Forest Bioenergy or Forest Carbon? Assessing Trade-Offs in

Forest Bioenergy or Forest Carbon?

Assessing Trade-Offs in Greenhouse

Gas Mitigation with Wood-Based

Fuels

JON MCKECHNIE,

†

STEVE COLOMBO,

‡

JIAXIN CHEN,

‡

WARREN MABEE,

AND

HEATHER L. MACLEAN*

,†,|

Department of Civil Engineering, University of Toronto, 35 St.

George Street, Toronto, Ontario M5S 1A4, Canada, Ontario

Forest Research Institute, 1235 Queen Street East, Sault Ste.

Marie, Ontario P6A 2E5, Canada, School of Policy Studies and

Deptartment of Geography, Queen’s University, 423-138

Union St. Kingston, Ontario K7L 3N6, Canada, and

Department of Chemical Engineering & Applied Chemistry,

School of Public Policy and Governance, University of

Toronto, Toronto, Ontario M5S 1A4, Canada

Received July 15, 2010. Revised manuscript received

November 11, 2010. Accepted November 18, 2010.

The potential of forest-based bioenergy to reduce greenhouse

gas (GHG) emissions when displacing fossil-based energy

must be balanced with forest carbon implications related to

biomass harvest. We integrate life cycle assessment (LCA) and

forest carbon analysis to assess total GHG emissions of

forest bioenergy over time. Application of the method to case

studies of wood pellet and ethanol production from forest

biomass reveals a substantial reduction in forest carbon due

to bioenergy production. For all cases, harvest-related

forest carbon reductions and associated GHG emissions

initially exceed avoided fossil fuel-related emissions, temporarily

increasing overall emissions. In the long term, electricity

generation from pellets reduces overall emissions relative to

coal, although forest carbon losses delay net GHG mitigation by

16-38 years, depending on biomass source (harvest residues/

standing trees). Ethanol produced from standing trees

increases overall emissions throughout 100 years of continuous

production: ethanol from residues achieves reductions after

a 74 year delay. Forest carbon more signiﬁcantly affects bioenergy

emissions when biomass is sourced from standing trees

compared to residues and when less GHG-intensive fuels are

displaced. In all cases, forest carbon dynamics are signiﬁcant.

Although study results are not generalizable to all forests, we

suggest the integrated LCA/forest carbon approach be

undertaken for bioenergy studies.

Introduction

Forests can contribute to greenhouse gas (GHG) mitigation

strategies through capturing and storing atmospheric CO

in live biomass, dead organic matter, and soil pools, supplying

a source for wood products that both stores carbon and can

displace more GHG-intensive alternatives, and providing a

feedstock for bioenergy to displace fossil fuel use. While the

merit of each of these options has been individually

investigated, trade-offs associated with forest resource

utilization decisions must also be considered. Of particular

interest is the relationship between harvest and forest carbon

storage and how this impacts the GHG mitigation perfor-

mance of forest products, including bioenergy. Existing tools

employed to evaluate emissions associated with different

forest resource use decisions are not individually well suited

to considering such interactions.

Life cycle assessment (LCA) has been applied to bioenergy

options, including electricity generation and transportation

fuels. The GHG mitigation potential of bioenergy products

depends on activities throughout the entire life cycle (LC),

making such a perspective necessary for a comprehensive

evaluation. Numerous LCAs have focused on agricultural

biomass as feedstock for bioenergy, e.g., reviewed in ref (1).

Comparatively few LCAs have evaluated bioenergy from forest

biomass; those that have examined electricity generation (e.g.,

ref (2)), heating (e.g., ref (3)), and transportation (e.g., ref

(4)). Bioenergy LCAs have generally found that the substitu-

tion of fossil fuel-derived energy with biomass-derived

alternatives reduces GHG emissions, owing in part to the

assumption that biomass-based CO

emissions do not

increase atmospheric CO

Conventional wisdom has generally accepted this as-

sumption of biomass ‘carbon neutrality’, and thus, most of

the LC GHG emissions associated with bioenergy production

are attributed to fossil carbon inputs into the system (5). In

practice, however, the assumption of carbon neutrality may

not accurately represent carbon cycling related to biomass

growth (e.g., ref (6)). The practice of annual or semiannual

harvest in agriculture means that carbon uptake by biomass

may reasonably match carbon release in bioenergy systems

within a short time frame, although land use change impacts

resulting from biomass production can upset this balance

(7). In temperate forests, the harvest cycle can range from

60 to 100 or more years due to the relatively slow growth of

forest species. It could therefore take a century for carbon

stocks to be replaced, particularly under a clearcutting regime

(harvest of all merchantable trees). Harvest patterns and

associated implications for forest carbon stocks vary exten-

sively, ranging from clearcuts to variable retention patterns,

including shelterwood and selection cuts. Some variable

retention approaches may actually increase forest regenera-

tion, increasing the potential to recover carbon (8). Bioenergy

production from harvest residues (tree tops and branches)

also impacts forest carbon stocks; left uncollected, residues

continue to store carbon until released by decomposition or

treatment for forest regeneration. While sustainable forest

management should ensure that harvest does not impair the

long-term productivity of forests, harvest and other forest

management activities clearly impact present and future

forest carbon stocks. LCA, in its current form, is not well

suited to consider the complexities of forest carbon dynamics.

