17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GCB Bioenergy (2011) 3, 413–426, doi: 10.1111/j.1757-1707.2011.01102.x

CO2 emissions from biomass combustion for bioenergy:

atmospheric decay and contribution to global warming
F R A N C E S C O C H E R U B I N I *, G L E N P. P E T E R S w , T E R J E B E R N T S E N w z,
A N D E R S H . S T R Ø M M A N * and E D G A R H E R T W I C H *
*Department of Energy and Process Engineering, Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim,
Norway, wCenter for International Climate and Environmental Research – Oslo (CICERO), Oslo, Norway, zDepartment of
Geosciences, University of Oslo, Norway

Abstract
Carbon dioxide (CO2) emissions from biomass combustion are traditionally assumed
climate neutral if the bioenergy system is carbon (C) flux neutral, i.e. the CO2 released
from biofuel combustion approximately equals the amount of CO2 sequestered in
biomass. This convention, widely adopted in life cycle assessment (LCA) studies of
bioenergy systems, underestimates the climate impact of bioenergy. Besides CO2 emis-
sions from permanent C losses, CO2 emissions from C flux neutral systems (that is from
temporary C losses) also contribute to climate change: before being captured by biomass
regrowth, CO2 molecules spend time in the atmosphere and contribute to global
warming. In this paper, a method to estimate the climate impact of CO2 emissions from
biomass combustion is proposed. Our method uses CO2 impulse response functions
(IRF) from C cycle models in the elaboration of atmospheric decay functions for biomass-
derived CO2 emissions. Their contributions to global warming are then quantified with a
unit-based index, the GWPbio. Since this index is expressed as a function of the rotation
period of the biomass, our results can be applied to CO2 emissions from combustion of
all the different biomass species, from annual row crops to slower growing boreal forest.
Keywords: bioenergy, carbon neutral, CO2 accounting, global warming potential, LCA

Received 14 October 2010; revised version received 31 January 2011 and accepted 7 February 2011

the land use, land-use change and forestry (LULUCF)

Introduction
sector, according to country-specific regulations
(UNFCCC, 2003; IPCC, 2006). Stemming from this con-
Background
vention, primary research life cycle assessment (LCA)
In 1991, the first comprehensive guidelines for estimat- studies tend to implicitly assume CO2 emissions from
ing national greenhouse gas (GHG) emissions and sinks biomass combustion climate neutral if the bioenergy
compiled by the Organization for Economic Coopera- system is C flux neutral, i.e. CO2 emissions from tem-
tion and Development (OECD) states that ‘CO2 emis- porary C losses are traditionally ignored.
sions resulting from bioenergy consumption should not In LCA studies of bioenergy systems, the OECD
be included in a country’s official emission inventory’ convention is implemented following two basic ac-
(OECD, 1991). This convention is motivated by the counting procedures. The majority of case studies
consideration of the carbon (C) neutrality of bioenergy: ignore the CO2 flux within a bioenergy system, assum-
because growing forests sequester C, then as long as ing that CO2 absorbed equals CO2 emitted, so giving a
areas harvested for biomass are kept forested, the C is net flux balance of zero; these studies simply assign a
again absorbed in growing trees and consequently the global warming potential (GWP) equal to zero to direct
net impact on GHG emissions is zero (Manomet, 2010). CO2 emissions (e.g., Carpentieri et al., 2005; Petersen
For this reason, in national GHG inventories direct Raymer, 2006; Huo et al., 2008; Kim & Dale, 2008). Other
carbon dioxide (CO2) emissions from bioenergy are studies follow the EcoInvent database (Werner et al.,
not reported in the energy sector (as for fossils) but in 2003) and offset CO2 emissions from biomass combus-
tion with an upstream sequestration credit that is nearly
Correspondence: Francesco Cherubini, tel. 1 477 359 8942, equal to the combustion emission. In this case, a GWP
e-mail: [email protected] equal to 1 is assigned to CO2, which is considered to be

r 2011 Blackwell Publishing Ltd 413

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
414 F . C H E R U B I N I et al.

offset by the sequestration of the same amount of CO2 which focus on bioenergy from fast growing biomass
that occurred to grow biomass (Reijnders & Huijbregts, generally tend to account for permanent changes in
2008; Luo et al., 2009). terrestrial C pools only, while basically ignoring the
These accounting conventions are so widely adopted climate impact of CO2 from temporary changes (i.e.
that in the majority of LCA studies it is not even biomass harvested for bioenergy and then regrown).
mentioned which one of the two is used (van der Voet This is a reasonable assumption for fast growing spe-
et al., 2010; Cherubini & Strmman, 2011). A recent cies, but may not apply in the case of biofuels from
paper reports that in only four of the 67 case studies slower growing biomass, like forests (Johnson, 2009;
reviewed the exclusion of the climate effect of biomass- Marland, 2010). A forest may take up to 100 years to
derived CO2 emissions is explicitly indicated, while in regrow, and the system can be defined C neutral only at
two cases it is clearly mentioned that emissions and the end of proper time boundaries: CO2 is emitted in
removals are both included and offset (van der Voet one point in time when biomass is burnt but the
et al., 2010). Most of the studies generally find a reduc- sequestration in the new vegetation is spread over
tion in the contribution to climate change when bio- several years, depending on the specific rotation period.
energy systems are compared to fossil reference Even if in these cases the fact that C neutral does not
systems, provided that permanent changes in terrestrial mean climate neutral is straightforward, this aspect has
C pools are minimized (Quirin et al., 2004; Searcy & been seldom considered in LCA, despite the importance
Flynn, 2008). One of the main reasons for this result is of the issue being thoroughly acknowledged from the
the absence in GHG balances of the climate impact of early 1990s (Harmon et al., 1990; Marland & Schlama-
CO2 emitted from biomass combustion. dinger, 1995; Schlamadinger & Marland, 1996b). Studies
Both in past and recent literature, an increasing that considered the time dimensions of forest growth
perception of the inadequacy of this accounting con- are essentially studies of forest C dynamics. These
vention and its implementation in LCA can be identi- studies usually report an increase in GHG emissions
fied. Already some years ago, Börjesson & Gustavsson of forest bioenergy systems in the short term, in favor of
(2000) did not presume wood to be C neutral and a decrease in net GHG emissions in the longer term; in
accounted for CO2 emissions from biomass as those some cases, a specific C deficit and pay-back time (up to
from fossils. Rabl et al. (2007) advocated ‘that emission some decades, depending on site-specific parameters
and removal of CO2 be accounted explicitly at each and reference system) is identified (Marland & Schla-
stage of the life cycle’ . Even if they realized that the net madinger, 1995; Schlamadinger & Marland, 1996b;
effect at the end would be almost zero, they claim that Manomet, 2010; McKechnie et al., 2010). Many analyti-
using this approach allows a dynamic modeling of cal models are available to perform this type of tempor-
emissions and removals. Others have questioned the al analysis (Schlamadinger & Marland, 1996a; Masera
distinction between fossil and biomass-derived CO2 in et al., 2003; Schelhaas et al., 2004; Kurz et al., 2009). A
national GHG accounting, emphasizing that ‘all CO2 is common feature of these assessments is to show results
equal in the atmosphere’ and IPCC only provides vague as a trend of cumulative CO2 emissions over century
guidance concerning this crucial matter, and further timescales, and do not elaborate yearly unit based
detailed analysis would be highly desirable to accu- indicators. The work performed in this paper bridges
rately account for all CO2 fluxes (Möllersten & Grönk- this type of analysis with LCA methodology, providing
vist, 2007). Johnson states that we should say ‘goodbye a methodology to estimate the contribution to global
to C neutral’ for bioenergy from forests (Johnson, 2009), warming of CO2 flux neutral bioenergy systems in
while other researchers have focused on fixing ‘a critical terms of GWP, so to provide an index which can be
climate accounting error’ (Searchinger et al., 2009; promptly included in LCA.
Searchinger, 2010). Searchinger et al. (2009) moved a
step forward, stating that ‘replacing fossil fuels with
Aims and objectives
bioenergy does not by itself reduce C emissions’, since
the CO2 released by tailpipe emissions ‘is roughly the All CO2 emissions, both from combustion of fossil fuels
same per unit of energy’: in order to mitigate climate or biomass, alter the C cycle and hence the earth’s
change, bioenergy must ensure that ‘the growth and radiative balance, thus causing a climate impact that
harvesting of the biomass for energy captures more C should be estimated. Our main aim in this paper is to
above and beyond what would be sequestered anyway quantify the climate impact of biomass-derived CO2
and thereby offset emissions from energy use’. emissions with a unit-based indicator to be used in LCA
A further distinction can be seen between LCA based or C accounting studies. We focus on a single biomass
on forest wood and fast growing biomass species (an- rotation where an existing aboveground C stock, either
nual crops and lignocellulosic energy crops). Studies a crop or a forest, is harvested for bioenergy and later

