Diverting residual biomass to energy use: Quantifying the global warming potential of biogenic CO2 (GWPbCO2)

To calculate the global warming potential of biogenic carbon dioxide emissions (GWPbCO2) associated with diverting residual biomass to bioenergy use, the decay of annual biogenic carbon pulses into the atmosphere over 100 years was compared between biomass use for energy and its business‐as‐usual decomposition in agricultural, forestry, or landfill sites. Bioenergy use increased atmospheric CO2 load in all cases, resulting in a 100GWPbCO2 (units of g CO2e/g biomass CO2 released) of 0.003 for the fast‐decomposing agricultural residues to 0.029 for the slow, 0.084–0.625 for forest residues, and 0.368–0.975 for landfill lignocellulosic biomass. In comparison, carbon emissions from fossil fuels have a 100GWP of 1.0 g (CO2e/g fossil CO2). The fast decomposition rate and the corresponding low 100GWPbCO2 values of agricultural residues make them a more climate‐friendly feedstock for bioenergy production relative to forest residues and landfill lignocellulosic biomass. This study shows that CO2 released from the combustion of bioenergy or biofuels made from residual biomass has a greenhouse gas footprint that should be considered in assessing climate impacts.

competition between food and fuel as well as adverse impacts on biodiversity (Patel et al., 2016;Reid et al., 2020). Moreover, plant residues are readily available, and do not require additional inputs to be generated; therefore, they can be sourced at low or negative cost (Gustavsson et al., 2015;Repo et al., 2012). In addition, plant residues are abundant in Canada, especially in the prairie provinces for agricultural residues and in British Columbia and Quebec for forest residues (Adetona & Layzell, 2019;Wood & Layzell, 2003). Diverting these residues to bioenergy production may increase the area of cropland available for cultivation and reduce the risk of forest wildfire caused by the spontaneous combustion of residues (Kurz et al., 2013;O'Connell, 1997).
Lignocellulosic materials (e.g. wood, cardboard, paper, straw) are the predominant form of residual biomass from agriculture, forestry, and landfills and are increasingly seen as a source of renewable energy in the transition pathway to net-zero energy systems (Sandalow et al., 2021). Furthermore, some agricultural and landfill residues are highly metabolizable biomass waste streams (e.g. animal by-products, food) that are readily converted into CH 4 , the primary component in natural gas and a potent GHG (Weiland, 2010).
For residual biomass that is diverted to energy use, the common assumption is that the biogenic CO 2 released to the atmosphere is 'closing the carbon (C) cycle' and therefore, it does not contribute to GHG emissions (Ragauskas, 2006;Weldemichael & Assefa, 2016). This may not be the case as the decision to divert residual biomass to bioenergy could also reduce biosphere C stocks (Law et al., 2018;Pingoud et al., 2016).
Under the United Nations Framework Convention on Climate Change (UNFCCC), nations are expected to monitor and report on biosphere C stock changes in agricultural soils and managed forest lands (Iversen et al., 2014). It is crucial to develop tools to quantify anthropogenic impacts on biosphere C stocks, especially because of the important roles the biological systems will play as a source of negative emission technologies required in meeting the net-zero emission target (Kemper, 2015;Sandalow et al., 2021;United Nations, 2020). Such a tool should account for the biogenic CO 2 that is emitted into the atmosphere when biomass is diverted from natural decomposition into bioenergy (Shapiro-Bengtsen et al., 2022).
