et al., 2009; Felipe-Morales, 2002). These contain nutrients
in relatively available forms and are often applied to foliage or soil, as a means
of rapidly correcting minor crop nutrient deficiencies. Organic materials can
also be processed through non-biological processes. For example, burning of
residues is common and provides the advantage of reducing biomass, suppres-
sing pests, and increasing the availability of some key elements (e.g., K), but
can also lead to high losses of others (e.g., N) and is generally ill-advised. A
promising alternative for low quality, lignocellulosic residues (e.g., wood,
wheat straw) is to pyrolyze or heat materials in a low-oxygen environment
to produce biochar. Production and application of biochar supply some
nutrients directly (Novak et al., 2009), but is perhaps more relevant for the
indirect benefits to soil fertility, C storage, and biological functioning of soils
(Chan et al., 2007; Noguera et al., 2010).
Despite a considerable research on the efficacy of synthetic and organic
fertilizers in Andean crop production, relatively few studies have considered
the combined application of these resources or integrated soil fertility
management (ISFM; Vanlauwe et al., 2001). This approach recognizes the
importance of organic matter inputs for supplying nutrients, maintaining
SOM, and promoting healthy soil food webs (Craswell and Lefroy, 2001;
Moore et al., 2004), but also allows for the strategic application of inorganic
nutrients to meet crop nutrient demand at critical crop growth stages
(Fig. 4). This strategy ultimately allows farmers to optimize the use of locally
available organic resources with relatively small quantities of imported farm
nutrients. Although this approach has been advocated by several researchers
in the Andean region (Sarmiento et al., 2001; Valente and Oliver, 1993),
few field trials have adequately addressed this approach (e.g., use of appro-
priate controls, equivalent nutrient additions) in the Andes and the potential
is largely unknown.
142 Steven J. Fonte et al.

In addition to considerations regarding the quality of nutrient resources

and the dynamics of nutrient release, other factors such as climate, soil type,
and management play a key role in determining optimal nutrient manage-
ment strategies for a particular agroecosystem. For example, Bottner et al.
(2006) compared the decay of buried crop residues between contrasting
Andean environments ( páramo soils of Venezuela vs. dry puna soils of
Bolivia) and found that, despite having warmer and moister conditions,
the finer texture and more acid soils of the páramo resulted in slower decay
and a greater contribution of residues to SOM (Pansu et al., 2007). This
suggests that the potential role of organic residues in meeting crop nutrient
demand is distinct for these two systems and may be more important for
crop productivity in the sandier, drier, SOM depleted soils of the Altiplano.
Corroborating this idea, Chivenge et al. (2010), who reviewed ISFM trials
across sub-Saharan Africa, found residue additions to contribute substan-
tially more to crop yield in coarse (as opposed to fine) textured soils and
suggested that organic matter additions may temporarily contribute to key
SOM functions (i.e., water- and nutrient-holding capacity) in sandy soils
with low SOM content. Another example comes from the Wachu rozado
system from the northern Andes, where soil is inverted and potatoes are
planted between two grass layers placed on top of one another (Sherwood
et al., 1999). The organic matter in the grass layers begins to decompose
and isolates the growing crop from the acidic, P-fixing Andosols, while
providing a source of nutrients and an ideal growing environment for the
potato. This system is well suited for the Ecuadorian páramo soils as means
to improve nutrient availability, but may not be appropriate for other areas
or soil types in the Andes. The great number of possibilities for such
environment–management interactions emphasizes the need for developing
locally adapted strategies that best utilize available nutrient resources and
strategically manipulate the soil environment.

2.2.3. Physiochemical environment of soils

Although the concept of soil fertility typically emphasizes the availability of
nutrients for plant growth, the physical and chemical conditions of the soil
environment play an important role in regulating nutrient processes as well
as the ability of plant roots to explore the soil and assimilate nutrients. The
physical status of soils concerns the structure of soils and how individual soil
particles are arranged and combined to form soil aggregates. Aggregation
in soils can have important impacts on a number of key soil properties. For
example, aggregation directly impacts key soil functions related to soil
porosity, aeration, water storage, and movement (Wu et al., 1990) and
also affects the ability of roots to penetrate and explore the soil volume for
water and nutrient resources. Soil aggregation also plays a fundamental role
in SOM turnover and regulation of soil biotic communities (Brussaard et al.,
2006; Six et al., 2002). In many soils, aggregates are closely associated with
Andean Soil Fertility 143

organic matter, as larger aggregates (>50mm diameter) typically form

around organic residues and/or are held together by microbial mucilages,
fine roots, and fungal hyphae (Tisdall and Oades, 1982). Although SOM is a
critical factor for aggregate formation in most soils, mechanisms of aggrega-
tion can vary depending on soil type (Bronick and Lal, 2005). For example,
carbonates may dominate aggregation processes in Aridisols, while Al/Fe-
humus complexes may be more important in Andosols (Bronick and Lal,
2005; Dahlgren et al., 2004). Soil ecosystem engineers (e.g., earthworms and
ants) are key agents of aggregation in soil with favorable conditions for their
activity (Lavelle et al., 1997). Texture also plays a role in aggregation, with
organic residues relatively more important for aggregation in sandy soils as
oppose to those dominated by clays (Bronick and Lal, 2005). Although
proper management of soil structure is often critical for crop growth, few
studies have looked at the effects of soil preparation and management on soil
structure in the Andes. One exception is the relatively well-documented
influence of management impacts on aggregation and soil degradation in
the hardened ash (Cangahua) soils of Ecuador. This research indicates that
Cangahua soils have highly unstable aggregates, are severely prone to crust-
ing and erosion following cultivation, and should ultimately be left untilled
with permanent vegetative cover (Buytaert et al., 2002; Podwojewski and
Germain, 2005; Poulenard et al., 2001). Although not true for all soils (e.g.,
Cangahua soils), soil structure can generally be improved via reduced soil
disturbance (e.g., tillage) and increased inputs of organic matter. The
paucity of information on soil structure from other parts of the Andes
suggests that further research on soil C and aggregate dynamics is needed
to improve the long-term soil fertility management in the region.
The chemical environment of soils has perhaps received more attention
(with respect to fertility) and refers to the ability of soils to maintain available
nutrients and support basic plant growing processes. There exist a number
of relevant chemical measures that influence soil fertility; however, soil pH
(acidity) is probably the most important, as it affects multiple other chemi-
cal, biological, and physical processes of the soil. For example, acidic soils
(pH<5) can have severe problems with Al toxicity and low nutrient
availability (including P, N, K, and Ca). Along with pH, cation exchange
capacity (CEC) is another important soil factor that relates to the ability of a
soil to retain available plant nutrients, while in drier parts of the Bolivian
Altiplano, soil salinity can also become a serious issue limiting agricultural
productivity (Valente and Oliver, 1993). Although these factors are closely
linked with site characteristics (i.e., soil type, parent material, and climate),
management can greatly exacerbate or ameliorate these soil chemical prop-
erties. For example, tillage and fertilization during the cultivation stage in
a sectoral fallow system of the Venezuelan páramo have been shown to
decrease soil pH in these already acidic soils (Abadı́n et al., 2002; Abreu
et al., 2009), and while pH can increase when soils are returned to fallow, it
144 Steven J. Fonte et al.

may not fully recover to the precultivation state for many years. Liming
(e.g., application of calcium carbonate) is perhaps the most commonly
prescribed means to increase the pH in acid soils, but required inputs can
be substantial and the costs prohibitively high for most smallholder farms.
Additions of ash and organic matter have also been shown to reduce acidity
(Aguilera, 2010; Garcı́a et al., 2010) and may represent a more affordable
alternative for Andean farmers in some circumstances. Organic matter
inputs have also been shown to increase CEC (Garcı́a et al., 2010; Oorts
et al., 2003) and can significantly improve other physical and chemical soil
properties such as aggregation, water-holding capacity, and soil buffering
capacity (Craswell and Lefroy, 2001; Kong et al., 2005). Biochar offers
many of the same benefits of organic matter, but with potentially more
immediate and longer lasting impacts. For example, biochar has been shown
to increase soil pH, CEC, water-holding capacity (Chan et al., 2007; Jha
et al., 2010; Karhu et al., 2011) and alter pore size distribution with multiple
impacts for plant growth and soil biological communities (see section 2.2.4).
As for soil structure, research concerning the characterization and manage-
ment of soil chemical environments is notably lacking across much of the
Andes and merits further emphasis in future research endeavors.

2.2.4. Biological functioning of soils

Understanding and managing biologically mediated nutrient cycling in soils
are critical for sustaining the productivity of smallholder agricultural systems.
Soil microbial and faunal communities ultimately depend upon carbon
derived from plant residues, root exudates, and other processed, more stable
forms of SOM as an energy source and habitat (Moore et al., 2004). In using
these resources, symbiotic and associative rhizosphere organisms can: (1) pro-
vide plants access to nutrient pools that are typically unavailable to them (e.g.,
atmospheric N2, recalcitrant P pools); (2) defend plants against pests and
disease; and (3) enhance plant growth via hormonal or other physiological
mechanisms. Here, we address two well-studied themes that link soil biology
to agroecological intensification in the region. First, proper organic matter
management by smallholder farmers is fundamental to soil biological com-
munities, as SOM is a key substrate for promoting soil food webs and
associated ecosystem services (Powlson et al., 2011). Second, inoculation of
soil with carefully selected microbes from key functional groups (that can
leverage substantial improvements in nutrient cycling or plant health) offers a
promising and low cost means of improving soil functioning and system
productivity. The hope (and hypothesis) is that improved management of
SOM, along with the targeted introduction and/or promotion of beneficial
organisms, can sustainably increase the productivity of crop rotations with
relatively low levels of farmer investment.
In farmed soils, microbial and faunal communities reflect the quantity,
quality, and diversity of plant C inputs, as well as the intensity and frequency
Andean Soil Fertility 145

of soil disturbance (Carney and Matson, 2005; Thies and Grossman, 2006).
Intensification on smallholder farms often leads to a decrease in the diversity
of residue inputs and an increase in the tillage intensity, thus reducing the
abundance and diversity of soil communities and the SOM which they
depend on (Postma-Blaauw et al., 2010). For example, in maize and wheat
cropping systems of the Mexican highlands, residue retention and reduced
tillage have been shown to foster greater overall bacterial populations,
including beneficial actinomycetes and fluorescent pseudomonad bacteria
(Govaerts et al., 2008). Improved residue management (greater quantity and
diversity of inputs) has also been shown to increase microbial biomass and
diversity when compared to a conventionally managed agriculture (with
low residue inputs) in California (Briar et al., 2011). Root derived C can also
contribute to SOM and soil aggregation and is a key resource for soil
microbes, in addition to the direct role of soluble root exudates from living
plants (Kong and Six, 2010; Puget and Drinkwater, 2001). A closer exami-
nation of different residue management strategies is particularly important
for the Andean region where residues are predominantly removed for
livestock, yet little is known about the consequences of this practice for
soil biological functioning. In addition to management, temporal changes
over the fallow cycle in the Andean region also drive microbial community
dynamics and can provide benchmarks for the functional microbial and
faunal diversity that managed fallows need to provide. For example, in the
Bolivian Altiplano, Sivila and Herve (1999) found changes in the microbial
communities along a chronosequence of fallowed fields, such that mycor-
rhizal spore counts increased with fallow age and SOM levels, but decreased
following rotations of non-mycorrhizal quinoa and with the removal of
Baccharis shrub vegetation.
The remarkable diversity and range of functions attributed to microbes
and fauna in the rhizosphere of crops are now widely acknowledged, growing
out of early knowledge of important plant symbionts such as rhizobia
and mycorrhizae, to learning that microbes mediate many of the processes
in soils that control plant growth (Murphy et al., 2007; Osorio-Vega, 2007;
Rodrı́guez et al., 2006). Symbioses between legumes and rhizobia/bradyrhi-
zobia, and associations between Frankia actinomycetes and trees such as the
agroforestry species Alnus spp., currently account for most of the biologically
fixed N on Earth (Freiberg et al., 1997). Legume species in Andean cropping
systems include endemic Andean lupine (Lupinus mutabilis) infected by Bra-
dyrhizobium lupini; vetch, pea, and fava bean infected by Rhizobium legumino-
sarum bv. vicieae; and alfalfa and various endemic and introduced clover species
(Trifolium spp.). Arbuscular mycorrhizae (AM) are another well-studied sym-
biont that infects the roots of most crop species. They expand the volume
of soil explored by the plant and increase access to P and micronutrients,
especially under dry soil conditions (Aroca and Ruiz-Lozano, 2009). In the
Andes, potatoes, maize, cereals, and legumes like Vicia spp. are hosts for AM.
146 Steven J. Fonte et al.

Interestingly, L. mutabilis and the Andean Chenopodium crops (Quinoa,

Kañawa, etc.) do not host AM. Another potentially important symbiont,
Trichoderma, is a fungal genus known to inhibit root pathogens (Lorito
et al., 2010) via antifungal toxins and enzymes (see Verma et al., 2007).
Other modes of action may also contribute, as some Trichoderma species inhabit
the interface between the root cortex and soil as plant growth-promoting
symbionts, inducing resistance to disease in the host plant with chemical
signaling and occupying a niche at the root surface which crowds out patho-
gens (Vinale et al., 2008). Trichoderma may also improve nutrient uptake and
induce positive responses to drought and pest stresses (Lorito et al., 2010).
Rhizobia, actinomycetes, AM, and Trichoderma are increasingly consid-
ered as part of a class of plant-growth-promoting microbes (PGPM) that act
either in the rhizosphere or endophytically to improve plant growth via
mobilization of scarce nutrients, hormonal action (e.g., auxin synthesis),
and/or plant protection from disease and pest attack (Martinez-Viveros
et al., 2010). Hormonal modes of action are likely important for PGPM
impacts on crop growth in the Andes. For example, the production of
indole acetic acid (IAA) promotes plant growth via effects on cell elongation
and greater branching of root systems. Oswald et al. (2010) found that over
half of the Bacillus and Actinomyces species collected in one Peruvian location
produced IAA in vitro and may have improved growth of potato in a pot
study. Other rhizosphere microbes degrade ethylene, a plant hormone that
deters plant growth at high or persistent levels (Hayat et al., 2010; Loon,
2007). Microbes that improve access to P inhabit many soils and include
phytate-mineralizers, which act on organic P substrates that form a substantial
P reservoir in many soils (Oberson et al., 2006). Meanwhile, P-solubilizers
can dissolve inorganic calcium phosphates and apatite minerals such as rock
phosphate using organic acids and other mechanisms (Rodrı́guez et al., 2006;
Sharma, 2003; Takeda and Knight, 2006). In addition to the root cortex (e.g.,
Rhizobia and AM), endophytic microbes colonize other plant parts such as
xylem vessels where they may favor plant growth under stress conditions,
including greater tolerance of drought and salt stress (Compant et al., 2010;
Hahn et al., 2008; Kane, 2011). It should also be noted that there is frequent
overlap of these functions within a single PGPM species (e.g., microbes that
simultaneously synthesize auxin hormones and solubilize P; Bashan and
Bashan, 2010). Despite advances in understanding of microbe–root interac-
tions, most microbes inhabiting the rhizosphere and their functional genes
may still be unknown. Leveau (2007) thus suggested that soil metagenomics,
the attempt to directly sequence and derive functional information from
DNA of the entire microbial community of a particular rhizosphere or soil,
can help to probe novel and yet unknown modes of action for PGPM.
Metagenomics could form part of a basic research agenda that may yield
benefits for smallholders over the long-term, in contrast to direct trials of
microbes as inoculants that we address next.
Andean Soil Fertility 147

Interest in microbial inoculants grew naturally from knowledge of the

beneficial effects of soil microbes for crops presented above. Nevertheless, it
may be harder to improve beneficial soil communities via inoculation of
seeds or soil than is suggested by experiments that test the effect of one
microbial species on a crop of interest (Benizri et al., 2001; Oswald et al.,
2010). Successful inoculants must exceed the positive impact of similar
microbes already in the rhizosphere and also compete well enough to persist
there. In vitro screening results are thus often not well correlated to field trials
(Martinez-Viveros et al., 2010; Oswald et al., 2010). In spite of these
challenges, microbial inoculants have shown promise in the Andes and in
other smallholder systems. In order to understand where inoculants might
contribute positively to agroecological intensification, research efforts
should employ rigorous and standardized screening and on-farm trials and
develop hypotheses about the soil and management conditions (e.g.,
degraded ones) where inoculants have greatest impact and those where
their effects are redundant or inconsistent (Oswald et al., 2010).
Research to date on microbial inoculation in the Andes confirms the need
to select for both effective and competitive inoculant microbes. Differences
in the effectiveness of Rhizobia strains for Andean legumes have been docu-
mented in several trials in the region (Barba et al., 2000; Conde et al., 2000;
Lagacherie et al., 1983). Although high levels of endemic Rhizobia are
generally thought to reduce the benefit of inoculation (Evans et al., 1996),
Mnasri et al. (2007) found that a competitive rhizobium strain on beans could
displace native strains even when native strains were abundant. Findings of
Vanek (2010) in a Bolivian smallholder system suggest that N-fixation rates
may be more limited by available P than by a lack of effective rhizobial
symbionts. For a number of common legumes in the Andes, it appears that
symbiont populations are maintained as long as legumes remain in the
rotation (Meneses et al., 2000) and that continued inoculation may not be
so important. Research on the potential for AM inoculants better illustrates
the challenge of improving on endemic microbes and also supports the idea
that gradients in soil fertility determine the impact of AM on crop nutrition.
Mycorrhizae can be mutualistic, neutral, or even parasitic in their impacts
on plant hosts (see Table 1), depending both on soil conditions (e.g., P
availability) and on the particular AM species ( Johnson, 2010; Moreno
Diaz, 1988). In low-P experiments testing only the impact of one AM strain,
crops typically benefit from mycorrhizal inoculation, a benefit the crop may
already receive in a biologically diverse soil. For example, Davies et al. (2005a)
found 44–57% increases in total dry matter with AM inoculation of a
Peruvian potato variety in sterile, low-P soil. By contrast, results of field
experiments with intact soil communities were mixed to positive (Table 1)
because inoculant AM do not guarantee a favorable plant C investment
compared to native AM species (Moreno Diaz, 1988; Rodriguez and
Ortuño, 2007; Rodriguez et al., 2010). Trichoderma continue to be attractive
Table 1 Example of impacts of mycorrhizal inoculation in sterile-control, greenhouse/nursery, and field inoculation of crops with
mycorrhizae reviewed for the Andean region

Sterile control/ Colonization Yield or disease Interactions with other inputs/

Source Country Crop field/greenhouse increase benefit (þ/
/0) treatments; comments
Davies et al. (2005a) Peru Potato Sterile control þ þ Mixed cultures from the field
outperform G. intraradices
Ortuño et al. (2010) Bolivia Maize Field nd þ
Rodriguez and Bolivia Onion Field þ þ Greater positive impact when
Ortuño (2007) combined with compost and
poultry manure
Rodriguez and Bolivia Potato Field þ

Ortuño (2007)
Ibarra (2008) Ecuador Pepper Field þ
Orna (2009) Ecuador Tomato Field þ/0 Additive effect of AM
inoculation at seeding and
transplant and P addition
Moreno Diaz (1988) Peru Potato Field 0
Medrano Echalar Bolivia Onion Greenhouse nd þ/0 Greater positive impact when
and Ortuño, 2007 seedlings combined with
vermicompost
Arandia et al. (2007) Bolivia Onion Greenhouse þ þ Mycorrhizal impact greater
than addition of rock
phosphate or wood vinegar
Ferrufino and Bolivia Heart of Nursery þ þ
Sanchez (2006) palm
Urgiles et al. (2009) Ecuador Tree spp.— Nursery þ þ Tropical trees, but shows
tropical viability of mycorrhizal root
inoculation
Moreno Diaz (1988) Peru Potato Seedbed na þ G. fasciculatum outperforms
other two strains
Andean Soil Fertility 149

inoculants since they are strong rhizosphere competitors via their antibiosis
and predation of other fungi, thus addressing a key challenge for inoculants
(Schuster and Schmoll, 2010; Verma et al., 2007). Trials in Bolivia have
found improved health and vigor of onion transplants when inoculated with a
Trichoderma harzianum (Medrano Echalar and Ortuño, 2007). As for other
inoculant research to date, trials on Trichoderma have focused frequently on
greenhouse or horticultural production, whereas future efforts should evaluate
their effectiveness in intensified smallholder crop rotations, especially where
soil disease pressure is high and positive effects would be most expected.
Research on PGPM that improve crop access to P is of particular interest
throughout the region and worldwide (Barroso and Nahas, 2005; Chen et al.,
2006; Jorquera et al., 2008; Keneni et al., 2010; Oliveira et al., 2009) and may
prove especially useful in P-sorbing volcanic soils (Andosols) in the northern
Andes. In Peru, Oswald et al. (2010) demonstrated the P-solubilizing activity
of 44% of Azotobacter and 58% of Bacillus bacterial strains isolated from the
rhizosphere in a single potato field, while 28% of bacterial isolates from
quinoa and one-quarter of rhizobial strains from fava bean in an Altiplano
collecting area had P-solubilizing ability (Ortuño, 2010). Despite the appar-
ently common occurrence of P-solubilizing bacteria in the Andes, their
impact on crop growth is not always clear. While Faccini et al. (2007)
found that P-solubilizing bacteria allowed equivalent potato yields with half
the P fertilizer, other researchers in the northern Andes observed little impact
of P-solubilizing bacteria on crop growth (Nustez and Acevedo, 2005;
Rodriguez et al., 2010). Previous work in the region has largely focused on
solubilizers of inorganic P forms, yet the prevalence of manure use in the
Andes suggests that increased focus on mineralization of organic P by PGPM
might be productive. Interestingly, P mineralizing and solubilizing bacteria
may have utility in both low-P environments and also in intensified farming
environments where fixation of applied fertilizer or manure P occurs (Chabot
et al., 1996).
Plant growth-promoting bacteria with hormonal modes of action pro-
vide glimpses of fascinating mechanisms of plant microbe interaction that
may prove useful in field settings. For example, inoculation with ethylene-
degrading PGPM has been found to increase yield of pea, tomatoes, and
peppers under in-season drought, by removing the ethylene stress signal to
roots and allowing a quicker resumption of growth when soil moisture
conditions improve (Arshad et al., 2008; Belimov et al., 2009; Mayak et al.,
2004). These results suggest that some soil microbes play a mediating role in
plant signaling to respond to stresses (Glick et al., 2007) and are especially
relevant given the risk of in-season drought for smallholders in the Andes.
For the inoculant organisms reviewed here, domains of soil properties
and management where inoculation is warranted need to be better char-
acterized. For the example of AM inoculants, high P availability in soil can
make mycorrhizal inoculation redundant or even inefficient for crops.
150 Steven J. Fonte et al.

However, agroecological intensification by smallholders will likely occupy

regimes of low-to-moderate soil fertility, as well as a mixture of inorganic
and organic nutrient inputs. Here, AM may in fact be well suited to
exploring available P pools with mutualistic outcomes for crop hosts
(Douds et al., 2007). In support of this idea, Mader et al. (2011) found
dramatic improvements in yields from the use of AM and PGPM, when
strains isolated from low-fertility fields in Indian rice–wheat systems were
used to inoculate low-fertility fields with degraded soils after flooding and
rice cropping. Of course, highly degraded soils may also prove unresponsive
to microbial inoculation, since soil C may be too depleted to provide for
proper microbial function and additional restorative measures must be taken
(e.g., organic matter inputs). Research on PGPM in the Andes may benefit
by isolating microbes from environments that select for desired microbial
traits as well as critical approaches to screening ( Jorquera et al., 2008;
Oswald et al., 2010). To avoid costly failures, this includes testing inoculants
across gradients of soil fertility and degradation to define where they have
greatest impact. It must also be noted that PGPM are highly multifunc-
tional, with their modes of action depending greatly on other rhizosphere
factors (Martinez-Viveros et al., 2010). Thus, producing a consistent inocu-
lant effect with these associative PGPM can be challenging, and research on
PGPM needs to acknowledge this possibility for prospective inoculant
microbes. A useful concept readily adapted from the pathology of soil-
borne plant diseases is that of soil receptivity and suppression for particular
microbes or microbial symbioses. Studies by Oyarzun et al. (1998) and
Herrera-Peraza et al. (2011) address this concept and exemplify how multi-
variate statistical techniques can relate microbial effectiveness to biotic and
physicochemical properties of soils and soil management and offer promising
insights for future research.
In addition to the direct effects of inoculation, research has focused on
factors that catalyze or facilitate impacts of PGPM or native AM on crop
growth. For example, Davies et al. (2005b) showed that simple additions of
the flavonoid signal formononetin triggered greater soil sporulation of
native AM fungi and increased in potato yield in Peru. Vanek (2010)
found that mycorrhizal legumes in highland Bolivian crop fields had a
higher ratio of uptake N:P and also AM colonization than did forage oat,
suggesting that legumes promote mutualistic AM function in crop rotations.
For Chilean Andosols, Borie et al. (2010) found that partially acidulated
rock phosphate was particularly suited to mycorrhizal uptake and bypassed
the P-fixation capacity of these soils. Additionally, biomass-derived charcoal
(biochar) represents a means of altering the habitat for rhizosphere microbes
in ways that can benefit crops and microbes alike. Pores in biochar derived
from the original vessels of plant material may provide protection from
drying and predation by other microbes and soil fauna, as well as changes in
pH, which for acidic soils and for most biochars will create more neutral-pH
Andean Soil Fertility 151

microsites that are favorable as bacterial habitat (Thies and Rillig, 2009). In
reviewing impacts of biochar on AM, Warnock et al. (2007) documented
neutral to very positive responses from char application on mycorrhizal
colonization. As possible mechanisms for AM–charcoal relations, they
advanced local alteration of soil physicochemical properties, indirect effects
via impacts on other soil microbes, plant–fungal signaling interference, and
protection from fungal grazers within charcoal. In addition to biochar,
zeolites, nanoporous aluminosilicates of varied chemical composition,
have the potential to increase the cation exchange and water-holding
capacity of soils (Ramesh et al., 2010). The combination of small pore size
and nutrient-holding capacity may make them especially active in micro-
bially mediated nutrient transformations that prolong the availability of P,
N, and other nutrients, especially via sorption of ammonium and other
cations (Flores Macias et al., 2007; Ippolito et al., 2011).
While functioning as a source of carbon and nutrients to support
decomposer food webs and improve soil physiochemical properties, com-
post and vermicomposts can also provide beneficial microbial inoculants
to soils that can competitively suppress soil-borne diseases and have similar
effects to inoculation with other rhizosphere microbes (Cardoza, 2011;
Litterick et al., 2004; Quilty and Cattle, 2011). In addition, composts and
other biomedia have been tested as vehicles to carry inoculant microbes into
the soil (Seneviratne et al., 2011; Singhai et al., 2011). Rock phosphate or
other minerals are sometimes mixed with the compost to foster their
solubilization (Lange, 1994). In effect, compost may be thought of as a
concentrated version of the effect desired in the whole soil for improving
crop access to nutrients or protecting crop roots from disease. Despite some
promising results with compost as inoculant promoters, their potential may
be limited in rural, extensive smallholder farms in the Andes, where labor to
produce compost may be at a premium and composting feedstocks have
competing uses. However, in intensive areas of smallholder production near
cities where waste feedstocks may be more available, composting as an
integrated source of beneficial microbes and available nutrients for crops
may hold greater promise.
The case of composting illustrates the need for inoculant technologies to
be accessible for smallholders in the Andes. Cost and local access must be
considered alongside technical optimization so that inoculant technologies
can achieve maximum impact for smallholders, and thus research on inocu-
lants from the region often includes analyses of economic and environmen-
tal costs and benefits. To avoid what Rosset and Altieri (1997) termed the
threat of import substitution in smallholder sustainable agriculture, researchers
have demonstrated promising results with local, on-farm methods for
inoculating crops with AM using roots of infected crops, often with supe-
rior results to commercial inoculants (Davies et al., 2005a, Douds et al.,
2007, Mäder et al., 2011). This approach is well suited to AM and Rhizobia,
152 Steven J. Fonte et al.

where propagules like spores are necessarily associated with roots and may
be more difficult with associative fungi and bacteria that have no intraradical
structures. Regardless, such field-to-field approaches for inoculation carry
risks for transmission of soil pathogens, which should be taken very seriously
given their potential for drastic impact in the Andes.
While the role of soil microbial communities has been well studied in
the Andes, the impacts of soil fauna in the region remain largely unknown,
despite their significant impact on numerous key soil processes in diverse
ecosystems around the globe. Soil fauna comprise a diverse group of
organisms and alter soil functioning in a variety of ways. Soil invertebrates
generally increase nutrient availability by (1) comminution of plant residues
and accelerating decomposition and mineralization processes and (2) feed-
ing on soil bacteria and fungi, thus releasing of nutrients bound within
microbial tissues (Lavelle and Spain, 2001). The maintenance of healthy soil
food webs (diverse and active soil fauna) has also been shown to control
plant pathogens and may have important implications for crop growth
(Blouin et al., 2005; Sanchez-Moreno and Ferris, 2007). Soil macrofauna,
particularly those representing ecosystem engineers (e.g., earthworms, ants,
termites), can drastically alter soil microbial communities via effects on soil
structure, water movement, nutrient dynamics, and SOM (Lavelle et al.,
1997). These activities can have varied implications for soil functioning but
have generally proven to be beneficial for plant growth (Brown et al., 1999;
Evans et al., 2011). Preliminary findings suggest that macrofauna can be
quite relevant for at least some agroecosystems in the Andes. For example,
Tonneijck and Jongmans (2008) examined biostructures and SOM distri-
bution in páramo soils of Ecuador and concluded that bioturbation by
earthworms was a fundamental process regulating the vertical distribution
and turnover of SOM in these soils. Findings of Morales and Sarmiento
(2002) suggest that soil disturbance associated with cropping reduces soil
macrofauna abundance and diversity in Venezuelan páramo. Although some
reports on soil fauna exist for other regions (Righi and van der Hammen,
1996; Zurita, 1997), information for highland Andean agroecosystems is
notably scarce. Given the global importance of soil fauna and their potential
to offer rapid and inexpensive indicators of soil health (Ferris et al., 2001;
Velasquez et al., 2007), improved understanding of soil faunal activity and
the factors affecting their abundance in agroecosystems remains a critical gap
for sustainable soil fertility management in the Andes.

2.2.5. Plant breeding for agroecological intensification and

climate change
Breeding of Andean crops to improve a large suite of desirable traits (for
example, disease resistance, product qualities, or nutrient use efficiency) is a
topic large enough to merit a separate review. Here, we focus on research
efforts related to breeding for improved adaptation to two major edaphic
Andean Soil Fertility 153

and climatic stresses, notably low nutrient availability and drought stress,
both of which will reduce crop yields under the likely scenarios of climate
change and soil degradation in the region. Great environmental and social
heterogeneity in the Andes somewhat limits the broad applicability of crop
breeding as an approach to agroecological intensification in the region.
However, we suggest that careful consideration of this diversity and
attempts to understand it can provide important opportunities for improv-
ing crop productivity and yield stability through breeding at both local and
regional scales.
With a few exceptions highlighted below, there are substantial gaps in
breeding and selection of highland Andean crops (native and introduced)
for improved performance under nutrient or drought stress. At the same
time, the range of variation of nutrient acquisition efficiency or drought
resistance already present in existing native varieties is poorly documented.
This lack of focus on stress tolerance for crops in the region may be traced to
activities that compete for breeders’ attention. These include expenditures
in resources and expertise toward assembling and maintaining important
agrobiodiversity collections (e.g., potatoes, quinoa); emphasis on forage
quality, market quality characteristics, and disease or pest resistance that
are more easily selected for at experiment stations (Sciarascia Mugnozza
et al., 1987); and a breeder bias toward wealthier farmers who have greater
access to fertility inputs and irrigation. Nevertheless, substantial interest in
abiotic stress breeding as well as a lack of adequate technical preparation for
local staff to undertake novel breeding approaches was revealed by a 2006
survey of Bolivian plant breeding programs (Gabriel, 2006), thus suggesting
high potential for new targeted breeding programs. Local biodiversity banks
and contacts with the CGIAR center breeding resources offer tremendous
opportunity to characterize existing genotypes for adaptation to low nutri-
ent supply and drought stress, and select new genotypes in stress environ-
ments along gradients of soil fertility and/or degradation. It should also
be noted that the lack of attention to breeding of Andean crops mirrors the
worldwide “orphaning” of these plants in comparison to other crops (e.g.,
maize, soybean, rice).
With the exception of soils of the Altiplano that have been characterized as
C- and N-limited, Andean valleys and particularly Andosols in the northern
Andes suffer from P limitation (Cárdenas et al., 2008; Dahlgren et al., 2004;
Valente and Oliver, 1993). Further, the soil P challenge in the Andes is one of
availability rather than absolute supply, suggesting that improving P foraging
ability of crops is a plausible strategy in this region. While N limitation is also
important for nonlegume crops in Andean agroecosystems (prompting inter-
est in N use efficiency for crops such as potatoes; Errebhi et al., 1999; Zebarth
et al., 2008), we suggest that N deficiencies are more easily addressed through
management (e.g., enhanced incorporation of legumes, see Sections 2.2.1 and
2.2.6) and merit less focus here.
154 Steven J. Fonte et al.

Substantial breeding investments have improved P foraging ability in

maize, bean, and soybean for lowland tropical P-limited environments
(Beebe et al., 2006; Ramaekers et al., 2010; Zhu et al., 2005). One transfer-
able element of these efforts is the concept that visible phenotypic char-
acteristics of crops such as root hair length, basal root number, aerenchyma
formation, and root shallowness (basal root gravitropism) predict P foraging
ability and can be used to rapidly screen promising genotypes for trials under
differential P availability (Ao et al., 2010; Beebe et al., 2006; Lynch, 2007).
Other traits such as organic acid exudation are also important in mobilizing
unavailable P and are known to occur in fava bean and Andean lupine
(Hocking and Jeffery, 2004; Pearse et al., 2006). Researchers have also
focused on mechanisms of plant tolerance to aluminum toxicity in weath-
ered soils of the lowland tropics (Beebe et al., 2009; Kochian et al., 2005),
which is of lower prevalence in the highland tropics, but may be an
important constraint to plant growth in Andosols of the Northern Andes
(Dahlgren et al., 2004). The link between root characteristics and P foraging
ability provides a logical entry point for efforts to evaluate existing varieties
and improve P access of Andean crops such as fava bean, lupine, or maize.
At the inception of breeding for P foraging in common bean, landraces
from a variety of geographic origins were compared for their P-foraging
ability (Beebe et al., 1997). This process may provide a blueprint for testing
germplasm of Andean legumes, quinoa, potato, and maize already gathered
by Andean breeders, as well as promising accessions of crops from outside
the region. Landraces selected by farmers in low-fertility environments may
already have promising traits for P foraging, and little is known about how
existing landraces are being selected by farmers as an adaptation to changing
climate and weather shocks. In situ conservation of existing stress-tolerant
varieties might be a useful complement to breeding approaches. However,
crosses and selection based on this characterization would be an important
further contribution of breeders, targeted toward gradients of fertility envir-
onments managed by subsistence farmers in the region.
As Noguera et al. (2011) suggest by their study of cultivar interactions
with earthworm and biochar management, it will be important that regional
breeding efforts shift focus from green revolution strategies that tested yields
under high soluble fertility regimes to breeding that interfaces more effec-
tively with the framework of agroecological intensification, for example,
microbial symbioses, direct and successional plant impacts on the rhizosphere,
biochar, and soil fauna (see previous sections of this review; Drinkwater and
Snapp, 2007). For example, Snoek et al. (2003) reviewed breeding approaches
for beans to improve rhizobial symbioses and N-fixation. Meanwhile,
Boomsma and Vyn (2008) reviewed the mechanisms by which breeding
maize for more effective symbiosis with AM could foster greater drought
tolerance. Andean crop varieties could be bred for (or might already exhibit)
better performance with mycorrhizae or organic nutrient sources, or even to
Andean Soil Fertility 155

enhance biological control of soil pests (Degenhardt et al., 2009). The positive
agroecological feedbacks from P foraging ability and other abiotic stress traits
in smallholder systems should be recognized. For example, Henry et al.(2009)
demonstrated that far from mining the soil for additional P, bean and soybean
varieties that scavenged unavailable P covered soil better in a smallholder
hillside context, reducing erosion P losses and creating a better P balance than
P-inefficient varieties.
Unpredictable weather extremes, including frost and in-season drought,
have always threatened crop productivity and food security in the Andes.
However, increases in average temperature at high elevations due to climate
change are predicted to exceed the global average (Bradley et al., 2006), so
that evapotranspiration may increasingly outstrip the water-supplying ability
of soil and both in-season and terminal drought stress may increase. There is
already evidence that farmers perceive increased drought risk and are adapting
their crop mixtures to climate change, for example, by increasing the propor-
tion of early-maturing barley relative to other crops on the Bolivian Altiplano
to avoid crop failure (A. Bonifacio, personal communication). Crop breeding
for drought resistance and yield stability to confront climate change is already
a worldwide effort, and much can be gained from linking Andean crop
breeders to these global efforts (Khan et al., 2010). It should also be noted
that plant traits targeted for agroecological intensification and climate change
are also likely to be appropriate for possible alterations in the soils of Andean
agroecosystems (discussed in Section 2.2.1), as most soil consequences of
climate change are likely to occur through alterations to plant cover.
Mechanisms of drought resistance in crops encompass drought escape
or earliness, dehydration avoidance (conservation of water in the plant, or
ways to access more water in the soil profile), and drought tolerance (Khan
et al., 2010). Drought resistance includes plant physiological mechanisms
that improve yields under drought stress, such as deep rooting and osmotic
adjustment of plant cells that allow photosynthesis and other metabolic
processes to support plant productivity even under drought-stressed condi-
tions (Blum, 2005; Chaves et al., 2003). Breeders and crop physiologists have
elucidated these and other traits conferring drought resistance at a physiologi-
cal level (Cattivelli et al., 2008), and are beginning to manipulate these traits
with marker-assisted breeding and other molecular approaches (Araus et al.,
2008; Miklas et al., 2006). Breeding methods for physiological drought
resistance apply across crop species and could be useful for breeding efforts
in the Andean region. Nevertheless, it should be recognized that drought
tolerance mechanisms like osmotic adjustment have had limited success in
explaining yield stability under drought, in comparison to avoidance of
drought stress via plant size, earliness, or root architecture for water capture.
A number of studies in the region and elsewhere have examined differ-
ences in drought resistance among genotypes of Andean crops. For legumes,
Siddique et al. (2001) tested a wide array of Mediterranean-type legumes
156 Steven J. Fonte et al.

and concluded that earliness of flowering and pod formation (drought escape)
was the dominant trait increasing yield under low rainfall. Studies on fava
bean suggest that drought resistance and better recovery from drought are
more due to stomatal closure and water conservation within the plant, rather
than osmotic adjustment (Amede et al., 1999; Khan et al., 2010; Sau and
Minguez, 2000). Analogous to root morphology traits used in screening for P
foraging ability, Khan et al. (2007) suggested a number of physiological
parameters of fava bean that are promising as physiological screening criteria
for drought resistance. Drought physiology has not been greatly studied in L.
mutabilis (see Carvalho et al., 2004), but Australian research on a number of
old-world lupines showed that yield under drought stress was related to
earliness and the speed of pod and bean filling rather than osmotic adjustment
of cells (French and Buirchell, 2005; Palta et al., 2007). Others have argued
that maximal water capture in the soil profile (e.g., deep rootedness observed
for L. mutabilis) and effective use of that water in plant productivity are more
important under drought stress than a sole focus on photosynthesis per unit of
water (Blum, 2009). Meanwhile, Tourneux et al. (2003a,b) examined six
potato varieties under drought stress at tuberization in Bolivia and suggested
that yield differences were not correlated with physiological parameters, but
rather were due to reduced leaf area as well as the inability to translocate
carbohydrates to tubers in sensitive varieties.
Regardless of physiological mechanisms of drought tolerance or avoid-
ance, approaches to drought stress breeding for rainfed farming systems such
as those in the Andes have recently emphasized the importance of selecting
environments for breeding. Selection of varieties in environments with a
particular drought stress, including the timing of the stress in the life cycle of
the plant, is important to success in breeding for drought tolerance (Ceccarelli
et al., 1991). For example, potatoes are highly sensitive to drought at tuber-
ization (DallaCosta et al., 1997); breeding efforts for drought tolerance in
maize focus on flowering and grain-filling stages (Banziger et al., 2006); and
fava beans are most sensitive to drought at the pod set and early pod-filling
phase (Khan et al., 2010). Ceccarelli et al. (1991) argued that the unpredictable
stresses found in many smallholder farm environments mean that a suite of
drought resistance traits must be selected for as a whole, a critique that has
fostered an evolutionary strategy of breeding with farmer participation that is
of interest to breeding efforts in the Andes (Ceccarelli et al., 2010). Never-
theless, it still makes sense to target the most vulnerable phases of crop growth
for a given stress as a first priority in stress breeding, to foster mechanistic
understanding of stress resistance and also since varieties that resist drought
stress during these vulnerable phases will likely fare better in the average year
when water becomes limiting.
Breeders working to select for stress tolerance in multi-environment trials
have developed expertise in the efficient design and sophisticated interpre-
tation of these trials, which often present challenges in teasing apart
Andean Soil Fertility 157

genotype by environment interactions (Banziger et al., 2006; Bucheyeki

et al., 2008; Setimela et al., 2005). One central and transferable concept from
these efforts is that multienvironment variety trials can be grouped by
“mega-environments” that embody similar soil types, climates, and agroe-
cosystems (Setimela et al., 2005) within which other sources of variation in
the trial (e.g., soil fertility and texture, drought stress) offer meaningful
gradients to compare the performance of varieties. A classification scheme
for Andean breeding environments and other smallholder environments
worldwide would summarize similarities and differences between Andean
and similar Mediterranean or highland climates elsewhere (e.g., east Africa,
southern Asia) and soils that could supply germplasm for Andean breeding.
Crop modeling using the performance of different genotypes may also be a
useful predictive tool in assessing impacts across environments. For exam-
ple, Condori et al. (2010) used crop modeling with experimental validation
in a number of environments across the Andes to assess the performance of
major potato varieties. This showed that crop growth modeling could be
used to predict the performance and usefulness of potato and bitter potato
cultivars with a wide range of characteristics, given weather data and a
number of scenarios of frost risk, and has applicability in rapid assessment
of new stress-tolerant crop varieties across rainfall and temperature gradients
in the Andes, and in the generation of hypotheses and research designs for
multienvironment trials.
A very different approach to breeding for sustainability, which must
be fit carefully to rotations in the Andean region, is the development of
perennial cereals such as those being tested in the United States and
Australia. These could dramatically reduce erosion and improve soil struc-
ture, water use from fluctuating precipitation, and benefits of rhizosphere
processes (Bell et al., 2010; Cox et al., 2002).
Last, we reiterate the importance of the efforts to collect and preserve
crop genetic diversity within the region for potato, quinoa, lupine, and
other Andean crops, and stress that in order for this genetic resource to have
maximal utility in confronting abiotic stresses for the region, it needs to be
more than just a repository of agricultural heritage. Andean germplasm
banks (and those elsewhere) should be screened and used in recurrent
breeding and other efforts that generate new, stress resistant varietal options
for smallholder farmers in the face of climate variability and change.