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G L O B A L WA R M I N G P O T E N T I A L O F C O 2 F R O M B I O E N E R G Y 415

allowed to regrow. We use this schematic case to retain in the oceans have been formulated (Oeschger et al.,
the focus on the key research question, without adding 1975; Siegenthaler & Joos, 1992; Blanke & Delecluse,
the complexity and additional assumptions linked to 1993; Caldeira & Kasting, 1993).
the possibilities of using specific factors like local con- Modeling the terrestrial components of the C cycle is
ditions and biomass management strategies. more challenging because of the natural variability of
This paper is structured as follows. The current some basic parameters (Enting et al., 2001). The most
method used to estimate the atmospheric decay of common way of modeling terrestrial C transfers is to
anthropogenic CO2 emissions is firstly described to- use discrete compartments as leaves, branches, soil C,
gether with a metric for measuring their contribution etc., characterized by an initial C content and turnover
to GWP. Afterwards, the climate impact of CO2 emis- times. The C transfers from the air to the plants is
sions from biomass combustion (bio CO2, from this described by a net primary production, which may
point forward) is investigated through the formulation depend on specific parameters like temperature, nutrient
of proper atmospheric decay functions, which are used levels, water supply and others. The terrestrial part of the
in the GWPbio index. Finally, results are presented as a different climate models usually differ in the number of
function of the biomass rotation period, and the most physiological compartments, feedback effects and the
relevant implications related to this methodology are degree of disaggregation (Friedlingstein et al., 1994, 1995;
discussed in the final section. Prentice et al., 2000; Cramer et al., 2001; McGuire et al.,
2001).
Materials and methods
Atmospheric decay. Thanks to the elaboration of these CC
Anthropogenic CO2 emissions models it is possible to predict the atmospheric decay of
CO2 emissions (Maier-Reimer & Hasselmann, 1987;
Lashof & Ahuja, 1990; Caldeira & Kasting, 1993; Joos
C cycle climate models. CO2 emissions play a key role et al., 1996, 2001; Enting et al., 2001). In all the cases, CO2
in the earth’s C cycle and climate system. Those which does not follow a simple decay according to one single
are classified as anthropogenic (i.e. from fossil fuel lifetime (as it is for the two other main GHG, N2O and
combustion, cement production, deforestation and CH4), but its decay is described by several time
land-use change) are one of the main responsible for constants and there is a fraction of the initial emission
anthropogenic climate change (Forster et al., 2007). that always remains in the atmosphere. The fraction
Complex C cycle climate (CC) models, which establish of CO2 remaining in the air following a CO2 release
the link between atmospheric CO2 concentration and depends on future atmospheric CO2 concentrations,
anthropogenic C emissions by modeling uptake and because the partial pressure of CO2 in the ocean surface
exchange fluxes of the atmosphere with the oceans and is a nonlinear function of surface total dissolved inorganic
the terrestrial biosphere, are used to model the time C concentration (Caldeira & Kasting, 1993).
evolution of airborne CO2. In order to make analysis The analytical form of the atmospheric decay of
easier for smaller case studies, such as LCA, impulse anthropogenic CO2 is given by a superposition of a
response functions (IRF) are often used to represent number of exponentials of different amplitude Ai and
CO2 atmospheric decay under given assumptions relaxation time ti
(Tubiello & Oppenheimer, 1995; Joos & Bruno, 1996;
Enting et al., 2001). X
i¼n
yCO2 ðtÞ ¼ A0 þ Ai eðt=ti Þ : ð1Þ
The oceans play an important role for the removal of
i¼1
anthropogenic C. They are generally distinguished into
the upper layer, which has a very fast turnover rate
(Wanninkhof, 1992), and the deep ocean, to which C is The value of this function at any time represents the
transported through oceanic circulation (Joos, 2003). fraction of the initial emission which is still found in the
This latter process is the limiting factor for the ocean’s atmosphere, and the removed fraction corresponds to the
uptake capacity, which is determined by ocean volume ocean/biosphere uptake. The amplitude A0 represents
and sea water chemistry. This uptake capacity is the asymptotic airborne fraction of CO2 which remains in
only reached after several centuries, and it takes the atmosphere because of the equilibrium response of
millennia to equilibrate ocean water and sediments the ocean–atmosphere system. The amplitudes Ai may be
after a perturbation in oceanic C content. Changes in interpreted as the relative capacity of the other sinks,
the land biosphere and in the upper ocean influence which are filled up by the atmospheric input at rates
atmospheric CO2 concentrations on seasonal to century characterized by the relaxation time scales ti. These time
time scales. Several models dealing with the C cycle scales determine the redistribution of anthropogenic CO2

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
416 F . C H E R U B I N I et al.