Like fossil fuels, the combustion of biomass for energy results in instantaneous release (i.e., pulse) of CO 2 into the atmosphere (Cherubini et al., 2011;Sterman et al., 2022). Although a fraction of CO 2 released into the atmosphere is removed by natural sinks, namely, ocean and terrestrial ecosystems (Bright et al., 2012;Cherubini et al., 2012), there is a time lag between when CO 2 is released from biomass combustion and when the CO 2 is removed by the sinks (Haberl, 2013;Searchinger et al., 2009). Some of the emitted CO 2 may remain in the atmosphere for up to 1000 years due to the equilibrium that occurs following the interaction between the ocean and the atmosphere (Archer et al., 1998;Cherubini et al., 2011). While the CO 2 remains in the atmosphere, it has some warming impacts which should be considered in life cycle assessment (LCA) studies of bioenergy production (Cherubini et al., 2011;Gustavsson et al., 2015).
In 2011, the term 'global warming potential of biogenic CO 2 (GWP bCO2 )' was coined to describe the radiative forcing of biomass-derived CO 2 emissions relative to fossil-derived CO 2 emissions (Cherubini et al., 2011). This metric allows an impartial assessment when comparing the potential effects of combusting biomass for energy relative to the effect of CO 2 emissions from the combusting fossil fuels for energy (Cherubini et al., 2011;van Kooten et al., 2021). Based on a 100year time horizon, the GWP bCO2 ranges from −1.00  to 1.79 g CO 2 e/g b CO 2 (Holtsmark, 2015), depending on factors such as the type of biomass, growth condition, as well as rotation and biomass C storage period (Cherubini et al., 2011;Guest et al., 2013a). The rotation period is the time it takes a plant to grow before it is harvested for bioenergy and the storage period is how long C is stored in plant biomass before it is combusted for energy (Cherubini et al., 2011;Guest et al., 2013a).
The GWP bCO2 value of 0.00 g CO 2 e/g b CO 2 means that there is no climate impact of biogenic CO 2 emissions Guest et al., 2013a). Values of GWP bCO2 that are below 0.00 have a cooling effect on the climate, whereas values between 0.00 and 0.99 imply that the emissions have a warming impact. However, this effect is not as high as the global warming potential for CO 2 from fossil fuel combustion (GWP fCO2 ) of 1.00 g CO 2 e/g f CO 2 (Cherubini et al., 2011;Guest et al., 2013a;. Values for GWP greater than 1.00 have a worse impact relative to fossil CO 2 and are commonly associated with methane (CH 4 ) emissions (25 g CO 2 e/g CH 4 , Forster et al., 2007) or nitrous oxide (N 2 O) emissions (298 g CO 2 e/g N 2 O, Forster et al., 2007). Biomass CO 2 emissions can also be higher than 1.0 g CO 2 e/g b CO 2 when forests or grasslands are replaced with fast-growing bioenergy crops that provide little C storage (Cherubini et al., 2011;Guest et al., 2013a).
Most studies that have used the GWP bCO2 metric to assess the climate impact of combusting biomass focused on plants that are dedicated (grown or harvested) for bioenergy production (Cherubini et al., 2011Guest et al., 2013a). Thus, there is little information on the impact of the decision to divert residual biomass from plants that are grown for other purposes such as food and fibre generation into bioenergy production.
Although Guest et al. (2013b) determined the effect of residual biomass decomposition on GWP bCO2, they focused on residues generated from plants that are grown for bioenergy production. Repo et al. (2012) and Gustavsson et al. (2015) examined the climate impact of removing residues from the forest for bioenergy production, but they did not calculate the GWP bCO2 associated with the decision.
To assess climate impacts associated with diverting residual lignocellulosic biomass to bioenergy, it is important to first understand the destiny of the biomass if it is not diverted to bioenergy. While biomass with high decomposition rates such as agricultural residues releases C into the atmosphere within a few years, biomass with slow decomposition rates such as woody forest residues tends to keep the C in the biosphere for decades (Beyaert & Voroney, 2011;Micales & Skog, 1997;Russell et al., 2014).
The objective of this study was to develop a method to calculate GWP bCO2 associated with the decision to divert residual biomass from agricultural, forestry, and landfill sites to bioenergy production. This study compared, over both a 20-year and 100-year period, the impact on atmospheric CO 2 load of combusting one megagram (Mg) of biomass C versus letting it decompose.