2.2.6. Spatial and temporal organization of farms

The above sections have focused on a variety of input-based technologies
and improved management options that offer promising advances to
improve soil fertility, productivity, and yield stability in the Andes. How-
ever, these technologies have largely been considered in isolation, over-
looking how farming practices might be integrated to advance overall
benefits for smallholder farmers. The spatial and temporal integration of
158 Steven J. Fonte et al.

agricultural practices at multiple scales plays a fundamental role in optimiz-

ing agroecosystem productivity and environmental services. For example,
the temporal integration of crops and livestock at the plot scale is a wide-
spread and highly successful practice, commonly observed in sectoral fallow
systems of the Andes (Molinillo and Monasterio, 2006; Pestalozzi, 2000).
The cropping period provides the majority of the food, while fallow
components serve the dual purpose of supplying fodder for animal produc-
tion and the recovery of soil nutrient stores, hydrologic function, and
biological communities (Sarmiento, 2000; Sarmiento and Bottner, 2002;
Sivila de Cary and Hervé, 1994). Combining these two phases on the same
plot of land yields higher productivity than either practice would in isolation.
Similarly, the production and transfer of manure from pasture areas to crops
represent another common, spatial integration of livestock with cropping
systems at the farm scale (Caycho-Ronco et al., 2009; Giller et al., 2006) that
improves overall productivity via redistribution of nutrients and organic
matter to where they are most needed. While integration of agricultural
production strategies may be common in the Andes, there remains consider-
able room for improvement. Further, there is a growing need to reinforce
these concepts, as agricultural intensification (i.e., shortened fallows, use of
agrochemicals) may be disrupting the balance and benefits associated with
long-established integrative practices (Claverı́as, 1994; Mayer, 1979).
In both fallow-based and continually cropped systems of the Andes, crop
rotation represents a widely utilized form of temporal diversification at the
plot scale (Caycho-Ronco et al., 2009; Tapia, 1994). Crop rotation (includ-
ing the use of cover crops) can provide key benefits for pest and disease
control, improved nutrient use, and to soil restoration (Liebman and Dyck,
1993; Smith et al., 2008; Snapp et al., 2005) and offers an indispensible tool of
sustainable management. A number of crops (e.g., Brassica sp., Crotalaria sp.)
are included in rotations specifically for their pest suppression effects
(Kirkegaard and Sarwar, 1998; Wang et al., 2002); however, less strategic
rotations of common crops from divergent plant families can also help to
break pest cycles, with more diverse rotations proving the most effective
(Miller et al., 2006). Even simple maize and soybean rotations have been
shown to outperform monocultures of either crop (Crookston et al., 1991),
though mechanisms are not always clear. Although there has been some
research on rotations in the Andes, much of this work has focused on residual
fertility for crops following heavily fertilized potatoes (e.g., Condori et al.,
1997). Research by Nieto-Cabrera et al. (1997) looking specifically at the role
of crop sequence in the Ecuadorean Andes found potatoes and quinoa
production to increase significantly following plantings of L. mutabilis, pre-
sumably due to additional N inputs associated with the legume. Other
benefits of legumes were noted for fava bean in Bolivian potato systems,
where application of residues has been suggested as a valuable tool for the
management of plant-parasitic nematodes (Iriarte et al., 1999).
Andean Soil Fertility 159

A promising area for improved spatial integration at the plot scale is the
implementation of polycultures, or intercropping. Combining multiple species
within the same field may complicate management in some cases but can
increase overall yield through increased species complementarity of resource
use (Fornara and Tilman, 2008). For example, Li et al. (2007) found a maize–
fava bean intercropping systems in China to overyield (produce more than
either crop would in monoculture) due to differing rooting depth and timing
of nutrient uptake (reduced competition) as well as improved availability of P
resulting from organic acid production by fava bean roots (facilitation). In this
same cropping system, increased competition for available N by maize forced
fava beans to fix more N from atmospheric sources, thus augmenting overall N
inputs to the system via N-fixation (Li et al., 2009). Just as for crop rotation,
increased spatial diversification in Andean agriculture can also help alleviate
pest issues, but through increases in predation/parasitism of pests and decreased
density of resources (Gianoli et al., 2006; Poveda et al., 2008). Although
offering clear potential for cropping situations, such benefits of intercropping
have been demonstrated in multiple ecosystem types and seem to be especially
well suited for grassland systems (Fornara and Tilman, 2008; Oberson et al.,
1999). Such practices could perhaps be most easily implemented during non-
cropping stages of sectoral fallow systems, as management requirements are
minimal and risks to farmers relatively low. However, there is also great
potential for more intensified crop–pasture rotations. The idea of improved
fallows and pastures (where species composition is intentionally managed) for
the Andes has been suggested previously (Barrios et al., 2005; Garcı́a, 2011;
Sarmiento et al., 2001), but principally with a focus on leguminous plants.
Others have suggested mixtures of legumes and grasses as an option to maxi-
mize N-fixation and productivity for fodder production (Bartl et al., 2009;
Bentley et al., 2007). Such mixtures could incorporate exotic plants that are
well adapted to local climates (e.g., cold tolerant, able to set seed before arrival
of the dry season; Wheeler et al., 1999), as well as native species that could help
reduce farmer investment and ensure survival (Bartl et al., 2009; Nezomba
et al., 2010). Further, improved fallows (with optimal varieties and species
mixtures) combined with strategic fertilization (particularly P additions) could
greatly enhance N-fixation, fodder production, and SOM stabilization, ulti-
mately accelerating the recovery of soil functioning and nutrient reserves.
Another means of enhancing spatial integration at the plot and farm level is
through agroforestry practices. Although potential for tree growth in some
parts of the Andes (i.e., very dry regions and/or high elevations) may be low,
evidence suggests that for large parts of the Andean highlands trees may have
been much more common than at present (Chepstow-Lusty and Winfield,
2000; Hensen, 2002), and that potential for agroforestry technologies may be
higher than often acknowledged (Cotler and Maass, 1999; Mahboubi et al.,
1997; Reynel and Felipe-Morales, 1990). For example, Mahboubi et al.
(1997) compared tree survival of candidate species in the Bolivian Altiplano
160 Steven J. Fonte et al.

and found several species (e.g., the native Buddleja coriacea) to tolerate condi-
tions well at elevations as high as 4200m, with a greater number of tree species
suitable for lower elevations. They also suggested that residues obtained from
these trees (all except Eucalyptus sp.) could provide valuable nutrients to
enhance crop growth. Comparisons of contour row species near Cocha-
bamba, Bolivia, suggest that woody species can grow well and also provide
effective erosion control (Sims et al., 1999). Additionally, the incorporation of
legumes (particularly woody species) into grass contour barriers has been
suggested as an effective means of contributing to soil fertility and rapid erosion
control, while reducing plant resource competition (Mutegi et al., 2008; Sims
and Rodriguez, 2001). Along with contour hedgerows, alternative options for
integrating woody species into agroecosystems include fodder banks, wood
lots, cut and carry systems, and/or windbreaks at field margins (Reynel and
Felipe-Morales, 1990; Young, 1997), as well as trees intermixed in the field, a
highly successful strategy for lower elevation hillsides used in the Quesungual
agroforestry system of Honduras (Fonte et al., 2010; Hellin et al., 1999).
Ultimately, the incorporation of trees into the landscape offers great potential
to provide farmers with supplementary fuel, fodder, and soil fertility amend-
ments (Barrios et al., 2005), while contributing key ecosystem services (e.g.,
erosion control, climate change mitigation, improved water dynamics) and
enhancing regional biodiversity (Antle et al., 2007; Otero and Onaindiam,
2009; Smukler et al., 2010). It should be noted, however, that proposed
agroforestry technologies must seek to minimize the potential negative
impacts of trees, such as increased competition with crops for light, water,
and nutrients; greater labor requirement; and promotion of crop pests (e.g.,
birds). Based on the relative paucity of materials encountered in our literature
search, research to date has yet to fully explore the full potential of tree-based
options in the Andes. Although the flexibility of agroforestry practices in the
Andean highlands is perhaps more limited than their lowland counterparts,
large potential exists and future research needs to address these possibilities.
In addressing spatial organization of farms, soil heterogeneity frequently
represents a major challenge for soil fertility management. Andean agroeco-
systems often have extreme and diverse topography which can generate steep
and somewhat predictable fertility gradients over the landscape (Buytaert
et al., 2007). However, fertility gradients are also generated by anthropogenic
factors at the farm scale, with improved nutrient status typically found closer
to the household (Tittonell et al., 2005; Vanek, 2010). In farming systems of
the east African highlands, Vanlauwe et al. (2006) suggested that fields further
from the farmer homes (outfields) should receive greater nutrients invest-
ment, as they are likely show a greater response to fertilizer additions.
However, in more mountainous regions such as the Andes, outfields may
represent fields with steeper slopes and/or rockier sites, where soils do not
have the same yield potential as flatter infield sites and increased nutrient
applications might be at greater risk to loss, via erosion (Vanek, 2010). In
either situation, there may be distinct soil fertility management strategies that
Andean Soil Fertility 161

reflect differential labor inputs associated with far versus near farm plots. For
example, in the Andes it may be better to apply manure and compost
generated near the home to infield plots, so as to minimize transport labor.
At the same time more distant plots, especially those prone to erosion, are
perhaps better candidates for cover crops and/or improved fallows combined
with strategic fertilizer applications in order to generate vegetative cover and
organic matter inputs on-site, and thus avoid transport. Such fertility gradients
are equally important at smaller scales as well and perhaps more manageable.
For example, Dercon et al. (2006) demonstrated that erosion within contour
hedgerows in Ecuador generated gradients of SOM, P, texture, such that
fertility was lowest on the upper portion of slopes and highest at the bottom,
where live barriers slowed runoff leading to soil deposition. Crop yield within
terraces varied significantly, with up to fourfold differences for wheat growing
at the top versus the bottom of a single terrace. They conclude that erosion
control is critical in slowing the formation of such gradients. However, it may
also be necessary to improve organic matter management (via manure or plant
residue additions) on the upslope portions of the terrace to restore SOM and
the capacity of these soils to provide and retain nutrients.
Organization at larger scales is also important for optimizing the
efficiency of farm management and ensuring a range of ecosystem func-
tions at the farm and landscape level. Although climate, topography, and
soil type largely determine where different crop and livestock elements are
placed, there is likely to be some room to better organize farms and
landscapes for increasing farm efficiency and the production of ecosystem
services and reducing labor. For example, active landscape management in
the Andes has been observed whereby farmers strategically disperse their
crop fields as means of distributing risks and improving overall yield
stability (Goland, 1993). At the same time, the physical arrangement of
crops and other farm attributes can greatly influence less obvious agroe-
cosystem functions in a variety of ways. The maintenance of biodiversity
and pollination services probably offers the clearest case for the role of
landscape organization in the provision of agroecosystem services. Polli-
nators are critical for the production of numerous crops around the globe
(Klein et al., 2007), including a number of important Andean crops (e.g.,
onions, fava beans, Andean lupine), and are greatly influenced by land-
scape structure and composition (Ricketts et al., 2008). Pest control agents
are also impacted by the landscape (Bianchi et al., 2006), but often in
counterintuitive ways (Parsa et al., 2011). Perhaps more relevant for soil
fertility, landscape diversity can influence soil fauna biodiversity and
distribution. For example, patches of native vegetation may function as a
reservoir or refuge of earthworms for adjacent agricultural plots, thus
maintaining their diversity, abundance, and activity on farmed plots
(Marichal, 2011). The composition of management attributes at the farm
level can also impact key ecosystem services such as C storage, N loss, and
erosion control (Robertson and Swinton, 2005; Smukler et al., 2010).
162 Steven J. Fonte et al.

3. Additional Considerations for Soil

Fertility Interventions
3.1. Need to incorporate co-limiting crop growth factors
While the availability of nutrients, optimal soil physical, chemical and
biological conditions, and appropriate cultivars are critical for agroecosys-
tem productivity, these factors alone are not sufficient to ensure high crop
yields. As mentioned earlier, water represents a fundamental limiting
constraint for much of the Andes, particularly for rainfed agroecosystems
of the Altiplano. The technologies advocated here for improving nutrient
balances thus need to consider impacts on crop–soil water dynamics. For
example, the use of live barriers for erosion control, N-fixation, and/or
fodder production can create competition for water and may ultimately
impact crop yields (Dercon et al., 2006; Pansak et al., 2007). At the same
time, enhanced inputs of organic matter may improve nutrient availability
and soil faunal activity with beneficial impacts for water infiltration and
storage (Capowiez et al., 2009). Along with water, other co-limiting
factors include light, temperature, as well as pest and disease. For example,
mulch-based technologies that work well for lower elevation, dry tropical
sites (Erenstein, 2003; Fonte et al., 2010) may not be appropriate for
cooler, high elevation sites where a residue layer may inhibit growth by
maintaining cooler soil temperatures (M. Scurrah, personal communica-
tion). Pests and disease factors present still more complex considerations
for soil fertility practices, as their responses to environmental change are
less predictable and often difficult to control. Although there is some
indication that improved soil fertility and biological functioning improve
plant resistance to pests (Bennet and Bever, 2007; Blouin et al., 2005;
Thamer et al., 2011), a number of management practices could potentially
exacerbate pest problems (Alyokhin et al., 2005; Conklin et al., 2002;
Parsa, 2010). While the relative importance of each of these co-limiting
factors depends on the environmental and management context, all factors
are essential to consider when developing agroecological options to
improve soil fertility.

3.2. Integrating local needs and knowledge into soil

fertility research
Limited adoption of new agricultural technologies has long been a challenge
for scientists and development professionals who work to refine and validate
these technologies, and lack of adoption is often conceptualized as an
important obstacle to improving livelihoods in rural communities (Ashby
et al., 1997). Some researchers have examined the way that implementation
Andean Soil Fertility 163

costs, lack of short-term payoffs, insufficient or unstable land tenure, or

access to information can hinder adoption (Bayard et al., 2006; Tenge et al.,
2004). While these economic and information limitations to adoption
certainly play a role, scientists studying extension and rural development
suggest that adoption or adaptation of new exogenous technologies will
only result when greater attention is given to existing knowledge structures
and perceived needs of farmers, as well as processes by which farmer
learning and innovation take place in rural communities (Bentley, 1989;
Deugd et al., 1998; Paredes, 2010). As mentioned earlier, many farming
communities in the Andes benefit from a long history of agricultural
knowledge and often possess an intimate and complex understanding of
their land (e.g., Sandor and Furbee, 1996). Despite this attribute, there often
exists a general disconnect between farmer needs and researcher objectives
(Horton, 1983). This can lead to low adoption rates and, further, suggests
the need for greater farmer involvement in the development of new
management strategies (Altieri, 2004; Barrios and Trejo, 2003; Bentley
et al., 2007). While increased farmer interaction is critical, farmers are not
always able to express their demands in the same knowledge system used by
researchers (Bentley et al., 2007). Thus, learning from both sides and mutual
involvement in the research process are critical for the development of
relevant agroecological intensification strategies that have a high probability
of acceptance among farmers (Ashby et al., 1997; Winklerprins, 1999). At
the same time, farmer involvement alone may not be sufficient for the
development of successful technologies. There exists a clear need to develop
flexible strategies that allow for farmers to better understand the mechanisms
behind new technologies and adjust them to their local needs as well as
to adapt them to rapidly changing socioeconomic and environmental con-
ditions (Bentley et al., 2007). Further, new technologies need to account
for community dynamics and local decision-making processes, as well as
regional policies and political structures. Ultimately, the integration of local
knowledge and farmer needs into the research process is fundamental for
developing practical soil fertility technologies that have a high probability of
adoption and achieving regional impact.

4. Conclusions and Recommendations

4.1. Outlook for agroecological intensification in the
Andean context
The Andean region faces key challenges in coming years, yet offers a number
of unique attributes that may allow for successful agroecological intensifica-
tion and betterment of rural livelihoods throughout the region. As discussed
earlier, this region is endowed with a large native crop diversity and
164 Steven J. Fonte et al.

demonstrates a strong interest in maintaining local varieties and indigenous

management techniques (Brush et al., 1995). Further, many smallholder
agroecosystems in the Andes already display considerable within-farm spatial
and temporal diversity, complex crop rotations (or fallow sequences), and
diverse crop mixtures, suggesting that farmers are already accustomed to key
principles of sustainability and perhaps more flexible in their farm manage-
ment strategies. Such attributes combined with well-developed local knowl-
edge have the potential to facilitate innovation and improve adaptation to
new conditions. Despite these attributes, the region faces many challenges
and the need for change is clear. While increased reliance upon agrochemical
inputs and ‘green revolution’ technologies represents one alternative to
traditional practices that has occurred elsewhere, this route may not be eco-
nomically or ecologically viable in the Andes due to the relative fragility of
soils and landscapes, low yield potential, and the minimal financial resources
of rural farmers. Modifying classic short-term, yield-driven definitions of
agricultural success may allow us to develop more appropriate development
strategies for the region. The success of a particular intervention must con-
sider not only short-term yield and economic returns but also long-term
stability of yields, ecological resilience, and the environmental consequences
of a given practice. Taking these factors into consideration, agroecological
intensification, as discussed here, has the potential to significantly improve the
productivity and stability of Andean agriculture.
This chapter attempts to take a pragmatic approach to soil fertility man-
agement options in the Andes. We consider a balance of applied research,
with more immediate prospects to improve livelihoods of Andean farmers
(e.g., ISFM, cover cropping), versus innovations associated with more basic
science that are less certain and have a longer return on investment, but may
yield potentially greater impacts (e.g., metagenomics, perennialization of
crops). Although we focus principally on ecologically based interventions,
we also consider the prudent use of agrochemicals, new cultivars, and
mechanization as valuable tools to help farmers increase overall agroecosystem
function on their farms. From the above discussion of past research and
technology options in the Andean highlands, we attempt to offer a brief
and targeted list of the most promising areas for research and implementation.
We want to emphasize that there exists no single technology that is likely to
lead to successful agroecological intensification of Andean farms across all
sites. Rather, we propose that a suite of technology options is likely required
and the precise combination likely varies by climate, soil type, as well as
political and cultural setting. For each technology, we stress the importance of
supporting farmers in experimentation and adaptation to their local needs. In
this review, we aimed to sort through the various options and offer research-
ers and extension workers the basic information and guidelines for future
research and implementation projects that will contribute to poverty allevia-
tion, improved food security, and environmental sustainability (for a more
Andean Soil Fertility 165

detailed list of recommendations, see http://mcknight.ccrp.cornell.edu/

projects/andes_cop/andescop.html).

4.2. Recommendations for future research and interventions

1. Soil conservation: There is a clear need for continued investment in soil
conservation, particularly for hillside soils and areas where mechanized
tillage has become more common. Erosion prevention strategies should
focus on barriers with low capital and labor investment (e.g., live
barriers) as well as improved management of fallow plots (e.g., cover
cropping). Better understanding of how current management practices
impact erosion processes and participatory development of new small
holder technologies are needed to combat erosion losses and maintain
more positive farm nutrient balances.
2. Biomass production: Organic matter has become highly limiting in Andean
agroecosystems (particularly in the Altiplano) due to competing uses of
fuel, fodder, and soil fertility amendments, thus indicating a need for
greater farm level biomass production. Efforts to address this issue should
emphasize the incorporation of N-fixing species both within crop fields
(e.g., polycultures, live barriers) and at field margins as a means to
enhance soil fertility and improve the overall farm N balances with
minimal farmer inputs.
3. Alternative nutrient sources: Large scale nutrient imbalances suggest the
need for the identification of new potential sources of nutrient inputs.
Waste streams of existing industries (e.g., poultry, floriculture) and
rock phosphate offer promising fertility sources for many farming
communities. Improved market access and transportation of these
materials (as well as conventional agricultural inputs such as fertilizer)
to more isolated farmers would contribute greatly to restoring farm
nutrient balances.
4. Improved management of organic residues would help to improve nutrient-
use efficiency, while contributing to soil biological functioning and
SOM storage. Further, the strategic application of inorganic fertilizers
with organic resources offers a particularly useful tool for controlling
nutrient release from a variety of organic materials and rapidly correcting
nutrient imbalances, thus allowing farmers to maximize the potential of
locally available organic resources.
5. Soil ecology: Continued research must strive to understand and optimize
soil microbial and faunal communities and their activity through proper
management (e.g., residue or charcoal additions, rotation) and inocula-
tive strategies. For microbial inoculants, rigorous and standardized
screening in the lab must be accompanied by field testing under diverse
conditions to fully evaluate the potential of these organisms to contribute
to improved crop performance. Soil metagenomics and other basic
166 Steven J. Fonte et al.

research on plant–microbe relations in the rhizosphere may also offer

longer-term contributions to principles and solutions for research.
6. Targeted breeding efforts should take advantage of the high agrobiodiversity
associated with the local domestication of multiple indigenous crops in
the Andes and seek to optimize interactions with key root symbionts
(e.g., Rhizobia, mycorrhizae, soil fauna) and alternative nutrient sources
that form the basis of agroecological intensification. Additionally, breed-
ing efforts should draw upon approaches from other crops and regions
with severe edaphic and/or climatic constraints to more rapidly improve
the productivity and stability of key orphan crops, including supporting
crops such as fava bean and Andean lupine.
7. Improved spatial and temporal organization of farms would contribute to
more efficient use of land and the improved provision of ecosystem
services (e.g., C sequestration, N-fixation, erosion control, pest control).
Research on diversification at the plot level should explore promising
plant synergies between functionally distinct species in order to enhance
crop utilization of resources. Additionally, improved understanding of
soil fertility gradients and options for strategic management of within-
farm heterogeneity would help farmers to improve overall farm produc-
tivity and yield stability in Andean agroecosystems.
8. Long-term and basic soils research: There is a need for basic research and
detailed soil maps to improve our understanding of soil physical, chemical,
and biological processes in the Andes and the potential impacts of new
management strategies throughout the region. Additionally, long-term
agricultural research and monitoring is notably lacking in the region, yet
would greatly improve our ability to understand the more far-reaching
impacts of land management and global change.

ACKNOWLEDGMENTS
We thank a number of people who provided assistance and advice at various stages in the
development of this chapter including Sady Garcia, Carlos Meneses, Maria Scurrah, Julio
Alegre, Ruddy Meneses, Karl Zimmerer, Noel Ortuño, Marcela Quintero, Edward
Guevara, Oscar Ortiz, Javier Aguilera, Alejandro Bonifacio, Claire Nicklin, Carlos Perez,
and many others. This work was made possible by support from the McKnight Foundation
Collaborative Crop Research Program.

REFERENCES
Abadı́n, J., González-Prieto, S. J., Sarmiento, L., Villar, M. C., and Carballas, T. (2002).
Successional dynamics of soil characteristics in a long fallow agricultural system of the
high tropical Andes. Soil Biol. Biochem. 34, 1739–1748.
Andean Soil Fertility 167

Abreu, Z., Llambı́, L. D., and Sarmiento, L. (2009). Sensitivity of soil restoration indicators
during Páramo succession in the high tropical Andes: Chronosequence and permanent
plot approaches. Restor. Ecol. 17, 619–628.
Aguilera, J. (2010). Impacts of Soil Management Practices on Soil Fertility in Potato-Based
Cropping Systems in the Bolivian Andean Highlands. University of Missouri, Columbia.
Alegre, J. C., Felipe-Morales, C., and LaTorre, B. (1990). Soil erosion studied in Peru. J. Soil
Water Conserv. 45, 417–420.
Altieri, M. (2004). Linking ecologists and traditional farmers in the search for sustainable
agriculture. Front. Ecol. Environ. 2, 35–42.
Alyokhin, A., Porter, G., Groden, E., and Drummond, F. (2005). Colorado potato beetle
response to soil amendments: A case in support of the mineral balance hypothesis? Agric.
Ecosyst. Environ. 109, 234–244.
Amede, T., Kittlitz, E. V., and Schubert, S. (1999). Differential drought responses of faba
bean (Vicia faba L.) inbred lines. J. Agron. Crop Sci. 183, 35–45.
Antle, J. M., Stoorvogel, J. J., and Valdivia, R. O. (2007). Assessing the economic impacts of
agricultural carbon sequestration: Terraces and agroforestry in the Peruvian Andes. Agric.
Ecosyst. Environ. 122, 435–445.
Ao, J., Fu, J., Tian, J., Yan, X., and Liao, H. (2010). Genetic variability for root morph-
architecture traits and root growth dynamics as related to phosphorus efficiency in
soybean. Funct. Plant Biol. 37, 304–312.
Arandia, W., Ortuño, N., Gutierrez, E., and Cáceres, A. (2007). Evaluación en invernadero
de la respuesta del cultivo de cebolla (Allium cepa) de la variedad rosada criolla al uso
combinado con aciduantes y micorrizas (m.a.). Facultad de Agronomı́a. Universidad
Mayor San Simon, Cochabamba, Bolivia.
Araus, J. L., Slafer, G. A., Royo, C., and Serret, M. D. (2008). Breeding for yield potential
and stress adaptation in cereals. Crit. Rev. Plant Sci. 27, 377–412.
Aroca, R., and Ruiz-Lozano, J. M. (2009). Induction of plant tolerance to semi-arid environ-
ments by beneficial soil microorganisms—A review. Sustain. Agric. Rev. 2, 121–135.
Arshad, M., Shaharoona, B., and Mahmood, T. (2008). Inoculation with Pseudomonas spp.
containing ACC-deaminase partially eliminates the effects of drought stress on growth,
yield, and ripening of pea (Pisum sativum L.). Pedosphere 18, 611–620.
Ashby, J., Beltran, A. J., Guerrero, M., and Ramos, H. F. (1997). Improving the acceptabil-
ity to farmer of soil conservation practices. J. Soil Water Conserv. 51, 309–312.
Bale, J. S., Masters, G. J., Hodkinson, I. D., Awmack, C., Bezemer, T. M., Brown, V. K.,
Butterfield, J., Buse, A., Coulson, J. C., Farrar, J., Good, J. E. G., Harrington, R., et al.
(2002). Herbivory in global climate change research: Direct effects of rising temperature
on insect herbivores. Glob. Change Biol. 8, 1–16.
Banziger, M., Setimela, P., Hodson, D., and Vivek, B. (2006). Breeding for improved
abiotic stress tolerance in maize adapted to southern Africa. Agric. Water Manage. 80,
212–224.
Barba, R., Coca, G., Tuma, N., and Pijnenborg, J. (2000). Pre-selección de cepas de (Brady)
rhizobium para leguminosas de grano en Bolivia. In “Compendio de trabajos presentados
por el proyecto rizobiologı́a (Cochabamba) en eventos y publicaciones de otras institu-
ciones” (R. Meneses, Ed.), pp. 47–49. Proyecto Rizobiologı́a Bolivia (CIAT-CIF-
PNLG-CIFP-UAW/DHV), Cochabamba, Bolivia.
Barrios, E., and Trejo, M. T. (2003). Implications of local soil knowledge for integrated soil
management in Latin America. Geoderma 111, 217–231.
Barrios, E., Cobo, J. G., Rao, I. M., Thomas, R. J., Amezquita, E., Jimenez, J. J., and
Rondon, M. A. (2005). Fallow management for soil fertility recovery in tropical Andean
agroecosystems in Colombia. Agric. Ecosyst. Environ. 110, 29–42.
Barroso, C. B., and Nahas, E. (2005). The status of soil phosphate fractions and the ability of
fungi to dissolve hardly soluble phosphates. Appl. Soil Ecol. 29, 73–83.
168 Steven J. Fonte et al.

Bartl, K., Gamarra, J., Gómez, C. A., Wettstein, H. R., Kreuzer, M., and Hess, H. D.
(2009). Agronomic performance and nutritive value of common and alternative grass and
legume species in the Peruvian highlands. Grass Forage Sci. 64, 109–121.
Bashan, Y., and Bashan, L. E.d. (2010). Chapter two—How the plant growth-promoting
bacterium Azospirillum promotes plant growth—A critical assessment. Adv. Agron. 108,
77–136.
Bayard, B., Curtis, M. J., and Dennis, S. (2006). The economics of adoption and manage-
ment of alley cropping in Haiti. J. Environ. Manage. 84, 62–70.
Bebbington, A. J., Carrasco, H., Peralbo, L., Ramon, G., Trujillo, J., and Torres, V. (1993).
Fragile lands, fragile organizations: Indian organizations and the politics of sustainability
in Ecuador. Trans. Inst. Br. Geogr. 18, 179–196.
Beebe, S., Lynch, J., Galwey, N., Tohme, J., and Ochoa, I. (1997). A geographical approach
to identify phosphorus-efficient genotypes among landraces and wild ancestors of com-
mon bean. Euphytica 95, 325–336.
Beebe, S. E., Rojas-Pierce, M., Yan, X., Blair, M. W., Pedraza, F., Munoz, F., Tohme, J.,
and Lynch, J. P. (2006). Quantitative trait loci for root architecture traits correlated with
phosphorus acquisition in common bean. Crop. Sci. 46, 413–423.
Beebe, S., Rao, I. M., Blair, M. W., and Butare, L. (2009). Breeding for abiotic stress tolerance
in common bean: Present and future challenges. SABRAO J. Breed. Genet. 41, 1–11.
Belimov, A. A., Dodd, I. C., Hontzeas, N., Theobald, J. C., Safronova, V. I., and Davies, W. J.
(2009). Rhizosphere bacteria containing 1-aminocyclopropane-1-carboxylate deaminase
increase yield of plants grown in drying soil via both local and systemic hormone signalling.
New Phytol. 181, 413–423.
Bell, L. W., Wade, L. J., and Ewing, M. A. (2010). Perennial wheat: A review of environmental
and agronomic prospects for development in Australia. Crop Pasture Sci. 61, 679–690.
Benizri, E., Baudoin, E., and Guckert, A. (2001). Root colonization by inoculated plant
growth-promoting rhizobacteria. Biocontrol Sci. Techn. 11, 557–574.
Bennet, A. E., and Bever, J. D. (2007). Mycorrhizal species differentially alter plant growth
and response to herbivory. Ecology 88, 210–218.
Bentley, J. W. (1989). What farmers don’t know can’t help them: The strengths and weak-
nesses of indigenous technical knowledge in Honduras. Agric. Human Values 6, 25–31.
Bentley, J., Velasco, C., Rodrı́guez, F., Oros, R., Botello, R., Webb, M., Devaux, A., and
Thiele, G. (2007). Unspoken demands for farm technology. Int. J. Agric. Sustain. 5, 70–84.
Bianchi, F. J. J. A., Booij, C. J. H., and Tscharntke, T. (2006). Sustainable pest regulation in
agricultural landscapes: A review on landscape composition, biodiversity and natural pest
control. Proc. R. Soc. B. Biol. Sci. B 273, 1715–1727.
Blouin, M., Zuily-Fodil, Y., Pham-Thi, A. T., Laffray, D., Reversat, G., Pando, A.,
Tondoh, J., and Lavelle, P. (2005). Belowground organism activities affect plant above-
ground phenotype, inducing plant tolerance to parasites. Ecol. Lett. 8, 202–208.
Blum, A. (2005). Drought resistance, water-use efficiency, and yield potential—Are they
compatible, dissonant, or mutually exclusive? Aust. J. Agr. Res. 56, 1159–1168.
Blum, A. (2009). Effective use of water (EUW) and not water-use efficiency (WUE) is the
target of crop yield improvement under drought stress. Field Crop. Res. 112, 119–123.
Boomsma, C. R., and Vyn, T. J. (2008). Maize drought tolerance: Potential improvements
through arbuscular mycorrhizal symbiosis? Field Crops Res. 108, 14–31.
Borie, B. F., Rubio, R., Morales, A., Curaqueo, G., and Cornejo, P. (2010). Arbuscular
mycorrhizae in agricultural and forest ecosystems in Chile. J. Soil Sci. Plant Nut. 10,
185–206.
Bossio, D., and Cassman, K. (1991). Traditional rainfed barley production in the Andean
highlands of ecuador: Soil nutrient limitations and other constraints. Mt. Res. Dev. 11,
115–126.
Bottner, P., Pansu, M., Sarmiento, L., Dominique Hervé, D., Ruben Callisaya-Bautista, R.,
and Metselaar, K. (2006). Factors controlling decomposition of soil organic matter in
Andean Soil Fertility 169

fallow systems of the high tropical Andes: A field simulation approach using 14C- and
15N-labelled plant material. Soil Biol. Biochem. 38, 2162–2177.
Bradley, R. S., Vuille, M., Diaz, H. F., and Vergara, W. (2006). Threats to water supply in
the Tropical Andes. Science 312, 1755–1756.
Briar, S. S., Fonte, S. J., Park, I., Six, J., Scow, K. M., and Ferris, H. (2011). The distribution
of nematodes and soil microbial communities across soil aggregate fractions and farm
management systems. Soil Biol. Biochem. 43, 905–914.
Briones, M. J. I., Ostle, N. J., McNamara, N. P., and Poskitt, J. (2009). Funtional shift
of grassland communities in response to soil warming. Soil Biol. Biochem. 41, 315–322.
Bronick, C. J., and Lal, R. (2005). Soil structure and management: A review. Geoderma 124,
3–22.
Brown, G. G., Pashanasi, B., Villenave, C., Patrón, J. C., Senapati, B. K., Giri, S., Barios, I.,
Lavelle, P., Blanchart, E., Blakemore, R. J., Spain, A. V., and Boyer, J. (1999). Effects of
earthworms on plant production in the tropics. In “Earthworm Management in Tropical
Agroecosystems” (P. Lavelle, L. Brussaard, and P. Hendrix, Eds.), pp. 87–147. CAB
International, Oxon.
Brush, S., Kesseli, R., Ortega, R., Cisneros, P., Zimmerer, K., and Quiros, C. (1995). Potato
diversity in the Andean center of crop domestication. Conserv. Biol. 19, 1189–1198.
Brussaard, L., Pulleman, M. M., Ouedraogo, E., Mando, A., and Six, J. (2006). Soil fauna
and soil function in the fabric of the food web. Pedobiologia 50, 447–462.
Bucheyeki, T. L., Shenkalwa, E. M., Mapunda, T. X., and Matata, L. W. (2008). On-farm
evaluation of promising groundnut varieties for adaptation and adoption in Tanzania. Afr.
J. Agric. Res. 3, 531–536.
Bunch, R. (2004). Adopción de abonos verdes y cultivos de cobertura. LEISA Revista de
Agroecologı´a 19(4), 11–13.
Buytaert, W., Deckers, J., Dercon, G., De BieÁvre, B., Poesen, J., and Govers, G. (2002).
Impact of land use changes on the hydrological properties of volcanic ash soils in South
Ecuador. Soil Use Manage. 18, 94–100.
Buytaert, W., Rolando Célleri, R., De Bièvre, B., Cisneros, F., Wyseure, G., Deckers, J.,
and Hofstede, R. (2006). Human impact on the hydrology of the Andean páramos. Earth
Sci. Rev. 79, 53–72.
Buytaert, W., Deckers, J., and Wyseure, G. (2007). Regional variability of volcanic ash soils in
south Ecuador: The relation with parent material, climate and land use. Catena 70, 143–154.
Buytaert, W., Cuesta-Camacho, F., and Tobón, C. (2010a). Potential impacts of climate
change on the environmental services of humid tropical alpine regions. Glob. Ecol.
Biogeogr. 20, 19–33.
Buytaert, W., Vuille, M., Dewulf, A., Urrutia, R., Karmalkar, A., and Célleri, R. (2010b).
Uncertainties in climate change projections and regional downscaling in the tropical Andes:
Implications for water resources management. Hydrol. Earth Syst. Sci. 14, 1247–1258.
Capowiez, Y., Cadoux, S., Bouchant, P., Ruy, S., Roger-Estrade, J., Richard, G., and
Boizard, H. (2009). The effect of tillage type and cropping system on earthworm
communities, macroporosity and water infiltration. Soil Till. Res. 105, 209–216.
Cárdenas, J., Choque, W., and Guzman, R. (2008). Fertilidad, uso, y manejo de suelos en la
zona del intersalar, departamentos de: Oruro y Potosı́. Fundación AUTAPO, programa
Quinoa sur, Oruro, Bolivia.
Cardoza, Y. J. (2011). Arabidopsis thaliana resistance to insects mediated by an earthworm-
produced organic soil amendment. Pest Manag. Sci. 67, 233–238.
Carney, K. M., and Matson, P. A. (2005). Plant communities, soil microorganisms, and soil
carbon cycling: Does altering the world belowground matter to ecosystem functioning?
Ecosystems 8, 928–940.
Carvalho, I. S., Ricardo, C. P., and Chaves, M. (2004). Quality and distribution of
assimilates within the whole plant of lupines (L-albus and L-mutabilis) influenced by
water stress. J. Agron. Crop Sci. 190, 205–210.
170 Steven J. Fonte et al.

Cassman, K. G., Dobermann, A., and Walters, D. T. (2002). Agroecosystems, nitrogen-use

efficiency, and nitrogen management. Ambio 31, 132–140.
Cattivelli, L., Rizza, F., Badeck, F. W., Mazzucotelli, E., Mastrangelo, A. M., Francia, E.,
Mare, C., Tondelli, A., and Stanca, A. M. (2008). Drought tolerance improvement in
crop plants: An integrated view from breeding to genomics. Field Crop Res 105, 1–14.
Caycho-Ronco, J., Arias-Mesia, A., Oswald, A., Esprella-Elias, R. A., Rivera, A.,
Yumisaca, F., and Andrade-Piedra, J. (2009). Tecnologı́as sostenibles y su uso en la
producción de papa en la región altoandina. Revista Latinoamericana de la Papa 15, 20–37.
Ceccarelli, S., Acevedo, E., and Grando, S. (1991). Breeding for yield stability in unpredict-
able enviroments—Single traits, interaction between traits, and architecture of geno-
types. Euphytica 56, 169–185.
Ceccarelli, S., Grando, S., Maatougui, M., Michael, M., Slash, M., Haghparast, R.,
Rahmanian, M., Taheri, A., Al-Yassin, A., Benbelkacem, A., Labdi, M., Mimoun, H.,
et al. (2010). Plant breeding and climate changes. J. Agric. Sci. 148, 627–637.
Chabot, R., Antoun, H., and Cescas, M. P. (1996). Growth promotion of maize and lettuce
by phosphate-solubilizing Rhizobium leguminosarum biovar phaseoli. Plant Soil 184,
311–321.
Chalk, P. M. (2000). Integrated effects of mineral nutrition on legume performance. Soil
Biol. Biochem. 32, 577–579.
Chan, K. Y., Van Zwieten, L., Meszaros, I., Downie, A., and Joseph, S. (2007). Agronomic
values of greenwaste biochar as a soil amendment. Aust. J. Soil Res. 45, 629–634.
Chaves, M. M., Maroco, J. P., and Pereira, J. S. (2003). Understanding plant responses to
drought: From genes to the whole plant. Funct. Plant Biol. 30, 239–264.
Chen, Y. P., Rekha, P. D., Arun, A. B., Shen, F. T., Lai, W. A., and Young, C. C. (2006).
Phosphate solubilizing bacteria from subtropical soil and their tricalcium phosphate
solubilizing abilities. Appl. Soil Ecol. 34, 33–41.
Chepstow-Lusty, A., and Winfield, M. (2000). Inca agroforestry: Lessons from the past.
Ambio 29, 322–328.
Chivenge, P., Vanlauwe, B., and Six, J. (2010). Does the combined application of organic and
mineral nutrient sources influence maize productivity? A meta-analysis. Plant Soil 342, 1–30.
Claverı́as, R. (1994). Causas de la reducción del periodo de descanso de las tierras en
comunidades campesinas de Puno: Alternativas para la sostenibilidad. In “Dinámicas del
descanso de la tierra en los Andes” (D. Hervé, Ed.), pp. 249–258. IBTA—OSTROM, La
Paz, Bolivia.
Cobo, J. G., Dercon, G., and Cadisch, G. (2010). Nutrient balances in African land use
systems across different spatial scales: A review of approaches, challenges and progress.
Agric. Ecosyst. Environ. 136, 1–15.
Cofie, O. O., Kranjac-Berisavljevic, G., and Drechsel, P. (2005). The use of human waste
for peri-urban agriculture in Northern Ghana. Renew. Agric. Food Syst. 20, 73–80.
Compant, S., Clement, C., and Sessitsch, A. (2010). Plant growth-promoting bacteria in the
rhizo- and endosphere of plants: Their role, colonization, mechanisms involved and
prospects for utilization. Soil Biol. Biochem. 42, 669–678.
Conde, M., Gutierrez, S., and Meneses, R. (2000). Respuesta a fertilizantes orgánicos y
cepas del género Rhizobium, en el cultivo de haba en invernadero y campo.
In “Compendio de trabajos presentados por el proyecto rizobiologı́a (Cochabamba) en
eventos y publicaciones de otras instituciones” (R. Meneses, Ed.), pp. 237–240. Proyecto
Rizobiologı́a Bolivia (CIAT-CIF-PNLG-CIFP-UAW/DHV), Cochabamba, Bolivia.
Condori, B., Devaux, A., Mamani, P., Vallejos, J., and Blajos, J. (1997). Efecto residual de la
fertilización del cultivo de papa sobre el cultivo de haba (Vicia faba L.) en el sistema de
rotación. Revista Latinoamericana de la Papa 9(10), 171–187.
Condori, B., Hijmans, R. J., Quiroz, R., and Ledent, J.-F. (2010). Quantifying the
expression of potato genetic diversity in the high Andes through growth analysis and
modeling. Field Crop Res. 119, 135–144.
Andean Soil Fertility 171

Conklin, A. E., Erich, M. S., Liebman, M., Lambert, D., Gallandt, E. R., and
Halteman, W. A. (2002). Effects of red clover (Trifolium pratense) green manure and
compost soil amendments on wild mustard (Brassica kaber) growth and incidence of
disease. Plant Soil 238, 245–256.
Coppus, R., Imeson, A. C., and Sevink, J. (2003). Identification, distribution and characteristics
of erosion sensitive areas in three different Central Andean ecosystems. Catena 51, 315–328.
Córdoba, J. J., and Novoa, V. (1997). Problematica, experiencias y enfoque sobre la erosión,
manejo y conservación de suelos de ladera en Ecuador. In “Manejo Integral de Micro-
cuencas” (M. Tapia, Ed.), pp. 123–133. International Potato Center (CIP), Lima.
Cotler, H., and Maass, J. M. (1999). Tree management in the northwestern Andean
Cordillera of Peru. Mt. Res. Dev. 19, 153–160.
Coûteaux, M., Hervé, D., and Mita, V. (2008). Carbon and nitrogen dynamics of potato
residues and sheep dung in a two-year rotation cultivation in the Bolivian altiplano.
Commun. Soil Sci. Plant 39, 475–498.
Cox, T. S., Bender, M., Picone, C., Van Tassel, D. L., Holland, J. B., Brummer, E. C.,
Zoeller, B. E., Paterson, A. H., and Jackson, W. (2002). Breeding perennial grain crops.
Crit. Rev. Plant Sci. 21, 59–91.
Craswell, E. T., and Lefroy, R. D. B. (2001). The role and function of organic matter in
tropical soils. Nutr. Cycl. Agroecosyst. 61, 7–18.
Craswell, E. T., Sajjapongse, A., Howlett, D. B. J., and Dowling, A. J. (1998). Agroforestry
in the management of sloping lands in Asia and the Pacific. Agroforest. Syst. 38, 121–137.
Crookston, R. K., Kurle, J. E., Copeland, P. J., Ford, J. H., and Lueschen, W. E. (1991).
Rotational cropping sequence affects yield of corn and soybean. Agron. J. 83, 108–113.
Dahlgren, R. A., Saigusa, M., and Ugolini, F. C. (2004). The nature, property, and
management of volcanic soils. Adv. Agron. 82, 113–182.
DallaCosta, L., DelleVedove, G., Gianquinto, G., Giovanardi, R., and Peressotti, A. (1997).
Yield, water use efficiency and nitrogen uptake in potato: Influence of drought stress.
Potato Res. 40, 19–34.
Davidson, E. A., and Janssens, I. A. (2006). Temperature sensitivity of soil carbon decom-
position and feedbacks to climate change. Nature 440, 165–173.
Davies, F. T. Jr., Calderon, C. M., and Huaman, Z. (2005a). Influence of arbuscular
mycorrhizae indigenous to Peru and a flavonoid on growth, yield, and leaf elemental
concentration of ‘Yungay’ potatoes. HortScience 40, 381–385.
Davies, F. Jr., Calderon, C., Huaman, Z., and Gomez, R. (2005b). Influence of a flavonoid
(formononetin) on mycorrhizal activity and potato crop productivity in the highlands of
Peru. Sci. Hortic. (Amsterdam) 106, 318–329.
de Koning, G. H. J., Van de Kop, P. J., and Fresco, L. (1997). Estimates of sub-national
nutrient balances as sustainability indicators for agro-ecosystems in Ecuador. Agric.
Ecosyst. Environ. 65, 127–139.
de la Torres, C., and Burga, M. (1986). Andenes y camellones en el Perú andino: Historia,
presente y futuro. Consejo Nacional de Cencia y Tecnologı́a, Lima.
Degenhardt, J., Hiltpold, I., Koellner, T. G., Frey, M., Gierl, A., Gershenzon, J.,
Hibbard, B. E., Ellersieck, M. R., and Turlings, T. C. J. (2009). Restoring a maize
root signal that attracts insect-killing nematodes to control a major pest. Proc. Natl. Acad.
Sci. USA 106, 13213–13218.
Dehn, M. (1995). An evaluation of soil conservation techniques in the Ecuadorian Andes.
Mt. Res. Dev. 15, 175–182.
Dercon, G., Deckers, J., Poesen, J., Govers, G., Sánchez, H., Ramı́rez, M., Vanegas, R.,
Tacuri, E., and Loaiza, G. (2006). Spatial variability in crop response under contour
hedgerow systems in the Andes region of Ecuador. Soil Till. Res. 86, 15–26.
Deugd, M., Roling, N., and Smaling, E. M. A. (1998). A new praxeology for integrated
nutrient management, facilitating innovation with and by farmers. Agric. Ecosyst. Environ.
71, 269–283.
172 Steven J. Fonte et al.