emissions in the climate system and are linked to the time value of the radiative efficiency for CO2 is 1.41  105
scales of the natural C cycle. Because of this exponential W m2 ppb1.
decay trend, more than half of the initial input is removed Since the decay of a CO2 pulse emission has a non-
from the atmosphere within few decades after emissions zero asymptote, its integral from zero to infinity is
through uptake by the upper ocean layer and the fast infinite. To avoid this, several attempts to define an
overturning reservoirs of the land biosphere. However, a effective residence time for CO2 in the air have been
certain fraction is still found in the atmosphere after 1000 formulated (Houghton et al., 1990; Lashof & Ahuja,
years; this fraction is only very slowly reduced further by 1990; Rodhe, 1990; Moore & Braswell, 1994). In the
ocean–sediment interaction and the weathering cycle 1990s, the IPCC introduced finite time horizons (THs)
(Archer et al., 1998). (20, 100 and 500 years) for integration in the GWP,
where the CO2 decay function by Joos et al. (1996) was
used (Schimel et al., 1996). As specified by the IPCC
Metrics for climate change itself, these different THs should not be considered of
The climate impact of GHG emissions needs to be any scientific significance (Fuglestvedt et al., 2003; For-
compared with a consistent metric. In this paper the ster et al., 2007). GWPs were then elaborated for all the
GWP is used, rather than other possible metrics (Fugle- different GHGs (denoted as i) according to this equation
R TH
stvedt et al., 2003; Shine et al., 2005). This metric was AGWPi C0 0 ai yi ðtÞdt
developed as a relative measure of the potential effects GWPi ¼ ¼ R : ð4Þ
AGWPCO2 C0 TH aCO2 yCO2 ðtÞdt
0
on climate of a GHG compared with CO2. GWP heavily
relies on the concept of radiative forcing which gives GWP then acts as a metric able to aggregate emission of
the perturbation of the earth energy balance at the top the various gases to a common unit (kg CO2-eq.). In
of the atmosphere by a climate change mechanism. The Table 1, GWPs for given THs are shown for the three
cumulative radiative forcing for a pulse emission, most important GHGs, together with their lifetime and
which is often referred to as the absolute global warm- radiative efficiency.
ing potential (AGWP), is given by the integral over time
of the product between the radiative efficiency of the
gas (a) and the decay function, y(t), that defines the CO2 emissions from biomass combustion
fraction of the gas remaining in the atmosphere after a The atmospheric decay of CO2 emissions from biomass
unit pulse (C0) combustion can be predicted with the IRF from C
Z 1 climate models only if biomass is not replanted (i.e.
AGWP ¼ C0 ayðtÞdt; ð2Þ deforestation), or a LUC occurs. Even if consistent
0
results were achieved in upgrading the modeling of
where the radiative efficiency (a) of CO2 is (Forster et al., the biosphere compartment (Gerber et al., 2004), the basic
2007) principles remain unchanged: if biomass is replanted,
0h i1 emissions from combustion are neutralized by CO2 re-
CO2 moval during regrowth; if biomass is not replanted, bio
aCO2 ¼ 5:35 ln@   A: ð3Þ
CO2 CO2 emissions become anthropogenic CO2 (Strassmann
et al., 2008). Then, a new IRF needs to be elaborated to
Where [CO2*] is the concentration in the atmosphere predict the atmospheric decay of bio CO2.
after small perturbation and [CO2] is the initial concen-
tration of CO2 in the atmosphere. If the background Modeling assumptions. The method developed in this
concentration of 378 ppm provided by the IPCC report paper is applied to a well-defined schematic case
is used, and a perturbation of 1 ppm is applied, the study that is suitable to demonstrate the approach

Table 1 Lifetime, radiative efficiency, and global warming potentials (GWPs) for different time horizons of the three most
important greenhouse gases (GHGs)

Radiative efficiency
GHG Lifetime (years) (W m2 ppb1) GWP 20 years GWP 100 years GWP 500 years

Carbon dioxide (CO2) na 1.4  105 1 1 1

Methane (CH4) 12 3.7  104 72 25 7.6
Nitrous oxide (N2O) 114 3.03  103 289 298 153

na, not available.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G L O B A L WA R M I N G P O T E N T I A L O F C O 2 F R O M B I O E N E R G Y 417

proposed (see Fig. 1). It is assumed that all biomass is means of an IRF which refers to the reaction (as a
burnt in one time step so that the CO2 emission is function of time) of any dynamic system in response
modeled as a pulse. The biomass harvested is from an to some external change. In our case, this means that the
even-aged vegetation stand (representing the starting atmospheric decay of bio CO2 is derived through
condition) which is clear cut and the land is combination of the biomass regrowth sink (the
immediately revegetated with the same biomass Gaussian curve, modeled as a negative emission) with
species after harvesting. We assume that the regrowth, the IRF modeling the removal of CO2 by the ocean and/
at the end of the rotation period, captures the same or terrestrial biosphere sinks. In mathematical terms,
amount of CO2 that was released by combustion (i.e., this is a convolution between two functions, based on a
we assume the entire process is C flux neutral). Only conventional and widely used approach (Siegenthaler
one rotation is assumed. CO2 emissions from loss of C & Oeschger, 1978). Then, the atmospheric CO2 con-
pools other than aboveground vegetation, like soil and centration f(t) after a pulse emission can be re-
litter, are not considered at this stage. presented as the sum of earlier emissions g at time
According to the most common practice in biomass t 0 multiplied by the fraction still remaining in the
growth modeling (Swallow et al., 1990; Rossi et al., atmosphere after time tt 0
2009), the rate of biomass growth (or regrowth, in our Z t
case) can be modeled as a normal distribution f ðtÞ ¼ ½C0 dðt0 Þ  gðt0 Þyðt  t0 Þdt0 ; ð6Þ
0
(Gaussian), expressed as atmospheric C uptake in
vegetation as a function of the rotation period of the where C0 is the pulse emission of bio CO2 to the
biomass. This is a probability density function that has atmosphere, d(t 0 ) is the delta function (which is zero
the following analytical form: everywhere except at the origin) g(t 0 ) is the rate of
biomass regrowth which removes the CO2 originally
1 2 2
gðtÞ ¼ pffiffiffiffiffiffiffiffiffiffi eðtmÞ =2s ; ð5Þ released, and y(t) is the IRF from the C cycle climate
2ps 2
model. Equation (6) can be written as follows:
where the parameters m and s (mean and variance) can Z t Z t
be used to represent characteristics of forest growth. It is f ðtÞ ¼ C0 dðt0 Þyðt  t0 Þdt0  gðt0 Þyðt  t0 Þdt0 : ð7Þ
0 0
assumed that the mean occurs in the year with the
maximum C uptake and is taken as half of the Since C0 5 1, we can write
rotation period (m 5 r/2). The variance determines the Z t
width of the distribution, and it is here assumed to be f ðtÞ ¼ yðtÞ  gðt0 Þyðt  t0 Þdt0 : ð8Þ
0
equal to half mean (s 5 m/2).