| Residual biomass decomposition/ combustion and carbon transfer to the atmosphere
Data from the scientific literature (Beyaert & Voroney, 2011;Micales & Skog, 1997;Russell et al., 2014) were used to quantify the rate at which one Mg of residual biomass C from agricultural, forestry, and landfill sources is decomposed and converted to atmospheric CO 2 over a 100-year period. A detailed description of the quantification is shown in the Supplemental Material.
While the decomposition of biomass under anaerobic conditions can also produce CH 4 , the current study is focused solely on the calculation of the Global Warming Potential associated with the CO 2 emissions (GWP bCO2 ) from residual biomass. GWP CH4 values for CH 4 emissions are well documented (ECCC, 2020a(ECCC, , 2020bForster et al., 2007) and not part of the current study.
CO 2 released from residual biomass decomposition is highly variable both between and within biomass types. This variability can be attributed to the type of biomass and environmental factors such as temperature and moisture (Chapin et al., 2012;Fraver et al., 2013;Harmon et al., 2009). To address variability within each type of residual biomass, the lowest and the highest biomass decomposition profiles were chosen for each biomass type to generate six 'business-as-usual (BAU)' scenarios, namely agricultural residues slow (A-slow) and fast decomposition (A-fast), forestry residues slow (F-slow) and fast (F-fast) decomposition as well as landfill lignocellulosic biomass slow (F-slow) and fast (F-fast) decomposition. For each scenario, the annual pulse of C to the atmosphere ( pulse C[t]) in Mg C per year was estimated over 101 years (i.e. t = 0-100 years) where year 0 was the year in which a decision was made to either leave the residual biomass to decompose or to divert it into energy (heat or fuels) for human use.
In the bioenergy scenario, one Mg of the diverted biomass C was assumed to be fully converted to atmospheric CO 2 in year zero (i.e. pulse C[t = 0] = 1 Mg C) while in the decomposition scenarios, a fraction of the one Mg C was assigned to each year from t = 0 to t = 100 (based on published evidence of decomposition rates) such that the sum of all pulses was less than or equal to one Mg C (Equation 1).

| Atmospheric decay of CO 2 pulses to the atmosphere
In the year that each pulse was released into the atmosphere, no atmospheric decay was assumed to occur. However, for all subsequent years, the pulse to the atmosphere was assumed to decay in a pattern consistent with the Bern 2.5 C cycle-climate (Bern 2.5 CC) model (Bright et al., 2012;Cherubini et al., 2011;Joos et al., 2001) as defined by Equation (2).
where f C n is the fraction of a pulse of C to the atmosphere that remains in the atmosphere in year 'n' and 'n' ranges from 1 to 100 years after the occurrence of the pulse at year t. Constants in this equation include A 0 (value of 0.2173) as the fraction of CO 2 that will not decay but remain in the atmosphere because of the equilibrium that is reached between the ocean and the atmosphere (Archer et al., 1998;Bright et al., 2012). The other constants are A 1 (value of 0.259), A 2 (value of 0.338), and A 3 (value of 0.186), representing the relative capacity of the terrestrial ecosystem and the ocean to remove atmospheric CO 2 (Bright et al., 2012). This capacity is based on the corresponding relaxation timescales for the sink, denoted as β 1 (value of 172.9), β 2 (value of 18.51), and β 3 (value of 1.186) (Bright et al., 2012).
Therefore, the C remaining in the atmosphere 'n' years after a pulse that occurred in year t (AC[t] n ) can be calculated using Equation (3). where C[t]pulse is the annual pulse of C (Mg C per year) to the atmosphere and f C n is the fraction of a pulse of C to the atmosphere that remains in the atmosphere in year 'n'.
For each modelled scenario, and each pulse of C to the atmosphere (C[t]pulse) associated with that scenario, AC[t] n was calculated for values from n = 0 to t + n = 100 years. The results were laid out in a matrix created using Microsoft Excel (Office 365, version 2019) (Figure 1), where each row in the matrix represents the annual pulse to the atmosphere and the amount of that pulse that remains in the atmosphere in the future years up to t + n = 100.
To calculate the atmospheric load of CO 2 each year over 100 years (AC load t+n ), the columns in the matrix were added as shown in Figure 1 and summarized in Equation (4). where AC load t+n is the atmospheric load of CO 2 associated with a given scenario in year t + n and AC[t] n is the C remaining in the atmosphere 'n' years after a pulse that occurred in year t.
Therefore, AC load t+n is the sum of the C pulse to the atmosphere in year 't' and the fraction of all previous pulses to the atmosphere that remains after n years.

