Chapter 1

Introduction

Abstract In simple terms, hydrology is the study of the waters of the Earth,
including their occurrence, distribution, and movement. The constant circulation of
water and its change in physical state is called the hydrologic cycle. The study of
water started at least a few thousands years ago, but the modern scientific approach
to the hydrologic cycle started in the seventeenth century. Since then, hydrology
has witnessed a tremendous growth, especially over the last century, with signifi-
cant advances in computational power and hydrologic data measurements. This
chapter presents a general and introductory account of hydrology. First, the concept
of the hydrologic cycle is described. Next, a brief history of the scientific devel-
opment of hydrology is presented. Then, the concept of hydrologic system is
explained, followed by a description of the hydrologic system model and model
classification. Finally, the role of hydrologic data and time series modeling as well
as the physical basis of time series modeling are highlighted.

1.1 Definition of Hydrology

The name ‘hydrology’ was derived from the Greek words ‘hydro’ (water) and
‘logos’ (study), and roughly translates into ‘study of water.’ Different textbooks
may offer different definitions, but all of them generally reflect the following
working definition:
Hydrology is the science that treats the waters of the Earth, their occurrence, circulation and
distribution, their chemical and physical properties, and their interactions with their envi-
ronments, including their relations to living things.

Within hydrology, various sub-fields exist. In keeping with the essential ingre-
dients of the above definition, these sub-fields may depend on the region (e.g. over
the land surface, below the land surface, mountains, urban areas) or property (e.g.
physical, chemical, isotope) or interactions (e.g. atmosphere, environment,
ecosystem) or other aspects (e.g. tools used for studies) of water. There may also be
significant overlaps between two or more sub-fields, and even inter-change of
terminologies depending on the emphasis for water in studies of the

© Springer Science+Business Media Dordrecht 2017 3

B. Sivakumar, Chaos in Hydrology, DOI 10.1007/978-90-481-2552-4_1
4 1 Introduction

Earth-ocean-atmospheric system. Some of the popular sub-fields within hydrology

are:
• Surface hydrology—study of hydrologic processes that operate at or near the
Earth’s surface
• Sub-surface hydrology (or Groundwater hydrology or Hydrogeology)—study of
the presence and movement of water below the Earth’s surface
• Vadose zone hydrology—study of the movement of water between the top of
the Earth’s surface and the groundwater table
• Hydrometeorology—study of the transfer of water and energy between land and
water body surfaces and the lower atmosphere
• Hydroclimatology—study of the interactions between climate processes and
hydrologic processes
• Paleohydrology—study of the movement of water and sediment as they existed
during previous periods of the Earth’s history
• Snow hydrology—study of the formation, movement, and effects of snow
• Urban hydrology—study of the hydrologic processes in urban areas
• Physical hydrology—study of the physical mechanisms of hydrologic processes
• Chemical hydrology—study of the chemical characteristics of water
• Isotope hydrology—study of the isotopic signatures of water
• Ecohydrology (or Hydroecology)—study of the interactions between hydrologic
processes and organisms
• Hydroinformatics—the adaptation of information technology to hydrology and
water resources applications.

1.2 Hydrologic Cycle

The constant movement of water and its change in physical state on the Earth (in
ocean, land, and atmosphere) is called the hydrologic cycle or, quite simply, water
cycle. The hydrologic cycle is the central focus of hydrology. A schematic repre-
sentation of the hydrologic cycle is shown in Fig. 1.1. A description of the
hydrologic cycle can begin at any point and return to that same point, with a
number of processes continuously occurring during the cycle; however, oceans are
usually considered as the origin. In addition, depending upon the scope or focus of
the study, certain processes (or components) of the hydrologic cycle may assume
far more importance over the others and, hence, such may be described in far more
detail. In what follows, the hydrologic cycle is described with oceans as the origin
and processes on and above/below the land surface assuming more importance. For
further details, including other descriptions of the hydrologic cycle, the reader is
referred to Freeze and Cherry (1979), Driscoll (1986), Chahine (1992), Maidment
(1993), and Horden (1998), among others.
Water in the ocean evaporates and becomes atmospheric water vapor (i.e.
moisture). Some of this water vapor is transported and lifted in the atmosphere until
1.2 Hydrologic Cycle 5

Fig. 1.1 Schematic representation of hydrologic cycle (source US Geological Survey, http://
water.usgs.gov/edu/watercycle.html; accessed May 5, 2015)

it condenses and falls as precipitation, which sometimes evaporates or gets inter-

cepted by vegetation before it can reach the land surface. Of the water that reaches
the land surface by precipitation, some may evaporate where it falls, some may
infiltrate the soil, and some may run off overland to evaporate or infiltrate elsewhere
or to enter streams. The water that infiltrates the ground may evaporate, be absorbed
by plant roots, and then transpired by the plants, or percolate downward to
groundwater reservoirs (also called aquifers). Water that enters groundwater
reservoirs may either move laterally until it is close enough to the surface to be
subject to evaporation or transpiration, reach the land surface and form springs,
seeps or lakes, or flow directly into streams or into the ocean. Stream water can
accumulate in lakes and surface reservoirs, evaporate or be transpired by riparian
vegetation, seep downward into groundwater reservoirs or flow back into the ocean,
where the cycle begins again.
Although the concept of the hydrologic cycle is simple, the phenomenon is
enormously complex and intricate. It is not just one large cycle but rather composed
of many inter-related cycles of continental, regional, and local extent. Each phase of
the hydrologic cycle also provides opportunities for temporary accumulation and
storage of water, such as snow and ice on the land surface, moisture in the soil and
groundwater reservoirs, water in ponds, lakes, and surface reservoirs, and vapor in
the atmosphere. Although the total volume of water in the global hydrologic cycle
6 1 Introduction

remains essentially constant, the distribution of this water is continually changing

on continents, in regions, and within local drainage basins.
The hydrologic cycle is also the basis for the concept of water as a renewable
resource. Hydrology recognizes that the natural hydrologic cycle can be altered by
human and natural activities, as the following examples indicate. Geologic forces
that raise mountains can increase orographic precipitation on one side of the
mountains and decrease precipitation on the other side with all of the attendant
changes in streamflow, flooding, etc. The development and use of water modify the
natural circulatory pattern of the hydrologic cycle, such as the use of surface water
for irrigation possibly resulting in downward seepage from reservoirs, canals, dit-
ches, and irrigated fields, adding to the groundwater. Diversions of streamflows
impact downstream flows which, if transferred to other watersheds, impact the
streamflows and groundwater systems in the other watersheds. Pumping from wells
may reduce the flow of water from springs or seeps, increase the downward
movement of water from the land surface and streams, reduce the amount of natural
groundwater discharge by evaporation and transpiration, induce the inflow of poorer
quality water to the groundwater reservoir, or have a combination of all these effects.
The natural circulation of the hydrologic cycle may also be changed by actions
not related to direct water use. Among these actions are weather modification
activities (e.g. cloud seeding), drainage of swamps and lakes, water-proofing of the
land surface by buildings and pavements, and major changes in vegetative cover
(e.g. removal of forests).

1.3 Scientific Development of Hydrology

Humans have been concerned with managing water at least since the first civi-
lizations developed along river banks over 8000 years ago. Hydraulic engineers
built functioning canals, levees, dams, water conduits, and wells along the Indus in
Pakistan, the Tigris and Euphrates in Mesopotamia, the Hwang Ho in China, and
the Nile in Egypt as early as 5000–6000 B.C. Flow monitoring was started by the
Egyptians around 3800 B.C., and the first rainfall measurements were made by
Kautilya of India around 2400 B.C. (Eagleson et al. 1991).
The concept of global hydrologic cycle started perhaps around 3000 B.C. (Nace
1974), when King Solomon wrote in Ecclesiastes 1:7 that
All the rivers run into the sea; yet the sea is not full; unto the place from whence the rivers
come, thither they return again.

