• Open Access

Net atmospheric impacts of forest bioenergy production and utilization in Finnish boreal conditions



The net CO2 exchange of forests was investigated to study net atmospheric impact of forest bioenergy production (BP) and utilization in Finnish boreal conditions. Net CO2 exchange was simulated with a life cycle assessment tool over a 90-year period and over the whole Finland based on National Forest Inventory data. The difference in the net exchanges between the traditional timber production (TP) and BP regime was considered the net atmospheric impact of forest bioenergy utilization. According to the results, forests became net sources of CO2 after about 20 years of simulation, and the net exchange was higher in the BP regime than in the TP regime until the middle of the simulation period. From 2040 onwards, the net exchange started to decrease in both regimes and became higher in the TP regime, excluding the last decade of the simulation. The shift of forests to becoming a CO2 source reflected the decrease in CO2 sequestration due to the increasing share of recently harvested and seedling stands that are acting as sources of CO2, and an increase of emissions from degradation of wood products. When expressed in terms of radiative forcing, the net atmospheric impact was on average 19% less for bioenergy compared with that for coal energy over the whole simulation period. The results show the importance of time dependence when considering dynamic forest ecosystems in BP and climate change mitigation. Furthermore, the results emphasize the dualistic role and possibilities of forest management in controlling the build and release of carbon into and from the stocks and in controlling the rate of the build speed, i.e. growth. This information is needed in identifying the capability and possibilities of ecosystems to produce biomass for energy, alongside other products and ecosystem services (e.g. pulp wood and timber), and simultaneously to mitigate climate change.


The European Union is committed to reducing its green house gas emissions by 20% and to raising the share of renewable energy (including bioenergy) to 20% by 2020, which will evidently increase the utilization of various sources of bioenergy. Global warming and growing demand for energy have also triggered the global interest in bioenergy utilization and biofuel crop development. The rationale behind this is that bioenergy produced in various ecosystems will not increase the net atmospheric greenhouse gas emissions, as the carbon emitted in combustion is sequestered into ecosystems by growth of new generations.

The carbon neutrality of renewable biomass has recently been questioned globally, due to high indirect greenhouse gas emissions consequent upon land-use changes associated with bioenergy production (BP) (Searchinger et al., 2008; Melillo et al., 2009). BP can be affected by climatic stress, releasing carbon emissions into the atmosphere comparable with coal emissions, as evidenced for agricultural crops at an organic site in Finland (Shurpali et al., 2010). In the case of forest bioenergy (i.e. logging residues, roots, and stumps), the importance and time dependence of indirect emissions have also been emphasized in recent studies, in which carbon emissions into the atmosphere have been studied via carbon stock changes in Nordic forests (Melin et al., 2010; Repo et al., 2010).

Biomass harvesting for energy alters the carbon balance of an ecosystem, as it decreases the amount of organic matter entering the soil. However, carbon sequestration, carbon stocks, and changes in them through time are also strongly dependent on existing forest resources, forest management practices (e.g. planting density, thinning intensity, fertilization), and the rotation periods followed. Conversely, wood products store carbon, and the size and lifespan of this storage is dependent on the forest management practices followed during the rotation and on the end-use of individual wood-based products. All these factors interact and affect the capacity of forests to reduce carbon emissions, to compensate for fossil fuel use, to increase sequestration of carbon from the atmosphere, and to increase long-term carbon storage. To date, there is no comprehensive knowledge of long-term greenhouse gas balances and their dynamics in forest ecosystems, especially related to energy biomass production.

At the regional level, the structure of managed forests (e.g. species composition and age class distribution) reflects the existing carbon stocks and the potential for bioenergy (Alam et al., 2010; Gustavsson & Sathre, 2011). In addition, existing resources and their future development define the timber yield which can be harvested within a certain time period. From the point of view of greenhouse gas balances, recently regenerated stands and over-mature stands lose carbon, and the net sequestration is strongly dependent on the growth rate of the stands in a region. Thus, the carbon sink and source considerations concerning the whole production chain, and the simultaneous effects of management and carbon stock changes in forest ecosystems, should be considered at the same time as assessment of net atmospheric impacts of BP. Holistic analysis of the mitigation capacity of forests includes not only components of the ecosystem dynamics and management but also carbon emissions of biomass harvesting, processing, and transportation. This helps in identifying the capability and possibilities of ecosystems to produce biomass for energy, alongside other products and ecosystem services (e.g. pulp wood and timber), and simultaneously to mitigate climate change.