| Global warming of biogenic CO 2
For each scenario, the values for AC load t+n were plotted against time (0-100 years) and used to calculate the atmospheric CO 2 loads over both a 20-year (ACscXload20) and 100-year ( 100 AC load scX ) period. This was done by summing the total C remaining for each year over the period to give values for the Mg C-years (Equations 5 and 6).
where 'scX' refers to the seven identified scenarios. The atmospheric CO 2 load for one-Mg b CO 2 pulse in the bioenergy scenario would be identical to the atmospheric CO 2 load for one-Mg b CO 2 pulse from a fossil fuel source (Haberl et al., 2012;Mollersten & Gronkvist, 2007) even though the bio-based CO 2 is not typically counted as a GHG. While GWP bCO2 (g CO 2 e/g b CO 2 ) is assumed to be zero compared to a value of 1.0 for the global warming potential of fossil CO 2 (GWP fCO2 ), the decision to divert residual biomass C to bioenergy use (rather than leaving it to more slowly decompose) would be expected to have a higher climate change impact and should have a GWP bCO2o value that is greater than 0 (Cherubini et al., 2011). To quantify this value, the following equations were used: where 20 AC load Bioenergy Scenario and 20 AC load BAU Scenario are the atmospheric loads for a 20-year time horizon ( 20 AC) for the bioenergy and BAU biomass decomposition, respectively.
(3) ) and the fraction of carbon that is removed in the atmosphere in year 'n' pC n . The matrix was also used to calculate the total carbon in the atmosphere in year t AC Total t , which was determined as the sum of the current year's pulse C where 100 AC load Bioenergy Scenario and 100 AC load BAU Scenario are the atmospheric loads for a 100-year time horizon ( 100 AC) for the bioenergy and BAU biomass decomposition, respectively.

| RESULTS
Differences were observed in the amount of C transferred to the atmosphere per year for every Mg of biomass C that was combusted (Figure 2) or left to decompose (Figure 3). The differences led to variations in the atmospheric load of CO 2 (i.e. the fraction of CO 2 remaining after atmospheric decay) over a 20-year or 100-year time horizon as well as their corresponding GWP bCO2 values.

| Carbon transfer to the atmosphere from residual biomass combustion and decomposition
When residual biomass was combusted for energy at year 't', there was an instantaneous release of C (as a single pulse of CO 2 ) from the biosphere (Figure 2, orange-shaded area) to the atmosphere (Figure 2, blue-shaded area).
However, for the six BAU decomposition scenarios, the amount of C released from the biosphere to the atmosphere was governed by the rate of decomposition of the residual biomass, and it varied with biomass type (Figure 3). Of the three examined biomass types, the agricultural residues had the fastest rate of decomposition, releasing up to 90% of the biomass C to the atmosphere within the first year of decomposition (Figure 3a).
Following the two-component exponential decomposition model, there were differences in the decomposition profiles of agricultural residues, especially during the first 2 years of decomposition (Figure 3a). While the fastdecomposing residues (e.g. soybean residues) emitted up to 90% of the biomass C to the atmosphere within the first year of decomposition, the slow-decomposing residues (e.g. wheat residues) emitted about 60% of its C. Over 85% of C in agricultural residues was emitted within 10 years of decomposition (Figure 3a).
Decomposition profiles also varied between fast (e.g. hardwood) and slow-decomposing (e.g. softwood) forestry residues (Figure 3b). The fast-decomposing forestry residues emitted over 95% of C to the atmosphere within 40 years of decomposition, whereas only 30% of the C was emitted for slow-decomposing residues within the same year. About 40% of the C in the slowdecomposing forestry biomass remained after 100 years of decomposition.