Devaux, A., Vallejos, J., Hijmans, R., and Ramos, J. (1997). Respuesta agronómica de dos
variedades de papa (spp. tuberosum y andigena) a diferentes niveles de fertilización mineral.
Revista Latinoamericana de la Papa 9(10), 123–139.
Dick, R. P., Sandor, J. A., and Eash, N. S. (1994). Soil enzyme activities after 1500 years of
terrace agriculture in the Colca Valley of Peru. Agric. Ecosyst. Environ. 50, 123–131.
Douds, D. D., Nagahashi, G., Reider, C., and Hepperly, P. R. (2007). Inoculation with
arbuscular mycorrhizal fungi increases the yield of potatoes in a high P soil. Biol. Agric.
Hort. 25, 67–78.
Drinkwater, L. E., and Snapp, S. S. (2007). Nutrients in agroecosystems: Rethinking the
management paradigm. Adv. Agron. 92, 163–186.
Erenstein, O. (2003). Smallholder conservation farming in the tropics and sub-tropics:
A guide to the development and dissemination of mulching with crop residues and
cover crops. Agric. Ecosyst. Environ. 100, 17–37.
Errebhi, M., Rosen, C. J., Lauer, F. L., Martin, M. W., and Bamberg, J. B. (1999).
Evaluation of tuber-bearing Solanum species for nitrogen use efficiency and biomass
partitioning. Am. J. Potato Res. 76, 143–151.
Espinosa, J. (1991). Efecto residual de fosforo en andisoles. Revista Facultad de Agronomia
(Maracay) 17, 39–47.
Evans, J., Gregory, A., Dobrowolski, N., Morris, S. G., O’Connor, G. E., and Wallace, C.
(1996). Nodulation of field-grown Pisum sativum and Vicia faba: Competitiveness of
inoculant strains of Rhizobium leguminosarum bv. viciae determined by an indirect, com-
petitive ELISA method. Soil Biol. Biochem. 28, 247–255.
Evans, T. A., Dawes, T. Z., Ward, P. R., and Lo, N. (2011). Ants and termites increase crop
yield in a dry climate. Nat. Commun. 2, 1–7.
Faccini, G., Garzon, S., Martı́nez, M., and Varela, A. (2007). Evaluation of the effect of a dual
inoculum of phosphate-solubilizing bacteria and Azotobacter chroococcum, in crops of creole
potato (“papa criolla”) “yema de huevo” variety (Solanum phureja). In “First International
Meeting on Microbial Phosphate Solubilization” (E. V.a.C. Rodrı́guez-Barrueco, Ed.),
pp. 301–308. Dordrecht, The Netherlands.
Felipe-Morales, C. (2002). Manejo agroecologico del suelo en sistemas andinos. Agroeco-
logia. Ediciones Cientificas Sudamericanas, Buenos Aires.
Fernandes, E. C. M., Motavalli, P. P., Castilla, C., and Mukurumbira, L. (1997). Management
control of soil organic matter dynamics in tropical land-use systems. Geoderma 79, 49–67.
Ferris, H., Bongers, T., and de Goede, R. G. M. (2001). A framework for soil food
web diagnostics: Extension of the nematode faunal analysis concept. Appl. Soil Ecol. 18,
13–29.
Ferrufino, A., and Sanchez, M. (2006). Efecto de la inoculación con micorrizas en el
crecimiento de plántulas de tembe para palmito (Bactris gasipaes Kunth) en fase de vivero.
Rev. Agric. 37, 1–10.
Flores Macias, A., Galvis Spinola, A., Hernandez Mendoza, T. M., De Leon Gonzalez, F.,
and Payan Zelaya, F. (2007). Effect of zeolite (clinoptilolite and mordenite) amended
andosols on soil chemical environment and growth of oat. Interciencia 32, 692–696.
Fonte, S. J., Barrios, E., and Six, J. (2010). Earthworms, soil fertility and aggregate-
associated soil organic matter dynamics in the Quesungual agroforestry system.
Geoderma 155, 320–328.
Fornara, D. A., and Tilman, D. (2008). Plant functional composition influences rates of soil
carbon and nitrogen accumulation. J. Ecol. 96, 314–322.
Freiberg, C., Fellay, R., Balroch, A., Broughton, W. J., Rosenthal, A., and Peret, X.
(1997). Molecular basis of symbiosis between Rhizobium and legumes. Nature 387,
394–401.
French, R. J., and Buirchell, B. J. (2005). Lupin: The largest grain legume crop in Western
Australia, its adaptation and improvement through plant breeding. Aust. J. Agr. Res. 56,
1169–1180.
Andean Soil Fertility 173

Fuentes, M., Govaerts, B., De Leon, F., Hidalgo, C., Dendooven, L., Sayre, K. D., and
Etchevers, J. (2009). Fourteen years of applying zero and conventional tillage, crop
rotation and residue management systems and its effect on physical and chemical soil
quality. Eur. J. Agron. 30, 228–237.
Gabriel, J. (2006). Plant Breeding and Biotechnology Capacity Survey: BOLIVIA. Funda-
ción Proinpa, Cochabamba.
Garcı́a, S. (2011). Evaluating the Biophysical Resource Management Strategies of the Agro-
ecosystems in Farm Communities of the Mantaro Valley, Central Andes of Peru. p. 276.
Bioscience Engineering. Katholieke Universiteit Leuven, Belgium.
Garcı́a, S., Rodrı́guez, J., Vera, J., and Schrevens, E. (2010). Effect of compost application on
soil chemical and biological properties under potato crop in the Mantaro Valley—Peru.
In “Proceedings of the International Soil Science Congress, Samsun, Turkey,” pp. 211–217.
Garrett, K. A., Dendy, S. P., Frank, E. E., Rouse, M. N., and Travers, S. E. (2006). Climate
change effects on plant disease: Genomes to ecosystems. Annu. Rev. Phytopathol. 44,
489–509.
Geerts, S., Raes, D., Garcia, M., Del Castillo, C., and Buytaert, W. (2006). Agro-climatic
suitability mapping for crop production in the Bolivian Altiplano: A case study for
quinoa. Agric. Forest Meteorol. 139, 399–412.
Gianoli, E., Ramos, I., Alfaro-Tapia, A., Valdéz, Y., Echegaray, E. R., and Yábar, E. (2006).
Benefits of a maize-bean-weed mixed cropping system in Urubamba Valley, Peruvian
Andes. Int. J. Pest Manag. 52, 283–289.
Giller, K. E., Rowe, E. C., de Ridder, N., and van Keulen, H. (2006). Resource use dynamics
and interactions in the tropics: Scaling up in space and time. Agric. Syst. 88, 8–27.
Glick, B. R., Cheng, Z., Czarny, J., and Duan, J. (2007). Promotion of plant growth by
ACC deaminase-producing soil bacteria. Eur. J. Plant Pathol. 119, 329–339.
Goland, C. (1993). Field scattering as agricultural risk management: A case study from Cuyo
Cuyo, Department of Puno, Peru. Mt. Res. Dev. 13, 317–338.
Goodman-Elgar, M. (2008). Evaluating soil resilience in long-term cultivation: A study of
pre-Columbian terraces from the Paca Valley, Peru. J. Archaeol. Sci. 35, 3072–3086.
Govaerts, B., Mezzalama, M., Sayre, K. D., Crossa, J., Lichter, K., Troch, V., Vanherck, K.,
Corte, P.d., and Deckers, J. (2008). Long-term consequences of tillage, residue manage-
ment, and crop rotation on selected soil micro-flora groups in the subtropical highlands.
Appl. Soil Ecol. 38, 197–210.
Hahn, H., McManus, M. T., Warnstorff, K., Monahan, B. J., Young, C. A., Davies, E.,
Tapper, B. A., and Scott, B. (2008). Neotyphodium fungal endophytes confer physio-
logical protection to perennial ryegrass (Lolium perenne L.) subjected to a water deficit.
Environ. Exp. Bot. 63, 183–199.
Harden, C. (1988). Mesoscale estimation of soil erosion in the Rio Ambato drainage,
Ecuadorian Sierra. Mt. Res. Dev. 8, 331–341.
Harden, C. (1993). Land use, soil erosion, and reservoir sedimentation in an Andean
drainage basin in Ecuador. Mt. Res. Dev. 13, 177–184.
Harden, C. (1996). Interrelationships between land abandonment and land degradation: A
case from the ecuadorian andes. Mt. Res. Dev. 16, 274–280.
Harris, P. J. C., Allison, M., Smith, G., Kindness, H. M., and Kelley, J. (2001). The Potential
Use of Waste-Stream Products for Soil Amelioration in Peri-Urban Interface Agricul-
tural Production Systems. CAB International, Wallingford.
Hayat, R., Ali, S., Amara, U., Khalid, R., and Ahmed, I. (2010). Soil beneficial bacteria and
their role in plant growth promotion: A review. Ann. Microbiol. 60, 579–598.
Hellin, J., Welchez, L. A., and Cherrett, I. (1999). The Quezungual system: An indigenous
agroforestry system from western Honduras. Agroforest. Syst. 46, 229–237.
Henry, A., Kleinman, P. J. A., and Lynch, J. P. (2009). Phosphorus runoff from a phosphorus
deficient soil under common bean (Phaseolus vulgaris L.) and soybean (Glycine max L.)
genotypes with contrasting root architecture. Plant Soil 317, 1–16.
174 Steven J. Fonte et al.

Hensen, I. (2002). Impacts of anthropogenic activity on the vegetation of Polylepis wood-

lands in the region of Cochabamba, Bolivia. Ecotropica 8, 183–203.
Herbas, J. E. (2000). Elaboracion de compost por los metodos indore y bocashi en la
comunidad de millu mayu, municipio de tiraque. Programa de Agronomı́a. Universidad
Mayor de San Simón, Cochabamba.
Herrera-Peraza, R. A., Hamel, C., Fernandez, F., Ferrer, R. L., and Furrazola, E. (2011).
Soil-strain compatibility: The key to effective use of arbuscular mycorrhizal inoculants?
Mycorrhiza 21, 183–193.
Hersh, N. (2000). Landrace Adaptation and Andean Farming Systems: Distribution and
Weevil (Mycrotrypes spp.) Infestation of Oca (Oxalis tuberosa) Tuber Varieties in Cusco.
Pennsylvania State University, Peru.
Hervé, D. (1994). Respuesta de los componentes de la fertilidad del suelo a la duración
del descanso. In “Dinámicas del descanso de la tierra en los Andes” (D. Hervé, Ed.),
pp. 155–169. IBTA—OSTROM, La Paz, Bolivia.
Hijmans, R. J., Cameron, S., Parra, J. L., Jones, P. G., and Jarvis, A. (2005). Very
high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25,
1965–1978.
Hocking, P. J., and Jeffery, S. (2004). Cluster-root production and organic anion exudation
in a group of old-world lupins and a new-world lupin. Plant Soil 258, 135–150.
Hofstede, R. (1995). The effects of grazing and burning on soil and plant nutrient concen-
trations in Colombian páramo grasslands. Plant Soil 173, 111–132.
Hofstede, R., Coppus, R., Mena Vásconez, P., Segarra, P., Wolf, J., and Sevink, J. (2002).
The conservation status of tussock grass paramo in Ecuador. Ecotropicos 15, 3–18.
Horton, D. (1983). Potato farming in the Andes: Some lessons from on-farm research in
Peru’s Mantaro Valley. Agric. Syst. 12, 171–184.
Ibarra, E. (2008). Efecto de especie y dosis de micorrizas en la producción de pimiento.
Pontificia Universidad Católica de Ecuador, Quito, Ecuador.
Inbar, M., and Llerena, C. A. (2000). Erosion processes in high mountain agricultural
terraces in Peru. Mt. Res. Dev. 20, 72–79.
Ippolito, J. A., Tarkalson, D. D., and Lehrsch, G. A. (2011). Zeolite soil application method
affects inorganic nitrogen, moisture, and corn growth. Soil Sci. 176, 136–142.
Iriarte, L., Franco, J., and Ortuño, N. (1999). Efecto de abonos orgánicos sobre las poblacione
de nematodos y la producción de la papa. Revista Latinoamericana de la Papa 11, 149–163.
Jaimes, V., and Sarmiento, L. (2002). Regeneración de la vegetación de paramo después de
un disturbio agrı́cola en la cordillera oriental de colombia. Ecotropicos 15, 61–74.
Jennings, P. R., and Cock, J. H. (1977). Center of origin of crops and their productivity.
Econ. Bot. 31, 51–54.
Jha, P., Biswas, A. K., Lakaria, B. L., and Rao, A. S. (2010). Biochar in agriculture—
Prospects and related implications. Curr. Sci. India 99, 1218–1225.
Johnson, N. C. (2010). Resource stoichiometry elucidates the structure and function of
arbuscular mycorrhizas across scales. New Phytol. 185, 631–647.
Jorquera, M. A., Hernandez, M. T., Rengel, Z., Marschner, P., and Mora, M.d.l.L. (2008).
Isolation of culturable phosphobacteria with both phytate-mineralization and phosphate-
solubilization activity from the rhizosphere of plants grown in a volcanic soil. Biol. Fertil.
Soils 44, 1025–1034.
Kane, K. H. (2011). Effects of endophyte infection on drought stress tolerance of Lolium
perenne accessions from the Mediterranean region. Environ. Exp. Bot. 71, 337–344.
Kanmegne, J., Smaling, E. M. A., Brussaard, L., Gansop-Kouomegne, A., and Boukong, A.
(2006). Nutrient flows in smallholder production systems in the humid forest zone of
southern Cameroon. Nutr. Cycl. Agroecosyst. 76, 233–248.
Andean Soil Fertility 175

Karak, T., and Bhattacharyya, P. (2011). Human urine as a source of alternative natural
fertilizer in agriculture: A flight of fancy or an achievable reality. Resour. Conserv. Recycl.
55, 400–408.
Karhu, K., Mattila, T., Bergström, I., and Regina, K. (2011). Biochar addition to agricultural
soil increased CH4 uptake and water holding capacity—Results from a short-term pilot
field study. Agric. Ecosyst. Environ. 140, 309–313.
Keneni, A., Assefa, F., and Prabu, P. C. (2010). Isolation of phosphate solubilizing bacteria
from the rhizosphere of faba bean of Ethiopia and their abilities on solubilizing insoluble
phosphates. J. Agric. Sci. Technol. 12, 79–89.
Kessler, C. A., and Stroosnijder, L. (2006). Land degradation assessment by farmers in
Bolivian mountain valleys. Land Degrad. Dev. 17, 235–248.
Khan, H. U. R., Link, W., Hocking, T. J., and Stoddard, F. L. (2007). Evaluation of
physiological traits for improving drought tolerance in faba bean (Vicia faba L.). Plant Soil
292, 205–217.
Khan, H. R., Paull, J. G., Siddique, K. H. M., and Stoddard, F. L. (2010). Faba bean
breeding for drought-affected environments: A physiological and agronomic perspective.
Field Crop. Res. 115, 279–286.
Kirkegaard, J. A., and Sarwar, M. (1998). Biofumigation potential of brassicas. I. Variation in
glucosinolate profiles of diverse field-grown brassicas. Plant Soil 201, 71–89.
Klein, A.-M., Vaissiere, B. E., Cane, J. H., Steffan-Dewenter, I., Cunningham, S. A.,
Kremen, C., and Tscharntke, T. (2007). Importance of pollinators in changing landscapes
for world crops. Proc. R. Soc. Lond. B. Biol. 274, 303–313.
Kochian, L. V., Piñeros, M. A., and Hoekenga, O. A. (2005). The physiology, genetics and
molecular biology of plant aluminum resistance and toxicity. Plant Soil 274, 175–195.
Kong, A. Y. Y., and Six, J. (2010). Tracing root vs. residue carbon into soils from
conventional and alternative cropping systems. Soil Sci. Soc. Am. J. 74, 1201–1210.
Kong, A. Y. Y., Six, J., Bryant, D. C., Denison, R. F., and van Kessel, C. (2005). The
relationship between carbon input, aggregation, and soil organic carbon stabilization in
sustainable cropping systems. Soil Sci. Soc. Am. J. 69, 1078–1085.
Kramer, S. B., Reganold, J. P., Glover, J. D., Bohannan, B. J. M., and Mooney, H. A.
(2006). Reduced nitrate leaching and enhanced denitrifier activity and efficiency in
organically fertilized soils. Proc. Natl. Acad. Sci. USA 103, 4522–4527.
Lagacherie, B., Bours, M., Giraud, J. J., and Sommer, G. (1983). Interaction between
Rhizobium lupini strains and species or cultivars of lupin (Lupinus albus, Lupinus luteus
and Lupinus mutabilis). Agronomie 3, 809–816.
Lange, F. R. (1994). Mejoramiento de la disponibilidad de fósforo por procesos biológicos de
la roca fosfórica. Facultad de Agronomı́a. Universidad Mayor San Simon, Cochabamba.
Lavelle, P., and Spain, A. V. (2001). Soil Ecology. Kluwer Academic Publishers, Dordretch,
Netherlands.
Lavelle, P., Blanchart, E., Martin, A., Martin, S., Spain, A., Toutain, F., Barois, I., and
Schaefer, R. (1993). A hierarchical model for the decomposition in terrestrial ecosystems:
Application to soils of the humid tropics. Biotropica 25, 130–150.
Lavelle, P., Bignell, D., Lepage, M., Wolters, V., Roger, P., Ineson, P., Heal, O. W., and
Dhillion, S. (1997). Soil function in a changing world: The role of invertebrate ecosystem
engineers. Eur. J. Soil Biol. 33, 159–193.
Lawler, J. J., Shafer, S. L., White, D., Kareiva, P., Maurer, E. P., Blaustein, A. R., and
Bartlein, P. J. (2009). Projected climate-induced faunal change in the Western Hemi-
sphere. Ecology 90, 588–597.
Leveau, J. H. J. (2007). The magic and menace of metagenomics: Prospects for the study of
plant growth-promoting rhizobacteria. Eur. J. Plant Pathol. 119, 279–300.
176 Steven J. Fonte et al.

Li, L., Li, S. M., Sun, J. H., Zhou, L. L., Bao, X. G., Zhang, H. G., and Zhang, F. S. (2007).
Diversity enhances agricultural productivity via rhizosphere phosphorus facilitation on
phosphorus-deficient soils. Proc. Natl. Acad. Sci. USA 104, 11192–11196.
Li, Y. Y., Yu, C. B., Cheng, X., Li, C. J., Sun, J. H., Zhang, F. S., Lambers, H., and Li, L.
(2009). Intercropping alleviates the inhibitory effect of N fertilization on nodulation and
symbiotic N-2 fixation of faba bean. Plant Soil 323, 295–308.
Liebman, M., and Dyck, E. (1993). Crop rotation and intercropping strategies for weed
management. Ecol. Appl. 3, 92–122.
Litterick, A. M., Wallace, P., Watson, C. A., and Wood, M. (2004). Compost extracts
in reducing pest and disease incidence and severity in sustainable temperate agricultural
and horticultural crop production—A review. Crit. Rev. Plant Sci. 23, 453–479.
Loon, L. C. (2007). Plant responses to plant growth-promoting rhizobacteria. Eur. J. Plant
Pathol. 119, 243–254.
Lorion, R. (2004). Rock Phosphate, Manure and Compost Use in Garlic and Potato Systems
in a High Intermontane Valley in Bolivia. Dept. of Crop and Soil Sciences, Washington
State University, Pullman, WA. p. 63.
Lorito, M., Woo, S. L., Harman, G. E., and Monte, E. (2010). Translational research on
Trichoderma: From ’omics to the field. Annu. Rev. Phytopathol. 48, 395–417.
Lynch, J. P. (2007). Roots of the second Green Revolution. Aust. J. Bot. 55, 493–512.
Machado, D., Sarmiento, L., and Gonzalez-Prieto, S. (2010). The use of organic substrates
with contrasting C/N ratio in the regulation of nitrogen use efficiency and losses in a
potato agroecosystem. Nutr. Cycl. Agroecosyst. 88, 411–427.
Mäder, P., Kaiser, F., Adholeya, A., Singh, R., Uppal, H. S., Sharma, A. K., Srivastava, R.,
Sahai, V., Aragno, M., and Wiemken, A. (2011). Inoculation of root microorganisms
for sustainable wheat–rice and wheat–black gram rotations in India. Soil Biol. Biochem. 43,
609–619.
Mahboubi, P., Gordon, A. M., Stoskopf, N., and Voroney, R. P. (1997). Agroforestry in the
Bolivian Altiplano: Evaluation of tree species and greenhouse growth of wheat on soils
treated with tree leaves. Agroforest. Syst. 37, 59–77.
Mamani, P., Botello, R., Condori, B., Moya, H., and Devaux, A. (2001). Efecto del tipo de
labranza con tracción animal en las caracterı́sticas fı́sicas del suelo, conservación de la
humedad y en el crecimiento y producción del cultivo de la papa. Revista Latinoamericana
de la Papa 12, 130–151.
Marichal, R. (2011). Impact de la déforestation sur les communautés de macrofaune et de
vers de terre en Amazonie—Relation avec les services écosystémiques. Université Pierre
et Marie Curie, Paris. p. 155.
Martinez-Viveros, O., Jorquera, M. A., Crowley, D. E., Gajardo, G., and Mora, M. L.
(2010). Mechanisms and practical considerations involved in plant growth promotion by
Rhizobacteria. J. Soil Sci. Plant Nutr. 10, 293–319.
Matson, P. A., Naylor, R., and Ortiz-Monasterio, I. (1998). Integration of environmental,
agronomic, and economic aspects of fertilizer management. Science 280, 112–115.
Mayak, S., Tirosh, T., and Glick, B. R. (2004). Plant growth-promoting bacteria that confer
resistance to water stress in tomatoes and peppers. Plant Sci. 166, 525–530.
Mayer, E. (1979). Land-Use in the Andes: Ecology and Agriculture in the Mantaro Valley of
Peru with Special Reference to Potatoes. International Potato Center, Lima, Peru.
Medrano Echalar, A. M., and Ortuño, N. (2007). Control de Damping off mediante la
aplicación e bioinsumos en almácigos de cebolla en el Valle Alto de Cochabamba,
Bolivia. Acta Nova 3, 660–679.
Mehari, A., Van Steenbergen, F., and Schultz, B. (2011). Modernization of spate irrigated
agriculture: A new approach. Irrig. Drain. 60, 163–173.
Meneses, R., Oller, V., and Waiijenberg, H. (2000). Inoculación y fertilización en el cultivo
de alfalfa en valles y alturas de Bolivia. In “Compendio de trabajos presentados por el
Andean Soil Fertility 177

proyecto rizobiologı́a (Cochabamba) en eventos y publicaciones de otras instituciones”

(R. Meneses, Ed.), pp. 66–67. Proyecto Rizobiologı́a Bolivia (CIAT-CIF-PNLG-
CIFP-UAW/DHV), Cochabamba, Bolivia.
Miklas, P. N., Kelly, J. D., Beebe, S. E., and Blair, M. W. (2006). Common bean breeding
for resistance against biotic and abiotic stresses: From classical to MAS breeding. Euphytica
147, 105–131.
Miller, D. R., Chen, S. Y., Porter, R. M., Johnson, G. A., Wyse, D. L., Stetina, S. R.,
Klossner, L. D., and Nelson, G. A. (2006). Rotation crop evaluation for management of
the soybean cyst nematode in Minnesota. Agron. J. 98, 569–578.
Misra, R. V., Roy, R. N., and Hiraoka, H. (2003). On-Farm Composting Methods. FAO,
Rome.
Mnasri, B., Tajini, F., Trabelsi, M., Aouani, M. E., and Mhamdi, R. (2007). Rhizobium
gallicum as an efficient symbiont for bean cultivation. Agron. Sustain. Dev. 27, 331–336.
Molinillo, M., and Monasterio, M. (2006). Vegetation and grazing patterns in andean
environments: A comparison of pastoral systems in punas and páramos. In “Land Use
Change and Mountain Biodiversity”. (E. M. Spehn, M. Liberman, and C. Körner, Eds.),
1–15. CRC Press, Boca Raton, FL.
Montgomery, D. R. (2007). Soil erosion and agricultural sustainability. Proc. Natl. Acad. Sci.
USA 104, 13268–13272.
Moore, J. C., Berlow, E. L., Coleman, D. C., de Ruiter, P. C., Dong, Q., Hastings, A.,
Johnson, N. C., McCann, K. S., Melville, K., Morin, P. J., Nadelhoffer, K.,
Rosemond, A. D., et al. (2004). Detritus, trophic dynamics and biodiversity. Ecol. Lett.
7, 584–600.
Morales, J., and Sarmiento, L. (2002). Dinámica de los macroinvertebrados edáficos y su
relación con la vegetación en una sucesión secundaria en el paramo venezolano. Ecotro-
picos 15, 99–110.
Moreno Diaz, P. (1988). Inoculación de micorrizas MVA en papa (Solanum tuberosum):
respuesta en el crecimiento y nutrición de plantas inoculadas en invernadero y en campo.
Revista Latinoamericana de la Papa 1, 84–103.
Murphy, D. V., Stockdale, E. A., Brookes, P. C., and Goulding, K. W. T. (2007). Impact of
microorganisms on chemical transformations in soil. In “Biological Fertility—A Key to
Sustainable Land Use in Agriculture” (L. K. Abbott and D. V. Murphy, Eds.), pp. 37–59.
Springer, Berlin.
Mutegi, J. K., Mugendi, D. N., Verchot, L. V., and Kung’u, J. B. (2008). Combining napier
grass with leguminous shrubs in contour hedgerows controls soil erosion without
competing with crops. Agroforest. Syst. 74, 37–49.
Nezomba, H., Tauro, T. P., Mtambanengwe, F., and Mapfumo, P. (2010). Indigenous
legume fallows (indifallows) as an alterative soil fertility resource in smallholder maize
cropping systems. Field Crop. Res. 115, 149–157.
Nieto-Cabrera, C., Francis, C., Caicedo, C., Gutiérrez, P. F., and Rivera, M. (1997).
Response of four Andean crops to rotation and fertilization. Mt. Res. Dev. 17, 273–282.
Noguera, D., Rondón, M., Laossi, K., Hoyos, V., Lavelle, P., Cruz de Carvalho, M. E., and
Barot, S. (2010). Contrasted effect of biochar and earthworms on rice growth and
resource allocation in different soils. Soil Biol. Biochem. 42, 1017–1027.
Noguera, D., Laossi, K. R., Lavelle, P., Cruz de Carvalho, M. H., Asakawa, N., Botero, C.,
and Barot, S. (2011). Amplifying the benefits of agroecology by using the right cultivars.
Ecol. Appl. 21, 2349–2356.
Novak, J. M., Busscher, W. J., Laird, D. L., Ahmedna, M., Watts, D. W., and
Niandou, M. A. S. (2009). Impact of biochar amendment on fertility of a southeastern
coastal plain soil. Soil Sci. 174, 105–112.
NRC. (1989). Lost Crops of the Incas: Little-Known Plants of the Andes with Promise for
Worldwide Cultivation. National Academy Press, Washington, DC.
178 Steven J. Fonte et al.

Nustez, C. E., and Acevedo, J. C. (2005). Evaluating using Penicillium janthinellum Biourge
on the efficiency of phosphoric fertilisation of potato crops (Solanum tuberosum L. var.
Diacol Capiro). Agron. Colomb. 23, 290–298.
Nziguheba, G., Palm, C. A., Buresg, R. J., and Smithson, P. C. (1998). Soil phosphorus fractions
and adsorption as a affected by organic and inorganic sources. Plant Soil 198, 159–168.
Oberson, A., Friesen, D. K., Tiessen, H., Morel, C., and Stahel, W. (1999). Phosphorus
status and cycling in native savanna and improved pastures on an acid low-P Colombian
Oxisol. Nutr. Cycl. Agroecosyst. 55, 77–88.
Oberson, A., Bünemann, E. K., Friesen, D. K., Rao, I. M., Smithson, P. C., Turner, B. L.,
and Frossard, E. (2006). Improving phosphorus fertility in tropical soils through
biological interventions. “Biological Approaches to Sustainable Soil Systems” (Uphoff
et al., Eds.), pp. 531–546. CRC Press, Boca Raton.
Oliveira, C. A., Alves, V. M. C., Marriel, I. E., Gomes, E. A., Scotti, M. R.,
Carneiro, N. P., Guimaraes, C. T., Schaffert, R. E., and Sa, N. M. H. (2009). Phosphate
solubilizing microorganisms isolated from rhizosphere of maize cultivated in an oxisol of
the Brazilian Cerrado Biome. Soil Biol. Biochem. 41, 1782–1787.
Oorts, K., Vanlauwe, B., and Merckx, R. (2003). Cation exchange capacities of soil organic
matter fractions in a ferric lixisol with different organic matter inputs. Agric. Ecosyst.
Environ. 100, 161–171.
Orlove, B., and Godoy, R. (1986). Sectoral fallowing systems in the central Andes.
J. Enthnobiology 6, 169–204.
Orna, A. R. (2009). Evaluación del efecto de la aplicación de micorrizas en la producción de
tomate riñon (Solanum lycopersicon) bajo invernadero. Recursos Naturales. ESPOCH,
Riobamba, Ecuador.
Orsag, V. (2009). Degradación de suelos en el Altiplano Boliviano. Análisis—Instituto
Boliviano de Economı´a y Polı´tica Agraria 1, 27–30.
Ortuño, N., Navia, O., Medrano, A., Rojar, K., and Torrico, L. (2010). Desarrollo de
bioinsumos: Un aporte a la soberanı́a alimentaria de Bolivia. Rev. Agric. 47, 30–38.
Ortuño, N. (2010). Desarrollo de bioinsumos para la producción sosteniblle de hortalizas con
pequeños agricultores para una soberanı́a alimentariia en los Andes. CIAT, Colombia.
Osman, A. (1999). Soil Fertility Management in Cajamarca, Peru. Centro Internacional de la
Papa (CIP), Lima.
Osorio-Vega, N. W. (2007). Review of beneficial effects of rhizosphere bacteria on soil
nutrient availability and plant nutrient uptake. Revista de la Facultad Nacional de Agronomı´a
Medellin 60, 3621–3643.
Oswald, A. (2010). Soil Fertility Management and Crop Productivity of Potato Based
Cropping Systems of the Andean Highlands of Peru. Working Paper. Centro Interna-
cional de la Papa, pp. 1–29.
Oswald, A., Haan, S., Sanchez, J., and Ccanto, R. (2009). The complexity of simple tillage
systems. J. Agric. Sci. 147, 399–410.
Oswald, A., Calvo Velez, P., Zúñiga Dávila, D., and Arcos Pineda, J. (2010). Evaluating soil
rhizobacteria for their ability to enhance plant growth and tuber yield in potato. Ann.
Appl. Biol. 157, 259–271.
Otero, J., and Onaindiam, M. (2009). Landscape structure and live fences in Andes Colombian
agroecosystems: Upper basin of the Cane-Iguaque River. Int. J. Trop. Biol. 57, 1183–1192.
Oyarzun, P. J., Gerlagh, M., and Zadoks, J. C. (1998). Factors associated with soil receptivity
to some fungal root rot pathogens of peas. Appl. Soil Ecol. 10, 151–169.
Pacheco, V., Zelada, A., and Navarro, C. (1992). Recuperación de tierras en el Proyecto
Norte Chuquisaca. Proyecto Norte Chuquisaca, Sucre, Bolivia.
Palm, C. A., Gachengo, C. N., Delve, R. J., Cadisch, G., and Giller, K. E. (2001). Organic
inputs for soil fertility management in tropical agroecosystems: Application of an organic
resource database. Agric. Ecosyst. Environ. 83, 27–42.
Andean Soil Fertility 179

Palta, J. A., Turner, N. C., French, R. J., and Buirchell, B. J. (2007). Physiological responses
of lupin genotypes to terminal drought in a Mediterranean-type environment. Ann. Appl.
Biol. 150, 269–279.
Pansak, W., Dercon, G., Hilger, T., Kongkaew, T., and Cadisch, G. (2007). 13C isotopic
discrimination: A starting point for new insights in competition for nitrogen and water
under contour hedgerow systems in tropical mountainous regions. Plant Soil 298, 175–189.
Pansu, M., Sarmiento, L., Metselaar, K., and Bottner, P. (2007). Modelling the transforma-
tions and sequestration of soil organic matter in two contrasting ecosystems of the Andes.
Eur. J. Soil Sci. 58, 775–785.
Paredes, M. (2010). Peasants, Potatoes and Pesticides: Heterogeneity in the Context of
Agricultural Modernization in the Highland Andes of Ecuador. School of Social
Sciences. Wageningen Univiersity, Wageningen. p. 322.
Parsa, S. (2010). Native herbivore becomes key pest after dismantlement of a traditional
farming system. Am. Entomol. 56, 242–251.
Parsa, S., Ccanto, R., and Rosenheim, J. A. (2011). Resource concentration dilutes a key
pest in indigenous potato agriculture. Ecol. Appl. 21, 539–546.
Pearse, S. J., Veneklaas, E. J., Cawthray, G. R., Bolland, M. D. A., and Lambers, H. (2006).
Carboxylate release of wheat, canola and 11 grain legume species as affected by phospho-
rus status. Plant Soil 288, 127–139.
Perez, C., Nicklin, C., D’angles, O., Vanek, S. J., Sherwood, S., Halloy, S., Garrett, K., and
Forbes, G. (2010). Climate change in the high Andes: Implications and adaptation
strategies for small-scale farmers. Int. J. Environ. Cult. Econ. Soc. Sustain. 6, 71–88.
Pestalozzi, H. (2000). Sectoral fallow systems and the management of soil fertility: The rationality
of indigenous knowledge in the high Andes of Bolivia. Mt. Res. Dev. 20, 64–71.
Pimentel, D. (2006). Soil erosion: A food and environmental threat. Environ. Dev. Sustain. 8,
119–137.
Podwojewski, P., and Germain, N. (2005). Short-term effects of management on the soil
structure in a deep tilled hardened volcanic-ash soil (Cangahua) in Ecuador. Eur. J. Soil
Sci. 39, 39–51.
Posthumus, H., and De Graaff, J. (2005). Cost-benefit analysis of bench terraces. A case study
in Peru. Land Degrad. Dev. 16, 1–11.
Posthumus, H., Gardebroek, C., and Ruben, R. (2010). From participation to adoption:
Comparing the effectiveness of soil conservation programs in the Peruvian Andes. Land
Econ. 86, 645–667.
Postma-Blaauw, M., De Goede, R. G. M., Bloem, J., Faber, J. H., and Brussard, L. (2010).
Soil biota community structure and abundance under agricultural intensification and
extensification. Ecology 91, 460–473.
Poulenard, J., Podwojewski, P., Janeau, J.-L., and Collinet, J. (2001). Runoff and soil
erosion under rainfall simulation of Andisols from the Ecuadorian Páramo: Effect of
tillage and burning. Catena 45, 185–207.
Poveda, K., Gómez, M. I., and Martı́nez, E. (2008). Diversification practices: Their effect on
pest regulation and production. Rev. Colomb. Entomol. 34, 131–144.
Powlson, D. S., Gregory, P. J., Whalley, W. R., Quinton, J. N., Hopkins, D. W.,
Whitmore, A. P., Hirsch, P. R., and Goulding, K. W. T. (2011). Soil management in
relation to sustainable agriculture and ecosystem services. Food Policy 36, S72–S87.
Puget, P., and Drinkwater, L. E. (2001). Short-term dynamics of root- and shoot-derived
carbon from a leguminous green manure. Soil Sci. Soc. Am. J. 65, 771–779.
Quilty, J. R., and Cattle, S. R. (2011). Use and understanding of organic amendments in
Australian agriculture: A review. Aust. J. Soil Res. 49, 1–26.
Quintero, M. (2009). Effects of Conservation Tillage in Soil Carbon Sequestration and Net
Revenues of Potato-Based Rotations in the Colombian Andes. University of Florida,
Gainesville, FL. p. 103.
180 Steven J. Fonte et al.

Quiroz, R., León-Velarde, C., Valdivia, R., Zorogastúa, F. P., Baigorria, G., Barreda, C.,
Reinoso, J., Holle, M., and Li Pun, H. (2003). Making a difference to Andean livelihoods
through an integrated research approach. In “Research Towards Integrated Natural
Resources Management—Examples of Research Problems, Approaches and Partnerships
in Action in the CGIAR” (R. R. Harwood and A. H. Kassam, Eds.), pp. 111–122. FAO,
Rome.
Ramaekers, L., Remans, R., Rao, I. M., Blair, M. W., and Vanderleyden, J. (2010).
Strategies for improving phosphorus acquisition efficiency of crop plants. Field Crop.
Res. 117, 169–176.
Ramesh, K., Biswas, A. K., Somasundaram, J., and Rao, A. S. (2010). Nanoporous zeolites
in farming: Current status and issues ahead. Curr. Sci. India 99, 760–764.
Reed, S. C., Seastedt, T. R., Mann, C. M., Suding, K. N., Townsend, A. R., and
Cherwin, K. L. (2007). Phosphorus fertilization stimulates nitrogen fixation and increases
inorganic nitrogen concentrations in a restored prairie. Appl. Soil Ecol. 36, 238–242.
Reynel, C., and Felipe-Morales, C. (1990). Agroforesteria tradicional en los Andes del Perú:
un inventario de tecnologı́as y especies para la integración de la vegetación leñosa a la
agricultura. Proyecto FAO/Holanda/INFOR, Peru.
Ricketts, T. H., Regetz, J., Steffan-Dewenter, I., Cunningham, S. A., Kremen, C.,
Bogdanski, A., Gemmill-Herren, B., Greenleaf, S. S., Klein, A. M., Mayfield, M. M.,
Morandin, L. A., Ochieng, A., et al. (2008). Landscape effects on crop pollination
services: Are there general patterns? Ecol. Lett. 11, 499–515.
Righi, G., and van der Hammen, T. (1996). Distribución de especies de lombrices en las dos
vertientes de la cordillera central (transecto parque Los Nevados, Colombia). In “Estudios
de ecosistemas tropandinos 4” (T. Van der Hammen and A. G. dos Santos, Eds.),
pp. 475–483. Cramer (Brontraeger), Berlin-Stuttgart.
Robertson, G. P., and Swinton, S. M. (2005). Reconciling agricultural productivity
and environmental integrity: A grand challenge for agriculture. Front. Ecol. Environ. 3,
38–46.
Rodriguez, K. R., and Ortuño, N. (2007). Evaluación de micorrizas arbusculares en
interacción con abonos orgánicos como coadyuvantes del crecimiento en la producción
hortı́cola del Valle Alto de Cochabamba, Bolivia. Acta Nova 3, 697–720.
Rodriguez, E. A., Bolaños, B. M. M., and Menjivar Flores, J. C. (2010). Efecto de la
fertilización en la nutrición y rendimiento de ajı́ (Capsicum spp.) en el Valle del Cauca,
Colombia. Acta Agron. 59, 55–64.
Rodrı́guez, H., Fraga, R., Gonzalez, T., and Bashan, Y. (2006). Genetics of phosphate
solubilization and its potential applications for improving plant growth-promoting bac-
teria. Plant Soil 287, 15–21.
Romero-León, C. C. (2005). A Multi-Scale Approach for Erosion Assessment in the Andes.
Wageningen University, Wageningen, Netherlands. p. 162.
Rosset, P. M., and Altieri, M. A. (1997). Agroecology versus input substitution: A funda-
mental contradiction of sustainable agriculture. Soc. Natur. Resour. 10, 283–295.
Sanchez-Moreno, S., and Ferris, H. (2007). Suppressive service of the soil food web: Effects
of environmental management. Agric. Ecosyst. Environ. 119, 75–87.
Sandor, J., and Eash, N. (1995). Ancient agricultural soils in the Andes of southern Peru. Soil
Sci. Soc. Am. J. 59, 170–179.
Sandor, J., and Furbee, L. (1996). Indigenous knowledge and classification of soils in the
Andes of southern Peru. Soil Sci. Soc. Am. J. 60, 1502–1512.
Sarmiento, L. (2000). Water balance and soil loss under long fallow agriculture in the
Venezuelan Andes. Mt. Res. Dev. 20, 1–9.
Sarmiento, L., and Bottner, P. (2002). Carbon and nitrogen dynamics in two soils with
different fallow times in the high tropical Andes: Indications for fertility restoration. Appl.
Soil Ecol. 19, 78–79.
Andean Soil Fertility 181

Sarmiento, L., Monasterio, M., and Montilla, M. (1993). Ecological bases, sustainability, and
current trends in traditional agriculture in the Venezuelan high Andes. Mt. Res. Dev. 13,
167–176.
Sarmiento, L., Acea, M., Barrios, E., Bowen, W., Herrera, R., Llambı́, L., Ortuño, N.,
Sivila, R., and Varela, A. (2001). Un marco conceptual y metodológico para estudios de
fertilidad del suelo en los Andes tropicales. In “Memorias del IV Simposio Internacional
de Desarrollo Sustentable en los Andes: La estrategia para el siglo XXI”.
Sau, F., and Minguez, M. I. (2000). Adaptation of indeterminate faba beans to weather and
management under a Mediterranean climate. Field Crop. Res. 66, 81–99.
Schuster, A., and Schmoll, M. (2010). Biology and biotechnology of Trichoderma. Appl.
Microbiol. Biotnechnol. 87, 787–799.
Sciarascia Mugnozza, G. C., Avila, G., Claure, T., Rios, R., Pierola, L., and Crespo, M.
(1987). Investigaciones sobre el mejoramiento genético y cultural de trigo duro, girasol,
maı́z, frijol, lupino y haba en Bolivia. Instituto Italo-Latinoamericano Fundación
Pro-Bolivia, Centro de Investigaciones Fitoecogenéticos Pairumani, Rome.
Seneviratne, G., Jayasekara, A. P. D. A., De Silva, M. S. D. L., and Abeysekera, U. P. (2011).
Developed microbial biofilms can restore deteriorated conventional agricultural soils. Soil
Biol. Biochem. 43, 1059–1062.
Setimela, P., Chitalu, Z., Jonazi, J., Mambo, A., Hodson, D., and Banziger, M. (2005).
Environmental classification of maize-testing sites in the SADC region and its implica-
tion for collaborative maize breeding strategies in the subcontinent. Euphytica 145,
123–132.
Sharma, S. N. (2003). Effect of phosphate-solubilizing bacteria on efficiency of Mussoorie
rockphosphate in rice (Oryza sativa)-wheat (Triticum aestivum) cropping system. Indian J.
Agric. Sci. 73, 478–481.
Sherwood, S. G., Monar, C., and Suquillo, J. (1999). Wachu rozado: Vestigio del pasado,
oportunidad para el futuro. Centro Internacional de la Papa (CIP), Quito.
Siddique, K. H. M., Regan, K. L., Tennant, D., and Thomson, B. D. (2001). Water use and
water use efficiency of cool season grain legumes in low rainfall Mediterranean-type
environments. Eur. J. Agron. 15, 267–280.
Sims, B. G., and Rodriguez, F. (2001). Forage production and erosion control as a comple-
ment to hillside weed management. In “International Workshop in Integrated Manage-
ment for Sustainable Agriculture, Forestry and Fisheries” Centro Internacional de
Agricultura Tropical (CIAT) Cali, Colombia.
Sims, B., Rodrı́guez, F., Eid, M., and Espinoza, T. (1999). Biophysical aspects of vegetative
soil and water conservation pratices in the inter-Andean valles of Bolivia. Mt. Res. Dev.
19, 282–291.
Singhai, P. K., Sarma, B. K., and Srivastava, J. S. (2011). Biological management of
common scab of potato through Pseudomonas species and vermicompost. Biol. Control
57, 150–157.
Sivila de Cary, R., and Hervé, D. (1994). El estado microbiologico del suelo, indicador de
una restauracion de la fertilidad. In “Dinámicas del descanso de la tierra en los Andes”
(D. Hervé, Ed.), pp. 185–197. IBTA—OSTROM, La Paz, Bolivia.
Sivila, R., and Herve, D. (1999). Análisis de la microbiota en suelos cultivados del altiplano
central. In “Primer Congreso Boliviano de la Ciencia del Suelo Sociedad Boliviana de la
Ciencia del Suelo, La Paz” pp. 5–14.
Six, J., Conant, R. T., Paul, E. A., and Paustian, K. (2002). Stabilization mechanisms of soil
organic matter: Implications for C-saturation of soils. Plant Soil 241, 155–176.
Smaling, E. M. A., and Fresco, L. O. (1993). A decision-support model for monitoring
nutrient balances under agricultural land-use (NUTMON). Geoderma 60, 235–256.
Smaling, E. M. A., Stoorvegel, J. J., and Windmeijer, P. N. (1993). Calculating soil nutrient
balances in Africa at different scales: II District Scale. Fertil. Res. 35, 237–250.
182 Steven J. Fonte et al.

