Forest harvest residues are important raw materials for bioenergy in regions practicing forestry. Removing these residues from a harvest site reduces the carbon stock of the forest compared with conventional stem-only harvest because less litter in left on the site. The indirect carbon dioxide (CO2) emission from producing bioenergy occur when carbon in the logging residues is emitted into the atmosphere at once through combustion, instead of being released little by little as a result of decomposition at the harvest sites. In this study (1) we introduce an approach to calculate this indirect emission from using logging residues for bioenergy production, and (2) estimate this emission at a typical target of harvest residue removal, i.e. boreal Norway spruce forest in Finland. The removal of stumps caused a larger indirect emission per unit of energy produced than the removal of branches because of a lower decomposition rate of the stumps. The indirect emission per unit of energy produced decreased with time since starting to collect the harvest residues as a result of decomposition at older harvest sites. During the 100 years of conducting this practice, the indirect emission from average-sized branches (diameter 2 cm) decreased from 340 to 70 kg CO2 eq. MWh−1 and that from stumps (diameter 26 cm) from 340 to 160 kg CO2 eq. MWh−1. These emissions are an order of magnitude larger than the other emissions (collecting, transporting, etc.) from the bioenergy production chain. When the bioenergy production was started, the total emissions were comparable to fossil fuels. The practice had to be carried out for 22 (stumps) or four (branches) years until the total emissions dropped below the emissions of natural gas. Our results emphasize the importance of accounting for land-use-related indirect emissions to correctly estimate the efficiency of bioenergy in reducing CO2 emission into the atmosphere.
Bioenergy, i.e. energy derived from renewable biomass, is used to replace fossil fuels in energy production in order to decrease greenhouse gas emissions into the atmosphere. The rationale behind this practice is that bioenergy does not cause any net carbon dioxide (CO2) emissions since the amount of CO2 released into the atmosphere in combustion is taken up again by the next generation of growing plants (Wihersaari, 2005; Stupak et al., 2007; Lattimore et al., 2009).
Following this idea, as a means to cut down greenhouse gas emissions, the Council of the European Union (EU) adopted a directive on the promotion of renewable energy, including bioenergy. This directive set targets to produce 20% of the final energy consumption using renewable energy sources in the EU by the year 2020. This target is higher for member states already producing a lot of renewable energy, for example as a by-product of pulping industry. Consequently, the national commitment is 38% for Finland and 49% for Sweden (Directive 2009/28/EC). During the reference year of the directive 2005, renewable sources represented already 28% of the total energy production in Finland and 39% in Sweden, while the EU-average was 11%.
These high targets for renewable energy are increasing the focus on biomass for energy production. Worldwide, this growing interest in bioenergy puts pressure on land use changes, including deforestation and consequent conversion of the forest land to energy crop cultivation (Melillo et al., 2009).
Recently, indirect CO2 emissions from bioenergy production associated with these land use changes have caused concern (Searchinger et al., 2008, 2009; Melillo et al., 2009). These indirect emissions occur when bioenergy production reduces the carbon stocks of biomass or soil. These carbon losses may be remarkable, and it is even possible that replacing fossil fuels with bioenergy increases net greenhouse gas emissions into the atmosphere as a consequence of large indirect emissions. The assumed CO2 neutrality of biofuels like ethanol and their actual potential to mitigate climate change have already been questioned because of large negative impacts on the carbon stock of soil (Fargione et al., 2008; Searchinger et al., 2008, 2009; Melillo et al., 2009).
It is important to realize that the indirect emissions of bioenergy production are not limited to the cases of land use change but may also be caused by new practices of ecosystem management within the same land use. In countries with extensive forest cover and an already high share of renewable energy, an appealing way to produce bioenergy is to intensify biomass removals from forests. Forested countries Finland and Sweden are pioneers in the field of using forest residues for energy production (Mälkki & Virtanen, 2003). Still, in order to meet the EU commitment of renewable energy, Finland plans to increase the use of logging residues for energy production from 3.6 Mm3 yr−1 in 2006 to 12 Mm3 yr−1 by 2020 (Ministry of Employment and the Economy of Finland, 2008).