This equation describes the atmospheric decay

Calculation procedures. The concentration in the
of a pulse of bio CO2 over time. The term represen-
atmosphere of bio CO2 over time can be described by
ting the biomass regrowth, g(t 0 ), is defined in Eqn (5),
while three alternative options are possible for the
IRF y(t):

1. Following the OECD convention, bio CO2 emissions

are removed from the atmosphere by the onsite
biomass growth. If this closed perspective is
adopted, bio CO2 will decay from the air only
because of the biomass regrowth. This means that
there are no contributions from the rest of the C cycle
components, and y(t) 5 1. Since it is totally unphysi-
cal to neglect any CO2 uptake from the oceans or
other sinks, this option is considered here only to
analytically demonstrate the inadequacy of the
Fig. 1 Simplified scheme of the carbon flux neutral system
OECD convention through the inconsistent results
modeled in this paper. (a) Biomass stand at steady state; (b) all
aboveground carbon is harvested and emitted to the atmosphere obtained. This approach will be referred to as the
as CO2. Simultaneously, the same biomass is replanted and starts vegetation IRF (VIRF).
growing by sequestering the CO2 released from combustion; (c) 2. As we have mentioned previously, oceans play a key
the same quantity of carbon originally released is sequestered role in the removal of CO2 from the atmosphere. In
once again in the vegetation at the end of the rotation. this second case, the ocean sink is added to the

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
418 F . C H E R U B I N I et al.

vegetation regrowth sink by considering a proper distribution, while a proper CC model is to be used to
climate model, so giving a specific profile for the predict the atmospheric decay due to ocean uptake. The
atmospheric decay of bio CO2, the ocean and vegeta- IRF of scenario #4 from the ocean model described in
tion IRF (OVIRF). Caldeira & Kasting (1993) is selected. This is a box-
3. As considered in CC models, when a CO2 molecule diffusion ocean model appropriate only on time scales
is released to the atmosphere can be removed by lower than 1000 years, when interaction with sediments
both the ocean and terrestrial biosphere. In this case, and rock cycles is of secondary importance. In this case,
a complete IRF is used and the resulting atmospheric atmospheric CO2 content is stabilized at 550 ppm by
decay is referred to as the full IRF (FIRF). year 2150, the 1990 growth rate in atmospheric CO2
content is 1.66 ppm yr1 and the growth rate at the
In all the cases, the resulting function f(t) is used in Eqn
stabilization date is zero. The IRF resulting from this
(4) to get an index of the relative climate impact of CO2
ocean model has the analytical form of Eqn (1), whose
emissions from biomass combustion
parameters are reported in Table 2 and profile is shown
R TH
AGWPbio CO2 C0 0 aCO2 f ðtÞdt in Fig. 2.
GWPbio ¼ ¼ R TH : ð9Þ If Eqn (1) is included in Eqn (8), we have
AGWPCO2 C0 aCO2 yðtÞdt
0
Xi¼4 Z t
t=ti 1 0 2 2
f ðtÞ ¼A0 þ Ai e  pffiffiffiffiffiffiffiffiffiffi eðt r=2Þ =2ðr=4Þ
0 2ps 2
i¼1
! :
VIRF. In this case, the biomass C cycle is independently Xi¼4
tt0 =ti 0
modeled as a closed system, from combustion to A0 þ Ai e dt
i¼1
removal by vegetation regrowth, which is the only
ð12Þ
sink considered. This option appears consistent with
the convention currently used in bioenergy LCA, where The integral is estimated by numerical approximation.
CO2 emissions from biomass combustion are assumed
to be offset by biomass growth. FIRF. CO2 emissions from biomass combustion are here
In mathematical terms, this means that y(t) 5 1, and considered to be removed from all the possible sinks,
Eqn (8) can be written as the oceans, the terrestrial biosphere and the onsite
Z t biomass regrowth. This integrates bio CO2 emissions
f ðtÞ ¼ 1  gðt0 Þdt0 : ð10Þ into the global C cycle. A complete IRF should be
0
therefore used. Among the existing models, the IPCC
The integral of this function is the cumulative density Fourth Assessment Report selected the IRF derived
function, which is the total C accumulated in the from an updated version of the Bern 2.5CC model
biomass stand along the full rotation. This integral can (Forster et al., 2007). In this paper, the same IRF is
be expressed in terms of the error function erf, so that considered. A detailed description of this model can
Eqn (10) becomes be found elsewhere (Joos et al., 1996, 2001). The analytic
  
1 tm form of this IRF has been shown in Eqn (1), while its
fðtÞ ¼ 1  1 þ erf pffiffiffi ;
2 s 2
Z t ð11Þ
2 x 2
erfðtÞ ¼ pffiffiffi e dx:
p 0
Table 2 Parameters to be used in Eqns (1) (Bern CC model
IRF) and (12) (ocean only IRF)
This allows the calculation of the atmospheric decay
Ocean only Bern 2.5CC
for CO2 emissions from combustion of different
Parameters IRF model IRF
biomass species, according to the rotation period r.
A0 0.297 0.217
OVIRF. This case models the removal of bio CO2 from A1 0.321 0.259
the atmosphere because of two compartments, the A2 0.266 0.338
oceans and the vegetation sink due to biomass A3 0.083 0.186
A4 0.033
regrowth. The rest of the terrestrial biosphere is not
t1 335.8 172.9
considered here as a possible sink. The same approach
t2 18.4 18.51
has been considered in the past to predict the
t3 2.8 1.186
contribution to climate change of CO2 emissions from t4 0.8
a forest fire (Randerson et al., 2006). As in the VIRF case,
the vegetation sink is modeled with the Gaussian IRF, impulse response function.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G L O B A L WA R M I N G P O T E N T I A L O F C O 2 F R O M B I O E N E R G Y 419

Fig. 2 Atmospheric decay of a pulse CO2 emission according to the two different complex carbon cycle climate (CC) models
considered.

parameters are reported in Table 2 and the curve is Results and discussion
shown in Fig. 2. The profile of this function should not
be directly compared with that of the ocean-only IRF Bio CO2 atmospheric decay
presented in the previous section, because they are
based on different conditions and parameters (even In Fig. 3, the three different IRF describing the decay of
though a slowest decay is predictable when oceans bio CO2 emissions from the atmosphere are compared
are the only sink). for selected rotation periods of 1, 10, 20, 50 and 100
In this case, Eqn (8) can be explicitly written as years, as well as when r ! 1 (that is trees are not
follows: replanted). The decay of anthropogenic CO2 according
to the Bern 2.5CC model is also shown for comparison.
! This decay applies in case of deforestation or perma-
X
i¼3
t=ti
fðtÞ ¼ A0 þ Ai e nent terrestrial C losses.
i¼1 For VIRF, OVIRF and FIRF, the longer the biomass
Z !
t
1 X
i¼3
rotation period, the longer is the mean stay of CO2 in
ðt0 r=2Þ2 =2ðr=4Þ2 tt0 =ti
 pffiffiffiffiffiffiffiffiffiffi e A0 þ Ai e dt0 :
0 2ps2 the atmosphere. The effect of the rotation length on the
i¼1
FIRF-based decay is shown in Fig. 4, where the bio CO2
ð13Þ
fraction remaining in the air after a pulse emission is
reported as a function of time and biomass rotation
The integral is estimated by numerical approximation. period. In the long term, all the decays asymptotically
The inclusion of the terrestrial biosphere component tend to zero, since a C flux neutral system is modeled.
among the sinks allows the uptake in the natural As already mentioned, the VIRF curve is based on the
biosphere, but will potentially include a (small) form of OECD convention of a closed cycle for biomass-derived
double counting of the vegetation compartment, since also CO2 (from combustion to uptake in new trees). There-
the onsite vegetation regrowth is considered. However, fore, the resulting atmospheric decay simply represents
this should not be the case because the Bern 2.5CC model the inverse (from an atmospheric point view) of the
only considers the potential CO2 uptake from stimulation sigmoid cumulative C accumulation curve describing
of plant growth by elevated atmospheric CO2 levels biomass regrowth. This is clearly inconsistent with CC
and enhanced nutrient supply, and ‘does not include models: if trees are not replanted bio CO2 would never
formulation for forestry management nor bioenergy decay, as shown in Fig. 4 (VIRF) with r ! 1. Such a
production’ (Strassmann et al., 2008). result is obviously a paradox, and can be seen as an

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Fig. 3 CO2 atmospheric decay following the VIRF, OVIRF and FIRF method for selected rotation periods (r, years). VIRF, vegetation

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

impulse response function; OVIRF, ocean and vegetation impulse response function; FIRF, full impulse response function.
420 F . C H E R U B I N I et al.
 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G L O B A L WA R M I N G P O T E N T I A L O F C O 2 F R O M B I O E N E R G Y 421

in the ocean upper layer will be released back to the

atmosphere at a low rate to compensate the initial
overabsorption (out-gassing). In the long term, the air-
borne fraction of bio CO2 approaches zero.