Smith, R. G., Gross, K. L., and Robertson, G. P. (2008). Effects of crop diversity on
agroecosystem function: Crop yield response. Ecosystems 11, 355–366.
Smukler, S. M., Sánchez-Moreno, S., Fonte, S. J., Ferris, H., Klonsky, K., O’Geen, A. T.,
Scow, K. M., Steenwerth, K. L., and Jackson, L. E. (2010). Biodiversity and multiple
ecosystem functions in an organic farmscape. Agric. Ecosyst. Environ. 139, 80–97.
Snapp, S. S., and Silim, S. N. (2002). Farmer preferences and legume intensification for low
nutrient environments. Plant Soil 245, 181–192.
Snapp, S. S., Mafongoya, P. L., and Waddington, S. (1998). Organic matter technologies for
integrated nutrient management in smallholder cropping systems of southern Africa.
Agric. Ecosyst. Environ. 71, 185–200.
Snapp, S., Swinton, S. M., Labarta, R., Mutch, D., Black, J. R., Leep, R., Nyiraneza, J., and
O’Neil, K. (2005). Evaluating cover crops for benefits, costs and performance within
cropping system niches. Agron. J. 97, 322–332.
Snoek, C., Vanderleyden, J., and Beebe, S. (2003). Strategies for genetic improvement
of Common Bean and Rhizobia: Towards efficient interactions. Plant Breed. Rev. 23,
21–72.
Stadel, C. (1991). Environmental stress and sustainable development in the tropical andes.
Mt. Res. Dev. 11, 213–223.
Stanish, C. (2007). Agricultural intensification in the Titicaca basin. In “Seeking a Richer
Harvest” (T. Thurston and C. Fisher, Eds.), pp. 125–139. Springer, New York.
Swinton, S., and Quiroz, R. (2003). Poverty and the deterioration of natural soil capital in
the Peruvian Altiplano. Environ. Dev. Sustain. 5, 477–490.
Takeda, M., and Knight, J. D. (2006). Enhanced solubilization of rock phosphate by
Penicillium bilaiae in pH-buffered solution culture. Can. J. Microbiol. 52, 1121–1129.
Tapia, M. (1994). Rotación de cultivos y su manejo en los Andes del Perú. In “Dinámicas del
descanso de la tierra en los Andes” (D. Hervé, Ed.), pp. 37–53. IBTA—OSTROM, La
Paz, Bolivia.
Tenge, A. J., De Graaff, J., and Hella, J. P. (2004). Social and economic factors affecting the
adoption of soil and water conservation in west Usambara highlands, Tanzania. Land
Degrad. Dev. 15, 99–114.
Terrazas, F., Suarez, G., Gardner, G., Thiele, G., Devaux, A., and Walker, T. (1998). Diagnosing
Potato Productivity in Farmers Fields in Bolivia. International Potato Center (CIP), Lima.
Thamer, S., Schädler, M., Bonte, D., and Ballhorn, D. J. (2011). Dual benefit from a
belowground symbiosis: Nitrogen fixing rhizobia promote growth and defense against
a specialist herbivore in a cyanogenic plant. Plant Soil 341, 209–219.
Thibeault, J., Seth, A., and Garcia, M. (2010). Changing climate in the Bolivian Altiplano:
CMIP3 projections for temperature and precipitation extremes. J. Geophys. Res. 115, 1–18.
Thies, J., and Grossman, J. (2006). The soil habitat and soil ecology. In “Biological
Approaches to Sustainable Soil Systems” (N. Uphoff, Ed.), pp. 59–78. CRC Press,
Boca Raton.
Thies, J., and Rillig, M. (2009). Characteristics of Biochar: Biological Properties. In “Bio-
char for Environmental Management”, (J. Lehmann and S. Joseph, Eds), pp. 85–105.
Earthscan, London.
Tisdall, J. M., and Oades, J. M. (1982). Organic matter and water-stable aggregates in soils.
J. Soil Sci. 62, 141–163.
Tittonell, P., Vanlauwe, B., Leffelaar, P. A., Shepherd, K. D., and Giller, K. E. (2005).
Exploring diversity in soil fertility management of smallholder farms in western Kenya: II.
Within-farm variability in resource allocation, nutrient flows and soil fertility status. Agric.
Ecosyst. Environ. 110, 166–184.
Tonneijck, F., and Jongmans, G. (2008). The influence of bioturbation on the vertical
distribution of soil organic matter in volcanic ash soils: A case study in northern Ecuador.
Eur. J. Soil Sci. 59, 1063–1075.
Andean Soil Fertility 183

Tonneijck, F., Jansen, B., Niero, K. G. J., Verstraten, J. M., Sevink, J., and De Lange, L.
(2010). Towards understanding of carbon stocks and stabilization in volcanic ash soils in
natural Andean ecosystems of northern Ecuador. Eur. J. Soil Sci. 61, 392–405.
Tourneux, C., Devaux, A., Camacho, M. R., Mamani, P., and Ledent, J. F. (2003a). Effect
of water shortage on six potato genotypes in the highlands of Bolivia (II): Water relations,
physiological parameters. Agronomie 23, 181–190.
Tourneux, C., Devaux, A., Camacho, M. R., Mamani, P., and Ledent, J. F. (2003b). Effects
of water shortage on six potato genotypes in the highlands of Bolivia (I): Morphological
parameters, growth and yield. Agronomie 23, 169–179.
Urgiles, N., Loján, P., Aguirre, N., Blaschke, H., Günter, S., Stimm, B., and Kottke, I.
(2009). Application of mycorrhizal roots improves growth of tropical tree seedlings in the
nursery: A step towards reforestation with native species in the Andes of Ecuador. New
Forest. 38, 229–239.
Valdivia, C., and Quiroz, R. (2003). Coping and adapting to increased climate variability in
the Andes. In “Annual Meeting, Presentation at the American Agricultural Economics
Association”.
Valdivia, C., Seth, A., Gilles, J., Garcia, M., Jiménez, E., Cusicanqui, J., Navia, F., and
Yucra, E. (2010). Adapting to climate change in Andean ecosystems: Landscapes, capitals,
and perceptions shaping rural livelihood strategies and linking knowledge systems. Ann.
Assoc. Am. Geogr. 100, 818–834.
Valente, J. F., and Oliver, R. (1993). Fertisuelos: Evaluacion de la fertilidad de los suelos del
antiplano, valle central y los llanos de Bolivia. FAO, Rome.
van de Kop, P. (1996). Regional Scale Nutrient Balances for Agro-Ecosystems in Ecuador.
Department of Agronomy. Wageningen Agricultural University, Wageningen. p. 43.
van Groenigen, K. J., Six, J., Hungate, B. A., de Graaff, M., van Breemen, N., and van
kessel, C. (2006). Elemental interactions limit soil carbon storage. Proc. Natl. Acad. Sci.
USA 103, 6571–6574.
Vanek, S. (2010). Legume-Phosphorus Synergies in Mountain Agroecosystems: Field
Nutrient Balances, Soil Fertility Gradients, and Effects on Legume Attributes and Nutri-
ent Cycling in the Bolivian Andes. Cornell University, Ithaca, NY.
Vanlauwe, B., Wendt, J., and Diels, J. (2001). Combined application of organic matter and
fertilizer. In “Sustaining Soil Fertility in West-Africa” (G. Tian, Ed.), pp. 247–279. Soil
Science Society of America, Madison, WI.
Vanlauwe, B., Tittonell, P., and Mukalama, J. (2006). Within-farm soil fertility gradients
affect response of maize to fertiliser application in western Kenya. Nutr. Cycl. Agroecosyst.
76, 171–182.
Vargas, C. A. (2009). Sistema de innovación agropecuaria y forestal. Revista de Agricultura
(Cochabamba, Bolivia) 45, 2–9.
Velasquez, E., Lavelle, P., and Andrade, M. (2007). GISQ, a multifunctional indicator of soil
quality. Soil Biol. Biochem. 39, 3066–3080.
Verma, M., Brar, S. K., Tyagi, R. D., Surampalli, R. Y., and Valero, J. R. (2007). Antagonistic
fungi, Trichoderma spp.: Panoply of biological control. Biochem. Eng. J. 37, 1–20.
Vinale, F., Krishnapillai, S., Ghisalberti, E. L., Marra, R., Woo, S. L., and Lorito, M. (2008).
Trichoderma-plant-pathogen interactions. Soil Biol. Biochem. 40, 1–10.
Vitousek, P. M., Naylor, R., Crews, T., David, M. B., Drinkwater, L. E., Holland, E.,
Johnes, P. J., Katzenberger, J., Martinelli, L. A., Matson, P. A., Nziguheba, G.,
Ojima, D., et al. (2009). Nutrient imbalances in agricultural development. Science 324,
1519–1520.
Wall, P. (1999). Experiences with crop residue cover and direct seeding in the Bolivian
highlands. Mt. Res. Dev. 19, 313–317.
Wang, K. H., Sipes, B. S., and Schmitt, D. P. (2002). Crotalaria as a covercrop for nematode
management: Review. Nematrópica 32, 35–57.
184 Steven J. Fonte et al.

Warnock, D. D., Lehmann, J., Kuyper, T. W., and Rillig, M. C. (2007). Mycorrhizal
responses to biochar in soil—Concepts and mechanisms. Plant Soil 300, 9–20.
Wheeler, T. R., Keatinger, H., Ellis, R. H., and Summerfield, R. J. (1999). Selecting
legume cover crops for hillside environments in Bolivia. Mt. Res. Dev. 19, 318–324.
Wiegers, E. S., Hijmans, R. J., Herve, D., and Fresco, L. O. (1999). Land use intensification
and disintensification in the Upper Canete valley, Peru. Hum. Ecol. 27, 319–357.
Winklerprins, A. (1999). Local soil knowledge: A tool for sustainable land management. Soc.
Nat. Resour. 12, 151–161.
Winters, P., Espinosa, P., and Crissman, P. (1998). Manejo de los Recursos en los Andes
Ecuatorianos: Revisión de Literatura y Evaluación del Proyecto Manejo del Uso Sos-
tenible de Tierras Andinas (PROMUSTA) de CARE. International Potato Center
(CIP), Quito, Ecuador.
Woomer, P., and Swift, M. J. (1994). The Biological Management of Tropical Soil Fertility.
John Wiley, Chichester, UK.
Wu, L., Vomocil, J. A., and Childs, S. W. (1990). Pore-size, particle-size, aggregate size, and
water-retention. Soil Sci. Soc. Am. J. 54, 952–956.
Young, A. (1997). Agroforestry for Soil Management. CAB International, Oxon, UK.
Zebarth, B. J., Tarn, T. R., de Jong, H., and Murphy, A. (2008). Nitrogen use efficiency
characteristics of andigena and diploid potato selections. Am. J. Potato Res. 85, 210–218.
Zehetner, F., and Miller, W. P. (2006). Soil variations along a climatic gradient in an Andean
agro-ecosystem. Geoderma 137, 126–134.
Zhu, J. M., Kaeppler, S. M., and Lynch, J. P. (2005). Topsoil foraging and phosphorus
acquisition efficiency in maize (Zea mays). Funct. Plant Biol. 32, 749–762.
Zimmerer, K. (1993). Soil erosion and labor shortages in the Andes with special reference to
Bolivia, 1953-91: Implications for “conservation-with-development. World Dev. 21,
1659–1675.
Zimmerer, K. S. (2011). The landscape technology of spate irrigation amid development
changes: Assembling the links to resources, livelihoods, and agrobiodiversity-food in the
Bolivian Andes. Glob. Environ. Chang. 21, 917–934.
Zurita, G. L. (1997). Composición taxonómica y abundancia poblacional de Lombrices en
sistemas de Monocultivo y Rotación de cultivos en Suka Kollus. Universidad Mayor de
San Andrés, La Paz, Bolivia.
 C H A P T E R F I V E

Rethinking Internal Phosphorus

Utilization Efficiency: A New
Approach Is Needed to Improve
PUE in Grain Crops
Terry J. Rose* and Matthias Wissuwa†

Contents
1. Introduction 186
2. Defining PUE: Terms, Units, and Assumptions 189
2.1. Criteria with agronomic implications 189
2.2. Criteria with physiological implications 192
2.3. Defining the utilization of P as “efficiency” 195
3. Quantifying PUE of Crop Genotypes Using Criteria with
Physiological Implications 196
3.1. Screening for vegetative PUE 197
3.2. Screening for grain PUE 200
4. P-Stress Levels in Screening Studies and the Utility of PUE in Low,
Medium, and High P Input Systems 201
4.1. P-deficient crops suffer a range of stress levels 201
4.2. What are the likely outcomes of improved PUE in
P-deficient plants? 202
5. Mechanisms and Physiology of PUE 205
5.1. Remobilization and scavenging of P 205
5.2. Alternative glycolytic pathways and mitochondrial electron
transport pathways 206
5.3. Exploiting P-deficiency stress response mechanisms 206
6. Conclusions and Future Prospective 208
6.1. PAE versus PUE—Does one offer better chances of success? 208
6.2. Screening methods, targets, and possible results 208
6.3. Marker-assisted selection—A paradigm shift in breeding
suited for PUE 209
6.4. Remaining questions 210
References 211

* Southern Cross Plant Science, Southern Cross University, Lismore, NSW, Australia
{
Japan International Research Center for Agricultural Sciences (JIRCAS), Crop Production and Environment
Division, Ohwashi, Tsukuba, Ibaraki, Japan

Advances in Agronomy, Volume 116 # 2012 Elsevier Inc.

ISSN 0065-2113, DOI: 10.1016/B978-0-12-394277-7.00005-1 All rights reserved.
185
186 Terry J. Rose and Matthias Wissuwa

Abstract
Grain crops are a key driver of the current global phosphorus (P) cycle through
their continued demand for P fertilizer, and the subsequent removal of P from
fields in the harvested grain. Breeding crops that can yield well with fewer P
inputs (i.e., P-efficient crops) may reduce the impact of grain crops of the P cycle,
but to date breeding P-efficient cultivars has focused on enhancing P acquisition
efficiency (PAE). While the literature abounds in reported genotypic differences in
internal P utilization efficiency (PUE) across a range of crops, there has been little
progress in breeding crop cultivars with high PUE. This review critically analyzes
why drawing conclusions from the body of research on PUE over the past few
decades remains difficult and how progress in breeding crop cultivars high in PUE
has been impeded. Four aspects of research on PUE are highlighted as being
critical in limiting our understanding and exploitation of PUE in grain crops:
(i) poor definition of PUE and inconsistent use of terminology, (ii) inappropriate
methods used in genotypic screening for PUE that fail to account for the con-
founding effects of PAE on PUE, (iii) inadequate discussion on the level of P stress
suffered by plants and its influence on potential mechanisms conferring high PUE
and their utility in cropping systems, and (iv) a focus on P-stress response
mechanisms rather than mechanisms conferring genotypic P-tolerance when
investigating PUE. These factors are discussed in detail and new approaches
and future areas of research on PUE are proposed.

1. Introduction
Phosphorus (P) is the second most limiting macronutrient to plant
growth and lack of plant-available P constrains plant growth in over 5.7
billion ha of land worldwide (Batjes, 1997). Large amounts of soil P are
“locked up” in recalcitrant organic P fractions or nonlabile inorganic P
pools in complexes with iron/aluminum in acid soils or with calcium in
alkaline soils. While application of phosphorus fertilizer can overcome these
constraints, the lack of locally available P fertilizer sources and the high cost
of importing and transporting P fertilizers typically prevent resource-poor
farmers in developing countries from doing so (Wissuwa and Ae, 2001).
Even in Western countries where food and fiber production relies heavily
on the application of nonrenewable P fertilizers, much of this fertilizer is
gradually rendered unavailable to plants over time as it reacts with soil
constituents (Richardson et al., 2011).
In addition to P immobilization in soils, a second frequently overlooked
factor is that the demand for P fertilizer is further driven by the removal of
almost 10million tons of P across the globe each year in harvested produce
(Lott et al., 2000), with grain crops (cereals, oilseeds, and pulses) being
responsible for the vast majority of this P removal at harvest (Lott et al.,
2002). Estimates suggest that as much as 85% of P applied as fertilizer can be
Internal Phosphorus Utilization Efficiency in Grain Crops 187

removed from fields in harvest product each year (Lott et al., 2009), though
regional imbalances in P budgets are apparent such that excess P inputs are
applied in many areas while 30% of the globe suffer from deficits in P
budgets (MacDonald et al., 2011). This inefficiency of P use in agriculture
has been highlighted in many recent reviews because of environmental
concerns regarding P from field runoff and human waste entering water-
ways and because finite global rock phosphate reserves are being depleted
(e.g., Cordell et al., 2009; Richardson et al., 2011; Shen et al., 2011).
Improving the P efficiency of farming systems (higher crop yields per
unit of P fertilizer applied) can be achieved using agronomic strategies, for
example, increasing P fertilizer availability to crops using liquid fertilizers
(Holloway et al., 2001) or strategic fertilizer placement (Ma et al., 2009).
In addition, breeding P-efficient crop cultivars has been advocated for
improving the P efficiency of cropping systems because it is relatively low
cost and, unlike optimizing fertilizer formulations and application strategies, it
can provide benefits both in high-input systems and in low-input systems
where fertilizer application may be rare or nonexistent due to its high cost
(Rose et al., 2010). Scope also exists to breed more P-efficient cultivars
because modern cultivars are generally not highly efficient at acquiring and
utilizing P, having been bred under optimal conditions that selected against
P-efficiency traits often present in landrace genotypes (Wissuwa et al., 2009).
Traits that confer P efficiency have typically been divided into those that
improve P acquisition efficiency (PAE) from soils or those that enhance
(internal) P utilization efficiency (PUE) (Wang et al., 2010). However, to
the best of our knowledge, all advances in breeding P-efficient grain crop
cultivars have involved the exploitation of PAE traits (Wissuwa et al., 2009),
and the genetics and mechanisms conferring PAE have been widely reviewed
elsewhere (Hinsinger, 2001; Ramaekers et al., 2010; Shen et al., 2011; Vance
et al., 2003). In contrast, PUE is poorly understood, despite a plethora of
studies reporting variation in PUE among grain crop genotypes (Table 1), the
mapping of several QTLs for PUE in a range of grain crops (Chen et al., 2009;
Hammond et al., 2009; Su et al., 2006, 2009; Wissuwa et al., 1998; Yang et al.,
2011), and discussion on PUE in many reviews on P efficiency (Ahmad et al.,
2001; Batten, 1992; Richardson et al., 2011; Shen et al., 2011; Shenoy and
Kalagudi, 2005; Vance et al., 2003; Wang et al., 2010).
A number of factors have contributed to our poor understanding of PUE
and lack of advances in breeding crops with enhanced PUE. First and
foremost, PUE is not clearly defined. Typically, PUE is expressed as biomass
per unit P, but biomass may refer to entire plant biomass or just grain yield,
whereas the unit P may refer to fertilizer P applied, P taken up by a plant, or
P present in specific tissues. The lack of a single definition being adopted
across studies has made it difficult to draw conclusions from the literature.
Second, PUE has long been the “‘poor cousin” in studies on P efficiency,
typically investigated as an additional component in studies primarily
188 Terry J. Rose and Matthias Wissuwa

Table 1 Summary of studies that have reported differences in PUE of

grain crop genotypes

Crop Authors
Barley Górny (1999) and Römer and Schenk (1998)
Canola Akhtar et al. (2008, 2009), Aziz et al. (2006), Duan et al.
(2009), Hammond et al. (2009), and Yang et al. (2011)
Common bean Araújo et al. (1997) and Fageria and da Costa (2000)
Faba bean Stelling et al. (1996) and T. Rose (unpublished data)
Maize Chen et al. (2009), Corrales et al. (2007), Fageria and Baligar
(1997a), and Parentoni and Júnior (2008)
Pigeon pea Adu-Gyamfi et al. (1989) and Subbarao et al. (1997)
Rice Aziz et al. (2005), Fageria and Baligar (1997b), Fageria et al.
(1988), Gill et al. (2002), Hafeez et al. (2010), Hedley et al.
(1994), Saleque et al. (1998), Sahrawat et al. (1997), and
Wissuwa and Ae (2001)
Triticale Oracka and Lapinski (2006)
Wheat Batten (1986), Batten and Khan (1987), Cao et al. (2009),
Fageria and Baligar (1999), Gill et al. (2004), Górny and
Garczy nski (2008), Gunes et al. (2006), Jones et al. (1989,
1992), Korkmaz et al. (2009), Manske et al. (2001, 2002),
Osborne and Rengel (2002a,b), Ozturk et al. (2005),
Sepehr et al. (2009), Su et al. (2006, 2009), Wang et al.
(2005), and Yaseen and Malhi (2009a)
Soybean Furlani et al. (2002), Kakar et al. (2002), Li et al. (2005), and
Zhang et al. (2009)

designed to assess PAE and because of this, screening methodologies have

not always been appropriate. Advancing our understanding of PUE has
been further hampered by a failure to acknowledge and discuss the implica-
tions of the level of P in screening media and the subsequent degree of stress
that plants experience. Finally, there has been a disproportionate focus on
molecular research characterizing P-stress responses of individual genotypes
(including mutants), as opposed to exploring loci conferring higher geno-
typic PUE in crop germplasm. Consequently, it remains to be seen whether
exploitable genetic variation exists for those few genes and mechanisms
that have so far been implicated in efficient internal P utilization (Pariasca-
Tanaka et al., 2009).
These factors are the focus of discussion in this review, which places
emphasis on PUE in grain crops because of their overriding influence on the
global P cycle. Throughout this review, our basic premise is that PUE
deserves most attention in plants experiencing some degree of P deficiency,
as efficient use of P will be most important if P limits growth. Implication
Internal Phosphorus Utilization Efficiency in Grain Crops 189

across a broader range of P supply will be discussed briefly. A second

premise of this review is that the most achievable means by which to
improve the PUE of modern crop cultivars is, at present, to identify loci
and superior alleles within germplasm and use marker-assisted introgression
to breed high-yielding varieties with specific PUE traits/genes. With
that goal in mind, less emphasis is placed on PUE traits in noncultivated
plants that have high PUE such as Banksia sp. (Denton et al., 2007) or stress
response genes/mechanisms that are upregulated under P-deficiency stress
in single genotypes of model species such as Arabidopsis (Morcuende et al.,
2007) but are not necessarily differentially regulated among genotypes that
differ in P efficiency (Pariasca-Tanaka et al., 2009).

2. Defining PUE: Terms, Units, and Assumptions

Criteria describing the P efficiency of plants have long suffered from
poor definition (Gourlay et al., 1994; Jones et al., 1989), and similarly,
multiple definitions exist for terms quantifying the internal PUE of plants
(Table 2). In addition, many terms and acronyms (e.g., PUE, P-efficiency
ratio (PER)) are used by various authors to refer to different criteria (Table 2).
To further complicate matters, the same acronym can have multiple meanings
beyond measurements of internal PUE: for example, the acronym PUE is
also used to refer to P uptake efficiency ( Jones et al., 1992) or P use efficiency
(grain yield per soil available P; Parentoni and Júnior, 2008). This range of
definitions and lack of consistent terminology for a given criterion are
undoubtedly responsible for some apparent contradictions in the literature
and has made drawing conclusions difficult. In this review, PUE is defined
as the biomass produced per unit P accumulated in tissue (gtissueDMmg
1
tissue P), and the grain yield per unit of P accumulated in aboveground plant
material is referred to as the PER (ggrainDMmg
1 shoot P).
The criteria used to quantify various aspects of internal PUE can be
broadly classified into those that have agronomic relevance and those that
have physiological relevance. Criteria that measure grain yield as a compo-
nent evidently have agronomic implications with PER or the “critical shoot
P concentrations for 90% maximum yield” being typical examples, whereas
on face value PUE (gDMmg
1 P) measurements have little direct agro-
nomic application.

2.1. Criteria with agronomic implications

Citing the trend of increasing average global grain yields over the past
half a century, it is possible to argue that PER has been selected for and
improved in modern crop varieties (Hammond et al., 2009). Indeed,
Table 2 Definitions of criteria used to assess phosphorus utilization efficiency in literature

Criterion definition and units Terminology Authors

Grain yield per unit of P P-efficiency ratio (PER) Batten (1992) and Jones et al. (1992)
accumulated in shoots at P internal utilization efficiency (PUTIL) Parentoni and Júnior (2008)
maturity (kg grain g
1 P) P utilization efficiency quotient (PEQ) Römer and Schenk (1998)
P utilization efficiency (PUE) Kakar et al. (2002)
P physiological efficiency index (PPEI) Yaseen and Malhi (2009a)
Physiological P use efficiency (PPUE) Ahmad et al. (2001)
Phosphorus utilization efficiency (PUTE) Manske et al. (2001, 2002)
Shoot dry matter per unit P P-efficiency ratio (PER) Gourlay et al. (1994) and Yaseen and Malhi
accumulated (g DM mg
1 P) (2009b)
Internal phosphorus utilization efficiency (PUE) Rose et al. (2011)
Phosphorus utilization efficiency (PUE) Oracka and Lapinski (2006) and Osborne
and Rengel (2002a,b)
Phosphorus utilization efficiency (PUTE) Sepehr et al. (2009)
Shoot DM/shoot P Phosphorus utilization index (PUI) Yaseen and Malhi (2009b)
concentration (g2 DM mg
1 P) Phosphorus utilization efficiency (PUE) Górny (1999), Gourlay et al. (1994), and
Siddiqi and Glass (1981)
Physiological phosphorus use efficiency (PPUE) Hammond et al. (2009)
Tissue P concentration required Chen et al. (2009) and White and
for 90% maximum yield Hammond (2008)
Internal Phosphorus Utilization Efficiency in Grain Crops 191

modern varieties (MVs) of rice typically had higher PERs than tradi-
tional or landrace cultivars in an upland (aerobic) screening trial: how-
ever, the PER was significantly correlated to harvest index (HI) such
that the low PER of many traditional varieties simply reflected their
much lower HI and grain yield potential (T. Rose and M. Wissuwa,
unpublished data). Similarly, studies with wheat also concluded that
PER was more highly correlated to HI than with grain yield (Batten
and Khan, 1987). Perhaps the most telling study was that of Manske et al.
(2002): these authors studied a number of P-efficiency traits in near
isogenic wheat lines differing in dwarfing characteristics and showed
that the yield-increasing effect of two dwarfing genes led to higher
PER. Thus, the apparent improvements in PER in MVs often reflect
their higher HI rather than specific selection of physiological traits that
improve internal P utilization in plants.
Calculating the tissue P concentration required for a given percentage of
maximal grain yield, referred to as the “critical” tissue P concentration if this
is 90% of maximal yield (White and Hammond, 2008), avoids the issue of
confounding P efficiency with HI because HI becomes inconsequential
when the yield of a genotype is expressed as a percentage of its own
maximum yield. However, this criterion may only be useful when all
genotypes screened have a similar yield potential: comparing the critical
tissue P concentration of a landrace with low yield potential (e.g., 4tha
1)
to that of a high-yielding MV (e.g., 8tha
1) is likely to produce misleading
results that may favor low-yielding genotypes.
A further issue with using grain yield as a parameter in defining internal
PUE, either as PER or as “critical” tissue P concentration, is that yield
formation not only depends on the yield potential of a given genotype
but also equally on location-specific effects such as day length, tempera-
ture, or disease pressure. Thus, while critical tissue P concentrations and
PER could be used in breeding nurseries to evaluate locally adapted high-
yielding breeding lines, neither criterion would be suitable to evaluate P
efficiency across a broad range of genotypes. Further, it remains question-
able whether sufficient genetic variation exists for these P-efficiency traits
among a set of rather similar breeding lines that have typically been
developed in well-fertilized nurseries (Wissuwa et al., 2009). Ideally,
screening studies investigating internal P utilization traits should maximize
the portion of the gene pool assessed within a crop species in search of
novel genes/alleles, many of which may only be present in older varieties
that were developed prior to the widespread use of P fertilizers. To enable
valid comparisons of older varieties and MVs, and to investigate physio-
logical mechanisms, it would be preferable to define PUE as the biomass
produced per unit P accumulated in tissue (gDMmg
1 P) and to dissect
PUE into components such as grain PUE, shoot PUE, and root PUE and
look to improve them individually.
192 Terry J. Rose and Matthias Wissuwa

2.2. Criteria with physiological implications

2.2.1. PUE at the vegetative stage
PUE has typically referred to shoot PUE (shoot dry weight per unit of P in
shoot) and occasionally root PUE, and both have been quantified across a
range of species (Table 1). Reported shoot P concentrations (the inverse of
PUE) range from below 1 to above 10mgPg
1 DM, depending on the
species and level of P supply (Table 3). The range of P concentrations
observed in roots tends to be narrower than in shoots, and root P concen-
trations are typically lower than corresponding shoot P concentrations
across most crop species (Table 3), possibly implying that roots have a
lower “P cost of production.” However, interpreting root PUE on a P
investment per unit DM basis may be misleading because ultimately, the
volume of soil explored by roots for P acquisition is driven by root length
(or surface area) rather than root biomass (the denominator when calculat-
ing root PUE). If root PUE was defined as mm2 root surface area mg
1 P,
then screening for root attributes known to increase root length with
minimal P investment such as root fineness, cortical aerenchyma, or root
hair formation (Lynch and Ho, 2005) may be more fruitful than screening
for lower root P concentrations per se. A study in maize reported that an
acknowledged P-efficient genotype had lower root P concentration in
lateral roots in addition to higher specific root length than a P-inefficient
genotype (Zhu and Lynch, 2004). It may therefore be possible to combine
the desired root traits mentioned above with mechanisms that allow roots to
function at a lower P concentration (i.e., high PUE). Further, data pre-
sented by Rose et al. (2011) suggest that control of root PUE and shoot PUE
is genetically independent, so selecting for both traits should be possible.
Investigating genotypic differences in tissue-specific PUE would ide-
ally lead to the identification and selection of favorable alleles that confer
increased biomass production per unit P within each tissue studied. Pre-
liminary results with rice suggest that such loci may differ between root
and shoot (Rose et al., 2011) and possible mechanisms are discussed in
Section 5. The validity of approaches focusing on screening for vegetative
PUE has been questioned based on the absence of direct evidence linking
vegetative PUE to improved grain yield (e.g., Siddiqi and Glass, 1981);
however, as outlined above, the absence of such a positive link might be
expected, given confounding effects of HI and yield potential. What
would be of concern is a negative association between vegetative PUE
and grain yield, but we are unaware of any data suggesting such associa-
tion. In fact, in rice, high-yielding cultivars were represented among the
highest and lowest ranked genotypes for shoot PUE (Rose et al., 2011).
The relationship between shoot PUE and grain yield is nonetheless criti-
cal, and the identification of loci conferring high PUE in crops should
facilitate further research in this area.
Table 3 Variation in shoot and root PUE (expressed as tissue P concentrations) during vegetative growth of three major grain crops

P concentration (mg P g
1 DM)

Crop Plant age (DAS) Shoots Roots Reference

Wheat 29 1.04 (low P) 0.74 (low P) Fageria and Baligar (1999)

4.27 (medium P) 1.79 (medium P)
4.29 (high-P) 2.19 (high P)
35 1.2–1.9 (low P) Osborne and Rengel (2002a)
2.6–5.9 (high P)
39 1.57–2.29 (low P) Ozturk et al. (2005)
2.80–4.49 (high P)
41 1.2–1.8 (low P) Osborne and Rengel (2002b)
1.8–2.8 (medium P)
6.9–11.9 (high P)
42 2.0–3.1 (low P) 3.0–6.5 (high P) Yaseen and Malhi (2009a)
3.1–6.2 (high P)
49 1.55–2.01 (low P) Gunes et al. (2006)
2.53–3.76 (high P)
54 0.6–1.7 (low P) Korkmaz et al. (2009)
4.0–12.7 (high P)
56 1.4–2.6 (low P) Sepehr et al. (2009)
4.1–5.9 (high P)
(Continued)
Table 3 (Continued)

P concentration (mg P g
1 DM)

Crop Plant age (DAS) Shoots Roots Reference

Rice 35 3.44–6.92 3.73–7.16 Hafeez et al. (2010)

50 0.60–1.14 (low P) 0.34–0.82 (low P) T. Rose et al. (unpublished)
2.80–4.50 (high P) 2.03–3.61 (high P)
73 1.2 (low P) 0.9 (low P) Fageria et al. (1988)
2.2 (medium P) 1.8 (medium P)
2.7 (high P) 2.0 (high P)
Unknown 0.41–0.52 (low P) 0.54–0.66 (low P) Hedley et al. (1994)
1.94–2.53 (high P) 1.1–1.2 (high P)
Canola 40 0.62–0.87 0.8–2.65 Aziz et al. (2006)
41 1.98–2.35 (low P) 4.41–9.14 (low P) Akhtar et al. (2009)
7.25–8.97 (high P) 3.54–4.02 (high P)
Internal Phosphorus Utilization Efficiency in Grain Crops 195

2.2.2. PUE at the reproductive stage

Crop P requirements in postanthesis growth are dominated by two com-
peting processes: the P requirements of vegetative tissues to continue
normal cellular function (including function throughout the senescence
process in determinant crops) and the P demand of the developing grain
(Rose et al., 2007). Theoretical calculations suggest that levels of P in grains
are well above the P levels required for normal cellular function (Raboy,
2009), consistent with earlier physiological studies that reported that grain
yield was not limited by the amount of P in the grain (Batten et al., 1986).
Thus, genotypic variation in grain P levels would likely be due to reduced
sink strength of the grain as a P storage organ rather than a lower physio-
logical requirement. Reduced P sink strength of the developing grain
may therefore be an ideal PUE trait in reproductive growth phases, given
higher remobilization of P from vegetative tissues to grains does not
improve grain yields (Batten et al., 1986) but may impair leaf function
(Garcia and Hanway, 1976). While such a trait has gained recent attention
(Richardson et al., 2011; Rose et al., 2010), it remains to be seen whether it
can be exploited in breeding crops with high PUE and whether a certain
threshold grain P concentration needs to be maintained to avoid loss in
seedling vigor, particularly on low-P soils.

2.3. Defining the utilization of P as “efficiency”

PUE (gDMmg
1 P) or PER (ggrainmg
1 aboveground P) are typically
evaluated following the assumption that higher PUE/PER in some geno-
types ultimately translates to superior biomass or grain yield. However, since
PUE is defined as a ratio (gbiomassmg
1 P), a genotype “A” may show
higher PUE (i) if it has proportionally higher biomass than P content
compared to a genotype “B” or (ii) if it has proportionally lower P
content than biomass relative to genotype “B.” Obviously case (i) is what
would make genotype “A” efficient in a positive sense that could be
exploited in breeding, whereas case (ii) would likely indicate a higher
degree of P starvation in genotype “A” and not true PUE. Unfortunately,
both cases cannot be distinguished without further examining the data
beyond simple calculations of PUE and it should be strongly emphasized
that such further analysis is crucial in order to avoid misinterpretations of
genotypic rankings in PUE (Rose et al., 2011; Wissuwa et al., 1998). The
most straightforward way of establishing whether PUE/PER values represent
true efficiency is to test for a positive correlation with biomass; should the
correlation be insignificant or negative, then PUE/PER values do not reflect
“efficiency” per se but may indicate a higher degree of P starvation. Such
negative correlations are surprisingly common (Gunes et al., 2006; Rose et al.,
2011; Sepehr et al., 2009; Wissuwa et al., 1998) typically arising in screening
196 Terry J. Rose and Matthias Wissuwa

Table 4 Correlation coefficients between shoot P content, biomass, and shoot PUE
from four experiments conducted over a range of P supply from deficient to fully
fertilized

Shoot biomass Shoot PUE

(a) Low-P soil (Wissuwa et al., 1998)
P content 0.98**
0.72**
Biomass
0.61**
(b) Mildly P deficient (Rose et al., 2010)
P content 0.73**
0.52**
Biomass 0.12ns
(c) Fixed (low) P supply (Rose et al., 2011)
P content 0.50**
0.51**
Biomass 0.49**
(d) Excess P (Rose et al., 2008)
P content
0.17ns
0.92**
Biomass 0.52**
** denotes significant difference at P < 0.001

experiments that concurrently evaluate genotypic differences in PAE under

severe or mild P deficiency (Table 4a and b). As will be discussed in more
detail in following chapters, it will be necessary to adjust screening methods to
avoid such “nonsensical” genotypic rankings for PUE (Table 4c).

3. Quantifying PUE of Crop Genotypes Using

Criteria with Physiological Implications
The above considerations regarding various definitions of PUE
already indicate that evaluating PUE is not as trivial as it may seem.
Screening methods would obviously differ based on the preferred defini-
tion, which, in turn, would depend on the ultimate purpose of screening. As
outlined above, the potentially confounding effects of HI, yield potential,
and local adaptation need to be considered and would limit screening for
what we defined as agronomic PUE criteria to sets of rather homogeneous
material such as high-yielding advanced breeding lines. For the identifica-
tion of novel loci/genes in more broad genetic material, protocols need
to be devised that allow for genotypic comparisons that are unbiased by
genotypic differences in traits that are not directly related to PUE. As shall
be discussed below, even for the definition of PUE deemed most suitable
for this purpose, gDMmg
1 P for various plant tissues, screening methods
require careful consideration.
Internal Phosphorus Utilization Efficiency in Grain Crops 197

3.1. Screening for vegetative PUE

3.1.1. Traditional screening systems
The standard method for quantifying the PUE of crop genotypes has been
to screen any number of genotypes in field or glasshouse conditions and to
simultaneously assess a number of P-efficiency criteria including PUE. Such
studies typically use measurements of shoot biomass, grain biomass, and
shoot and grain P concentrations to calculate PUE, PAE, PER, or other
criteria (e.g., Hammond et al., 2009; Oracka and Lapinski, 2006; Osborne
and Rengel, 2002a,b).
However, we recently hypothesized that such screening systems were of
limited use because PUE is strongly affected by the level of P deficiency
suffered by the plant, which is directly linked to the amount of P taken up
(Rose et al., 2011). This hypothesis was based on the widely reported phe-
nomenon that PUE decreases (tissue P concentrations increase) with increas-
ing P supply in the screening media (Fageria et al., 1988; Hedley et al., 1994;
Osborne and Rengel, 2002a,b; Saleque et al., 1998). Further, this decrease is
not linear (Fig. 1), which means that any additional unit P taken up would be
converted to biomass slightly less efficiently. As can be derived from the data in
Fig. 1, the effect of plant P content (or P uptake) on PUE is particularly
pronounced under P-deficient conditions—which typically are the growth
conditions where genetic improvements in PUE would be most welcome
and, hence, where precise genotypic evaluations are crucially important.
Should genotypes to be evaluated for PUE be screened in a system that
simultaneously allows for genotypic differences in PAE to be expressed, it is
likely that such differences in PAE would confound any difference in PUE,
given the nonlinear relation between both parameters.
Results from QTL mapping experiments confirm this direct link
between PUE and P uptake (PAE), since QTL for PUE and P uptake was
mapped to the same loci, albeit with opposite genotypic effects (Yang et al.,
2011). Even the Pup1 locus, the most influential QTL for P uptake identi-
fied in rice to date, was mapped to the same location as the main QTL for
PUE, but whereas the Pup1 locus doubled P uptake, it strongly reduced
PUE (Wissuwa and Ae, 1999). Key correlations between PUE, P content,
and biomass for this QTL mapping experiment (Table 4a) confirm the
negative association between PUE and P uptake. The phenotypic data of
this mapping population shown in Fig. 2 thus serve to illustrate three
important points, namely, (i) comparisons between genotypes that concur-
rently differ in P uptake do not allow for unbiased estimates of PUE,
(ii) estimates of high PUE of certain genotypes may be of no practical or
genetic relevance if they merely reflect a higher degree of P starvation, and
(iii) defining g DM per unit P (PUE) as “efficiency” can be rather mislead-
ing if its contextual significance is ignored, that is, if the correlation between
PUE and biomass is negative (Table 4a).
198 Terry J. Rose and Matthias Wissuwa

2.5

y = 2.31x −0.49
R 2 = 0.96
PUE shoot (g DM mg P-1) 2.0

1.5

1.0

0.5

0.0
0 10 20 30 40
Shoot P content (mg plant −1)

Figure 1 Shoot PUE as affected by shoot P content over a wide range of P supply
levels. Data presented are based on experimental data of Wissuwa et al. (2005). Briefly,
the experiment compared biomass accumulation of two closely related genotypes
differing for the presence of the Pup1 locus in a nutrient solution experiment under
four levels of P supply (3.2, 6.4, 9.6, and 100mMP). Genotypic differences were not
significant, so data from both genotypes were combined. A power function provided
the best fit to the experimental data.

Other studies have assessed relationships between P-efficiency criteria

in screening experiments (e.g., Hammond et al., 2009) or have attempted
to determine the contribution of both PUE and PAE to overall P efficiency
across genotypes (Parentoni and Júnior, 2008). However, given the interde-
pendence of these variables, and the overriding impact of PAE on biomass
production in screening trials on low-P soils, the results of any attempts to
determine the relative contributions of PAE and PUE are questionable.

3.1.2. A modified nutrient solution screening technique to nullify

differences in P uptake among genotypes
To overcome the confounding influence of P uptake in screening studies,
we recently developed a simple screening method in which genotypes were
grown in nutrient solution in individual containers supplemented with a
Internal Phosphorus Utilization Efficiency in Grain Crops 199

4 10

Dry matter (g plant–1)

6

4
3
2
PUE (mg P g–1 DM)

0
0 1 2 3 4 5 6
–1
P content (mg plant )

R 2 = 0.64

0
0 1 2 3 4 5 6
P content (mg plant–1)

Figure 2 Shoot PUE as affected by shoot P content in a QTL mapping population

grown on low-P soil (Wissuwa et al., 1998). Lines with the lowest P uptake appear to
be most efficient in utilizing P; however, they also fail to produce adequate shoot
biomass (inlay), which indicates that the apparent high PUE is due to severe P
starvation and not true efficiency in a sense that could be harnessed in plant
improvement.

specific low amount of soluble P to nullify the influence of P uptake (Rose

et al., 2011). Nutrient solution experiments have been used repeatedly to
estimate genotypic variation in shoot PUE across a range of species (Akhtar
et al., 2008; Hafeez et al., 2010; Yaseen and Malhi, 2009b); however,
all genotypes were typically grown in the same containers in competition
for the amount of P supplied. Consequently, confounding effects due to
differences in rates of P absorption among genotypes existed. This was
avoided in our modified method that used individual containers per geno-
type with 500mgP added to each container.
Twenty-nine rice genotypes were screened using this technique, result-
ing in variation in root PUE and shoot PUE that showed no correlation
with genotypic PUE rankings of the same genotypes screened in soil (Rose
et al., 2011). This study confirmed that genotypic differences in PUE do
exist in the absence of large genotypic differences in PUE, and further,
shoot PUE was positively correlated to biomass production (Table 4c),
suggesting that efficiency of P utilization was responsible for biomass pro-
duction (Rose et al., 2011).
200 Terry J. Rose and Matthias Wissuwa

While this technique holds promise for screening of large numbers of

rice genotypes, screening of other crop species that require aerated nutri-
ent solution may be impractical. It is possible that PUE mechanisms/loci
might be specific to certain level of P deficiency with little effects above/
below certain thresholds. It would therefore be desirable to screen at more
than one level of P deficiency or at least to confirm results with selected
genotypes over a range of P supplies. High PUE genotypes identified
using such screening techniques would require further study in soil-
based growth conditions to determine the effect of high-PUE mechanisms
when feedback mechanisms between PAE and PUE exist. Another factor
that affects all vegetative screening studies is the duration of plant growth,
since ontogeny plays a role in the critical P requirement of plant tissues
(Ahmad et al., 2001). Screening for PUE at the seedling stage may avoid
this complication but plants must be grown for sufficient duration to
eliminate effects of genotypic differences in seedling vigor; that is, plants
must be grown for such a period of time that all plants are P deficient
in order for PUE to become the primary driver of biomass production
(Rose et al., 2011).

3.2. Screening for grain PUE

Several studies have investigated the possibility of reducing grain P levels
using criteria including the proportion of aboveground P located in
the grain at maturity, or phosphorus HI or the P concentration in grains
at maturity (Batten, 1992; Jones et al., 1992; Rose et al., 2010). While
genotypic variation for PHI was apparent in these studies, it was strongly
influenced by the HI of any given genotype and was therefore of
questionable validity. Batten (1992) also reported that grain P concen-
tration appeared to be a function of grain yield in wheat (i.e., a dilution
effect whereby greater grain yield resulted in lower grain P concentra-
tion). However, we found no such correlation in a screening study with
rice (Rose et al., 2010) and Stangoulis et al. (2007) mapped QTLs for
rice grain P concentration in an AzucenaIR64 mapping population,
suggesting that grain P is under independent genetic control to a degree.
In fact, a number of studies have mapped QTLs for total grain P in the
model species Arabidopsis and its crop relative Brassica rapa (Bentsink
et al., 2003; Ding et al., 2010; Ghandilyan et al., 2009; Zhao et al., 2008),
but what is not known is whether these QTLs coincided with either
QTL for plant P uptake or grain yields. Thus, while it is possible
that screening for grain PUE (grain P concentration) may be achieved
using conventional screening trials with a wide range of genotypes, or
mapping populations, further research is required into the extent to
which plant mineral uptake and GE interactions play a role in grain
P concentrations in such studies.
Internal Phosphorus Utilization Efficiency in Grain Crops 201

4. P-Stress Levels in Screening Studies and the

Utility of PUE in Low, Medium, and High P
Input Systems
4.1. P-deficient crops suffer a range of stress levels
The level of P stress suffered by plants in screening studies has typically
varied depending on native P levels in the growing media and the amount
of supplemental P fertilizer, and as such, a range of shoot and root P
concentrations have emerged across screening studies to date (Table 3).
Given that the critical shoot P concentration for wheat and rice is in the
range of 2–5 and 1–2mgPg
1 DM, respectively, during the vegetative stage
(Dobermann and Fairhurst, 2000; Reuter and Robinson, 1997), the shoot P
concentrations summarized in Table 3 (ranging from 0.6 to 7–12mgPg
1
DM) represent the full range from highly P-deficient to P-replete plants,
with luxury P consumption being likely toward the high end tissue P
concentrations. These wide ranges reported pose the question whether
the concept of PUE should be applied equally to this entire range. At the
outset of the review, we stated that our primary focus is on PUE under P-
deficient conditions, defining P deficiency as a situation where an additional
supply of P to the plant would increase plant biomass (in the physiological
sense) or crop yield (in the agronomic sense).
The focus on shoot and root PUE in P-deficient plants is justified
because in plants that are not P-deficient, biomass production is typically
not correlated to PUE (Rose et al., 2011) or P uptake (Table 4d). In these
circumstances, any differences among genotypes tend to reflect differences
in luxury uptake of P, that is, uptake of P beyond the requirements
necessary for maximum biomass or grain yields. When PUE is not the
causal factor for differences in biomass accumulation, measurements of g
DMmg
1 P no longer represent “efficiency.” A recent review on PUE by
Wang et al. (2010) concluded that PUE had a dominant influence on the P
efficiency of crops at adequate and high P supply, and therefore PUE would
become a bottleneck for breeding P efficiency in high-P input agriculture.
However, the premise for this conclusion was that PAE had little impact in
high-P conditions (because sufficient P is supplied as P fertilizer), and the
paper of Manske et al. (2001) was cited as evidence for the high contribution
of PUE to P efficiency in P-fertilized crops. However, the PER was not
significantly positively correlated with grain yield in any of the eight
treatment combinations in the study of Manske et al. (2001); hence, these
measurements do not satisfy the requirements of “efficiency” as defined in
Section 2.3. In contrast to the conclusion of Wang et al. (2010), we suggest
that in adequate or high-P conditions, shoot or root PUE may be of limited
practical relevance.
202 Terry J. Rose and Matthias Wissuwa

There is, however, a strong desire to enhance the P efficiency of highly

productive agricultural systems because heavy reliance on P fertilizer has led
to eutrophication of water bodies due to P runoff from fields (Sharpley et al.,
2001), as well as increased production costs. While enhancing shoot and
root PUE may be of little benefit in high-input agriculture, improving the
grain PUE may be promising for improving the P efficiency of such systems
(Richardson et al., 2011; Rose et al., 2010). However, little work has been
done in this field of research and its potential to improve the P efficiency of
farming systems will depend on whether genetic variation can be exploited
or mutant/transgenic approaches can be successfully applied, and the impact
of such a trait on other grain quality attributes and agronomic traits when
the grain is used as seed.
While at the opposite end of the spectrum are crops grown in poverty-
stricken nations in Africa that suffer from severe P-deficiency stress, the
majority of P-deficient crops across the globe tend to suffer from a more
moderate level of P-deficiency stress. One issue with the study of Rose et al.
(2011) was that the level of P used to screen genotypes (one-off dose of 500m
gP) resulted in severely P-deficient plants that are rarely observed in com-
mercial rice fields. Investigating PUE in P-starved plants has also been used to
elucidate P-stress response mechanisms in rice at the molecular level
(Morcuende et al., 2007; Pariasca-Tanaka et al., 2009), but whether any
genes upregulated or physiological mechanisms identified in such P-deficient
plants have any relevance in plants suffering mild P deficiency is not known.
Screening plants at a P level sufficient for production of 70–80% maxi-
mum biomass yield at a given growth stage may identify mechanisms that
are active under typical field conditions, and whether such a level of stress is
sufficient to allow phenotypic separation of genotypes for PUE needs to be
resolved. Perhaps the most comprehensive screening approach would be to
derive P response curves to allow determination of critical tissue P con-
centrations for any percentage of maximum shoot biomass yields. However,
screening large numbers of genotypes in this fashion would be prohibitively
expensive and impractical. Further studies on PUE are needed to address
this issue and elucidate which mechanisms that enhance PUE operate at
various P-stress levels, whether genotypic variation exists, and whether such
mechanism have any impact when plants are grown in field conditions.