Until now, research on the effects of logging residue removal has focused on nutrient balances (e.g. Wall, 2008; Luiro et al., 2009), socioeconomic impacts (e.g. Börjesson, 2000), profitability (e.g. Heikkiläet al., 2007), forest productivity (e.g. Peng et al., 2002) and properties of wood fuel (e.g. Alakangas, 2005), whereas the indirect emissions have received little attention. Lattimore et al. (2009) dealt with the indirect emissions to some extent in their recent review. They concluded that, in order to be sustainable, bioenergy production from forest residues must not have adverse effects on soil quality, hydrology and water quality, site productivity, or forest biodiversity but also not on greenhouse gas balances.
The indirect emissions from removing forest harvest residues, and using them for energy production, result from combusting the residues and releasing CO2 into the atmosphere soon after harvesting instead of letting them decompose slowly at the harvested site. As a consequence of such practice, the amount of carbon stored at the forest site decreases, possibly to a remarkable degree.
The indirect emission of using logging residues for energy production depend critically on the decomposition rate of the residues if they were left at the site. Studies based on extensive sets of measurements have been published recently making it possible to estimate the decomposition rate of the harvest residues more reliably than before (e.g. Tarasov & Birdsey, 2001; Palviainen et al., 2004; Mäkinen et al., 2006; Vávřováet al., 2009). We have used these measurements plus other measurements related to decomposition and carbon cycling in soil and developed a new soil carbon model Yasso07 (Tuomi et al., 2008, 2009). The large datasets used and advanced mathematical methods applied make the Yasso07 model particularly suitable for estimating the decomposition rate of woody litter in boreal forests.
In this study, we used this model to estimate the indirect emissions from using logging residues for bioenergy production. The objectives of the current study were to (1) introduce an approach to estimate the indirect CO2 emissions associated with bioenergy production from forest harvest residues, (2) estimate these emissions in a forested boreal landscape during the first 100 years after starting to produce bioenergy from harvest residues, and (3) compare the total CO2 emissions per unit of bioenergy produced to the emissions caused by using other fuels.
Materials and methods
Modelling decomposition of forest harvest residues
To estimate the indirect CO2 emissions from producing bioenergy from forest harvest residues in boreal conditions, we simulated the decomposition of logging residues using a user-interface of the dynamic soil carbon model Yasso07 (Tuomi et al., 2008, 2009, http://www.environment.fi/syke/yasso). The measurements of the decomposition of woody litter used to develop the model were taken in Finland and neighboring regions in Estonia and Russia (M. Tuomi, R. Laiho, A. Repo & J. Liski, unpublished results). These measurements used represent the majority of data on woody litter decomposition in this region. The data sets includes branches and stems ranging from 0.5 to 60 cm in diameter, and the mass loss of these woody biomass components has been followed for 1–70 years since the start of decomposition. In addition to these data, the Yasso07 model is based on an extensive data set on decomposition of nonwoody litter across Europe and North and Central America (Tuomi et al., 2009) plus data sets on the accumulation and stock of soil organic carbon (Liski & Westman, 1995, 1997; Liski et al., 1998). These additional measurements specially provide information on the cycling of recalcitrant organic carbon compounds in soil that are relevant for the long-term carbon balance of decomposing woody litter.
The parameter values of Yasso07 have been sampled using a Markov chain Monte Carlo method (Tuomi et al., 2009). This method has been used to make sure that, first, the model is not over-parameterized given the data and, second, there are unequivocal maximum likelihood values for each parameter combination. This mathematical approach and the data available in the development process of Yasso07 make this model suitable for this study because the uncertainty estimates of the model predictions are available and because the data covers the simulated scenarios well without extrapolation.
Using Yasso07, we simulated the decomposition of harvest residues at a typical site of harvest residue removal in Finland, namely an even-aged mature 81–100-year-old Norway spruce (Picea abies L.) forest stand located in the Pirkanmaa region of Southern Finland. Spruce stands cover some 40% of forest area in this region (Korhonen et al., 2000). Clear-cut spruce stands are favorable targets of harvest residue removals because there are more logging residues than in clear-cut Scots pine stands, which are also common in the region. Spruce stumps are also preferred over pine stumps because they are easier to extract from the soil because of a shallower root structure (Alakangas, 2005; Wihersaari, 2005).