Early Greek philosophers, such as Thales, Anaxagoras, Herodotus, Hippocrates,

Plato, and Aristotle also embraced the basic idea of the hydrologic cycle. However,
while some of them had reasonable understandings of certain hydrologic processes,
they postulated various fanciful underground mechanisms by which water returned
from sea to land and entered rivers. The Romans had considerable practical
knowledge of hydrology (and especially hydraulics) and constructed extensive
1.3 Scientific Development of Hydrology 7

aqueduct systems, but their scientific ideas were largely based on those of the
Greeks.
Independent thinking occurred in ancient Asian civilizations (UNESCO 1974).
The Chinese recorded observations of rain, sleet, snow, and wind on Anyang oracle
bones as early as 1200 B.C. They probably used raingages around 1000 B.C., and
established systematic raingaging around 200 B.C. In India, the first quantitative
measurements of rainfall date back to the fourth century B.C. The concept of a
dynamic hydrologic cycle may have arisen in China by 900 B.C, in India by 400 B.
C., and in Persia by the tenth century, but these ideas had little impact on Western
thought (Chow et al. 1988).
In the meantime, the theories of the Greek philosophers continued to dominate
Western thought until much of the Renaissance, which spanned roughly from the
12th to the 17th century A.D. Then, Leonardo da Vinci (about 1500 A.D.) in Italy
and Bernard Palissy (about 1550 A.D.) in France asserted, based on field obser-
vations, that the water in rivers comes from precipitation (Adams 1938; Biswas
1970). With this initiation, the modern scientific approach to the hydrologic cycle
was taken up in the seventeenth century by the Frenchmen Pierre Perrault and Edmé
Marriotte, who published, in the 1670s and 1680s, measurements and calculations
that quantitatively verified the rainfall origin of streamflow. Shortly after that
(around 1700), Edmund Halley, an English scientist, extended the quantification of
the hydrologic cycle through estimation of the amounts of water involved in the
ocean-atmosphere-rivers-ocean cycle of the Mediterranean Sea and surrounding
islands.
The eighteenth century witnessed significant advances in the applications of
mathematics to fluid mechanics and hydraulics, notably by Henri Pitot, Daniel
Bernoulli, Leonhard Euler, Antoine de Chézy, and other Europeans. In fact, use of
the term ‘hydrology’ in approximately its current meaning also began during this
time (around 1750). By about 1800, the nature of evaporation and the present
concepts of the global hydrologic cycle were firmly established by the English
physicist and chemist John Dalton (Dalton 1802), and Charles Lyell, James Hutton,
and John Playfair published scientific works on the fluvial erosion of valleys.
Routine network measurements of precipitation were begun before 1800 in Europe
and the United States, and established there and in India by 1820s.
Until mid-nineteenth century, one of the barriers to understanding the hydrologic
cycle was the ignorance of the groundwater flow process. This changed in 1856,
when the French engineer Henry Darcy established the basic phenomenological law
of flow through porous media (Darcy 1856). The nineteenth century also saw
further advances in fluid mechanics, hydraulics, and sediment transport by
Jean-Louis Poiseuille, Jules DuPuit, Paul DuBoys, George Stokes, Robert
Manning, William Reynolds, and others, whose names have become associated
with particular laws or principles. Details of these efforts can be seen in Manning
(1891), among others.
Treaties on various aspects of hydrology, beginning with Nathaniel Beardmore’s
Manual of Hydrology in 1851 (Beardmore 1851), appeared with increasing fre-
quency in the second half of the nineteenth century. Many of these works examined
8 1 Introduction

relations between rainfall amounts and streamflow rates, because of the need to
estimate flood flows for the design of bridges and other hydraulic structures. This
was also the beginning of a close association between hydrology and civil engi-
neering; as a matter of fact, the first English-language texts in hydrology by Daniel
Mead in 1904 and Adolf Meyer in 1919 were written for civil engineers (Eagleson
et al. 1991).
The first half of the twentieth century saw great progress in many aspects of
hydrology. With the formation of the Section of Scientific Hydrology in the
International Union of Geodesy and Geophysics (in 1922) and the Hydrology
Section of the American Geophysical Union (in 1930), hydrology received formal
scientific recognition for the first time. During this time, there were many notable
contributions to advances in specific areas: Allen Hazen, Emil Gumbel, Harold
Hurst, and Walter Langbein in the application of statistics to hydrologic data; Oscar
Meinzer, Charles Theis, Charles Slichter, and Marion King Hubbert in the devel-
opment of the theoretical and practical aspects of groundwater hydraulics, and
especially Lorenzo Richards in the development of governing equation for unsat-
urated flow; Ludwig Prandtl, Theodor von Kármán, Hunter Rouse, Ven Te Chow,
Grove Karl Gilbert, and Hans Einstein in stream hydraulics and sediment transport;
Robert Horton and Luna Leopold in understanding runoff processes and quantita-
tive geomorphology; Charles Warren Thornthwaite and Howard Penman in
understanding climatic aspects of hydrology and modeling evapotranspiration; and
Abel Wolman and Robert Garrels in the understanding and modeling of water
quality. Details of these developments can be seen in Richards (1931), Horton
(1933, 1945), Gumbel (1941), and Hurst (1951), among others.
The 1960s witnessed the beginning of stochastic concepts in hydrology, notably
applications of linear stochastic methods to hydrologic data (Thomas and Fiering
1962; Yevjevich 1963; Fiering 1967). Assisted by the discovery of self-similarity
concept during this decade (Mandelbrot 1967), the ideas of scale in hydrology also
gained more recognition (Mandelbrot and Wallis 1968, 1969). With advances in
stochastic time series methods and fractal concepts in the 1970s (Box and Jenkins
1970; Mandelbrot 1975), the linear stochastic and scaling concepts in hydrology
started to proliferate and are now prevalent in hydrology (e.g. Yevjevich 1972;
Mandelbrot 1977; Gupta et al. 1986; MacNeill and Umphrey 1987; Gelhar 1993;
Salas et al. 1995; Kalma and Sivapalan 1996; Rodriguez-Iturbe and Rinaldo 1997;
Govindaraju 2002).
The revolutionary advances since the 1970s in computer and measurement
technologies (e.g. supercomputers, remote sensors, and geographic information
systems) have facilitated the emergence of various nonlinear concepts and the
development of a host of nonlinear time series methods as well as others. In addition
to the nonlinear stochastic ones, these methods include: data-based mechanistic
models, artificial neural networks, wavelets, entropy theory, support vector machi-
nes, genetic programming, fuzzy logic, and nonlinear dynamics and chaos.
Applications of these concepts in hydrology roughly began in the late 1980s and
early 1990s and have tremendously amplified since then (e.g. Foufoula-Georgiou
and Kumar 1994; Young and Beven 1994; Bardossy and Duckstein 1995; Babovic
1.3 Scientific Development of Hydrology 9

1996; Singh 1997; Govindaraju and Rao 2000; Sivakumar 2000; Dibike et al. 2001;
see also Sivakumar and Berndtsson 2010), largely under the umbrella of
‘Hydroinformatics.’ These advances have also led to the development of numerous
lumped, semi-distributed, and distributed hydrologic models, such as the TANK
model, SWMM (Storm Water Management Model), TOPMODEL (Topographic
model), HEC-HMS (Hydrologic Engineering Center Hydrologic Modeling System)
and HEC-RAS (River Assessment System), SHE (Système Hydrologique Européen)
and MIKE-SHE, SLURP (Semi-distributed Land Use Runoff Process), and SWAT
(Soil and Water Assessment Tool) models. Extensive details about these models are
already available in the literature (e.g. Metcalf and Eddy 1971; HEC 1995, 1998;
Singh 1995; Abbott and Refsgaard 1996; Beven 1997; Neitsch et al. 2005). Since the
complex semi-distributed and distributed models incorporate more and more pro-
cesses and, thus, require calibration of more and more parameters, the problems of
parameter estimation and the associated uncertainties have also become significant,
and have been important areas of hydrologic studies since the 1980s (Sorooshian and
Gupta 1983; Beven 1993; Duan et al. 2003).
Advances in the above areas, and still in many others, are continuing at a much
faster rate than at any other time in the history of hydrology. Despite these, how-
ever, our understanding of hydrologic systems and the associated processes and
problems is still far from complete. In fact, it is fair to say that the advances we have
made thus far have brought in more questions than answers (e.g. Klemeš 1986;
Sivakumar 2008c). There are some major challenges in several areas, and so are
great opportunities. These include: simplification in our modeling practice, uncer-
tainy estimation in hydrologic models, formulation of a hydrologic classification
framework, scale issues, predictions in ungaged basins, assessment of the impacts
of global climate change on our future water resources, connections between
hydrologic data and system physics, translations and interpretations of our math-
ematical models and methods for better understanding of hydrologic systems and
processes (e.g. Beven 2002, 2006; Sivapalan et al. 2003; McDonnell and Woods
2004; Kirchner 2006; Gupta et al. 2007; Sivakumar 2008a, b, c; IPCC 2014). There
is no doubt that studying these issues will be an important part of hydrologic theory
and practice in the coming decades and centuries.