4.2. What are the likely outcomes of improved PUE in

P-deficient plants?
While identifying mechanisms and genes that enhance PUE in plants
suffering from varying degrees of P-deficiency stress has proved difficult,
predicting the phenotypic outcome of the exploitation of such mechanisms
on the P efficiency of commercial crop cultivars is equally difficult. As
discussed earlier, P uptake affects PUE and confounds PUE rankings in
Internal Phosphorus Utilization Efficiency in Grain Crops 203

genotypic screening studies on P-deficient soils. However, modeling studies

suggest that PUE also affects P uptake, either because higher root PUE
would enhance root growth directly since a greater root surface area could
be produced at equal P cost or because root growth is affected indirectly if
higher shoot PUE allows greater P distribution to roots (Wissuwa, 2003).
The theoretical outline of the model with regard to effects of changes in
PUE is shown in its most simplified form in Fig. 3: an increase in root PUE
of 20% would produce 20% higher root biomass and soil exploration
leading to 20% more P uptake above what would have been taken up in a
genotype without the PUE increase. At time txþ1, a portion of the additional
P taken up remains in roots (6%, assuming 70% of P is distributed to shoots),
and this additional P is also converted to root biomass at the higher internal
efficiency, resulting in 27.2% higher biomass and P uptake. Over an entire
cropping cycle, the effect of improving PUE would therefore be far greater
than the initial “physiological” PUE differential between two genotypes

Shoot
biomass

Assimilate distribution
PUE
tx : +20%
tx+1: +27.2%
tx+2: +29.8%
70%
Root:shoot
P uptake
P distribution
30% tx : -
tx+1: +6%
tx+2: +8.2%
PAE PUE

Soil Root
Architecture
exploration biomass

tx : +20%
tx+1: +20% +(6*1.2)%
tx+2: +20% +(8.2*1.2)%

Figure 3 Schematic outline of the principles behind the model “P-LIM-GROW,”

assuming a 20% increase in PUE of roots, as indicated by 20% higher root biomass at
time tx. Adapted from Wissuwa (2003) and Rose et al. (2011).
204 Terry J. Rose and Matthias Wissuwa

would suggest because increased PUE would also enhance P uptake (Rose
et al., 2011; Wissuwa, 2003). Small genotypic differences in PUE as detected
in our modified screening experiment (Rose et al., 2011) could therefore be
well worth exploiting in plant breeding, at least in P-deficient scenarios where
growth is limited directly by insufficient P availability rather than by assimilate
supply (Wissuwa et al., 2005).
In addition to the obvious reason for focusing on PUE, the efficient use
of a limited and costly resource, improving PUE rather than PAE alone
may be more sustainable since relying solely on increasing P uptake would
pose the danger of overmining soil-P reserves, particularly in low or zero
input agricultural systems (Henry et al., 2010). The theoretical considera-
tions put forward in Fig. 3, as well as additional experimental data on low-
cost root systems and their effect on soil exploration (Zhu and Lynch,
2004), indicate that a strict separation of PUE and P uptake may be
impossible since enhanced PUE, especially in roots, would also enhance
P uptake and mining of soil-P.
Thus, improving grain PUE is theoretically the only option for reducing
the amount of P taken up by crops in P-deficient scenarios, with the added
benefit of minimizing the removal of P from fields, which would be beneficial
in low-input and high-input cropping systems. However, any potential
reductions in P uptake by crops with high grain PUE will depend on whether
a reduction in competition for P between the grain and vegetative tissues
triggers a decline or cessation in P uptake from soil. Research to date has
produced no firm conclusions: for example, early work suggested that wheat
crops acquire sufficient P for maximum grain yields by anthesis (Batten et al.,
1986; Boatwright and Viets, 1966), yet studies have shown that wheat plants
may cease (net) soil uptake at anthesis, accumulate small additional amounts of
P beyond anthesis, or take up a substantial proportion of total plant P after
anthesis (Manske et al., 2002; Mohamed and Marshall, 1979; Römer and
Schilling, 1986; Rose et al., 2007) and this pattern is genotype specific ( Jones
et al., 1992). Indeterminate crops such as canola, or crops that are partially
perennial such as rice, continue to acquire P from the soil during grain filling
to satisfy the competing demands of actively growing vegetative tissue and the
developing grain (Rose et al., 2009, 2010).
Thus, it is unclear whether determinate or indeterminate crop species (or
genotypes within crop species) would reduce their net P uptake if postanthesis
P demand was reduced. Whether regulatory mechanisms exist that reduce P
uptake from soils when P demand has been satisfied during postanthesis
growth is therefore a key question that remains to be addressed in future
research. Interestingly, among the suite of mutants with reduced phytic acid
levels in grains (lpa mutants) developed over the past decade, one barley lpa
mutant also has lower total grain P (Raboy, 2009). However, to the best of
our knowledge, it is not known whether this mutant has lower postanthesis P
uptake from soil compared to wild-type barley plants.
Internal Phosphorus Utilization Efficiency in Grain Crops 205

5. Mechanisms and Physiology of PUE

Most of our knowledge on plant adaptation to P-deficiency stress
pertaining to internal P utilization has emerged from studies that have
examined the physiological or transcriptional changes that occur in single
genotypes of a species that have been subjected to P stress. As with geno-
typic screening studies, the range of P deficiency suffered by plants in such
studies has varied from severely P-deficient plants grown for weeks in a
P-deficient soil (Pariasca-Tanaka et al., 2009) to plants under less severe P
stress that were grown in optimum (200mM) P solution before being
transferred to low-P nutrient solution for 1–10 days (Calderón-Vázquez
et al., 2008). The duration or severity or P stress determines both the
number of type of genes that are up- or downregulated as in response to
this stress (Hammond et al., 2003; Pariasca-Tanaka et al., 2009).

5.1. Remobilization and scavenging of P

A major plant adaptation to P-deficiency stress is the enhanced remobilization
and recycling of P in plant tissues. Under P-replete conditions, the vast
majority of Pi is stored in vacuoles with only 5–15% of Pi in the cytoplasmic
pool, and in the early stages of P deficiency, inorganic P is redistributed from
the vacuoles to the cytoplasm to maintain minimum cytosolic P levels
(Raghothama and Karthikeyan, 2005). However, recent work suggests that
during P starvation Pi efflux from the vacuole cannot compensate for a rapid
decrease of the cytosolic Pi concentration, and this sudden drop may be the
first endogenous signal triggering the Pi starvation response (Pratt et al., 2009).
As P starvation sets in, both inorganic P and P remobilized from organic
compounds are redistributed from older tissues to metabolically active tissues
including growing points and young leaves and active roots (Akhtar et al.,
2008). Membrane lipid composition may be altered to reduce consumption
of P, and degradation of existing phospholipids into Pi and diacylglycerol is a
key process in the remobilization of P during P starvation (Li et al., 2006).
Because P is a key component of nucleic acids, phospholipids and meta-
bolites of energy-mediated metabolic processes, phosphatases, ribonucleases,
and several key genes involved in lipid metabolism including phospholipase,
UDP-sulfoquinovose synthase (SQD1), and sulfoquinovosyldiacylglycerol
(SQD2) are among those that are strongly upregulated under P-deficiency
stress (Calderón-Vázquez et al., 2008; Pariasca-Tanaka et al., 2009). Increased
expression of putative Pi-transporters would then allow for rapid redistribu-
tion of the P remobilized by above processes (Vance et al., 2003). Interest-
ingly, while it is generally thought that upregulation of genes such as
Pi-transporters will improve PUE, in the case of grain PUE, the opposite
206 Terry J. Rose and Matthias Wissuwa

may be the case because a reduction in translocation of P to the developing

grain is desirable (Rose et al., 2010). Thus, reduced expression of high-affinity
Pi-transporters in reproductive tissues may result in a decrease in grain P
levels and enable continual utilization of P in photosynthetic tissues (Lambers
et al., 2006).

5.2. Alternative glycolytic pathways and mitochondrial

electron transport pathways
Several enzymes of the glycolytic pathway are Pi dependent, but in plants
suffering P-deficiency stress Pi-dependent or adenosine-tri-phosphate
(ATP)-dependent steps can be bypassed (Theodorou and Plaxton, 1996).
Similarly, alternative mitochondrial electron transport pathways can be uti-
lized that bypass the adenylate and Pi reaction (Theodorou and Plaxton,
1993). For detailed discussion on the various alternative pathways and
the enzymes involved, the reader is referred to reviews by Theodorou and
Plaxton (1993) and Vance et al. (2003).
In addition to the physiology of alternative glycolytic and respiratory
pathways, much information regarding the genes involved in these processes
has been gleaned from transcriptomics studies on P-deficient plants. In
particular, genes such as phosphofructokinase (PFK), pyrophosphate:
fructose-6-phosphate phosphotransferase (PFP), pyruvate kinase, and phos-
phoenolpyruvate carboxylase (PEPC) have been implicated in modifications
to carbon metabolism that bypass P-requiring steps (Hammond et al., 2004;
Pariasca-Tanaka et al., 2009). In addition, a multitude of genes are involved in
regulating the above responses to P stress (Hammond et al., 2004).
There are a multitude of genes and physiological processes that are
affected by P-deficiency stress, and these have been the subject of a number
of review papers over recent years (Ahmad et al., 2001; Calderón-Vázquez
et al., 2011; Hammond et al., 2004; Shen et al., 2011; Shenoy and Kalagudi,
2005; Vance et al., 2003; Wang et al., 2010), and the reader is referred to
these reviews for detailed discussions. In this review, we focus on the key
question as to whether these genes can potentially be manipulated to
improve the PUE of grain crop cultivars.

5.3. Exploiting P-deficiency stress response mechanisms

Attempts have been made to enhance the PAE of crop cultivars through the
overexpression of genes that are responsive to P-deficiency stress, but to
date, the results have not been entirely promising. For example, transgenic
plants that have enhanced phytase exudation can effectively access organic
P sources in artificial conditions, but their usefulness in field situations
is limited (George et al., 2005). Similarly, overexpression of high-affinity
P transporters did not enhance P uptake rates from soil (Rae et al., 2004) and
Internal Phosphorus Utilization Efficiency in Grain Crops 207

overexpression of genes related to the synthesis of organic acids has not yet
proved to be a successful strategy for increasing the PAE of plants (summar-
ized in Richardson et al., 2011).
Similarly, the known genetic and physiological responses of plant to
P-deficiency stress related to PUE have yet to have made the leap from
the laboratory to the field. As yet, we are unaware of any studies that
have conclusively shown that the overexpression of any P-stress response
genes confers higher PUE in plants. Wang et al. (2009) overexpressed an
Arabidopsis purple acid phosphatase gene (AtPAP15) in soybean and
reported higher yields in the transgenic plants when grown on a low-P
acid soil, and concluded that higher internal PUE was the likely cause.
However, no data on P uptake by plants or tissue P concentrations were
presented, and in earlier experiments, the transgenic plants showed
greater phytase secretion from roots and enhanced P acquisition from
organic P sources; hence, improved P uptake may have been the causal
factor in yield differences.
The lack of progress in improving P efficiency through manipulation
of genes that are responsive to P stress suggests that in many instances
P-stress response mechanisms are not necessarily those that confer toler-
ance to P stress. A microarray study comparing a P-efficient rice near
isogenic line containing the Pup1 locus with its P-inefficient parent
Nipponbare came to similar conclusions with regard to many PUE stress
response genes (Pariasca-Tanaka et al., 2009). While many of the afore-
mentioned PUE-related genes were upregulated in root and shoot tissue,
virtually all were induced to a similar or higher degree in the P-inefficient
Nipponbare than the P-efficient NIL6-4 (Pariasca-Tanaka et al., 2009).
Thus, even genotypes that show poor tolerance to P deficiency in the field
have evolved many of the well-known adaptations to P stress. Questions
remain as to whether the manipulation of P-stress response genes using
biotechnology can facilitate improvements in the PUE of field-grown
crops, but given the rate at which global P resources are being depleted,
further research is warranted.
In terms of breeding crops higher in PAE, the most promising option to
date has been to exploit natural variation among genotypes (Lynch, 2007;
Wissuwa et al., 2009). It is therefore surprising that although the physiolog-
ical and molecular responses to P stress have been widely studied, little is
known regarding mechanisms that confer higher PUE in genotypes of the
same species. While there still remains much to be elucidated about the
genetics and physiology of P-stress responses, further research also needs to
target the identification of valid QTLs from appropriate screening studies
to determine the genes and mechanisms that confer enhanced genotypic
PUE. Genes identified in such studies may prove useful in breeding pro-
grams but may also provide targets for genetic manipulation in addition to
the well-known genes that are affected by P-deficiency stress.
208 Terry J. Rose and Matthias Wissuwa

6. Conclusions and Future Prospective

6.1. PAE versus PUE—Does one offer better chances
of success?
Breeding P-efficient crop cultivars and developing more P-efficient agro-
nomic practices are needed for future food security and environmental
protection because the current global P cycle is not sustainable (Cordell
et al., 2009). Breeding efforts in the past have disproportionally focused on
improving PAE of crops (Wissuwa et al., 2009), and while there is no doubt
about the importance of developing cultivars that are more efficient at
acquiring soil P, it is equally obvious that improving the PUE of crop cultivars
would complement PAE traits in an ideal fashion. Despite this, we are not
aware of any serious effort to specifically improve PUE in plant breeding.
This is to some extent due to the difficulty in evaluating PUE without having
genotypic rankings confounded by differences in HI and P uptake. A second
reason may have been that genotypic variation for PAE typically appears to be
higher than for PUE (Parentoni and Júnior, 2008; Wissuwa and Ae, 2001).
However, upon closer inspection of the data, this assessment no longer holds
because values typically shown for P uptake (or P content) do not represent
measurements of the uptake process itself but rather their aggregate effect
(Wissuwa et al., 2005) over extended periods up to an entire crop cycle. If two
genotypes differ twofold in P uptake after a 100-day growth period, this does
not mean that they differ by an equally large margin for the causal mechanisms
that trigger more efficient P uptake. Genotypic differences in causal mechan-
isms can be rather small (Wissuwa, 2003) and differences in P uptake may
even be caused by changes in PUE (Fig. 3), particularly by root morphologi-
cal adaptations that reduce the cost of producing and maintaining roots
(Lynch and Ho, 2005). Thus, while data from screening experiments cer-
tainly overestimate the genotypic variation present for mechanisms enhancing
PAE, the full benefits of increased PUE are likely underestimated. Further,
due to the impact of P uptake on PUE and vice versa, determining the
relative contributions of PUE and PAE to P efficiency may not be possible.

6.2. Screening methods, targets, and possible results

PAE has an overriding and unavoidable influence on PUE rankings in soil-
based screening experiments, and this has led to confounded results and
misleading conclusions in a variety of studies on PUE (Rose et al., 2011).
One would therefore ideally screen at equal P uptake or derive genotype-
specific P response curves over a target range of P content (Rose et al.,
2011). While the latter option would be ideal, the effort needed to produce
response curves for a large number of genotypes would in most cases be
Internal Phosphorus Utilization Efficiency in Grain Crops 209

prohibitive. We therefore tested of a modified nutrient solution screening

technique that uncouples PAE and PUE by ensuring that all screened
genotypes have the same P uptake. Results showed that genotypic differ-
ences in PUE do exist in the absence of differences in P uptake (Rose et al.,
2011), that genotypic rankings of the unbiased screening method differed
considerable from those in conventional soil-based method, and that rank-
ings for PUE in shoot and root tissue were not related.
While further confirmation is needed to establish that the genetic basis of
PUE differs from shoot to root tissue, such a notion makes intuitive sense
since shoots and roots perform entirely different functions. PUE mechanisms
could therefore relate to different physiological processes or morphological
attributes. The tissue specificity of PUE does become more obvious if PUE
was defined in a more root-specific way as root surface area (rather than
biomass) per milligram P, since root-specific morphological modifications
(root hairs, aerenchyma, proportion of vascular tissue) could be targeted with
the aim of producing a low-cost root system (Lynch and Ho, 2005).
Yet another target for improvements in PUE is the grain tissue. Since P
concentrations in grains tend to be an order of magnitude higher compared to
roots or shoots (Rose et al., 2010), and theoretically higher than needed for
normal physiological function (Raboy, 2009), it is likely that some P translo-
cation and loading mechanisms would determine grain PUE, rather than
physiological redistribution or substitution mechanisms expected to operate
in shoot or root tissue. Although QTL for grain P concentration have been
identified (Stangoulis et al., 2007; Zhao et al., 2008), such QTL may be related
to either yield formation (causing a dilution effect on grain P concentration;
Batten, 1992) or plant P uptake, which determines the available pool of P that
can be remobilized to developing grains. Another critical question is what are
the minimum P requirements of developing seeds for maximum yield forma-
tion and subsequent germination of seeds? These issues, and the more
fundamental question of which phosphate transporters are involved in loading
P into grains and how are they regulated at the molecular level, are critical
areas of research that need to be addressed. However, while this remains an
exciting emerging area of research from a plant physiological and P-efficiency
point of view, the potential impact of a low-grain-P trait on human health
and seedling vigor cannot be ignored. Further research is needed to address
these concerns, but such research cannot be adequately conducted until either
low-grain-P near isogenic lines or mutants become available.

6.3. Marker-assisted selection—A paradigm shift in breeding

suited for PUE
Traditional plant breeding relied almost entirely on phenotypic selection,
with grain yield being the far most important selection criterion. Selecting
for tissue-specific PUE would have had little appeal for breeders, given that
210 Terry J. Rose and Matthias Wissuwa

a direct positive effect of PUE on grain yield is difficult to prove considering

that factors like HI or PAE tend to obscure effects of higher PUE. The
recent paradigm shift in plant breeding toward marker-assisted selection
(MAS) now allows for the introgression of traits that are difficult to evaluate
and, further, assures that any undesirable attributes of a donor variety
are efficiently eliminated (Septiningsih et al., 2009). The yield potential of
the donor parent is therefore of little consequence in MAS, which now
enables breeders to make full use of the genetic variation present in gene-
bank collections.
The task at hand therefore is to evaluate a large number of genebank
accessions for specific PUE traits and to identify loci controlling these traits.
Genome-wide association mapping, where marker information of genebank
accessions is used in conjunction with some phenotypic data to identify loci
associated with the phenotype, is ideally suited to identify loci controlling
PUE in various tissues. Once superior alleles for such traits have been
identified, they can be transferred via MAS to a recipient variety with high
yield potential. Since this process is highly efficient, it is feasible to combine
(pyramid) the most promising PUE loci with loci for high PAE in a single
recipient variety, thus creating a variety with superior overall P efficiency.

6.4. Remaining questions

As plants suffer higher levels of P-deficiency stress, their PUE increases. It is
therefore tempting to speculate that alleles conferring higher PUE will be most
useful under strongly P-limiting conditions and that screening should also take
place under these conditions. While such conditions are common where
farmers have no access to P fertilizers or are too resource-poor to “invest” in
fertilizer application (Ismail et al., 2007), in reality the majority of crops
experiencing P deficiency only suffer a mild form, and estimating potential
benefits of enhanced PUE under such mild P deficiency is much less straight-
forward. Moreover, the utility of high-PUE traits identified in strongly P-
limiting conditions for improving PUE in plants suffering milder P-deficiency
stress is not known. A further critical question is whether genotypes (or QTL/
genes) identified in artificial screening methods that separate P uptake and
PUE ultimately enhance P efficiency under field conditions.
Improving root/shoot PUE in high-input or P-replete conditions
would also be desirable, but we believe this cannot be achieved through
plant improvement. Data presented in Table 4d indicated a very tight
negative correlation between PUE and P uptake under high-P input,
which essentially means that one would have to select against P uptake in
order to improve PUE. That would be counterproductive as fertilizer P
would remain in soil unused. Thus, in high-input systems, what should be
targeted is the overall system P efficiency and reductions in P fertilizer
application, possibly in combination with improved fertilizer timing or
Internal Phosphorus Utilization Efficiency in Grain Crops 211

placement, appear most suitable. Targeting grain PUE may be of additional

benefit because more P will be retained on-farm if the straw is not exported
from fields.
Much effort in breeding P-efficient crops has focused on improving
PAE because many resource-poor farmers cannot afford P fertilizers and
even when P fertilizers are used they become gradually immobilized in soils
over time (Richardson et al., 2011). The downside of focusing solely on
improving PAE is that such genotypes increase the rate of mining of soil P
by removing more P in harvested products (Henry et al., 2010) and little of
the P removed in harvested products is recycled back onto fields (Cordell
et al., 2009). While much remains unknown about mechanisms that could
enhance PUE beyond the level already present in most crops, and whether
they can be exploited in breeding, research to address the gaps is warranted:
if plant breeding is to contribute to mitigating an unsustainable global P
cycle, then high-PUE traits must ultimately complement high PAE traits in
crop cultivars.

REFERENCES
Adu-Gyamfi, J. J., Fujita, K., and Ogata, S. (1989). Phosphorus absorption and utilization
efficiency of pigeon pea (Cajanus cajan (L) Millsp.) in relation to dry matter production
and dinitrogen fixation. Plant Soil 119, 315–324.
Ahmad, Z., Gill, M. A., and Qureshi, R. H. (2001). Genotypic variations of phosphorus
utilization efficiency of crops. J. Plant Nutrition 24, 1149–1171.
Akhtar, M. S., Oki, Y., and Adachi, T. (2008). Genetic variability in phosphorus acquisition
and utilization efficiency from sparingly soluble P-sources by Brassica cultivars under
P-stress environment. J. Agron. Crop Sci. 194, 380–392.
Akhtar, M. S., Oki, Y., and Adachi, T. (2009). Mobilization and acquisition of sparingly
soluble P-sources by Brassica cultivars under P-starved environment I. Differential
growth response, P-efficiency characteristics and P-remobilization. J. Integr. Plant Biol.
51, 1008–1023.
Araújo, A. P., Teixeira, M. G., and De Almeida, D. L. (1997). Phosphorus efficiency of wild
and cultivated genotypes of common bean (Phaseolus vulgaris L.) under biological nitro-
gen fixation. Soil Biol. Biochem. 29, 951–957.
Aziz, T., Gill, M. A., Rahmatullah, M. A., and Mansoor, T. (2005). Differences in
phosphorus absorption, transport and utilization by twenty rice (Oryza saliva L.) cultivars.
Pak. J. Agric. Sci. 42, 8–15.
Aziz, T., Rahmatullah, M. A., Tahir, M. A., Ahmad, I., and Cheema, M. A. (2006).
Phosphorus utilization by six Brassica cultivars (Brassica juncea L.) from tri-calcium
phosphate; a relatively insoluble compound. Pak. J. Bot. 38, 1529–1538.
Batjes, N. H. (1997). A world data set of derived soil properties by FAO-UNESCO soil unit
for global modeling. Soil Use Manag. 13, 9–16.
Batten, G. D. (1986). The uptake and utilization of phosphorus and nitrogen by diploid,
tetraploid and hexaploid wheats (Triticum spp.). Ann. Bot. 58, 49–59.
Batten, G. D. (1992). A review of phosphorus efficiency in wheat. Plant Soil 146, 163–168.
Batten, G. D., and Khan, M. A. (1987). Uptake and utilisation of phosphorus and nitrogen
by bread wheats grown under natural rainfall. Aust. J. Exp. Agric. 27, 405–410.
212 Terry J. Rose and Matthias Wissuwa

Batten, G. D., Wardlaw, I. F., and Aston, M. J. (1986). Growth and the distribution of
phosphorus in wheat developed under various phosphorus and temperature regimes.
Aust. J. Agric. Res. 37, 459–469.
Bentsink, L., Yuan, K., Koornneef, M., and Vreugdenhil, D. (2003). The genetics of phytate
and phosphate accumulation in seeds and leaves of Arabidopsis thaliana, using natural
variation. Theor. Appl. Genet. 106, 1234–1243.
Boatwright, G. O., and Viets, J. F. G. (1966). Phosphorus absorption during various growth
stages of spring wheat and intermediate wheatgrass. Agron. J. 58, 185–188.
Calderón-Vázquez, C., Ibarra-Laclette, E., Caballero-Perez, J., and Herrera-Estrella, L.
(2008). Transcript profiling of Zea mays roots reveals gene responses to phosphate
deficiency at the plant- and species-specific levels. J. Exp. Bot. 59, 2479–2497.
Calderón-Vázquez, C., Sawers, R. J. H., and Herrera-Estrella, L. (2011). Phosphate depri-
vation in maize: Genetics and genomics. Plant Physiol. 156, 1067–1077.
Cao, H. X., Zhang, Z. B., Sun, C. X., Shao, H. B., Song, W. Y., and Xu, P. (2009).
Chromosomal location of traits associated with wheat seedling water and phosphorus use
efficiency under different water and phosphorus stresses. Int. J. Mol. Sci. 10, 4116–4136.
Chen, J., Xu, L., Cai, Y., and Xu, J. (2009). Identification of QTLs for phosphorus
utilization efficiency in maize (Zea mays L.) across P levels. Euphytica 167, 245–252.
Cordell, D., Drangert, J., and White, S. (2009). The story of phosphorus: Global food
security and food for thought. Glob. Environ. Chang. 19, 292–305.
Corrales, I., Amenós, M., Poschenreider, C., and Barceló, J. (2007). Phosphorus efficiency
and root exudates in two contrasting tropical maize varieties. J. Plant Nutr. 30, 887–900.
Denton, M. D., Veneklaas, E. J., Freimoser, F. M., and Lambers, H. (2007). Banksia species
(Proteaceae) from severely phosphorus-impoverished soils exhibit extreme efficiency in
the use and re-mobilisation of phosphorus. Plant Cell Environ. 30, 1557–1565.
Ding, G., Yang, M., Hu, Y., Liao, Y., Shi, L., Xu, F., and Meng, J. (2010). Quantitative trait
loci affecting seed mineral concentrations in Brassica napus grown with contrasting
phosphorus supplies. Ann. Bot. 105, 1221–1234.
Dobermann, A., and Fairhurst, T. H. (2000). Nutrient Disorders and Nutrient Management.
Potash and Phosphate Institute of Canada and International Rice Research Institute,
Singapore.
Duan, H. Y., Shi, L., Ye, X. S., Wang, Y. H., and Xu, F. S. (2009). Identification of
phosphorous efficient germplasm in oilseed rape. J. Plant Nutr. 32, 1148–1163.
Fageria, N. K., and Baligar, V. C. (1997a). Phosphorus-use efficiency by corn genotypes.
J. Plant Nutr. 20, 1267–1277.
Fageria, N. K., and Baligar, V. C. (1997b). Upland rice genotypes evaluation for phosphorus
use efficiency. J. Plant Nutr. 20, 499–509.
Fageria, N. K., and Baligar, V. C. (1999). Phosphorus-use efficiency in wheat genotypes.
J. Plant Nutr. 22, 331–340.
Fageria, N. K., and da Costa, J. G. C. (2000). Evaluation of common bean genotypes for
phosphorus use efficiency. J. Plant Nutr. 23, 1145–1152.
Fageria, N. K., Wright, R. J., and Baligar, V. C. (1988). Rice cultivar evaluation for
phosphorus use efficiency. Plant Soil 111, 105–109.
Furlani, A. M. C., Furlani, P. R., Tanaka, R. T., Mascarenhas, H. A. A., and
Delgado, M. D. P. (2002). Variability of soybean germplasm in relation to phosphorus
uptake and use efficiency. Sci. Agric. 59, 529–536.
Garcia, R. L., and Hanway, J. J. (1976). Foliar fertilization of soybean during the seed-filling
period. Agron. J. 68, 653–657.
George, T. S., Richardson, A. E., Smith, J. B., Hadobas, P. A., and Simpson, R. J. (2005).
Limitations to the potential of transgenic Trifolium subterraneum L. plants that exude
phytase, when grown in soils with a range of organic phosphorus content. Plant Soil
278, 263–274.
Internal Phosphorus Utilization Efficiency in Grain Crops 213

Ghandilyan, A., Ilk, N., Hanhart, C., Mbengue, M., Barboza, L., Schat, H., Koornneef, M.,
El-Lithy, M., Vreugdenhil, D., Reymond, M., and Aarts, M. G. M. (2009). A strong
effect of growth medium and organ type on the identification of QTLs for phytate and
mineral concentrations in three Arabidopsis thaliana RIL populations. J. Exp. Bot. 60,
1409–1425.
Gill, M. A., Mansoor, S., Aziz, T., Rahmatullah, M. A., and Akhtar, M. S. (2002).
Differential growth response and phosphorus utilization efficiency of rice genotypes.
Pak. J. Agric. Sci. 39, 83–87.
Gill, H. S., Singh, A., Sethi, S. K., and Behl, R. K. (2004). Phosphorus uptake and use efficiency
in different varieties of bread wheat (Triticum aestivum L.). Arch. Agron. Soil Sci. 56, 563–572.
Górny, A. G. (1999). Inheritance of nitrogen and phosphorus utilization efficiency in spring
barley at the vegetative growth stages under high and low nutrition. Plant Breed. 118,
511–516.
Górny, A. G., and Garczy nski, S. (2008). Nitrogen and phosphorus efficiency in wild and
cultivated species of wheat. J. Plant Nutr. 31, 263–279.
Gourlay, C. J. P., Allan, D. L., and Russelle, M. P. (1994). Plant nutrient efficiency:
A comparison of definitions and suggested improvement. Plant Soil 158, 29–37.
Gunes, A., Inal, A., Alpaslan, M., and Cakmak, I. (2006). Genotypic variation in phosphorus
efficiency between wheat cultivars grown under greenhouse and field conditions. Soil Sci.
Plant Nutr. 52, 470–478.
Hafeez, F., Aziz, T., Maqsood, M. A., Ahmed, M., and Farooq, M. (2010). Differences in
rice cultivars for growth and phosphorus acquisition from rock phosphate and mono-
ammonium phosphate sources. Int. J. Agric. Biol. 12, 907–910.
Hammond, J. P., Bennett, M. J., Bowen, H. C., Broadley, M. R., Eastwood, D. C.,
May, S. T., Rahn, C., Swarup, R., Woolaway, K. E., and White, P. J. (2003). Changes
in gene expression in Arabidopsis shoots during phosphate starvation and the potential for
developing smart plants. Plant Physiol. 132, 578–596.
Hammond, J. P., Broadley, M. R., and White, P. J. (2004). Genetic responses to phosphorus
deficiency. Ann. Bot. 94, 323–332.
Hammond, J. P., Broadley, M. R., White, P. J., King, G. J., Bowen, H. C., Hayden, R.,
Meacham, M. C., Mead, A., Overs, T., Spracklen, W. P., and Greenwood, D. J. (2009).
Shoot yield drives phosphorus use efficiency in Brassica oleracea and correlates with root
architecture traits. J. Exp. Bot. 60, 1953–1968.
Hedley, M. J., Kirk, G. J. D., and Santos, M. B. (1994). Phosphorus efficiency and the forms
of soil phosphorus utilized by upland rice cultivars. Plant Soil 158, 53–62.
Henry, A., Chaves, N. F., Kleinman, P. J. A., and Lynch, J. P. (2010). Will nutrient-efficient
genotypes mine the soil? Effects of genetic differences in root architecture in common
bean (Phaseolus vulgaris L.) on soil phosphorus depletion in a low-input agro-ecosystem in
Central America. Field Crop Res. 115, 67–78.
Hinsinger, P. (2001). Bio-availability of soil inorganic P in the rhizosphere as affected by root
induced chemical changes: A review. Plant Soil 237, 173–195.
Holloway, B., Bertrand, I., Frischke, A. J., Brace, D., McLaughlin, M., and Shepperd, W.
(2001). Improving fertiliser efficiency on calcareous and alkaline soils with fluid sources
of P, N and Zn. Plant Soil 236, 209–219.
Ismail, A. M., Heuer, S., Thomson, J. T., and Wissuwa, M. (2007). Genetic and genomic
approaches to develop rice germplasm for problem soils. Plant Mol. Biol. 65, 547–570.
Jones, G. P. D., Blair, G. J., and Jessop, R. S. (1989). Phosphorus efficiency in wheat:
A useful selection criterion? Field Crop Res. 21, 257–264.
Jones, G. P. D., Jessop, R. S., and Blair, G. J. (1992). Alternative methods for the selection of
phosphorus efficiency in wheat. Field Crop Res. 30, 29–40.
Kakar, K. M., Tariq, M., Taj, F. H., and Nawab, K. (2002). Phosphorus use efficiency
of soybean as affected by phosphorus application and inoculation. Pak. J. Agron. 1, 49–50.
214 Terry J. Rose and Matthias Wissuwa

Korkmaz, K., Ibrikci, H., Karnez, E., Buyuk, G., Ryan, J., Ulger, A. C., and Oguz, H.
(2009). Phosphorus use efficiency of wheat genotypes grown in calcareous soils. J. Plant
Nutr. 32, 2094–2106.
Lambers, H., Shane, M. W., Cramer, M. D., Pearse, S. J., and Veneklaas, E. J. (2006). Root
structure and functioning for efficient acquisition of phosphorus: Matching morphologi-
cal and physiological traits. Ann. Bot. 98, 693–713.
Li, Y. D., Wang, Y. J., Tong, Y. P., Gao, J. G., and Zhang, J. S. (2005). QTL mapping of
phosphorus deficiency tolerance in soybean (Glycine max L. Merr.). Euphytica 142, 137–142.
Li, M., Welti, R., and Wang, X. (2006). Quantitative profiling of Arabidopsis polar
glycerolipids in response to phosphorus starvation: Roles of phospholipases D z1 and D
z2 in phosphatidylcholine hydrolysis and digalactosyldiacylglycerol accumulation in
phosphorus-starved plants. Plant Physiol. 142, 750–761.
Lott, J. N. A., Ockenden, I., Raboy, V., and Batten, G. D. (2000). Phytic acid and
phosphorus in crop seeds and fruits: A global estimate. Seed Sci. Res. 10, 11–33.
Lott, J. N. A., Ockenden, I., Raboy, V., and Batten, G. D. (2002). A global estimate of
phytic acid and phosphorus in crop grains, seeds, and fruits. In “Food Phytates”
(N. R. Reddy and S. K. Sathe, Eds.), pp. 7–24. CRC Press, Boca Raton.
Lott, J. N. A., Bojarski, M., Kolasa, J., Batten, G. D., and Campbell, L. C. (2009). A review
of the phosphorus content of dry cereal and legume crops of the world. Int. J. Agric.
Resour. Gov. Ecol. 8, 351–370.
Lynch, J. P. (2007). Roots of the second green revolution. Aust. J. Bot. 55, 493–512.
Lynch, J. P., and Ho, M. D. (2005). Rhizoeconomics: Carbon costs of phosphorus acquisi-
tion. Plant Soil 269, 45–56.
Ma, Q., Rengel, Z., and Rose, T. J. (2009). The effectiveness of deep placement of fertilisers
is determined by crop species and edaphic conditions in Mediterranean-type environ-
ments: A review. Aust. J. Soil Res. 47, 19–32.
MacDonald, G. K., Bennett, E. M., Philip, A., Potter, P. A., Ramankutty, N., and
Agronomic phosphorus imbalances across the world’s croplands (2011). Proc. Natl.
Acad. Sci. USA 108, 3086–3091. 10.1073/pnas.1010808108.
Manske, G. G. B., Ortiz-Monasterio, J. I., van Ginkel, M., Gonzalez, R. M., Fischer, R. A.,
Rajaram, S., and Vlek, P. L. G. (2001). Importance of P uptake efficiency versus P utilization
for wheat yield in acid and calcareous soils in Mexico. Eur. J. Agron. 14, 261–274.
Manske, G. G. B., Ortiz-Monasterio, J. I., van Ginkel, M., Rajaram, S., and Vlek, P. L. G.
(2002). Phosphorus use efficiency in tall, semi-dwarf and dwarf near-isogenic lines of
spring wheat. Euphytica 125, 113–119.
Mohamed, G. E. S., and Marshall, C. (1979). The pattern of distribution of phosphorus and
dry matter with time in spring wheat. Ann. Bot. 44, 721–730.
Morcuende, R., Bari, R., Gibon, Y., Zheng, W., Pant, B. D., Blasing, O., Usadel, B.,
Czechowski, T., Mudvardi, M. K., Stitt, M., and Scheible, W. R. (2007). Genome-wide
reprogramming of metabolism and regulatory networks of Arabidopsis in response to
phosphorus. Plant Cell Environ. 30, 85–112.
Oracka, T., and Lapinski, B. (2006). Nitrogen and phosphorus uptake and utilization
efficiency in D(R) substitution lines of hexaploid triticale. Plant Breed. 125, 221–224.
Osborne, L. D., and Rengel, Z. (2002a). Screening cereals for genotypic variation in the
efficiency of phosphorus uptake and utilization. Aust. J. Agric. Res. 53, 295–303.
Osborne, L. D., and Rengel, Z. (2002b). Genotypic differences in wheat for uptake and
utilisation of P from iron phosphate. Aust. J. Agric. Res. 53, 837–844.
Ozturk, L., Eker, S., Torun, B., and Cakmak, I. (2005). Variation in phosphorus efficiency
among 73 bread and durum wheat genotypes grown in a phosphorus-deficient calcareous
soil. Plant Soil 269, 69–80.
Parentoni, S. N., and Júnior, C. L. S. (2008). Phosphorus acquisition and internal utilization
efficiency in tropical maize genotypes. Pesqui. Agropecu. Bras. 43, 893–901.
Internal Phosphorus Utilization Efficiency in Grain Crops 215

Pariasca-Tanaka, J., Satoh, K., Rose, T., Mauleon, R., and Wissuwa, M. (2009). Stress
response versus stress tolerance: A transcriptome analysis of two rice lines contrasting
in tolerance to phosphorus deficiency. Rice 2, 167–185.
Pratt, J., Boisson, A.-M., Gout, E., Bligny, R., Douce, R., and Aubert, S. (2009). Phosphate
(Pi) starvation effect on the cytosolic Pi concentration and Pi exchanges across the
tonoplast in plant cells: An in vivo 31P-nuclear magnetic resonance study using methyl-
phosphonate as a Pi analog. Plant Physiol. 151, 1646–1657.
Raboy, V. (2009). Approaches and challenges to engineering seed phytate and total phos-
phorus. Plant Sci. 177, 281–296.
Rae, A. L., Jarmey, J. M., Mudge, S. R., and Smith, F. W. (2004). Over-expression
of a high-affinity phosphate transporter in transgenic barley plants does not enhance
phosphate uptake rates. Funct. Plant Biol. 31, 141–148.
Raghothama, K. G., and Karthikeyan, A. S. (2005). Phosphate acquisition. Plant Soil 274,
37–49.
Ramaekers, L., Remans, R., Rao, I., Blair, M., and Vanderleyden, J. (2010). Strategies for
improving phosphorus acquisition efficiency of crop plants. Field Crop Res. 117, 169–176.
Reuter, D. J., and Robinson, J. B. (1997). Plant Analysis: An Interpretation Manual.
CSIRO, Collingwood, Victoria.
Richardson, A. E., Lynch, J. P., Ryan, P. R., Delhaize, E., Smith, F. A., Smith, S. E.,
Harvey, P. R., Ryan, M. H., Veneklaas, E. J., Lambers, H., Oberson, A.,
Culvenor, R. A., et al. (2011). Plant and microbial strategies to improve the phosphorus
efficiency of agriculture. Plant Soil 349, 121–156. 10.1007/s11104-011-0950-4.
Römer, W., and Schenk, H. (1998). Influence of genotype on phosphate uptake and
utilization efficiencies in spring barley. Eur. J. Agron. 8, 215–224.
Römer, W., and Schilling, G. (1986). Phosphorus requirements of the wheat plant in various
stages of its life cycle. Plant Soil 91, 221–229.
Rose, T. J., Rengel, Z., Ma, Q., and Bowden, J. W. (2007). Differential accumulation
patterns of phosphorus and potassium by canola cultivars compared to wheat. J. Plant
Nutr. Soil Sci. 170, 404–411.
Rose, T. J., Rengel, Z., Ma, Q., and Bowden, J. W. (2008). Post-flowering supply of P, but
not K, is required for maximum canola seed yields. Eur. J. Agron. 28, 371–379.
Rose, T. J., Rengel, Z., Ma, Q., and Bowden, J. W. (2009). Phosphorus accumulation
by field-grown canola crops and the potential for deep phosphorus placement in a
Mediterranean-type climate. Crop Pasture Sci. 60, 987–994.
Rose, T. J., Pariasca-Tanaka, J., Rose, M. T., Fukuta, Y., and Wissuwa, M. (2010).
Genotypic variation in grain phosphorus concentration; and opportunities to improve
P-use efficiency in rice. Field Crop Res. 119, 154–160.
Rose, T. J., Rose, M. T., Pariasca-Tanaka, J., Heuer, S., and Wissuwa, M. (2011). The
frustration with utilization: Why have improvements in internal phosphorus utilization
efficiency in crops remained so elusive? Front. Plant Nutr. 2, 1–5.
Sahrawat, K. L., Jones, M. P., and Diatta, S. (1997). Direct and residual fertilizer phosphorus
effects on yield and phosphorus efficiency of upland rice in an Ultisol. Nutr. Cycl.
Agroecosyst. 48, 209–215.
Saleque, M. A., Abedin, M. J., Panaullah, G. M., and Bhuiyan, N. I. (1998). Yield and
phosphorus efficiency of some lowland rice varieties at different levels of soil-available
phosphorus. Commun. Soil Sci. Plant Anal. 29, 2905–2916.
Sepehr, E., Malakouti, M. J., Kholdebarin, B., Samadi, A., and Karimian, N. (2009).
Genotypic variation in P efficiency of selected Iranian cereals in greenhouse experiment.
Int. J. Plant Prod. 3, 17–28.
Septiningsih, E. M., Pamplona, A. M., Sanchez, D. L., Neeraja, C. N., Vergara, G. V.,
Heuer, S., Ismail, A. M., and Mackill, D. J. (2009). Development of submergence-
tolerant rice cultivars: The Sub1 locus and beyond. Ann. Bot. 103, 151–160.
216 Terry J. Rose and Matthias Wissuwa

Sharpley, A. N., McDowell, R. W., and Kleinman, P. J. A. (2001). Phosphorus loss from
land to water: Integrating agricultural and environmental management. Plant Soil 237,
287–307.
Shen, J., Yuan, L., Zhang, J., Li, H., Bai, Z., Chen, X., Zhang, W., and Zhang, F. (2011).
Phosphorus dynamics: From soil to plant. Plant Physiol. 156, 997–1005.
Shenoy, V. V., and Kalagudi, G. M. (2005). Enhancing plant phosphorus use efficiency for
sustainable cropping. Biotechnol. Adv. 23, 501–513.
Siddiqi, M. Y., and Glass, A. D. (1981). Utilization index: A modified approach to the
estimation and comparison of nutrient utilization efficiency in plants. J. Plant Nutr. 4,
289–302.
Stangoulis, J. C. R., Huynh, B., Welch, R. M., Choi, E., and Graham, R. D. (2007).
Quantitative trait loci for phytate in rice grain and their relationship with grain micronu-
trient content. Euphytica 154, 289–294.
Stelling, D., Wang, S. H., and Römer, W. (1996). Efficiency in the use of phosphorus,
nitrogen and potassium in topless faba bean (Vicia faba L.)—Variability and inheritance.
Plant Breed. 115, 361–366.
Su, J. Y., Xiao, Y., Li, M., Liu, Q., Li, B., Tong, Y., Jia, J., and Li, Z. (2006). Mapping
QTLs for phosphorus-deficiency tolerance at wheat seedling stage. Plant Soil 281, 25–36.
Su, J. Y., Zheng, Q., Li, H. W., Li, B., Jing, R. L., Tong, Y. P., and Li, Z. S. (2009). Detection
of QTLs for phosphorus use efficiency in relation to agronomic performance of wheat
grown under phosphorus sufficient and limited conditions. Plant Sci. 176, 824–836.
Subbarao, G. V., Ae, N., and Otani, T. (1997). Genetic variation in acquisition, and
utilization of phosphorus from iron-bound phosphorus in pigeonpea. Soil Sci. Plant
Nutr. 43, 511–519.
Theodorou, M. E., and Plaxton, W. C. (1993). Metabolic adaptations of plant respiration to
nutritional phosphate deprivation. Plant Physiol. 101, 339–344.
Theodorou, M. E., and Plaxton, W. C. (1996). Purification and characterization of
pyrophosphate-dependent phosphofructokinase from phosphate-starved Brassica nigra
suspension cells. Plant Physiol. 112, 343–351.
Vance, C. P., Uhde-Stone, C., and Allan, D. L. (2003). Phosphorus acquisition and use: Critical
adaptations by plants securing a non-renewable resource. New Phytol. 157, 423–457.
Wang, Q. R., Li, J. Y., Li, Z. S., and Christie, P. (2005). Screening Chinese wheat
germplasm for phosphorus efficiency in calcareous soils. J. Plant Nutr. 28, 489–505.
Wang, X., Wang, Y., Tian, J., Lim, B. L., Yan, X., and Liao, H. (2009). Overexpressing
AtPAP15 enhances phosphorus efficiency in soybean. Plant Physiol. 151, 233–240.
Wang, X., Shen, J., and Liao, H. (2010). Acquisition or utilization, which is more critical for
enhancing phosphorus efficiency in modern crops? Plant Sci. 179, 302–306.
White, P. J., and Hammond, J. P. (2008). Phosphorus nutrition of terrestrial plants. In “The
Ecophysiology of Plant–Phosphorus Interactions” (P. J. White and J. P. Hammond,
Eds.), pp. 51–81. Springer, The Netherlands.
Wissuwa, M. (2003). How do plants achieve tolerance to phosphorus deficiency? Small
causes with big effects. Plant Physiol. 133, 1947–1958.
Wissuwa, M., and Ae, N. (1999). Molecular markers associated with phosphorus uptake and
internal phosphorus-use efficiency in rice. In “Plant Nutrition—Molecular Biology and
Genetics” (G. Gissel-Nielsen and A. Jensen, Eds.), pp. 433–439. Kluwer Academic
Publishers, The Netherlands.
Wissuwa, M., and Ae, N. (2001). Genotypic variation for tolerance to phosphorus defi-
ciency in rice and the potential for its exploitation in rice improvement. Plant Breed. 120,
43–48.
Wissuwa, M., Yano, M., and Ae, N. (1998). Mapping of QTLs for phosphorus-deficiency
tolerance in rice (Oryza sativa L.). Theor. Appl. Genet. 97, 777–783.
Internal Phosphorus Utilization Efficiency in Grain Crops 217

Wissuwa, M., Gamat, G., and Ismail, A. (2005). Is root growth under phosphorus deficiency
affected by source or sink limitations? J. Exp. Bot. 56, 1943–1950.
Wissuwa, M., Mazzola, M., and Picard, C. (2009). Novel approaches in plant breeding for
rhizosphere-related traits. Plant Soil 321, 409–430.
Yang, M., Ding, G., Shi, L., Xu, F., and Meng, J. (2011). Detection of QTL for phosphorus
efficiency at vegetative stage in Brassica napus. Plant Soil 339, 97–111.
Yaseen, M., and Malhi, S. S. (2009a). Variation in yield, phosphorus uptake, and physiolog-
ical efficiency of wheat genotypes at adequate and stress phosphorus levels in soil.
Commun. Soil Sci. Plant Anal. 40, 3104–3120.
Yaseen, M., and Malhi, S. S. (2009b). Differential growth response of wheat genotypes
to ammonium phosphate and rock phosphate phosphorus sources. J. Plant Nutr. 32,
410–432.
Zhang, D., Cheng, H., Geng, L. Y., Kan, G. Z., Cui, S. Y., Meng, Q. C., Gai, J. Y., and
Yu, D. Y. (2009). Detection of quantitative trait loci for phosphorus deficiency tolerance
at soybean seedling stage. Euphytica 167, 313–322.
Zhao, J., Jamar, D. C. L., Lou, P., Wang, Y., Wu, J., Wang, X., Bonnema, G.,
Koornneef, M., and Vreugdenhil, D. (2008). Quantitative trait loci analysis of phytate
and phosphate concentrations in seeds and leaves of Brassica rapa. Plant Cell Environ. 31,
887–900.
Zhu, J., and Lynch, J. P. (2004). The contribution of lateral rooting to phosphorus acquisi-
tion efficiency in maize (Zea mays) seedlings. Funct. Plant Biol. 31, 949–958.
 C H A P T E R S I X

Computer Simulation
in Plant Breeding
Xin Li,* Chengsong Zhu,* Jiankang Wang,† and Jianming Yu*

Contents
1. Introduction 220
1.1. An urgent need in plant breeding 220
1.2. Computer simulation 221
1.3. Joining computer simulation with plant breeding 222
2. Applying Computer Simulation to Plant Breeding 224
2.1. Comparing breeding methods 225
2.2. Gene mapping 237
2.3. Gene network and genotype-by-environment interaction 246
2.4. Crop physiology and crop modeling 247
3. Computation and Software Issues 254
4. Summary and Perspectives 255
Acknowledgments 257
References 257

Abstract
As a bridge between theory and experimentation, computer simulation has
become a powerful tool in scientific research, providing not only preliminary
validation of theories but also guidelines for empirical experiments. Plant
breeding focuses on developing superior genotypes with available genetic
and nongenetic resources, and improved plant-breeding methods maximize
genetic gain and cost-effectiveness. Computer simulation can lay out the
breeding process in silico and identify optimal candidates for various scenarios;
empirical validation can then follow. Insights gained from empirical studies, in
turn, can be incorporated into computer simulations. In this review, we discuss
different applications of computer simulation in plant breeding. First, we briefly
summarize the history of plant breeding and computer simulation and how
computer simulation can facilitate the breeding process. Next, we partition the
utility of computer simulation into different research areas of plant breeding,
including breeding method comparison, genetic mapping, gene network and

* Department of Agronomy, Kansas State University, Manhattan, Kansas, USA

{
Institute of Crop Science and CIMMYT China, Chinese Academy of Agricultural Sciences, Beijing, China

Advances in Agronomy, Volume 116 # 2012 Elsevier Inc.