Forest carbon studies have weighed the carbon balance

of harvest with the GHG mitigation potential of forest

products (e.g., refs 9-11). Some studies have utilized

sophisticated forest carbon models to track changes in carbon

stored in living biomass (above ground and below ground),

dead organic matter, and soil pools (e.g., refs 12, 13). These

studies, however, generally employ simpliﬁed assumptions

regarding the GHG emissions of forest products (including

bioenergy) and have not incorporated a full LC approach.

Given the dependence of emissions on speciﬁc system

* Corresponding author phone: (416) 946-5056, fax: (416) 978-

3674, e-mail: [email protected].

†

Department of Civil Engineering, University of Toronto.

‡

Ontario Forest Research Institute.

Queen’s University.

School of Public Policy and Governance, University of Toronto.

Environ. Sci. Technol. 2011, 45, 789–795

10.1021/es1024004  2011 American Chemical Society VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

789

Published on Web 12/10/2010

characteristics (e.g., biomass source, bioenergy production

process, fuel displaced), generalized assumptions regarding

the GHG mitigation potential of bioenergy are inadequate

for informing decision making and public policies.

State-of-the-art tools are available for independently

evaluating both the LC emissions of bioenergy systems and

forest carbon dynamics. Using these methods in isolation,

as has been general practice, stops short of the comprehensive

evaluation needed to properly assess the GHG emissions of

forest products. In an assessment of GHG mitigation

performance of structural wood products, ref (14) incorpo-

rated LCA with an analysis of forest carbon dynamics. While

the study did not consider bioenergy as a product, the results

illustrate the importance of considering forest carbon and

LC emissions simultaneously when evaluating forest prod-

ucts. Applied to bioenergy, integrating LCA with forest carbon

modeling would improve understanding of potential con-

tributions to climate change mitigation.

Bioenergy has been treated inconsistently across energy

and climate change policy initiatives in terms of how (or if)

GHG emissions are quantiﬁed. Forest bioenergy policies that

ignore carbon ﬂows in the forest may prove ineffective at

achieving actual emissions reductions (15). Exclusion of forest

carbon from current initiatives is in part due to data issues,

although emerging guidelines may ameliorate this situation

(16). Tools that are able to synthesize forest carbon data and

LCA and evaluate trade-offs between bioenergy and forest

carbon remain to be developed.

Forest bioenergy has the potential to signiﬁcantly reduce

GHG emissions compared with fossil fuel alternatives.

However, interactions between biomass harvest and forest

carbon and the resulting effect on the GHG mitigation

performance of bioenergy systems are inadequately under-

stood. The objectives of this study are to demonstrate the

integration of LCA and forest carbon modeling to assess the

total GHG emissions (referred to as “emissions”) of forest-

based bioenergy options and to determine how emissions

reductions associated with bioenergy are impacted when

forest carbon is taken into account. We demonstrate this

approach through a case study investigating two bioenergy

products (wood pellets, referred to as pellets, and ethanol)

from two biomass sources (standing trees and harvest

residues, referred to as residues) within the Great Lakes-St.

Lawrence (GLSL) forest region of Ontario, Canada.

Methods

We develop a framework integrating two analysis tools: life

cycle inventory (LCI) analysis and forest carbon modeling.

See Supporting Information for additional detail on all

methods. LCI analysis quantiﬁes emissions related to the

production and use of forest biomass-derived energy. The

LCI is based on the assumption of immediate biomass carbon

neutrality, as is common practice, and is therefore employed

to quantify the impact of all emissions on atmospheric GHGs

with the exception of biomass-based CO

Forest carbon modeling quantiﬁes the impact of biomass

harvest on forest carbon dynamics, permitting an evaluation

of the validity of the immediate carbon neutrality assumption.

If biomass-based CO

is fully compensated for by forest

regrowth, biomass harvest will have no impact on forest

carbon stocks. Reduced forest carbon indicates that a portion

of biomass-based CO

emissions contributes to increased

atmospheric GHGs and should be attributed to the bioenergy

pathway. The total emissions associated with a bioenergy

system are the sum of the two sets of GHG ﬂows (those

resulting from the LCI and those from the forest carbon

analysis)

GHG

Tot

(t) ) ∆FC(t) + GHG

Bio

(t) (1)

where GHG

Tot

(t) is the total emissions associated with

bioenergy, ∆FC(t) is the change in forest carbon due to

biomass harvest for bioenergy, and GHG

Bio

(t) is the GHG

emissions associated with bioenergy substitution for a fossil

fuel alternative [all reported in metric tonne CO

equivlent

(tCO

equiv)] at time t.

The change in forest carbon, ∆FC(t), is the difference in

forest carbon stocks between harvest scenarios: those ‘with’

and ‘without’ bioenergy production. While we present this

as a single parameter in eq 1, in reality forest carbon models

consider the complexity of carbon ﬂuxes between pools

within the forest and between the forest and atmosphere.

Carbon in biomass harvested for bioenergy is assumed to be

immediately released to the atmosphere. However, forest

regrowth will capture and store atmospheric CO

over time.

There is therefore a time dependency to the carbon impact

of forest harvest for bioenergy. Assessing the change in forest

carbon requires consideration of the forest response following

harvest and the fate of the biomass source if it is not harvested

for bioenergy (standing trees could be harvested for other

uses or never harvested; residues could decompose on site,

be burned as part of site preparation, or be collected for

other uses). Local conditions inﬂuence such factors and must

inform speciﬁc applications of this method. Information

relevant to the current case study is provided in the following

methods subsection.