The GWPbio index and its interpretation

The curves of Fig. 3 are used to get the climate effect of
CO2 emissions from biomass combustion after their
inclusion in Eqn (9). This is a metric relative to the
climate effect of anthropogenic CO2 and based on the
integration up to a defined TH. In Table 3, the GWPbio
index is reported as a function of the biomass rotation
period for the VIRF, OVIRF and FIRF. These results are
shown for the three most common THs (20, 100 and 500
years). The use of this index is identical to the other
Fig. 4 Bio CO2 atmospheric fraction as a function of time and GWP equivalency factors: it is to be multiplied by the
biomass rotation period for the FIRF case. FIRF, full impulse direct CO2 emissions from biomass combustion to get
response function. their relative contribution to global warming in terms of
kg CO2-eq. This allows an estimate of the climate impact
of CO2 flux neutral systems in LCA and other similar
analytical-derived evidence of the physical inaccuracy methodologies. Results are intended to be generally
of the OECD convention. applied to all biomass sources (specified with the rota-
The profile of the curves from OVIRF and FIRF are tion period) from annual row crops to fast growing
similar, since they are both the outcome of a convolu- biomass, tropical, temperate and boreal forests. For
tion operation between the Gaussian and an exponen- annual crops and for short rotation species, the rotation
tial function. As it would have been expected, the period is usually very short, from 1 to 5 years. The
OVIRF decay is slightly longer than the FIRF, where resulting GWPbio is small, since the average lifetime of
the CO2 sequestration is favored by the inclusion of the bio CO2 in the atmosphere in this case is so short that
terrestrial biosphere sink. This can be appreciated by the contribution to global warming is limited. When the
looking at the points where the curves turn to negative rotation period becomes longer, e.g. from fast growing
values: for r 5 100 years, the OVIRF becomes negative at species (r 5 5–20) to tropical (r 5 25–50), temperate
t  71 years, while the FIRF at t  65 years. When (r 5 55–80) and boreal (r 5 80–100) forest, the climate
r ! 1, the curves are equal to the respective function impact increases accordingly. The fact that GWPbio is
y(t) derived from the CC model considered in Eqns (12) larger for longer rotation periods should not be over
and (13), for the OVIRF and FIRF case, respectively. interpreted: it only means that short rotation biomass
Among the three methods, the FIRF appears as the most (e.g. annual crops, short rotation coppice) has less
physically and logically consistent, and the curve for climate impact than long rotation biomass (e.g. forest
r ! 1 coincides with the anthropogenic CO2 decay. wood) per unit of CO2 emitted from the combustion of
At first sight, the presence of negative values in the the biofuel. Before deriving general conclusions, there
atmospheric decay profiles of OVIRF and FIRF may are many other aspects to be considered like efficiency
appear as a contradiction, because the amount of CO2 in in biomass conversion processes, number of rotations,
the atmosphere is lower than the level before the emis- selection of proper time and spatial boundaries, land-
sion. The reason for this is that atmospheric CO2 is use changes and other life cycle implications (like
taken up in different biogeochemical sinks at different material and energy inputs for cultivation, harvesting,
time constants, as mathematically represented by Eqn processing and transport). Land-use changes could also
(1); as implicitly assumed by Eqn (8), the same time include factors such as changes in surface albedo (in
constants are also applied to CO2 uptake in biomass particular at latitudes with seasonal snow cover),
regrowth. Soon after the emission, when the biomass change in soil C content, and changes in fluxes of heat
growth rate is still slow, a significant fraction of the CO2 and humidity between the surface and the atmosphere.
originally released is quickly stored in the ocean upper Misleading conclusions can only be avoided by ac-
layer. The following transport of this C to the deep counting for all climate forcing agents, like GHG emis-
ocean layers is slower, and when the uptake by the sions, removals and, in some cases, substitutions,
onsite biomass regrowth increases, the C initially stored within a life-cycle perspective, preferably using case-

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
422 F . C H E R U B I N I et al.

Table 3 GWPBio index calculated with the three different methods and for three different time horizons: 20, 100 and 500 years

VIRF OVIRF FIRF

GWPbio GWPbio GWPbio GWPbio GWPbio GWPbio GWPbio GWPbio GWPbio

Rotation r (years) TH 5 20 TH 5 100 TH 5 500 TH 5 20 TH 5 100 TH 5 500 TH 5 20 TH 5 100 TH 5 500