1.4 Concept of Hydrologic System

Hydrologic phenomena are enormously complex, and are not fully understood. In
the absence of perfect knowledge, a simplified way to represent them may be
through the concept of system. There are many different definitions of a system, but
perhaps the simplest may be: ‘a system is a set of connected parts that form a
whole.’ Chow (1964) defined a system as an aggregate or assemblage of parts,
being either objects or concepts, united by some form of regular interaction or
inter-dependence. Dooge (1967a), however, defined a system as: “any structure,
device, scheme, or procedure, real or abstract, that inter-relates in a given time
10 1 Introduction

reference, an input, cause, or stimulus, of matter, energy, or information and an

output, effect, or response of information, energy, or matter.” This definition by
Dooge is much more comprehensive and instructive and it brings out, among
others, the following important characteristics of the system: (1) a system can
consist of more than one component; (2) these components are separate, and they
may be inter-dependent; (3) these components are put together following some sort
of scheme, i.e. a system is an ordered arrangement; (4) a system inter-relates input
and output, cause and effect, or stimulus and response; (5) a system does not require
that input and output be alike or have the same nature; and (6) a system can be
composed of a number of sub-systems, each of which can have a distinct
input-output linkage.
With this system concept, the entire hydrologic cycle may be regarded as a
hydrologic system, whose components might include precipitation, interception,
evaporation, transpiration, infiltration, detention storage or retention storage, sur-
face runoff, interflow, and groundwater flow, and perhaps other phases of the
hydrologic cycle. Each component may be treated as a sub-system of the overall
cycle, if it satisfies the characteristics of a system set out in its definition. Thus, the
various components of the hydrologic system can be regarded as hydrologic
sub-systems. To analyze the total system, the simpler sub-systems can be treated
separately and the results combined according to the interactions between the
sub-systems (especially with the assumption of linearity). Whether a particular
component is to be treated as a system or sub-system depends on the objective of
the inquiry (Singh 1988; see also Sivakumar and Singh 2012).
Considering in our aim a fair balance between reduction in system complexity
and incorporation of necessary system details, the global hydrologic cycle system
may roughly be divided into three sub-systems (Chow et al. 1988).
• The atmospheric water system containing the processes of precipitation, inter-
ception, evaporation, and transpiration. This sub-system is studied under the
sub-field of hydrometeorology;
• The surface water system containing the processes of overland flow, surface
runoff, sub-surface and groundwater flow, and runoff to streams and the ocean.
This sub-system is studied under the sub-field of surface hydrology; and
• The sub-surface water system containing the processes of infiltration, ground-
water recharge, sub-surface flow and groundwater flow. This sub-system is
studied under the sub-field of sub-surface hydrology. Sub-surface flow takes
place in the soil near the land surface, while groundwater flow occurs deeper in
the soil or rock strata.
In fact, for most practical problems, only a few processes of the hydrologic cycle
are considered at a time, and then only considering a small portion of the Earth’s
surface. For such treatment, a more restricted system definition than the global
hydrologic system may be appropriate, with the concept of the control volume, as is
1.4 Concept of Hydrologic System 11

the case in the field of fluid mechanics. In this context, a hydrologic system can be
defined as a structure or volume in space, surrounded by a boundary, that accepts
water and other inputs, operates on them internally, and produces them as outputs.
The structure (for surface or sub-surface flow) or volume in space (for atmospheric
moisture flow) is the totality of the flow paths through which the water may pass as
throughput from the point it enters the system to the point it leaves. The boundary is
a continuous surface defined in three dimensions enclosing the volume or structure.
A working medium enters the system as input, interacts with the structure and other
media, and leaves as output. Physical, chemical, and biological processes operate
on the working media within the system.
The procedure for developing working equations and models of hydrologic
phenomena is similar to that in fluid mechanics, where mass, momentum, and
energy principles serve as bases. In hydrology, however, there is generally a greater
degree of approximation in applying physical laws because the systems are larger
and more complex, and may involve several working media, whose properties may
change tremendously in time and/or space. It must also be noted that many
hydrologic systems are normally treated as random because their major input is
precipitation, which is a highly variable and often unpredictable phenomenon,
although there are non-random ways of treating precipitation behavior (e.g.
Rodriguez-Iturbe et al. 1989; Sivakumar et al. 2001; see also Chap. 9 for further
details). Consequently, statistical analysis plays a large role in hydrologic analysis.
Because of these complications, and many others, it is not possible to describe some
hydrologic processes with exact physical laws. The system concept helps in the
construction of a model that relates inputs and outputs, rather than the extremely
difficult task of exact representation of the system details, and thus has significant
practical advantage.

1.5 Hydrologic System Model

As mentioned just now, the goal of the system concept is to establish an

input-output relationship that can be used for reconstructing past events or pre-
diction of future events. In this systems approach, we are concerned with the system
operation, not the nature of the system itself (its components, their connection with
one another, and so on) or the physical laws governing its operation. A system
model is an approximation of the actual system (i.e. prototype); its inputs and
outputs are measurable variables and its structure is an equation (or a set of
equations) linking the inputs and outputs. Central to the model structure is the
concept of a system transformation.
Let us assume that precipitation (P) over a river basin produces some flow (Q) at
the outlet of the basin. It can then be said that the river basin system performs a
transformation of precipitation (P) into flow (Q), which can be represented by:
12 1 Introduction

Q ¼ f ðPÞ ð1:1Þ

where f is the transformation function, or simply transfer function. Generally

speaking, f is a transfer function between the input (cause) and the output (effect).
The cause and effect can be either internal to the system or external to the system or
a combination, depending on the ‘boundaries’ of the system. In Eq. (1.1), Q is the
‘dependent’ variable, and P is the ‘independent’ variable.
Equation (1.1) is the simplest form of the equation for the flow process in a river
basin, with precipitation serving as the only input variable. Since a host of other
variables also influences the flow process (depending upon the river basin char-
acteristics), this relationship is oftentimes a gross approximation at best. If, for
example, infiltration (I) and evaporation (E) also act as influencing variables, then
Eq. (1.1) must be changed to:

Q ¼ f ðP; I; EÞ ð1:2Þ

Assuming that data are available (and in good quality), Eq. (1.2) is certainly a more
accurate representation of the flow process than Eq. (1.1). At the same time,
however, Eq. (1.2) is also more complex than Eq. (1.1) and, thus, is more difficult
to solve.
It must be clear now that inclusion of any additional influencing variable(s) will
result in an even more complex equation, which will be even more difficult to solve.
It is important, therefore, to be mindful of the complexity of the equation (i.e.
model), not only for its solution but also for its data requirements. The problems
associated with the development of more and more complex hydrologic models
have been extensively discussed in the literature (e.g. Sorooshian and Gupta 1983;
Konikow and Bredehoeft 1992; Beven 1993, 2002; Young et al. 1996; Duan et al.
2003), especially under the topics of parameter estimation and uncertainty.
It must also be noted that the variables in Eq. (1.1) represent the total values (i.e.
total amount of precipitation and total volume of flow) over the basin over a period
of time. However, since precipitation changes (and, hence, the flow) with time (t),
Eq. (1.1) must be modified to incorporate this as:

Qt ¼ f ðPt Þ ð1:3Þ

where Pt is the rainfall intensity and Qt is the flow rate. This ‘time’ or ‘dynamic’
factor (e.g. precipitation intensity) brings in further complexity to the
precipitation-flow relationship, and so does the distribution of the variables in
‘space.’ This then requires a spatio-temporal perspective to the overall
precipitation-flow relationship. Consideration of both space and time factors in all
the variables influencing the flow process [such as an extension of Eq. (1.2)] will
result in a highly complex spatio-temporal relationship. This is why modeling and
prediction of hydrologic processes is often a tremendously difficult task.
The transfer function f in the above equations may be linear or nonlinear,
depending on the properties of the input variables and on the characteristics of the
1.5 Hydrologic System Model 13

river basin. In simple terms, ‘linear’ means output is proportional to the input (e.g.
double the amount of rainfall producing double the amount of flow), and ‘nonlin-
ear’ means output is not proportional to the input. Looking at the general
non-proportionality between hydrologic inputs and outputs, it is fair to say that
most, if not all, hydrologic processes are nonlinear in nature. The nonlinear nature
of hydrologic processes had indeed been recognized as early as in the 1960s (e.g.
Minshall 1960; Jacoby 1966; Amorocho 1967; Dooge 1967b; Amorocho and
Brandstetter 1971). However, much of early hydrologic analysis (during the 1960s–
1980s), especially based on time series methods (see Sect. 1.7), assumed the
transfer functions as linear, perhaps due to the lack of data and computational
power. This situation, however, changed dramatically in the 1980s, with the
development of nonlinear time series methods facilitated by the availability of more
data and computational power. At the current time, both linear and nonlinear
transfer functions are assumed in hydrology, depending upon whether linear or
nonlinear time series analysis method is employed. Further details about linearity
and nonlinearity (and several other characteristics of hydrologic systems and pro-
cesses) will be discussed in Chap. 2, and some of the popular linear and nonlinear
time series methods applied in hydrology will be discussed in Chaps. 3 and 4,
respectively.