LCI quantiﬁes emissions associated with all activities from

initial resource extraction and fuel production through to

the use of fuels, inclusive of transportation and distribution

stages. Emissions related to the production of inputs are

included based on their cradle-to-grave activities. Comparing

emissions of a bioenergy product with the relevant reference

fossil fuel alternative(s) determines the bioenergy GHG

mitigation performance. The output of the bioenergy LCI

models, emissions per functional unit, is not directly

compatible with the output of forest carbon models, which

quantify carbon stocks over relatively long time periods (e.g.,

100 years) in order to fully capture the impact of management

decisions. To integrate the assessment tools, we quantify the

cumulative emissions associated with bioenergy production

within the time period investigated with the forest carbon

model (e.g., 100 years), considering GHG mitigation from

fossil fuel displacement to be permanent. LCI results are

converted to a quantity of emissions by

GHG

Bio

(t) )

∫

(t) × GHG

dt (2)

where GHG

Bio

(t) represents emissions associated with bioen-

ergy substitution for fossil fuel alternative(s) at time t

(tCO

equiv), Q

(t) is the quantity of biomass used to produce

bioenergy product i at time t (e.g., oven dry tonne (odt)

biomass/year), and GHG

is the emissions associated with

bioenergy product i per unit biomass (tCO

equiv/odt).

Summing the bioenergy emissions (based on the LCI results)

and the forest carbon emissions gives the total emissions of

bioenergy utilization over time as shown in eq 1.

Considering emissions over a long time period is relevant

to the carbon dynamics of a forest; however, this introduces

uncertainty regarding future forest conditions, markets, and

the performance of the energy systems investigated. The LCI

and forest carbon analysis in this research consider that these

conditions remain static throughout the time frame due to

the difﬁculty of deriving reasonable estimates for these

parameters. These issues are further examined in the Results

and Discussion.

Application of LCI/Forest Carbon Model framework. We

apply the above framework to investigate the impact of forest

carbon dynamics on the total emissions associated with

several forest-based bioenergy pathways. Forest biomass is

assumed to be procured for the production of fuels for

790

9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011

electricity generation and light-duty vehicle (LDV) trans-

portation. Reference models are also developed for con-

ventional fuel sources to which the bioenergy pathways are

compared. We examine emissions of selected GHGs (CO

O), reported as CO

equiv based on 100 year global

warming potentials (17). See the Supporting Information for

additional case study details and data.

The pathways considered are as follows. (1) Electricity

generation: (a) Reference coal: production of electricity from

coal at an existing generating station (GS) in Ontario; (b)

Pellet coﬁring, harvest residue: production of electricity at

20% coﬁring rate (energy input basis) at retroﬁt coal GS,

pellets produced from residues; (c) Pellet coﬁring, standing

tree: production of electricity at 20% coﬁring rate (energy

input basis) at a retroﬁt coal GS, pellets produced from

standing trees. (2) Transportation: (a) Reference gasoline:

gasoline use in LDV; (b) E85, harvest residue: ethanol/gasoline

blended fuel use in LDV, ethanol produced from residues

(biomass is not pelletized); (c) E85, standing tree: ethanol/

gasoline blended fuel (85% ethanol by volume) use in LDV,

ethanol produced from standing trees (biomass is not

pelletized).

Biomass Sources. Biomass is supplied from standing trees

and residues from 5.25 million hectares within the GLSL forest

region in Ontario. This area represents 19% of provincially

owned forest managed for timber production. Trees allocated

for harvest that are not currently utilized for traditional

products could serve as a source of biomass for bioenergy

applications without impacting markets for conventional

wood products. Residues do not have a useful purpose in the

region’s conventional forest products industry and are left

to decompose in the forest. Competition for limited wood

resources can result in diversion from current uses (e.g., pulp)

to bioenergy (18) with potential indirect emissions conse-

quences (7). By limiting the present study to biomass sources

unutilized for conventional products, we avoid such market

interactions.

Standing tree harvest and related forest operations

(regeneration, road construction/maintenance, and transport

to the pellet/ethanol facility) are assessed using a model

developed in our previous work (6). Emissions related to

residue collection are calculated by treating the residues as

a byproduct of forest harvest. Only additional fuel use

required for collection beyond that of current harvest

operations is allocated to the residues; other forest operations

are allocated to the primary forest product and are therefore

not included in the present study. Residue collection consists

of roadside chipping and loading.

Electricity Pathways. LCI models representing electricity

generation from coal and coﬁring of pellets from standing

trees were developed in our prior work (6). The models

consider emissions associated with the full fuel LCs from

initial resource extraction through to combustion as well as

upstream emissions related to process inputs. One kWh is

selected as the functional unit for the analysis. We assume

that pellet production from residues and their use for coﬁring

is similar to that of pellets from standing trees but modify

the pelletization process to reﬂect that residues are chipped

in the forest (standing trees are delivered as logs). For both

sources, 15% of input biomass is assumed to be consumed

during pellet production to dry the biomass. Avoiding fossil

fuel use reduces emissions during the pelletization process

but increases biomass input to pellet production and

associated forest carbon impacts. Implications of this as-

sumption are considered in Results and Discussion.

