Effect of Bioenergy Demands and Supply Response on Markets,

Carbon, and Land Use

Karen L. Abt, Robert C. Abt, and Christopher Galik

Abstract: An increase in the demand for wood for energy, including liquid fuels, bioelectricity, and pellets, has
the potential to affect traditional wood users, forestland uses, management intensities, and, ultimately, carbon
sequestration. Recent studies have shown that increases in bioenergy harvests could lead to displacement of
traditional wood-using industries in the short run and intensive management, land use change, and sawtimber
market impacts in the long-run. We simulate timber markets, as well as land use response and carbon outcomes
resulting from projections of both traditional and bioenergy wood use in Alabama, Georgia, and Florida under
differing levels of market supply responses. Increased logging residue recovery had a moderating effect on
prices, although increased planting response led to higher carbon sequestration, and smaller effects on prices.
Increased forest productivity led to lower prices, but also led to reduced timberland and thus lower forest carbon
sequestration. Supply responses will be crucial to moderating market responses to increases in bioenergy wood
demands. FOR. SCI. 58(5):523–539.
Keywords: woody biomass, timber supply, timber demand, logging residue, carbon sequestration

T
HE POTENTIAL FOR INCREASED USE OF WOOD in changes in carbon resulting from biological changes without
producing energy has added uncertainty to forest including potential market responses (Manomet Center for
management and timber markets. Although the de- Conservation Sciences 2010, Biomass Energy Resource
mand for wood for bioenergy is primarily policy-driven at Center 2012).
this time, changes in the prices of fossil fuels could make The current low level of logging residue recovery (LRR)
the use of wood for energy a viable market alternative in the will probably increase to supply the bioenergy market. The
production of liquid fuels or electricity. Regardless of the level of LRR is addressed in most analyses of woody
impetus, however, increasing demands for wood will influ- bioenergy demand, in part because residues are specified in
ence both the existing demand from traditional wood users some federal and state policies as the only type of woody
as well as the potential supply responses. In this analysis, biomass that qualifies as renewable (Bracmort and Gorte
we use a market model of supply and demand, combined 2009). Depending on the temporal scale chosen, carbon
with biomass demands developed by Forisk Consulting accounting of forest residues is also a fairly straightforward
(Brooks Mendell, pers. comm., Jan. 17, 2011) to address exercise. Recent studies suggest that the use of LRR for
potential market responses. We evaluate how markets, land bioenergy may lead to a net increase in carbon emissions in
use, and on-site forest carbon are affected by supply re- the near term, but that the emission differential falls over
sponses. The potential supply responses we evaluate are time (Repo et al. 2011, Domke et al. 2012). Other studies
logging residue recovery rates, increased pine plantation (e.g., Abt et al. 2010b) assume that the long-term differen-
growth, and increased planting response to stumpage prices. tial is negligible, implicitly treating the pool as carbon
Typical approaches to evaluating the wood bioenergy neutral.
markets include the following: assessments of potential Existing forest landowners probably will also increase
available woody biomass (Perlack et al. 2005, Gan and their investments in intensive pine plantation management if
Smith 2006, Biomass Research and Development Board prices increase. A recent summary by Munsell and Fox
[BRDI] 2008, Perez-Verdin et al. 2009) that do not include (2010) provides support for the potential benefits resulting
models of biomass demand; market models of both supply from increased management of pine plantations. The vari-
and demand, using policy-based biomass demands (Galik et ous treatments, including tree improvement, genetics, till-
al. 2009,Rossi et al. 2010, Abt et al. 2010a, 2010b, Ince et age, fertilizer, and competition control, are shown to be
al. 2011); and facility location evaluations, often including economically beneficial to forest landowners (Allen et al.
optimal plant size in addition to plant locations (Wu et al. 2005), technologically possible (Fox et al. 2007), and sus-
2011). Carbon impacts of this changing market have also tainable (Fox 2000).
been assessed, focusing on national (Gan and Smith 2006) Land in timber production has been shown to increase
or regional implications (Abt et al. 2010b), or assessing the based on relative changes in agriculture and forest rents

Manuscript received May 1, 2011; accepted April 2, 2012; published online May 10, 2012; http://dx.doi.org/10.5849/forsci.11-055.
Karen L. Abt, Southern Research Station, USDA Forest Service, PO Box 12254, Research Triangle Park, NC 27709 —Phone (919) 549-4094; [email protected].
Robert C. Abt, North Carolina State University— [email protected]. Christopher Galik, Duke University— [email protected].
Acknowledgments: We thank Brooks Mendell and Amanda Lang, Forisk Consulting, for providing the bioenergy announcement data. We also thank Dan
Koehler, La Capra Associates, Jianbang Gan, Texas A&M, and two anonymous Forest Science reviewers for their helpful reviews and comments.
This article was written by U.S. Government employees and is therefore in the public domain.

Forest Science 58(5) 2012 523

(Hardie et al. 2000). In this analysis, we use the Hardie et al. Timber Market Region: Alabama, Florida, and
(2000) empirical model based on endogenous timber Georgia
price changes, while holding agriculture rents constant. If
timber acreage increases because of timber price increases, Alabama, Florida, and Georgia have an extensive forest
we assume that all gains are in pine plantations, although resource base with a concentration of productive pine plan-
natural forest types remain constant. For this analysis, if tations. More than 33% of the timberland in this region is in
timberland is decreasing, losses are distributed proportion- privately owned pine plantations (USDA Forest Service
ally across all broad management types (pine plantation, 2011) The forest resource currently supports a large forest
natural pine, oak-pine, upland hardwoods, and lowland product industry that consumed more than 90 million metric
hardwoods). tons annually from 1989 to 2007 (Figure 1) (USDA Forest
In the following, we discuss the methods we use to Service 2009). The resource base and logging residue avail-
evaluate the potential supply responses to meet varying ability have also been attractive to the emerging bioenergy
bioenergy wood demands in a subregion of the Southeastern industry. Of the announcements of wood-using bioenergy
Unites States, Alabama, Florida, and Georgia, which is the facilities recorded by Forisk Consulting for the southern
core of the existing southern forest industry and also has states, more than 60% of total announced demands in the
been a focus of the evolving bioenergy industry. We de- South were in Alabama, Florida, and Georgia. In the last
velop demands for both traditional and emerging bioenergy decade, however, lower prices have led to reduced planting
industries. The effects on prices, inventories, and harvest in Georgia and Florida, which will affect future timber
levels, as well as impacts on land use, traditional industries, supply (Figure 2) (USDA Forest Service 2011). The avail-
and carbon sequestration are discussed. The analysis of ability of logging residue is also being affected by the recent
carbon outcomes is limited to forest carbon inventory within drop in wood consumption shown in Figure 1.
the three states only. It does not include impacts of forests Although this region is dominated by pine forests, hard-
outside these states nor changes in emissions associated wood forests are also an important component of the forest
with wood products production or substitutes. landscape. In these states, hardwood utilization is domi-
nated by pulpwood, but hardwood pulpwood consumption
has been less than 20% of total pulpwood consumption in
Methods and Data the last decade (USDA Forest Service 2009). Because hard-
Below we describe our timber market region, describe woods are not managed intensively in this region, there is
the Sub-Regional Timber Supply (SRTS) model used in the no history of management intensification response due to
simulations, describe the scenario development, and discuss higher prices. Higher prices do tend to keep more land in
carbon sequestration calculations. hardwood forests, and this effect is captured in our analysis.