1.6 Hydrologic Model Classification

As of now, there is no universally accepted hydrologic model classification. As

different people perceive, conceptualize, and understand hydrologic systems in
different ways, hydrologic models may be classified in different ways too; see
Snyder and Stall (1965), Dawdy (1969), Dawdy and Kalinin (1969), DeCoursey
(1971), Snyder (1971), Woolhiser (1971, 1973, 1975), Miller and Woolhiser (1975)
for some early studies. Consequently, any classification or grouping of hydrologic
models can be only rather arbitrary. For purposes of simplicity and convenience,
however, hydrologic models may generally be grouped under two broad categories:
physical models and abstract models.

1.6.1 Physical Models

A physical model is a representation of the system in a reasonably physically

realistic manner. Physical models include scale models and analog models.
A scale model is a model that represents the system in a different size (enlarged
or reduced) than the prototype. In hydrology, scale models are normally on a
reduced size. A hydraulic model of a dam spillway is one example of a scale model,
and an open channel hydrologic laboratory model of a river is another.
14 1 Introduction

An analog model is a model that uses another physical system having properties
similar to those of the system under study but is much easier to work with. It does
not physically resemble the actual system but depends on the correspondence
between the symbolic models describing the prototype and the analog system. The
Hele-Shaw model (Hele-Shaw 1898) is one example of an analog model, as it uses
the movement of a viscous fluid between two closely-spaced parallel plates to
model seepage in an aquifer or embankment. An electrical analog model for
watershed response is another example.

1.6.2 Abstract Models

An abstract model is a representation of the actual system in mathematical form,

and thus it is also called as a mathematical model. In this model, the system
operation is described by a set of equations linking the input and the output vari-
ables. Abstract or mathematical models may be divided into three groups (Dooge
1977): (1) empirical models; (2) theoretical models; and (3) conceptual models.
An empirical model is merely a representation of the facts based on the available
data; if the conditions change, it has no predictive capability. Therefore, all
empirical models have some chance of being fortuitous and, in principle, should not
be used outside the range of data from which they were derived. One example of
empirical models is the rational method (Kuichling 1889)

Q ¼ CiA ð1:4Þ

where Q is flow rate (cubic feet per second), i is rainfall intensity (inches per hour),
A is area (acre), and C is a constant (runoff coefficient) that can range from 0 to 1.
Other examples include the unit hydrograph models based on harmonic analysis
(O’Donnell 1960), the least squares method (Snyder 1955), and the Laguerre
polynomials (Dooge 1965).
A theoretical model is presumably a consequence of the most important laws
governing the phenomena. It has a logical structure similar to the real-world system
and may be helpful under changed circumstances. Examples of theoretical models
may include infiltration models based on two-phase flow theory of porous media
(Morel-Seytoux 1978), evaporation models based on theories of turbulence and
diffusion (Brutsaert and Mawdsley 1976), and groundwater models based on fun-
damental transport equations (Freeze 1971).
A conceptual model is an intermediate between an empirical and a theoretical
model, although it can be used broadly to embrace both of these types of models.
Generally, conceptual models consider physical laws but in highly simplified form.
Examples of conceptual models may include rainfall-runoff models based on the
spatially lumped form of the continuity equation and the storage-discharge rela-
tionship (Nash 1958; Dooge 1959), and models derived from linear diffusion anal-
ogy and linearized versions of St. Venant equations (Harley 1967; O’Meara 1968).
1.6 Hydrologic Model Classification 15

1.6.3 Remarks

Further sub-grouping of hydrologic models may be made on the basis of their

nature or complexity or other aspects. These include: linear and nonlinear,
time-invariant and time-variant, lumped and distributed, and deterministic and
stochastic. Some of these will be discussed in Chaps. 2–4, as part of the charac-
teristics of hydrologic systems and processes and popular models that have been in
existence. An excellent essay on empirical and physical models in hydrology was
presented by Klemeš (1982), which provides in detail a philosophical perspective
on: (1) hydrologic modeling in general; (2) merits and demerits of empirical
modeling; (3) why hydrologic models work; and (4) drawbacks, dangers, and
potential benefits of causal modeling. Singh (1988) also presented an extensive
account of the various types of models in existence in hydrology.
It must be emphasized that models are only approximate representations of
actual systems. Their development depends not only on our limited understanding
of the actual systems but also on limited observations and computational/structural
powers. Therefore, it is fair to say that all models are wrong, or will be proven to be
wrong. The only perfect model of a physical system is the system itself. There is no
such thing as a perfect model because an abstract quantity cannot perfectly repre-
sent a physical entity (Singh 1988).

1.7 Hydrologic Data and Time Series Modeling

Whether physical or abstract hydrologic models, observations or data on relevant

hydrologic variables (e.g. rainfall, evaporation, streamflow) as inputs/outputs play a
key role in model formulation and model validation. A variable can be observed at a
particular location over time or at a particular time at different locations or at
different locations over time. A set of observations of a hydrologic variable made at
a particular location over time is called a hydrologic time series (Fig. 1.2). For
instance, if precipitation P is the variable of interest at a given location and is
observed over time N, then the time series of precipitation may be denoted as Pi,
i  N. Observations over time at a particular location are helpful for studying the
temporal dynamics; those made at a particular time at different locations are helpful
for studying the spatial patterns; and observations made at different locations over
time are helpful for studying the spatio-temporal dynamics.
The timescale of a hydrologic time series may be either discrete or continuous.
A discrete timescale would result from observations at specific times with the times
of the observations separated by i or from observations that are some function of the
values that actually occurred during i. Most hydrologic time series fall in this latter
category, such as the one shown in Fig. 1.2. Examples would be the annual peak
discharge (i = 1 year), monthly precipitation (i = 1 month), and average daily flow
in a stream (i = 1 day). A continuous timescale results when data is recorded
16 1 Introduction

Fig. 1.2 An example of a hydrologic time series: monthly rainfall data

continuously with time, such as the stage at a streamgaging location. Even when a
continuous timescale is used for collecting the data, the analysis is usually done by
selecting values at specific time intervals. For example, raingage charts are usually
analyzed by reading the data at selected times (i.e. every 5 min) or at ‘break points’
(here i is not a constant).
Depending upon the hydrologic variable and the period of observation, a
hydrologic time series may be composed of only deterministic events (e.g. no
precipitation over a certain period) or only stochastic events (e.g. precipitation with
significant variability over a period of time) or a combination of the two (e.g.
precipitation over a period of time with a certain amount of variability as well as
trend). Generally speaking, and especially considering a reasonably long period of
observations, a hydrologic time series is of the third category, oftentimes with
stochastic components superimposed on deterministic components; for example,
the series composed of average monthly streamflows at some location would
contain the deterministic component of seasonal cycle plus the stochastic compo-
nent arising from random deviations from the seasonal values.
Despite this knowledge, hydrologic modeling efforts have mostly adopted either
only a deterministic approach or only a stochastic approach but not their combi-
nation (e.g. Yevjevich 1968; Quimpo 1971; Vogel 1999; Sivakumar 2008a, c). In
the deterministic approach, the hydrologic system is described and represented by
theoretical and/or empircal physical relationships; that is, there is always a unique
correspondence between the input, say precipitation, and the output, say stream-
flow. On the other hand, in the stochastic approach, a type of model is assumed
aiming to represent the most relevant statistical characteristics of the historic series.
Further details on the deterministic and stochastic components in hydrologic time
series and the available models can be seen in Haan (1994), Salas et al. (1995), and
Beven (2001), among others.
Over the past half a century, analysis of hydrologic time series has become a
fascinating endeavor and an important part of hydrologic studies. Two factors have
1.7 Hydrologic Data and Time Series Modeling 17

contributed mainly to this: developments in measurement technology (e.g.