1 0.04 0.01 0.00 0.03 0.00 0.00 0.02 0.00 0.00

2 0.08 0.02 0.01 0.05 0.01 0.00 0.04 0.01 0.00
4 0.15 0.04 0.01 0.11 0.02 0.01 0.09 0.02 0.00
6 0.23 0.07 0.02 0.16 0.04 0.01 0.13 0.02 0.00
8 0.30 0.09 0.03 0.21 0.05 0.01 0.18 0.03 0.01
10 0.38 0.11 0.03 0.27 0.06 0.01 0.22 0.04 0.01
12 0.45 0.13 0.04 0.32 0.07 0.01 0.27 0.05 0.01
14 0.53 0.15 0.05 0.38 0.08 0.02 0.32 0.06 0.01
16 0.60 0.17 0.05 0.44 0.09 0.02 0.37 0.06 0.01
18 0.68 0.19 0.06 0.50 0.10 0.02 0.42 0.07 0.01
20 0.75 0.22 0.07 0.55 0.12 0.02 0.47 0.08 0.02
22 0.82 0.24 0.07 0.61 0.13 0.03 0.52 0.09 0.02
24 0.89 0.26 0.08 0.66 0.14 0.03 0.56 0.10 0.02
26 0.95 0.28 0.09 0.71 0.15 0.03 0.61 0.10 0.02
28 1.00 0.30 0.09 0.76 0.16 0.03 0.65 0.11 0.02
30 1.05 0.32 0.10 0.80 0.18 0.04 0.68 0.12 0.02
32 1.09 0.34 0.10 0.83 0.19 0.04 0.71 0.13 0.02
34 1.13 0.37 0.11 0.86 0.20 0.04 0.74 0.14 0.03
36 1.16 0.39 0.12 0.89 0.21 0.04 0.76 0.15 0.03
38 1.19 0.41 0.12 0.91 0.22 0.05 0.79 0.15 0.03
40 1.21 0.43 0.13 0.93 0.24 0.05 0.80 0.16 0.03
42 1.23 0.45 0.14 0.95 0.25 0.05 0.82 0.17 0.03
44 1.25 0.47 0.14 0.97 0.26 0.05 0.83 0.18 0.03
46 1.27 0.49 0.15 0.98 0.27 0.06 0.85 0.19 0.04
48 1.28 0.52 0.16 1.00 0.28 0.06 0.86 0.20 0.04
50 1.30 0.54 0.16 1.01 0.30 0.06 0.87 0.21 0.04
52 1.31 0.56 0.17 1.02 0.31 0.06 0.88 0.21 0.04
54 1.32 0.58 0.18 1.03 0.32 0.07 0.89 0.22 0.04
56 1.33 0.60 0.18 1.03 0.33 0.07 0.89 0.23 0.04
58 1.34 0.62 0.19 1.04 0.34 0.07 0.90 0.24 0.04
60 1.35 0.64 0.20 1.05 0.36 0.07 0.90 0.25 0.05
62 1.35 0.67 0.20 1.05 0.37 0.08 0.91 0.26 0.05
64 1.36 0.69 0.21 1.06 0.38 0.08 0.91 0.27 0.05
66 1.36 0.71 0.22 1.06 0.39 0.08 0.92 0.28 0.05
68 1.37 0.73 0.22 1.07 0.41 0.08 0.92 0.29 0.05
70 1.37 0.75 0.23 1.07 0.42 0.09 0.93 0.30 0.05
72 1.38 0.77 0.24 1.08 0.43 0.09 0.93 0.30 0.06
74 1.38 0.79 0.24 1.08 0.44 0.09 0.93 0.31 0.06
76 1.39 0.82 0.25 1.08 0.46 0.09 0.94 0.32 0.06
78 1.39 0.84 0.25 1.09 0.47 0.10 0.94 0.33 0.06
80 1.39 0.86 0.26 1.09 0.48 0.10 0.94 0.34 0.06
82 1.40 0.88 0.27 1.09 0.49 0.10 0.94 0.35 0.06
84 1.40 0.90 0.27 1.09 0.51 0.10 0.95 0.36 0.06
86 1.40 0.92 0.28 1.10 0.52 0.11 0.95 0.37 0.07
88 1.40 0.94 0.29 1.10 0.53 0.11 0.95 0.38 0.07
90 1.41 0.96 0.29 1.10 0.54 0.11 0.95 0.39 0.07
92 1.41 0.98 0.30 1.10 0.55 0.11 0.95 0.39 0.07
94 1.41 0.99 0.31 1.10 0.56 0.12 0.95 0.40 0.07
96 1.41 1.01 0.31 1.10 0.58 0.12 0.96 0.41 0.07
98 1.41 1.03 0.32 1.11 0.59 0.12 0.96 0.42 0.08
100 1.42 1.05 0.33 1.11 0.60 0.12 0.96 0.43 0.08

GWP, global warming potential; VIRF, vegetation impulse response function; OVIRF, ocean and vegetation impulse response
function; FIRF, full impulse response function; TH, time horizon.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G L O B A L WA R M I N G P O T E N T I A L O F C O 2 F R O M B I O E N E R G Y 423

specific parameters. Therefore, Table 3 does not expli- Concerning the OVIRF-based GWPbio, values slightly
citly mean that one biomass source is better than others: higher than one can be obtained for TH 5 20 years with
a lower value of the index does not necessarily reflect a rotation periods longer than 50 years. The reason can be
lower climate impact of the whole bioenergy system. seen in the corresponding graph in Fig. 3: in the first
Figure 5 shows the value of the GWPbio index as a years soon after the emission, the OVIRF with r larger
function of the biomass rotation period for the three than 50 years has a slower decay than the decay from
different cases and for the three selected THs. The the Bern 2.5CC model (used as reference in the metric),
curves have an exponential trend to a maximum, which thus affecting the GWPbio for TH 5 20 years. By con-
has the same value for each method and can be better trast, the FIRF-based GWPbio index ranges from 0 to 1,
appreciated for TH 5 20 years. since the same IRF is used as y(t) in Eqn (8) and
The GWPbio is bigger for shorter TH, because this reference in the metric. In this case, the climate impact
index considers the area below the decay curve of bio derived from biomass combustion and subsequently
CO2 relative to that of anthropogenic CO2. The latter has reabsorbed in the ocean and terrestrial sinks can never
a fast decay in the first years soon after the emission and be higher than the impact of the same quantity released
then a slow asymptotic trend towards the ocean/atmo- by fossil fuel combustion or deforestation. Owing to the
sphere equilibrium, while bio CO2 decay tends to zero. consideration of all the C cycles with terrestrial and
The fact that GWPbio are higher for TH 5 20 years rather ocean sinks, the FIRF method has the most consistent
than for TH 5 100 or 500 years confirms that bioenergy results which should be used in bioenergy LCA studies
is a climate change mitigation strategy particularly to estimate the climate impact of CO2 emissions from
effective for long-term targets. biomass combustion.
The VIRF-based GWPbio is larger than one for some
circumstances. This is a direct consequence of the
Conclusions and next outlook
OECD convention on which the VIRF decay is based:
the exclusion of the ocean and terrestrial biosphere The work performed in this paper brings a new con-
uptake other than the onsite regrowth can make the tribution to the rising discussion on the proper account-
climate impact of bio CO2 approximately 1.5 times ing of CO2 emissions from biomass combustion in
higher than that of anthropogenic CO2. This result is bioenergy systems. Even if perceived as urgent, a
further evidence about the shortcomings of the existing methodology able to quantify the effective climate im-
assumption on the closed cycle for biomass-derived pact of biomass-derived CO2 emissions with unit-based
CO2 emissions. indicators was not elaborated by LCA practitioners. The

Fig. 5 GWPbio for TH equal to 20, 100 and 500 years as a function of the biomass rotation period. GWP, global warming potential; TH,
time horizon.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
424 F . C H E R U B I N I et al.