F I G U R E 2
The fate of one megagram C as CO 2 (orange area) released to the atmosphere through combustion. At t = 0, the CO 2 enters the atmospheric pool (blue-shaded area) and then a portion is transferred to the biosphere carbon pool (green-shaded area) over 100 years. The atmospheric CO 2 load in megagram-carbon years (Mg C-years, i.e. blue-shaded area) was calculated for the 0-to 20-year and 0-to 100-year period and the values are shown on the chart.  For the landfill scenario, up to 97% of C embedded in the slow-decomposing biomass remained in the biosphere even after 100 years of decomposition (Figure 3c).

| Atmospheric CO 2 load from residual biomass combustion
Approximately 40% and 60% of the C was removed from the atmosphere at 20 and 100 years, respectively, returning the removed C to the biosphere (Figure 2, green shaded area). The area under the atmospheric removal curve (Figure 2, blue-shaded area) signifies the atmospheric CO 2 load (i.e. the cumulative amount remaining in the atmosphere CO 2 from the pulses released at a certain period). The CO 2 loads were 14.4 and 48.5 Mg C-years for 20-year ( 20 A) and 100-year ( 100 A) time horizon, respectively ( Figure 2).

| Atmospheric CO 2 load from residual biomass decomposition
For the BAU scenarios, a proportion of one Mg of C was released into the atmosphere yearly as a sequence of small pulses (Figures 3 and 4). In each year, this proportion was governed by the decomposition rates of the residual biomass, which varied with biomass type (Figure 3). In the case of the A-fast BAU scenario, the decomposing biomass released C into the atmosphere like a single pulse of C emitted to the atmosphere during the bioenergy scenario since over 90% of the residual C was emitted during the first year of decomposition ( Figure 4a). For this BAU scenario, the atmospheric loads were 14.2 and 48.4 Mg Cyears over a 20-and 100-year time horizon, respectively. For the A-slow BAU scenario, the atmospheric CO 2 loads (12.3 and 47.1 Mg C-years) were lower than those of the Afast BAU scenario (14.2 and 48.4 Mg C-years) for 20-and 100-year time horizon, respectively, due to the slow decomposition rate of this biomass (Figure 4b).
For the F-fast scenario, the atmospheric loads were 10.9 and 44.4 Mg C-years for a 20-and 100-year time horizon, respectively ( Figure 4c). The rate at which the biomass released C from the biosphere to the atmosphere was more gradual for the F-slow BAU scenario relative to the F-fast scenario (Figure 4d). As a result, the atmospheric loads (1.2 and 18.2 Mg C-years) of this scenario were lower than those of the F-fast scenario (10.9 and 44.4 Mg C-years) for 20-and 100-year time horizon, respectively.
Similar to the forest BAU scenario, there was a substantial difference between the atmospheric CO 2 load for the fast-decomposing lignocellulosic biomass in landfills (Figure 4e) relative to the slow-decomposing biomass ( Figure 4f). The atmospheric CO 2 loads for the fast BAU were 3.1 and 30.9 Mg C-years for 20-and 100-year time horizon, respectively, whereas the atmospheric CO 2 loads of the L-slow BAU biomass were 0.2 and 1.2 Mg C-years for a 20-and 100-year time horizon, respectively.

| Atmospheric CO 2 load from residual biomass combustion versus decomposition
An equivalent amount of C was released into the atmosphere when fast-decomposing agricultural residues were diverted to bioenergy production or left in the field to decompose (Figure 5a). Also, the difference between the atmospheric load for slow-decomposing agricultural residues and the bioenergy scenario was only 2.09 Mg Cyears and 1.43 Mg C-years based on the 20-and 100-year time horizon, respectively ( Figure 5b).
The differences between the atmospheric CO 2 loads of the F-fast BAU scenario and the bioenergy scenario ( 20 A = 3.44 Mg C-years and 100 A = 4.10 Mg C-years) were lower than the differences between the atmospheric CO 2 loads of the F-slow BAU and the bioenergy scenarios ( 20 A = 13.2 Mg C-years and 100 A = 30.3 Mg C-years based on 20-and 100-year time horizon, respectively) (Figures 5c,d).
The atmospheric CO 2 loads of the bioenergy scenario were 11.3 and 17.6 Mg C-years, higher than the L-fast BAU scenario based on a 20-and 100-year time horizon, respectively ( Figure 5e). The highest difference between the BAU and bioenergy scenario ( 20 A = 14.2 Mg C-years and 100 A = 47.3 Mg C-years based on 20-and 100-year time horizon, respectively) occurred when slow-decomposing landfill biomass was combusted for energy ( Figure 5f).