improved precipitation and streamgages, remote sensors, geographic information
systems) and advances in data analysis concepts and tools (e.g. novel mathematical
and statistical methods, computers, softwares). Numerous time series analysis
methods have found their applications in hydrology, such as the linear and non-
linear stochastic techniques, scaling and fractal methods, artificial neural networks,
data-based mechanistic methods, wavelets, entropy-based methods, support vector
machines, genetic programming, fuzzy logic, and nonlinear dynamics and chaos.
All of these techniques may be put under the broad umbrella of ‘data-based’ or
‘data-driven’ approaches.
A plethora of literature is already available on these time series analysis tech-
niques and their applications in hydrology. Examples include: Box and Jenkins
(1970), Yevjevich (1972), and Salas et al. (1995) for stochastic methods;
Mandelbrot and Wallis (1969), Mandelbrot (1983), and Rodriguez-Iturbe and
Rinaldo (1997) for scaling methods; Haykin (1999), ASCE Task Committee
(2000a, b), Govindaraju and Rao (2000) for artificial neural networks; Young
(1984), and Young and Beven (1994), Young et al. (1996) for data-based mech-
anistic models; Chui (1992), Foufoula-Georgiou and Kumar (1994), and Labat
(2005) for wavelets; Singh (1997, 1998, 2013) for entropy concepts; Vapnik
(1995), Cristianini and Shawe-Taylor (2000), and Dibike et al. (2001) for support
vector machines; Holland (1975), Koza (1992), and Babovic (1996) for evolu-
tionary algorithms and genetic programming; Zadeh (1965), Bardossy and
Duckstein (1995), and Zadeh et al. (1996) for fuzzy logic; Tsonis (1992),
Sivakumar (2000), and Kantz and Schreiber (2004) for nonlinear dynamics and
chaos. A comprehensive review of many of these data-based methods and their
applications in hydrology is presented in Sivakumar and Berndtsson (2010).
Among these methods, the linear stochastic methods are much more popular and
established, partly due to their developments earlier than the others and partly due
to our assumption that hydrologic processes are stochastic in nature. The stochastic
assumption of hydrologic processes, brought about mainly due to their ‘complex’
and ‘random’ behaviors, was a significant deviation from the deterministic para-
digm that was dominant during the first half of the twentieth century. However, fast
developments in data measurement and computer technologies have made the other
methods equally attractive as well, such as the ones based on nonlinear dynamic
and chaos theories, which are the main focus of this book.

1.8 Physical Basis of Time Series Modeling

Although time series models have become immensely popular and found wide
applications in hydrology, they are oftentimes treated only as ‘black-box’ models.
This is because, time series models, despite their ability to represent even highly
complex and nonlinear input-output relationships based on data alone, often lack
the physical connections between the model structure/function and the catchment
18 1 Introduction

details/physics. The difficulty comes from the fact that the actual mechanisms in a
catchment occur at various (and all) scales but the data are usually measured only at
the catchment scale. Numerous attempts have been made to establish connections
between time series models and catchment physics, and a short list is presented
below.
From the viewpoint of stochastic time series models in hydrology, Yevjevich
(1963) and Fiering (1967) tried to set the physical basis of stochastic modeling, at
least for the case of autoregressive models. Moss and Bryson (1974) tried to
establish the physical basis of seasonal stochastic models. O’Connor (1976)
attempted to relate the unit hydrograph and flood routing models to autoregressive
and moving average models. Pegram (1977) and Selvalingam (1977) provided the
physical justification of continuous stochastic streamflow models. Other studies
offering physical explanation for stochastic time series models include Klemeš
(1978), Salas and Smith (1981), Parlange et al. (1992).
From a scaling perspective, Gupta et al. (1996) offered a theory to establish
connections between physics of floods and power laws. Their goal was first to
understand how spatial peak-discharge power laws are connected to physical pro-
cesses during rainfall-runoff events and then to extend this understanding to longer
timescales. These results have then been generalized in many directions (e.g. Gupta
and Waymire 1998; Menabde and Sivapalan 2001; Menabde et al. 2001; Morrison
and Smith 2001; Veitzer and Gupta 2001; Ogden and Dawdy 2003; Furey and
Gupta 2005, 2007). Gupta et al. (2007) provide an overview of progress that had
been made during the previous 20 years in understanding the physical origins of
spatial power laws that are observed, on average, in floods.
A few studies have attempted to provide a physical basis for the use of artificial
neural networks in hydrology. In their attempt to model daily river flow using
neural networks, Jain et al. (2004) suggested that two of the hidden units used in
their network were clearly capturing the baseflow component, as they were strongly
correlated with past river discharge, baseflow, and soil moisture. Sudheer and Jain
(2004) explained the internal structure and behavior of neural networks similar to
the function of the flow-duration curve. See et al. (2008) performed a similar
investigation, and results from different methods of analysis applied to both internal
and external outputs depicted comparable organization of hidden units into base-
flow, surface flow, and quick flow components. For futher details, the reader is also
referred to Abrahart et al. (2010).
Many studies employing time series methods in hydrology have adopted the
concept of ‘thresholds’ in many different forms, such as ‘critical states’ in the search
for self-organization in landscapes and river networks (e.g. Rodriguez-Iturbe and
Rinaldo 1997), ‘characteristic patterns’ of rainfall behavior for prediction of
streamflow using self-organizing maps (e.g. Hsu et al. 2002), and ‘regimes’ in the
prediction of streamflow dynamics using nonlinear dynamic and chaos methods
(e.g. Sivakumar 2003). Sivakumar (2005) tried to offer a physical explanation on
the role of ‘thresholds’ in catchments through an analogy between catchment
behavior and human behavior.
1.8 Physical Basis of Time Series Modeling 19

Despite these attempts and advances, establishing connections between data and
catchment physics continues to be a tremendously challenging task, and a coherent
approach to deal with this issue remains by and large elusive (e.g. Kirchner 2006;
Sivakumar 2008a). There are sufficient grounds to believe that a general ‘discon-
nection’ that has been present between researchers employing ‘physics-based’
approaches and those employing ‘data-based approaches’ has partially contributed
to this problem (e.g. Sivakumar 2008a). The difficulties in ‘communication’ (even
among those employing time series methods), largely because of the use of different
‘jargons’ in the literature, has made the situation only worse (e.g. Sivakumar 2005).
Current efforts that attempt to reconcile the upward (process-based) approaches
and the downward (data-based) approaches are encouraging towards establishing
connections between data and catchment physics. As scaling properties are
essentially related to the physics of the basins, scaling theories could also lead to a
better understanding of hydrologic systems and processes (e.g. Gupta 2004; Dawdy
2007). Further, since different time series methods possess different advantages,
another possibility to establish relations between data and catchment physics may
be to integrate two or more methods, each of which is suitable for the purpose at
hand but not as effective as their combination. The study by Young and Ratto
(2009), introducing some extensions to the data-based mechanistic approach (e.g.
Young and Beven 1994) and coupling the hypothetico-deductive approach of
simulation modeling with the inductive approach of data-based modeling, is a good
example to this idea of integration. Such integration of methods may also lead to
simplification in hydrologic modeling (e.g. Sivakumar 2004b). However, research
in these directions is still in a state of infancy, and there is certainly some distance
to go.

1.9 Scope and Organization of the Book

The central focus of this book is the role and applications of chaos theory and
related ideas in the field of hydrology. In the nonlinear science literature, the term
‘chaos’ refers to situations where complex and random-looking behaviors arise
from simple nonlinear deterministic systems with sensitive dependence on initial
conditions (Lorenz 1963), and the converse also applies. The three fundamental
properties inherent in this definition: (1) nonlinear inter-dependence; (2) hidden
determinism and order; and (3) sensitivity to initial conditions are highly relevant in
hydrologic systems and processes. For example: (1) components and mechanisms
involved in the hydrologic cycle act in a nonlinear manner and are also
inter-dependent; (2) daily cycle in temperature and annual cycle in river flow
possess determinism and order; and (3) contaminant transport phenomena in surface
and sub-surface waters largely depend upon the time (e.g. rainy or dry season) at
which the contaminants are released at the source, which themselves may not be
known. The first property represents the ‘general’ nature of hydrologic phenomena,
whereas the second and third represent their ‘deterministic’ and ‘stochastic’ natures,
20 1 Introduction