In the above context, we simulated net carbon dioxide (CO2) exchange and estimated the consequent radiative forcing (RF) impacts of BP and utilization in Finnish boreal conditions.

Material and methods

General approach

The atmospheric impact of forest BP was approached using net ecosystem carbon dioxide (CO2) exchange simulations for the whole of Finland, utilizing the Sima ecosystem model (Kellomäki et al., 1992; Kolström, 1998) and the life cycle assessment (LCA) tool for forest production (Kilpeläinen et al., 2011). The tree data input for the simulations were accessed from the Finnish National Forest Inventory (NFI9). The forest management followed the practices recommended for the Finnish forestry (Tapio, 2006; Äijälä et al., 2010). Annual net ecosystem carbon exchange values were calculated over a 90 year period for the traditional timber production (TP) regime and BP regime. The difference between these two regimes was considered a net atmospheric impact of forest BP, and finally, this was expressed as RF of BP and utilization over the whole of Finland.

Net CO2 exchange calculation

LCA tool

The estimation of annual CO2 exchange of the stands was undertaken using the LCA tool (Kilpeläinen et al., 2011) according to Eqn (1). It includes components from both the ecosystem and technosystem related to forest production. The life cycle of energy biomass and timber starts from seedling production in a nursery, proceeds through management and harvests, and finishes in the yard of a pulp mill (pulpwood), sawmill (sawlogs), and power plant (bioenergy). Carbon sequestration into the forest ecosystem (growth, above-ground, and below-ground) has negative values (carbon flows from the atmosphere to the forest). Therefore, the net exchange (Cnet) is calculated by adding carbon emissions of forest management operations (Cman), litter and humus decomposition (Cdecomp), emissions from the combustion of bioenergy and degradation of wood-based items manufactured from timber (Charv) to the sequestrated carbon (Cseq). The release of carbon by combustion to produce bioenergy was assumed to take place immediately after the harvesting.

display math(1)

Produced timber (pulpwood, saw logs) was converted into useable wood-based products, and the carbon emissions from the items no longer in use were calculated, applying Eqn (2) (Karjalainen et al., 1994):

display math(2)

where PU is the proportion (0…100) of products in use and a, b, d are fixed parameters (120, 5, 120, respectively). C (a−1) is 0.5, 0.15, 0.065, and 0.03 for short, medium-short, medium-long, and long lifespan of a product, respectively. In our calculations, the pulpwood represented the items with a medium-short lifespan and the sawlogs the items with a medium-long lifespan. Accordingly, the half-lives for the pulpwood and sawlogs were 15 and 36 years, respectively. The carbon released from products that were no longer in use was assumed to convert completely into CO2.

Sima model

Estimation of ecosystem carbon fluxes

The Sima model was utilized to estimate Cseq, Cdecomp and the amounts of harvested timber and bioenergy as an input to the LCA tool. The details of the model are described in several earlier papers (Kellomäki et al., 1992, 2005, 2008; Kellomäki & Kolström, 1994; Talkkari & Hypén, 1996; Kolström, 1998; Alam et al., 2008, 2010, 2011; Kilpeläinen et al., 2011). Therefore, only the main outline is given here, focusing on the essential components of the model simulations and the earlier validation and performance of the model.

The Sima model is parameterized for the species growing between the latitudes N 60° and N 70° and longitudes E 20° and E 32° within Finland. In the model, the dynamics of the forest ecosystem is assumed to be determined by the dynamics of the number and mass of trees as regulated by their regeneration, growth, and death. All these processes are related to the availability of resources, which are in turn regulated by the dynamics of gaps in the canopy of the tree stand.