Figure 1. Historical removals for two softwood product groups from 1985 to 2009 for Alabama, Florida.
and Georgia (USDA Forest Service 2009).

524 Forest Science 58(5) 2012

 Figure 2. Plantation area by age class from the most recent FIA surveys for Alabama, Florida, and
Georgia (USDA Forest Service 2011).

Although hardwood harvests generate more logging residue wood pulpwood, softwood sawtimber, hardwood pulpwood,
per unit volume, pine utilization dominates these markets and hardwood sawtimber. The model grows and harvests
and is the focus of the supply response discussion below. trees in response to demands for pulpwood and sawtimber,
where the term pulpwood is used to describe the growing
Modeling Approach stock volume in softwood trees of ⬍9 inches dbh or hard-
wood trees of ⬍11 inches dbh. In addition, a portion of trees
We used the SRTS model to conduct a partial equilib-
in larger diameter classes, which includes top volume and
rium analysis of stumpage markets in the region. The model
cull, was added to pulpwood volume. Softwood and hard-
uses the USDA Forest Service Forest Inventory and Anal-
wood volumes in larger diameters, net of this cull propor-
ysis (FIA) (USDA Forest Service 2011) data set of inven-
tion, were considered sawtimber. The amount of harvest
tory, growth, removals, and acreage by forest type, private
that was considered logging residue was calculated based on
ownership category, species group, and age class for multi-
TPO removal proportions (Table 10 in Johnson et al. 2009)
county areas (FIA survey units). A supply function with
by state.
assumed price elasticity of 0.3 and inventory elasticity of
1.0 was equilibrated with a demand function with price Scenario Development
elasticity of ⫺0.5. Previous research by Pattanayak et al.
(2002) concluded that there is consensus in the literature
Traditional Wood Demand
that both the supply and demand price response is inelastic Demand represents the potential consumption of wood at
(meaning that a large change in price is needed to induce a different prices. To develop demand scenarios over time we
small change in harvest), although the precise values of project consumption of wood in various uses at existing
these elasticities vary by study. prices. This would represent future consumption only if
Once the equilibrium price and quantity of each of the supply were unconstrained. We then use these demand
four products were projected for a year, the model used a scenarios to shift the product demand curve in the model.
goal program to determine from which owners and man- This demand interacts with a supply curve, which is based
agement types the product harvest would come. The goal on current year product inventories and econometrically
program reconciles product volumes with observed harvest estimated supply-price relationships. The resulting price
patterns across forest types and age classes. The harvest was and harvest from supply-demand equilibrium is reported as
then passed to the biological accounting module, and inven- the market outcome. The change in inventory resulting from
tory was updated for the next period’s equilibrium calcula- the harvest and annual growth is also presented.
tion. The accounting module tracks inventory by 5-year age The projected demand for timber from traditional wood
classes for the five southern forest types (pine plantations, processing for four product types is shown in Figure 3. The
natural pine, oak-pine, upland hardwoods, and bottomland 2007–2010 recession can be seen to have had an effect by
hardwoods) and for two ownership classes (corporate and reducing demand and is followed by a recovery through
other private). Further details of the SRTS model can be 2018. We used downscaled national observations of paper
found in Abt et al. (2009). and wood products industry value of shipments (US De-
In this analysis, we evaluate four product types: soft- partment of Commerce, Bureau of the Census 2011) to

Forest Science 58(5) 2012 525

 Figure 3. Historical traditional demand for four product groups from 1997 to 2007 and projected
traditional demands for 2008 to 2037 for Alabama, Florida, and Georgia.

extend the 2007 timber product output data (USDA Forest et al. 2011, the U.S. Billion-Ton Update (US Department of
Service 2009) to 2010. We then used the Bureau of Labor Energy 2011), and US Energy Information Administration
Statistics 2008 –2018 forecast at the national industry level policy assessments (US Energy Information Administration
for total industry output (Woods 2009), also downscaled to 2007). For this analysis, we used a database of announced
the state level. Then, assuming a softwood-hardwood mix facilities in the region that would process woody biomass
consistent with state historical trends, we partitioned the into electricity, wood pellets, and liquid fuels. Forisk Con-
changes in industries year-over-year to each of the four sulting has developed a methodology to project medium-
products out to 2018. This assumes that the change in value term demand for woody biomass based on an evaluation of
of shipments translates directly to a change in input demand. viable technologies and the status of the announcements
Beyond 2018, we assume decade-over-decade increases of (Forisk Consulting (Brooks Mendell, pers. comm., Jan. 17,
10 and 5% for sawtimber and 5 and 2.5% for pulpwood for 2011)).
the second and third decades of the projection, respectively. Forisk Consulting gathers information on all announced
Because some policies encourage the use of logging facilities by type and location and then researches the level
residues and others allow only the use of logging residues, of commitment made to the facility. These announcements
the level of projected traditional wood demand is important include facilities for the production of liquid fuels, wood
in any assessment of the impact of woody bioenergy de- pellets, and electricity. Forisk Consulting assesses the an-
mand on timber markets. A lower level of timber demand nouncements for the projected start-up date and includes
from pulp and paper mills and sawmills, for example, will announcements out to 2020. We use the timber demands
lead to lower harvest levels and fewer available logging from all announced facilities as the bioenergy demand in the
residues. If only residues are allowed to qualify as renew- region (Figure 4). The start-up dates result in a steep in-
able, then the woody bioenergy industry is explicitly tied to crease in demand in early years and then a leveling off in
the future of the traditional wood industries. However, if later years, which will probably be smoothed to a more
roundwood is used for bioenergy, then the market outcome linear increase as projects are delayed or ended and new
is more complicated. A lower level of traditional harvest projects are announced. A higher level of bioenergy demand
could lead to fewer available residues (which could raise the could result if the technology for cellulosic ethanol develops
price of residues and set a physical upper limit on residue and wood-derived fuel is used to meet the Renewable Fuels
supply) but could also lead to higher inventory levels and Standard (US Energy Information Administration 2007) or
lower roundwood prices, which would favor increased if the United States adopts a carbon reduction policy or a
roundwood utilization for bioenergy. renewable electricity standard. Conversely, a lower level of
bioenergy demand could result if the technology for com-
mercial cellulosic ethanol does not develop, if additional
Bioenergy Wood Demand state-level renewable energy standards are not adopted, or if
Because bioenergy is an emerging industry, econometric the price of nonrenewable energy falls.
relationships have not been estimated for either supplies or Forisk Consulting provides estimates of actual
demands. The current biomass energy literature focuses on (2008 –2010) and potential (2011–2020) bioenergy demand.
projected consequences of bioenergy policy on bioenergy We assumed that the level will increase at a rate of 10% per
demands. Examples include Ince et al. (2011), Alavalapati decade for the following two decades of the projection. In

