The Economic Effects of
Climate Change
on U.S. Forests
Final Report
RCG/Hagler Bailly
-------
The Economic Effects of
Climate Change
on U.S. Forests
Final Report
Prepared by:
Mac Callaway
Joel Smith
Sally Keefe
RCG/Hagler Bailly
P.O. Drawer O
Boulder, CO 80306-1906
(303) 449-5515
Prepared for:
Adaptation Branch, Climate Change Division
Office of Policy, Planning, and Evaluation
U.S. Environmental Protection Agency
Washington, DC 20460
Contract No. 68-W2-0018
Project Manager
Dr. Neil Leary
Chief, Adaptation Branch
Dr Joel D Scheraga
May 18, 1995
-------
Preface
Global climate change poses threats to the environment, human welfare, and human health.
How the United States will respond to these threats is a challenging public policy decision
that will be influenced by a variety of considerations. Among the considerations are the
potential damages that may result from climate change, where damages are evaluated using
economic concepts of human welfare To advance our knowledge of potential damages, the
Climate Change Division of the United States Environmental Protection Agency has
conducted and supported a number of studies to examine different categories of climate
change effects. The Economic Effects of Climate Change on U.S. Forests reports results from
one of the studies supported by our office.
The study draws upon previous work on the potential physical responses of forests to changes
in climate The study moves our state of knowledge forward by evaluating how these changes
may influence the production and consumption of commercial forest products, standing forest
stocks, and the economic rewards reaped by producers and consumers of commercial forest
products The results provide an indication that the values that are derived by our society
from our forests, and that are vulnerable to climate change, are substantial. The magnitudes of
the damage estimates for commercial forestry are sufficiently great to warrant concern and
further consideration as we formulate strategies to address global climate change
The results of the study should not be interpreted as forecasts of the economic consequences
of the effects of climate change on U.S. forests. Information and understanding of how
climate may change in different regions, the responses of individual tree species and forest
communities to changes in climate, the potential for human responses to adapt forestry
practices to a changing climate, and how forestry and forest product markets may evolve in
the future in response to other forces are too imperfect to support reliable forecasts at this
time Instead, the results provide illustrations of the potential severity of the economic
consequences of selected scenarios of climate change impacts.
The study was performed by RCG/Hagler Bailly, under subcontract to ICF Incorporated, for
the Environmental Protection Agency. An earlier version of the report was reviewed by EPA
staff and by experts from outside the agency Participants in the review process are thanked
for their efforts The final report is a report to the EPA and does not necessarily reflect the
views or policies of the EPA.
Neil Leary,
Project Manager
U.S. Environmental Protection Agency
RCG/Hagler Bailly
Final Report
-------
Contents
Executive Summary S-l
Chapter 1 Introduction
1 1 Background 1-1
12 Review of Economic Studies 1-3
1.3 Scope and Objectives 1-4
1.4 References 1-6
Chapter 2 Sources of Information and Methods for Estimating Changes in
Forest Yields
2.1 Data Needs 2-1
2.2 Information Sources 2-1
2 3 Climate Change Scenarios 2-5
2.4 Using Site-Specific Information to Estimate Regional Changes in
Growing Stock . 2-9
2 5 References 2-11
Chapter 3 Results: Yield Changes
3 1 Southeast and South Central Regions ... 3-1
3 2 North Central . . 3-3
3 3 Northeast . ... .3-5
3 4 Pacific Northwest and Pacific Southwest 3-6
3 5 Rocky Mountains . . . . . 3-7
3.6 National Results 3-8
3 7 Conclusion .... ... 3-11
3 8 References 3-12
Chapter 4 FASOM Model and Economic Analysis Methods
4 1 Model Overview .... .... 4-1
4.2 Forest Sector Detail . . 4-4
4.2 1 Product Demand Functions 4-5
4 2.2 Inventory Structure and Dynamics 4-6
4.2.3 Production Technology, Costs, and Capacity Adjustment . . 4-8
4 2 4 The FASOM Tableau: Forestry Component . 4-9
RCG/Hagler Bailly
Final Report
-------
Contents ~ page ii
4.3 Approach, Inputs, and Limitations of Analysis 4-9
4 3.1 Approach 4-11
4.3.2 Inputs: Yield Changes by Region and Species 4-13
4.3.3 Limitations of the Economic Analysis 4-15
4.4 References 4-16
Chapter 5 Results of the Economic Analysis
5.1 The Base Case Summarized 5-1
5.2 Climate Change Scenarios 5-2
5.2.1 Welfare 5-2
5.2.2 Product Prices 5-8
5.2.3 Production (Harvests) 5-12
5.2.4 Inventory Levels 5-16
5.3 Substitution Issues 5-20
5.4 References 5-29
Chapter 6 Conclusions and Suggestions for Further Research
6 1 Conclusions 6-2
6 2 Suggestions for Further Research . . 6-3
Appendix A Details on Estimating Changes in Growing Stock of Softwoods in Pacific
Northwest, Pacific Southwest, and Rocky Mountain Regions
Appendix B Tables
Appendix C Harvest and Management Decision Model
RCG/Hagler Bailly
Final Report
-------
Tables
S-l Percent Changes in Yields by Region and Species for Four Climate Scenarios . . S-3
S-2 Annualized Value of Welfare Components for the Base Case and Percent Changes
from the Base Case for Four Climate Change Scenarios . . S-4
2-1 Gap Model Studies 2-3
2-2 GCM Runs Used in Gap Model Studies 2-8
3-la Change in Southeast Softwood Biomass 3-1
3-lb Change in Southeast Hardwood Biomass 3-2
3-2a Change in South Central Softwood Biomass 3-2
3-2b Change in South Central Hardwood Biomass 3-2
3-3a Change in North Central Softwood Biomass 3-4
3-3b Change in North Central Hardwood Biomass 3-5
3-4 Change in Northeast Biomass 3-6
3-5 Change in Softwood Biomass in Pacific Northwest and Pacific Southwest 3-7
3-6 Change in Softwood Biomass in Rocky Mountains .... 3-8
3-7a Regional Changes in Softwood Growing Stock 3-9
3-7b Regional Changes in Hardwood Growing Stock 3-9
3-8 Percent Change in National Growing Stock of Private Forests .... . . . . 3-10
4-1 Sample Tableau for Overall FASOM Model Emphasizing Forestry Component 4-10
4-2 Percent Changes in Yields by Region and Species for Four Climate Scenarios . 4-14
5-1 Annualized Values of Welfare Components for the Base Case and Percent
Changes from the Base Case for Four Climate Change Scenarios .... 5-3
5-2 Annualized Values of Welfare Components for the Base Case and Revised
Base Case Compared with Original and Revised 4 0 without COa Scenarios . . 5-28
RCG/Hagler Bailly
Final Report
-------
Figures
2-1 Typical Gap Model Relationship between Growing Degree Days and Growth . . 2-4
2-2 Location of Sites from Published Gap Model Studies 2-10
3-1 Percent of Change in National Growing Stock of Private Forests 3-11
4-1 FASOM Regions 4-2
5-1 Projected Producer Surplus for the North 5-5
5-2 Projected Producer Surplus for the South 5-6
5-3 Projected Producer Surplus for the West 5-7
5-4 Projected Tornquist-Theil Softwood Price Index 5-9.
5-5 Projected Tornquist-Theil Hardwood Price Index 5-10
5-6 Projected Tornquist-Theil Price Index for Forest Products Prices 5-11
5-7 Projected Softwood Production 5-13
5-8 Projected Hardwood Production 5-14
5-9 Projected Total Production 5-15
5-10 Projected Softwood Inventory Volume 5-17
5-11 Projected Hardwood Inventory Volume 5-18
5-12 Projected Total Inventory Volume 5-19
5-13 Projected Price Path Softwood Pulpwood 5-22
5-14 Projected Price Path Hardwood Pulpwood 5-23
5-15 Projected Softwood Pulpwood Price Path 5-25
5-16 Projected Hardwood Pulpwood Price Path 5-26
5-17 Projected Tornquist-Theil Price Index for Forest Products Prices 5-27
RCG/Hagler Bailly
Final Report
-------
Executive Summary
Global climate change poses risks to the forests of the United States. Rising temperatures,
changes in precipitation and soil moisture, and other changes in climate will affect both the
growth rates of different tree species and the competition among them for nutrients and
sunlight. This will result in changes in the species composition of forest stands and, with this,
perhaps wider changes in the distribution of forest types across the United States. The exact
character of these effects is uncertain and will be determined both by complex physical and
biological processes and by management responses. Some of the studies that have analyzed
the effects of climate change on forests suggest that climate change could adversely impact
both forest health and productivity in the United States The impacts projected in these studies
are most severe in the southeastern and south central states, where the growth rates of
plantation and natural pines could be severely reduced, leading to decreases or the
disappearance of these species in some areas. Such changes would have adverse consequences
on economic well-being in the U.S due to reduced timber supplies and higher wood and
paper product prices. These impacts on forests could also have potentially severe
environmental consequences in the form of reductions in wildlife habitat, biodiversity,
watershed protection, aesthetic values and forest-based recreation.
This study examines one category of potential losses from the effects of climate change on
U.S forests, the economic losses associated with the production and consumption of
commercial timber products Previous studies have explored the potential effects of climate
change on U S forests in physical terms, such as impacts on geographic distributions and
physiological effects on photosynthetic rates, leaf area, biomass, and yields. The objective of
the present study is to extend the previous research by evaluating the economic consequences
of the physical changes in commercial forests on timber markets in the United States
To evaluate the economic consequences of the effects of climate change on commercial
forests, scenarios of hypothesized yield changes are developed based on published results of
site specific forest responses to climate change The majority of published results relied upon
come from studies that use forest gap models to project the potential responses of forested
stands to changes in climate Other types of simulation models, such as biogeography and
biogeochemistry models, also can and have been used to assess the potential responses of
forests to climate change and the results have been found to vary significantly across the
various models used in these analyses. However, with a few exceptions, the results from these
studies were not published or available until all of the substantive research on this study had
been completed. The results now becoming available from these other methodologies indicate
that the projections of the forest gap models are within the range of those from other
methods, but are nearer to the lower end of the range (i.e., greater losses of forest biomass, or
smaller gains, depending upon the species and region).
RCG/Hagler Bailly
Final Report
-------
Executive Summary ~ S-2
The projected yield changes used in this study are presented in Table S-l. They vary by
region and by wood type—softwood and hardwood. Four scenarios were developed for the
analysis. The scenarios differ in the assumed amount of wanning and in the inclusion or
exclusion of beneficial effects of carbon dioxide on photosynthetic rates and water use
efficiency. The four scenarios can be characterized as displaying severe reductions in
softwood yields, particularly in the South. Hardwood yield changes are modest in comparison,
though still substantial, and range from yield increases in the Northeast and North Central
regions to yield decreases in the Southeast and South central regions.
The hypothesized yield changes are used as inputs to FASOM, a model of U.S. stumpage
markets. FASOM is a dynamic, nonlinear programming model of the forest sector in the
United States. It simulates the production and consumption of sawtimber, pulp, and fuelwood
in nine forest regions1 of the United States. It also projects product prices, forest inventories,
and consumer and producer surpluses. Simulations were performed for the four climate
change scenarios plus a base case scenario for comparison. FASOM incorporates adaptation to
climate change in the sense that it simulates the impacts of reduced yields on the relative
profitability of different timber products. Landowners respond by altering a wide range of
management decisions related to planting/regeneration, tending of planted stands, and
harvesting, consistent with intertemporal profit maximization. Buyers of stumpage optimize by
shifting the quantity and mix of forest products purchased in response to price signals.