The growth and development of forest are determined by the incorporated four environmental subsystems: temperature, light, soil moisture, and decomposition. These determine the site conditions regarding temperature sum (degree-days, +5 °C threshold), within-stand light conditions, soil moisture, and soil nitrogen. These factors therefore directly affect the regeneration and growth of trees and indirectly influence the death of trees in tree populations and communities.

Temperature controls the geographical thresholds and annual growth response of each species and their ecotypes. At the same time, competition for light controls tree growth and is dependent on tree species and their height distribution. Soil moisture indicates the balance between precipitation, evaporation, and drainage, and its effect is determined through the number of dry days, i.e. the number of days per growing season with soil moisture equal to or less than that of the wilting point specific to soil types and tree species. The availability of nitrogen is controlled by the decomposition of litter and soil organic matter (SOM, or humus) and is dependent on the quality of litter (lignin, carbon, and nitrogen content), humus, and evapotranspiration.

The amount of litter and humus on the site is defined on the basis of the humus layer measured during the inventory of permanent sample plots undertaken through the National Forest Inventory (NFI) in Finland. The thickness is converted into the mass of SOM using bulk density of SOM, considering the site type and tree species dominating the plot (Tamminen, 1991; Talkkari & Hypén, 1996). Thereafter, the mass of SOM is regressed against the prevailing temperature sum of the plot by the site types. These values are used in the simulation initialization. The values are further used in calculating the initial amount of nitrogen in soil, applying the values of the total nitrogen concentration of the humus layer by site type and tree species, as presented by Tamminen (1991).

In the Sima model, the decomposition of litter determines the weight loss, nitrogen immobilization, lignin decay, and CO2 loss from the decomposing litter cohort (Kellomäki et al., 1992). The decomposition of humus determines nitrogen mineralization, weight loss from humus, and CO2 loss from humus. In the calculations, a carbon content of 50% of the weight was assumed, and litter and humus were treated as cohorts. The litter cohort indicates the amount of dead material originating from trees and ground vegetation on annual basis. Litter is divided into foliage, twig, root, and woody litter (stems of dead trees). Woody litter is divided into stems of standing and fallen dead trees.

Input data and management scenarios

The tree data utilized in the model simulations were based on the NFI9 (1996–2003). The measurements in the inventory were undertaken in systematically located rectangular or L-shaped clusters, each cluster containing 10–18 sample plots. The distance between the clusters varied from 6 × 6 km in the southernmost part of the country to 10 × 10 km in northern Finland. For our study, data from one sample plot per cluster were used to represent variables such as tree species, diameter at breast height, site type, and location. The simulations included only sample plots in upland mineral soil sites.

In the simulations, management operations included energy biomass thinning (EBT), commercial thinning, final felling (FF), and regeneration. The thinning rules, based on the development of basal area and dominant height, followed those currently recommended for different tree species, site fertility types, and regions in Finland (Tapio, 2006). Thinning was accomplished from below to such a level that the remaining basal area was reduced to the expected value at a given dominant height. EBT was also undertaken based on site and species-specific recommendations for the dominant stand height, basal area, and remaining stand density. EBT followed the published recommendations for harvesting and growing of bioenergy in Finnish forestry (Äijälä et al., 2010). FF was undertaken whenever the mean diameter of the trees or stand age in the plots exceeded the given value used for indicating the maturity of the tree population for regeneration. The stand density of 2000 seedlings per hectare was utilized for regeneration of the stands.

Model performance

The Sima model validation has been previously discussed in detail in Kolström (1998) and Kellomäki et al. (2008). Furthermore, recently Routa et al. (2011b) compared the growth of parallel simulations using both the Sima model and the Motti model (Hynynen et al., 2002). The Motti model is a statistical growth and yield model in which tree growth estimation is based on data from a large number of forest inventory sample plots over the whole territory of Finland. This comparison (Routa et al., 2011b) showed that there is a fairly good correlation between the corresponding simulated values for the Motti and Sima simulators for different tree species. The Sima simulator would appear to underestimate slightly (10–20%) the growth when compared with the Motti simulator. Furthermore, the analysis in which the performance of the Sima model was analyzed using data from 10 Forest Centres in southern Finland based on NFI measurements (Finnish Forest Research Institute, 2005) showed a close correlation between the measured and simulated growths (Routa et al., 2011b).