Transportation Pathways. Ethanol production, trans-

portation, distribution, and use as E85 fuel in LDV are

modeled based on the wood-to-ethanol biochemical con-

version pathway in the Government of Canada’s “well-to-

wheel” model, GHGenius 3.17 (4). The gasoline portion of

E85 fuel and the reference gasoline pathway are also taken

from GHGenius. The functional unit for the transportation

pathways is 1 km driven. Signiﬁcant uncertainty exists in

evaluating ethanol production from cellulosic feedstock as

technological development and optimization is ongoing and

production not yet at commercial scale (19).

Forest Carbon. The forest carbon dynamics related to

biomass harvest are evaluated using FORCARB-ON, an

Ontario-speciﬁc adaptation of the FORCARB2 model (12).

FORCARB-ON quantiﬁes carbon stocks (in living trees, soil,

standing dead trees, down dead wood, forest ﬂoor, and

understory vegetation pools) based on harvest schedules and

inventories that producers are required to report to the

Province. Harvest schedules take into account species and

age composition of the forest, age classes eligible for harvest,

natural disturbance frequency, growth rates, and forest

succession. The model estimates forest carbon stocks over

100 years, a time frame relevant to the long-term perspective

of forest management planning.

We evaluate forest carbon stocks for three potential harvest

scenarios: (1) “current harvest” baseline, where biomass

(standing trees, residues) is not collected for bioenergy

production and therefore timber is removed solely to satisfy

the current demand for traditional wood products; (2)

“current + residue” harvest, with residue removal for

bioenergy production; and (3) “maximum allowable” harvest,

with additional standing tree harvest (compared to the

baseline) for bioenergy production (residues are not col-

lected). The difference in forest carbon stocks between the

bioenergy production scenarios and “current harvest” base-

line scenario is allocated to the bioenergy products. Additional

standing tree harvest for bioenergy occurs as scheduled under

forest management plans; following harvest, stands are

regenerated by planting or natural regeneration, varying by

site. If not harvested for bioenergy, standing trees eventually

undergo natural succession and are subject to a small

likelihood of natural disturbance. Residue collection is

assumed to not impact soil carbon stocks; uncollected

residues are assumed to decompose on site, either at the

roadside or near where trees were felled. The consequence

of collecting residues for bioenergy production is that this

temporary carbon store is ‘liquidated’ immediately (com-

busted during bioenergy production and use) rather than

decomposing slowly in the forest. Therefore, the associated

change in forest carbon is the difference between immediate

release (bioenergy) and decomposition over time if not

collected. As noted previously, these factors could vary by

location with a potentially signiﬁcant impact on the assessed

forest carbon emissions. We do not consider emissions related

to the current harvest for traditional wood products or their

use. Under the assumptions in this study, this is not affected

by the decision to undertake additional harvest or collect

residues for bioenergy production.

Results and Discussion

Life Cycle Inventory Results, Excluding Forest Carbon. LCI

results for the pathways are shown in Table 1, using the

assumption of immediate biomass carbon neutrality. LCI

emissions for biomass are greater when sourced from

standing trees than from residues. Upstream (fuel production)

stages, however, are minor contributors to LC emissions of

either pellets or ethanol. The majority of emissions arise from

the combustion of fossil fuels, both as the fossil portion during

bioenergy use and in the reference fossil pathways. Excluding

changes in forest carbon, 20% pellet coﬁring reduces LC

emissions by 18% compared to coal-only operation (kWh

basis) whether standing trees or residues are utilized, whereas

an E85-fueled LDV reduces LC emissions by 57% compared

to a gasoline LDV (km-driven basis). The greater emission

reduction of E85 relative to pellet coﬁring gives the appear-

VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9 791

ance that this pathway represents a preferred use of biomass

for reducing emissions, but this results primarily from the

coﬁring scenario utilizing a lower proportion of biomass fuel

(20%, energy basis) than E85 (79%, energy basis).

We convert the LC emissions from their initial functional

units (kWh, km driven) to a basis of one odt of biomass

removed from the forest for bioenergy production (odt

biomass

This makes the LCI and forest carbon model results compat-

ible and facilitates a comparison of the two bioenergy

pathways (electricity, ethanol) in terms of their effectiveness

of biomass utilization in reducing emissions (see Supporting

Information, equation S-3). Over their respective LCs, the

production and use of pellets from standing trees displaces

1.49 tCO

equiv/odt

biomass

, while ethanol production and use

displaces 0.51 tCO

equiv/odt

biomass

, exclusive of forest carbon

impacts. Utilizing residues as a feedstock for pellets and

ethanol displaces 1.50 and 0.53 tCO

equiv/odt

biomass

, respec-

tively. Substitution of coal with pellets provides a greater

mitigation beneﬁt than substitution of gasoline with ethanol,

primarily due to the higher GHG intensity of coal. To put

these values into perspective, the constituent carbon in

biomass is equivalent to 1.83 tCO

equiv/odt. The signiﬁcance

ofreleasingthisbiomass-basedCO

isconsidered subsequently.