However, adaptations in management practices that would mitigate the yield changes
themselves, such as the introduction of heat and drought tolerant species into a region, are not
reflected in the analysis.
The analysis was performed, assuming that the yield changes in Table S-l reflected an
equilibrium adjustment by forests to the changes in climate. The yield changes were
introduced during the first decade of the model projections, 1990-2000. Thus, the estimates of
annualized value of damages are higher than could be expected if, as one would expect, forest
yields change at a rate roughly in proportion to the rate of climate change, over time. On the
other hand, this approach gives a clearer picture of long-run impacts on welfare, harvest
levels and stumpage prices and forest inventory stocks.
The potential welfare impacts of the four scenarios are summarized in Table S-2. They are
consistent with the long-run consequences of a 2.5° and 4.0°C warming. The simulations
project substantial losses in economic welfare. On an annualized basis, losses vary from
$2.5 billion to $12 billion in 1990 dollars across the scenarios, or roughly 4 to 20% of the net
value of commercial forests to society The losses are very unevenly distributed across
consumers and producers. Consumers, facing prices that are 100 to 250% higher than in the
base case scenario, consume 20 to almost 50% less forest products. Consequently, consumer
welfare, measured by the change in ordinary consumer surplus, declines by $10 billion to
1 These regions are the Pacific Northwest-West, The Pacific Northwest-East, the Pacific Southwest,
the Rocky Mountains, the Lake States, Corn Belt, Southeast, South central, Northeast
RCG/Hagler Bailly
Final Report
-------
Executive Summary ~ S-3
Table S-l
Percent Changes in Yields by Region and Species
for Four Climate Scenarios
Region
Scenarios
2.5 with
C02 Change
2.5 without
C02 Change
4.0 with C02
Change
4.0 without
C02 Change
Softwood
Pacific Northwest-West
20
-10
-32
-53
Pacific Northwest-East
20
-10
-32
-53
Northeast
-40
-55
-68
-78
Lakes States
-51
-64
-66
-76 "
Corn Belt
-51
-64
-66
-76
Southeast
-56
-83
-76
-100
South Central
-75
-87
-76
-100
Rocky Mountains
70
30
79
33
Pacific Southwest
20
-10
-32
-53
Hardwood
Pacific Northwest-West
0
0
0
0
Pacific Northwest-East
0
0
0
0
Northeast
46
14
71
36
Lakes States
39
11
53
21
Corn Belt
38
11
53
21
Southeast
20
-10
-8
-30
South Central
0
-20
-44
-60
Rocky Mountains
0
0
0
0
Pacific Southwest
0
0
0
0
RCG/Hagler Bailly
Final Report
-------
Executive Summary ~ S-4
Table S-2
Annualized Value1 of Welfare Components
for the Base Case and Percent Changes from the Base Case
for Four Climate Change Scenarios
Base
2.5 with
2.5 without
4.0 with
4.0 without
Welfare Component
(S Millions)
co2
co2
co2
co2
Consumer Surplus
56,748
-17.42%
-33.47%
-28.34%
-42.09%
Producer Surplus
3,871
145.67%
237.40%
194.50%
207.69%
Foreign Trade Surplus
324
-21.79%
-39.34%
-35.62%
-50.10%
Terminal Inventory
1,805
-6.13%
-13.89%
-11.33%
-33.13%
Public Cut
2,023
79.56%
159.84%
137.05%
222 87%
Total Surplus
64,771
-4.35%
-10.73%
-9.42%
-18.68%
1 Annualized values calculated for 1990-2080 using a discount rate of 4%
$24 billion per year. In contrast, producers benefit from the higher prices, and reap a welfare
gain of $6 billion to $8 billion per year despite the reduced volume of marketed harvests. The
net effect in all scenarios, however, is to decrease the benefits to society provided by
commercial forest harvests by billions of dollars annually
In addition to the welfare losses, the scenario simulations also raise concerns about the
sustainability of softwood forest resources in the United States in a warmer climate Projected
harvests are typically greater than net growth, causing softwood forest inventories to decline
fairly dramatically by 2060. The projected trends also suggest that inventories will continue to
decline after 2060 for the most severe scenarios. Sustainable production of softwood forest
products would therefore require reductions in harvests that are considerably larger than those
projected to occur in response to market price increases under conditions of severe yield
decreases.
The results of the study indicate that the economic consequences of climate change for
commercial forestry are potentially severe The results, however, are not without qualification.
The hypothesized yield changes are based on gap model results for unmanaged forests, which
may overestimate the sensitivity of forests to climate change. The omission of forest
management adaptations that might mitigate yield losses may result in further overestimation
of impacts. Also, the potential effects of elevated carbon dioxide levels on forests are highly
uncertain and the speculative yield increases from carbon dioxide included in some scenarios
RCG/Hagler Bailly
Final Report
-------
Executive Summary ~ S-5
could either overestimate or underestimate these gains. Another limitation of the study is the
omission of changes in Canadian forest yields and their potential effects on U.S forest
product markets. Increases in softwood yields in Canada could offset domestic supply losses,
providing gains to consumers but possible losses to domestic producers. Other limitations
include the absence of adjustment in land uses between agricultural, forest, and other uses;
uncertainty about the possibilities for substitutions among species, among wood products, and
between wood products and other materials; and uncertainty about future technologies. Each
of these limitations represents directions in which the work presented in this report might be
extended to improve our understanding of the economic consequences of the effects of
climate change on commercial forests.
RCG/Hagler Bailly
Final Report
-------
Chapter 1
Introduction
1.1 Background
Climate is a major determinant of the range, type, and density of forests. The northern
geographic limit of forests is determined by minimum temperatures, while the southern limit
is influenced by such factors as soil moisture and competition from species better adapted to
warmer climates (Davis, 1989) Changes in average climate conditions in pre-history, such as
the Ice Ages, resulted in large scale shifts in location of forests. Oaks, for example, which are
now found all over the eastern United States, only existed in the Deep South at the height of
the last Ice Age (Webb, 1992)
Climate change will most likely have a significant effect on the range, type, and density of
forests in the United States (VEMAP Participants, submitted; Smith and Shugart, 1993,
Westman et al., 1990; Smith and Tirpak, 1989, Shands and Hoffman, 1987) Warmer
temperatures may enable forests to grow in higher latitudes and at higher altitudes. Warmer
temperatures may also result in drier soils and northward expansion of pests and diseases,
which could cause rapid dieback along the current southern and low elevation boundaries of
forests (Smith and Shugart, 1993) In addition, warmer adapted species will most likely
migrate north or upslope, driving out species better adapted to cooler climates
The literature on climate change impacts on forests contains very different conclusions about
whether forest density (i e., biomass) will increase or decrease Neilson (1993), for example,
using the MAPPS model found that forest cover in the coterminous United States could
decrease by 30 to 94%. Studies of forests in the South using gap models have found that
there could be significant reductions in forest biomass, particularly for softwood species
(Urban and Shugart, 1989, Solomon, 1986) On the other hand, studies such as Melillo et al
(1993), which used a bio-geochemical model found that productivity of many forests in the
United States could significantly increase (Many of these studies are discussed in more detail
in Chapter 2.) Thus, the future fate of forests in the United States appears to be uncertain.
Forests in the United States produce many different types of market and nonmarket services.
The most visible of these is timber supply from public and private lands. There are roughly
740 million acres of forestland1 in the United States. Of that amount, roughly 490 million
1 Forestland is any land stocked with at least 10 percent of trees of any kind or size.
RCG/Hagler Bailly
Final Report
-------
Introduction ~ 1-2
acres are in timberland.2 About 75 percent of all timberland in the United States is privately
held, either by industrial or nonindustrial private owners (Waddell et al., 1989). In 1986, the
value of timber harvested in the United States was approximately $12.67 billion, with
softwoods contributing over 80% of the value before manufacturing (USFS, 1990). Forests
also provide a variety of nonmarket services including recreation and aesthetics, wildlife
habitat and ecosystem diversity, and the control of surface water runoff. However, the value
of these services are more difficult to quantify and value in monetary terms.
Changes in the location, composition, and abundance of forests have the potential to affect
timber markets and the welfare of individuals and firms that own timberland, supply inputs
for timber harvesting and management and who purchase stumpage for milling and
transformation into primary and secondary products. In general, reductions in forest
productivity will make it more expensive to supply timber to the market in each period,
leading to higher stumpage prices and reduced harvests, while increases in productivity will
have the opposite effect. Forest inventories will be affected directly by changes in forest
yields, however, adverse impacts due to yield decreases will be choked off to some extent by
reductions in harvests. If forest yields decline, as suggested by the studies reviewed in this
report, higher prices will always lead to reductions in consumer welfare; however, the welfare
of producers may rise if products demands are inelastic and the reductions in inventories is
not too severe. If changes in productivity are unevenly distributed across species and regions,
the economic impacts will also be uneven. Commercial operations such as milling and
primary and secondary processing of wood products may move closer to timber areas that
have become relatively more productive Timber and wood product production could even
shift to Canada if warmer temperatures have a strong beneficial affect as one moves
northward
Climate change almost certainly will affect the nonmarket services provided by forest
ecosystems At the very simplest level, climate-induced increases in forest area are likely to
enhance such services as wildlife habitat, ecosystem diversity, flood control and carbon
storage, while decreases in forest area would, in all probability, adversely impact the ability
of forests to supply these and other kinds of nonmarket services. However, our knowledge
about the relationship between climate and forest ecosystems and the availability of data and
methods for quantifying and valuing these impacts are not well developed enough at this
stage to perform national assessments Therefore this study does not include them, nor does it
dismiss them as unimportant
Timberland is forested land that can produce at least 20 cubic feet per year of industrial wood and
is not reserved.
RCG/Hagler Bailly
Final Report
-------
Introduction »> 1-3
1.2 Review of Economic Studies
To date, no published study has examined the economic effects of climate change on timber
markets in the United States based on estimates of yield changes found in the science
literature.3 In that regard, this study appears to be a first However, there are a handful of
published studies that looked at the economic effects of changes in forest yields from air
pollution and other undifferentiated sources. This literature provides the background to the
current study.
An early study by Callaway et al. (1986) used the TAMM model (Adams and Haynes, 1980)
to conduct a sensitivity analysis of the economic impacts of hypothetical reductions in timber
yields, ranging from -10% to -20%. The authors speculated that these yield changes could be
due to a variety of sources, including acid rain and C02-induced changes in temperature.
They found that the present value of damages between 1985 and 2030 ranged from $3 4
billion to $5 0 billion ($0 34 to $0 52 billion in annualized terms), depending on the severity
of the hypothetical growth reduction. These welfare losses are in the range of 5% to 1% of
total welfare in the base case.