| Global warming potential of biogenic CO 2
For agricultural residues, the 20 GWP bCO2 values ranged between 0.014 g CO 2 e/g b CO 2 for fast-decomposing biomass and 0.146 for the slow-decomposing biomass while the 100 GWP bCO2 values ranged between 0.003 for the fast and 0.029 for the slow (Figure 6).
The 20 GWP bCO2 values for the forestry residues diverted to bioenergy production ranged between 0.240 g CO 2 e/g b CO 2 for the fast-decomposing biomass and 0.916 for the slow-decomposing biomass, whereas the 100 GWP bCO2 values ranged between 0.084 for the fast and 0.625 for the slow (Figure 6).
Considering all the BAU scenarios, the 20 GWP bCO2 values ranged between 0.014 g CO 2 e/g b CO 2 for the fastdecomposing agricultural residues and 0.989 for the slow-decomposing lignocellulosic biomass in landfills ( Figure 6). The 20 GWP bCO2 values were higher than those of the 100 GWP bCO2 , which ranged between 0.003 g CO 2 e/g b CO 2 for the fast-decomposing agricultural residues and 0.975 for the slow-decomposing lignocellulosic biomass landfills.

| Carbon transfer to the atmosphere from residual biomass combustion and decomposition
When C-based compounds are combusted, whether from fossil fuels or biomass, there is an instantaneous release of C as CO 2 to the atmosphere (Cherubini et al., 2011;Gustavsson et al., 2017) as shown in Figure 2. If the feedstock is residual biomass diverted to energy use, the destiny of the embedded C could differ greatly from the same biomass being left to decomposition under various BAU scenarios (Figure 3).
Agricultural residues tend to decompose quickly, releasing up to 90% of their biomass C to the atmosphere within the first year (Figure 3a; Lehmann et al., 2006). As a rich source of nutrients for microorganisms, agricultural biomass is highly susceptible to degradation (Beyaert & Voroney, 2011).
Forest residues tend to decompose more slowly (Figure 3b; Russell et al., 2014), depending on plant composition and environmental factors, including moisture and temperature (Kurz et al., 2013;Zhang et al., 2008). Conditions conducive to microbial degradation release biomass CO 2 more quickly.
Landfills are often an important C sink because the soil and environmental conditions are less favourable for microbial degradation of biomass (Micales & Skog, 1997;Ximenes et al., 2008). Fungi, which are F I G U R E 4 The effect of biomass (orange) decomposition on atmospheric (blue) and biosphere (green) C pools. Fast and slow decomposition rates (see Figure 3) were modelled for one megagram of biomass carbon from agricultural (a, b) and forestry (c, d) residues as well landfill lignocellulosic biomass (e, f) at t = 0 on the pools of atmospheric (blue-shaded area) and biosphere carbon (green-shaded area) over 100 years from the start of the decomposition. All decomposition occurring in year x to x + 1 was assumed to occur at t = x. The orangeshaded area shows the amount of biomass carbon remaining. The Mg C-years of atmospheric CO 2 load (i.e. blue-shaded area) was calculated for the 0-to 20-year and 0-to 100-year period and the values are shown on the chart. capable of decomposing woody biomass, do not thrive well under the anaerobic condition of most landfill sites, whereas bacteria that can thrive in this condition are not efficient in decomposing woody biomass (Ximenes et al., 2008). In fact, the rate of C accumulation in landfills is higher than the rate of accumulation in structural materials such as buildings (Apps et al., 1999;Harmon et al., 1996).