To illustrate differences in decomposition rate between different harvest residues, we simulated decomposition of spruce branches varying from 1 to 5 cm in diameter and stumps varying from 10 to 35 cm in diameter for a 100-year period after the start of decomposition. The mean diameters of spruce branches and stems in the study region, 2 and 26 cm, respectively (Korhonen et al., 2000; Kantola et al., 2007), were chosen for more detailed analyses of the indirect emissions caused by using the logging residues for bioenergy production. The other input variables of the Yasso07 model used in the simulations are shown in Table 1. When estimating the indirect emissions we assumed that needles were left at the site and only branches or stumps were removed. We assumed also that there was little or no delay in combusting the harvest residues at a power plant and thus CO2 was released to the atmosphere at once.
Table 1. The values of input variables used in the Yasso07 model
Mean annual temperature
Chemical composition of woody litter
Branch ± SD/stump ± SD (%)
The climate values represent averages for southern Finland between 1971 and 2000 (Drebs et al., 2002) and the chemical composition averages of several individual studies (Hakkila, 1989). The standard deviation (SD) values of the chemical composition are based on coefficients of variation calculated from a data base of foliage litter (Berg et al., 1991). The temperature amplitude means a half of the difference between the mean temperatures of the warmest and the coldest month of the year.
Acid hydrolysable compounds
59 ± 4.3/70 ± 5.0
Water soluble compounds
1 ± 0.3/1 ± 0.2
Ethanol soluble compounds
1 ± 0.3/1 ± 0.2
Klason lignin (neither hydrolysable nor soluble compounds)
37 ± 1.0/28 ± 0.8
Energy production estimation
The indirect CO2 emissions from using the harvest residues for bioenergy production were taken to be equal to amount of carbon remaining in the harvest residues if the residues were left to decompose at the site harvested. These emissions were also related to the amount of bioenergy produced. The cumulative indirect emissions caused by combusting the harvest residues until year i were calculated by summing up the amounts of carbon left in the harvest residues until this year (i) and relating these emissions to the cumulative amount of bioenergy produced. In other words, we simulated a case where the practice of removing the harvest residues and using them for bioenergy production was started and continued on a harvest area of similar size year after year. We applied biomass-compartment-specific net calorific heating values to estimate the amount of energy obtained by combusting the harvest residues, i.e. 19.30 MJ kg−1 for dry Norway spruce branches and 19.18 MJ kg−1 for dry stumps (Nurmi, 1997). These values of dry logging residues range commonly from 18 to 20 MJ kg−1 (Alakangas, 2005). The carbon content of the harvest residues was assumed to be equal to 50% of dry wood (m/m).
In order to estimate the full fuel cycle emissions from using the logging residues for energy production, we added other emissions from a typical wood chip fuel production chain using harvest residues to the calculated indirect emissions. These emissions result from (1) collecting, chipping, and transporting the harvest residues, (2) emitting methane (CH4) and nitrous oxide (N2O) from combustion, (3) fertilizing the forest to compensate for nutrient loss, and (4) recycling ash, and they range typically from 5 to 18 kg CO2 eq. MWh−1 produced (Palosuo et al., 2001; Wihersaari, 2005). We used a central value of this range, 12 kg CO2 eq. MWh−1 produced, in our calculations. Other estimates for direct emissions from wood fuel chain range from 5 to 20 kg CO2 eq. MWh−1 (Korpilahti, 1998; Mälkki & Virtanen, 2003; Berg, 2010) depending on background information and included operations in the estimates. Fossil fuel comparison values 280, 306, and 395 kg CO2 eq. MWh−1 were estimates of entire fuel cycle emissions of natural gas, oil, and diesel, and coal (Statistics Finland, 2006; Ecoinvent Centre, 2007).
Branches lost mass at a remarkably higher rate than stumps because the simulated decomposition rate of woody litter was dependent on the initial diameter (Fig. 1). For example, after 10 years of decomposition, the branches 1–5 cm in diameter had 30–55% of the initial mass still remaining (Fig. 1a) while stumps 10–35 cm in diameter had 63–81% (Fig. 1b). The simulated rate of mass loss decreased over time, and after 100 years of decomposition there was still 2–16% of the initial branch mass remaining and 19–28% of the initial stump mass remaining.