most important contributions of this work are the the temporary C loss needs to be established in order to
formulation of IRF for the atmospheric decay of reach the intended climate policy targets.
CO2 emissions from biomass combustion and the
adoption of an index, the GWPbio, to estimate
their climate impact. Three methods were formulated,
the VIRF, based on the closed cycle of bio CO2, the Acknowledgements
OVIRF, which includes the ocean uptake, and the FIRF, The work was funded by the Norwegian research council
which considers the full C cycle with ocean and terres- through the ‘Bio-energy Innovation Centre – CenBio’ (Cherubini,
trial sinks. The FIRF-based GWPbio is the most reliable Strmman and Hertwich), ‘Transport and Environment – Mea-
and accurate option, given its complete consideration sures and Policies’ (Peters and Berntsen) and ‘Terrestrial C
sequestration potential in Norway under present and future
of all the C components and biogeochemical sinks.
climate’ (Peters). We thank Ottar Michelsen, Ryan Matthew
The GWP equivalency factor currently used for CO2 Bright and Geoffrey Guest (NTNU) for the fruitful discussions
emissions from biomass combustion in LCA should and their critical review of this work.
be revised: rather than a value of 0 (when the
OECD convention is strictly followed) that underesti-
mates the climate impact of the bioenergy system, or References
1 (as performed by studies considering the initial Archer D, Kheshgi H, Maier-Reimer E (1998) Dynamics of fossil
CO2 sequestration during plant growth or by studies fuel neutralization by Marine CaCO3. Global Biogeochemical
based on forest C dynamics) that overestimate the Cycles, 12, 259–276.
climate impact of bio CO2, this work proposes a figure Blanke B, Delecluse P (1993) Variability of the tropical Atlantic
between 0 and 1, depending on the rotation period Ocean simulated by a general circulation model with two
of the biomass harvested. This is a first step towards different mixed-layer physics. Journal of Physical Oceanography,
the overcoming of the inadequacy of CO2 accounting 23, 1363–1388.
in LCA and the development of an accurate and Börjesson P, Gustavsson L (2000) Greenhouse gas balances in
standardized procedure. building construction: wood versus concrete from life-cycle
and forest land-use perspectives. Energy Policy, 28, 575–588.
This work acts as starting point for future research
Caldeira K, Kasting JF (1993) Insensitivity of global warming
activities and investigation of specific case studies. In
potentials to carbon dioxide emission scenarios. Nature, 366,
order to keep the focus of the paper on the development 251–253.
of a methodology to quantify the climate effect of bio Carpentieri M, Corti A, Lombardi L (2005) Life cycle assessment
CO2 emissions, a schematic case based on one single (LCA) of an integrated biomass gasification combined
rotation and with well defined initial conditions and cycle (IBGCC) with CO2 removal. Energy Conversion and Man-
parameters has been selected. The theoretical basis and agement, 46, 1790–1808.
calculations developed here can be expanded to model Cherubini F, Strmman AH (2011) Life cycle assessment of
more specific case studies, with customized biomass bioenergy systems: state of the art and future challenges.
growth curves, multiple rotations, particular manage- Bioresource Technology, 102, 437–451.
ment strategies, different starting conditions (e.g. affor- Cramer W, Bondeau A, Woodward FI et al. (2001) Global re-
sponse of terrestrial ecosystem structure and function to CO2
estation rather than deforestation) or other specific
and climate change: results from six dynamic global vegeta-
factors. These outcomes can also be integrated within
tion models. Global Change Biology, 7, 357–373.
software tools modeling the climate effects of biomass Enting IG, Wigley TML, Heimann M (2001). Future emissions and
production on terrestrial C pools and the environmental concentrations of carbon dioxide: key ocean/atmosphere/land ana-
impact of bioenergy systems. lyses. Technical paper no. 31, CSIRO Division of Atmospheric
Besides LCA-based applications, of particular interest Research.
is the possibility to include the outcomes of this work in Forster P, Ramaswamy V, Artaxo P et al. (2007) Changes in
national GHG accounting mechanisms, so to revise the atmospheric constituents and in radiative forcing. In: Climate
OECD convention presented at the beginning of this Change 2007: The Physical Science Basis. Contribution of Working
paper. The FIRF for bio CO2 is suitable to be combined Group I to the Fourth Assessment Report of the Intergovernmental
with the existing accounting of C stock changes to Panel on Climate Change (ed. Solomon S), Intergovernmental
Panel on Climate Change, Cambridge, UK.
develop a robust and thorough C accounting frame-
Friedlingstein P, Fung I, Holland E, John J, Brasseur G, Erickson
work. This application may have significant impacts on
D, Schimel D (1995) On the contribution of CO2 fertilization to
national GHG reporting for bioenergy production, and the missing biospheric sink. Global Biogeochemical Cycles, 9,
consequently needs to be investigated further to explore 541–556.
advantages and disadvantages. Implications at a policy Friedlingstein P, Müller JF, Brasseur GP (1994) Sensitivity of the
decision level can be also relevant: new strategies taking terrestrial biosphere to climatic changes: impact on the carbon
into account the climate impact of CO2 emissions from cycle. Environmental Pollution, 83, 143–147.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
G L O B A L WA R M I N G P O T E N T I A L O F C O 2 F R O M B I O E N E R G Y 425