A subsequent study by Haynes and Kaiser (1991) used a newer version of the TAMM model
to examine the effects of reductions in timber yields due to acidic deposition. These yield loss
estimates were based on expert opinions published by de Steigeur and Pye (1988) and were
-5% for hardwood species and in the range of -5% (North) to -10% (South) for softwoods
Haynes and Kaiser reported annual welfare losses (expressed in future values) ranging from
$0.6 billion in 2000 to $3.0 billion in 2040
This study was followed up by one by Callaway (1991) in which he used that newer version
of TAMM to estimate the economic impacts of reductions in southern pine yields (planted
and natural) in the range of +2% to -10% The annualized welfare impacts associated with
these yield reductions ranged from an increase of about +$40 million per year for the +2%
case to about -$110 million for the -10% case
All these studies shared two common features They were all sensitivity analyses and the
yield changes used in the analyses did not vary much, if at all, from species to species or
region to region In addition, all these studies used the TAMM model. TAMM is a spatial
price and equilibrium model, and as such it is basically a good tool for examining the
economic effects of yield changes from a variety of sources However, TAMM has one
important limitation in this regard. Management of the forest inventory, including harvesting,
is linked only to current period prices Thus, TAMM does not do a very good job of
simulating economic behavior in the market for timberland. This is an important limitation
3 An unpublished study by Sohngen and Mendlesohn, "Integrating Ecology and Economics: The
Timber Market Impacts of Climate Change," was not available until the final revisions to this report were
being made This study shows increases in both forest yields and economic welfare due to climate change
RCG/Hagler Bailly
Final Report
-------
INTRODUCTION 1-4
because it means that the behavior of landowners in current periods does not anticipate
reduced yields in future periods, through the signal of land prices. Thus, producers in TAMM
are always in a "reactive" mode, and welfare losses are larger than they would be if
timberland owners were able to anticipate the reductions in yields over longer periods.
Recently, Burton et al. (1994) presented a paper under the auspices of the Southern Global
Change Program on the economic effects of climate change on Southern forests The paper is
important because it relied on the same model being used in this study, FASOM, (Forest and
Agricultural Sector Optimization Model) as the vehicle for estimating these impacts.
However, like all the previous studies cited above, the yield changes used in this analyses
were not based on studies of the effects of changes in climate on forest yields, but were
instead framed in the context of a sensitivity analysis. The authors postulated changes of ±5%
and ±10% in the annual growth rate of all species in all regions of the United States. Changes
in the net present value of welfare ranged from +0.6% (+10% case) to -0.6% (-10% case)
This study showed that the market behavior simulated in FASOM could greatly mitigate the
effects of reductions in yields, although the authors did not describe in any detail the
underlying economic logic for this.
Finally, Cline (1992) estimated economic losses to forests as a result of climate change using
a fairly crude, but transparent, top-down approach. He very roughly estimated the annual
average value of stumpage at $10 billion He used studies cited by EPA to fix expected
reductions in forest productivity at about 40% per year and then reduced the stumpage value
accordingly to around $4 billion annually, in the absence of any mitigation. Cline then
estimated that forest management costs would have to double or triple to reduce forest losses
by one-third. Taking into account all three factors—losses without mitigation, plus the costs
of reducing damages through forest management, minus the losses avoided by management
measures—Cline estimated that net annual losses would be in the neighborhood of $3 3
billion.
1.3 Scope and Objectives
The study described in this report differs from previous ones in two distinct ways First, the
yield changes used in the analysis are based on results contained in the published literature,
based on a fairly sophisticated extrapolation methodology. Second, for the first time, the
analysis was performed with a dynamic economic model that simulates land owner behavior
in a way that is perfectly consistent with intertemporal optimization.
Accordingly, the objectives of this study were to use the FASOM model to estimate the
impacts of these changes in yields on:
*¦ the annualized value and distribution of consumer and producer surplus
~ the regional distribution of producer surpluses, over time
RCG/Hagler Bailly
Final Report
-------
Introduction ~ 1-5
~ stumpage product prices, over time
~ stumpage production, by product, over time
~ inventory levels, by species, over time.
In this study, we used published information on climate change impacts on yields of
hardwood and softwood species to estimate changes in consumer and producer surplus from
commercial use of forests The estimates of changes in yields are based on published results
of so-called "gap models." While these models do not represent the entire range of potential
changes in forest yields from climate change (see Chapter 2), they are the only set of studies
that give species specific results. Based on site specific results from the gap model studies,
we estimated changes in hardwood and softwood yields for four scenarios:
* a 2.5°C warming with carbon dioxide fertilization4
» a 2.5°C warming without carbon dioxide fertilization
~ a 4°C warming with carbon dioxide fertilization
* a 4°C warming without carbon dioxide fertilization.
We then used FASOM (Adams et al, 1994), to evaluate the impact of these changes in forest
productivity on welfare, production (i.e., harvests), and forest inventory stocks in stumpage
markets in the United States FASOM is an optimization model that simulates the growth,
harvesting, and sale of timber products, in nine regions of the coterminous United States. It
provides information about the effects of changes in forest yields on inventory volumes and
acreage, on timber harvest and stumpage prices for sawtimber pulpwood and fuelwood, and
on various components of producer and consumer welfare. A more complete discussion of the
model and definitions of the welfare measures used in this report are contained in Chapter 4
The important feature about FASOM that differentiates it from other economic models used
to look at the impacts of timber yield changes on stumpage markets is that FASOM is not a
static or a myopic economic model It incorporates the future price expectations into the
simulated investment behavior of timberland owners. What this means is that timberland
owners are able to look into the future and adjust their planting and harvesting decisions,
now, based on an understanding of future resource scarcity (or abundance). This allows for
greater flexibility in the decision-making behavior that is incorporated into the model and
more closely simulates how landowners actually behave. Because FASOM does include this
type of mechanism, welfare losses due to lower yields, as projected by FASOM, would be
reduced.
This report describes the methods and results from the research. The research was conducted
in two phases. In the first phase, we estimated changes in hardwood and softwood yields
based on results of studies in the literature. The first phase is described in Chapters 2 and 3.
4 The C02 effect assumes actual doubling of atmospheric C02 levels to about 700 ppm.
RCG/Hagler Bailiy
Final Report
-------
INTRODUCTION ~ 1-6
Chapter 2 describes the methods used to estimate the impacts of climate change on forest
productivity, including the scenarios. Chapter 3 describes how estimates of site specific
changes in yields due to climate change were used to estimate changes in hardwood and
softwood yields in nine U.S. Forest Service regions.
In the second phase of our analysis, we used the yield changes to estimate changes in timber
production and societal welfare. The methods and results from this phase are described in
Chapters 4 and 5. Chapter 4 describes the FASOM model, with special reference to the forest
sector. This chapter also describes the procedures for introducing the yield changes into the
FASOM model and shows how the yields vary by species and region for each scenario.
Chapter 5 presents the main economic results of the analysis in tabular and graphic form.
Chapter 6 provides the major conclusions of the study and outlines future research needs to
improve the quality of the estimated economic impacts. Appendix A outlines the specific
assumptions used to estimate changes in growing stock in the western United States
Appendix B contains the tabular information that was used to develop the figures presented in
Chapter 5. Lastly, Appendix C contains documentation on the mathematical structure of the
forest sector in FASOM.
1.4 References
Adams, D M and R.M Haynes. 1980. "The 1980 Softwood Timber Assessment Market
Model Structure, Projections, and Policy Simulations " Forest Science Monograph 22,
supplement to Forest Science. September. 1-64.
Adams, D., R Alig, J. Callaway, and B. McCarl. 1994. Forest Sector Optimization Model:
Model Description. RCG/Hagler Bailly, Boulder, CO
Burton, D.M., B.A McCarl, D.M Adams, R. Alig, J.M. Callaway, and S.M. Winnett. 1994.
"An Exploratory Study of the Economic Impacts of Climate Change on Southern Forests.
Preliminary Results," Paper prepared for the Southern Global Change Program, Raleigh, NC.
Callaway, J.M. 1991. "Economics," Section 4 9 in 1990 Integrated Assessment Report, The
NAPAP Office of the Director, 722 Jackson Place, NW, Washington, DC, pp. 392-398.
Callaway, JM., RF. Darwin, and R.J. Nesse. 1986. "Economics Effects of Hypothetical
Reductions in Tree Growth in the Northeastern and Southeastern United States " PNL-5939
Battelle Pacific Northwest Laboratory, Richland, WA.
Cline, W.R. 1992. The Economics of Global Warming. Washington, DC: Institute for
International Economics.
RCG/Hagler Bailly
Final Report
-------
Introduction ~ 1-7
Davis, M.B., 1989. "Lags in Vegetation Response to Greenhouse Warming." Climatic
Change. 15:75-82.
de Steigeur, J.E. and J.M. Pye. 1988. "Using Scientific Opinion to Conduct Forestry Air
Pollution Economic Analyses," in Proceedings of the Symposium on the Economic Assessment
of Damage Caused to Forests by Air Pollutants. IUFRO Working Party S4,04-02,
September 13-17, Gmunden, Austria.
Haynes, R.W and H.F. Kaiser. 1991. "Methods for Valuing Acidic Deposition/Air Pollution
Effects: Forests," in SOS-T 27: Methods for Valuing Acidic Deposition and Air Pollution
Effects, Brown, G.M. and J M Callaway, eds., The NAPAP Office of the Director, 722
Jackson Place, NW, Washington, DC, pp. 109-121.
Melillo, J.M., A.D. McGuire, D.W. Kicklighter, B. Moore, C.J. Vorosmarty, and A L.
Schloss 1993. "Global Climate Change and Terrestrial Net Primary Productivity " Nature
363:234-240.
Neilson, RP 1993 "Vegetation Redistribution: a Possible Biosphere Source of C02 During
Climate Change " Water, Air, & Soil Pollution. 70:659-673.
Shands, W.E. and J.S Hoffman (eds.) 1987. The Greenhouse Effect, Climate Change, and
U.S. Forests. Washington, DC: The Conservation Foundation.
Smith, J B. and D Tirpak (eds). 1989. The Potential Effects of Global Climate Change on the
United States. Washington, DC: U.S Environmental Protection Agency, EPA-230-05-89-050
Smith, T.M. and HH Shugart, 1993. "The Transient Response of Terrestrial Carbon Storage
to a Perturbed Climate." Nature 361:523-6.
Solomon, A.M. 1986. "Transient Response of Forests to C02-Induced Climate Change
Simulation Modeling Experiments in Eastern North America." Oecologia 68: 567-579
Urban, D.L. and HH. Shugart 1989. "Forest Response to Climatic Change: A Simulation
Study for Southeastern Forests." In The Potential Effects of Global Climate Change on the
United States, Appendix D: Forests, edited by J.B. Smith, and D. Tirpak. Washington, DC
U.S Environmental Protection Agency, EPA-230-05-89-054
(USFS) U.S. Forest Service. 1990. Rocky Mountain Forest and Range Experiment Station,
Fort Collins, Colorado, General Technical Report RM-199.
Waddell, K.L , D.D Oswald, and D S. Powell. 1989 Forest Statistics of the United States,
1987. U S. Forest Service. Portland, Oregon- Pacific Northwest Research Station, Resource
Bulletin RNW-RB-168.
RCG/Hagler Bailly
Fmal Report
-------
Introduction ~ 1-8
Webb, T. 1992. "Past Changes in Vegetation and Climate: Lessons for the Future." in: Peters,
R.L. and T.E. Lovejoy (eds.) Global Warming and Biological Diversity. New Haven: Yale
University Press.
Westman, W. et al. 1990. "Natural Terrestrial Ecosystems." in: Houghton, J.T., G.J. Jenkins
and J.J. Ephraums (eds.). Climate Change: The IPCC Scientific Assessment. Cambridge
University Press. New York.