F I G U R E 5
Comparison of biomass combustion (red line) and decomposition (black line) on atmospheric CO 2 pools. The grey regions show the differences between the combustion and decomposition curves and reflect the additional CO 2 added to the atmosphere following a decision to divert one Mg biomass C to energy production. Panels (a), (c), and (e) depict the fast decomposition rate for agricultural, forestry, and landfill biomass, respectively. Similarly, Panels (b), (d), and (f) depict the slow decomposition rate for the three biomass residues. The Mg C-years of atmospheric CO 2 load (i.e. grey-shaded area) was calculated for the 0-to 20-year and 0-to 100-year period and the values are shown on the chart.

| Atmospheric CO 2 load from residual biomass combustion versus decomposition
Whether from combustion or decomposition, pulses of CO 2 entering the atmosphere are gradually removed by the natural sinks, including the ocean and terrestrial ecosystem, as described by the Bern 2.5 CC model (Bright et al., 2012;Cherubini et al., 2011;Joos et al., 2001) and as shown in Figures 2 and 4. Within 20 years of a single CO 2 pulse, about 40% of the CO 2 is removed, and this rises to 60% after 100 years (Figure 2). The blue-shaded area in Figure 2 is a measure of the atmospheric CO 2 load, and for the combustion of 1 Mg biomass C, the calculated values were 14.4 and 48.5 Mg C-years for 20-year and 100-year time horizon, respectively. In the case of biomass decomposition, the atmospheric CO 2 loads were lower, especially for forest residues and landfill biomass (Figure 4c-e). The difference in the atmospheric CO 2 loads between the combustion and decomposition scenarios represents the climate impact of the decision to divert residual biomass to bioenergy use. Large differences were observed among decomposition scenarios when compared to residual biomass combustion ( Figure 5). These values were incorporated into Equations (7) and (8), to calculate GWP bCO2 .

| Global warming potential of biogenic CO 2
Based on a 100-year time horizon, diverting agricultural residues to bioenergy only had 0.3%-2.9% of the climate impact of CO 2 released from fossil fuel combustion ( Figure 6b). This low impact explains why the GWP bCO2 associated with combusting agricultural biomass to energy is usually assumed to be negligible (Liu et al., 2017;Liu & Rajagopal, 2019).
In contrast, on a 100-year time horizon, combusting forestry residues for bioenergy had 8.4%-63% of the climate impact of CO 2 from fossil fuel combustion (Figure 6b).
For lignocellulosic biomass in landfills, the slow BAU decomposition of this biomass meant that the decision to divert it to bioenergy use would be associated with a larger global warming impact. Over a 100-year period, the CO 2 released from combusting landfill lignocellulosic biomass residues were estimated to have 36.4%-97.5% of the climate impact of CO 2 from fossil fuel combustion (Figure 6b).
These positive GWP bCO2 values show that the diversion of residual biomass to energy had an adverse impact on climate, indicating that the assumption of zero impact (Ragauskas, 2006) is not valid (Booth, 2018;Matuštík & Kočí, 2022;Searchinger et al., 2009).
It was difficult to compare the GWP bCO2 values of this study with those reported by previous studies because of differences in objectives and parameters accounted for. For instance, on a 100-year time horizon, Guest et al. (2013b) reported GWP bCO2 values that ranged between 0.44 and 0.62 for no and complete residue removal, respectively. The study was conducted to determine the effect of biogenic CO 2 released from the decomposition of residues generated from plants that are primarily grown for bioenergy production. However, this study examined the climate impact of diverting residual biomass generated from plants grown for food and fibre generation into bioenergy production.
The results of this study show that there are little or no negative impacts associated with diverting agricultural residues to bioenergy use given the low GWP bCO2 associated with this decision ( Figure 6). However, care must be taken when crop residues are diverted into bioenergy production since a fraction of the plant biomass is required to recycle plant nutrients, maintain microbial activities, control erosion, and improve water storage (Smil, 1999). Thus, it is often recommended that at least 30% of residues are left in agricultural lands to maintain soil health and crop productivity (Li et al., 2012).
Compared with agricultural residues, the higher GWP bCO2 values associated with diverting residual biomass from forest and landfill sites to bioenergy production results in more substantive adverse climate impacts. This is particularly the case when diverting to bioenergy use, the slowly decomposing lignocellulosic biomass destined for landfills ( Figure 6).
Of course, the GWP bCO2 values derived here provide a means to calculate only one of the factors needed to determine the system-level GHG impacts associated with diverting residual biomass to bioenergy use. A full lifecycle accounting of the GHG emissions associated with diverting residual biomass to energy should include: • The fossil fuel emissions displaced by the biomass energy. This would account for differences in the efficiencies of conversion systems for biomass to energy compared to fossil fuels to energy (Zhang et al., 2010). Delivering the same energy service typically requires more biomass energy and C than that needed from fossil fuels (Gustavsson et al., 2015;Liu et al., 2017Liu et al., , 2020. • The CH 4 and N 2 O emissions avoided by diverting residual biomass away from landfill sites, or by avoiding the anaerobic decomposition of forestry or agricultural residues (Cherubini, 2010;Micales & Skog, 1997).
(Note that if any of the residues were to be burned, diverting biomass to energy use would avoid significant non-CO 2 emissions. Of course, the CO 2 emission to the atmosphere, in this case, would be rapid, so the GWP bCO2 values for diverting biomass would be near zero.) • The fossil fuel emissions associated with collecting, processing, and transporting the residual biomass to the site where it will be used Repo et al., 2012). In a net-zero future, fossil emissions associated with transportation and electricity use should be negligible (Lof et al., 2019). • The CH 4 and N 2 O emissions that occur during biomass combustion for energy use (Ministry of Environment British Columbia, 2014;Repo et al., 2012).
Not included in the list above are the fossil fuel, CH 4 , or N 2 O emissions associated with the production of the biomass (Thakur et al., 2014) since those emissions would normally be assigned to the reason the biomass was produced in the first place (e.g. food or fibre production), not to the residues from that production (Gustavsson et al., 2015).
Another important factor to consider is the effect of wildfire on residual biomass that is left on the forest floor. Wildfire instantaneously releases biomass CO 2 into the atmosphere (Buchholz et al., 2016;Kurz et al., 2013); thus, the climate impact of the emitted CO 2 would be similar to that of CO 2 emitted into the atmosphere when biomass is combusted for energy ( Figure 2). Therefore, diverting residual biomass from the forest floor into bioenergy may reduce fire frequency and severity (Buchholz et al., 2016) while reducing fossil GHG emissions. Thus, it will be valuable to incorporate the impact of natural disturbances like wildfire when determining the climate impact of diverting residual biomass to energy. The findings will help inform policy and investment decisions relating to GHG emissions management in Canada.
It is also important to recognize that the decision to divert residual biomass to bioenergy use can have other environmental impacts (positive or negative) in addition to the contribution to climate forcing. Positive impacts could include increasing the lifetime of landfill sites or opening forest land to tree growth. Negative impacts could include increased ocean acidification if diversion enhances the CO 2 load on the atmosphere, reduction in biodiversity, and C and nutrient removal from soils (potentially requiring more fertilizer with resulting impacts on energy use, N 2 O emissions, and eutrophication of surface waters) (Cherubini, 2010;Weldu & Assefa, 2016).