Layout of simulations

The Sima model produces the annual growth (stem, branches, foliage, coarse roots, and fine roots) (Cseq) and the amount of biomass harvested (Charv) in EBT (energy biomass in terms of foliage, branches, and stems), in the commercial thinnings (timber), and in FF (energy biomass and timber). In the case of needles, the loss in harvesting was assumed to be 30%. Furthermore, the model simulations produce the annual litter fall for the decomposition and consequent emissions of carbon from soil (Cdecomp) to be used in the LCA tool. In the LCA tool, technosystem emissions (Cman) were produced according to the output of the Sima model (see Kilpeläinen et al., 2011). In the calculations, wood density of 400 kg m−3 was utilized and carbon content of biomass was assumed to be 50%. The energy content of 3.2 MWh in Mg (dry biomass) was utilized.

The annual net CO2 exchange values were calculated for traditional TP and BP regimes over a 90 year simulation period. During this period, no change in climatic conditions and consequent direct effects on forest growth and development were assumed. The basic stand treatment for this study was TP in which bioenergy was not harvested, but only timber. Energy biomass was harvested integrated with timber in BP. Energy biomass was harvested in EBT (small-sized trees) and in FF (logging residues, needles, branches, tops of the stems, large roots, and stumps).

Radiative forcing calculations

The atmospheric impact of forest BP and utilization was calculated by comparing the net CO2 exchange (see Eqn 3) of BP regime to that of the TP regime over time. TP described the nonutilization situation in which only timber was produced. BP described the utilization situation and covered BP and utilization along with TP. The utilized unit is W m−2 a−1. In the comparison of forest bioenergy to coal, coal emission was assumed at 95 gCOMJ−1.

display math(3)

The net climatic impact, I, is approached by subtracting the climatic impact induced by the TP regime (Inu) from the climatic impact of forest BP and utilization induced by the BP regime (Iu).

The results were expressed as the RF (Eqn 4), which is a methodology to quantify impacts of greenhouse gases in the atmosphere (Ramaswamy et al., 2001).

display math(4)

In Eqn (4), α is 5.35 W m−2 and C is CO2 in ppm. C0 is the reference CO2 concentration. The atmospheric CO2 concentration was assumed to increase linearly from 372 to 817 ppm at the end of the simulation period (IPCC SRES A2). The estimation of the RF is conducted through the change in the net CO2 exchange in BP and TP over time (Eqn 5):

display math(5)

where τ is time, t is the time period in question (i.e. 90 years), and E is the change of CO2 concentration in the atmosphere (in kg). The function f is the lifetime function of CO2, expressing the decay of a pulse of CO2 with time (Eqn 6):

display math(6)

where a0, a1, a2, a3 are 0.217, 0.259, 0.338, and 0.186, respectively. τ1, τ2, and τ3 are 172.9, 18.5, and 1.2, respectively (IPCC, 2007).


BP potential

In the BP, bioenergy was produced in EBT and in FF. The amount of bioenergy produced changed annually and peaked in 2028. At its highest, 156 PJ of bioenergy was produced annually. On average, over the whole simulation period 78 PJ a−1 was produced (Fig. 1). This corresponds to about 17 million m3 a−1.

Figure 1.

Annual energy biomass production potential in PJ over the 90-year simulation period.

Net CO2 exchange in BP and TP

The net CO2 exchange between BP and TP is shown in Fig. 2. At the beginning of the period, there was a net sequestration of CO2 in the forests in both regimes. The net sequestration was higher in TP compared with BP. After about 20 years of simulation, forests became net sources of CO2. After 40 years of simulation, the net exchanges began to decrease and BP emitted less CO2 than TP. This continued for 40 years, and after 80 years of simulation, emissions started to increase, being at that point higher in BP than in TP.

Figure 2.

Annual net CO2 exchange in bioenergy production (BP) and timber production (TP) regimes. See Eqn (1) for the calculation of the net CO2 exchange.