ISSN 0065-2113, DOI: 10.1016/B978-0-12-394277-7.00006-3 All rights reserved.

219
220 Xin Li et al.

genotype-by-environment interaction simulation, and crop modeling. Then we

discuss computational issues involved in simulation. Finally, we offer some
perspectives on the future of computer simulation in plant breeding.

1. Introduction
1.1. An urgent need in plant breeding
1.1.1. Three eras in plant-breeding history
What is plant breeding? A definition in Bernardo (2002) captured its
essence: “Plant breeding is the science, art, and business of improving plants
for human benefit.” Plant breeding is a discipline initiated by our ancestors
tens of thousands years ago. Without any knowledge of biology or genetics,
our ancestors provided domesticated crops for us today based only on
phenotypic selection (PS), or “art.” But scientific elements are also involved
in this process because the superior characteristics can be inherited from
generation to generation. Crossing “good” by “good,” chances are we get
better progeny next season. Comparing the crops we have today with their
wild relatives, we can see what great progress our ancestors achieved, even
though it took a time as long as civilization itself.
In the early twentieth century, with the rediscovery of Mendel’s Law and
the milestone paper of R. A. Fisher, plant breeding entered the scientific era.
Geneticists began to investigate the underlying principles of plant improve-
ment, especially for quantitative traits, which are the major features of most
economically important characteristics such as yield, biotic and abiotic stress
resistance, and flowering time. Classical textbooks of Mather (1949), Allard
(1960), and Falconer (1960) provide systematic guidelines for plant breeding.
With scientific theories in hand, plant breeders made significant progress in
answering questions pertinent to the plant-breeding process: Which parents
should I use to get the most variance and simultaneously a higher mean value?
Which is the best, single, double, or triple cross? How many generations does
it take to get the desired cultivar or population? What about the selection
intensity? Early testing or late testing? How do I control genotype-by-
environment interaction (GEI)? The complexity of the mechanisms of quan-
titative traits, unsurprisingly, means that these questions have no absolute
answers, which is why plant breeding is often called “a numbers game.” The
more plants, the more likely we can find a superior genotype. But exact
answers do not exist; they depend on the type of crops, populations, and most
importantly, the genetics of the traits. Probabilities are intrinsic to all of the
elements controlling the final result.
The third era came along with the discovery of molecular markers.
Although phenotypic or biochemical markers can also be used indirectly
to select the desired traits, most of which are difficult or expensive to
Computer Simulation in Plant Breeding 221

measure; they are not diverse enough and are too tedious and expensive to
implement into plant-breeding practice. Molecular markers (AFLF, RFLP,
RAPD, SSR, SNP, etc.) are abundant in organisms and easy to apply to
selection and gene mapping (Mackay, 2001). Thanks to those molecular
makers, today we have a better understanding of the genetic mechanism
controlling many quantitative traits, both in humans and plants. At the
beginning of the second decade of the twenty-first century, we are
equipped with powerful tools such as the high-throughput sequencing
platform and bioinformatics to help us tap into research areas impossible
only decades ago.

1.1.2. A gap between our knowledge and plant-breeding practice

With such data explosion, the more information we have, the more over-
whelmed we feel. We have to acknowledge that a gap exists between the
huge amount of information and plant-breeding practice. Plant breeding is a
time-consuming and resource-intensive process, during which decision-
making and resource allocation are critical to the final result. How to reduce
trial and error and exploit those results from genetic research is a new task
for crop scientists. Meanwhile, we face food and fuel crises as well as climate
change and other biotic and abiotic stresses that can damage crop produc-
tion. Thus, we need innovative tools to help us integrate the knowledge we
have gained on plant breeding during the past three eras into plant-breeding
practices that allow us to tackle those problems and accelerate plant
improvement. Computer simulation is one of these tools.

1.2. Computer simulation

1.2.1. Origin of computer simulation
The origin of computer simulation can be traced to the invention of
computers. After the first digital computer was invented in 1945, John
von Neumann, one of the greatest mathematicians in modern history,
with the help of Nicholas Metropolis and Stanislaw Ulam, investigated a
computational model of thermonuclear reaction (Winsberg, 2010). Today,
computer simulation is the third component of the scientific research triad,
along with theoretical and experimental approaches. Computer simulation
allows us to do pilot research, speed research and development, transfer
discoveries from the laboratory to practice, and so on.
The essence of computer simulation is to sample as many conditions as
possible that might be encountered in practice, which may be infinite or
astronomically large (e.g., selecting 100,000 random numbers, which is
nearly impossible by hand). Computers make such sampling and calcula-
tion fast and easy. By subjecting our models to this representative sam-
pling, we can have some confidence in the performance of our models in
actual practice.
222 Xin Li et al.

1.2.2. Classification of computer simulation

Computer simulations can be classified in many different ways. From the
viewpoint of output, computer simulations can be deterministic or stochas-
tic. The results from deterministic simulation are clearly defined; that is, one
input value corresponds to a single output. On the other hand, the outputs
from stochastic simulation are probabilistic, or a distribution around the true
value. From the perspective of models used, computer simulation can be
classified into (1) iconic, such as the model used in clinical simulation to
train surgeons; (2) analogous, such as treating electricity current as water
flow in the pipe; and (3) symbolic, using a mathematical model, an abstract
replication of the real system, to analyze a problem. Most simulations
relevant to plant breeding are the third type, also termed Monte Carlo
simulation.

1.2.3. An example
To illustrate the process and components involved in computer simulation,
we provide a simple computer simulation to sample progeny phenotypic
values (Fig. 1A). Two parental inbred lines are crossed to form segregating
populations (for simplicity, only F1 and F2 are shown). In a completely
additive genetic model, the phenotypic mean of the F1 generation should
be the mean of the two parental lines. In this model, what we can control
are the overall mean and genetic effects. Other nongenetic effects are
assumed to follow a normal distribution with a mean of zero and predefined
variance. Computer simulation can be used to randomly sample nongenetic
effects and add them into the genetic model. Other subsequent generations,
such as BC1, DH, or RIL, also can be generated when recombination is
simulated to control for genetic effects.
Thanks to this random sampling and repeat feature of computer simula-
tion, the models used to simulate a real system are good approximations
of what would happen in reality. That is also why plant breeding, which
is time-consuming and resource-intensive, can benefit from computer
simulation.

1.3. Joining computer simulation with plant breeding

Because computer simulation is fast and uses few physical resources, we can
easily see why simulations and plant breeding are compatible. Computer
simulation can accommodate genetic models with multiple genes, multiple
alleles, pleiotropic and epistatic effects, and GEI. An analytical approach
makes it difficult to account for parental selection, pedigree relationships,
linkage, and recombination, but a simulation experiment can easily accom-
modate all of these and more. Thus, simulation can be a valuable tool for
breeders to find the most efficient path to the target cultivar using parental
Computer Simulation in Plant Breeding 223

A B
P1 ⫻ P2

P1 F1 P2 F1
F2
Density

F3
F2 F4
F5
F6
F7
F8
Phenotype value
Computer simulation Plant breeding

C 90
Genotype value

80
(% of max)

70
H=0.1 ES
H=0.1 SSD
H=0.5 ES
H=0.5 SSD
60

50
0 100 200 300 400
Number of F2 plants
Comparative simulation of ES and SSD

Figure 1 Joining computer simulation with plant breeding. (A) Computer simulation
can be used to generate random numbers from the underlying distribution. One
random sample is an approximation of what would happen in reality. (B) The aim of
plant breeding is to create superior genotypes for new cultivars. The more plants are
available for selection, the more likely we are to obtain superior genotypes. A quantita-
tive assessment can be very helpful for directing plant breeding. (C) Applying computer
simulation to plant breeding can integrate the number of plants, heritability, number of
genes, selection method and intensity, and number of generations to predict possible
outcomes of a breeding program and help us find the optimal combination of these
factors to accelerate the breeding process. The graph in (C) is based on data from
Vanoeveren and Stam (1992).

selection, predicting cross performance, comparing selection strategies,

breeding by design, etc. (Peleman and van der Voort, 2003; Wang and
Pfeiffer, 2007). From the perspective of plant breeding, finding the opti-
mum breeding strategy and reducing the breeding cycle are very helpful.
This leads us to the questions asked in the first part of this review. Computer
simulation, by predicting what will happen based on a model, can provide
answers to those questions (Fig. 1).
224 Xin Li et al.

In this review, we focus on the various applications of computer simu-

lation in plant breeding and summarize how computer simulation can take
advantage of results from genetic research to facilitate plant improvement.

2. Applying Computer Simulation to

Plant Breeding
Based on the literature, the major applications of computer simulation
can be partitioned into four areas: (1) breeding method comparison: finding
the best breeding scheme taking account of multiple factors; (2) gene
mapping: validating the effectiveness of new mapping methods, calculating
LOD score and confidence interval (CI); (3) plant-breeding platforms:
integrative simulation programs that can simulate the whole plant-breeding
process; and (4) crop modeling: combining crop models, genetic architec-
ture of traits, and environmental information to fill the gap between
genotype and phenotype (Fig. 2).

Breeding method
• Compare breeding strategy
• Assess factors influencing
marker-assisted selection

Gene mapping
Crop modeling
• Assess factors influencing Computer • Use gene information as
mapping power simulation in model parameters
• Determine significance
plant breeding • Predict crop performance in
threshold and confidence
target environments
interval of QTL position

Gene network and GEI

• Combine genetic and
genotype-by-environment
interaction to simulate the
whole plant breeding process

Figure 2 Applications of computer simulation in plant breeding. Computer simulation

can be used to (1) compare breeding methods, (2) compare gene mapping strategies and
calculate significance threshold and confidence interval, (3) simulate gene networks and
genotype-by-environment interactions, and (4) simulate crop growth using crop mod-
els to assess the influence of genes, environment, and climate change, among other
things.
Computer Simulation in Plant Breeding 225

2.1. Comparing breeding methods

In most situations, extensive experimental evaluation of breeding strategies
for manipulating complex yield adaptation traits is impractical or beyond the
resources of breeding programs. Therefore, simulation tools are important
in designing and testing breeding strategies (Chapman et al., 2003), and this
role becomes increasingly important as our understanding of the genetic
architecture of quantitative traits improves.

2.1.1. Finding the best breeding scheme

The aim of plant breeding is to find superior cultivars from available genetic
resources. Evaluating the performance of a breeding method or comparing
the efficiencies of different breeding methods within an ongoing breeding
program is difficult, cumbersome, and expensive. Also, depending only on
field trials to compare different breeding strategies is risky. A single field trial
is a random sample of all possible outcomes; thus, random effects could
misrepresent the actual situation. Computer simulation, on the other hand,
uses a large number of replications to provide a reliable probabilistic state-
ment of different schemes. Simulation models also allow us to test a range of
input parameter values and subsequently provide data so that we can assess
the influences of these parameters and their value ranges on the efficiency of
selection (Vanoeveren and Stam, 1992). A remarkable utility of computer
simulation in plant breeding is combining knowledge from quantitative
genetics, molecular genetics, and plant breeding to predict the efficiency of
different selection methods (Cooper et al., 2007).
Vanoeveren and Stam (1992) compared two selection strategies for self-
pollinated crops: early selection (ES) and single seed descent (SSD). They
investigated the effects of heritability, number of F2 plants, segregating loci,
number of crosses, and effects of dominance on the phenotypic mean of the
best 10 plants in final generation (F6 in ES and F7 in SSD). Simulation
results showed that for single crosses, the advantage of ES over SSD is largest
when heritability is high and the number of F2 plants is low. If the traits are
controlled by more loci, ES has less advantage over SSD because of the
difficulty of accumulating all positive alleles into a single genotype. How-
ever, for multiple crosses, except with high heritability, the advantage of ES
is lost because of selection errors between cross-selection. Effect of domi-
nance is not important in SSD, because at higher generations of SSD
approach, a higher homozygosity level is reached. But for ES, the effect is
not so explicit because dominance could increase both variances of cross and
level of heterozygosity, so the overall effect is trivial. When the total cost is
considered, the SSD method is cheaper and needs only a small area.
Wang and Pfeiffer (2007) used QU-GENE (plant-breeding simulation
software; see Section 3) to compare two plant-breeding schemes, modified
pedigree/bulk selection method (MODPED), and selected bulk selection
226 Xin Li et al.

method (SELBLK), in CIMMYT’s wheat breeding program. The simula-

tion result showed that genetic gain from SELBLK was on average 3.9%
higher than from MODPED, and genetic gain adjusted by target genotypes
from SELBLK was on average 3.3% higher than MODPED for a wide
range of genetic models. More crosses are retained with SELBLK than
MODPED, and from F1 to F8, SELBLK requires one-third less land than
MODPED and produced fewer families, which means SELBLK is more
cost-effective. Wang and Pfeiffer (2007) also used computer simulation to
find the best trade-off when alleles have pleiotropic effects on target traits. In
their research, rice quality, for example, QTL M35 has a negative effect on
both ACE (area of chalky endosperm) and AC (amylose content). Simula-
tion results showed that under these conditions, the best genotype would
have an ACE value of 9.2% (rather than 0) and an AC value of 17.73%
(rather than the theoretical value of 22.3%).

2.1.2. Factors influencing marker-assisted selection

Because marker and QTL information used in MAS is the output from QTL
mapping studies, whatever influences the results of QTL mapping will have a
direct effect on the results of marker-assisted selection (MAS; for plant
breeding, we can bypass the mapping step and use a regression-based method
because gene discovery is not necessary). The difference between QTL
mapping and MAS is that, most of the time, QTL is detected in a single
population from a specific generation. MAS, on the other hand, uses the
outcome from QTL mapping for the whole breeding process—multiple
generations, populations, and environments. In other words, QTL mapping
analysis is a more static process, and MAS is more dynamic. In this section, we
focus on how factors such as population size, heritability, and marker density
affect selection response.
The efficiency of MAS, an indirect selection method, is determined by the
genetic correlation between the target traits and the markers used for selec-
tion. This correlation is determined by the strength of linkage disequilibrium
(LD) between the markers and the QTL as well as the percentage of additive
genetic variance explained by the markers ( Johnson, 2004). Theoretical
studies have shown that heritability and the effect of the QTL linked to the
markers can influence the efficiency of MAS (Lande and Thompson, 1990).
Zhang and Smith (1993) investigated other factors that could influence
the outcome of MAS. They considered method and population size to
estimate marker effects, average marker effects versus individual estimates,
size of selection population, and regenerating LD by crossing selected lines.
Least square method yielded a much smaller response than Mixed Model
because least square method overestimated the marker effects and thus had
more prediction errors; however, the two differed little if QTL effects were
large. Large population sizes led to more accurate QTL effects and higher
selection response. The response achieved using the average of the best
Computer Simulation in Plant Breeding 227

marker effects was less than using individual effects. With an effective
population size of about 60, selecting large effect markers could not gener-
ate significant genetic drift, so crossing the selected lines to regenerate LD is
not effective in producing further genetic response. Zhang and Smith
suggested that the issue of decreased LD with selection could be solved by
identifying markers that are themselves the target QTL.
Population size was important in determining the efficiency of MAS
(Gimelfarb and Lande, 1994) for two possible reasons. First, with a large
population, the estimation of QTL effects is more accurate; second, finding
the desirable genotype is more likely with a large population. For a popula-
tion of at least 100 individuals of each sex, MAS is most efficient if the
phenotype does not contribute to the selection index. When several QTLs
are linked in coupling phase, the regression method may detect the QTLs as
a block, but when QTLs are linked in repulsion, the combined effects of
these QTLs decrease, making detection of the underlying QTLs difficult.
The optimal number of markers depends on the degree of LD between the
marker and QTL. LD is determined by the average number of recombinations
per generation in that region of the genome, the number of generations since
the original mutation, and the population size (Mackay, 2001). If LD is large,
more markers will not necessarily lead to a higher selection response. This
result occurs for two possible reasons. First, when LD is large, multiple makers
are in LD with each other, creating colinearity, meaning that their effects
overlap; second, covariance in these markers makes estimating QTL effects
inaccurate, thus decreasing the efficiency of MAS. Increasing the number of
markers in a finite population could even result in spurious linkages between
markers and unlinked QTLs (Gimelfarb and Lande, 1995).
For most MAS experiments, the effect of QTL is fixed at the beginning
of a breeding program; however, because of recombination between
marker and QTL, epistasis, or gene-by-environment interaction, the con-
tribution of QTL to a specific trait is not necessarily the same over the entire
breeding process. Some researchers (Gimelfarb and Lande, 1994; Peccoud
et al., 2004; Podlich et al., 2004) suggest reevaluating the QTL effect every
generation would be more efficient. This also highlights the dynamic
property of breeding methods involving MAS.
Podlich et al. (2004) proposed a method called mapping as you go
(MAYG) to tackle the issue of context dependency. Because plant breeding
is a dynamic process, the QTL values used in MAS would change from
generation to generation (or cycle to cycle) because of gene-by-gene and
gene-by-environment interaction. Computer simulation based on the E(N:
K) model (see Section 3) showed that compared with the mapping start only
(MSO) approach, which keeps the initial QTL values unchanged in the
breeding process, the MAYG method led to higher genetic gains in almost
all simulated scenarios. When an additive model was assumed (both E and K
are zero), these two methods do not differ significantly, but when epistasis
228 Xin Li et al.

or GEI effects occurred, the MAYG approach has the advantage because, in
different genetic backgrounds or different environments, the genetic effect
of a QTL can be distinct or even reversed. The MAYG approach can take
this into account and update the gene effect in every cycle or generation.
Thus, the gene effects remain relevant to the current breeding population.

2.1.3. Phenotypic selection versus marker-assisted selection

A debate has persisted for many years about the advantages of MAS over PS.
The fundamental question in this debate is whether the markers used in
MAS can represent the genetic structure of a complex trait. Using computer
simulation, Edwards and Page (1994) found that MAS is better than con-
ventional PS only in the first 2–3 years of recurrent selection. Thus, they
proposed combining MAS in early generations with PS in later generations.
The number of QTLs and the recombination frequency between marker
and QTL are the two most important factors affecting the response to MAS.
Another simulation experiment by Hospital et al. (1997) reached similar
conclusions. In addition, they found that MAS had the disadvantage of more
unstable variability at low heritability; using multiple regression of pheno-
type on marker genotype, a higher Type I error can increase the efficiency
of MAS. This strategy entails the trade-off of including more false positives
in exchange for higher coverage of potential markers or QTLs that do have
functional effects on the desired phenotype. van Berloo and Stam (1998)
investigated the effects of trait heritability, selection intensity, genetic archi-
tecture, and uncertainty in QTL mapping on the results of MAS. They
found that MAS is advantageous because it overworks better than PS when
dominant alleles at QTLs are present and linked in coupling phase. Inaccu-
racy in QTL position reduces the effectiveness of MAS.
Conventional PS is based on the phenotype of one individual, not
including any phenotypic information from other individuals in a pedigree.
Best linear unbiased prediction (BLUP) is a general procedure that allows
comparisons of phenotypes among genotypes developed from different
breeding populations and evaluated in different sets of environments
(Bernardo, 2002). Computer simulation allows us to compare BLUP and
MAS. Bernardo (1999a) used computer simulation to compare the effec-
tiveness of BLUP based on trait data alone (T-BLUP) and trait and marker
data combined (TM-BLUP). The correlation between predicted and true
performance of untested single crosses ranged from 0.60 for traits with
lower heritability to 0.96 when the heritability is 1. The usefulness of
TM-BLUP for predicting single cross performance increased as the propor-
tion of QTL with known linkage to flanking markers approached p¼1, and
the map distance between a QTL and its flanking markers increased from
2.5 to 10cM. The advantage of TM-BLUP over T-BLUP decreased as
heritability and number of QTLs controlling the trait increased.
Computer Simulation in Plant Breeding 229

Zhang and Smith (1992) compared the selection efficiency of MAS

(based on marker information only), conventional BLUP, and a combina-
tion of these two methods (COMB). Their simulation design used only
markers with large effects because markers with minor effects may add noise
to the estimation of marker effects in the regression procedure, a conclu-
sion Bernardo (2001) also reached. Their simulation showed that conven-
tional BLUP led to higher genetic gains than MAS, but MAS contributed to
even higher selection response in the COMB method. As in other studies,
the authors also suggested reevaluating marker effects after a number of
generations. The LD between marker and QTL changes because of recom-
bination, random genetic drift, and selection in the breeding process. Thus,
reevaluating the association every few generations is necessary (Lande and
Thompson, 1990).
Based on gene information (such as number of genes controlling traits,
their positions in the genome, their genetic effects, and epistasis and gene-
by-environment interaction), selecting parental material could be based
on simulation of all the possible crosses between parental lines and pre-
dicting the corresponding progeny performance (Wang et al., 2005). van
Berloo and Stam (1998, 2001) used computer simulation to investigate the
efficiency of MAS in selecting parent inbred lines that could lead to the
best single genotype containing as many desirable alleles as possible. Again,
MAS is favored when heritability is low (0.1 or 0.3). At higher heritability,
MAS has fewer advantages over PS, even disappearing when heritability is
0.9. MAS is not very sensitive to incorrectly mapped or missing QTLs.
Even when only 25% of the QTLs are mapped, the selection response with
MAS is still 4% higher than PS. In selecting for multiple traits, however,
the difference is dramatic. For selection in an RIL population, only when
more than 80% of the QTLs were mapped could MAS outperform PS.
Selection response was higher when other parental populations such as
BC1 or F2 were used because MAS could capture the genetic diversity in
these populations.
With pedigree structure information and field trial data in hand, BLUP
can also be used to select breeding populations with the highest mean
performance (Bernardo, 2003). Simulation results showed that, unlike the
number of breeding populations and the size of each population, the ability
to identify the best breeding population is more important to obtain higher
genetic gain in the RIL population. With effective parental selection, a large
number of breeding populations (not a large population size) could lead to
higher selection response.
MAS has not been widely used for improving polygenic traits because
QTL mapping remains insufficiently precise and because QTL information
cannot be easily extrapolated from mapping populations to breeding popu-
lations. MAS will remain a specialized breeding tool until QTL mapping
can estimate breeding values across many diverse breeding crosses and
230 Xin Li et al.

subpopulations such as those in typical plant-breeding programs (Holland,

2004). Lande and Thompson (1990), in their seminal paper, pointed out
three issues that could slow the implementation of MAS: (1) the number of
molecular makers needed to capture as many additive genetic effects as
possible, (2) sample size needed to detect traits with low heritability, and
(3) sampling errors in estimating relative weights in the selection index
combining molecular and phenotypic information.
Using a large population size could ameliorate these three problems.
Increasing sample size can improve estimates of gene effects, especially for
genes with small effects, but this can simultaneously increase the efficiency
of T-BLUP PS. Some researchers thus recommend ignoring genes with
smaller effects when using gene information to select the best hybrids or
inbred lines. In a similar paper, Bernardo and Charcosset (2006) investigated
the usefulness of gene information in marker-assisted recurrent selection.
Simulation results showed that if the trait is controlled by only 10 genes, the
more gene information we have, the higher selection response we can
obtain; however, when 40 or 100 genes control the target trait, including
all genes in the model provides the lowest selection response. In this case
(large number of QTLs), PS is always better than MAS. The authors also
compared the differences of flanking marker and markers of the QTL per se.
The conclusion is that using markers for the QTLs per se is always better than
using flanking makers, but this advantage is not significant when a small
number of QTLs control the trait because flanking makers can fix the
favorable allele in a few generations, leaving little room to further increase
favorable alleles through markers for QTLs themselves.
From all this information, we can see neither MAS nor PS is perfect, so
using an approach, combining these two (using MAS in early generation to
select large effects and thus significantly reduce the number of individuals
that must be phenotyped in later generations to capture genes with minor
effects) may be more effective (Wang et al., 2009a). Another possibility is to
detect significant markers in an environment in which the trait has a high
heritability, thus ensuring the estimates of QTL or marker effect are accu-
rate. This marker information can be subsequently used in other environ-
ments where heritability is low, perhaps in an off-season nursery that is not
the same as the target environment.
To date, several criteria determine whether to use MAS instead of PS:
(1) the target trait is controlled by a few genes with large effects,
(2) shortening the breeding cycle is one priority, and (3) PS is difficult. As
our knowledge of various biological phenomena increases, however, the
shortcomings of MAS may be overcome. For example, the expense of
genotyping has decreased significantly because of the high-throughput
genotyping platform, and we can use polymorphisms within the genes as
markers (functional markers) to monitor the transmission of the target gene
(s) from generation to generation.
Computer Simulation in Plant Breeding 231

2.1.4. Marker-assisted recurrent selection

Marker-assisted recurrent selection, or MARS, refers to improving an F2
population by 1 cycle of MAS (i.e., based on phenotypic data and marker
scores) followed by 3 cycles of marker-based selection (i.e., based on marker
scores only; Bernardo and Charcosset, 2006). Marker-assisted recurrent
selection should enrich favorable genes in a single cultivar. The principle
is to generate gene combinations as quickly as possible in the recurrent
selection process, selecting superior plants and intercrossing them. This
process may take several generations. The result is similar to backcross (see
next section 2.1.5); the difference is that for MARS, the goal is to combine
favorable genes from both parents into a single cultivar. On the other hand,
backcross introgresses a single gene or a limited number of favorable genes
from the donor parent into the recipient parent and eliminates as much of
the other donor parent’s genome as possible. Population size is important in
recurrent selection because theoretically, with an unlimited population size,
we can always get the ideal cultivar in one generation! Unfortunately, this is
impossible in real breeding programs with an upper limit of thousands of
plants. Charmet et al. (1999) assessed the effects of number of QTLs (both
additive and epistatic QTLs) and CIs for QTL location on the population
size needed to obtain the best line with 95% confidence. Their results
showed that the accuracy of QTL location greatly influenced the population
size needed. With no interaction between 10 QTLs that control the trait as
well as a CI of 10cM for QTL location, the minimum number of plants
needed is 8669; when the CI increases to 30cM, the number is nearly
70,000. In practice, a population size of 200 would probably give us a near-
ideal cultivar.
For recurrent selection, the selection intensity should be high for short-
term selection and low for long-term selection. Bernardo et al. (2006)
provided a rule of thumb for the number of individuals selected in each
generation: Nsel. Nsel should approximately equal the number of cycles for
which selection would be conducted. Assortative mating (crossing the pair
of individuals with the highest and second-highest phenotypic or genotypic
values and so on) between selected individuals, either based on marker
scores or phenotype, could lead to even higher selection response in the
short term, but higher selection intensity is also coupled with faster decrease
in genetic variance (Bernardo, 1999b).

2.1.5. Marker-assisted backcross

Using marker information to select a quantitative trait depends on the
extent to which we understand the target trait; however, as mentioned,
deciphering genes and their interactions in controlling a quantitative trait is
a daunting task. Marker-assisted introgression of a single or a few genes with
major effects could bypass this problem. Because this type of introgression
232 Xin Li et al.

involves genomic composition from two distinct sources (donor or recipi-

ent/recurrent genomes) of the species, breeds, strains, lines, or populations,
theoretical predictions can be quite realistic (Hospital et al., 1992).
Marker-assisted backcross (MABC) is routinely used in plant breeding
for gene introgression (Frisch and Melchinger, 2005). The essence of
this method is selecting a target gene (foreground selection) coupled with
recovery of a recurrent parent genome (RPG; background selection). The
problem of linkage drag is a great concern for backcrossing (Hospital, 2001).
Most simulations focus on the selection of foreground and background
chromosome segments under different marker densities, number of genes
or QTLs introgressed, and generations needed to recover the RPG. Com-
puter simulation shows that the most important parameter is the distance
between the introgressed gene and the flanking markers, which should be as
closely linked to the introgressed gene as possible (Hospital, 2001).
Simulations (Hospital and Charcosset, 1997) show that three markers,
one at the center of a QTL CI and two flanking markers, are necessary to
show the existence of a target gene after several backcross generations.
Population size is proportional to the number of genes to be introgressed.
For example, to introgress one QTL with CI of 40cM with two flanking
markers, the minimum number of individuals needed to obtain at least one
target genotype is 9. But when the QTL number increases to six, the
minimum number of individuals needed is 901. So Hospital and Charcos-
set’s results suggest that no more than four QTLs should be manipulated at
once. Background selection is more effective when applied after foreground
selection because, if foreground selection comes first, background selection
requires fewer individuals to be genotyped. This can save one or two
generations in reaching the level of recipient genome.
Visscher et al. (1996a) found that using a marker spacing of 10–20cM
could save one to two generations compared with random selection or PS.
For precisely mapped QTLs, flanking markers should cover 10–20cM
around the estimated position to maintain the linkage between marker
and QTL. For less accurately mapped QTLs, a whole segment containing
the target gene should be introgressed, but in that case, more deleterious
genes might be introduced into the elite lines, thus decreasing selection
efficiency.
Frisch et al. (1999) showed computer simulation results that compared
with a constant population size across all generations; increasing population
sizes from generation BC1 to BC3 reduced the number of required marker
data points (MDP) by as much as 50% without affecting the proportion of
the RPG. A four-stage selection approach, selecting in the first generations
for recombinants on the carrier chromosome of the target allele, reduced
the required number of MDPs by as much as 75% over a selection index
taking into account all markers across the genome. Ribaut et al. (2002), in
similar research, also found that selecting only for target loci in BC1 and
Computer Simulation in Plant Breeding 233

BC2 and background selection in BC3 in a maize population is more

efficient than other selection schemes. Selectable population size of 10
in BC1 and BC2 and 100 in BC3 can reduce the donor genome to <5%.
Through more generation of backcross, the proportion of the RPG
increases even without selection. Some empirical research supports this
conclusion (Kuchel et al., 2007).
For multistage selection in an MABC program, increasing the number of
individuals genotyped in each generation could reduce the required MDP
by as much as 50% (Frisch et al., 1999). This supported Hospital et al.’s
(1992) conclusion that single-generation background selection is most
efficient if selection is performed in the last backcross generation. Selecting
the RPG is not necessary in the first generation because the donor parent
genome is only 1/4 in subsequent generations after first backcross, so it can
be done in advanced generations; however, foreground selection in BC1
could significantly reduce the number of individuals to be genotyped in the
next generation (culling effect). Increasing the density of markers does not
necessarily lead to higher recovery of RPG on chromosomes that do not
carry the introgressed gene. In practice, two markers per 100cM should be
sufficient to get the highest possible response in the early generations. In the
carrier chromosome, the recombination probability between the intro-
gressed gene and donor genome segments is higher in later generations.
Thus, low-density markers should suffice in early generations, and high-
density markers should be used in later generations.
Other researchers have focused on backcrossing more than one gene into
elite lines (Frisch and Melchinger, 2001; Wang et al., 2007). For simultaneous
introgression of several genes into one recipient, gene–gene interaction (e.g.,
epistasis and pleiotropy) requires some care. A single combination of alleles
may not improve all the desired traits at the same time. For example, one
allele can improve drought tolerance or herbicide resistance, but it simulta-
neously decreases yield (e.g., Roundup Ready soybean). Consequently, a
breeder must balance different factors, such as number of genes to be pyr-
amided and population size, to find an ideal genotype, because the probability
of getting the desired genotype decreases exponentially with the increase in
number of genes. Simulation suggests that gene enrichment during early
generations (selection of homozygotes and heterozygotes with desirable
alleles) can greatly reduce resource requirements when combining nine
genes into one genotype through MAS (Chapman et al., 2007). Two different
approaches for multi-QTL introgression were also compared (Hospital and
Charcosset, 1997): simultaneous design (introgress several QTLs at the same
time) and pyramidal design (a single QTL is introgressed at the first stage and
then put into a single genotype). Simulation showed that to introgress four
QTLs at the same time, 1600 individuals are necessary to obtain a target
genotype with 99% confidence. With a pyramidal design, the total number of
individuals would be 580.
234 Xin Li et al.

Wang et al. (2007) compared three marker selection schemes to enrich

nine genes into a single genotype. Simulation showed that a top cross
scheme with approximately equal selection intensity at three stages (top
cross F1, top cross F2, and double haploid (DH)) required the minimum
total number of lines. Two of the nine genes in their study are linked in
repulsion phase, and the final allele frequency for one gene is lower than
expected because recombination is more rare than for two unlinked genes.
For gene introgression, the genes or QTLs transferred from the donor
parent would be assumed to behave the same in the recipient parent, but
due to epistatic effects, this assumption does not always hold. Therefore,
revealing the QTLs that interact with genetic background would help
improve plant breeding by sidestepping this issue ( Jannink and Jansen,
2001). The advanced backcross QTL (AB-QTL) mapping method couples
QTL mapping with introgression of the desired QTL to recipient lines
(Bernacchi et al., 1998a,b; Pillen et al., 2003; Schauer et al., 2006). In
Tanksley and Nelson’s (1996) proof of concept paper, computer simulation
examined the differences between self- and backcross populations for
detecting QTLs with different effects. Except for inefficiency in detecting
recessive genes and QTLs with epistasis effect, backcrossing population
(BC1, BC2) is more effective than selfing population in mapping dominate
or additive QTLs. For linkage drag, the simulation showed that the shorter
the segment of the donor QTL (15cM compared with 20, 25, or 30cM),
the more individuals were needed to get the target genotype. Yang et al.
(2007) also simulated combining QTL detection and introgression in
outbred species. The results showed that the allele substitution effect is
underestimated when the favorable QTL is not fixed in the donor line. In
linkage drag, AB-QTL is similar to the traditional introgression method.

2.1.6. Genome-wide selection

Genome-wide selection (GS, also as Genomic Selection) refers to marker-
based selection without significance testing and without identifying a subset
of markers associated with the trait under study. All genotyped markers are
fitted as random variables, and their effects are obtained in a linear model
across individuals with both marker and trait data. Breeding values (genomic
estimated breeding value, GEBV) are then predicted using the sum of effects
across all markers for individuals with marker data only (Bernardo and Yu,
2007; Meuwissen et al., 2001). Briefly, GS has two parts: model training and
selection. A model training population is first genotyped with an appropri-
ate density of markers and phenotyped under a set of environmental con-
ditions. Then statistical models can link marker information to breeding
values of individuals in the training population. With this model in hand,
the breeding values of new germplasm can be predicted using only marker
information.
Computer Simulation in Plant Breeding 235

With the development of cheaper and faster, high-throughput genotyp-

ing technologies, genome-wide selection has great potential to expedite
breeding through increased selection cycles per unit time and reduced
phenotyping efforts (Bernardo, 2008). The goal of genome-wide selection
is to reduce the burden of phenotyping as much as possible, when the model
can capture all of the genetic variance (additive variance) and the cost of
genotyping is low enough. As we discussed in section 2.1.3, traditional
MAS is never perfect because, to date, we have not captured the whole
genetic structure of most quantitative traits (yield, flowering time, drought
tolerance, etc.). The markers used in MAS are just a small subset of all
genetic determinants controlling the phenotype. Meanwhile, some of these
markers from mapping studies may be false positives or influenced by the
environment. In applying MAS to plant breeding, two distinct options will
help avoid this problem. One is to choose only markers with large genetic
effects (Bernardo and Charcosset, 2006); on the other end of the spectrum,
another option is to include all markers, even if their functions remain
unknown (Bernardo and Yu, 2007; Jannink et al., 2010; Meuwissen et al.,
2001; Xu, 2003). GS may avoid the shortcomings of traditional MAS with
its low percentage of variance captured, overestimated QTL effects, and
gaps between mapping population and breeding population. Although GS
is still generally considered marker-assisted breeding, generating accurate
prediction using genome-wide markers differentiates GS from other, more
traditional methods described in previous sections.
Bernardo and Yu (2007) compared GS and MARS in computer simula-
tions where the breeding strategy was one generation of phenotyping and
genotyping of DH lines in conjunction with 2 cycles of marker-based
selection in an off-season nursery. Simulations showed that with 20 QTL
and a high heritability (H¼0.80), the advantage of GS over MARS was 6%.
When QTL numbers are large and heritability is low, GS has an 18%
advantage over MARS. When the selection response due to PS in the
first generation is excluded, the maximum response was 43% higher with
GS than with MARS. The optimum range of markers used in GS is
128–256 and 64–128 in MARS (Fig. 3A). This number depends on the
population size. In a small population, the number of recombination events
is small, so more markers may not be helpful. The authors also investigated
the effect of decreasing phenotyping in cycle 0. Decreasing the DH lines
from 144 to 96 led to responses similar to a larger population in cycles 1 and
2, but decreasing the number of DH lines phenotyped in cycle 0 to 48 led to
decreased selection response regardless of the number of markers or plants
genotyped in cycles 1 and 2 (Fig. 3B). Consequently, a minimum of about
100 DH lines needs to be phenotyped to ensure the response to GS.
Mayor and Bernardo (2009) compared the usefulness of DH and F2
populations in MARS and GS. They considered the number of QTLs,
trait heritability, and population size. Simulation showed that the advantage
236 Xin Li et al.

A B
4.2 4.2 128 markers 48 DHs
128 markers 96 DHs
256 markers 48 DHs
256 markers 96 DHs

3.8 3.8
Response

Response
20 QTL MARS
3.4 20 QTL GS 3.4
100 QTL MARS
100 QTL GS

3.0 3.0
32 64 128 256 512 768 288 576 1152
Number of markers used Number of plants used
in two selection cycles in genomic selection

Figure 3 Comparison of marker-assisted recurrent selection (MARS) and genomic

selection (GS). (A) Response to testcross phenotypic selection in cycle 0 plus 2 cycles of
selection based on either on markers with significant effects (MARS) or on all markers (GS)
in simulated maize populations. (B) Response to GS for different numbers of markers,
double haploids evaluated in cycle 0 (NDH), and individual plants in GS (N) in simulated
maize populations. QTL¼40. Heritability is assumed to be 0.50 for both (A) and (B).
The horizontal coordinate is not proportional, but for illustration only. Responses are in
units of the testcross genetic standard deviation in the base population. The LSD 0.05 was
approximately 0.10. This figure is based on data from Bernardo and Yu (2007).

of DH over F2 is most significant in MARS, not GS, because the DH

population has higher genetic variance than the F2 population or selfed
(e.g., F3 or F4) families. This result leads to higher power and precision in
QTL mapping. Having an accurate estimate of the QTL effect in MARS is
important because genetic gain is expected only from a subset of QTLs with
significant effects. This contrast is most obvious when many QTLs control
the trait (e.g., grain yield), heritability is low, and population size is small.
For GS, the difference between DH and F2 population is not significant
because for GS, in attempting to include all genetic variances, a good
estimate of the QTL effect is less crucial because genetic gain is spread
over a larger number of QTLs.
Even with a small population, as for some tree species (e.g., N¼50 for oil
palm), GS for 3–4 cycles is better than both MARS and PS in terms of gain
per unit cost and time (Wong and Bernardo, 2008). Selection response was
higher with PS than with GS and MARS per cycle; however, per unit time,
GS and MARS are superior to PS (Bernardo and Yu, 2007; Wong and
Bernardo, 2008).
Bernardo (2009) simulated using GS to facilitate introgression of exotic
desirable genes into an adapted cultivar. In his research, unlike common
MABC, which involves successive backcrosses to an elite parent, only one
Computer Simulation in Plant Breeding 237

cross was made between adapted and exotic parents. Then multiple cycles of
recurrent selection were used to select superior lines. Surprisingly, simula-
tion showed that starting from an F2, not a BC1 or BC2 population, could
lead to higher selection response because this increases the frequency of
favorable alleles inherited from the adapted parent while maintaining a
moderate frequency of favorable alleles inherited from the exotic parent.
On the other hand, selection starting from a BC1 or BC2 population, which
has a lower frequency of exotic genes, could lose opportunities to incorpo-
rate favorable alleles from the exotic parents. As shown in another study
(Bernardo and Yu, 2007), responses increased as heritability increased, but
unlike the previous study, which showed that 144 individuals in cycle 0 was
sufficient for adapted by adapted cross, more selection response was
observed with larger populations for adapted by exotic cross.
Selection responses to GS for self-pollinated crops were 81–87% of the
corresponding responses to GS for cross-pollinated crops on an equal-time basis
(Bernardo, 2010). To achieve approximately the same response as GS in cross-
pollinated crops, the number of individuals that should be genotyped in self-
pollinated crops should be roughly twice as large as in cross-pollinated crops.