respectively. Further, despite their complexity and random-looking behavior,

hydrologic phenomena may also be governed by a very few degrees of freedom
(e.g. runoff in a well-developed urban catchment depends essentially on rainfall),
another fundamental idea of chaos theory (e.g. Sivakumar 2004a).
In view of these, chaos theory has found a growing number of applications in
hydrology during the last two decades or so. Sivakumar (2000, 2004a, 2009)
present extensive reviews of these applications and also address the current issues
and future challenges [see also Sivakumar and Berndtsson (2010)]. Much of the
advances thus far on chaos theory in hydrology has come in terms of time series
analysis. Therefore, a good understanding of the salient characteristics of hydro-
logic systems and processes and of the fundamentals of time series analysis is
needed in the first place to appreciate the role of chaos theory in hydrology and the
advances made (this is one reason for highlighting, in this chapter, the time series
analysis methods and their physical basis). With this in mind, the rest of this book is
organized as follows.
Chapter 2 discusses the salient characteristics of hydrologic systems and pro-
cesses, including complexity, correlation, trend, periodicity, cyclicity, and season-
ality, intermittency, stationarity and nonstationarity, linearity and nonlinearity,
determinism and randomness, scale and scale-invariance, self-organization and
self-organized criticality, threshold, emergence, feedback, sensitivity to initial
conditions, and the class of nonlinear determinism and chaos. Chapter 3 reviews
some popular conventional linear stochastic time series analysis methods (both
parametric and nonparametric; e.g. autoregressive, autoregressive moving average,
Markov, and k-nearest neighbor) and their applications in hydrology. In Chap. 4, a
brief account of some popular modern nonlinear time series methods (e.g. nonlinear
stochastic methods, data-based mechanistic methods, artificial neural networks,
support vector machines, wavelets, evolutionary computing, fuzzy logic, entropy
methods, nonlinear dynamic and chaos methods) and their hydrologic applications
is presented.
Part B is devoted to the theoretical aspects of nonlinear dynamic and chaos
concepts, with Chap. 5 on the fundamental ideas (e.g. dynamic systems, attractors,
bifurcations and chaos, phase space, dimension), Chap. 6 on the methods for
identification and prediction of chaos (e.g. phase space reconstruction, correlation
dimension method, false nearest neighbor algorithm, close returns plot, nonlinear
local approximation method), and Chap. 7 on the important issues associated with
the application of chaos techniques to finite (especially small) and noisy time series,
which are often the type of time series observed in hydrology.
Part C provides an extensive review of chaos theory applications in hydrology.
After an overview in Chap. 8, applications to rainfall and river flow are presented in
Chaps. 9 and 10, respectively. Applications to other processes (e.g. rainfall-runoff,
lake volume and level, sediment transport, ground water) are reviewed in Chap. 11,
while Chap. 12 presents details of studies on methodological and data issues in the
applications of chaos theory in hydrology (e.g. delay time, data size, data noise,
presence of zeros).
1.9 Scope and Organization of the Book 21

Part D looks to the future of chaos theory in hydrology. After re-visiting the
current status (e.g. successes, failures, limitations, concerns) in Chap. 13, some
potential directions for further advances (e.g. parameter identification,
multi-variable analysis, model simplification and integration, reconstruction of
system equations, linking data and physics), especially in light of future challenges
(including the impacts of climate change), are highlighted in Chap. 14. Finally,
Chap. 15 offers some thoughts on the philosophy and pragmatism in studying
hydrology and argues in favor of chaos theory as a balanced and middle-ground
approach to our dominant extreme views of determinism and stochasticity.

References

Abbott MB, Refsgaard JC (ed) (1996) Distributed hydrological modelling. Kluwer Academic
Publishers, Dordrecht, The Netherlands
Abrahart RJ, See LM, Dawson CW, Shamseldin AY, Wilby RL (2010) Nearly two decades of
neural network hydrologic modeling. In: Sivakumar B, Berndtsson R (eds) Advances in
data-based approaches for hydrologic modeling and forecasting. World Scientific Publishing
Company, Singapore, pp 267–346
Adams FD (1938) The birth and development of the geological sciences. Williams and Wilkins,
Baltimore, MD
Amorocho J (1967) The nonlinear prediction problems in the study of the runoff cycle. Water
Resour Res 3(3):861–880
Amorocho J, Brandstetter A (1971) Determination of nonlinear functional response functions in
rainfall-runoff processes. Water Resour Res 7(5):1087–1101
ASCE Task Committee (2000a) Artificial neural networks in hydrology. I: Preliminary concepts.
ASCE. J Hydrol Eng 5(2):115–123
ASCE Task Committee (2000b) Artificial neural networks in hydrology. II: Hydrologic
applications. ASCE. J Hydrol Eng 5(2):124–137
Babovic V (1996) Emergence, evolution, intelligence; hydroinformatics. Balkema Publishers,
Rotterdam
Bardossy A, Duckstein L (1995) Fuzzy rule-based modeling with applications to geophysical,
biological and engineering systems. CRC Press, Boca Raton, FL, USA, p 256
Beardmore N (1851) Manual of hydrology. Waterlow & Sons, London
Beven KJ (1993) Prophecy, reality and uncertainty in distributed modeling. Adv Water Resour
16:41–51
Beven KJ (1997) Distributed hydrological modelling: applications of the TOPMODEL concept.
John Wiley, Chichester, UK
Beven KJ (2001) Rainfall-runoff modelling: the primer. Wiley, Chichester, UK
Beven KJ (2002) Uncertainty and the detection of structural change in models of environmental
systems. In: Beck MB (ed) Environmental foresight and models: a manifesto. Elsevier, The
Netherlands, pp 227–250
Beven KJ (2006) On undermining the science? Hydrol Process 20:3141–3146
Biswas AK (1970) History of hydrology. North-Holland Publishing, Amsterdam
Box GEP, Jenkins G (1970) Time series analysis, forecasting and control. Holden-Day, San
Francisco
Brutsaert W, Mawdsley JA (1976) The applicability of planetary boundary layer theory to
calculate regional evapotranspiration. Water Resour Res 12:852–858
22 1 Introduction

Chahine MT (1992) The hydrological cycle and its influence on climate. Nature 359:373–380
Chow VT (ed) (1964) Handbook of applied hydrology. McGraw-Hill, New York
Chow VT, Maidment DR, Mays LW (1988) Applied hydrology. McGraw-Hill, Singapore
Chui CK (1992) An introduction to wavelets. Academic Press, Boston
Cristianini N, Shawe-Taylor J (2000) An introduction to support vector machines. Cambridge
University Press, Cambridge, UK
Dalton J (1802) Experimental essays on the constitution of mixed gases; on the force of steam or
vapor from waters and other liquids, both in a Torricellian vacuum and in air; on evaporation;
and on the expansion of gases by heat. Mem Proc Manch Lit Phil Soc 5:535–602
Darcy H (1856) Les fontaines publiques de la ville de Dijon. V. Dalmont, Paris
Dawdy DR (1969) Considerations involved in evaluating mathematical modeling of urban
hydrologic systems. U. S. Geological Survey Water Supply Paper 1591-D, U. S. Department of
Interior, Washington, D. C., D1-D18
Dawdy DR (2007) Prediction versus understanding (The 2007 Ven Te Chow Lecture). ASCE J
Hydrol Eng 12:1–3
Dawdy DR, Kalinin GP (1969) Mathematical modeling in hydrology. International Association of
Scientific Hydrology Report, Mid-Decade Conference of the International Hydrological
Decade, held in August in Surány, Hungary
DeCoursey DG (1971) The stochastic approach to watershed modeling. Nordic Hydrol 11:186–
216
Dibike YB, Velickov S, Solomatine DP, Abbott M (2001) Model induction with support vector
machines: introduction and applications. ASCE J Comput Civil Eng 15(2):208–216
Dooge JCI (1959) A general theory of the unit hydrograph. J Geophys Res 64(2):241–256
Dooge JCI (1965) Analysis of linear systems by means of Laguerre functions. Soc Indust Appl
Math (SIAM) J Cont 2(3):396–408
Dooge JCI (1967a) The hydrologic cycle as a closed system. Int Assoc Sci Hydrol Bull 13(1):58–
68
Dooge JCI (1967b) A new approach to nonlinear problems in surface water hydrology: hydrologic
systems with uniform nonlinearity. Int Assoc Sci Hydrol Publ 76:409–413
Dooge JCI (1977) Problems and methods of rainfall-runoff modeling. In: Ciriani TA, Maione U,
Wallis JR (eds) Mathematical models for surface water hydrology. John Wiley, New York,
pp 71–108
Driscoll FG (ed) (1986) Groundwater and wells. Johnson Division, St. Paul, Minnesota
Duan Q, Gupta HV, Sorooshian S, Rousseau AN, Turcotte R (2003) Calibration of watershed
models. Water Science and Application Series. American Geophysical Union, Washington vol
6, pp 1–346
Eagleson PS, Brutsaert WH, Colbeck SC, Cummins KW, Dozier J, Dunne T, Edmond JM,
Gupta VK, Jacoby JC, Manabe S, Nicholson SE, Nielsen DR, Rodríguez-Iturbe I, Rubin J,
Sposito G, Swank WT, Zipser EJ, Burges S (1991) Opportunities in the hydrologic sciences.
National Academy Press, Washington, DC, p 348
Fiering MB (1967) Streamflow synthesis. Harvard University Press, Cambridge, Massachusetts
Foufoula-Georgiou E, Kumar P (eds) (1994) Wavelets in geophysics. Academic Press, San Diego,
California, p 373
Freeze RA (1971) Three-dimensional, transient, saturated-unsaturated flow in a groundwater basin.
Water Resour Res 7(2):347–366
Freeze RA, Cherry JA (1979) Groundwater. Prentice-Hall, Englewood Cliffs, New Jersey
Furey PR, Gupta VK (2005) Effects of excess-rainfall on the temporal variability of observed peak
discharge power laws. Adv Water Resour 28:1240–1253
Furey PR, Gupta VK (2007) Diagnosing peak-discharge power laws observed in rainfall–runoff
events in Goodwin Creek experimental watershed. Adv Water Resour 30:2387–2399
Gelhar LW (1993) Stochastic subsurface hydrology. Prentice-Hall, Englewood Cliffs, New Jersey
References 23