Radiative forcing of BP and utilization

The atmospheric impact of forest BP and utilization over the whole of Finland was expressed by marginal RF (Fig. 3). RF increased from the beginning of the simulation and reached its maximum value at 29 years; 0.17 mW m−2 a−1. Thereafter, marginal RF of BP and utilization decreased to a smoothed rate of under 0.04 mW m−2 a−1 after 70 years of simulation.

Figure 3.

Marginal radiative forcing of forest bioenergy production and utilization over the whole Finland.

Cumulative RF over the whole of Finland calculated from the marginal RF per produced PJ of energy is shown in Fig. 4. It reached 0.23 mW m−2a−1 in 2071 and was 0.10 mW m−2a−1 at the end of the 90-year simulation period. Production and utilization of forest bioenergy was also compared with utilization of coal in constant production of energy (PJ a−1). This showed that the RF was higher in forest BP and utilization than that of coal at the beginning of the simulation period and for the period between years 65 and 80. Over the whole simulation period, RF was about 19% lower for forest bioenergy than that for utilization of coal in energy production.

Figure 4.

Cumulative radiative forcing of forest bioenergy production and utilization over the whole Finland compared with that for coal energy. Values are expressed per PJ of produced energy.


The net atmospheric impacts of forest BP and utilization were analyzed in a dynamic manner with the help of the LCA tool (Kilpeläinen et al., 2011). The annual net CO2 exchange was studied over a 90-year period for traditional TP and BP regimes based on NFI data, with consequent bioenergy and timber potentials. The simulations of growth and development of forest stands were completed for the whole country, and the difference in the net CO2 exchanges between the BP and TP regime was considered the net atmospheric impact of forest BP and utilization. This was finally expressed in terms of RF. The LCA approach (i.e. cradle to gate) considered all the annual flows of CO2 into and from the atmosphere and ecosystems (Kilpeläinen et al., 2011). Bioenergy was harvested from EBT (small-sized trees) and from FF (logging residues, including needles, branches, stumps, and roots) and timber was harvested integrated with bioenergy. Thus, the main deviation in the CO2 emissions between the two regimes originated from the combustion of biomass and the consequent changes in decomposition of SOM in the longer term.

The CO2 exchange values over the whole of Finland showed that, at the beginning of the simulation, forests acted as sinks of CO2 in both regimes. This sink was formed due to the large share of young and middle-aged stands coupled with the high growth rate of trees (Peltola, 2007; Alam et al., 2010). The sink turned into a source of CO2 in the both management regimes in 2020, and the source was at its highest, 340 gCO2 m−2 a−1, in 2040. Due to instant combustion of energy biomass and consequent release of CO2, the net exchange was logically higher in the BP than in the TP regime until the middle of the simulation period. In general, the shift to becoming a CO2 source reflected the decrease in CO2 sequestration due to the increasing share of recently harvested and seedling stands that are acting as sources of CO2, and an increase of emissions from degradation of wood products. From 2040 onwards, the net exchange started to decrease in both regimes and became higher in the TP than in the BP regime, when the last decade of the simulation period was excluded from the analysis. Accordingly, this change in forest age structure over time could be seen in annual bioenergy and TP potential, which increased at the beginning, peaked in 2028, but then decreased till 2080. Since the amount of bioenergy from logging residues in FF is greater compared with that from EBT, combustion-related CO2 emissions also decreased until the end of the simulation period. Earlier, the effect of age structure on forest growth and BP potential in Finland was also highlighted by Alam et al. (2010), who found that the effect of forest structure on forest growth and BP and TP in Finnish conditions, is more pronounced than that of climate change, for example.

The forest management followed the rules recommended for Finnish forestry (Tapio, 2006; Äijälä et al., 2010) in both the BP and TP regimes, even though changes in forest management can enhance the net CO2 exchange of forest production, according to recent results (Alam et al., 2011; Routa et al., 2011a). In addition, Alam et al. (2010) have noted that increasing the thinning thresholds to 30% will simultaneously increase carbon sequestration and energy biomass and TP. Despite these results, the management regimes in our study were kept as similar as possible, to concentrate on the effects of BP and utilization of existing and developing forest resources. Furthermore, rotation length (about 80–100 years) followed in this study is widely used in Finnish forestry to maximize the production of sawlogs. However, these rotation lengths have also been found to be fairly beneficial for reducing the emissions of BP at stand level simulations in Norway spruce, the main contributor of forest bioenergy in Finnish conditions (Routa et al., 2011a).