526 Forest Science 58(5) 2012

 Figure 4. Bioenergy demand from announced facilities for liquid fuels, pellets, and electricity production
for Alabama, Florida, and Georgia from 2007 to 2037 (adapted from Forisk Consulting).

addition, although Forisk Consulting separates the an- database was split into hardwood and pine demand based on
nouncements by end use, they do not estimate how the current pulpwood harvest. A portion of this demand was
various uses will use hardwood and softwood species. We met by logging residue based on the availability and the
use their aggregate woody bioenergy demand value in met- LRR factor for that species group and year. After residue
ric tons (Figure 4) as our bioenergy demand. utilization was deducted, the remaining bioenergy demand
We implemented bioenergy demand by calculating avail- quantity was used to shift the current year pulpwood de-
able logging residue from existing harvests and assumed mand curve. Because demand and supply are price-inelastic
that these residues would reduce bioenergy demand for in these markets, demand shifts lead to larger price changes
roundwood. The utilization of residues will depend on the than harvest changes.
mix of bioenergy consumers (e.g., liquid fuels, pellets, and The difference between the demand shift and the result-
electricity) and the technologies they use. There are also ing market equilibrium harvest change for both industries is
cost, logistics, and logging capacity questions associated what we call “displacement.” If we had information on the
with residue supply. To the extent that these factors con- demand price elasticity of emerging bioenergy demand, it
strain residue utilization below our assumptions, our esti- would be possible to simulate the change in the price
mates of roundwood market impacts are conservative and responsiveness of the combined (existing and bioenergy)
vice versa.
industry demand. Because we do not have this information,
Residue availability was based on two calculations. The
however, we compare this total displacement quantity to the
first is an estimate of residue production from harvest op-
demand quantities in bioenergy and existing pulpwood con-
erations. The second is LRR over time, which is discussed
sumer. Three potential outcomes could occur.
below under supply responses. For residue production, we
First, if both industries had the same demand price re-
used the state-level data recorded in Table 10 in Johnson et
sponse, the displacement would be proportional, e.g., de-
al. (2009) to estimate growing stock and nongrowing stock
logging residues per ton of growing stock harvest. Table 10 mand might be reduced by 20% in each sector because of
estimates were adjusted to remove stump volumes, which higher prices. Second, if bioenergy were more price-inelas-
are included in the Forest Service definition of residue but tic (less price-responsive) than traditional industries, the
are not usually considered recoverable. Although utilization higher market prices would affect the existing wood con-
standards may evolve over time, we held these residue sumers disproportionately; i.e., a larger than proportionate
production rates constant over the projection. share of displacement would be due to reduced consump-
Although the above procedure gives us supply estimates tion from the existing industry. Price-inelastic bioenergy
of logging residue by size class and species group, our demand would be consistent with the view that these de-
bioenergy demands are not specific to species or size mands are policy-driven and that some power markets are
classes. For these projections, we assumed that bioenergy regulated so that costs may be passed through to power
demand would follow the current pulpwood harvest alloca- consumers.
tion between pine and hardwood. As noted above, this Third, if bioenergy demand were more price-responsive
assumes that approximately 80% of bioenergy demand in (less price-inelastic) than existing wood consumers, higher
the region would come from pine. wood prices would have a larger impact on bioenergy
The procedure for adding bioenergy demand to tradi- consumption, and a larger than proportionate share of dis-
tional demand was the following. The Forisk Consulting placement would be due to reduced consumption by the

Forest Science 58(5) 2012 527

bioenergy sector. This view would be consistent with bio- lation to reflect income from both pulpwood and sawtimber
energy firms updating their feedstock price expectations as price changes using a net present value calculation. Our
markets evolve so that fewer announcements become actual calculation takes into account the price and timing differ-
facilities. Availability of renewable or carbon-friendly sub- ential in these products. This calculation continues to put the
stitute feedstocks at relatively lower prices could also lead primary weight on sawtimber prices, but in times of lower
to greater bioenergy demand response to wood prices. sawtimber prices and high pulpwood prices, a larger pro-
For these projections, we assumed that the price elastic- portion of rent is based on pulpwood price.
ity of wood demand did not change with the addition of Current growth rates in the model are developed from
bioenergy demand, so that the difference between the de- regression equations for each management type and phys-
mand shift due to bioenergy and the resulting harvest is iographic region. These rates represent averages across
based on a demand price elasticity of ⫺0.5. In our results large areas, and different owners. To model increased
we show the proportion of bioenergy or pulpwood demand growth that is likely to occur on some stands, but not all,
that this displacement represents. This gives some perspec- due to increases in fertilization, thinning, or genetic selec-
tive to how changes in the allocation of displacement would tion, we augmented the growth rates on new plantations
affect the sectors. beginning in 2008, rising to a regionwide average increase
of 25% by 2037. Again, assuming a regionwide average of
25% implies that although some individual stands could
Supply Response have growth rates higher than 25% over 30 years, some
Three potential supply responses were evaluated: in- stands will also have lower growth rates. An earlier study,
creased LRR, increased plantation growth, and increased the SFRA (Prestemon and Abt 2002), used an assumed
timberland area response. For LRR and growth responses, increase of 50% in growth over 50 years compared with our
there are no empirical data to determine market response to 25% growth increase over 30 years.
demand/price changes. Thus, we make assumptions about
the rate of adoption of these responses. For timberland, the
model calculates endogenous planting response (Hardie et
Scenarios
al. 2000) as described in the introduction, but these re- We simulated six scenarios to evaluate the consequences
sponses are sensitive to the definition of forest rent, which of the three potential supply responses individually or in
we explore below. combination (Table 1). The scenarios include (1) a baseline
We evaluated two end levels of LRR, 33% and 66%, including only traditional demands, no new bioenergy de-
with increases from zero to this level over the first 5 years mands, and no additional supply responses; (2) BIO 1 added
of the projection. These recovery rates are applied to all bioenergy demand and a supply response of a 33% LRR;
harvests, implying that we expect some harvests to be (3) BIO 2 doubled the residual supply response by assuming
recovered at a higher rate and some at a lower rate to reach a 66% LRR; (4) BIO 3 added a potential increase in planting
this regional average on all harvests. The higher level, 66%, response by including pine pulpwood prices in the timber
is similar to levels assumed to be the operational maximum rent calculation; (5) BIO 4 added a 25% increase in growth
level of LRR (Perlack et al. 2005, BRDI 2008) on individual on new plantations but did not include pulpwood in the rent
pine stands. We assume that the adjustment process will not calculation, and (6) BIO 5 added both the increased planting
be immediate because the logging sector will need to adjust response of BIO 3 and the increased growth on new plan-
equipment, transportation, and employment capacity to ac- tations of BIO 4.
commodate the new markets.
The initial hectares of pine plantations are derived from
the FIA database (USDA Forest Service 2011). Timberland
Carbon Sequestration Calculations
area is projected for each scenario based on Hardie et al. We converted SRTS inventory projections into estimates
(2000) as a function of pine sawtimber prices. Other inputs of on-site forest carbon through the use of FIA-derived
are used in Hardie et al. (2000), such as agricultural rents ecosystem-level equations (Foley et al. 2009, as based on
and county population forecasts. Agricultural rents are held Smith and Heath 2002 and Smith et al. 2006). These pro-
constant and the loss of rural land to urbanization is based vide carbon estimates for each of the five management
on the county-level population forecasts used in the South- types by age class included in the SRTS model. For each
ern Forest Resource Assessment (SFRA) (Prestemon and year of SRTS output, we calculated the total amount of
Abt, 2002). The timber price used in the land use forecast is carbon contained in live tree, dead tree, understory, down
the previous year’s SRTS model output, so that the response deadwood, and forest floor carbon pools across all forested
of planting in time t is made in response to the modeled acres in each of the five management types.
softwood pulpwood and sawtimber prices in time t ⫺ 1. In Note that this analysis accounts only for changes in
this analysis, overall timberland change was initially tied forest carbon and is not a complete carbon accounting of the
primarily to pine sawtimber prices, as developed in the greenhouse gas (GHG) effects, which would require a as-
Hardie et al. model. When timberland increases, the in- sessment of fossil fuel substitution and a life cycle analysis
creases are assumed to occur in pine plantations, although of bioenergy GHG emissions (see, e.g., Mann and Spath
decreases in timberland are assumed to be proportional 2001). We also take a simplified approach to LRR carbon
across all five management types. To better reflect biomass accounting, assuming a negligible difference between near-
market impacts on forest rents, we modified the rent calcu- term use of residues for bioenergy and their long-term