Govindaraju RS (ed) (2002) Stochastic methods in subsurface contaminant hydrology. ASCE,

New York
Govindaraju RS, Rao AR (2000) Artificial neural networks in hydrology. Kluwer Acadmic
Publishers, Amsterdam
Gumbel EJ (1941) The return period of flood flows. Ann Math Stat 12(2):163–190
Gupta VK (2004) Emergence of statistical scaling in floods on channel networks from complex
runoff dynamics. Chaos Soliton Fract 19:357–365
Gupta VK, Waymire E (1998) Spatial variability and scale invariance in hydrologic regional-
ization. In: Sposito G (ed) Scale dependence and scale invariance in hydrology. Cambridge
University Press, Cambridge, pp 88–135
Gupta VK, Rodríguez-Iturbe I, Wood EF (eds) (1986) Scale problems in hydrology: runoff
generation and basin response. Reidel Publishing Company, FD, p 244
Gupta VK, Castro SL, Over TM (1996) On scaling exponents of spatial peak flows from rainfall
and river network geometry. J Hydrol 187:81–104
Gupta VK, Troutman BM, Dawdy DR (2007) Towards a nonlinear geophysical theory of floods in
river networks: an overview of 20 years of progress. In: Tsonis AA, Elsner JB (eds) Twenty
years of nonlinear dynamics in geosciences. Springer Verlag
Haan CT (1994) Statistical methods in hydrology. Iowa University Press, Iowa
Harley BM (1967) Linear routing in uniform flow channels. M. Eng. Sc. Thesis, National
University of Ireland, University College, Cork, Ireland
Haykin S (1999) Neural networks: a comprehensive foundation. Prentice Hall International Inc,
New Jersey
Hele-Shaw HS (1898) The flow of water. Nature 58:34–36
Holland JH (1975) Adaptation in natural and artificial systems. University of Michigan Press, Ann
Arbor, MI
Horden RH (1998) The hydrologic cycle. In: Herschy RW, Fairbridge RW (eds) Encyclopedia of
hydrology and water resources. Kluwer Academic Publishers, Dordrecht, The Netherlands,
pp 400–404
Horton RE (1933) The role of infiltration in the hydrologic cycle. Trans Am Geophys Union
14:446–460
Horton RE (1945) Erosional development of streams and their drainage basins: Hydrophysical
approach to quantitative morphology. Bull Geol Soc Am 56:275–370
Hsu KL, Gupta HV, Gao X, Sorooshian S, Imam B (2002) Self-organizing linear output map
(SOLO): an artificial neural network suitable for hydrologic modeling and analysis. Water
Resour Res 38(12). doi:10.1029/2001WR000795
Hurst HE (1951) Long-term storage capacity of reservoirs. Trans Am Soc Civ Eng 116:770–799
Hydrologic Engineering Center (1995) HEC-HMS: Hydrologic modeling system user’s manual.
U. S, Army Corps of Engineers, Davis, CA
Hydrologic Engineering Cener (1998) HEC-RAS: river analysis system user’s manual. U.S. Army
Corps of Engineers, Davis, CA
IPCC (2014) Climate change 2014—impacts, adaptation and vulnerability. Contribution of
Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate
Change [Field CB, Barros VR, Dokken DJ, Mach KJ, Mastrandrea MD, Bilir TE,
Chatterjee M, Ebi KL, Estrada YO, Genova RC, Girma B, Kissel ES, Levy AN,
MacCracken S, Mastrandrea PR, White LL (eds.)], Cambridge University Press, Cambridge
Jacoby SLS (1966) A mathematical model for nonlinear hydrologic systems. J Geophys Res 71
(20):4811–4824
Jain A, Sudheer KP, Srinivasulu S (2004) Identification of physical processes inherent in artificial
neural network rainfall runoff models. Hydrol Process 18(3):571–581
Kalma JD, Sivapalan M (1996) Scale issues in hydrological modeling. Wiley, London
Kantz H, Schreiber T (2004) Nonlinear time series analysis. Cambridge University Press,
Cambridge
24 1 Introduction

Kirchner JW (2006) Getting the right answers for the right reasons: linking measurements,
analyses, and models to advance the science of hydrology. Water Resour Res 42:W03S04.
doi:10.1029/2005WR004362
Klemeš V (1978) Physically based stochastic hydrologic analysis. Adv Hydrosci 11:285–352
Klemeš V (1982) Empirical and causal models in hydrology. Scientific basis of water-resources
management. National Academic Press, Washington, DC, pp 95–104
Klemeš V (1986) Dilettantism in hydrology: Transition or destiny? Water Resour Res 22(9):177S–
188S
Konikow LF, Bredehoeft JD (1992) Ground-water models cannot be validated. Adv Water Resour
15:75–83
Koza JR (1992) Genetic programing: on the programming of computers by natural selection. MIT
Press, Cambridge, MA
Kuichling E (1889) The relation between the rainfall and the discharge of sewers in populous
districts. Trans Am Soc Civ Eng 20:1–56
Labat D (2005) Recent advances in wavelet analyses: Part 1. A review of concepts. J Hydrol
314:275–288
Lorenz EN (1963) Deterministic nonperiodic flow. J Atmos Sci 20:130–141
MacNeill IB, Umphrey GJ (1987) Stochastic hydrology. Kluwer Academic Publishers, Boston
Maidment DR (ed) (1993) Handbook of hydrology. McGraw-Hill, New York
Mandelbrot BB (1967) How long is the coast of Britain? Statistical self-similarity and fractional
dimension. Science 156:636–638
Mandelbrot BB (1975) On the geometry of homogeneous turbulence, with stress on the fractal
dimension of the isosurfaces of scalars. J Fluid Mech 72:401–416
Mandelbrot BB (1977) Fractals: form, chance and dimension. W. H, Freeman and Co, New York
Mandelbrot BB (1983) The fractal geometry of nature. Freeman, New York
Mandelbrot BB, Wallis JR (1968) Noah, Joseph and operational hydrology. Water Resour Res 4
(5):909–918
Mandelbrot BB, Wallis JR (1969) Some long run properties of geophysical records. Water Resour
Res 5(2):321–340
Manning R (1891) On the flow of water in open channels and pipes. Trans Inst Civ Eng Ireland
20:161–207
McDonnell JJ, Woods RA (2004) On the need for catchment classification. J Hydrol 299:2–3
Menabde M, Sivapalan M (2001) Linking space-time variability of river runoff and rainfall fields:
a dynamic approach. Adv Wat Resour 24:1001–1014
Menabde M, Veitzer S, Gupta VK, Sivapalan M (2001) Tests of peak flow scaling in simulated
self-similar river networks. Adv Wat Resour 24:991–999
Metcalf and Eddy, Inc., University of Florida, and Water Resources Engineers, Inc., (1971) Storm
water management model, vol. 1. Final Report, 11024DOC07/71 (NTIS PB-203289), U.S.
EPA, Washington, DC
Miller WA, Woolhiser DA (1975) Choice of models. In: Yevjevich VM (ed) Unsteady flow in
open channels. Water Resources Publications, Littleton, Colorado
Minshall NE (1960) Predicting storm runoff on small experimental watersheds. J Hydraul Div Am
Soc Eng 86(HYB):17–38
Morel-Seytoux HJ (1978) Derivation of equations for variable rainfall infiltration. Water Resour
Res 14(4):561–568
Morrison JE, Smith JA (2001) Scaling properties of flood peaks. Extremes 4(1):5–22
Moss ME, Bryson MC (1974) Autocorrelation structure of monthly streamflows. Water Resour
Res 10(4):737–744
Nace RL (1974) General evolution of the concept of the hydrological cycle. Three centuries of
scientific hydrology. UNESCO-World Meteorological Organization-International Association
of Hydrological Sciences, Paris, pp 40–48
Nash JE (1958) Determining runoff from rainfall. Proc Instit Civil Eng Ireland 10:163–184
References 25