The indirect CO2 emissions, caused by combusting logging residues after harvesting instead of letting them decompose at the harvested site, were equal to the CO2 emissions from combustion, 340 kg CO2 MWh−1, when the practice was started but these emissions decreased over time as a result of decomposition of the harvest residues (Fig. 2). The indirect emissions of using branches for bioenergy decreased faster than the emissions of using stumps because the branches decomposed faster (Figs 1 and 2). After the first 10 years of conducting this practice, the indirect emissions from branches were equal to 200 kg CO2 MWh−1 and those from stumps 310 kg CO2 MWh−1 (Fig. 2). After the first 100 years, these emissions from the branches and stumps were equal to 70 and 160 kg CO2 MWh−1, respectively.
The estimated greenhouse gas emissions from the rest of the bioenergy production chain, i.e. the emissions from collecting, transporting, chipping, and combusting the harvest residues plus the emissions from fertilizing the forest and recycling the ash, were equal to 12 kg CO2 eq. MWh−1 (Fig. 3). At the time of starting this practice, these direct emissions represented only 3% of the total emissions caused by using the harvest residues for bioenergy. After the 100-year period of conducting this practice, this share increased to 15% if bioenergy was produced from branches and to 7% if bioenergy was produced from stumps (Fig. 3). The increased contribution of the direct emissions was a result of the decreased indirect emissions (Fig. 2).
At the time of starting to use the harvest residues for energy production, the total emissions were comparable to the emissions caused by using fossil fuels (Fig. 3). After 10 years of producing bioenergy from branches, the total emissions caused were 210 kg CO2 eq. MWh−1. These emissions are 24%, 30%, or 46% lower than the emissions caused by producing the energy from natural gas, oil, or coal, respectively. If bioenergy was produced from stumps, it took 22 years for the average emissions to decrease below the emissions of producing the energy from natural gas or 14 years for the emissions to decrease below the emissions of oil. After 100 years of producing energy by combusting branches, the emissions were lower by 71%, 74%, or 79% than the emissions caused by producing the energy from natural gas, oil, or coal, respectively. For bioenergy produced by combusting stumps, these percentages of emission reductions were 40%, 46%, or 58%, over 100 years, compared with the emissions of natural gas, oil, or coal, respectively.
Using logging residues for energy production decreases the amount of carbon stored in forest and causes thus indirect CO2 emissions into the atmosphere. These indirect emissions occur because combustion releases carbon of logging residues to the atmosphere at once, otherwise the residues would form a slowly decomposing carbon stock at the harvest site. The indirect emissions are an order of magnitude larger than the direct emissions from the rest of this bioenergy production chain (Fig. 3), not accounting for the CO2 emissions of combusting the harvest residues. The indirect emissions depend on the decomposition rate of the harvest residues, with the decomposition rate being lower with an increasing size of woody litter (Fig. 1, Harmon et al., 1986; Janisch et al., 2009; M. Tuomi, R. Laiho, A. Repo & J. Liski, unpublished results). The use of bigger-sized stumps for energy production causes therefore larger indirect emissions than the use of smaller-sized branches (Fig. 2). The average indirect emissions per unit of energy produced decrease over time since the start of this form of bioenergy production (Fig. 2) but during the first few years, or a couple of decades in the case of stumps, the total emissions caused by energy production from harvest residues are comparable to the emissions of fossil fuels (Fig. 3).
The reliability of the current results depends critically on the decomposition estimates of the harvest residues. The Yasso07 decomposition and soil carbon model used in this study is based on a large collection of mass loss measurements taken on woody litter in boreal forests across Finland and neighboring countries plus large sets of other relevant measurements from across the world (see ‘Materials and methods’). This model has been shown to give unbiased estimates for the decomposition of woody (M. Tuomi, R. Laiho, A. Repo & J. Liski, unpublished results) and nonwoody litter (Tuomi et al., 2009). Compared with other studies on the decomposition of woody litter carried out under comparable conditions, the estimates of this study are similar except for the end of the 100-year study period; for this late phase of decomposition, the current estimates of mass remaining are higher. Mäkinen et al. (2006) estimated that Norway spruce stems lost about 20% of their initial mass during the first 10 years of decomposition and the stems disappeared completely in some 60–80 years of time. On the other hand, according to the residuals reported by Mäkinen et al. (2006), their model seems to underestimate the mass remaining after 30 years of decomposition. Melin et al. (2009) compared several decomposition models developed for Norway spruce logs, snags, and stumps. They concluded that after 10 years of decomposition some 60–75% of the initial mass was still remaining whereas after 100 years of decomposition practically none or <10% of the initial mass was still left. Our higher estimates of mass remaining during the late phases of decomposition can be explained by a difference between the methods used. The two earlier studies were based on measurements of woody litter mass remaining. These measurements may not capture the formation and translocation of well-decayed soil organic matter originating from woody litter. The Yasso07 model, on the other hand, is additionally based on measurements of formation of soil organic matter (Tuomi et al., 2009). For this reason, we think that the estimates of Yasso07 model are probably more realistic for the late phases of woody litter decomposition.