2.2. Gene mapping

Finding the genetic bases of complex traits is an essential task of plant
breeding. This process is greatly accelerated by the discovery of molecular
markers (Botstein et al., 1980) and QTL mapping methods (Doerge, 2002;
Mackay, 2001). Computer simulation can verify the efficacy of mapping
methods and explore the factors that determine mapping power, precision,
and accuracy (Beavis, 1998). Unfortunately, mapping power, precision, and
accuracy cannot be examined in an experimental population because we do
not know the true genetic structure of a specific trait, which is why computer
simulation is always used to verify the usefulness of a new mapping strategy.
The principle of computer simulation for this purpose is to artificially
design the genetic structure: QTL number, position, effect, epistasis, chro-
mosome number, length, marker density, heritability of the trait, popula-
tion type, and population size. This virtual mapping population can then be
used to test the mapping method and check whether results confirm any
advantages to the new method. This review is not dedicated to the details of
different mapping methods, so we will provide only a brief summary of
various methods and emphasize the role of computer simulation in experi-
mental design and data analysis of gene mapping.

2.2.1. Factors influencing mapping efficiency

The accuracy of any QTL mapping procedure depends on a number of
factors: the heritability of the trait, the number of genes involved, the
interactions of the genes, the distribution of the genes over the genome,
238 Xin Li et al.

the statistical distribution of the random nongenetic factors, the type of

segregating population studied, the size of this population, the genome size,
the number of marker loci, and marker distribution over the genome
(Ooijen, 1992). The principle of QTL mapping is to detect significant
phenotypic differences between different genotypes as revealed by markers.
Thus, broadly speaking, detecting QTLs in large populations is easier when
a few QTLs have large effects on the trait. A large population always means
more available information for estimating statistical parameters, in this case,
QTL position and effect. Fewer QTLs means fewer parameters we need to
estimate simultaneously, which also can increase precision. The larger the
QTL effect or higher heritability of the trait, the less environmental noise,
which can increase the power to detect QTLs. Marker density and popula-
tion type (F2, BC1, RIL, or DH) also affect mapping efficiency. Computer
simulation is used extensively to quantify the effect of these factors because
an analytical solution is not practical at the population scale.

2.2.1.1. Population size and marker density Increasing sample size and
testing environments could make estimating QTL effects more accurate.
With large populations, more recombination events can be sampled, thus
increasing mapping power. With small sample sizes, the QTL effects are
inflated and mapping results are less likely to be repeated in other environ-
ments or experiments (Beavis, 1998; Schon et al., 2004). A population of at
least 2000 individuals is necessary to map a QTL with a small additive effect
(variance explained is 1%; Ooijen, 1992). Ooijen’s (1992) simulation results
showed that a minimum population of 200 backcross or F2 individuals is
needed to detect a QTL that explains at least 5% of the total phenotypic
variance. Moreover, the larger the genetic effect of the QTL, the more
likely it can be located precisely.
In a classic computer simulation in QTL mapping study, Beavis (1998)
investigated the potential power, precision, and accuracy of QTL analysis
(Fig. 4). He used three factors in his simulation: number of QTLs (10 or 40),
phenotypic variability explained by these QTLs (30%, 63%, 95%), and
number of F2 progeny (100, 500, 1000). These 18 simulation combinations
were repeated 200 times each, so the total number of simulations was 3600.
Simulation results showed that identifying all QTLs was possible if the
number of QTLs was small and the population size was as large as 1000
(Fig. 4A). For more QTLs (e.g., 40), even with 1000 progeny and a
heritability of 0.95, detecting all the QTLs was impossible. With large
numbers of progeny and higher heritability, the standard error of the
estimated genetics effect of any QTL decreased. The average estimated
magnitudes of genetic effects associated with a correctly identified QTL
were greatly overestimated if only 100 progeny were evaluated, slightly
overestimated with 500 progeny, and fairly close to the actual magnitude
when 1000 progeny were evaluated (Fig. 4B).
Computer Simulation in Plant Breeding 239

100 2
10 QTL h = 0.3 2
2 10 QTL h = 0.3
10 QTL h = 0.95

Estimated–simulated additive effects

2
2 10 QTL h = 0.95
40 QTL h = 0.3
2
40 QTL h = 0.95
3 2
40 QTL h = 0.3
2
80 40 QTL h = 0.95
Power (%)

60 2

40
1
20

0 0

100 500 1000 100 500 1000

Size of mapping population Size of mapping population

Figure 4 Effects of heritability and sample size on the power, precision, and accuracy
of QTLs identified in F2 progeny with either 10 or 40 QTLs. (A) Power to detect
simulated QTLs given a¼0.25. Power is the proportion of QTLs detected among all
the simulated QTLs. (B) Differences between estimated and simulated additive genetic
effects and standard errors. The genetic effect of each QTL is the percentage of the
phenotypic variability that each contributes. Only additive genetic effects are simulated,
and all QTLs have equal genetic effects. This figure is based on data from Beavis (1998).

Moreover, for linkage analysis, increasing marker intensity does not

necessarily lead to increased resolution for QTL position, but increasing
population size and heritability could improve accuracy (Darvasi et al., 1993;
Schon et al., 2004; Visscher et al., 1996a). To position a QTL, linkage
mapping uses only the recombinations that occurred within the data set,
which typically contains two to three generations. Darvasi et al. (1993)
noted that with closely linked markers, only a few recombinations between
adjacent markers occur during these two to three generations; hence, a
dense marker map provides little extra information about the position of the
QTL unless the number of individuals per generation is very large. Further,
the effect of marker density on mapping resolution depends on the resolving
power itself. Marker spacing narrower than the resolving power of the
experiment did not estimate QTL position more accurately. Computer
simulation may, however, estimate the mapping resolution potential of a
given experiment with a specific population size and QTL effect. The
marker density can then be set according to the simulated results. Darvasi
and Soller’s (1997) simulation showed that, for BC and F2 mapping popu-
lations, two simple expressions can be used to estimate marker spacing and
resolving power. One is 3000/(mNd2) (where m¼1 for BC and m¼2 for
F2; N¼population size, and d¼allele substitution effect), and the other is
530/Nv (where v is the proportion of variance explained).
240 Xin Li et al.

2.2.1.2. Missing marker information Missing markers are common in

QTL mapping experiments. For example, dominant markers cannot differ-
entiate homozygosity and heterozygosity in an F2 or BC1 population.
Experimental errors also generate missing markers. The missing marker
issue was investigated by Jiang and Zeng (1997). Their method used marker
information from other genomic positions of the individual and related
individuals as well as multiple generations. Simulation results showed that
having some codominant markers in a linkage group is beneficial. Codomi-
nant markers can provide accurate genotypic information in various positions
of the genome and can substantially increase the accuracy of the analysis.
Zhang et al. (2010) investigated the effects of missing markers and
segregation distortion on QTL mapping efficiency in a rice F2 population.
The QTL number and effects as well as the linkage map in their simulation
study derive from a real rice mapping population. Simulation results showed
that missing markers have the most impact on QTLs with small effects. For
instance, if 10% of 137 markers were missing, the corresponding detection
power would be 91.8% for a QTL with a PVE (phenotypic variation
explained) of 25.57%; however, for a QTL with a PVE of 3.86%, the
detection power is only 7.7%. This trend also held true when the simulated
population size increased from 180 to 500. In essence, the randomness in
imputation of missing markers means missing markers have the same effect as
a smaller mapping population. In other words, a population of n individuals
with p percent missing markers is the same as a population of n(1p), but
without missing markers. The influences of segregation depend not only on
the effect of the QTL but also on the linkage distance between the distorted
marker and the QTL. The simulation showed that when the distorted marker
was closely linked to a QTL, detection power would either increase or
decrease, depending on the QTL effect (additive or dominant), but when
the distance between an SDM (segregation distortion marker) and a QTL is
more than 40cM, the effect on QTL mapping is not significant.

2.2.1.3. Meta-analysis Most QTL mapping studies were conducted in a

single mapping population that was derived from a single pair of parental lines,
developed in a limited number of environments, and genotyped by a specific
set of markers. This makes the experimental results nonreplicable, even in
other similar experiments. Beavis (1994) pointed out a number of sources of
difference between QTL experiments that reduce reproducibility. First, when
the power to detect a QTL is less than one, the QTL may be detected in some
experiments but not others. Detecting a QTL is enhanced or reduced by
colinearity with error residuals caused by random measurement variation and
by colinearity with other QTLs caused by random sampling of particular
genotypes among segregating progeny. Given proper randomization, these
sampling variations will not be repeated from experiment to experiment.
Second, a QTL will be detected only in populations where it segregates:
Computer Simulation in Plant Breeding 241

even in the absence of known coancestry, some parents may carry identical-in-
state alleles. Third, QTL-by-environment interaction may affect QTL substi-
tution effects in experiments conducted in different environments. Finally,
QTL-by-genetic background interaction may affect QTL substitution effects
in experiments conducted with different populations. These sources of differ-
ence not only explain problems with replication but also could explain the
discrepancies between mapping results and MAS results.
From these possible causes of unrepeatable results, we can logically
assume that more testing environments and multiple mapping populations
could increase mapping accuracy. Simulation and experimental results both
show that more samples are more important than more testing environ-
ments (Schon et al., 2004).
Meta-analysis combines multiple QTL studies to improve the estimation
of QTL effects and positions. Xu (1998) pointed out that using more than a
single family for QTL mapping would allow detection of some QTLs that
may not be segregating in a particular family. In addition, mapping popula-
tions from a pair of parents with opposite extreme values for the trait of
interest are commonly used for population development. This is a powerful
method of mapping a single trait, but it may not help in detecting QTLs
responsible for other traits. For multiple-family QTL mapping, a random
model approach that treats each effect of gene substitution as a random
variable is preferred.

2.2.2. Significance threshold and confidence interval

In QTL mapping analysis, the threshold used to declare a significant QTL
and CI of QTL position can be determined analytically only if we assume
that the number of markers is infinite, which is not the case in practice.
Thus, computer simulation can better calculate the threshold and CI for
QTL detection. The critical values for significance tests and interval esti-
mates of the parameters have to be established using a repeated sampling
technique, for example, a permutation test or bootstrapping analysis.
The determination of critical value and CI is difficult for two reasons
(Churchill and Doerge, 1994). The first is that the distribution of the test
statistic under an appropriate null hypothesis is implicit. In most cases, the
regularity conditions that ensure an asymptotic chi-square distribution for
the likelihood ratio test statistic are not satisfied. Second, multiple tests are
performed in QTL detection, most of which are not independent of each
other. The dependence structure of these tests makes analysis difficult in
cases other than the extremes of very dense or very sparse genetic maps.

2.2.2.1. Significance threshold Computer simulation can also determine

threshold values (Lander and Botstein, 1989). Based on the genetic structure
(number of chromosomes, length of each chromosome, and number of
markers), a population is generated, and the phenotype is randomly selected
242 Xin Li et al.

from a normal distribution. This assumes no QTLs in the genome (the null
hypothesis of maximum likelihood test). If this process is repeated a large
number of times (e.g., 10,000), all these LOD scores can be used to generate
a distribution of test statistics under the null hypothesis. Subsequently, this
distribution can be used to determine the critical value.
The threshold deduced in this approach is case-sensitive, that is, a unique
character of the genetic model under study (Churchill and Doerge, 1994;
Van Ooijen, 1999). Churchill and Doerge (1994) suggested the permuta-
tion test to approximate the distribution of LOD statistics under the null
hypothesis. The genetic information is kept untapped, the trait values are
randomly assigned to the genotype without replacement, and the LOD
scores are calculated. This process is repeated 1000 times (or more, depend-
ing on the Type I error rate that investigators use; the lower the Type I error
rate, the more repetitions) to obtain the distribution of LOD scores.
Because the genetic maps and trait values are not altered by the shuffling
procedure, this distribution automatically takes into account the particular
characteristics of the experiment at hand.
To get the LOD score that can be applied to various mapping studies, Van
Ooijen (1999) used computer simulation to determine the cumulative distri-
bution function of the maximum LOD score under the null hypothesis that
no segregating QTL is present. The simulation was based on a diploid species
with one chromosome on which no QTL was segregating. The chromosome
map length was 50–250cM with a fully informative marker every 1cM. Four
different kinds of population (BC1, F2, RIL, FS [full sibling] family of a cross
between non-inbred genotypes of an outbreeding species) were used with a
population size of 100. The individuals were assigned a random phenotypic
value sampled from a normal distribution regardless of the genotype, that is,
no QTL. This method means calculating the LOD score, in effect, as if you
were looking up the test statistics distribution tables at the back of a statistics
textbook, and it needs less computation time.
The false discovery rate (FDR) is a concept usually confused with P-
value. P-value is used in hypothesis tests to control Type I error rates (i.e.,
rejecting the null hypothesis when it is true). FDR is a measure of the
quality of a mapping experiment. FDR is the number of false-positive
QTLs/number of total QTLs detected. In extreme cases, finding one
significant QTL that turned out to be a false positive gives an FDR of
100%, even though we may have used 0.01 as the P-value. Bernardo (2004)
used computer simulation to test the relationship between these two distinct
criteria for QTL detection. The results showed that a P-value of 0.05 led to
a low FDR when many QTLs (30 or 100) controlled the trait, but this
lower FDR was accompanied by a diminished power to detect QTLs.
Larger mapping populations led to both a lower FDR and an increased
power. To control false positives, Bernardo (2004) recommended a P-value
of 0.0001 to determine a QTL as true positive.
Computer Simulation in Plant Breeding 243

2.2.2.2. Confidence interval of QTL position The commonly used

method for calculating the 95% CI for QTL map position is the LOD
drop-off method. The CI is calculated by finding the location at either side
of the estimated QTL location that corresponds to an LOD score decrease
of 1 or 2 units. The total width around this potential QTL is then taken as
the CI. This CI, however, may be biased for small and medium population
sizes (Visscher et al., 1996b). Visscher et al. (1996b) used the bootstrapping
method to calculate the empirical CI for QTL position. Their method first
resamples (withdrawing N observations with replacements) the original
population (size N) 1000 times. Second, the 1000 estimated QTL positions
are ordered and the bottom and top 5th and 2.5th percentile are taken. This
method could make CI more conservative, which can be useful in marker-
assisted introgression, because the possibility of introducing large parts of the
donor parent genome around the target gene is decreased.
The debate on permutation or bootstrap randomization continues. A
permutation may retain the summary information of the trait, whereas the
bootstrap changes the mean and variance of the original sample (Doerge,
2002). The disadvantage of the permutation test is that it is computationally
intensive.

2.2.3. Brief summary of mapping methods

As we mentioned at the beginning of this section, the true genetic structures
of most quantitative traits are unknown, so using an existing population to
test the power of new mapping methods is inappropriate. Computer simu-
lation, with a genetic model predefined by the researcher, can be used to
generate phenotypic and genotypic data sets. Based on a large number of
data sets, the simulated mapping results can show the behavior of the
mapping method in the long run. Most preliminary validations of new
mapping methods come from simulation studies. Here, we briefly summa-
rize the progress of mapping methods and their merits.

2.2.3.1. Linkage analysis Detecting marker-trait association can be car-

ried out through t-tests based on a single marker (Soller et al., 1976). Single
marker approaches (t-test, ANOVA, or simple regression) use the differ-
ences in phenotype means of different genotype groups to detect the
association of marker and trait, but this method cannot separate the effect
and location of the QTL. Interval mapping (IM; Lander and Botstein, 1989)
uses a genetic map to test the likelihood of QTL existence in each interval,
a segment of chromosome bracketed by two flanking markers. The likeli-
hood ratio statistics, LOD scores, from all tests in the genome constitute a
smooth plot. A significant peak from this plot profile (depending on the
threshold value; see the previous discussion) indicates a potential QTL at
that position. As a single QTL search algorithm, however, IM cannot
separate multiple QTLs. This often can result in “ghost QTL” near a peak
244 Xin Li et al.

in the LOD plot (Doerge, 2002). Composite interval mapping (CIM;

Jansen, 1993; Zeng, 1994) is a method of linking regression with the
maximum likelihood test used in IM. The effects of other QTLs outside
the current interval are treated as cofactors in the model. CIM can substan-
tially improve mapping efficiency if the cofactors are properly chosen.
Multiple interval mapping (MIM) can simultaneously map multiple QTLs
(Kao et al., 1999). Although IM and CIM are statistical methods for mapping
individual QTLs (i.e., the genetic effect and genomic location of a single
QTL), MIM is better suited for simultaneously identifying and estimating the
genetic architecture parameters, including the number, genomic positions,
effects, and interactions of significant QTLs as well as their contributions to
genetic variance (Zeng et al., 1999). Because MIM includes epistasis in the
analysis, it can improve the statistical power to identify more minor and
complex QTLs and is more precise in estimating QTL positions.
Inclusive composite interval mapping (ICIM) is a simplified form of
CIM (Li et al., 2007). It uses stepwise regression to select significant markers,
then phenotypic values are adjusted by all markers retained in the regression
equation except the two markers flanking the current mapping interval.
CIM is used to map QTLs with the adjusted phenotypic values. Computer
simulation shows that ICIM works better than CIM at increasing power,
reducing false-positive rates, and estimating QTL effects with less bias.

2.2.3.2. Association mapping An intrinsic shortcoming of linkage analysis

for QTL mapping is a low resolution for gene discovery. Linkage analysis in
plants typically localizes QTLs to 10–20-cM intervals because of the limited
number of recombination events during population construction and because
of limited sample size (Zhu et al., 2008). Association mapping using a
collection of germplasm rather than controlled biparental population can
give higher mapping resolution (Risch and Merikangas, 1996); however,
unknown population structure, a notorious property of natural populations,
can result in spurious associations (Pritchard et al., 2000; Yu et al., 2006).
Thus, accounting for population structure is a critical step in controlling false
positives in association mapping (Price et al., 2006; Yu et al., 2006; Zhu and
Yu, 2009). Relationship estimation can use genome-wide background mar-
kers. Simulation results (Yu et al., 2009) show that 300–600 biallelic markers
provide a robust estimation of kinship for association mapping with diverse
germplasm.
In silico mapping (Grupe et al., 2001) is a good combination of computer
science and the abundant data resources available in the public domain,
although its efficiency is debated (Chesler et al., 2001). In silico mapping in
plants has four advantages over designed mapping experiments (Parisseaux
and Bernardo, 2004; Yu et al., 2005). First, in silico mapping uses larger
populations than designed mapping experiments. Second, the phenotypic
data used in in silico mapping are obtained through more extensive testing
Computer Simulation in Plant Breeding 245

under multiple, diverse environments. Third, the hybrids and inbreds tested
typically represent a broader genetic background. Fourth, the data used for
in silico mapping are available at no extra cost. Yu et al. (2005) investigated
the statistical power of in silico mapping in a hybrid breeding program via
computer simulation. They found that a large sample size, high marker
density, high heritability, and small numbers of QTLs led to the highest
power for in silico mapping via a mixed-model approach, but a compromise
between the power of detecting QTL and FDR would be necessary.

2.2.3.3. Joint linkage and linkage disequilibrium mapping Joining LD

mapping with linkage analysis was initially suggested as a fine mapping
approach (Meuwissen et al., 2002; Wu and Zeng, 2001). With linkage
mapping, we can assume the QTL was narrowed down to a region of 20
cM. Then, within this smaller region, marker density of one marker each
CentiMorgn and LD mapping were used to narrow the position of the QTL
to 3cM. Next, this 3-cM region would be covered by markers 0.25cM
apart, and the position of the QTL is narrowed to a 0.75-cM region. Finally,
with even higher marker density, the underlying gene could be cloned.
Nested association mapping (NAM) may take advantage of both linkage
mapping (balanced population and fewer markers) and association mapping
(higher resolution; Myles et al., 2009; Yu et al., 2008). The principle of
NAM is similar to QTL mapping in that it uses identical by descendent
(IBD) inbred lines from breeding programs (Crepieux et al., 2004; Jansen
et al., 2003). In a proof-of-concept paper, Yu et al. (2008) compared QTL
detection in a maize NAM population derived from crossings between 25
diverse founders and the common parent B73. Three marker schemes were
considered: complete marker information for all 5000 RILs, all SNP geno-
typed for the 26 founders but only common parent specific (CPS) markers
scored for the RILs, and a traditional method without projecting founder
SNP information between CPS makers. Simulation results showed that
with 5000 RILs and QTL numbers of 20, the average power achieved by
scoring only CPS markers for RILs was 94% of the complete marker
scheme, 90% when QTLs numbered 50. Decreasing the number of RILs
or the crosses selected as founders for the RIL population led to decreased
mapping power and higher FDR. The mating design for NAM population
development was investigated by Stich (2009).

2.2.3.4. Bayesian mapping Recently, Bayesian methods for QTL

mapping and association mapping have become more attractive than tradi-
tional methods such as maximum likelihood or regression-based approaches
(Bauer et al., 2009; Bogdan et al., 2004; Chen et al., 2010; Han and Xu,
2010; Manichaikul et al., 2009; Sahana et al., 2010; Satagopan et al., 1996;
Vieland, 1998; Xu, 2010; Yang et al., 2007). In Bayesian analysis, inferences
about the parameters of interest are based on the joint posterior distribution
246 Xin Li et al.

of all the unknowns. Because the joint posterior distribution does not have a
standard form, MCMC samplers (e.g., Jibbs sampler) can generate samples
from the joint posterior distribution. Bayesian mapping can perform linkage
analysis with any number of marker loci, multiple trait loci, and multiple
genomic segments (Xu, 2010).

2.3. Gene network and genotype-by-environment interaction

Quantitative traits are difficult to study for at least two reasons: they are
influenced by a small number of genes with large effects and a large number
of genes with small effects (Lynch and Walsh, 1997) and they are influenced by
environmental factors. These difficulties have two consequences. First, our
efforts to dissect complex traits are hindered because finding all of the genes
controlling a quantitative trait is almost impossible; second, even if we could
illuminate the function of a single or several genes involved in a genetic
network, predicting the phenotype would still be difficult because the genes
might interact with each other (epistasis) and function differently in different
environments. However, predicting phenotypes given knowledge of the
genome and modeling dynamic properties of a biological system is fundamen-
tal if we try to accelerate the process of plant breeding (Cooper et al., 2002).
Because quantitative traits are influenced by environment, a cultivar that
grows well in one specific environment does not necessarily have high yield
in another environment. Thus, potential cultivars are tested for performance
over several years and at different locations. A new commercial maize
hybrid, for example, may be tested in 120–2100 environments before it is
released to farmers (Bernardo, 2002). Using our knowledge about geno-
typic and environmental effects to predict the performance of a cultivar in a
specific environment can help shorten the duration of a breeding program.
In this section, we will describe the E (N:K) model and QU-GENE
simulation platform, both of which are used to simulate the gene network
and GEI. In the next section, we discuss how crop models can tackle GEI
and how to couple gene information with crop modeling.

2.3.1. E(N:K) model

The E(N:K) model developed by Podlich and Cooper (1998) is a frame-
work that considers adaptation and fitness landscape. In this model, E
denotes the number of environment types in the target population of
environments (TPEs); N is the number of genes determining the perfor-
mance of the genotype; K is the degree of epistasis between the genes. In
different environments, the values of N and K can differ. From the simplest
model with only one environment and no epistasis (e.g., 1(12:0)), to the
most complicated model with different types of environments in the TPE
and epistasis (e.g., 5(12:2)), using the E(N:K) framework, families of models
can be defined to generate genetic scenarios ranging from simple to highly
Computer Simulation in Plant Breeding 247

complex. This helps us to use simulation to investigate the efficiencies of

different breeding strategies (Kang, 2001). Software developed by Podlich
and Cooper (1998), called QU-GENE, accompanies this model. This
software has two components: (1) the engine, which is used to define the
genetic model for the genotype–environment system and (2) the application
modules used to investigate, analyze, or manipulate populations of geno-
types within the defined genotype–environment system.

2.3.2. Plant-breeding simulations with QU-GENE

Many simulation experiments have used this platform. Podlich et al. (1999)
used a selection strategy that weights the data from the multienvironment
trial (MET) according to their expected frequency of occurrence in the
TPE. The simulation results showed that when the MET matched TPE, the
response to selection is similar between weighted and unweighted strategies,
but when MET did not match the real TPE, the weighted selection method
is more effective.
Wang et al. (2009b) investigated the efficiency of a backcross breeding
strategy with different genetic compositions of parental lines, crossing
strategies, and selection schemes. The desirable traits in their study were
controlled by multiple genes, and selection was based on phenotype, not a
simple trait controlled by one or two genes with maker-based selection.
The simulation results showed that the single backcross breeding strategy
(SBBS) can transfer more than 60% of favorable genes from the donor
parent and improve the adaptation of the recipient parent. Two backcrosses
would have the advantage when the adaptation of donor parent is poorer
than the adapted parent, but three backcrosses led to minimal advantages
over the two-backcross breeding strategy.
Kuchel et al. (2005) compared three MAS strategies with PS in a wheat
breeding program in Australia. The three strategies (MAS1, MAS2, and
MAS3) differed by the stage in which marker selection was implemented.
Their target was to accumulate all desirable genes, both disease resistance genes
and grain quality genes, into the high-yield cultivar Stylet. Marker-based
selection plus PS in the TPE were used to select the target genotypes. Applying
MAS at all three stages (BC1F1, haploid, and recurrent parent background
selection) led to the most genetic gain. One of the most significant contribu-
tions of marker selection to the breeding program was to discard undesirable
genotypes in the haploid type stage. This saved $AUD 65,000. Economic
analysis clearly showed the usefulness of MAS in resource allocation.

2.4. Crop physiology and crop modeling

All efforts to decipher the genetic code of specific traits, either in plants or
human beings, have two benefits: (1) understanding the mechanism of a trait
or disease to provide possible treatment or explanation and (2) predicting
248 Xin Li et al.

the phenotype by examining the genotype. As mentioned in the previous

section, the relationship between genotype and phenotype is not straight-
forward. The QTLs appear to be developmental stage-specific (Yin et al.,
1999), and the results of one experiment cannot be extrapolated to other
situations. One reason for this is that many QTL experiments focus on a
complex trait as a whole rather than its underlying component dynamic
traits. For example, most QTL studies treat yield as a single trait, but when
this complex trait was dissected into three components (number of spikes
per unit area, number of grains per spike, and grain weight), the QTL
mapping result differed from those focused on yield alone (Yin et al., 1999).
Another reason for low consistency between QTL experiments conducted
in different environments is that most QTL studies take the genotype-by-
environment effect into account statistically, but environmental factors and
how they influence the phenotype are not explicit (Reymond et al., 2003;
Yin et al., 2004). Physiological dissection and modeling of traits can con-
tribute to integrating molecular genetic technologies in crop improvement
(Hammer et al., 2002).

2.4.1. Crop modeling

Crop simulation models use quantitative descriptions of ecophysiological
processes to predict plant growth and development as influenced by envi-
ronmental conditions and crop management, which are specified for the
model as input data (Hodson and White, 2010). Crop modeling has been
used primarily as a decision-making tool for crop management, but crop
modeling, coupled with crop physiology and molecular biology, also could
be useful in breeding programs (Slafer, 2003).
Of the two different, but not exclusive, approaches to crop modeling,
one focuses on dissected parts of plants, like leaf canopy (Leuning et al., 1995)
or root structure (de Dorlodot et al., 2007; Dupuy et al., 2010). The other
approach integrates models for different physiological traits to simulate
the growth of the whole plant. Whole-crop computer simulation models
use plant physiology and environmental variables to calculate plant growth,
or more specifically, yield and dry matter production. Both directions are
increasingly coupled with molecular genetics to facilitate crop modeling.
Crop model parameters are usually determined by iterative parameter
adjustment and comparison with observed data from field trials; however,
parameters estimated in this way are inaccurate because of the inherent experi-
mental errors associated with field observation. Researchers now dissect model
parameters into genetic factors to make cultivar-specific models. Crop model-
ing, which incorporates all (if possible) parameters affecting a trait into a model,
can combine both genetic and environmental elements to predict phenotype
(Hoogenboom et al., 2004; Messina et al., 2006; White and Hoogenboom,
2003; Yin et al., 2003, 2004). This approach has several advantages over
conventional methods of investigating a single trait without considering the
impact of environment. First, by separating the underlying QTLs into different
Computer Simulation in Plant Breeding 249

parts (parameters) of the ecophysiological model, some QTLs otherwise have

only small effects if using the target trait (e.g., yield or resistance to drought)
could be detected by QTL mapping (Letort et al., 2008). In this situation, the
ecophysiological model is like a gene network linking all of the relevant genes
into a complete picture. Some QTLs may have a small effect statistically, but
that does not mean the QTL is unimportant to the target trait, even though we
neglect these QTLs in almost all circumstances and turn to others that can
explain more variance. When we examine the underlying traits (leaf elongation
rate, leaf angle, biomass allocation, etc.) one by one, more QTLs might be
discovered. Second, because of GEI, some QTLs identified as beneficial to the
desired trait might have the opposite effect in another environment (Chenu
et al., 2009). Third, under the influence of pleiotropy, selecting for one trait
may cause an adverse effect if that trait is coupled to another trait. All these
issues can be reflected in a well-verified crop model.
With crop models, model input parameters represent genetic coeffi-
cients. QTL mapping is carried out according to the parameters of the
ecophysiological model; that is, each parameter of the ecophysiological
model is computed as the sum of QTL effects (Yin and Struik, 2010).
Moreover, in multienvironmental trials (METs), the values of the para-
meters are not necessarily the same because of GEI (Chapman, 2008); thus,
crop models are like bridges over the genotype–phenotype gap. QTL
information and ecophysiological models theoretically can be combined
to predict a physiological trait of any cultivar under study in any climate
scenario (Reymond et al., 2003).
Yin et al. (2005) investigated flowering time of a recombinant inbred line
population of spring barley (Hordeum vulgare L.). CIM was conducted to
identify QTLs controlling four input traits. The model accounted for 72%
of the observed variation among 94 RILs and 94% of the variation among
the two parents across eight environments; however, because some QTLs
have only small effects on the input traits, the ecophysiological model is less
accurate in predicting phenotypes of the RILs than the original phenotypic
input values. The authors also compared the QTL-based model with QTLs
that directly control the days to flowering. The results showed that the latter
method accounted for 85.5% of overall variation, higher than the variation
explained by the QTL-based model (72.6%), mainly because the prediction
using the QTL-based model is independent of the greenhouse experiments
calibrating the parameters of the model. This is also an intrinsic problem
with QTL-based crop modeling (Messina et al., 2006). Further, QTLs
cannot explain all the variation; this means the ecophysiological model,
like other genetic models, cannot totally predict the phenotype of a cultivar
with a specific genetic background.
APSIM (The Agricultural Production Systems sIMulator) is a widely
used simulation tool for cropping systems that was designed to combine
accurate predictions of economic products (e.g., grain, biomass, or sugar
yield) for many crop species in response to climate and management
250 Xin Li et al.

conditions (Keating et al., 2003; McCown et al., 1996). Asseng et al. (2003)
used a derivative of APSIM, APSIM-Nwheat, to investigate the relationship
between specific leaf area (SLA) and yield under the environmental condi-
tions of Australia. Simulation results showed that increasing SLA could
increase yield only on the sandy, low water-holding soils in a Mediterranean
environment when the N supply was unlimited. On loamy soils, the
increased SLA trait actually reduced grain yields.
Using sorghum as an example, Chapman et al. (2003) simulated the
selection for yield in three drought environment types in sorghum produc-
tion regions in Australia. In their simulation, 15 genes were assumed to have
an influence on yield. These 15 genes controlled transpiration efficiency
coefficient (TE), flowering time (FT), osmotic adjustment (OA), and stay
green (SG). APSIM generated the yield value for all 4235 genotypes in the
target environments, and QU-GENE monitored the selection response of
the 15 genes in different drought conditions in an S1 family recurrent
selection-breeding program (Fig. 5). In general, selection for higher values

Selection in MET

Population Yield value

Soil
Gene number
Water
Gene effect
QU-GENE CO2
Epistasis Trait value APSIM
engine Nitrogen
Environment
Radiation
GEI
Temperature

Figure 5 Link QU-GENE with Agricultural Production Systems sIMulator (APSIM).

Gene information and expression states in TPE are input into the QU-GENE engine to
generate a simulated breeding population and the corresponding trait values (e.g.,
flowering time) of each genotype. The trait values are subsequently used by APSIM
to predict the yield value in TPE, based on environmental information. With the
breeding population and yield value for each individual, the breeding process can be
simulated with QU-GENE simulation modules (e.g., QU-Line).
Computer Simulation in Plant Breeding 251

for these four traits resulted in higher yield, but not in all environments.
This study demonstrated the advantages of combining a crop physiology
model with a breeding simulation program: to investigate selection response
in different environments. Without the biophysical crop simulation model,
and assuming that all the 15 genes have equal additive effects on yield in
different environments, each of these genes would be fixed at a similar rate.
The introduction of crop models has modified the relative yield value of the
different genes in the environment types and thus has influenced the rate
and timing of fixation of the favorable alleles for these four traits. Simulation
results showed that increasing the level of trait expression (by increasing the
number of favorable alleles for a trait) led to higher grain yield, except for trait
PH in the Severe-Terminal drought environment type. Higher values for the
PH and TE traits had more effect on yield as the environments changed from
Severe-Terminal stress to Mid-season stress to Mild-Terminal stress. On the
other hand, higher values of TE and OA had more effect on yield in the
Severe-Terminal drought environment type. This is a kind of epistasis effect,
which has important implications for MAS programs for the sequence of gene
fixation that can result in the highest selection response.
Moreover, crop simulation models can be used to characterize environ-
ments based on crop performance data by connecting GIS systems and crop
models (Chapman et al., 2000; Loffler et al., 2005). With 108 years of
weather and soil data from six locations in major sorghum growth regions
in Australia, Chapman et al. (2000) used a simulation model to characterize
the environments into three drought patterns: Mid-season, Mild-Terminal,
and Severe-Terminal. The frequency with which these three types of
environment occur is an approximation of TPE. In another later study,
based on geographic information for 50 years (1952–2002) for each U.S.
Corn Belt Township, six major environment classes (EC) were identified
(Loffler et al., 2005). This classification of maize growth environments,
rather than climate and soil data, can be useful in interpreting GEI and
predicting cultivar performance in TPE.

2.4.2. Virtual plants and E-cell

Elucidating the many mysteries of life science requires first looking at the
microview, by dissecting a complex organism into single traits or metabolic
pathways and the genetic factors controlling these traits, then taking the
macroview, by integrating all of our accumulated knowledge about a
specific phenomenon (e.g., photosynthesis of plants) into a whole picture
of the metabolic pathway. The second stage allows us to evaluate the
influences of interactions between these small elements on the growth of
a whole organism, to model an organism as a whole, which is the primary
task of systematic biology. Once again, computer simulation can be a
powerful tool to accomplish this task.
252 Xin Li et al.

Grafahrend-Belau et al. (2009) investigated barley seed metabolism using

a computational approach. They constructed a metabolic network of the
primary metabolism in the developing endosperm of barley (H. vulgare),
including 257 biochemical and transport reactions across four different
compartments. The simulation output of growth rate and the active meta-
bolic pathway patterns under anoxic, hypoxic, and aerobic conditions
showed good agreement with experimental results.
Based on real structural and morphological data, a virtual plant is more
intuitively like simulation than data alone (Barczi et al., 2008; Evers et al.,
2010). A virtual plant can generate a model (a vivid picture) that mimics the
structure of real plants and a microclimate for individual plant organs. A 3D
plant model is useful for research in physiology, forestry, and landscape
management (Barczi et al., 2008), but unlike crop biophysical models, which
are based upon complex trait dissection, the virtual plant growth models are
usually based on the physiological age concept or botanical architectural
theory. One important modeling method is based on the theory of Linden-
mayer systems (L-systems; Prusinkiewicz, 2004; Prusinkiewicz et al., 1996).
Because most of the literature is dedicated to simulating the effect of plant
architecture on plant responses to the environment, not plant breeding, we
provide only a basic introduction to virtual plants in this review.
E-cell System (Takahashi et al., 2003; Tomita et al., 1999) is a software
system used for modeling and reconstructing biological phenomena in silico
(http://www.e-cell.org/ecell/, verified August 8, 2011). Unlike a virtual
plant, which focuses on the plant architecture and plant growth, E-cell
attempts to simulate the biochemical reactions in a single cell to sustain a real
(although simulated) life, from transcription, translation to protein interac-
tion, and metabolism; however, it depends heavily on existing knowledge
of metabolism regulation and quantitative data for biochemical reactions,
such as concentrations of metabolites and enzymes, flux rates, kinetic para-
meters, and dissociation constants (Tomita, 2001). Before we figured out
most, if not all, of the mechanisms in genome, proteome, transcriptome,
and metabolome, E-cell might not have been considered a virtual replica-
tion of a real cell. But simulation for some simple organisms (e.g., Myco-
plasma genitalium, from which the first E-cell was constructed) have shown
some novel phenomena not discovered through experiments. Thus, this
system exploits the huge amount of information available to discover how
the cell would behave under certain conditions, which, in turn, would be
helpful in genome engineering.

2.4.3. Climate change

Increasingly severe environmental stress, such as heat and drought, can
affect plant growth and pose a threat to sustainable agriculture (Ahuja
et al., 2010). Predicting climate change has become a major research area
for meteorologists. An Intergovernmental Panel on Climate Change
Computer Simulation in Plant Breeding 253

(IPCC) report projects that the global surface temperature is likely to rise a
further 1.1–6.4  C (2.0–11.5F) during the twenty-first century (Christensen
et al., 2007). How would the crops respond to this climate change, and what
kind of cultivars could adapt to the changing environment? These questions
are of great concern to plant breeders and policymakers because food
security is strategically important to society.
Climate change research in plant breeding is not intended to provide
information about how to improve the efficiency of plant breeding, but to
determine our direction. For example, the advanced global general circula-
tion models (GCMs) can assess the effects of different emissions on climate
change. These data can be used to determine the possible future importance
of specific traits in a geographic region (Masutomi et al., 2009). GIS
technologies could benefit plant breeding by improving our understanding
of crop production areas and environments at both the global and local
scales, providing insights into potential shifts in production environments
with changing climate and predicting the possible spread of epidemics
(Hodson and White, 2010). Modeling, using this kind of information,
would allow us to make early decisions and prepare for the likely influences
of climate change on crop growth.
Regression-based models and dynamic process-based models are used to
assess the influence of climate change on crop production. Masutomi et al.
(2009) used 49 GCM models to predict the impact of climate change on rice
production in Asia. A large number of GCMs were used to account for the
inherent process/parameter uncertainty in GCMs. Under all future climate
scenarios, assuming no technological development, average production fell
for 2020 and 2080. CO2 fertilization was offset by the detrimental effect of
climate change. Under the scenario with the lowest atmospheric CO2
concentration, however, the simulation showed a small decrease in produc-
tion and a much lower probability of production decrease. Technological
development, when included in assessing the effects of climate change,
influenced production by only 10%.
Zou and Zeng (2009) investigated the combined effects of climatic con-
ditions, CO2 concentration, and technology development on wheat produc-
tion in Europe from 2000 to 2080. Technology development refers to all
measures related to crop management (e.g., improved machinery, pesticides,
herbicides, etc., and agronomic knowledge of farmers) and breeding (devel-
opment of higher yielding varieties through improved stress resistance and/or
yield potential) that result in yield increase. The estimates suggested increases
in crop productivity ranging from 25% to 163% depending on time slice and
scenario compared to the baseline year of 2000. The predicted changes in
productivity were mostly due to the effects of technology development.
Surprisingly, the CO2 increase did not have pronounced effects on wheat
production, and the differences between regions were also not significant,
even though climate change might make the northern part of Europe more
254 Xin Li et al.

amenable to wheat growth than the south. Typically, the further the predic-
tions are in the future, the less accurate the estimate. Both these studies (Lobell
et al., 2008; Masutomi et al., 2009) used multiple GCMs (20 and 49, respec-
tively) to control the uncertainty in the estimations.
Lobell et al. (2008) showed that South Asia and Southern Africa would
suffer from climate change by 2030 based on different models of climate
change and crop response. Some priorities were set: Southern Africa: maize,
wheat, and sugarcane; Western Africa: yams and groundnut; and Sahel: wheat.

3. Computation and Software Issues

Advances in computer science have made computer simulation possi-
ble. The exponential increase in computational speed, as predicted by
Moore’s law (which holds that the number of transistors on a single chip
doubles roughly every 2 years), has made significant progress in computer
simulation in various research areas. With parallel computing, the time
needed for a single simulation experiment can be reduced from several
months or even years to several hours or days. However, the demand for
higher computational speed never ceases. For commercial seed companies,
shortening the time to release a new cultivar is important to remain
competitive. In some commercial seed companies, scientists are using
high-performance computers to simulate the breeding process and predict
the results, instead of waiting for months for field trials to be complete.
The software commonly used for computer simulation can be divided
into two categories: off-the-shelf software and custom-designed simulation
programs (Table 1). Off-the-shelf software is usually user-friendly but is not
flexible enough because it cannot accommodate all possible scenarios. This
kind of software, however, is suitable for industry, such as using QU-GENE
for selection. On the other hand, most researchers in academic institutions
prefer custom programs designed for a specific research question, like the
simulation programs in the gene-mapping section. After these custom
programs are verified, the researchers may release the software to the public.
Authors often write their programs for computer simulation using a
programming language they already know such as Cþþ, FORTRAN,
Python, and Java. For plant breeding, the core of the simulation program is
the algorithm used to simulate meiosis (Frisch et al., 2000). In PLABSIM, this
process uses a pseudorandom number generator coupled with Haldane’s
mapping function. The pseudorandom number generator provides a uni-
formly distributed random variable 0Z1 for each pair of adjacent loci. If,
and only if, r, the recombination frequency, is greater than or equal to Z, a
crossover between these two loci can occur. A large number of runs are
needed to acquire the statistical properties of each simulation scenario.
Computer Simulation in Plant Breeding 255

Table 1 A list of software commonly used in computer simulation in plant breeding

Name Utility References Online access

APSIM Crop modeling McCown et al. http://www.apsim.
(1996), info/Wiki/
Keating et al.
(2003)
PLABSIM Marker-assisted Frisch et al.
backcrossing (2000)
PLABSOFT Plant breeding Maurer et al. http://www.r-project.
(2008) org/
QU-GENE Genotype-by- Podlich and http://www.uq.edu.
environment Cooper au/lcafs/qugene/
interaction and (1998)
plant breeding
E-CELL Whole-cell Tomita et al. http://www.e-cell.
simulation (1999) org/ecell/

Commonly used variables for different runs include the number of QTLs and
markers, QTL position on the genetic map, genome size, the effect of each
QTL, population size, heritability, degree of epistasis, and environment types.
After meeting the computational challenge, researchers must analyze and
interpret computer output. As with statistical software, computer simulation
gives us only the numbers, but the numbers require interpretation to be
meaningful and practical. Computer simulation must be coupled with
human intelligence to be useful.

4. Summary and Perspectives

In summary, the main points we presented in this review included:
1. Computer simulation, a bridge between theory and experimentation,
has become a powerful tool in scientific research. It can be used to
conduct pilot or virtual experiments to verify new theories or provide
guidelines for empirical experimentation.
2. Huge amounts of information have been generated in crop improve-
ment research during the past several decades, especially significant
advances in molecular dissecting of complex traits and high-throughput
genotyping techniques. Computer simulation can transfer these advance-
ments into plant-breeding practice.
3. Computer simulation can compare different breeding strategies, incorpor-
ating gene information, cross scheme, propagation method, population
256 Xin Li et al.

size, selection intensity, and number of generations simultaneously; thus,

we can use computer simulation to decide which breeding strategy could
lead to higher genetic gain.
4. Computer simulation can be applied to gene mapping to validate the
effectiveness of new mapping methods or assess the factors influencing
mapping power (e.g., population type and size, marker number and
density, heritability, and number of QTLs). Computer simulation can
also help us determine the significance threshold and CI, which other-
wise would be difficult to calculate analytically.
5. Plant-breeding simulation platforms are potent tools that can simulate the
whole plant-breeding process. They use genetic and GEI information to,
for instance, predict cross performance and compare selection methods,
thus enhancing our ability to make decisions about plant breeding.
6. Computer simulation can integrate crop physiological models, environ-
mental information, and genetic composition of different crops to fill the
gap between genotype and phenotype. We can use computer simulation
to predict the performance of different cultivars in TPEs and thus
facilitate the plant-breeding process. When coupled with climate simu-
lation models, crop models can be used to predict the possible influences
of climate change on crop production, which can subsequently provide
guidelines for plant breeding.
7. With the exponential increase in computational power and decrease in
the price of that power, computer simulation will become more com-
mon, but custom programs tailored for specific purposes will still need to
be developed.
At the beginning of the second decade of the twenty-first century, we
are closer than at any other point in history to deciphering various mysteries
in life science. Huge amounts of information in genetics, genomics, bio-
chemistry, molecular biology, and bioinformatics are now available, but
only some of this information has been applied to plant breeding. The
practical goal of scientific research is more than just explaining the mechan-
isms underlying life phenomena—it is learning to manipulate those
mechanisms to benefit humankind. Plant breeders have the challenge of
determining how to take advantage of this knowledge to make crop
improvement more efficient and enhance genetic gain.
Research in establishing genotype–phenotype relationship and develop-
ing new breeding methods has been proposed to realize the potential brought
about by ultrahigh-throughput genomic technologies in plant breeding (Yu,
2009), and computer simulation undoubtedly will be a key part of this
process. As a tool, computer simulation will aid decision-making and resource
allocation through transferring experimental results from laboratory to agri-
cultural production and by predicting the outcome of breeding decisions,
directing gene mapping, and tackling GEI and climate change.
Computer Simulation in Plant Breeding 257

ACKNOWLEDGMENTS
This work was supported by the Agriculture and Food Research Initiative Competitive
Grant (2011-03587) from the USDA National Institute of Food and Agriculture, the Plant
Feedstock Genomics Program (DE-SC0002259) of the U.S. Department of Energy, and the
Plant Genome Program (DBI-0820610) of the National Science Foundation, the Targeted
Excellence Program of Kansas State University, and the Kansas State University Center for
Sorghum Improvement.