In addition to the existence of BP and utilization, the only difference in forest management operations between the regimes was the EBT/first thinning as the first operation. However, EBT (harvesting of only energy biomass), done at an earlier phase than first thinning, has not been found to cause major changes in the future stand development as regards to carbon sequestration and harvested timber yields (Alam et al., 2011). Furthermore, the amount of energy biomass harvested in EBT is relatively small compared to that from logging residues from FF. In the other management operations, timber was produced in thinning and FFs, and utilization of wood products was assumed to release CO2 according to specific time-spans of wood-based products in use (see Kilpeläinen et al., 2011) in both regimes. While the utilization of bioenergy worsened the net CO2 exchange at the early phase of the simulation period, degradation of wood-based products caused more delayed emissions in both regimes in the second half of the model period, due to increased timber harvesting. However, the main CO2 contributor in TP, and the cause of the difference when compared to BP at the second half of the simulation, was the increased decomposition of logging residues. These results show the importance of time dependence when considering dynamic forest ecosystems in BP, greenhouse gas reduction, and climate change mitigation (Sathre & Gustavsson, 2011). On the one hand, the results emphasize the dualistic role and possibilities of forest management in controlling the build and release of carbon into and from the stocks and, on the other hand, in controlling the rate of the build speed, i.e. growth. The importance of existing and developing stocks has also been noted as important when considering the substitution effects of wood in construction (e.g. Gustavsson & Sathre, 2011).

The technosystem emissions related to production and utilization (e.g. chipping) followed the management operations and transportation emissions and were low in comparison to ecosystem flows and emissions from combustion of biomass. Their effect in the calculation of atmospheric impact was marginal, since the difference in management-related emissions between the TP and BP regimes differed only slightly. Earlier, management-related emissions have ranged from 2% to 3% from the produced energy and 4–20 kg CO2 MW h−1 (Korpilahti, 1998; Mälkki & Virtanen, 2003; Wihersaari, 2005; Alam et al., 2011; Kilpeläinen et al., 2011).

As the difference of the net CO2 exchange values of the BP and TP regimes was converted to marginal RF of forest BP and utilization, the figures showed increasing net atmospheric impacts during the first 30 years of simulations. Thereafter, the impact started to decrease and remained below 0.05 mW m−2 a−1 at the end of the simulation period. When expressed as cumulative values, RF showed a constant increase in forest BP and utilization over the whole of Finland. At the same time, bioenergy potential ranged between 32 and 156 PJ a−1 during the simulation period, in agreement with earlier estimations (Alam et al., 2008, 2010; Kärkkäinen et al., 2008). When cumulative values were further converted per produced PJ, the values varied from 0.0003 to 0.23 mW m−2 a−1 PJ−1. Earlier, Liski et al. (2011) reported RF values of constant energy production (PJ a−1) less than 0.2 mW m−2 for different components of energy biomass at the end of the century, in line with our calculations. When compared to coal energy, cumulative RF was less in forest bioenergy utilization over the whole simulation period, except for the years at the beginning of the simulation (i.e. the years 1–11), and between the years 65 and 80. Over the whole simulation period, the RF was on average 19% less in BP and utilization compared to that for coal energy.

The calculation of net atmospheric impacts revealed the key role of net greenhouse gas balances, their dynamics in BP and in assessing the total ecological efficiency and substitution capacity of various biomasses. Assessing the role and possibilities of the forests in climate change mitigation, the effects of changing climate and forest management on growth of forest and decomposition of SOM should be also considered carefully in the future. These effects on the net CO2 exchange will be highly important in approaching net climate impacts of energy biomass production and utilizations, and also in assessing changes in land use, fossil fuel substitution impacts, and climate change mitigation options. This helps in identifying optimized BP and management systems and revealing the most appropriate ways to produce and utilize bioenergy and jointly reduce greenhouse gas emissions in the future.