In addition to the decomposition estimates, the reliability of the current results depends also on the other parameters of our calculations such as the reference energy systems or combustion techniques chosen as well as variation in the chemical composition of litter (affecting the decomposition rate estimates of the harvest residues) and calorific values. Furthermore, the simulations were done for climate conditions prevailing in Southern Finland today. The results do not thus account for the effects of climate change or those of increased atmospheric CO2 concentration on forest growth. In addition, forest soil disruption associated with stump removal may release additional CO2 into the atmosphere. Currently, empirical research on the magnitude of these emissions is few in number (Jandl et al., 2007; Walmsley & Godbold, 2010). Hope (2007) found that stump removal together with forest floor scarification reduced soil carbon stocks in the first year of stump harvesting and 9 years later. This conclusion is apparent only if during stump removal the forest soil is completely scarified by removal or mixing with mineral soil. Although the quantitative data on stump extraction and emissions associated with this practice is scarce, generally it is known that soil disturbance can change the microclimate and stimulate the decomposition of litter (Johansson, 1994). In a Finnish study site preparation after a clear-cut with mixing organic matter with mineral soil increased CO2 efflux from soil but this effect leveled off rapidly (Pumpanen et al., 2004). Despite of these uncertainties, we think that the current estimates are reliable enough to demonstrate the magnitude of indirect emissions associated with producing bioenergy from harvest residues in boreal coniferous forest.
The indirect emissions, caused by the reduced carbon stock of decomposing harvest residues, represented 85–97% of the total emissions of this bioenergy production chain, not accounting for the CO2 emissions from combusting the harvest residues. These emissions are thus highly significant for the full fuel cycle emissions of logging residues. When comparing the present results to earlier ones it is important to acknowledge differences in system boundaries, energy use-technology, reference energy system, conversion technology and type and management of raw material (Cherubini et al., 2009). Still, it is possible to conclude that a common outcome of the earlier studies is that the greenhouse gas emissions from logging, collecting, chipping, and transporting harvest residues are relatively low (Börjesson, 1996; Palosuo et al., 2001; Mälkki & Virtanen, 2003; Wihersaari, 2005). These estimates have ranged from 4 to 20 kg CO2 MWh−1 depending on the details of the bioenergy production chains studied. The current study shows that if indirect CO2 emissions are accounted for, the total CO2 emissions are at least an order of magnitude higher than these emissions which have been considered to represent the total emissions (cf. Mälkki & Virtanen, 2003).
The decreasing effect of logging residue removal on soil carbon stock has been demonstrated earlier but this has not been considered to be problematic as long as this removal practice does not jeopardize the carbon sink of soil (Börjesson, 2000; Ågren & Hyvönen, 2003; Mälkki & Virtanen, 2003; Petersen Raymer, 2006; Eriksson et al., 2007; Sievänen et al., 2007; Eriksson & Gustavsson, 2008). Sievänen et al. (2007) calculated that increasing the removals of logging residues from 4 to 15 Mm3 yr−1 in Finland will not turn the Finnish forests from net carbon sinks to net sources. However, the intensified removals of the logging residues would decrease the annual carbon sink of these forest soils by 3.1 million tons of CO2 eq. (Sievänen et al., 2007). Assuming that 1 m3 of harvest residues gives 2 MWh of energy (Alakangas, 2005), the indirect emission from the decreasing carbon stock of soil is equal to about 100 kg CO2 eq. MWh−1. This estimate is somewhat lower than the estimates of the present study because Sievänen et al. (2007) applied an earlier version of Yasso soil carbon and decomposition model in their study (Liski et al., 2005). This earlier model version gave less reliable, higher estimates for the decomposition rate of woody litter because it was based on a substantially smaller number of measurements. Nevertheless, these figures demonstrate that it is important to relate the indirect and direct emissions of bioenergy production to the amount of energy obtained in order to get a correct picture on the efficiency of using different energy sources in decreasing greenhouse gas emissions to the atmosphere. The current results demonstrate that if the indirect CO2 emissions are counted in, the total fuel chain emissions from using spruce branches or stumps for energy production may cause even bigger CO2 emissions during the first years or decades of starting this practice than producing energy from oil or natural gas.