Forest Carbon Analysis Results: Impact of Biomass

Harvest. Sustainable biomass sources in the study area could

provide, on average, 1.8 million odt/year from standing trees

and 0.38 million odt/year from residues. Combined, these

sources could provide 2.2% of annual electricity generation

in the province or reduce gasoline consumption by 3.3%

(see Supporting Information). Forest carbon loss due to

undertaking biomass harvest in the study area over a 100

year period is shown in Table 2. For both sources (residues,

standing trees), harvest reduces forest carbon asymptotically

toward a “steady state”. For standing trees, as more stands

are harvested for bioenergy over time, the rate of carbon

accumulation in regrowing stands increases toward a point

where, under ideal conditions, carbon accumulation balances

removals associated with continued harvest. For residues, a

similar steady state is eventually achieved when the rate of

carbon removals at harvest is matched by the expected rate

of residue decomposition if harvest is not undertaken.

Continuing biomass harvest once a steady state has been

reached would not impact forest carbon stocks; however,

initiating biomass harvest beyond current removals has

signiﬁcant emissions consequences in the near to medium

term. Forest carbon loss due to harvest residue collection

approaches a maximum of ∼15MtCO

equiv, whereas stand-

ing tree harvest for bioenergy results in a carbon loss

exceeding 150 MtCO

equiv after 100 years. Proportional to

the quantity of biomass provided, standing tree harvest results

in a greater impact on forest carbon than harvest residue

collection because live trees would generally continue to

sequester carbon if not harvested, whereas carbon in

uncollected residues declines over time.

Total GHG Emissions: Combined LCI and Forest Carbon

Analysis Results. Summing the cumulative emissions of the

bioenergy options (LCI results Figure 1, dashed lines) and

the forest carbon emissions (Figure 1, dotted lines) results

in the total emissions of bioenergy production and use (Figure

1, solid lines). When reductions in forest carbon are included,

emission mitigation is delayed and reduced compared to

the case where immediate biomass carbon neutrality is

assumed. For all scenarios investigated, total emissions from

the bioenergy pathways initially exceed those of the reference

fossil fuel pathways, indicating an initial increase in emissions

resulting from bioenergy use. Emissions associated with forest

carbon loss due to biomass harvest exceed the reduction of

fossil fuel-based emissions provided by bioenergy substitu-

tion. The emissions increase associated with bioenergy,

however, is temporary: the rate of forest carbon loss decreases

with time, whereas the emissions reduction associated with

utilizing bioenergy in place of fossil alternatives continues

to increase throughout the 100 year period, proportional to

the cumulative quantity of pellets or ethanol produced. A

TABLE 1. Life Cycle GHG Emissions Associated with Bioenergy Product (wood pellets, ethanol) Blended for Use and Substitution

for Fossil Reference Pathway

electricity generation pathways transportation pathways

life cycle stage

coal

c,d

(g CO

equiv/kWh)

20% pellet

coﬁring, residue

(g CO

equiv/kWh)

20% pellet

coﬁring, standing tree

(g CO

equiv/kWh)

gasoline

(g CO

equiv/km)

E85, residue

(g CO

equiv/km)

E85, standing tree

(g CO

equiv/km)

forest operations 1.9 4.3 5.1 11.7

bioenergy production, distribution

9.5 9.6 46 46

upstream fossil energy component 62 50 50 77 16 16

fuel use (combustion)

939 760 760 211 48 48

total life cycle emissions 1001 821 824 288 116 123

Values assume immediate carbon neutrality and do not take into consideration forest carbon implications.

Includes

transport of biomass to the production facility, bioenergy production, electricity coproduct credit from biochemical

production of ethanol, and bioenergy transportation/distribution stages.

Reference (6).

Surface coal mining removes

biomass and disturbs soil, which results in GHG emissions due to direct land use change. These emissions along with

other mining process emissions are considered in our analysis.

Fuel use consists of GHG emissions from the fossil

component of fuel (coal, gasoline) and non-CO

GHG emissions associated with bioenergy (pellet, ethanol) combustion.

Reference (4).

TABLE 2. Forest Carbon Impacts of Continuous Biomass Harvest

forest carbon stock change (MtCO

equiv)

year

biomass source 0 10 20 30 40 50 60 70 80 90 100

residues 0

a,b

-8.2 -11.8 -13.0 -13.5 -13.9 -14.3 -14.7 -15.0 -15.2 -15.2

standing trees 0 -43.6 -80.9 -106.3 -112.5 -113.4 -112.7 -132.8 -143.6 -150.8 -150.7

Negative values indicate a GHG emission source (forest carbon stocks are reduced due to biomass harvest) that is

attributable to bioenergy production.

Reported values are the total stock change due to continuous harvest. For example,

50 years of continuous standing tree harvest reduces total forest carbon stocks by 113.4 MtCO

equiv.

792

9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011

time delay therefore exists before bioenergy systems reach

a “break-even” point where total emissions for the bioenergy

and reference fossil pathways are equal. Only after the break-

even point are net emissions reductions achieved.

Figure 1a and 1b shows the total emissions resulting from

continuous use of residues for pellet and ethanol production,

respectively, over a 100 year period. Excluding forest carbon,

the emissions reduction associated with utilizing bioenergy

in place of fossil alternatives increases steadily over time.

The reduction of forest carbon stocks due to residue collection

slows toward a steady state. Co-ﬁring with pellets produced

from residues reduces cumulative emissions relative to coal

only after an initial period of increased emissions lasting 16

years. Forest carbon impacts of residue removal reduce the

total emission mitigation at year 100 from 57 MtCO

equiv

(expected assuming immediate biomass carbon neutrality)

to 42 MtCO

equiv.