Fuglestvedt JS, Berntsen TK, Godal O, Sausen R, Shine KP, of Energy Resources (ed. Walker T), Contributors: Cardellichio,
Skodvin T (2003) Metrics of climate change: assessing radia- P, Colnes, A, Gunn, J, Kittler, B, Perschel, R, Recchia, C, Saah,
tive forcing and emission indices. Climate Change, 58, 267–331. D, and Walker, T. Manomet Center for Conservation Sciences,
Gerber S, Joos F, Prentice C (2004) Sensitivity of a dynamic global Brunswick, ME.
vegetation model to climate and atmospheric CO2. Global Marland G (2010) Accounting for carbon dioxide emissions from
Change Biology, 10, 1223–1239. bioenergy systems. Journal of Industrial Ecology, 14, 866–869.
Harmon ME, Ferrell WK, Franklin JF (1990) Effects on carbon Marland G, Schlamadinger B (1995) Biomass fuels and forest-
storage of conversion of old-growth forests to young forests. management strategies: how do we calculate the greenhouse-
Science, 247, 699–702. gas emissions benefits? Energy, 20, 1131–1140.
Houghton JT, Jenkins GJ, Ephraums JJ (1990) Climate Change. The Masera O, Garza-Caligaris JF, Kanninen M et al. (2003) Modelling
IPCC Scientific Assessment, New York, USA. carbon sequestration in afforestation, agroforestry and forest
Huo H, Wang M, Bloyd C, Putsche V (2008) Life-cycle assess- management projects: the CO2FIX V.2 approach. Ecological
ment of energy use and greenhouse gas emissions of soybean- Modelling, 164, 177–199.
derived biodiesel and renewable fuels. Environmental Science McGuire AD, Sitch S, Clein JS et al. (2001) Carbon balance of the
and Technology, 43, 750–756. terrestrial biosphere in the Twentieth Century: analyses of
IPCC (2006) Guidelines for National Greenhouse Gas Inventories, CO2, climate and land use effects with four process-based
Volume 4 – Agriculture, Forestry and other Land Use. Intergovern- ecosystem models. Global Biogeochemical Cycles, 15, 183–206.
mental Panel on Climate Change, Hayama, Japan. McKechnie J, Colombo S, Chen J, Mabee W, MacLean HL (2010)
Johnson E (2009) Goodbye to carbon neutral: getting biomass Forest bioenergy or forest carbon? Assessing trade-offs in
footprints right. Environmental Impact Assessment Review, 29, greenhouse gas mitigation with wood-based fuels. Environ-
165–168. mental Science and Technology, 45, 789–795.
Joos F (2003) The anthropogenic perturbation of atmospheric Möllersten K, Grönkvist S (2007) All CO2 is equal in the atmo-
CO2 and the climate system. In: Recent research developments in sphere – a comment on CDM GHG accounting standards for
biotechnology and bioengineering. Special Issue: Biotechnology and methane recovery and oxidation projects. Energy Policy, 35,
bioengineering of CO2 fixation (eds Bhattacharya SK, Mal TK, 3675–3680.
Chakrabarti S), Trivandrum, India, 183pp. Moore B, Braswell BH (1994) The lifetime of excess atmospheric
Joos F, Bruno M (1996) Pulse response functions are cost-efficient carbon dioxide. Global Biogeochemical Cycles, 8, 23–28.
tools to model the link between carbon emissions, atmospheric OECD (1991) Estimation of greenhouse gas emissions and sinks. Final
CO2 and global warming. Physics and Chemistry of the Earth, 21, report from the OECD experts meeting, 18–21 February, 1991,
471–476. OECD, Paris.
Joos F, Bruno M, Fink R, Stocker TF, Siegenthaler U, Le Quéré C, Oeschger H, Siegenthaler U, Schotterer U, Gugelmann A (1975)
Sarmiento JL (1996) An efficient and accurate representation of A box diffusion model to study the carbon dioxide exchange
complex oceanic and biospheric models of anthropogenic in nature. Tellus, 27, 168–192.
carbon uptake. Tellus, 48B, 397–417. Petersen Raymer AK (2006) A comparison of avoided green-
Joos F, Prentice IC, Sitch S et al. (2001) Global warming feedbacks house gas emissions when using different kinds of wood
on terrestrial carbon uptake under the Intergovernmental energy. Biomass and Bioenergy, 30, 605–617.
Panel on Climate Change (IPCC) emission scenarios. Global Prentice C, Heimann M, Sitch S (2000) The carbon balance of the
Biogeochemical Cycles, 15, 891–907. terrestrial biosphere: ecosystem models and atmospheric ob-
Kim S, Dale BE (2008) Life cycle assessment of fuel ethanol servations. Ecological Applications, 10, 1553–1573.
derived from corn grain via dry milling. Bioresource Technology, Quirin M, Gartner SO, Pehnt M, Reinhardt GA (2004). CO2
99, 5250–5260. mitigation through biofuels in the transport sector: status and
Kurz WA, Dymond CC, White TM et al. (2009) CBM-CFS3: a perspectives. In: I. f. E. a. E. R. (IFEU) (Ed.), Main Report,
model of carbon-dynamics in forestry and land-use change Heidelberg, Germany. 57 pp.
implementing IPCC standards. Ecological Modelling, 220, Rabl A, Benoist A, Dron D, Peuportier B, Spadaro JV, Zoughaib A
480–504. (2007) How to account for CO2 emissions from biomass in an
Lashof DA, Ahuja DR (1990) Relative contributions of LCA. International Journal of Life Cycle Assessment, 12, 281.
greenhouse gas emissions to global warming. Nature, 344, Randerson JT, Liu H, Flanner MG et al. (2006) The impact of
529–531. boreal forest fire on climate warming. Science, 314, 1130–1132.
Luo L, van der Voet E, Huppes G, Udo de Haes H (2009) Reijnders L, Huijbregts MAJ (2008) Biogenic greenhouse gas
Allocation issues in LCA methodology: a case study of corn emissions linked to the life cycles of biodiesel derived from
stover-based fuel ethanol. The International Journal of Life Cycle European rapeseed and Brazilian soybeans. Journal of Cleaner
Assessment, 14, 529–539. Production, 16, 1943–1948.
Maier-Reimer E, Hasselmann K (1987) Transport and storage of Rodhe H (1990) A comparison of the contribution of various
CO2 in the ocean – an inorganic ocean-circulation Carbon gases to the greenhouse effect. Science, 248, 1217–1219.
Cycle Model. Climate Dynamics, 2, 63–90. Rossi S, Tremblay M-J, Morin H, Savard G (2009) Growth and
Manomet (2010) Biomass Sustainability and Carbon Policy productivity of black spruce in even- and uneven-aged stands
Study. In: Massachusetts Biomass Sustainability and Carbon Policy at the limit of the closed boreal forest. Forest Ecology and
Study: Report to the Commonwealth of Massachusetts Department Management, 258, 2153–2161.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426

 17571707, 2011, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1757-1707.2011.01102.x by National Medical Library The Director, Wiley Online Library on [03/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
426 F . C H E R U B I N I et al.

Schelhaas MJ, van Esch PW, Groen TA (2004). CO2FIX V 3.1 – a Siegenthaler U, Joos F (1992) Use of a simple model for studying
modelling framework for quantifying carbon sequestration in oceanic tracer distributions and the global carbon cycle. Tellus,
forest ecosystems. ALTERRA Report 1068, Wageningen, the 44B, 186–207.
Netherlands. Siegenthaler U, Oeschger H (1978) Predicting future atmospheric
Schimel D, Alves D, Enting I et al. (1996) Radiative forcing of carbon dioxide levels. Science, 199, 388–395.
climate change. In: Climate Change 1995. The Science of Climate Strassmann KM, Joos F, Fischer G (2008) Simulating effects of
Change (eds Houghton JT, Meira Filho LG, Callander BA land use changes on carbon fluxes: past contributions to
et al.), pp. 65–131. Cambridge University Press, Cambridge. atmospheric CO2 increases and future commitments due to
Schlamadinger B, Marland G (1996a) The role of forest and losses of terrestrial sink capacity. Tellus B, 60, 583–603.
bioenergy strategies in the global carbon cycle. Biomass and Swallow SK, Parks PJ, Wear DN (1990) Policy-relevant noncon-
Bioenergy, 10, 275–300. vexities in the production of multiple forest benefits. Journal of
Schlamadinger B, Marland G (1996b) Full fuel cycle carbon Environmental Economics and Management, 19, 264–280.
balances of bioenergy and forestry options. Fuel and Energy Tubiello FN, Oppenheimer M (1995) Impulse-response functions
Abstracts, 37, 197–197. and anthropogenic CO2. Geophysical Research Letters, 22,
Searchinger T (2010) Biofuels and the need for additional carbon. 413–416.
Environmental Research Letters, 5, 024007, doi: 10.1088/1748- UNFCCC (2003) Estimating, reporting, and accounting of harvested
9326/5/2/024007. wood products. Technical paper FCCC/TP2003/7.
Searchinger TD, Hamburg SP, Melillo J et al. (2009) Fixing a van der Voet E, Lifset RJ, Luo L (2010) Life-cycle assessment of
critical climate accounting error. Science, 326, 527–528. biofuels, convergence and divergence. Biofuels, 1, 435–449.
Searcy E, Flynn PC (2008) Processing of straw/corn stover: Wanninkhof R (1992) Relationship between wind speed and gas
comparison of life cycle emissions. International Journal of Green exchange over the ocean. Journal of Geophysical Research –
Energy, 5, 423–437. Oceans, 97, 7373–7382.
Shine K, Fuglestvedt J, Hailemariam K, Stuber N (2005) Alter- Werner F, Althaus H-J, Künniger T, Richter K (2003) Life Cycle
natives to the global warming potential for comparing climate Inventories of Wood as Fuel and Construction Material. Final
impacts of emissions of greenhouse gases. Climatic Change, 68, report ecoinvent 2000 No. 9, EMPA Dübendorf, Swiss Centre
281–302. for Life Cycle Inventories, Dübendorf, CH.

r 2011 Blackwell Publishing Ltd, GCB Bioenergy, 3, 413–426