RCG/Hagler Bailly
Final Report
-------
Chapter 2
Sources of Information and Methods for
Estimating Changes in Forest Yields
This chapter describes the methods used to estimate changes in yields of commercially
important hardwood and softwood forests The estimates are based solely on published
scientific literature on the biophysical effects of climate change on forests. The chapter
includes discussions on data needs for conducting this analysis, the published studies on
biophysical effects of climate change, the climate change scenarios used in those studies,
uncertainties about climate change and its effects on forests, and how regional estimates of
changes in hardwood and softwood growing stock for different climate change scenarios were
derived.
2.1 Data Needs
FASOM requires information on hardwood and softwood yields for the nine U.S Forest
Service regions (see Chapter 4). To be useful for this study, information from the literature on
climate change effects on forests had to indicate hardwood and softwood yields at the
regional level Studies that estimated changes in the range of forests or total forest biomass or
only gave information at a national level would not be used for this study. The next section
describes the types of studies of climate change effects on vegetation, whether they provide
the information needed for this analysis, and results and limitations of the studies we used
2.2 Information Sources
We evaluated the published literature on climate change impacts on forests based on whether
it provided information on hardwood and softwood yields in the nine U.S Forest Service
regions Three general types of models have been used to examine climate change impacts on
forests:
~ biogeography models, which simulate the dominance of competing forms of
vegetation based on ecophysiological and resource constraints
~ biogeochemistry models, which simulate the cycling of carbon, nutrients and
water as influenced by climate and environmental conditions
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-2
~ community competition (gap) models, which simulate competition among tree
species in response to changes in temperature and moisture.
Since FASOM requires yield information by species and region, highly aggregated estimates
of total forest productivity were not sufficient. Biogeography models (e.g., Smith et al.,
1992a) and biogeochemical models (e.g., Melillo et al., 1993) estimate changes in conditions
of major forest types, such as cool temperate mixed forests1 and, do not produce information
on changes in yields in hardwoods and softwoods. Although some forest types, such as
temperate deciduous forests, are either hardwood or softwoods, many of the classifications
contain mixtures of both hardwoods and softwoods. Thus, neither type of model gives
sufficient detail about changes in hardwood and softwood yields.
To obtain the information needed for this analysis, we used the results from community
competition, or gap models. This class of models derives its name from the fact that they
simulate the growth of trees when gaps in a forest canopy are created by the death of a large
tree (Shugart, 1987) These also simulate the site specific growth of tree species over time 2
The models are physiologically based and have been used to estimate the effects of climate
change on hardwood and softwood trees over much of North America. Table 2-1 lists the gap
model studies used in this analysis.
Gap models are considered to be highly credible tools for examining forest response to
environmental stress (Dale and Rauscher, 1994) These models have been applied and
validated for different forest types around the world (Shugart et al., 1992). Yet there are a
number of limitations in using gap models as a basis for estimating how climate change will
affect private forests The first is that they have been used to examine only unmanaged
forests Gap models have not been used to examine forests that are harvested, thinned, or
replanted—that is, forests that are used for timber, among other purposes. Timber forests are
the subject of this study 3
The second limitation is that gap models tend to be very sensitive to changes in temperature.
They use an inverse parabola to describe the relationship between temperature and tree
species growth (see Figure 2-1). Using this approach, a 4°C warming would translate into an
1 See VEMAP Participants, submitted. "Vegetation/Ecosystem Modeling and Analysis Project
(VEMAP). Assessing Biogeography and Biogeochemistiy Models in a Regional Study of Terrestrial
Ecosystem Responses to Climate Change and C02 Doubling." Manuscript obtained from J.M Melillo,
VEMAP coordinator, The Ecosystem Center, Marine Biological Laboratory, Woods Hole, MA 22903.
2 This feature also makes gap models useful as in input into FASOM. As is discussed below, most
of the gap model results are for static climate change scenarios.
3 Most studies of climate change impacts on forests have focused on forests or vegetation types
unaffected by human activities
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-3
Table 2-1
Gap Model Studies
Study
Region
Climate Change Scenarios
Solomon, 1986
Eastern
United States
UKMO 2XC02 and 4XC02
Pastor and Post, 1988
Northeast
United States
UKMO 2XC02
Botkin et al., 1989
Great Lakes States
GISS, GFDL, OSU, 2XC02, GISS Transient
Urban and Shugart, 1989
Southeast
GISS, GFDL, OSU, 2XC02, GISS Transient
Urban et al, 1993
Pacific Northwest
GISS, OSU, 2XC02
King and Tingey, 1992
Pacific Northwest
GISS, OSU, 2xC02, GISS Transient
Running and Nemani, 1991
Rocky Mountains
+4°C, +10% Precipitation
increase of approximately 1,000 growing degree days, which can result in a dramatic
reduction in tree growth (Urban et al, 1993). The specification of the parabola is based on the
observed range of tree species, with the temperatures at the northern boundary used to
determine the minimum temperature tolerance and the temperatures at the southern boundary
used to determine the maximum temperature tolerance Although the northern or upper
elevation boundaries of trees are often limited by cold temperatures, the southern or lower
elevation limits are often influenced by reduced soil moisture or competition from species
better adapted to warmer climates Both soil moisture and the range of competitive species
are influenced by temperature. Raising temperatures and leaving precipitation unchanged will
most likely result in drier soils Yet, climate change could cause both temperature and
precipitation to increase, which could result in wetter soils. Many tree species could survive
warmer and wetter conditions, and perhaps even prosper, at least until more southern species
migrate into the region and outcompete them (Smith et al., 1992b) 4 On the other hand, the
gap models do not account for such climate change induced factors as increased wildfires,
pests, and pathogens, which could accelerate dieback of forests (Shugart et al., 1992)
On the whole, gap models may overestimate the extent to which forests may die back in
response to higher temperatures. This problem is particularly acute in the Southeast. The
observed southern boundary of many forest species in that region may be limited by the
4 This situation could be exacerbated because some gap models may overstate the degree of drought
stress (Bugman and Martin, 1995).
RCG/Hagler Bailly
Fmal Report
-------
Sources and Methods for Estimating Changes *¦ 2-4
Figure 2-1
Typical Gap Model Relationship between Growing Degree Days and Growth
1.2
Species 1
~" Species 2
0.8
£
"5
oc
•C
%
O
5
0.6
o
>
"3
0)
cr
0.4
0.2
0
1
2
3
4
5
6
7
Growing Degree-Days (Thousands)
Source: Based on Urban and Shugart, 1989.
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-5
presence of the Gulf of Mexico. So it is uncertain whether these forests could survive in
hotter conditions (Smith and Tirpak, 1989).
The third limitation is that the gap model studies, for the most part, did not estimate whether
heat-and drought-tolerant species not currently found in the subject region would grow under
climate change. Forest managers may replant forests with more heat- and drought-tolerant
species Thus, the studies may have overestimated the decrease in yields, particularly of the
softwood species grown in managed forests One exception to this limitation is Solomon
(1986), who found that some northern sites could support southern pines under some
scenarios.
The final limitation is that, with the exception of the Running and Nemani (1991), all of the
studies used in this analysis did not account for C02 fertilization. Increased concentrations of
atmospheric C02 will, without climate change, increase growth and decrease demand for
water by trees (Bazzaz and Fajer, 1992). Most studies on C02 fertilization have been
conducted in laboratories or under controlled conditions There is significant uncertainty about
the degree to which enhanced C02 concentrations will affect trees outside the laboratory
Box 2-1 summarizes the results of the studies using gap models and Box 2-2 briefly describes
the results of studies using other approaches Generally, the gap models tend to estimate
reductions in forest biomass, particularly for softwoods in the Southeast. To be sure, the
models estimate that hardwood growth in the north could increase The biogeography and
biogeochemical models described in Box 2-2 tend to have more mixed results. Some show
significant declines in total national biomass of forests that are of even larger magnitude than
those estimated by the gap models (Some also show potentially severe declines in
southeastern forests). Others estimate that forest productivity could increase under climate
change In fact, the VEMAP study (VEMAP Participants, submitted) found there is no basis
to determine which of these biogeography or biogeochemical models is more accurate Thus,
the gap models are consistent with results of other studies, but they tend to be on the
pessimistic side of the range of all forest studies
2.3 Climate Change Scenarios
Since FASOM is a dynamic model, that is it simulates change in forest production over time,
it would be desirable to use transient scenarios of climate change. Transient scenarios
estimate how climate might change over time as greenhouse gas concentrations in the
atmosphere gradually increase. In contrast, equilibrium scenarios estimate static climate
conditions based on a specific level of atmospheric greenhouse gases, such as a doubling of
carbon dioxide levels (often referred to as 2XC02) In determining how climate change affects
forests, we were limited by the choice of climate change scenarios in the gap modeling
studies. Table 2-2 lists the changes in global temperatures from the general circulation model
GCM runs used in the gap model studies Most studies of climate change impacts on forests
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-6
Box 2-1
Results of Gap Model Studies
This table briefly summarizes the results of the studies used as the basis for estimating changes
in yields. All studies, except for Running and Nemani (1991), used gap models and assumed no
C02 fertilization effect.
Solomon (1986) examined effects on vegetation across all of the eastern half of North America.
He found that along northern border states, biomass did not change significantly, but composition
did. South of those sites, climate change significantly reduced biomass.
Pastor and Post (1988) modeled two areas in the United States, northeastern Minnesota and
Maine. They found that biomass would increase on wet soils in Minnesota and decrease on dry
soils Biomass in Maine would increase in both types of soils. Softwoods generally disappeared,
while hardwoods increased in abundance.
Botkin et al. (1989) modeled a site in northeastern Minnesota and another in south-central
Michigan. They found that the climate change scenarios resulted in the replacement of southern
boreal species in Minnesota (softwoods and hardwoods) with northern deciduous species (mostly
hardwoods). In the Michigan site, the abundance of oaks and maples is significantly reduced.
Urban and Shugart (1989) examined the effects of climate change on softwoods in sites across
the Southeast. In Tennessee, the oak-pine forest would shift to a loblolly pine forest, with little
change in total biomass. In South Carolina, Georgia, and Mississippi, forest biomass would
decline, with the most severe decline in Mississippi.
Urban et al. (1993) studied a site on the western slope of the Cascade Range in central Oregon.
They found that forest zones would shift 500 to 1000 meters higher in elevation. Both the OSU
and GISS scenarios caused reductions in biomass for forests at 1000 meter elevation.
King and Tingey (1992) examined several sites in Oregon and Washington. Like Urban et al
(1993), they also found that a 2°C to 5°C warming would cause vegetation zones to shift 500 to
1000 meters upslope. King and Tingey concluded that forest cover in the Pacific Northwest
could be reduced by 5 to 25% Biomass changes were not reported.
Running and Nemani (1991) used the FOREST-BGC model, a biogeochemical model, to
examine changes in net primary productivity (NPP) and leaf area index (LAI) for Missoula
Montana. We used this study because no gap model results were available for the Rocky
Mountain Regions. Assuming a 4°C wanning, a 10% increase in precipitation, and a carbon
fertilization effect, they estimated that NPP would increase 88% and LAI 10 to 20%
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-7
Box 2-2
Simulated Effects of Climate Change Using
Biogeography and Biogeochemical Models
Smith et al. (1992a), using a biogeography model, estimated that forest cover over the lower 48
states could be reduced by 1 to 7% in the long run. The extent of wet forest cover would be
reduced, and dry forest cover would increase.