| CONCLUSION
This paper reports on a new method to quantify the Global Warming Potential associated with biogenic CO 2 released to the atmosphere (GWP bCO2 ) when residual biomass from agriculture, forestry, and landfill sites is diverted to bioenergy use. On a 100-year time horizon, the GWP bCO2 values ranged between 0.003 for fast-decomposing agricultural biomass to 0.975 for slow-decomposing lignocellulosic landfill biomass.
This anthropogenic contribution to the atmospheric CO 2 load, and therefore to climate forcing, is currently not accounted for in most LCA of residual biomass use for energy. In a world striving for net-zero GHG emissions, GWP bCO2 must be considered in assessing the contribution of bioenergy to future energy systems. Since forest residues and lignocellulosic biomass destined for landfills are commonly diverted for use as energy resources, the findings provided here should be of importance to project developers and policy makers.
Of course, the GWP bCO2 factor defined here is only one of the factors that must be considered when assessing the system-level and lifecycle GHG implications of diverting residual biomass to energy use. Other factors include non-CO 2 and fossil fuel energy emissions avoided, and any other energy emissions associated with collecting, processing, transporting, and using the residual biomass for energy. Such an analysis has been carried out in a companion paper (Adetona et al., 2023).

ACKNO WLE DGE MENTS
This publication was supported by funding provided through scholarships to AA from the University of Calgary Faculty of Graduate Studies and grants received by DBL from the Ivey Foundation, the Edmonton Community Foundation, and the Transition Accelerator.