Compared to the electricity pathway results, utilization

of residues for ethanol production is more greatly impacted

by changes in forest carbon, due to the lower GHG intensity

of the displaced fuel (gasoline compared to coal). An overall

emission reduction occurs only after 74 years of continuous

production of ethanol; total GHG reductions by year 100 are

reduced by 76% from expected performance assuming

immediate biomass carbon neutrality.

Due to the greater forest carbon impact of standing tree

harvest compared to residue collection, bioenergy production

from standing trees performs worse in terms of reducing

emissions (Figure 1c and 1d). Pellet production from standing

trees results in a greater initial emissions increase, reaching

a break-even point only after 38 years of continuous

production and use when displacing coal for electricity

generation. The total emissions reductions from utilizing

wood pellets from standing trees over a 100 year period,

expected under the assumption of biomass carbon neutrality,

is reduced by 56% when forest carbon impacts are considered.

As in the residue cases, for the standing tree cases forest

carbon more signiﬁcantly impacts total emissions of ethanol

than those associated with pellets for electricity generation.

Ethanol production from standing trees (Figure 1d) does not

reduce emissions at any point within the 100 year period;

instead, overall emissions to the atmosphere increase relative

to the gasoline reference pathway. Disregarding biobased

emissions, as is common to most LCAs, would return

an opposite, and erroneous, result. This contradiction, also

identiﬁed elsewhere (15), illustrates the misleading conse-

quence of assuming immediate biomass carbon neutrality

when quantifying emissions of some bioenergy pathways.

Simply adding biobased CO

emissions associated with

bioenergy production and use to the LCI totals presented in

Table 1 would increase emissions associated with bioenergy.

Pellet coﬁring (at 20%) would result in (all in gCO

equiv/

kWh) 1039 (residue) and 1042 (standing tree) compared to

1001 for coal only. E85 would emit (all in gCO

equiv/km) 711

(residue) and 718 (standing tree) compared to 288 for

gasoline. This approach, however, would not accurately assess

the impact of bioenergy production and use on the atmo-

sphere. By only considering carbon in harvested biomass,

near-term emissions would be underestimated (decomposi-

tion of uncollected biomass, for example, below ground

biomass, is omitted). Mid- to long-term emissions would be

overestimated as compensation for biobased CO

emissions

within the forest (e.g., regrowth) is not considered.

Sensitivity Analysis. A sensitivity analysis is performed

to assess the impact of key sources of uncertainty/variability

in the LCI and forest carbon model parameters on the study

FIGURE 1. Cumulative GHG emissions from continuous biomass harvest for bioenergy production: (a) pellets produced from residues,

displacing coal (20% coﬁring), (b) ethanol produced from residues, displacing gasoline (E85 fuel), (c) pellets produced from standing

trees, displacing coal (20% coﬁring), and (d) ethanol produced from standing trees, displacing gasoline (E85 fuel). Positive values

indicate an increase in GHG emissions to the atmosphere.

VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9 793

results (see Supporting Information). The results are not

sensitive to most parameters, and the general trends of the

impacts of biomass harvest on carbon stocks and their

contribution to overall emissions were not found to be

impacted by uncertainty in the parameters. The pellet

pathway results were found to be most sensitive to assump-

tions related to the quantity of biomass used for drying during

pelletization (15% of input biomass in base case) (see

Supporting Information Figure S-3). Reducing the consump-

tion of biomass during the drying stage increases pellet output

and fossil fuel displacement per unit of input biomass. Co-

location of pelletization facilities with processes generating

waste heat could reduce the drying energy requirement. If

no input biomass is required for drying, there are larger

emissions reductions associated with pellet use and the time

before reaching break even with the fossil energy system is

reduced from 16 to 11 years (residues) and from 38 to 29

years (standing trees). When forest carbon is excluded from

the analysis, biomass utilization for drying energy has a

minimal impact on LC emissions (6).

Study Implications. The simpliﬁed assumption of im-

mediate biomass carbon neutrality has been commonly

employed in bioenergy studies, owing in part to emissions

from the energy and forest sectors being reported separately

in national inventories (17). This study, however, shows that

increasing biomass removals from the forest signiﬁcantly

reduces carbon stocks and delays and lessens the GHG

mitigation potential of the bioenergy pathways studied.

Ignoring the complex relationship between forest carbon

stocks and biomass harvest by employing the carbon

neutrality assumption overstates the GHG mitigation per-

formance of forest bioenergy and fails to report delays in

achieving overall emissions reductions.

Combining LCI analysis and forest carbon modeling as

an analytical approach provides a more accurate represen-

tation of the role of forest bioenergy in GHG mitigation. When

forest carbon dynamics are included in the case study, the

use of forest-based bioenergy increases overall emissions

for many years and, in the worst-performing scenario

(standing tree harvest for ethanol production), does not yield

any net climate mitigation beneﬁt over the 100 year period.

Carbon implications of bioenergy production are not limited

to forests, and these results should not be taken to suggest

that agricultural biomass is inherently preferable. Land use

impacts associated with agriculture-sourced bioenergy can

greatly increase LC emissions (7). Nonbioenergy systems can

also impact carbon stocks (e.g., overburden removal in coal

mining). While the contribution to total emissions may not

be signiﬁcant in all situations, a comprehensive evaluation

of any fossil or renewable system should consider impacts

of life cycle activities on terrestrial carbon stocks.