Melillo et al. (1993) used the terrestrial ecosystem model (TEM), a biogeochemical model, to
estimate climate change effects on global NPP. They found that assuming climate change from
four GCMs and C02 doubling, global NPP would increase by 19 to 27%. Regional results for
the United States were not included in this article.
VEMAP Study (VEMAP Participants, submitted) used common scenarios of climate change
and elevated C02 in the United States to compare three biogeography (BIOME2, DOLY, and
MAPPS) and three biogeochemical models (TTiM, CENTURY, and BIOME-BGC). All six
models accurately estimate the current distribution of vegetation across the lower 48 states under
current climate Yet, their estimates of climate change impacts on vegetation diveige
substantially.
~ BIOME2 (biogeography model) projected an expansion of forests in
North America with C02-induced warming.
~ MAPPS (biogeography model) showed a significant reduction in forest cover, particularly
m the southeast
~ TEM (biogeochemical model) projected 12 to 16% increases in biomass due to a
doubling of C02.
~ BIOME-BGC (biogeochemical model) indicated a 9 to 32% reduction in biomass for the
same C02 doubling.
All three biogeochemical models, however, estimate that doubling C02 concentrations without
changing climate only increases carbon storage by 2 to 9%. The study also "coupled"
biogeographical models to estimate changes in location of vegetation types and biogeochemical
models to estimate changes in plant functioning in response to climate change. These couplings
produced very widely divergent results, with the combination of BIOME2 and TEM yielding a
32 to 56% increase in carbon storage, while the combination of MAPPS and BIOME-BGC
estimated a 8 to 39% reduction in carbon. The main source of the differences among the model
results concerns how they respond to increase evaporative demands from higher temperature. The
study concluded that there is no basis for determining which model results are most reliable
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-8
Table 2-2
GCM Runs Used in Gap Model Studies
Model
Citation
Global AT (°C)
UKMO 2XC02
Mitchell, 1983
2.3
UKMO 4XC02
Mitchell and Lupton, 19841
4.6
GISS 2XC02
Hansen et al., 1983
4.2
GFDL 2XC02
Manabe and Wetherald, 1987
4.0
OSU 2XC02
Schlesinger and Zhao, 1989
2.8
GISS Transient
Hansen et al., 1988
0-4.2
' We assumed that 4XC02 concentrations in the UKMO model result in double the temperature
change from the 2XC02 run using the UKMO model (Mitchell, 1983).
used 2XCOz equilibrium scenarios. Note that the GISS transient was the only transient
scenario used in gap model studies and it was used in only three studies.Thus, this study only
uses equilibrium climate change scenarios.
The GCM scenarios in Table 2-2 tend to be clustered either around a 2.5°C or a 4°C global
warming. Based on this selection of models, we decided to estimate the forest impacts under
scenarios a global 2 5°C and 4°C warming These changes in temperature are within the
range of increase in global temperatures estimated by the Intergovernmental Panel on Climate
Change (Houghton et al., 1992) We used results based on the UKMO 2XC02 and OSU
models as representative of a 2.5 °C warming and results based on the UKMO 4XC02, GISS
and GFDL models as representative of a 4°C warming. The GISS transient provides
information about both amounts of climate change. Relative changes in regional temperatures
may vary between models.
We also developed scenarios with and without the effects of C02 fertilization. The without
C02 scenarios uses the results directly from all of the gap modeling studies, except for
Running and Nemani (1991) (see Chapter 3). The with C02 scenarios assumes there is a C02
fertilization effect. We assumed a doubling of atmospheric C02 levels from approximately
350 ppm to 700 ppm.
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-9
The assumption about the C02 fertilization effect is our educated guess based on
conversations with gap modelers (e.g., Ron Neilson, Corvallis Lab, Corvallis, Oregon; Dean
Urban, Colorado State University, Fort Collins; George King, Mantech Inc., Corvallis,
Oregon) and a review of the literature on the C02. fertilization effect. The literature indicates
that a doubling of C02 concentrations in the atmosphere could increase plant growth by up to
40% For example, Drake found that elevated C02 levels increase plant growth by 30 to 40%
(Drake, 1992). Since there is significant uncertainty about the extent to which elevated C02
concentrations will enhance forest growth in relatively unman aged conditions (e.g., Bazzaz
and Fajer, 1992; Luxmoore et al., 1993), we assumed that C02 increases yields by 25% to
33%
To calculate the C02 effect we multiplied the change in biomass under the no C02 scenario
by 1.25 to 1.33. If, in the no C02 scenario, biomass was estimated to be reduced, we assumed
that the C02 fertilization effect increased yields by one-third. For example, if the no C02
scenario had biomass decrease to 30% of no climate change level (i e., a 70% decrease in
yield), the with C02 case would have biomass decrease to 40% of current level (1.33 X 0.3 =
0 4; a 60% decrease in yield). To have a minimum positive C02 fertilization effect, in all
cases we assumed at least an absolute change of 10%. So if biomass were estimated to
decrease by 90% in a no C02 case, we assumed that it would decrease by 80% in a with C02
case. If in a no C02 scenario, biomass is unchanged or increases, we assumed the C02 effect
increased yields by one-quarter. We assumed a lower positive effect of C02 when yields
increase to avoid unduly large increases in biomass. So if biomass were estimated to be 200%
of current biomass under a no C02 case, the C02 effect would result in yield of 250%
(1.25 x 2 = 2 50).
2.4 Using Site-Specific Information to Estimate Regional
Changes in Growing Stock
As noted above, gap models were used to estimate changes in growth of trees on specific
sites. Figure 2-2 displays the sites modeled by the gap model studies The map also displays
the U S Forest Service regions, which are the same regions used in FASOM. There are more
modeled sites and more extensive coverage of forests in the eastern half of the United States
than in the western half. This section describes how we used site-specific results to develop
estimates of changes in hardwood and softwood growing stock for the U.S. Forest Service
regions.
We used the results from the gap model studies to develop estimates of changes in hardwood
and softwood growing stock Where a study gave estimates for many hardwood or softwood
species, we used the change in total biomass for all hardwood or softwood species as
indicative of changes in hardwood or softwood biomass. For example, Solomon (1986)
reports results for several dozen hardwood and softwood species To determine changes in
hardwoods, we summed his estimated biomass changes for all hardwoods at a site. If only a
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-10
Figure 2-2
Location of Sites from Published Gap Model Studies
Great
Plains
North C<
Rocky Mountain
Pacific\
Southwest
Southeast
South Central
a - Solomon
b - Pastor and Post
c - Botlrin et al
d • Urban and Shugait
e - Urban et al
f - King and Tingey
g - Running and Nemani
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-11
limited number of species were reported, we used results for commercially important species
For example, in the South, Urban and Shugart (1989) report results for loblolly and shortleaf
pines. We used those results to develop estimates of changes in softwood production in the
Southeast and South Central regions.
If two or more studies had sites in the same area, we averaged their results together to create
a composite result. For example, Solomon (1986), Pastor and Post (1988), and Botkin et al.
(1989) estimated changes in northern Minnesota forests. The results of all three were
averaged to estimate yield changes in Minnesota. In some cases, we gave less weight to
studies that appeared to have extreme results. We assumed that each site was representative of
the forests in the entire state For example, the results of Urban and Shugart's (1989) study of
softwood forests in Macon, Georgia, were assumed to be representative of all Georgia
softwoods. We had more than one site in only one state, Michigan, where sites in upper and
lower Michigan were modeled separately. In that case, we developed state-wide estimates
based on the distributions of hardwoods and softwoods in the state (see Chapter 3).
Shifts in elevation are not relatively important in affecting yield of eastern forests and we did
not take them into account in our analysis. In the West such shifts could have significant
influence on the habitable area and yield of commercially important species, such as Douglas
fir For the West, we assumed that forests would shift upslope with higher temperatures We
extrapolated state results to Forest Service regions based on the state's percentage of regional
volume of hardwood or softwood growing stock (Waddell et al., 1989). For example, Urban
and Shugart estimated change in yields for softwoods in Georgia and South Carolina These
states respectively have 33 and 17% of softwood growing stock in the Southeast region We
weighted the state changes in softwood by their relative shares, (66%) for Georgia and (34%)
for South Carolina, to calculate change in softwood production for the region We describe
specific regional extrapolation issues in Chapter 3.
We also calculated the changes in the national growing stock of hardwoods and softwoods
These were based on the relative shares of hardwoods and softwoods in each region
2.5 References
Bazzaz, F.A. and ED. Fajer. 1992. "Plant Life in a C02-Rich World." Scientific American
January :68-74.
Botkin, D.B., R.A Nisbet, and T.E Reynales. 1989. "Effects of Climate Change on Forests of
the Great Lake States." In The Potential Effects of Global Climate Change on the United
States, Appendix D: Forests, edited by J.B Smith and D Tirpak. Washington, DC.: U.S
Environmental Protection Agency, EPA-230-05-89-054
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~ 2-12
Bugmann, H. and P. Martin, 1995. "How Physics and Biology Matter in Forest Gap Models.
An Editorial Comment." Climatic Change. 29:251-257.
Dale, V.H. and H.M. Rauscher, 1994. "Assessing Impacts of Climate Change on Forests The
State of Biological Modeling." Climatic Change 28:65-90.
Drake, B G. 1992. "The Impact of Rising C02 on Ecosystem Production.11 Water, Air, and
Soil Pollution 64:25-44.
Hansen, J. et al. 1983. "Efficient Three-Dimensional Global Models for Climate Studies:
Models I and n." Monthly Weather Review 3:609-622.
Hansen, J. et al. 1988. "Global Climate Changes as Forecast by the GISS 3-D Model."
Journal of Geophysical Research 93:9341-9364.
Houghton, J.T., B.A. Callander, and S.K. Varney, 1992. Climate Change 1992-The
Supplementary Report to the IPCC Scientific Assessment. WMO/UNEP Intergovernmental
Panel on Climate Change. Cambridge, England' Cambridge University Press.
King, G A and D.T. Tingey. 1992. Potential Impacts of Climate Change on Pacific
Northwest Forest Vegetation Prepared for the U.S. Environmental Protection Agency March.
EP A600/R-92/095.
Luxmoore, R.J., S.D. Wullschleger, and P.J. Hanson. 1993. "Forest Responses to C02
Enrichment and Climate Warming." Water, Air, and Soil Pollution 70:309-323.
Manabe, S and RT. Wetherald. 1987. "Large Scale Changes in Soil Wetness Induced by an
Increase in Carbon Dioxide " Journal of Atmospheric Sciences 44* 1211-1235
Melillo, J M., A.D McGuire, D.W Kicldighter, B. Moore, C.J. Vorosmarty, and A.L.
Schloss. 1993. "Global Climate Change and Terrestrial Net Primary Productivity." Nature
363:234-240.
Mitchell, J.F.B. 1983. "The Seasonal Response of a General Circulation Model to Changes in
C02 and Sea Temperatures." Quarterly Journal of the Royal Meteorological Society
109:113-152.
Mitchell, J.F B. and G. Lupton 1984 "A 4XC02 Integration with Prescribed Changes in Sea
Surface Temperatures." Progress in Biometeorology 3:353-374.
Pastor, J. and W.M. Post. 1988 "Response of Northern Forests to C02-Induced Climate
Change." Nature 334:55-58.