528 Forest Science 58(5) 2012

Table 1. Definition of baseline and bioenergy demand scenarios with varying supply responses.

Supply response
Residue Increased planting Growth
Scenario recovery response increase
Baseline: traditional demand only (no NA Timber rent based on pine sawtimber prices None
bioenergy demand or supply response)
BIO 1: traditional ⫹ bioenergy demand; 33% Timber rent based on pine sawtimber prices None
33% LRR
BIO 2: traditional ⫹ bioenergy demand; 66% Timber rent based on pine sawtimber prices None
66% LRR
BIO 3: traditional ⫹ bioenergy demand; 66% Timber rent based on pine pulpwood and None
66% LRR; increased planting response sawtimber prices
BIO 4: traditional ⫹ bioenergy demand; 66% Timber rent based on pine sawtimber prices 25% on new planting by end
66% LRR; 25% growth increase of projection
BIO 5: traditional ⫹ bioenergy demand; 66% Timber rent based on pine pulpwood and 25% on new planting by end
66% LRR; increased planting sawtimber prices of projection
response; 25% growth increase
NA, not applicable; LRR, logging residue recovery.

decomposition on the ground. We therefore exclude carbon (Figure 5a) shows a decline in price from 2007, with an
stored in harvest residues from our estimates of forest eventual recovery, although prices never return to the 2007
carbon, because the pool is effectively canceled out by level. Although acres of young stands are affected by re-
including it in both the bioenergy and baseline scenarios. duced planting, this effect is offset by the “cull” proportion
Other research has explored the effect of displaced fossil of the growing sawtimber inventory. Removals and inven-
emissions on net GHG balance (e.g., Abt et al. 2010b); tory stay fairly steady throughout the projection. When
therefore, this analysis is limited to the carbon implications bioenergy demand is introduced, but supply response is
of shifts in timberland and inventory only. Finally, we limited (BIO 1, with only a 33% LRR) (Figure 5b), pulp-
evaluate only the three states listed above, and do not wood prices rise through the end of the projection, with
include impacts on forests outside these states, nor do we 2037 97% higher than 2007. Pulpwood inventory rises
include any changes in emissions associated with the pro- during the recession, but quickly falls to hover below the
duction of wood products or their substitutes. Our analysis 2007 level, and removals show a rise after the recession, but
focuses on the net forest carbon impacts of increased de- then level off, consistent with our expectations of a market
mand for bioenergy and a set of expected supply responses with inelastic price response.
to this demand. Doubling the logging residual recovery rate to 66% (Fig-
ure 5c) reduces the impact on the pulpwood market, with
Results and Discussion prices rising only by 70% over 2007 levels and smaller
The results are presented in Figures 5⫺12. Figure 5 effects on both inventory and removals. Adding an en-
shows softwood pulpwood market responses from 2007 to hanced response of planting to pulpwood prices over the 30
2037, and Figure 6 shows softwood sawtimber responses for year projection moderates the market effects, again leveling
that same period. There are few impacts on the hardwood off the rising trend so that the projected 2037 value is now
market, both because of the initial effect of changes in only 63% higher than the 2007 value (Figure 5d, BIO 3).
demand, based on the proportion of the market fulfilled by Under a supply response of high LRR and an increased
hardwoods, and because the hardwood forest types are not growth rate on new plantations (Figure 5e, BIO 4), prices
typically intensively managed to improve growth nor are rise as in BIO 2 and BIO 3 through 2023 but then begin to
they planted. Plantation acres over time are shown in Figure moderate as the new growth on plantations begins to influ-
7 for all six scenarios, and Figure 8 shows the land use ence the inventory values. Inventory increases, removals
changes by planted and natural forest types over the pro- increase, and prices fall after 2023, with prices projected at
jection by scenario. Figure 9 shows the bioenergy feed- only 32% higher in 2037 than in 2007. Figure 5f shows the
stocks from forests including residue use, harvest change, combined supply response of 66% LRR plus 25% growth
and harvest displacement, and Figure 10 illustrates the range increase plus the increased planting response (BIO 5) and
of how total displacement of demand could affect the two has the lowest price level in 2037 of all of the bioenergy
industries. Figures 11 and 12 show the three-state forest scenarios (at 127% of the 2007 level). An increase in the
carbon storage from 2007 to 2037, with Figure 11 showing LRR reduces the overall effect of the bioenergy demands;
the total and Figure 12 showing the storage by both natural increased plantation response to pulpwood prices leads to a
and planted forest types. Each of the figure and scenario moderating of price increases and increased growth on
results is discussed in more detail below. plantations leads to a reversal of the rising trend in prices.
Softwood pulpwood markets respond, as expected, more Figure 6 shows the response of the softwood sawtimber
than the other three product types because any bioenergy market to the six supply response scenarios.
demand not met through residue recovery is assumed to be Pulpwood and sawtimber are linked in the model through
met by using pulpwood. The baseline pulpwood projection product definitions because a percentage of the sawtimber