The indirect greenhouse gas emissions resulting from using logging residue for bioenergy production are highest per unit of energy produced immediately when the practice is started. The average emissions per energy unit decrease, however, over time. As a result of this temporal pattern, this form of bioenergy production is not efficient in decreasing emissions to the atmosphere in the near future. Our results stress the importance of considering time perspective when assessing the potential of different bioenergy options to mitigate climate change (Schlamadinger & Marland, 1996; Petersen Raymer, 2006). The issue of which temporal approach is appropriate depends on the management and policy strategies and whether the selection of energy systems is made to meet long-term or short-term greenhouse gas reduction objectives (Schlamadinger et al., 1997).
It is possible to reduce the indirect emissions of logging residue removals by collecting quickly decomposing harvest residues for bioenergy production, for example branches instead of stumps. However, it may be still tempting to extract stumps from harvest sites because the gain of primary energy per hectare may be twice that compared with collecting branches (Eriksson & Gustavsson, 2008). Leaving the needles at the harvest site, which helps to avoid nutrient loss, has a marginal effect on the carbon balance at clear-cut sites, although needle and fine root litter produce more than two-thirds of the soil carbon stock in growing forests (Ågren & Hyvönen, 2003).
The indirect emissions have a remarkable effect on the total greenhouse gas emissions from some systems of bioenergy production, as demonstrated in this study for harvest residues of boreal forests, and emphasized for several other systems by Johnson (2009) and Searchinger et al. (2009). For this reason, to account for the actual greenhouse gas effect of various alternative bioenergy options, it would be essential to include the indirect emissions adequately in guidelines of greenhouse gas inventorying and reporting. Currently, the rules of carbon accounting applied under the Kyoto Protocol do not count all indirect emissions, which among other things distorts the accounting of net emissions (Johnson, 2009; Searchinger et al., 2009). A particular shortcoming of the accounting rules under the Kyoto Protocol is that a party to this protocol may choose not to account for changes in one or several of the agreed carbon pools (aboveground biomass, belowground biomass, litter, dead wood, and soil organic carbon) as long as it can reliably show that the pool is not decreasing. Owing to this threshold, some of the indirect emissions caused by using logging residues for bioenergy production may be excluded from the inventory figures. On the other hand, in inventory reports of greenhouse gases under the United Nations Framework Convention on Climate Change, the carbon release resulting from forest harvesting must be counted and reported as a land-use emission or as a energy emission but not both (IPCC, 2000). Today, the land-use-related greenhouse gas emissions from bioenergy production systems are recognized and investigated (IPCC, 2000; Melillo et al., 2009) but one of the practical problems is that measuring methods for the indirect effects and feasible means to bring these impacts to regulatory policies are still lacking (Mathews & Tan, 2009).
Using logging residues as a source of bioenergy causes net CO2 emissions into the atmosphere and a great majority (85–97%) of these emissions are indirect emissions resulting from a decline in the carbon stock of harvest residues in forest. The amount of the indirect emissions increases with a decreasing decomposition rate of the harvest residues. Norway spruce stumps decay at slower rate than branches and consequently the energy use of the stumps causes 1.5–2 times larger indirect emissions than the use of branches. Production of bioenergy from forest harvest residues causes emissions that are comparable to the emissions of fossil fuel over the first few years (branches) or first few decades (stumps) of the practice. After 50 years, bioenergy produced of Norway spruce stumps decreases average emissions per unit of energy produced by some 20% and bioenergy from branches by some 60% compared with entire fuel cycle emissions of natural gas.
This study was partially funded by the European Commisson through research project ‘Greenhouse gas management in European land use systems’ (grant agreement no. 244122). We thank two anonymous reviewers for their comments and suggestions.