RCG/Hagler Bailly
Final Report
-------
Sources and Methods for Estimating Changes ~2-13
Running, S.W. and R.R. Nemani. 1991. "Regional Hydrologic and Carbon Balance Responses
of Forests Resulting from Potential Climate Change." Climatic Change 19:349-368
Schlesinger, M.E and Z.C. Zhao. 1989. "Seasonal Climatic Changes Induced by Double C02
as Simulated by the OSU Atmospheric GCM/Mixed-Layer Ocean Model." Journal of Climate
2:459-495.
Shugart, H.H., T M. Smith, and W.M. Post. 1992. "The Potential for Application of
Individual-Based Simulation Models for Assessing the Effects of Global Change." Annual
Review of Ecological Systems 23.15-38.
Shugart, H.H 1987. "Dynamic Ecosystem Consequences of Tree Birth and Death Patterns "
Bioscience 37:596-602.
Smith TM, R. Leemans, and H.H Shugart 1992a. "Sensitivity of Terrestrial Carbon Storage
to C02-Induced Climate Change: Comparisons of Four Scenarios Based on General
Circulation Models." Climatic Change 21:367-84.
Smith, TM, H.H. Shugart, G.B. Bonan, and J.B. Smith, 1992b "Modeling the Potential
Response of Vegetation to Global Climate Change " Advances in Ecological Research
22.93-116
Solomon, AM 1986. "Transient Response of Forests to C02-Induced Climate Change
Simulation Modeling Experiments in Eastern North America." Oecologia 68 567-579.
Urban, D L and H.H Shugart. 1989 "Forest Response to Climatic Change: A Simulation
Study for Southeastern Forests " In The Potential Effects of Global Climate Change on the
United States, Appendix D: Forests, edited by J.B Smith, and D Tirpak. Washington, D C
U S Environmental Protection Agency, EPA-230-05-89-054.
Urban, D L., M E. Harmon, and C B Halpern 1993. "Potential Responses of Pacific
Northwestern Forests to Climatic Change, Effects of Stand Age and Initial Composition."
Climatic Change 23 247-266.
VEMAP Participants. Submitted "Vegetation/Ecosystem Modeling and Analysis Project
(VEMAP) Assessing Biogeography and Biogeochemistry Models in a Regional Study of
Terrestrial Ecosystem Responses to Climate Change and C02 Doubling " Submitted to
Biogeochemical Cycles. Woods Hole, Massachusetts. The Ecosystem Center, Marine
Biological Laboratory.
Waddell, K L., D.D Oswald, and D.S. Powell. 1989. Forest Statistics of the United States,
1987. U S Forest Service. Portland, Oregon: Pacific Northwest Research Station, Resource
Bulletin RNW-RB-168.
RCG/Hagler Bailly
Final Report
-------
Chapter 3
Results: Yield Changes
In this chapter, we discuss regional and national changes in hardwood and softwood yields
Only the regional changes were used as inputs into FASOM. The national results were not
used in FASOM, but are useful for comparing percentage changes in yields with percentage
changes in welfare. State by state results are reported in the discussions on each region.
Weighted average results for hardwood and softwood yields by region are displayed in
Tables 3-7a and 3-7b at the end of this chapter.
3.1 Southeast and South Central Regions
The Forest Service and FASOM consider the Southeast and South Central regions to be
distinct regions Results are reported for both regions because the same gap model studies
included sites in both regions The Southeast region consists of Virginia, North Carolina,
South Carolina, Georgia, and Florida The South Central region consists of Kentucky,
Tennessee, Alabama, Mississippi, Louisiana, Arkansas, Oklahoma, and Texas
We developed estimates of changes in softwood yields based on Urban and Shugart (1989)
and for hardwood yields based on Solomon (1986) State results for softwoods and hardwoods
are displayed in Tables 3-1 (a and b) and 3-2 (a and b).
Table 3-la
Change in Southeast Softwood Biomass
Climate Scenario
% Change in Biomass
South Carolina
Georgia
2 5°C without C02
-90
-80
2 5°C with C02
-30
-70
4°C without C02
-100
-100
4°C with C02
-50
-90
RCG/Hagler Bailly
Final Report
-------
Results: Yield Changes ~> 3-2
Table 3-lb
Change in Southeast Hardwood Biomass
Climate Scenario
% Change in Biomass
Virginia
N. Carolina
2 5°C without C02
-10
-10
2.5°C with C02
+20
+20
4°C without C02
-20
-40
4°C with C02
+5
-20
Table 3-2a
Change in South Central Softwood Biomass
Climate Scenario
% Change in Biomass
Tennessee
Arkansas
Mississippi
2.5°C without C02
0
-100
-100
2 5°C with C02
+25
-90
-90
4°C without C02
-100
-100
-100
4°C with C02
+10
-90
-90
Table 3-2b
Change in South Central Hardwood Biomass
Climate Scenario
% Change in Biomass
Tennessee
Arkansas
2.5°C without C02
+5
-50
2.5°C with COa
+30
-33
4°C without C02
-35
-90
4°C with C02
-15
-80
RCG/Hagler Bailly
Final Report
-------
results: Yield Changes ~ 3-3
Urban and Shugart estimated changes in loblolly and shortleaf pine yields in South Carolina,
Georgia, Tennessee, and Mississippi under the GISS, GFDL, and OSU 2XC02 scenarios and
the GISS transient scenario. We used the GFDL scenario to estimate the 4°C without C02
case and the OSU scenario to estimate the 2.5°C without C02 case. To estimate the 4°C and
2 5°C with C02 cases, we used the GISS 2XC02 scenario and the GISS transient scenario
(combined with the C02 fertilization effect). We included the GISS results in the with C02
case because the GISS scenario is significantly wetter than the other two scenarios (Smith and
Tirpak, 1989). Thus, the without C02 scenarios in the South can be thought of as relatively
dry scenarios without a C02 fertilization effect and the with C02 scenarios can be thought of
as relatively wet scenarios with a C02 fertilization effect.
The results for the with and without C02 scenarios for softwoods in Tennessee differ widely
This is because there were large differences between GISS, OSU, and GFDL from Urban and
Shugart Urban and Shugart estimated that, under the GISS and OSU scenarios, there would
be no change in biomass, while under the GFDL scenario, all forests would disappear
Solomon (1986) examined sites in Virginia, North Carolina, Tennessee, and Arkansas He
only used two scenarios. UKMO 2XC02 and UKMO 4XC02 The first scenario was used to
represent a 2.5°C warming. Solomon reported that all hardwoods in Arkansas would be
eliminated under the 4XC02 scenario, which has a 4.6°C warming. To arrive at an estimate
for 4 0°C, we interpolated between the 4.6°C and the 2XC02 warming (2.3°C) to arrive at the
90% reduction in biomass. It is quite possible that the threshold for eliminating hardwoods in
Arkansas is below 4 6°C and we were too generous in assuming that some hardwoods would
survive. If so, the without C02 case would have a 100% reduction in biomass.
The results generally show decreases, particularly for softwoods The 4°C without C02
scenario eliminates all softwoods in the regions Hardwoods are not as adversely affected The
more southerly and westerly sites in the two regions tend to be the most negatively affected
by climate change
3.2 North Central
The North Central region consists of Minnesota, Wisconsin, Michigan, Ohio, Indiana, Illinois,
Iowa, and Missouri. We used three studies to develop biomass change estimates for this
region: Solomon (1986), Pastor and Post (1988), and Botkin et al. (1989). Solomon had sites
in Minnesota, Wisconsin, upper and lower Michigan, Ohio, and Missouri. Pastor only had
species-specific results in Minnesota (he estimated changes in total biomass in other sites).
Botkin provided estimates for Minnesota and lower Michigan.
As noted above, Solomon used two scenarios, the UKMO 2XCOz and the UKMO 4XC02.
Pastor and Post also used the same UKMO 2XC02 scenario That scenario was used to
represent the 2 5°C warming, and the UKMO 4XC02 scenario was used (with some
RCG/Hagler Bailly
Final Report
-------
results: Yield Changes ~ 3-4
interpolation) to represent the 4°C scenario. Botkin et al. used the OSU (2.8°C), GFDL
(4.0°C), and GISS (4.2°C) 2XC02 scenarios and the GISS transient (0 to 3.7°C). The OSU
and the GISS transient were used to help estimate biomass changes for 2.5°C without C02
scenario , and the GFDL GISS 2XC02 and GISS transient results were used to estimate
changes for a 4°C warming without C02 scenario.
All three studies provided estimates for Minnesota. Botkin et al., however, gave estimates for
only one hardwood species, sugar maples. Those estimates showed increases in biomass of S
to 10 times current levels. Solomon and Pastor and Post had hardwood increases of less than
100% Because they considered more species and were consistent with each other, we gave
more weight to Solomon and Pastor and Post than to Botkin et al. for the Minnesota
estimates.
Estimates were given for sites in the upper and lower parts of Michigan. Botkin estimated
changes for southern Michigan, as did Solomon. We averaged results from both studies to
develop a composite estimate for lower Michigan. Solomon also estimated changes in upper
Michigan. By examining the distribution of hardwoods and softwoods in the Great Lakes
states, we assumed that hardwoods were split evenly between upper and lower Michigan, but
that 90% of softwoods are in upper Michigan. We used these assumptions to weight the
results from upper and lower Michigan. State results are displayed in Table 3-3 (a and b).
Table 3-3a
Change in North Central Softwood Biomass
Climate Scenario
% Change in Biomass
Minnesota
Wisconsin
Michigan
Ohio
2.5°C without C02
-75
-70
-60
-10
2.5°C with C02
-65
-60
-45
+20
4°C without C02
-70
-80
-75
-70
4°C with C02
-60
-70
-65
-60
The North Central region presents some interesting differences between northern and southern
states in the region. Softwood yields tend to decrease across all the region, although the 2.5°C
sensitivities for Ohio have very little change in yield. It is not evident to us why the
softwoods in Ohio respond differently to the 2.5°C warming than did softwoods elsewhere in
the region. In contrast, hardwood yields increase markedly in the northern states of
Minnesota, Wisconsin, and Michigan, but decline in the more southern states of Ohio and
Missouri In the more northern states, the larger the temperature increase, the more hardwood
yields rise.
RCG/Hagler Bailly
Final Report
-------
Results: Yield Changes > 3-5
Table 3-3b
Change in North Central Hardwood Biomass
Climate Scenario
% Change in Biomass
Minnesota
Wisconsin
Michigan
Ohio
Missouri
2.5°C without C02
+40
+40
+15
-10
-40
2.5°C with C02
+75
+75
+40
+20
-20
4°C without C02
+100
+60
+35
-25
-70
4°C with C02
+150
+100
+70
0
-60
3.3 Northeast
Estimates of changes in yields from the Northeast regions were based on Solomon (1986) and
Pastor and Post (1988). The states in the region are the New England states, New York,
Pennsylvania, New Jersey, Delaware, Maryland, and West Virginia. Solomon modeled sites in
New York and Maine, and Pastor and Post reported species results only for Maine Both
studies used the same UKMO scenarios discussed above, and we used the same procedures to
examine the 2 5°C and 4°C scenarios. Both studies showed very similar reductions in
softwood biomass. Although both studies estimated that hardwoods would increase in
biomass, Solomon estimated a much larger increase than Pastor and Post did In developing
the combined estimate of hardwood changes in Maine, we gave slightly greater weight to
Pastor and Post's results The results for Maine and New York are displayed in Table 3-4.