Forest Science 58(5) 2012 529

 Figure 5. Softwood pulpwood: comparison of prices, inventory, and removals from simulations of the
baseline and five bioenergy supply response scenarios for Alabama, Florida, and Georgia from 2007 to
2037.

size class is classified as pulpwood, and increased harvest of effect of the recession on softwood sawtimber markets, with
pulpwood reduces ingrowth into the sawtimber category a return of prices to 2007 levels not projected until 2027.
[1]. Sawtimber harvests are a key component of logging This is a result of only our traditional demand projection
residue production. Because sawtimber prices are currently and the buildup in sawtimber inventories, because bioen-
at about three times the level of pulpwood prices [2], we do ergy demand is not included in the baseline projections.
not expect that sawtimber will be used to provide bioenergy Inventory levels rise sufficiently in the early years to allow
over the projection period. Even in the lowest supply re- removals to recover with slower recovery in price levels. In
sponse scenario (BIO 1), pulpwood prices rise only to about the scenario with the largest harvesting impact (BIO 1, the
double their current level, not high enough to induce the use smallest supply response of the five we examined), prices
of sawtimber for pulpwood or biomass except at the margin. rise more rapidly, returning to 2007 levels 4 years earlier
The baseline (Figure 6a) shows a large and long-lasting and continuing at a higher rate of increase through the end

530 Forest Science 58(5) 2012

 Figure 6. Softwood sawtimber: comparison of prices, inventory, and removals from simulations of the
baseline and five bioenergy supply response scenarios for Alabama, Florida, and Georgia from 2007 to
2037.

of the projection (Figure 6b, BIO 1). Correspondingly, the 2007 levels. The addition of a 25% growth increase,
levels of removals and inventory begin to decline somewhat shown in Figure 6e (BIO 4), instead of the change in
toward the last decade of the projection. planting response, results in very similar outcomes, and
Doubling the residue recovery rate has effects similar to combining both growth and planting responses leads to the
the effects on the pulpwood market, although not as dra- least increase in prices over the projection, similar to the
matic because as noted above this market is assumed to not response of the pulpwood market discussed above (Figure
be directly affected by the changes in roundwood demand 6f, BIO 5).
caused by the changes in residue recovery rates (Figure 6c, Although not displayed in detailed figures, the scenarios
BIO 2). The year in which price returns to the 2007 level is included projections for the hardwood pulpwood and hard-
delayed by more than a year, and the ending price increase wood sawtimber markets. Overall, these markets were little
is only 31% compared with 36% in BIO 1. Addition of the affected by most of the scenarios, in part because hard-
increased response of planting to pulpwood prices further woods comprise a smaller portion of the overall timber
moderates the increase (Figure 6d, BIO 3) to 27% above markets in these three states but largely because the supply

Forest Science 58(5) 2012 531

 Figure 7. Area of planted and natural timberland from simulations of the baseline and five bioenergy
supply response scenarios for Alabama, Florida, and Georgia from 2007 to 2037.

responses of growth increase and planting response do not Figure 8 provides more detail on the changes in pine plan-
affect hardwoods. Only the change in LRR had noticeable tations. After the recession, the baseline plantation area
effects on the hardwood simulations. returns to a stable level approximately 0.1 million ha below
The baseline projections for hardwood pulpwood show a the 2007 value of 7.1 million ha. This is partially due to an
continuation of the rises in hardwood inventory and corre- assumption that plantation gains occur at the extensive
sponding declines in prices. Removals recover from the margin, but it is primarily due to low pine sawtimber prices
recession in 2023. Adding bioenergy demands and the 33% for much of the projection period. All of the bioenergy
LRR leads to a spike in price as demands increase, but then scenarios lead to higher planting, resulting from the in-
price smooths out, ending at 30% over 2007 levels, although creases in pulpwood and/or sawtimber prices as a result of
both inventory and removals are also above 2007 levels the new bioenergy demands. The variation results from the
because of increased growth in the hardwood inventory. The differing supply responses and their effects on pulpwood
doubling of LRR to 66% removes the large price increase, and sawtimber prices. Increasing the planting responsive-
with prices recovering only to 90% of the 2007 levels. None ness to pulpwood prices leads to higher projected land area
of the other supply responses change the hardwood pulp- in plantations (BIO 3 and BIO 5), with the addition of
wood projections noticeably. For hardwood sawtimber, the growth tempering the increase in area as a result of the
baseline reflects the recession and recovery, with prices lower softwood product prices. Total timberland acres de-
slightly below 2007 levels, but inventory and removals creased 8% in the baseline and 5– 6% in the bioenergy
slightly higher. The addition of bioenergy demand keeps scenarios. These results have less plantation area increases
prices slightly higher, but there are few other changes over than prior regional assessments including the SFRA (Pres-
the supply response scenarios. Because hardwoods are not temon and Abt 2002) and more recently the Southern Forest
expected to have increased growth or to be planted in Futures Project (Huggett et al. 2011). SFRA was conducted
substantial numbers, these results are expected. when the relevant policy questions were related to how the
Overall changes in both pine plantation and natural tim- southern resource would respond to anticipated higher de-
berland area for the three states over the projection period mands for wood products. Those studies suggested an ex-
for the six scenarios are shown in Figure 7. Although the pansion of the plantation base consistent with current plan-
differences are small relative to overall timberland area, tation acreage. The Southern Forest Futures Project is based

532 Forest Science 58(5) 2012

 Figure 8. Detailed area of planted pine timberland from simulations of the baseline and five bioenergy
supply response scenarios for Alabama, Florida, and Georgia from 2007 to 2037.