Do our results support continued reliance on fossil fuels

for electricity generation and transportation? Fossil fuel use

transfers carbon from the Earth’s crust to the atmosphere;

moving beyond reliance on these energy sources is imperative

to address climate change and nonrenewable resource

concerns. Bioenergy offers advantages over other renewable

options that are limited by supply intermittency and/or high

cost. However, effective deployment of bioenergy requires

the thoughtful selection of appropriate pathways to achieve

overall emissions reductions. Harvesting standing trees for

structural wood products has been reported to reduce overall

emissions: storing carbon in wood products and displacing

GHG-intensive materials (steel, concrete) exceeds associated

forest carbon impacts (14). In comparison, using standing

trees for bioenergy immediately transfers carbon to the

atmosphere and provides a relatively smaller GHG beneﬁt

from displacing coal or gasoline, increasing overall emissions

for several decades. Identifying biomass supply scenarios

that minimize forest carbon loss will improve the emission

mitigation performance of forest bioenergy. Residues em-

ployed for bioenergy reduce emissions from coal after a much

smaller delay than standing trees, while other forest biomass

sources (e.g., processing residuals) could offer near-term

emission reductions if used to replace GHG-intensive fossil

fuels. Industrial ecology approaches (e.g., utilizing end-of-

life wood products as a biomass source; integrating bioenergy

production with other wood products to utilize waste heat

for processing) could reduce forest carbon implications of

bioenergyproduction andare deservingoffurther consideration.

Utilizing bioenergy to displace the most GHG-intensive

fossil fuels minimizes initial emissions increases and reduces

the time required before net GHG beneﬁts are achieved.

Ethanol production for gasoline displacement, under the

modeled conditions, is not an effective use of forest biomass

for GHG reductions. Displacing coal in electricity generation,

in comparison, is superior in reducing emissions. However,

this does not indicate that electricity applications are always

preferable. The mitigation performance of biomass-derived

electricity depends on the displaced generation source.

Further, these results represent the expected near-term state

of energy system technologies and do not consider changes

in either the reference or the bioenergy pathways over the

time frame studied. Performance improvements are inevi-

table with technological maturation and commercialization.

Technological developments regarding thermal electricity

generation (e.g., efﬁciency improvements; viable carbon

capture and storage) would be applicable to both biomass

and coal, while improvements in pellet production would

not greatly inﬂuence total emissions. Emissions from pro-

ducing ethanol, regarding both the ethanol production

process and the appropriate reference pathway in the future

given the limited petroleum supply and associated price

volatility, is uncertain and in the future could prove a more

effective means of emissions reductions than reported here.

Ethanol can also play an important role in addressing

economic and energy security concerns related to petroleum

dependency.

Although the method demonstrated in this research is

generalizable, site-speciﬁc characteristics of forests prevent

the generalization of speciﬁc results from this study. Numer-

ous factors would inﬂuence forest carbon dynamics and must

be considered in speciﬁc analyses. Intensifying silvicultural

practices (e.g., planting instead of natural regeneration,

utilization of fast-growing species) could shorten, but not

eliminate, the period of net emission increase found in our

results. In some jurisdictions, residues are burned during

site preparation for forest regrowth. Using such residues for

bioenergy would not signiﬁcantly impact forest carbon stocks.

While GHG mitigation is an important consideration of

forest resource utilization, numerous other factors must be

considered in the decision-making process. In particular,

declines in Ontario’s forest sector have negatively impacted

communities that would welcome the investment and

employment opportunities associated with bioenergy. Other

environmental factors and technical constraints must be

considered before implementing bioenergy production.

The potential of forest-based bioenergy to reduce emis-

sions from fossil fuels must be balanced with forest carbon

impacts of biomass procurement. This perspective is of

particular importance as policies related to climate change

mitigation, deployment of renewable energy, and the forest

bioeconomy are developed and implemented. Considering

bioenergy in isolation of its impact on forest carbon could

inadvertently encourage the transfer of emissions from the

energy sector to the forest sector rather than achieve real

reductions. Accounting methods must be designed to

measure the complete impact of mitigation options on the

atmosphere. By considering the broader impacts of bioenergy

production on the forest, particularly forest carbon pools,

794

9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 2, 2011

policy can lend support to effective uses of forest resources

for climate change mitigation.

Acknowledgments

This research was supported by the Ontario Ministry of

Natural Resources and the Natural Sciences and Engineering

Research Council. We thank Michael Ter-Mikaelian and Denis

Cormier for data and insights for the study.

Supporting Information Available

Additional detail on biomass sources, life cycle inventory of

bioenergy systems, forest carbon analysis, and additional

results and discussion. This material is available free of charge

via the Internet at http://pubs.acs.org.

Literature Cited

(1) Larson, E. D. A review of life-cycle analysis studies on liquid

biofuel systems for the transport sector. Energy Sustainable Dev.

2006, X (2), 109-126; DOI:10.1016/S0973-0826(08)60536-0.

(2) Damen, K.; Faaij, A. A greenhouse gas balance of two existing

international biomass import chains. Mitigation Adapt. Strate-

gies Global Change 2006, 11 (5-6), 1023-1050; DOI:10.1007/

s11027-006-9032-y.