Since both states are in the northern part of the region, we assumed they are representative of
forests in New York and New England
Even though Ohio is not in the region, we used the results for Ohio forests as representative
of the other states in the region (Pennsylvania, New Jersey, West Virginia, Maryland, and
Delaware) because it was the closest and most representative site
The results differ in sign for hardwoods and softwoods. While hardwood show increasing
yields with increasing temperatures, softwoods have decreasing yields with higher
temperatures.
RCG/Hagler Bailly
Final Report
-------
Results: Yield Changes » 3-6
Table 3-4
Change in Northeast Biomass
Climate Scenario
% Change in Biomass
Softwoods
Hardwoods
New York
Maine
New York
Maine
2.5°C without C02
-33
-70
0
+100
2.5°C with C02
-10
-60
+25
+150
4°C without C02
-75
-80
+60
+150
4°C with C02
-65
-70
+100
+200
3.4 Pacific Northwest and Pacific Southwest
As is shown in Figure 2-1, only one site in the West Coast states of the United States has
been the subject of a gap modeling study. The site, in central Oregon, is approximately in the
middle latitudes of the hardwood and softwood range in the region, and thus we assumed that
it is representative of the Pacific Northwest and Pacific Southwest regions. These regions
consist of Washington, Oregon, and California.
Results are reported in Urban et al (1993) and King and Tingey (1992). Since Urban
provided forest yield changes to King and Tingey, for the sake of simplicity, we only cite
Urban et al. in the rest of this section Only the OSU and GISS scenarios were used. We used
the GISS scenario, which produced yield declines of 90%, to estimate the 4°C without C02
sensitivity The OSU scenario resulted in yield declines of 50% and was used to determine
the 2 5°C without C02 case Urban et al estimated biomass only for softwoods, so no
changes in hardwood yields for these regions were developed. Hardwood yields in the West
are relatively small so not including changes in yields of those species did not significantly
affect results from FASOM
Urban et al estimated changes in yields for softwood forests at current elevations. In the
West, the elevation ranges where trees are found will most likely change as a result of
climate Although yield could decline at current elevations, warmer temperatures will most
likely enable forests to survive at higher elevations.
To estimate the effects of climate change on western forests, we had to develop two set of
assumptions The first concerned the changes in habitable area as temperatures and elevations
rise The second concerned how yields in the different elevations would change. We used a
RCG/Hagler Bailly
Final Report
-------
Results: Yield Changes ~ 3-7
typical mountain profile of the Cascades to calculate changes in habitable area due to shifts in
elevation. We further assumed that for each incremental increase in temperature (current to
2.5°C; 2.5°C to 4.0°C), estimated changes in yields would shift upslope by 500 meters. Both
assumptions are discussed in detail in Appendix A.
Results are displayed in Table 3-5 The higher reductions in yields from the 4°C warming
scenario appears to substantially offset the increase in elevation for softwoods. For the 2.5°C
scenarios, the increase in elevation approximately offsets decrease in yields at current
elevations.
Table 3-5
Change in Softwood Biomass in Pacific Northwest and Pacific Southwest
Climate Scenario
% Change in Biomass
2 5°C without C02
-10%
2.5°C with C02
+20%
4°C without C02
-53%
4°C with C02
-32%
3.5 Rocky Mountains
As is shown in Figure 2-1, the Rocky Mountain region had only one site, located in
northwestern Montana This region, however, includes Montana, Idaho, Wyoming, Colorado,
Utah, Nevada, Arizona, and New Mexico. We assumed this site is representative of softwoods
in the Rocky Mountains region Although almost all of the softwoods are in the northern part
of the region, the site still may be too far north to be representative of most softwoods in the
region. As can be seen in other regions, the farther north a site, the more likely it is to have
less negative or more positive impacts Thus, the Running and Nemani study (1991) may
have results that are too positive for the entire Rocky Mountains region.
Running and Nemani used only a single scenario of climate change, a 4°C increase in
temperature and a 10% increase in precipitation. Running and Nemani also assumed that
photosynthesis increases 30% and stomatal conductance decreases 30% because of C02
fertilization. They only estimated changes in softwood yields.
Unlike the other studies, Running and Nemani used a biogeochemical model and estimated
changes in leaf area index (LAI), not biomass Since LAI is not the same as biomass, we
RCG/Hagler Bailly
Final Report
-------
Results: Yield Changes * 3-8
needed to convert it to biomass to be consistent with the other studies. Based on Carlson
(1987), we assumed that increases in biomass are one-third less than increases in LAI.
Running and Nemani estimated that LAI would increase 10 to 20% under the 4°C scenario.
The midpoint of that estimate is 15%, and a one-third reduction gives an estimate of a 10%
increase in biomass.
Since Running and Nemani include C02 fertilization, their results were used for the with C02
case. Assuming that C02 increases growth by 33%, we calculated that without the C02 effect,
yields would have decreased by 20% in the 4°C low sensitivity case. We assumed that the
2.5°C warming would halve the reduction in yields to 10%. This latter result was used for the
high sensitivity case. We increased the biomass by one-third to give the 2.5°C high sensitivity
case
As in the West Coast, we accounted for shifts in elevation. We assumed that the relative
shifts in area elevation and habitable area for the 2.5 and 4°C warmings are the same as the
assumptions used in the Pacific Northwest and Pacific Southwest (George King, personal
communication). We also assumed that, for each incremental increase in temperature, yields
would shift upslope by 500 meters (see Appendix)
Results are displayed in Table 3-6. Unlike the other regions, yields for softwoods increase
under all four scenarios.
Table 3-6
Change in Softwood Biomass in Rocky Mountains
Climate Scenario
% Change in Biomass
2 5°C without C02
+30%
2 5°C with C02
+70%
4°C without C02
+33%
4°C with C02
+79%
3.6 National Results
Tables 3-7a and 3-7b display, respectively, regional changes in softwood and hardwood
growing stock. Table 3-7b does not display results for the western regions because hardwood
growing stock is quite small. As discussed in Chapter 2, regional changes for regions with
multiple sites from the gap model studies were derived based on the current distribution of
growing stock in the region.
RCG/Hagler Bailly
Final Report
-------
73
O
g ^
.§ S
I s?
Table 3-7a
Regional Changes in Softwood Growing Stock
(Percent Change)
Scenario
Southeast
South
Central
Northeast
North
Central
Rocky
Mountains
Pacific
Northwest
Pacific
Southwest
2.5°C without C02
-83
-87
-55
-64
+30
-10
-10
2.5°C with C02
-56
-75
-40
-51
+70
+20
+20
4°C without C02
-100
-100
-78
-76
+33
-53
-53
4°C with C02
-76
-77
-68
-66
+79
-32
-32
Table 3-7b
Regional Changes in Hardwood Growing Stock
(Percent Change)
Scenario
Southeast
South Central
Northeast
North Central
2.5°C without C02
-10
-20
+14
+11
2.5°C with C02
+20
-1
+46
+39
4°C without C02
-30
-60
+36
+21
4°C with C02
-8
-44
+71
+53
£
oo
H
u
~<
H
r
a
o
z
o
m
c/a
u>
I
\o
-------
Results- Yield Changes ~ 3-10
Softwoods were estimated to decline in the eastern half of the United States under all four
scenarios In some regions such as the South, softwoods were estimated to be completely
eliminated under some scenarios. In contrast, softwoods were estimated to increase in western
areas in some scenarios. The potential for increase is due in part to higher elevations,
allowing the softwood growing area to expand.
Hardwood growing stock also shows mixed results with increases in some regions and
decreases in others. In contrast to the softwoods, latitude appears to be a more important
factor than longitude in determining whether hardwood growing stock will increase or
decrease The northern regions consistently show increased hardwood yields, while the
southern regions tend to have decreased hardwood yields.
Summing across regions, the South is estimated to face decreased hardwood yields and severe
decreases in softwood yields. The northeastern quadrant of the country could have significant
decreases in softwood yields, but significant increases in hardwood yields. The West is
estimated to have mixed results for softwood yields.
We calculated changes in national yields of hardwoods, softwoods, and all forests by
weighting current growing stock in each region. The results are displayed in Table 3-8 and
Figure 3-1. In spite of the effect of increased softwood yields in the western regions in some
scenarios, nationwide softwood yields are estimated to decrease in all scenarios. The results
for hardwood yields are very sensitive to assumptions about the C02 fertilization effect. When
no C02 effect is assumed, national hardwood yields decline. When a C02 fertilization effect
is assumed, national hardwood yields are estimated to increase. The combined national effects
on forests tend to show either little change or reductions in growing stock. The most negative
scenario, a 4°C warming with no C02 fertilization effect, has an almost 40% decline in
yields Incorporating C02 reduces the change by about two-thirds
Table 3-8
Percent Change in National Growing Stock of Private Forests
Scenario
Hardwoods
Softwoods
Total
2.5°C without C02
-2
-47
-23
2 5°C with C02
+25
-24
+3
4°C without C02
-10
-70
-37
4°C with C02
+16
-48
-13
RCG/Hagler Bailly
Final Report
-------
Results: Yield Changes ~ 3-11
Figure 3-1
Percent of Change in National Growing Stock of Private Forests
40
20
0
-20
-40
•60
-80
The lower warming scenario, 2.5°C without C02, results in an almost one-quarter decline in
total yields For this amount of warming, incorporating C02 fertilization results in a slight
increase in national yields. The results reflect the importance of the degree of wanning as
well as the uncertainty regarding the C02 fertilization effect.
3.7 Conclusion
The estimated changes in yields vary quite considerably across the country. The variance
appears to depend on such factors as magnitude of temperature change, whether the C02
fertilization effect is considered, latitude, elevation, and whether hardwoods or softwoods are
being examined These results should be interpreted with caution because there is significant
uncertainty about the sensitivity of forests to higher temperatures and elevated C02 levels.
Other vegetation models have shown a wider range of potential climate change impacts.
Biogeography and biogeochemical models have found that biomass in the United States could
significantly decrease or increase. The results for this study range from virtually no change in
forest yields to a significant decrease. To be consistent with the range of results from the lit-
erature, future studies should also include scenarios with significant increases in forest yields.
Percent Change
HHardwoods OSoftwoods tibial
1
IX W V1
§
1
1
2 J°C Low 2S°C High 4°C Low 4°C High
Scenario
RCG/Hagler Bailly
Final Report
-------
results: Yield Changes *> 3-12
3.8 References
Botkin, D.B., R.A. Nisbet, and T.E. Reynales. 1989. "Effects of Climate Change on Forests of
the Great Lake States." In The Potential Effects of Global Climate Change on the United
States, Appendix D: Forests, edited by J.B. Smith and D. Tirpak. Washington, DC: U.S.
Environmental Protection Agency, EPA-230-05-89-054.
Carlson, L.J. 1987. Leaf Area Index and Forest Community Properties: Linking Ground-Based
and Remotely Sensed Data for a Northern Minnesota Boreal Forest. Thesis for Master of
Science. Charlottesville, Virginia: Department of Environmental Sciences, University of
Virginia.
King, G.A. and D.T. Tingey. 1992. Potential Impacts of Climate Change on Pacific
Northwest Forest Vegetation Corvallis, Oregon: U.S. Environmental Protection Agency
Pastor, J. and W.M. Post. 1988. "Response of Northern Forests to C02-induced Climate
Change." Nature 334: 55-58.
Running, S.W. and R.R. Nemani. 1991. "Regional Hydrologic and Carbon Balance Responses
of Forests Resulting from Potential Climate Change." Climatic Change 19: 349-368.