Neitsch SL, Arnold JG, Kiniry JR, Williams JR (2005) Soil and water assessment tool theoretical
documentation, Version 2005. USDA.ARS Grassland, Soil and Water Research Laboratory,
Temple, TX
O’Connor KM (1976) A discrete linear cascade model for hydrology. J Hydrol 29:203–242
O’Donnell T (1960) Instantaneous unit hydrograph derivation by harmonic analysis. Int Assoc Sci
Hydrol Publ 51:546–557
O’Meara WA (1968) Linear routing of lateral inflow in uniform open channels. M. E. Sc. Thesis,
The National University of Ireland, University College, Cork, Ireland
Ogden FL, Dawdy DR (2003) Peak discharge scaling in a small hortonian watershed. J Hydrol
Eng 8(2):64–73
Parlange MB, Katul GG, Cuenca RH, Kavvas ML, Nielsen DR, Mata M (1992) Physical basis for
a time series model of soil water content. Water Resour Res 28(9):2437–2446
Pegram GGS (1977) Physical justification of continuous streamflow model. In: Morel-Seytoux HJ,
Salas JD, Sanders TG, Smith RE (eds) Modeling hydrologic processes. Proceedings of Fort
Collins III international hydrol Symposium, pp 270–280
Quimpo R (1971) Structural relation between parametric and stochastic hydrology models. In:
Mathematical models in hydrology, Warsaw Symposium (IAHS Publication 100) 1:151–157
Richards LA (1931) Capillary conduction of liquids through porous mediums. Physics A 1:318–
333
Rodriguez-Iturbe I, Rinaldo A (1997) Fractal river basins: chance and self-organization.
Cambridge University Press, Cambridge
Rodriguez-Iturbe I, De Power FB, Sharifi MB, Georgakakos KP (1989) Chaos in rainfall. Water
Resour Res 25(7):1667–1675
Salas JD, Smith RA (1981) Physical basis of stochastic models of annual flows. Water Resour Res
17(2):428–430
Salas JD, Delleur JW, Yevjevich V, Lane WL (1995) Applied modeling of hydrologic time series.
Water Resources Publications, Littleton, Colorado
See LM, Jain A, Dawson CW, Abrahart RJ (2008) Visualisation of hidden neuron behaviour in a
neural network rainfall-runoff model. In: Abrahart RJ, See LM, Solomatine DP (eds) Practical
hydroinformatics: computational intelligence and technological developments in water
applications. Springer-Verlag, Berlin, Germany, pp 87–99
Selvalingam S (1977) ARMA and linear tank models. In: Morel-Seytoux HJ, Salas JD,
Sanders TG, Smith RE (eds) Modeling hydrologic processes. Proceedings of Fort Collins III
international hydrol symposium, pp 297–313
Singh VP (1988) Hydrologic systems: vol 1. Rainfall-runoff modeling, Prentice Hall, New Jersey
Singh VP (1995) Computer models of watershed hydrology. Water Resources Publications,
Highlands Ranch CO
Singh VP (1997) The use of entropy in hydrology and water resources. Hydrol Process 11:587–
626
Singh VP (1998) Entropy-based parameter estimation in hydrology. Kluwer Academic, Dordrecht,
The Netherlands
Singh VP (2013) Entropy theory and its application in environmental and water engineering. John
Wiley and Sons, Oxford, UK
Sivakumar B (2000) Chaos theory in hydrology: important issues and interpretations. J Hydrol 227
(1–4):1–20
Sivakumar B (2003) Forecasting monthly streamflow dynamics in the western United States: a
nonlinear dynamical approach. Environ Model Softw 18(8–9):721–728
Sivakumar B (2004a) Chaos theory in geophysics: past, present and future. Chaos Soliton Fract 19
(2):441–462
Sivakumar B (2004b) Dominant processes concept in hydrology: moving forward. Hydrol Process
18(12):2349–2353
26 1 Introduction

Sivakumar B (2005) Hydrologic modeling and forecasting: role ofthresholds. Environ Model
Softw 20(5):515–519
Sivakumar B (2008a) Dominant processes concept, model simplification and classification
framework in catchment hydrology. Stoch Env Res Risk Assess 22:737–748
Sivakumar B (2008b) Undermining the science or undermining Nature? Hydrol Process 22
(6):893–897
Sivakumar B (2008c) The more things change, the more they stay the same: the state of hydrologic
modeling. Hydrol Process 22:4333–4337
Sivakumar B (2009) Nonlinear dynamics and chaos in hydrologic systems:latest developments and
a look forward. Stoch Environ Res Risk Assess 23:1027–1036
Sivakumar B, Berndtsson R (2010) Advances in data-based approaches for hydrologic modeling
and forecasting. World Scientific Publishing Company, Singapore
Sivakumar B, Singh VP (2012) Hydrologic system complexity and nonlinear dynamic concepts
for a catchment classification framework. Hydrol Earth Syst Sci 16:4119–4131
Sivakumar B, Sorooshian S, Gupta HV, Gao X (2001) A chaotic approach to rainfall
disaggregation. Water Resour Res 37(1):61–72
Sivapalan M, Takeuchi K, Franks SW, Gupta VK, Karambiri H, Lakshmi V, Liang X,
McDonnell JJ, Mendiondo EM, O’Connell PE, Oki T, Pomeroy JW, Schertzer D,
Uhlenbrook S, Zehe E (2003) IAHS decade on predictions in ungauged basins (PUB),
2003–2012: shaping an exciting future for the hydrological sciences. Hydrol Sci J 48(6):
857–880
Snyder WM (1955) Hydrograph analysis by the method of least squares. J Hyd Div, Proc Ame Soc
Civ Eng 81(793), 25 pp
Snyder WM (1971) The parametric approach to watershed modeling. Nordic Hydrol 11:167–185
Snyder WM, Stall JB (1965) Men, models, methods, and machines in hydrologic analysis. J Hyd
Div, Proc Ame Soc Civ Eng 91(HY2):85–99
Sorooshian S, Gupta VK (1983) Automatic calibration of conceptual rainfall-runoff models: the
question of parameter observability and uniqueness. Water Resour Res 19(1):251–259
Sudheer KP, Jain A (2004) Explaining the internal behavior of artificial neural network river flow
models. Hydrol Process 18(4):833–844
Thomas HA, Fiering MB (1962) Mathematical synthesis of streamflow sequences for the analysis
of river basins by simulation. In: Mass A et al (eds) Design of water resource systems. Harvard
University Press, Cambridge, Massachusetts, pp 459–493
Tsonis AA (1992) Chaos: from theory to applications. Plenum Press, New York
UNESCO (1974) Contributions to the development of the concept of the hydrological cycle. Sc.
74/Conf.804/Col. 1, Paris
Veitzer SA, Gupta VK (2001) Statistical self-similarity of width function maxima with
implications to floods. Adv Wat Resour 24:955–965
Vapnik V (1995) The nature of statistical learning theory. Springer-Verlag, Germany
Vogel RM (1999) Stochastic and deterministic world views. J Water Resour Plan Manage 125
(6):311–313
Woolhiser DA (1971) Deterministic approach to watershed modeling. Nordic Hydrol 11:146–166
Woolhiser DA (1973) Hydrologic and watershed modeling: state of the art. Trans Ame Soc Agr
Eng 16(3):553–559
Woolhiser DA (1975) The watershed approach to understanding our environment. J Environ Qual
4(1):17–21
Yevjevich VM (1963) Fluctuations of wet and dry years. Part 1. Research data assembly and
mathematical models. Hydrology Paper 1, Colorado State University, Fort Collins, Colorado,
pp 1–55
Yevjevich VM (1968) Misconceptions in hydrology and their consequences. Water Resour Res 4
(2):225–232
References 27

Yevjevich VM (1972) Stochastic processes in hydrology. Water Resour Publ, Fort Collins,
Colorado
Young PC (1984) Recursive estimation and time-series analysis. Springer-Verlag, Berlin
Young PC, Beven KJ (1994) Database mechanistic modeling and rainfall-flow non-linearity.
Environmetrics 5(3):335–363
Young PC, Ratto M (2009) A unified approach to environmental systems modeling. Stoch Environ
Res Risk Assess 23:1037–1057
Young PC, Parkinson SD, Lees M (1996) Simplicity out of complexity in environmental systems:
Occam’s Razor revisited. J Appl Stat 23:165–210
Zadeh LA (1965) Fuzzy sets. Inform. Control 8(3):338–353
Zadeh LA, Klir GJ, Yuan B (1996) Fuzzy sets, fuzzy logic and fuzzy systems. World Scientific
Publishers, Singapore