on demand scenarios derived from Intergovernmental Panel ing response is linked to price changes, but growth re-
on Climate Change global change scenarios and uses tran- sponses are not, increased growth lowers the price impetus
sition matrices and planting rates consistent with price sce- for increased planting. The combined effect (Figure 9e, BIO
narios embedded in a set of cornerstone solutions. Each of 5) is much smaller than the sum of the separate effects
these scenarios project higher acreage of plantations than (Figure 9c and d).
we do, but the study does not include the current recession, The impact of displacement depends on the relative
which dominates the short-term outlook in this study. demand price elasticities for wood by the bioenergy and
The harvest outcomes and residue recovery assumptions pulpwood sectors. Figure 10 shows the relationship between
are reflected in the projected bioenergy feedstocks seen in pine pulpwood and bioenergy demand projections and the
Figure 9, which shows the portion of bioenergy demand met resulting harvest change. In Figure 10a, which has the
by softwood and hardwood used residues, new harvest, and lowest supply response (BIO 1), the price effect on reduced
displacement. As defined above, displacement is the differ- wood consumption is more than 80% of original biomass
ence between expected demand with no price change and demand. In other words, if all of the reduction in sector
equilibrium harvest that incorporates supply. Only in BIO 1 demand was allocated to the bioenergy industry, less than
(Figure 9a), which has the smallest supply response (33% 20% of announced bioenergy capacity would be built. Al-
LRR), does hardwood harvest change and displacement ternatively, if the entire displacement was fully absorbed by
play a role; otherwise hardwood residues fulfill most of the the pine pulpwood industry, there would be a 40% reduction
demand for hardwoods. In Figure 9b– e, softwood residues in projected sector demand. Figure 10b shows the direct
show steep increases in the beginning of the projection as impact of LRR increases (BIO 2), and although the effect is
LRR rises to 66% in 2012 (BIO 2–5). Actual levels fluctuate immediate, it does not change the path of displacement.
in concert with harvest levels. After depletion of available Figure 10c and d (BIO 3 and BIO 4) shows the reduction in
residues, pulpwood harvest increases, but inelastic supply displacement over time associated with planting responses
implies that equilibrium harvest will be significantly smaller and growth responses, respectively, and Figure 10e shows
than the demand increase. Planting has a small influence on the combined responses (BIO 5). These figures also illus-
the feedstocks (Figure 9c), but the impact does not occur trate that although increased harvest directly affects pine
until a decade after the biomass demand enters the pulp- pulpwood displacement, the additional residues from in-
wood market (BIO 3). The planting impact is muted because creased harvest allow bioenergy displacement to decline
low sawtimber prices dominate the rent calculation even faster for a given supply response. With the assumption of
when pine pulpwood prices are included. There are larger, similar demand-price responses, both industries would
more immediate, cumulative impacts from assumed growth experience a 28% reduction from expected demand based
increases (Figure 9d, BIO 4). Higher supply leads to in- on supply price effects at the end of the projection in the low
creased pine harvest and less displacement. Because plant- supply response BIO 1 scenario (Figure 10a). In the high

Forest Science 58(5) 2012 533

 Figure 9. Wood use for bioenergy feedstocks from simulations of five bioenergy supply response scenarios
showing displacement, new harvest, and used residues by species group for Alabama, Florida, and Georgia
from 2007 to 2037.

supply response BIO 5 scenario (Figure 10e), with similar additional planting or reduced loss of natural stands to
price responses between the sectors, displacement would agriculture due to higher bioenergy demand. The difference
peak at 19% in 2022 but decline to 12% by the end of the between BIO 1and BIO 2, which reflects only higher resi-
projection when increased growth and planting rates are due utilization, in BIO 2, shows BIO 2 with small but
incorporated into pulpwood supply. cumulative carbon advantage over time. This effect is
Forest carbon stock is driven by growth, removals, age slightly overstated as the reduction in the transient carbon
class distribution, and land use change. The market impacts stock in residuals is not estimated. The advantage here is
were concentrated in pine plantations, but they represent due to decreased harvest of roundwood only. Note the scale
only 26% of the initial carbon stock, although the hardwood in Figure 11: the maximum carbon disadvantage over the
types (oak-pine, upland hardwood, and lowland hardwood) baseline is approximately 2% in BIO 4 in the second decade
represent 57%. Figure 11 shows the aggregate on-site forest and the maximum carbon advantage over the baseline is
carbon inventory across all forest types by scenario by year. approximately 1.5% in BIO 3 and BIO 5 in the first decade.
The total carbon impacts indicate that scenarios where for- Differences between carbon impacts from plantation
est rent included pine pulpwood and had net carbon in- growth rate increases (BIO 4 and BIO 5) and land use
creases over the baseline scenario. Scenarios that linked effects due to forest rents (BIO 3 and BIO 5) are best
forest rents only to pine sawtimber prices showed net re- understood by examining the differential impacts on planted
ductions because the increased harvest was not offset by and natural stands shown in Figure 12. As modeled here,

534 Forest Science 58(5) 2012

 Figure 10. Range of displacement possible for bioenergy sector and traditional sector resulting from
assumptions that all displacement will fall on selected sector from simulations of five bioenergy supply
response scenarios for Alabama, Florida, and Georgia from 2007 to 2037.

land use change responds to price, but the growth increase carbon stocks in the plantations with increasing growth
is not price-responsive, which might correspond to a con- rates to exceed baseline carbon by the end of the projection.
tinued trend in genetic improvement. Figure 12 shows that Planting responses in these scenarios are depressed by
the carbon differences among scenarios are driven primarily continued low pine sawtimber prices. Pine sawtimber is
by differences in natural stand carbon. Over time, market the only rent driver in BIO 1, BIO 2, and BIO 4 but has
responses tend to offset changes in supply or demand on a greater influence than pine pulpwood prices in all
plantations. During the recession, fewer acres were both scenarios.
harvested and planted. Although decreased prices led to a The natural stands are also affected by price in the
loss of plantation acres, the decreased harvest allowed total model. When forest rents rise, the timberland increase rel-
inventory and carbon to increase slightly. Higher demand ative to agriculture is attributed to pine plantations. With
leads to increased harvest and price, which reduces the increased rents, however, there may also be less conversion
inventory and carbon stock but leads to more planting and of natural stands to agriculture. No scenario increased tim-
intensive management. As shown in Figure 12, this allows berland at the agriculture margin enough to offset the effect

Forest Science 58(5) 2012 535

 Figure 11. On-site carbon sequestration from simulations of five bioenergy supply response scenarios for
Alabama, Florida, and Georgia from 2007 to 2037.

of urbanization, but compared with the baseline run, re- market results indicate that the price effect may displace up
duced loss of natural stands has a significant carbon effect. to 28% of projected demand in the low supply response
Scenarios BIO 3 and BIO 5, which use forest rents that scenarios or 12% of projected demand in the high supply
include biomass price effects on pine pulpwood, lead to less response scenarios. How this displacement is distributed
loss in all forestland and an increase in plantations to depends on the wood demand price responsiveness of the
prerecession levels as described above. This advantage ap- emerging industry relative to that of the existing industry.
pears before the biomass markets begin because pine pulp- Higher product prices are linked to land use through the
wood prices are less affected by the recession in the pre- planting response function such that an increase in prices
biomass market period. The BIO 5 scenario, which also will lead to an increase in new pine plantations. Thus, in
includes the higher plantation growth rate, however, begins scenarios in which prices increase, there is more timberland
to lose some of this advantage over time as higher pulpwood area than occurs under the baseline scenario without bioen-
inventories lower prices and rents, which lower replanting ergy demand. This, in turn, leads to a higher level of carbon
rates. Scenarios BIO 1 and BIO 4 reduce the carbon stock in sequestration in the standing forest than occurs under the
natural stands relative to the baseline. In BIO 1 with low baseline. Growth increases, assumed at a regional average
residue utilization, natural hardwood stands are harvested to of 25% on new pine plantations, lead to an increase in
meet bioenergy demand. BIO 4 includes plantation growth inventory volume that will have a moderating effect on
increases, but forest rent is linked to only pine sawtimber so pulpwood prices in the simulations. The regional carbon
that natural timberland is decreasing faster than in other outcome is lower in the scenario with growth increases than
scenarios at the end of the projection. in the scenario with only increased planting response. In all
scenarios, including the baseline, the three-state forest car-
Conclusions bon sequestration is higher at the end of the projection than
Regardless of the forces that create new bioenergy wood in 2007, because carbon in aging existing stands accrues at
demands, there will be impacts on traditional wood-using a faster rate than what is expected to be lost through harvest
industries. The magnitude of these impacts, however, will or conversion to other land uses.
depend on both the level of demand and the level of supply One crucial assumption regarding the market effects of
response. The time path and market consequences of the bioenergy demands is the projection of the traditional in-
growth and planting responses are quite different, so that dustries. We modeled as if traditional demand for stumpage
although these are the primary levers by which management would continue to increase at a rate of about 1% per year for
can influence supply, the scale, timing, and market impacts sawtimber and 0.5% per year for pulpwood for the second
of all three responses lead to different land use and carbon decade and half that for the third decade of the projection.
outcomes. If demand stays at a level coincident with an- Under an alternative assumption that demands from these
nouncements of facilities and if logging residue recovery industries fall in the absence of bioenergy demands, then
rates can average 66% or higher across the region, then prices will fall, potentially leading to declines in timberland