REFERENCES
Ahuja, I., de Vos, R. C. H., Bones, A. M., and Hall, R. D. (2010). Plant molecular stress
responses face climate change. Trends Plant Sci. 15(12), 664–674.
Allard, R. W. (1960). Principles of Plant Breeding. John Wiley and Sons Inc., New York.
Asseng, S., Turner, N. C., Botwright, T., and Condon, A. G. (2003). Evaluating the impact
of a trait for increased specific leaf area on wheat yields using a crop simulation model.
Agron. J. 95(1), 10–19.
Barczi, J. F., Rey, H., Caraglio, Y., De Reffye, P., Barthelemy, D., Dong, Q. X., et al.
(2008). AmapSim: A structural whole-plant simulator based on botanical knowledge and
designed to host external functional models. Ann. Bot. 101(8), 1125–1138.
Bauer, A. M., Hoti, F., Korff, M.v., Pillen, K., Leon, J., and Sillanpaa, M. J. (2009).
Advanced backcross-QTL analysis in spring barley (H. vulgare ssp. spontaneum) com-
paring a REML versus a bayesian model in multi-environmental field trials. Theor. Appl.
Genet. 119(1), 105–123.
Beavis, W. D. (1998). QTL analyses: Power, precision, and accuracy. In “Molecular Dissec-
tion of Complex Traits” (A. H. Paterson, Ed.), pp. 145–162. CRC press, New York.
Beavis, W. (1994). The power and deceit of QTL experiments: Lessons from comparative
QTL studies. In “Proceedings of the 49th Annual Corn and Sorghum Industry Research
Conference,” pp. 250–266.
Bernacchi, D., Beck-Bunn, T., Emmatty, D., Eshed, Y., Inai, S., Lopez, J., et al. (1998a).
Advanced backcross QTL analysis of tomato. II. Evaluation of near-isogenic lines
carrying single-donor introgressions for desirable wild QTL-alleles derived from lyco-
persicon hirsutum and L. pimpinellifolium. Theor. Appl. Genet. 97(1), 170–180.
Bernacchi, D., Beck-Bunn, T., Eshed, Y., Lopez, J., Petiard, V., Uhlig, J., et al. (1998b).
Advanced backcross QTL analysis in tomato. I. Identification of QTLs for traits of
agronomic importance from lycopersicon hirsutum. Theor. Appl. Genet. 97(3), 381–397.
Bernardo, R., and Yu, J. (2007). Prospects for genomewide selection for quantitative traits
in maize. Crop Sci. 47(3), 1082–1090.
Bernardo, R. (1999a). Marker-assisted best linear unbiased prediction of single-cross perfor-
mance. Crop Sci. 39(5), 1277–1282.
Bernardo, R. (1999b). Selection response with marker-based assortative mating. Crop Sci. 39
(1), 69–73.
Bernardo, R. (2001). What if we knew all the genes for a quantitative trait in hybrid crops?
Crop Sci. 41(1), 1–4.
Bernardo, R. (2002). Breeding for Quantitative Traits in Plants. Stemma Press, Woodbury,
MN.
Bernardo, R. (2003). Parental selection, number of breeding populations, and size of each
population in inbred development. Theor. Appl. Genet. 107(7), 1252–1256.
Bernardo, R. (2004). What proportion of declared QTL in plants are false? Theor. Appl.
Genet. 109(2), 419–424.
258 Xin Li et al.

Bernardo, R. (2008). Molecular markers and selection for complex traits in plants: Learning
from the last 20 years. Crop Sci. 48(5), 1649–1664.
Bernardo, R. (2009). Genomewide selection for rapid introgression of exotic germplasm in
maize. Crop Sci. 49(2), 419–425.
Bernardo, R. (2010). Genomewide selection with minimal crossing in self-pollinated crops.
Crop Sci. 50(2), 624–627.
Bernardo, R., and Charcosset, A. (2006). Usefulness of gene information in marker-assisted
recurrent selection: A simulation appraisal. Crop Sci. 46(2), 614–621.
Bernardo, R., Moreau, L., and Charcosset, A. (2006). Number and fitness of selected
individuals in marker-assisted and phenotypic recurrent selection. Crop Sci. 46(5), 1972–
1980.
Bogdan, M., Ghosh, J. K., and Doerge, R. W. (2004). Modifying the schwarz bayesian
information criterion to locate multiple interacting quantitative trait loci. Genetics 167(2),
989–999.
Botstein, D., White, R. L., Skolnick, M., and Davis, R. W. (1980). Construction of a
genetic-linkage map in man using restriction fragment length polymorphisms. Am. J.
Hum. Genet. 32(3), 314–331.
Chapman, S., Cooper, M., Podlich, D., and Hammer, G. (2003). Evaluating plant
breeding strategies by simulating gene action and dryland environment effects. Agron.
J. 95(1), 99–113.
Chapman, S. C. (2008). Use of crop models to understand genotype by environment
interactions for drought in real-world and simulated plant breeding trials. Euphytica 161
(1), 195–208.
Chapman, S. C., Cooper, M., Hammer, G. L., and Butler, D. G. (2000). Genotype by
environment interactions affecting grain sorghum. II. Frequencies of different seasonal
patterns of drought stress are related to location effects on hybrid yields. Aust. J. Agr. Res.
51(2), 209–221.
Chapman, S. C., Wang, J., Rebetzke, G. J., and Bonnett, D. G. (2007). Accounting for
variability in the detection and use of markers for simple and complex traits. In “Scale and
Complexity in Plant Systems Research: Gene-Plant-Crop Relations” ( J. H. J. Spiertx,
P. Struik, and H. H. van Laar, Eds.), pp. 37–44. Springer, Dordrecht, Netherlands.
Charmet, G., Robert, N., Perretant, M. R., Gay, G., Sourdille, P., Groos, C., et al. (1999).
Marker-assisted recurrent selection for cumulating additive and interactive QTLs in
recombinant inbred lines. Theor. Appl. Genet. 99(7–8), 1143–1148.
Chen, X., Zhao, F., and Xu, S. (2010). Mapping environment-specific quantitative trait loci.
Genetics 186(3), 1053–1066.
Chenu, K., Chapman, S. C., Tardieu, F., McLean, G., Welcker, C., and Hammer, G. L.
(2009). Simulating the yield impacts of organ-level quantitative trait loci associated with
drought response in maize: A “gene-to-phenotype” modeling approach. Genetics 183(4),
1507–1523.
Chesler, E. J., Rodriguez-Zas, S. L., and Mogil, J. S. (2001). In silico mapping of mouse
quantitative trait loci. Science 294(5551), 2423.
Christensen, J., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, R., et al. (2007).
Regional climate projections. In “Climate change, 2007: The physical science basis.
Contribution of working group I to the fourth assessment report of the intergovernmen-
tal panel on climate change” (S. Solomon, D. Qni, M. Manning, Z. Chen, M. Marquis,
K. B. Averyt, M. Tignor, and H. L. Miller, Eds.), Cambridge University Press, Cam-
bridge, United Kingdom and New York, NY, USA.
Churchill, G. A., and Doerge, R. W. (1994). Empirical threshold values for quantitative trait
mapping. Genetics 138(3), 963–971.
Cooper, M., Chapman, S. C., Podlich, D. W., and Hammer, G. L. (2002). The GP
problem: Quantifying gene-to-phenotype relationships. In Silico Biol. 2(2), 151–164.
Computer Simulation in Plant Breeding 259

Cooper, M., Podlich, D. W., and Luo, L. (2007). Modeling QTL effects and MAS in plant
breeding. In “Genomics-Assisted Crop Improvement” (R. K. Varshney and R. Tuberosa,
Eds.), pp. 57–95. Springer, Dordrecht, The Netherlands.
Crepieux, S., Lebreton, C., Servin, B., and Charmet, G. (2004). Quantitative trait loci
(QTL) detection in multicross inbred designs: Recovering QTL identical-by-descent
status information from marker data. Genetics 168(3), 1737–1749.
Darvasi, A., and Soller, M. (1997). A simple method to calculate resolving power and
confidence interval of QTL map location. Behav. Genet. 27(2), 125–132.
Darvasi, A., Weinreb, A., Minke, V., Weller, J. I., and Soller, M. (1993). Detecting marker-
qtl linkage and estimating qtl gene effect and map location using a saturated genetic-map.
Genetics 134(3), 943–951.
de Dorlodot, S., Forster, B., Pagčs, L., Price, A., Tuberosa, R., and Draye, X. (2007). Root
system architecture: Opportunities and constraints for genetic improvement of crops.
Trends Plant Sci. 12(10), 474–481.
Doerge, R. W. (2002). Mapping and analysis of quantitative trait loci in experimental
populations. Nat. Rev. Genet. 3(1), 43–52.
Dupuy, L., Gregory, P. J., and Bengough, A. G. (2010). Root growth models: Towards a
new generation of continuous approaches. J. Exp. Bot. 61(8), 2131–2143.
Edwards, M. D., and Page, N. J. (1994). Evaluation of marker-assisted selection through
computer-simulation. Theor. Appl. Genet. 88(3–4), 376–382.
Evers, J. B., Vos, J., Yin, X., Romero, P., van der Putten, P. E. L., and Struik, P. C. (2010).
Simulation of wheat growth and development based on organ-level photosynthesis and
assimilate allocation. J. Exp. Bot. 61(8), 2203–2216.
Falconer, D. S. (1960). Introduction to Quantitative Genetics. Oliver and Boyd, London.
Frisch, M., Bohn, M., and Melchinger, A. (2000). Computer note. PLABSIM: Software for
simulation of marker-assisted backcrossing. J. Hered. 91(1), 86.
Frisch, M., Bohn, M., and Melchinger, A. E. (1999). Comparison of selection strategies for
marker-assisted backcrossing of a gene. Crop Sci. 39(5), 1295–1301.
Frisch, M., and Melchinger, A. E. (2001). Marker-assisted backcrossing for simultaneous
introgression of two genes. Crop Sci. 41(6), 1716–1725.
Frisch, M., and Melchinger, A. E. (2005). Selection theory for marker-assisted backcrossing.
Genetics 170(2), 909–917.
Gimelfarb, A., and Lande, R. (1994). Simulation of marker-assisted selection in hybrid
populations. Genet. Res. 63(1), 39–47.
Gimelfarb, A., and Lande, R. (1995). Marker-assisted selection and marker-qtl associations
in hybrid populations. Theor. Appl. Genet. 91(3), 522–528.
Grafahrend-Belau, E., Schreiber, F., Koschutzki, D., and Junker, B. H. (2009). Flux balance
analysis of barley seeds: A computational approach to study systemic properties of central
metabolism. Plant Physiol. 149(1), 585–598.
Grupe, A., Germer, S., Usuka, J., Aud, D., Belknap, J. K., Klein, R. F., et al. (2001). In silico
mapping of complex disease-related traits in mice. Science 292(5523), 1915–1918.
Hammer, G., Kropff, M., Sinclair, T., and Porter, J. (2002). Future contributions of crop
modelling—From heuristics and supporting decision making to understanding genetic
regulation and aiding crop improvement. Eur. J. Agron. 18(1–2), 15–31.
Han, L., and Xu, S. (2010). Genome-wide evaluation for quantitative trait loci under the
variance component model. Genetica 138(9–10), 1099–1109.
Hodson, D., and White, J. (2010). GIS and crop simulation modelling applications in climate
change research. In “Climate Change and Crop Production” (M. P. Reynolds, Ed.),
pp. 245–262. CABI, Oxfordshire, UK.
Holland, J. B. (2004). Implementation of molecular markers for quantitative traits in breeding
programs—challenges and opportunities. In “New Directions for a Diverse Planet: Pro-
ceedings for the 4th International Crop Science Congress, Brisbane, Australia, 26”.
260 Xin Li et al.

Hoogenboom, G., White, J. W., and Messina, C. D. (2004). From genome to crop:
Integration through simulation modeling. Field Crop Res 90(1), 145–163.
Hospital, F., and Charcosset, A. (1997). Marker-assisted introgression of quantitative trait
loci. Genetics 147(3), 1469–1485.
Hospital, F., Chevalet, C., and Mulsant, P. (1992). Using markers in gene introgression
breeding programs. Genetics 132(4), 1199–1210.
Hospital, F., Moreau, L., Lacoudre, F., Charcosset, A., and Gallais, A. (1997). More on the
efficiency of marker-assisted selection. Theor. Appl. Genet. 95(8), 1181–1189.
Hospital, F. (2001). Size of donor chromosome segments around introgressed loci and
reduction of linkage drag in marker-assisted backcross programs. Genetics 158(3),
1363–1379.
Jannink, J. L., and Jansen, R. (2001). Mapping epistatic quantitative trait loci with one-
dimensional genome searches. Genetics 157(1), 445–454.
Jannink, J. L., Lorenz, A. J., and Iwata, H. (2010). Genomic selection in plant breeding:
From theory to practice. Brief. Funct. Genomics 9(2), 166–177.
Jansen, R. C. (1993). Interval mapping of multiple quantitative trait loci. Genetics 135(1),
205–211.
Jansen, R. C., Jannink, J. L., and Beavis, W. D. (2003). Mapping quantitative trait loci in
plant breeding populations: Use of parental haplotype sharing. Crop Sci. 43, 829–834.
Jiang, C. J., and Zeng, Z. B. (1997). Mapping quantitative trait loci with dominant and
missing markers in various crosses from two inbred lines. Genetica 101(1), 47–58.
Johnson, R. (2004). Marker-assisted selection. Plant Breed. Rev. 24(1), 293–309.
Kang, M. S. (2001). Quantitative genetics, genomics and plant breeding. In “Quantitative
Genetics Genomics and Plant Breeding” (M. S. Kang, Ed.). xvi þ 400 pp. CABI,
Wallingford, UK.
Kao, C. H., Zeng, Z. B., and Teasdale, R. D. (1999). Multiple interval mapping for
quantitative trait loci. Genetics 152(3), 1203–1216.
Keating, B. A., Carberry, P., Hammer, G., Probert, M. E., Robertson, M., Holzworth, D.,
et al. (2003). An overview of APSIM, a model designed for farming systems simulation.
Eur. J. Agron. 18(3–4), 267–288.
Kuchel, H., Fox, R., Reinheimer, J., Mosionek, L., Willey, N., Bariana, H., et al. (2007).
The successful application of a marker-assisted wheat breeding strategy. Mol. Breed. 20(4),
295–308.
Kuchel, H., Ye, G., Fox, R., and Jefferies, S. (2005). Genetic and economic analysis of a
targeted marker-assisted wheat breeding strategy. Mol. Breed. 16(1), 67–78.
Lande, R., and Thompson, R. (1990). Efficiency of marker-assisted selection in the
improvement of quantitative traits. Genetics 124(3), 743–756.
Lander, E. S., and Botstein, D. (1989). Mapping mendelian factors underlying quantitative
traits using rflp linkage maps. Genetics 121(1), 185–199.
Letort, V., Mahe, P., Cournede, P. H., De Reffye, P., and Courtois, B. (2008). Quantitative
genetics and functional-structural plant growth models: Simulation of quantitative trait
loci detection for model parameters and application to potential yield optimization. Ann.
Bot. 101(8), 1243–1254.
Leuning, R., Kelliher, F., Pury, D., and Schulze, E. D. (1995). Leaf nitrogen, photosynthe-
sis, conductance and transpiration: Scaling from leaves to canopies. Plant Cell Environ. 18
(10), 1183–1200.
Li, H. H., Ye, G. Y., and Wang, J. K. (2007). A modified algorithm for the improvement of
composite interval mapping. Genetics 175(1), 361–374.
Lobell, D. B., Burke, M. B., Tebaldi, C., Mastrandrea, M. D., Falcon, W. P., and
Naylor, R. L. (2008). Prioritizing climate change adaptation needs for food security in
2030. Science 319(5863), 607–610.
Computer Simulation in Plant Breeding 261

Loffler, C. M., Wei, J., Fast, T., Gogerty, J., Langton, S., Bergman, M., et al. (2005).
Classification of maize environments using crop simulation and geographic information
systems. Crop Sci. 45(5), 1708–1716.
Lynch, M., and Walsh, B. (1997). Genetics and Analysis of Quantitative Traits. Sinauer
Associates, Inc., Sunderland, MA.
Mackay, T. F. C. (2001). The genetic architecture of quantitative traits. Annu. Rev. Genet.
35, 303–339.
Manichaikul, A., Moon, J. Y., Sen, S., Yandell, B. S., and Broman, K. W. (2009). A model
selection approach for the identification of quantitative trait loci in experimental crosses,
allowing epistasis. Genetics 181(3), 1077–1086.
Masutomi, Y., Takahashi, K., Harasawa, H., and Matsuoka, Y. (2009). Impact assessment of
climate change on rice production in Asia in comprehensive consideration of process/
parameter uncertainty in general circulation models. Agric. Ecosyst. Environ. 131(3–4),
281–291.
Mather, K. (1949). Biometrical Genetics. Methuen, London.
Maurer, H. P., Melchinger, A. E., and Frisch, M. (2008). Population genetic simulation and
data analysis with Plabsoft. Euphytica 161(1–2), 133–139.
Mayor, P. J., and Bernardo, R. (2009). Genomewide selection and marker-assisted recurrent
selection in doubled haploid versus F-2 populations. Crop Sci. 49(5), 1719–1725.
McCown, R., Hammer, G., Hargreaves, J., Holzworth, D., and Freebairn, D. (1996).
APSIM: A novel software system for model development, model testing and simulation
in agricultural systems research. Agric. Syst. 50(3), 255–271.
Messina, C., Jones, J., Boote, K., and Vallejos, C. (2006). A gene-based model to
simulate soybean development and yield responses to environment. Crop Sci. 46(1),
456–466.
Meuwissen, T. H. E., Karlsen, A., Lien, S., Olsaker, I., and Goddard, M. E. (2002). Fine
mapping of a quantitative trait locus for twinning rate using combined linkage and
linkage disequilibrium mapping. Genetics 161(1), 373–379.
Meuwissen, T. H. E., Hayes, B. J., and Goddard, M. E. (2001). Prediction of total genetic
value using genome-wide dense marker maps. Genetics 157(4), 1819–1829.
Myles, S., Peiffer, J., Brown, P. J., Ersoz, E. S., Zhang, Z. W., Costich, D. E., et al. (2009).
Association mapping: Critical considerations shift from genotyping to experimental
design. Plant Cell 21(8), 2194–2202.
Ooijen, J. W. (1992). Accuracy of mapping quantitative trait loci in autogamous species.
Theor. Appl. Genet. 84(7), 803–811.
Parisseaux, B., and Bernardo, R. (2004). In Silico Mapping of Quantitative Trait Loci in
Maize. Springer, Berlin/Heidelberg.
Peccoud, J., Vander Velden, K., Podlich, D., Winkler, C., Arthur, L., and Cooper, M.
(2004). The selective values of alleles in a molecular network model are context depen-
dent. Genetics 166(4), 1715–1725.
Peleman, J. D., and van der Voort, J. R. (2003). Breeding by design. Trends Plant Sci. 8(7),
330–334.
Pillen, K., Zacharias, A., and Leon, J. (2003). Advanced backcross QTL analysis in barley
(Hordeum vulgare L.). Theor. Appl. Genet. 107(2), 340–352.
Podlich, D., Cooper, M., Basford, K., and Geiger, H. (1999). Computer simulation of a
selection strategy to accommodate genotype environment interactions in a wheat recur-
rent selection programme. Plant Breed. 118(1), 17–28.
Podlich, D. W., and Cooper, M. (1998). QU-GENE: A simulation platform for quantitative
analysis of genetic models. Bioinformatics 14(7), 632–653.
Podlich, D. W., Winkler, C. R., and Cooper, M. (2004). Mapping as you go: An effective
approach for marker-assisted selection of complex traits. Crop Sci. 44(5), 1560–1571.
262 Xin Li et al.

Price, A. L., Patterson, N. J., Plenge, R. M., Weinblatt, M. E., Shadick, N. A., and
Reich, D. (2006). Principal components analysis corrects for stratification in genome-
wide association studies. Nat. Genet. 38(8), 904–909.
Pritchard, J. K., Stephens, M., Rosenberg, N. A., and Donnelly, P. (2000). Association
mapping in structured populations. Am. J. Hum. Genet. 67(1), 170–181.
Prusinkiewicz, P., Hammel, M., Hanan, J., and Mech, R. (1996). L-systems: From the
theory to visual models of plants. In “Proceedings of the Second CSIRO Symposium on
Computational Challenges in Life Sciences”.
Prusinkiewicz, P. (2004). Art and science for life: Designing and growing virtual plants
with L-systems. In “Nursery Crops Development, Evaluation, Production and Use”
(C. Davidson and T. Fernandez, Eds.), Vol. 630, pp. 15–28. ISHS, Leuven, Belgium.
Reymond, M., Muller, B., Leonardi, A., Charcosset, A., and Tardieu, F. (2003). Combining
quantitative trait loci analysis and an ecophysiological model to analyze the genetic
variability of the responses of maize leaf growth to temperature and water deficit. Plant
Physiol. 131(2), 664–675.
Ribaut, J., Jiang, C., and Hoisington, D. (2002). Simulation experiments on efficiencies of
gene introgression by backcrossing. Crop Sci. 42(2), 557–565.
Risch, N., and Merikangas, K. (1996). The future of genetic studies of complex human
diseases. Science 273(5281), 1516–1517.
Sahana, G., Guldbrandtsen, B., Janss, L., and Lund, M. S. (2010). Comparison of association
mapping methods in a complex pedigreed population. Genet. Epidemiol. 34(5), 455–462.
Satagopan, J. M., Yandell, Y. S., Newton, M. A., and Osborn, T. C. (1996). A Bayesian
approach to detect quantitative trait loci using Markov chain Monte Carlo. Genetics 144
(2), 805–816.
Schauer, N., Semel, Y., Roessner, U., Gur, A., Balbo, I., Carrari, F., et al. (2006).
Comprehensive metabolic profiling and phenotyping of interspecific introgression lines
for tomato improvement. Nat. Biotechnol. 24(4), 447–454.
Schon, C. C., Utz, H. F., Groh, S., Truberg, B., Openshaw, S., and Melchinger, A. E.
(2004). Quantitative trait locus mapping based on resampling in a vast maize testcross
experiment and its relevance to quantitative genetics for complex traits. Genetics 167(1),
485–498.
Slafer, G. A. (2003). Genetic basis of yield as viewed from a crop physiologist’s perspective.
Ann. Appl. Biol. 142(2), 117–128.
Soller, M., Brody, T., and Genizi, A. (1976). Power of experimental designs for detection of
linkage between marker loci and quantitative loci in crosses between inbred lines. Theor.
Appl. Genet. 47(1), 35–39.
Stich, B. (2009). Comparison of mating designs for establishing nested association mapping
populations in maize and Arabidopsis thaliana. Genetics 183(4), 1525–1534.
Takahashi, K., Ishikawa, N., Sadamoto, Y., Sasamoto, H., Ohta, S., Shiozawa, A., et al.
(2003). E-cell 2: Multi-platform E-cell simulation system. Bioinformatics 19(13), 1727–
1729.
Tanksley, S., and Nelson, J. (1996). Advanced backcross QTL analysis: A method for the
simultaneous discovery and transfer of valuable QTLs from unadapted germplasm into
elite breeding lines. Theor. Appl. Genet. 92(2), 191–203.
Tomita, M. (2001). Whole-cell simulation: A grand challenge of the 21st century. Trends
Biotechnol. 19(6), 205–210.
Tomita, M., Hashimoto, K., Takahashi, K., Shimizu, T. S., Matsuzaki, Y., Miyoshi, F., et al.
(1999). E-CELL: Software environment for whole-cell simulation. Bioinformatics 15(1),
72–84.
van Berloo, R., and Stam, P. (1998). Marker-assisted selection in autogamous RIL popula-
tions: A simulation study. Theor. Appl. Genet. 96(1), 147–154.
Computer Simulation in Plant Breeding 263

van Berloo, R., and Stam, P. (2001). Simultaneous marker-assisted selection for multiple
traits in autogamous crops. Theor. Appl. Genet. 102(6–7), 1107–1112.
Van Ooijen, J. W. (1999). LOD significance thresholds for QTL analysis in experimental
populations of diploid species. Heredity 83, 613–624.
Vanoeveren, A. J., and Stam, P. (1992). Comparative simulation studies on the effects of
selection for quantitative traits in autogamous crops—Early selection versus single seed
descent. Heredity 69, 342–351.
Vieland, V. J. (1998). Bayesian linkage analysis, or: How I learned to stop worrying and love
the posterior probability of linkage. Am. J. Hum. Genet. 63(4), 947–954.
Visscher, P. M., Haley, C. S., and Thompson, R. (1996a). Marker-assisted introgression in
backcross breeding programs. Genetics 144(4), 1923–1932.
Visscher, P. M., Thompson, R., and Haley, C. S. (1996b). Confidence intervals in QTL
mapping by bootstrapping. Genetics 143(2), 1013–1020.
Wang, J. K., Chapman, S. C., Bonnett, D. G., and Rebetzke, G. J. (2009a). Simultaneous
selection of major and minor genes: Use of QTL to increase selection efficiency of
coleoptile length of wheat (Triticum aestivum L.). Theor. Appl. Genet. 119(1), 65–74.
Wang, J. K., and Pfeiffer, W. H. (2007). Simulation modeling in plant breeding: Principles
and applications. Agric. Sci. China 6(8), 908–921.
Wang, J. K., Singh, R. P., Braun, H. J., and Pfeiffer, W. H. (2009b). Investigating the
efficiency of the single backcrossing breeding strategy through computer simulation.
Theor. Appl. Genet. 118(4), 683–694.
Wang, J. K., Eagles, H. A., Trethowan, R., and van Ginkel, M. (2005). Using computer
simulation of the selection process and known gene information to assist in parental
selection in wheat quality breeding. Aust. J. Agric. Res. 56(5), 465–473.
Wang, J. K., Chapman, S. C., Bonnett, D. G., Rebetzke, G. J., and Crouch, J. (2007).
Application of population genetic theory and simulation models to efficiently pyramid
multiple genes via marker-assisted selection. Crop Sci. 47(2), 582–590.
White, J. W., and Hoogenboom, G. (2003). Gene-based approaches to crop simulation: Past
experiences and future opportunities. Agron. J. 95(1), 52–64.
Winsberg, E. (2010). Science in the Age of Computer Simulation. University of Chicago
Press, Chicago.
Wong, C., and Bernardo, R. (2008). Genomewide selection in oil palm: Increasing selection
gain per unit time and cost with small populations. Theor. Appl. Genet. 116(6), 815–824.
Wu, R. L., and Zeng, Z. B. (2001). Joint linkage and linkage disequilibrium mapping in
natural populations. Genetics 157(2), 899–909.
Xu, S. Z. (2010). An expectation-maximization algorithm for the lasso estimation of
quantitative trait locus effects. Heredity 105(5), 483–494.
Xu, S. Z. (1998). Mapping quantitative trait loci using multiple families of line crosses.
Genetics 148(1), 517–524.
Xu, S. Z. (2003). Estimating polygenic effects using markers of the entire genome. Genetics
163(2), 789–801.
Xu, Y. (2010). Bayesian mapping. In “Molecular Plant Breeding” (Y. Xu, Ed.), pp. 219–
223. CAB International, Wallingford, UK.
Yang, J., Zhu, J., and Williams, R. W. (2007). Mapping the genetic architecture of complex
traits in experimental populations. Bioinformatics 23(12), 1527–1536.
Yin, X., Kropff, M. J., and Stam, P. (1999). The role of ecophysiological models in QTL
analysis: The example of specific leaf area in barley. Heredity 82(4), 415–421.
Yin, X., Stam, P., Kropff, M. J., and Schapendonk, A. (2003). Crop modeling, QTL
mapping, and their complementary role in plant breeding. Agron. J. 95(1), 90–98.
Yin, X., Struik, P. C., and Kropff, M. J. (2004). Role of crop physiology in predicting gene-
to-phenotype relationships. Trends Plant Sci. 9(9), 426–432.
264 Xin Li et al.

Yin, X., Struik, P. C., Van Eeuwijk, F. A., Stam, P., and Tang, J. (2005). QTL analysis and
QTL-based prediction of flowering phenology in recombinant inbred lines of barley.
J. Exp. Bot. 56(413), 967.
Yin, X., and Struik, P. C. (2010). Modelling the crop: From system dynamics to systems
biology. J. Exp. Bot. 61(8), 2171–2183.
Yu, J. (2009). Realizing the potential of ultrahigh throughput genomic technologies in plant
breeding. Plant Genome 2, 2.
Yu, J., Holland, J. B., McMullen, M. D., and Buckler, E. S. (2008). Genetic design and
statistical power of nested association mapping in maize. Genetics 178(1), 539.
Yu, J., Pressoir, G., Briggs, W. H., Bi, I. V., Yamasaki, M., Doebley, J. F., et al. (2006).
A unified mixed-model method for association mapping that accounts for multiple levels
of relatedness. Nat. Genet. 38(2), 203–208.
Yu, J., Zhang, Z., Zhu, C., Tabanao, D. A., Pressoir, G., Tuinstra, M. R., et al. (2009).
Simulation appraisal of the adequacy of number of background markers for relationship
estimation in association mapping. Plant Genome 2(1), 63–77.
Yu, J., Arbelbide, M., and Bernardo, R. (2005). Power of in silico QTL mapping from
phenotypic, pedigree, and marker data in a hybrid breeding program. Theor. Appl. Genet.
110(6), 1061–1067.
Zeng, Z. B. (1994). Precision mapping of quantitative trait loci. Genetics 136(4), 1457.
Zeng, Z. B., Kao, C. H., and Basten, C. J. (1999). Estimating the genetic architecture of
quantitative traits. Genet. Res. 74(3), 279–289.
Zhang, L. Y., Wang, S. Q., Li, H. H., Deng, Q. M., Zheng, A. P., Li, S. C., et al. (2010).
Effects of missing marker and segregation distortion on QTL mapping in F-2 populations.
Theor. Appl. Genet. 121(6), 1071–1082.
Zhang, W., and Smith, C. (1992). Computer-simulation of marker-assisted selection utiliz-
ing linkage disequilibrium. Theor. Appl. Genet. 83(6–7), 813–820.
Zhang, W., and Smith, C. (1993). Simulation of marker-assisted selection utilizing linkage
disequilibrium—The effects of several additional factors. Theor. Appl. Genet. 86(4),
492–496.
Zhu, C., Gore, M., Buckler, E. S., and Yu, J. (2008). Status and prospects of association
mapping in plants. Plant Genome 1(1), 5–20.
Zhu, C., and Yu, J. (2009). Nonmetric multidimensional scaling corrects for population
structure in association mapping with different sample types. Genetics 182(3), 875–888.
Zou, W., and Zeng, Z. B. (2009). Multiple interval mapping for gene expression QTL
analysis. Genetica 137(2), 125–134.
Index

Note: Page numbers followed by “f ” indicate figures, and “t” indicate tables.

A APSIM. See Agricultural Production Systems

sIMulator (APSIM)
Active landscape management, 161 Arachis hypogaea L. See Groundnut crop
Agricultural Production Systems sIMulator Arbuscular mycorrhizae (AM), 145–146
(APSIM), 249–251, 250f Assess site-specific yield potential (ASYP), 89
Agroecological intensification Average yields, palm oil production, 92–93, 92t
Andean region, 163–165
plant breeding, 152–157 B
Alternative glycolytic pathways, PUE, 206
American Palm Oil Council (APOC), 85 Barley seed metabolism, 252
Andean soil fertility Bayesian mapping, 245–246
biological function Best linear unbiased prediction (BLUP), 228, 229
biochar, 150–151 Biochar, 143–144, 150–151
composts, 151 Biological function, soil
import substitution, 151–152 biochar, 150–151
inoculation, 144 composts, 151
microbial and faunal communities, 144–145 import substitution, 151–152
microbial inoculants, 147 inoculation, 144
organic matter management, 144 microbial and faunal communities, 144–145
P-solubilizing activity, 149 microbial inoculants, 147
soil fauna, 152 organic matter management, 144
symbioses, 145–146 P-solubilizing activity, 149
biophysical limitations and risks soil fauna, 152
climate, 127–128, 128f symbioses, 145–146
soil environment, characterization of, Bornean orangutans, 75–76
127–128, 129f Breeding methods
challenges and threats, 131–132 early selection, 225
crop growth factors, 162 genome-wide selection
cropping systems, 126–127 goal of, 235
ecologically based intensification, 132–133 vs. MARS, 235–236, 236f
mass balance marker-assisted backcross, 231–234
erosion, 135–137 marker-assisted selection
fertilizer application, 138–139 influencing factors, 226–228
nitrogen flow diagram, 133–135, 134f vs. phenotypic selection, 228–230
nutrient availability/uptake vs. time, 137, MARS, 231
137f MODPED, 225–226
nutrient dynamics and synchronization SELBLK, 225–226
composting, 140–141 SSD, 225
fertilizer application, 139
nutrient management strategies, 142 C
organic resources, 139–140
physiochemical environment Cangahua soils, 142–143
aggregation, 142–143 Cation exchange capacity (CEC), 143–144
liming, 143–144 Composite interval mapping (CIM), 243–244
pH, 143–144 Composting, anerobic methods, 140–141
plant breeding, 152–157 Computer simulation, plant breeding
socioeconomic and cultural setting, 129–130 breeding methods
spatial and temporal organization, farms, early selection, 225
157–161 genome-wide selection, 234–237

265
266 Index

Computer simulation, plant breeding (cont.) D

marker-assisted backcross, 231–234
marker-assisted selection, 226–230 Deforestation, 75–76
MARS, 231 Deterministic simulation, 222
MODPED, 225–226
SELBLK, 225–226 E
SSD, 225 E-cell system, 252
classification of, 222 Elaeis guineensis cultivation. See Oil palm
climate change, 252–254 cultivation
computation and software issues, 254–255, 255t E(N:K) model, 246–247
crop modeling Environmental stakes, oil palm cultivation
advantages, 248–249 agricultural policies, 80–81
APSIM, 249–251, 250f deforestation and loss of biodiversity, 75–76
uses, 248 GHG emissions and carbon storage, 78
example, 222 peatland degradation, 76–77
gene mapping water pollution, 79–80
association mapping, 243–244
Bayesian mapping, 245–246 F
confidence interval, 243
joint linkage and linkage disequilibrium False discovery rate (FDR), 242
mapping, 245 Fertilizer management, oil palm cultivation, 86
linkage analysis, 243–244 Food production, 30–31
meta-analysis, 240–241 Forest clearing, hydrological impacts of, 111, 112f
missing markers, 240 Foster system, 90
population size and marker density,
238–239, 239f G
significance threshold, 241–242
gene network and genotype-by-environment Gene mapping, plant breeding
interaction association mapping, 243–244
E(N:K) model, 246–247 Bayesian mapping, 245–246
QU-GENE, 247 confidence interval, 243
origin of, 221 joint linkage and linkage disequilibrium
random number generation, 223f mapping, 245
virtual plants and E-cell, 251–252 linkage analysis, 243–244
Critical tissue P concentration, 191 meta-analysis, 240–241
Crop breeding, 155 missing markers, 240
CROPGRO models, groundnut population size and marker density, 238–239,
climatic effects on root growth, 64 239f
pod addition, seed growth, and partitioning significance threshold, 241–242
intensity, 63–64 General circulation models (GCMs), 253
reproductive progression, 61–62 Grain crop genotypes, PUE, 187, 188t
vegetative development, 61 Grain PUE, screening for, 200
vegetative expansion and photosynthesis Greenhouse gas (GHG) emissions, 78
processes, 62–63 Groundnut crop
Crop modeling canopy expansion and growth processes
advantages, 248–249 leaf area and stem elongation, 46–47
APSIM, 249–251, 250f leaf senescence, 47
uses, 248 leaf thickness, 45
Cropping system, INM net assimilation and growth rates, 49–50
fertilizer rates, 16–17 photosynthesis, 48–49
management strategy, 17 stomatal conductance and transpiration,
nitrogen efficiency, 16 47–48
nutrient resource characteristics, 16, 17t CROPGRO models
orchards, 26–29 climatic effects on root growth, 64
paddy rice, 21–22 pod addition, seed growth, and partitioning
vegetable systems, 23–26 intensity, 63–64
wheat and maize, 18–21 reproductive progression, 61–62
Crude palm oil (CPO), 74 vegetative development, 61
Index 267

vegetative expansion and photosynthesis nitrogen efficiency, 16

processes, 62–63 nutrient resource characteristics, 16, 17t
harvest index, 58–59 orchards, 26–29
reproductive development and growth paddy rice, 21–22
appearance of flowers, pegs, and pods, vegetable systems, 23–26
50–51 wheat and maize, 18–21
number of pegs, pods, and seeds, 54–55 fertilizer rate per hectare, 3–4, 4f
pod and seed growth rates and their size, grain production and chemical fertilizer inputs,
55–56 2, 3f
pollen production and viability and fruit-set, large-scale dissemination
53–54 environmental pollution, 30–31
rate of flower production, 51–52 factors, 29–30
root growth, 59–60 partial factor productivity, 30, 30f
root-to-shoot ratio, 60–61 principles of
shelling percentage, 59 crop yields, 15–16
total dry matter, pod, and seed yield, 56–58 mobilization and acquisition, 6–7
vegetative development nitrogen fertilization, 7–8
germination and emergence processes, N loss reduction, Chinese cropping systems,
43–44 12–15
leaf appearance and leaf number, 44–45 NO-3, 7–8
nutrient input optimization, 9–11
H rhizosphere/root-zone nutrient
management, 8, 9f
Harvest index (HI), 189–191
soil nutrient supply matching, 11–12
Hydrological processes, oil palm
evapotranspiration, 95–96
L
forest clearing, hydrological impacts of, 111,
112f Leaf analysis, 89
hydrological cycle, 93, 94f Least square method, 226–227
interception, 94–95 LOD drop-off method, 243
leaching and goundwater facility Lupinus mutabilis, 145–146, 155–156
annual rainfall, annual runoff and nutrient
losses, 97–98, 99t, 100t M
plot-scale study, 101–103
soil texture, 98–101, 102t Mapping as you go (MAYG) method, 227–228
precipitation, 94 Marker-assisted backcross (MABC), 231–234
qualitative description, 111, 113t Marker-assisted recurrent selection (MARS), 231
soil infiltration Marker-assisted selection (MAS)
deforestation, 96–97 breeding methods
hydraulic conductivities, 96 influencing factors, 226–228
soil types and locations, 97, 97t vs. phenotypic selection, 228–230
stream flow, 108–109 PUE, 209–210
stream water quality, 109–111, 110t MARS. See Marker-assisted recurrent selection
surface runoff and soil erosion, 103–108, 104t, (MARS)
106t, 107t Microbial symbioses, 154–155
Mitochondrial electron transport pathways, PUE,
I 206
Modified pedigree/bulk selection method
Import substitution, 151–152 (MODPED), 225–226
Inclusive composite interval mapping (ICIM), Multiple interval mapping (MIM), 244
244 Mycorrhizal inoculation, impacts of, 147–149,
Industrial vs. smallholder plantations, oil palm, 82 148t
In silico mapping, 244–245
Integrated nutrient management (INM) N
chemical fertilizer consumption, 3–4, 4f
conceptual model, 6, 6f Nested association mapping (NAM), 245
in cropping system Nucleus estate scheme (NES), 82
fertilizer rates, 16–17 Nutrient dynamics and synchronization, soil
management strategy, 17 fertility
268 Index

Nutrient dynamics and synchronization, soil P

fertility (cont.)
composting, 140–141 Paddy rice
fertilizer application, 139 basal and top-dressing, 22
nutrient management strategies, 142 fertilizer N application rate, 21
organic resources, 139–140 INM performance, China, 22, 23f
Nutrient use efficiency (NUE), 15 soil N supply capacity, 22
Palm kernel oil, 74
O Palm oil mill effluent (POME), 79–80
P-deficiency stress response mechanisms,
Off-the-shelf software, 254 206–207
Oil palm cultivation Peatland degradation
climate and soil conditions, 81–82 ecological functions, 77
environmental stakes oil palm conversion, 77
agricultural policies, 80–81 peatland formation, 76
deforestation and loss of biodiversity, PER. See Phosphorus-efficiency ratio (PER)
75–76 Phalaris sp, 136–137
GHG emissions and carbon storage, 78 Phosphorus acquisition efficiency (PAE)
peatland degradation, 76–77 crop cultivars, 206–207
water pollution, 79–80 vs. PUE, 208
expansion of Phosphorus-efficiency ratio (PER), 189,
future expansion, 74–75 195–196
in Indonesia, 74, 75f Phosphorus utilization efficiency (PUE)
palm oil utilization, 74 agronomic implications, 189–191
fertilizer management classification, 189
chemical fertilizer, 90 definitions, 189, 190t, 195–196
organic fertilizer, 90–92 grain crop genotypes, 187, 188t
hydrological processes marker-assisted selection, 209–210
evapotranspiration, 95–96 mechanism and physiology
forest clearing, hydrological impacts of, alternative glycolytic pathways and
111, 112f mitochondrial electron transport
hydrological cycle, 93, 94f pathways, 206
interception, 94–95 P-deficiency stress response mechanisms,
leaching and goundwater facility, 97–103 206–207
precipitation, 94 remobilization and scavenging, of P,
qualitative description, 111, 113t 205–206
soil infiltration, 96–97, 97t vs. PAE, 208
stream flow, 108–109 physiological implications
stream water quality, 109–111, 110t reproductive stage, 195
surface runoff and soil erosion, 103–108, vegetative stage, 192–194, 193t
104t, 106t, 107t screening
land clearing and site preparation, grain PUE, 200
82–83, 84f vegetative PUE, 197–200
nutrient-demand assessment Physiochemical environment, soil
fertilizer management, 86 aggregation, 142–143
soil nutrient supply, 86–89, 88t liming, 143–144
positive and negative aspects, 73 pH, 143–144
production systems, 82 Plant-growth-promoting microbes (PGPM), 146,
synthesis, 92–93 149–150
water and soil management, 83–85 Plasma, 82
On-field crop residue retention, 137–138 P-LIM-GROW model, 202–204, 203f
Orchards Potassium fluxes, oil palm plantation,
basal fertilization, 28 86–89, 89f
controlling N and P fertilization, 27–28 Pseudorandom number generator, 254–255
deficiency, 27 P-stress levels
management, 28 outcomes, PUE, 202–204
misuse of fertilizers, 27 P-deficient crops, 201–202
nutrient inputs, 27 PUE. See Phosphorus utilization efficiency (PUE)
Index 269

R Stochastic simulation, 222

Sumatran orangutans, 75–76
Remobilization and scavenging, of P, 205–206
Root PUE
T
definition, 192
variation, vegetative crop growth, 193t Trichoderma, 145–146
Roundtable on Sustainable Palm Oil (RSPO), 80 Trifolium spp, 145–146
S V
Screening Vegetable systems
grain PUE, 200 characteristics, 23–24
P-stress levels, 201–204 chili pepper and cucumber yield, 24–25, 24f
vegetative PUE fertigation techniques, 26
P uptake nullification, genotypes, 198–200 in-season root-zone N management, 25, 25f
shoot PUE vs. shoot P content, 197, 198f, P management strategy, 26
199f Vegetative PUE, screening
Selected bulk selection method (SELBLK), 225–226 P uptake nullification, genotypes, 198–200
Shoot PUE shoot PUE vs. shoot P content, 197,
control of, 192 198f, 199f
correlation coefficients, 196t Vermicomposting, 140–141
vs. shoot P content, 198f, 199f Virtual plants, 252
variation, vegetative crop growth, 193t
Single seed descent (SSD), 225 W
Soil aggregation, 142–143
Soil degradation, 131 Wacho rozado, 136–137
Soil infiltration Water management, oil palm cultivation, 83–85
deforestation, 96–97 Wheat and maize system
hydraulic conductivities, 96 factors, 20–21
soil types and locations, 97, 97t grain yields, 18
Spatial and temporal organization, farms, 157–161 INM performance, North China Plain, 19–20,
SSD. See Single seed descent (SSD) 20f
Standing stock biomass, in oil palm plantation, model of, in-season root-zone, 18–19, 19f
86, 87t soil N tests, 18–19