(3) Gustavsson, L.; Karlsson, A. A System Perspective on the Heating

of Detached Houses. Energy Policy 2002, 30, 553-574; DOI:

10.1016/S0301-4215(01)00128-8.

(4) GHGenius-a model for lifecycle assessment of transportation fuels,

version 3.17; National Resources Canada: Ottawa, Ontario, Canada,

2010; available at http://www.ghgenius.ca/downloads.php.

(5) Cherubini, F.; Bird, N. D.; Cowie, A.; Jungmeier, G.; Schla-

madinger, B.; Woess-Gallasch, S. Energy- and greenhouse gas-

based LCA of biofuel and bioenergy systems: Key issues, ranges,

and recommendations. Resour. Conserv. Recycl. 2009, 53, 434-

447; DOI:10.1016/j.resconrec.2009.03.013.

(6) Zhang, Y.; McKechnie, J.; CormierD. Lyng, R.; Mabee, W.; Ogino,

A.; MacLean, H. L. Life cycle emissions and cost of producing

electricity from coal, natural gas, and wood pellets in Ontario,

Canada. Environ. Sci. Technol. 2010, 44, 538-544; DOI:10.1021/

es902555a.

(7) Searchinger, T.; Heimlich, R.; HoughtonR. A. Dong, F.; Elobeid,

A.; Fabiosa, J.; Tokgoz, S.; Hayes, D.; Yu, T. Use of US croplands

for biofuels increases greenhouse gases through emissions from

land-use change. Science 2008, 319, 1238-1240; DOI:10.1126/

science.1151861.

(8) Archambault, L.; Delisle, C.; Larocque, G. R. Forest regeneration

50 years following partial cutting in mixedwood ecosystems of

southern Quebec, Canada. Forest Ecol. Manag. 2008, 257 (2),

703-711; DOI:10.1016/j.foreco.2008.09.056.

(9) Schlamadinger, B.; Marland, G. The Role of Forest and Bioenergy

Strategies in the Global Carbon Cycle. Biomass Bioenergy 1996,

10, 275-300; DOI:10.1016/0961-9534(95)00113-1.

(10) Hennigar, C. R.; MacLean, D. A.; Amos-Binks, L. J. A novel

approach to optimize management strategies for carbon stored

in both forests and wood products. Forest Ecol. Manag. 2008,

256, 786-797; DOI:10.1016/j.foreco.2008.05.037.

(11) Biomass sustainability and carbon policy study; Manomet Center

for Conservation Sciences: Manomet, MA, 2010; available at http://

manomet.org/sites/manomet.org/ﬁles/Manomet_Biomass_Report_

Full_LoRez.pdf.

(12) Chen, J.; Colombo, S. J.; Ter-Mikaelian, M. T.; Heath, L. S. Future

carbon storage in harvested wood products from Ontario’s

Crown forests. Can. J. Forest Res. 2008, 38, 1947-1958; DOI:

10.1139/X08-046.

(13) Kurz, W. A.; Dymond, C. C.; White, T. M.; Stinson, G.; Shaw,

C. H.; Rampley, G. J.; Smyth, C.; Simpson, B. N.; Neilson, E. T.;

Trofymow, J. A.; Metsaranta, J.; Apps, M. J. CBM-CFS3: a model

of carbon-dynamics in forestry and land-use change imple-

menting IPCC standards. Ecol. Model. 2009, 220, 480-504; DOI:

10.1016/j.ecolmodel.2008.10.018.

(14) Perez-Garcia, J.; Lippke, B.; Comnick, J.; Manriquez, C. An

assessment of carbon pools, storage, and wood products market

substitution using life-cycle analysis results. Wood Fibre Sci.

2005, 37, 140–148.

(15) Searchinger, T. D.; Hamburg, S. P.; Melillo, J.; Chameides, W.; Havlik,

P.; Kammen, D.; Likens, G. E.; Lubowski, R. N.; Obersteiner, M.;

Oppenheimber, M.; Robertson, G. P.; Schlesinger, W. H.; Tilman,

G. D. Fixing a critical climate accounting error. Science 2009, 326,

527528; DOI:10.1126/science.1178797.

(16) Baker, D. J.; Richards, G.; Grainger, A.; Gonzalez, P.; Brown, S.;

DeFries, R.; Held, A.; Kellndorfer, J.; Ndunda, P.; Ojima, D.;

Skrovseth, P.-E.; Souza, C.; Stolle, F. Achieving forest carbon

information with higher certainty: A ﬁve-part plan. Environ. Sci.

Policy 2010, 13 (3), 249-260; DOI:10.1016/j.envsci.2010.03.004.

(17) Intergovernmental Panel on Climate Change (IPCC) In Guide-

lines for National Greenhouse Gas Inventories; Eggleston, S.,

Eds.; Institute for Global Environmental Strategies: Hayama,

2006.

(18) Abt, R. C.; Abt, K. L.; Cubbage, F. W.; Henderson, J. D. Effect of

policy-based bioenergy demand on southern timber markets:

A case study of North Carolina. Biomass Bioenergy 2010, 34

(12), 1679-1686.

(19) Spatari, S.; Bagley, D. M.; MacLean, H. L. Life Cycle Evaluation

of Emerging Lignocellulosic Ethanol Conversion Technologies.

Bioresour. Technol. 2010, 101, 654-667; DOI:10.1016/j.biortech.

2009.08.067.

ES1024004

VOL. 45, NO. 2, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9 795