536 Forest Science 58(5) 2012

 Figure 12. On-site forest carbon sequestration from simulations of five bioenergy supply response
scenarios for planted and natural stands for Alabama, Florida, and Georgia from 2007 to 2037.

area and thus in sequestered carbon. Adding in the new These results also illustrate the importance of the recov-
bioenergy demands, however, means that declines in tradi- ery of pine sawtimber demand in this region. Although a
tional harvest will reduce feedstocks available from logging relatively modest plantation growth increase is sufficient to
residues, regardless of recovery rates. This, in turn, will lead stabilize prices and restore inventories given bioenergy de-
to higher harvests of pulpwood for bioenergy with resulting mand examined here, the key forest rent driver of supply
price increases for pulpwood and thus smaller declines in response, pine sawtimber price, may be largely unaffected
timberland area and higher forest carbon sequestration in by bioenergy demand in the near term. These factors imply
these three states. that increased pine pulpwood demand in the absence of pine
Additional research is needed on the price responsive- sawtimber market recovery will increase pressure to use
ness of several of the relationships modeled in the scenarios small sawtimber to meet these demands. Further research is
for which we made assumptions, including residue recovery needed to investigate how price-sensitive product defini-
response to prices, plantation productivity response to tions would affect these results.
prices, and changes in pine plantation conversion from other
forest types. Finally, improved carbon accounting would
better account for carbon stored in end-use wood products,
Endnote
the carbon dynamics of residue recovery or decay over time, [1] An increase in the demand for residues for wood energy has the
and emissions attributable to shifts in harvest activity out- potential to result in an increased harvest of roundwood, dependent on
a complementary relationship between residues and roundwood (either
side of the three-state-region. Although we do not expect the sawtimber or pulpwood), such that an increase in the price of residues
inclusion of these components to change the magnitude or (from zero to a positive value) would lead to an increased harvest of
direction of our results, it would add additional detail to our either sawtimber or pulpwood. Using a theoretical stand-level model of
soil expectation value, a new product such as residues could lead to a
estimates and allow for increased scrutiny of those scenarios small decline of the optimal rotation age, depending on the presence of
yielding marginal carbon benefits (e.g., BIO 5). other costs and the production rates of residues from the different

Forest Science 58(5) 2012 537

 products (all else held constant). Whether this leads to an increase or Responsiveness of rural and urban land uses to land rent
decrease in aggregate harvest volumes is dependent on these optimal determinants in the U.S. South. Land Econ. 78(4):659 – 673.
stand decisions, the number of stands affected, and the ultimate effects
on market prices from changes in stand management. We are not aware
HUGGETT, R.J. JR., D.N. WEAR, R. LI, J. COULSTON, AND S. LIU.
of any research on the complementarity between roundwood and 2011. Forecasts of forest conditions. Chapter 5 in Southern
residues, and little research on the complementarity of the pulpwood forest futures project, Wear, D.N., and J. Greis. (eds.). 71 p.
and sawtimber. The results from these latter studies (Newman 1987, Available online at www.srs.fs.usda.gov/futures/reports/draft/
Newman and Wear 1993, Prestemon and Wear 2000, Polyakov et al. Frame.htm; last accessed on Mar. 5, 2012.
2010) provide mixed results on complementarity between forest
stumpage products. INCE, P.J., A.D. KRAMP, K.E. SKOG, D. YOO, AND V.A. SAMPLE.
[2] According to Timber-Mart South, fourth quarter average southwide 2011. Modeling future U.S. forest sector market and trade
prices for pine sawtimber were $23.54/ton and for pine pulpwood were impacts of expansion in wood energy consumption. J. For.
$8.20/ton (Timber Mart South 2012). Econ. 17(2):142–156.
JOHNSON, T.G., J.W. BENTLEY, AND M. HOWELL. 2009. The
Literature Cited South’s timber industry—An assessment of timber product out-
put and use, 2007. US For. Serv. Resour. Bull. SRS–164.
ABT, R.C., K.L. ABT, F.W. CUBBAGE, AND J.D. HENDERSON.
Southern Research Station, Asheville, NC. 52 p.
2010a. The impact of bioenergy demands on southern forests:
MANN, M.K., AND P.L. SPATH. 2001. A life cycle assessment of
A case study of North Carolina. Biomass Bioenergy
biomass cofiring in a coal-fired power plant. Clean Products
34:1679 –1686.
ABT, R.C., F.W. CUBBAGE, AND K.L. ABT. 2009. Projecting south- Processes 3(2):81–91.
ern timber supply for multiple products by subregion. For. MANOMET CENTER FOR CONSERVATION SCIENCES. 2010. Biomass
Prod. J. 59(7/8):7–16. sustainability and carbon policy study. Available online at
ABT, R.C., C.S. GALIK, AND J.D. HENDERSON. 2010b. The near- www.manoxmet.org/sites/manomet.org/files/Manomet_Bio-
term market and greenhouse gas implications of forest biomass mass_Report_Full_LoRez.pdf; last accessed Apr. 5, 2011.
utilization in the Southeastern United States. Climate Change MUNSELL, J.F., AND T.R. FOX. 2010. An analysis of the feasibility
Policy Partnership, Duke University, Durham, NC. 49 p for increasing woody biomass production from pine planta-
ALAVALAPATI, J.R.R., P. LAL, A. SUSAETA, R.C. ABT, AND D.N. tions in the southern United States. Biomass Bioenergy 34(12):
WEAR. 2011. Forest Biomass-based energy. Chapter 10 in 1631–1642.
Southern forest futures project, Wear, D.N., and J. Greis. (eds.). NEWMAN, D.H. 1987. An econometric analysis of the southern
155 p. Available online at www.srs.fs.usda.gov/futures/reports/ softwood stumpage market 1950 –1980. For. Sci. 33:932–945.
draft/Frame.htm; last accessed on Mar. 5, 2012. NEWMAN, D.H., AND D.N. WEAR. 1993. Production economics of
ALLEN, H.L., T.R. FOX, AND R.G. CAMPBELL. 2005. What is ahead private forestry: A comparison of industrial and nonindustrial
for intensive pine plantation silviculture in the south? South. forest owners. Am. J. Agric. Econ. 75:674 – 684.
J. Appl. For. 29(2):62– 69. PATTANAYAK, S., B. MURRAY, AND R. ABT. 2002. How joint is
BIOMASS ENERGY RESOURCE CENTER. 2012. Biomass supply and joint forest production? An econometric analysis of timber
carbon accounting for southeastern forests. Biomass Energy supply conditional on endogenous amenity values. For. Sci.
Resource Center, the Forest Guild, and Spatial Informatics 48(3x):479 – 491.
Group, Montpelier, VT. 132 p. PEREZ-VERDIN, G., D.L. GREBNER, C. SUN, I.A. MUNN, E.B.
BIOMASS RESEARCH AND DEVELOPMENT BOARD. 2008. Increasing SCHULTZ, AND T.G. MATNEY. 2009. Woody biomass availabil-
feedstock production for biofuels: Economic drivers, environ- ity for bioethanol conversion in Mississippi. Biomass Bioen-
mental implications, and the role of research. Available on- ergy 33:492–503.
line at www.usbiomassboard.gov/pdfs/8-increasing-biofuels- PERLACK, R.D., L.L. WRIGHT, A.F. TURHOLLOW, R.L. GRAHAM,
feedstock-production.pdf; last accessed on Apr. 23, 2012. B. STOKES, AND D.C. ERBACH. 2005. Biomass as feedstock for
BRACMORT, K.S., AND R.W. GORTE. 2009. Biomass: Comparison a bioenergy and bioproducts industry: The technical feasibility
of definitions in legislation. Congressional Res. Serv. Rep. of a billion-ton annual supply. Oak Ridge National Laboratory,
R40529, Washington, DC. 12 p. Oak Ridge, TN. 78 p.
DOMKE, G.M., D.R. BECKER., A.W. D’AMATO, A.R. EK, AND POLYAKOV, M., D.N. WEAR, AND R.N. HUGGETT JR. 2010. Harvest
C.W. WOODALL. 2012. Carbon emissions associated with the choice and timber supply models for forest forecasting. For.
procurement and utilization of forest harvest residues for energy, Sci. 56(4):344 –355.
northern Minnesota, USA. Biomass Bioenergy 36:141–150. PRESTEMON, J.P., AND R.C. ABT. 2002. The Southern Timber
FOLEY, T., D. DEB. RICHTER, AND C.S. GALIK. 2009. Extending Market to 2040. J. For. 100(7):16 –22.
rotation age for carbon sequestration: A cross-protocol com- PRESTEMON, J.P., AND D.N. WEAR. 2000. Linking harvest choices
parison of North American forest offsets. For. Ecol. Manage. to timber supply. For. Sci. 46(3):377–389.
259:201–209. REPO, A., M. TUOMI, AND J. LISKI. 2011. Indirect carbon dioxide
FOX, T.R. 2000. Sustained productivity of intensively managed emissions from producing bioenergy from forest harvest resi-
forest plantations. For. Ecol Manage. 138(1–3):187–202. dues. GCB Bioenergy 3:107–115.
FOX, T.R., E.J. JOKELA, AND H.L. ALLEN. 2007. The development ROSSI, F.J., D.R. CARTER, AND R.C. ABT. 2010. Final report:
of pine plantation silviculture in the southern United States. Woody biomass for electricity generation in Florida: Bioeco-
J. For. 105:337–347. nomic impacts under a proposed renewable portfolio standard
GALIK, C.S., R.C. ABT, AND Y. WU. 2009. Forest biomass supply (RPS) mandate, prepared for Florida Department of Agricul-
in the southeastern united states—Implications for industrial ture and Consumer Services Division of Forestry. Available
roundwood and bioenergy production. J. For. 107(2009):69 –77. online at www.floridaforestservice.com/forest_management/
GAN, J., AND C.T. SMITH. 2006. Availability of logging residues fm_pdfs/Final%20Report%20on%20Woody%20Biomass
and potential for electricity production and carbon displace- %20Demand%20and%20Supply.pdf; last accessed on Mar. 12,
ment in the USA. Biomass Bioenergy 30:1011–1020. 2011.
HARDIE, I.W., P.J. PARKS., P. GOTTLIEB, AND D.N. WEAR. 2000. SMITH, J.E., AND L.S. HEATH. 2002. A model of forest floor carbon

538 Forest Science 58(5) 2012

 mass for United States forest types. US For. Serv. Res. Paper mass supply for a bioenergy and bioproducts industry, Perlack,
RP-NE-722. Northeastern Research Station, Newtown Square, R.D., and B.J. Stokes (leads). ORNL/TM-2011/224. Oak Ridge
PA. National Laboratory, Oak Ridge, TN. 227 p.
SMITH, J.E., L.S. HEATH, K.E. SKOG, AND R.A. BIRDSEY. 2006. US ENERGY INFORMATION ADMINISTRATION. 2007. Energy and
Methods for calculating forest ecosystem and harvested carbon economic impacts of implementing both a 25-percent RPS and
with standard estimates for forest types of the united states. US. a 25-percent RFS by 2025. Rep. No. SR/OIAF/2007-05. US
Gen. Tech. Rep. GTR-NE-343. Northeastern Research Station, Department of Energy. Washington, DC. 86 p.
Newton Square, PA. USDA FOREST SERVICE. 2009. Forest Inventory and Analysis
TIMBER MART SOUTH. 2012. South-wide average prices. Available National Program: TPO database. Available online at www.fia.
online at www.timbermart-south.com/prices.html; last ac- fs.fed.us/program-features/tpo/; last accessed Mar. 19, 2009.
cessed Mar. 5, 2012. USDA FOREST SERVICE. 2011. Forest Inventory and Analysis
US DEPARTMENT OF COMMERCE, BUREAU OF THE CENSUS. 2011. NationalProgram:FIAdatabase.Availableonlineatfia.fs.fed.us/
Industry statistics sampler: NAICS 321: Wood product manu- tools-data/default.asp; last accessed Jan. 8, 2011.
facturing. Available online at www.census.gov/econ/industry/ WOODS, R.A. 2009. Industrial output and employment projections
current/c321.htm. Industry statistics sampler: NAICS 322: Pa- to 2018. Monthly Labor Rev. 132(11):52– 81.
per manufacturing. Available online at http://www.census.gov/ WU, J., J. WANG, AND J. MCNEEL. 2011. Economic modeling of
econ/industry/current/c322.htm; last accessed Mar. 11, 2009. woody biomass utilization for bioenergy and its application in
US DEPARTMENT OF ENERGY. 2011. U.S. billion-ton update: Bio- central Appalachia, USA. Can. J. For. Res. 41:165–179.

Forest Science 58(5) 2012 539