Solomon, AM. 1986 "Transient Response of Forests to C02-Induced Climate Change
Simulation Modeling Experiments in Eastern North America." Oecologia 68. 567-579.
Smith, J.B. and Tirpak, D. 1989 (eds). The Potential Effects of Global Climate Change on the
United States Washington, DC. U S Environmental Protection Agency, EPA-230-05-89-050
Urban, D.L andHH Shugart 1989 "Forest Response to Climatic Change: A Simulation
Study for Southeastern Forests." In The Potential Effects of Global Climate Change on the
United States, Appendix D: Forests, edited by J.B. Smith and D. Tirpak. Washington, DC.
U.S Environmental Protection Agency, EPA-230-05-89-054.
Urban, D.L., M.E Harmon, and CB Halpern 1993. "Potential Responses of Pacific
Northwestern Forests to Climatic Change, Effects of Stand Age and Initial Composition."
Climatic Change 23' 247-266.
RCG/Hagler Bailly
Final Report
-------
Chapter 4
FASOM Model and Economic Analysis Methods
This chapter provides a brief description of the major features and important assumptions of
the FASOM model, with special attention focused on the forest sector portion of the model
This chapter also describes how the forest sector portion of the model was used to simulate
the economic effects of changes in forest yields and presents the yield changes used in each
of the forest effects scenarios.
4.1 Model Overview
This economic analysis performed in this report was conducted using the timber supply
portion of FASOM (Adams et al., 1994) The U.S. timber market in FASOM is modeled as a
multi-regional nonlinear programming problem. FASOM maximizes the net present value of
the total economic surplus associated with timber production and consumption subject to
constraints that characterize flows of products between domestic supply demand regions and
which account for changes in the forest inventory due to natural processes and management.
As such, FASOM simulates two inter-related sets of processes:
~ investment in, and management of, private timberland and the production,
consumption and price formation for products from that land
»• the evolution of regional forest inventories on those lands, over time, as a
result of regeneration, growth, and management.
These two sets of processes are linked by virtue of the fact that, because it is an economic
optimization model, FASOM solves simultaneously for economically-efficient production
levels and prices in the product markets in each period and for the optimal levels of
investment in regeneration and management of timberland.
FASOM contains nine timber supply regions1 and a single national demand region. These
regions coincide with those used by the USDA Forest Service, with minor exceptions (see
Figure 4-1). Production, consumption and price formation are simulated for two species
1 The supply regions, in Figure 4-1, are Pacific Northwest-West and -East, Pacific Southwest,
Rocky Mountains, Lake States, Corn Belt, S. Central, Northeast, and Southeast.
RCG/Hagler Bailiy
Final Report
-------
pacific Northwest
Figure 4-1
FASOM Regions
W°rthT%nnW
Agriculture Only
Rocky Mountains
Forestry Only
Pacific Southwest
Lake States
Northeast \ ^
Southeast
South Central
Southern
Agriculture Only
-------
FASOM Model and Economic Analysis Methods *¦ 4-3
(softwood and hardwood) and three product types (sawtimber, pulpwood, and fuelwood), for a
total of six products. The model is designed to simulate market behavior over a 100-year
period, with explicit accounting on a decade by decade basis. The model incorporates linear
national demand curves for forest products, by decade, for the projection period 1990-2080
Production on public timberland is treated as exogenous, which is consistent with public
policy Changes in public harvest volumes influence product prices, harvest levels and timber
inventories because through their effects on demand. If public harvests are increased, the total
supply available for consumption also increases. As a result, stumpage prices fall, harvests on
private timberland are reduced, and the inventory on private timberland increases.
The forest sector of FASOM simulates the use of existing private timberland as well as the
reforestation decision on harvested land. The outflow flow of land from timberland to other
land uses is exogenous in the forest sector version of the model (but endogenous in the joint
sector model). Forested land is differentiated by region, the age cohort of trees,2 ownership
class, cover type, site condition, management regime, and suitability of land for agricultural
uses FASOM accounts for carbon accumulation in forest ecosystems on private timberland,
as well as the fate of this carbon, both at and after harvest.
A unique feature of FASOM is the way it addresses the problem of terminal conditions Trees
may have rotation lengths that would carry them beyond the explicit time frame of the model.
To get around this problem, terminal conditions must be specified At the time of planting,
producers should anticipate a discounted flow of net returns that justify stand establishment
costs This is done with "terminal conditions," which represent the projected net present value
of an asset for all time periods beyond the end of the model projection Terminal conditions
in FASOM are resolved using downward sloping demand curves for the terminal inventory
The model provides the following output information
~ The net present value of total surplus in U.S timber markets. This consists of
° Consumer surplus of product buyers. Conceptually, this is the
maximum amount of money these buyers would be willing to pay for a
product, rather than do without it, less the amount they actually have to
pay.
d Producer surplus (for both domestic and foreign producers).
Conceptually, this is the amount of money that a producer receives for a
product less the minimum cost of producing that output.
Forest lands are grouped in ten 10-year age cohorts' 0 to 9 years, 10 to 19, ... , 90 + years.
Harvesting is assumed to occur at the midyear of the cohort
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-4
a Value of the terminal inventory. This is equal to the terminal
inventory volume times its marginal value.
a Production (harvest) levels. Timber volumes (cu ft) by product,
species, and region.
° National product consumption and price levels. Volumes by product,
species, region, and dollars per cu ft by product and species.
a Inventory volumes and acres. By region, owner group, species and
various indicators of productivity, land class, and management regime.
For the purposes of this study, FASOM can simulate adaptation to climate change by timber
producers and consumers in a number of different ways. These include:
*¦ Changes in the location where trees are planted. FASOM can shift planting away
from lower growth to high growth regions.
~ Changes in management intensity. FASOM can increase yields of a species in a
given region by shifting reforested acres into a higher management intensity.
~ Changes in species. FASOM can shift the species distribution in a region through
planting.
~ Changes in rotation lengths. FASOM can harvest existing stands earlier to avoid
exposure to reduced growth at later ages and to allow replanting of stands into
different species and management intensities. FASOM can also vary the rotation
lengths of re-forested stands consistent with the economic optimum.
~ Changes in harvesting capacity. FASOM can increase or reduce investment in the
amount of harvesting capacity in a region.
~ Changes in timberland investment. FASOM can simulate disinvestment in
timberland, by idling the land and reducing management costs to zero.
~ Price-based resource conservation. Physical changes in yields directly and indirectly
affect current and future stumpage prices and land rents. As seen in this study,
decreases in yields make it more costly for producers to supply stumpage, leading
them to cut back harvesting. Consumers react to higher prices by reducing their
consumption.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-5
4.2 Forest Sector Detail
Although FASOM can be run as a linked, two-sector model, this analysis was performed
using a stand-alone version of the forest sector. The decision to use this version, as opposed
to the linked model, was based on the fact that consistent climate effects scenarios for the two
sectors are not readily available Generating these scenarios using climate and physical effects
models in the sectors was well beyond the scope of this research. Preliminary analyses
conducted using hypothetical estimates of forest and agricultural yield changes, using the
linked model showed highly-dependent scenario results.
A mathematical exposition of the forest sector portion of FASOM is contained in
Appendix C. The forest sector in FASOM consists of the following basic building blocks
~ demand functions for forest products
*¦ timberland area and inventory structure and dynamics
~ production technology and costs.
4.2.1 Product Demand Functions
FASOM employs a single national demand region for forest products that treats only the log
market portion of the sector. The parameters for these demand functions were derived from
the regional stumpage demand equations in TAMM.3 There is, in fact, very little
interregional shipment of logs in the U.S. forest sector. Competitive price relations between
regions at the log and stumpage market levels are maintained through extensive trade and
competition at the secondary product level (e.g., lumber, plywood, pulp) Use of a single
consuming region for logs emulates the effects of competition at higher market levels without
the use of an explicit representation of activity at these levels
The demand for logs derives from the manufacture of products at higher market levels In
FASOM, log demands are aggregated into six categories sawlogs, pulp wood, and fuelwood
for both softwoods and hardwoods Log volumes are adjusted to exclude all but the crowing
stock portion 4 Thus, demand is for growing stock log volumes delivered to processing
facilities Log demand curves are derived from solutions of the TAMM and NAPAP (Ince,
1994) models by summing regional derived demands for logs from manufacturing at higher
market levels (sawlogs from TAMM, pulpwood from NAPAP). Fuelwood demand, which is
not price sensitive in TAMM, is represented by a fixed minimum demand quantity and a
fixed price National fuelwood demand volumes by decade were derived from appropriate
3 The parameters of the demand equations m TAMM are estimated using econometric methods and
are re-estimated periodically by the USDA Forest Service.
4 Nongrowing stock volumes are included only for carbon accounting.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-6
scenarios in the 1993 Timber Assessment Update report (Haynes et al., 1994a). Demand
curves are linearized about the point of total decade quantity and average decade price. These
demand functions are shifted, exogenously, for the period 1990-2040 based on
macroeconomic forecasts of key driver variables that reflect the underlying secondary product
demand environment (e.g., population, GNP, housing starts, expenditures on repair and
maintenance costs ), secondary processing technology, and secondary product capacity
adjustment across regions. The forecast values for these variables are held constant after 2040.
These forecasts are presented more fully in Haynes et al. (1994a).
Off-shore trade in forest products occurs at the supply region level and includes both
softwood and hardwood sawlogs and pulpwood. Fuel wood is not traded. Price-sensitive, linear
demand (export) or excess supply (import) relations were developed for the various regions
and products as appropriate for their current trade position. For example, the Pacific
Northwest-West region faces a net export demand function for softwood sawlogs but no
offshore trade demand for hardwood products or other softwood log products.
4.2.2 Inventory Structure and Dynamics
Forest inventory in FASOM is divided into a number of strata, each representing a particular
set of region, forest type (species), private ownership, site class, age, agricultural use
suitability, and timber management intensity characteristics. Each stratum is characterized in
terms of the number of timberland acres and the growing stock volume per unit area (in cubic
feet per acre) it contains Inventory estimates for the existing forest inventory on private
timberland are drawn from data used in the 1993 Timber Assessment Update (Haynes et al.,
1994a)
The descriptors used in FASOM to characterize the structure of the inventory on private
timberland in each region are as follows'
~ Owner groups. Forest industry and other private.5
*¦ Forest type/species. Softwood and hardwood in the current rotation and immediately
preceding rotation.
*¦ Suitability for agricultural use. Land is classified by type of alternative use for
which it is suited (crop, pasture, or forest) and by its present use (crop, pasture, or
forest). Multiple suitability classes are only used for the other private ownership,
because all forest industry timberland is classified as forest only.
3 Unlike Powell et al. (1993), the other private inventory in FASOM does not include Native
American lands. Harvests on these lands are included with the other public exogenous harvest group.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods »• 4-7
*¦ Site quality classes. High, medium, and low based on the Forest Service productivity
classification scheme.
~ Management intensity classes (MIC). Three current classes for both private owner
groups (high, medium, and low) are based on a qualitative characterization drawn from
the management intensity classes used in the ATLAS system (Haynes et al., 1994b) A
fourth class, low-low, is used to characterize any future harvested timberland that is
passively managed thereafter and is not converted to agriculture.
~ Age cohorts. Ten-year intervals, based on FLA descriptions, from the 0-9 year class
(just regenerated) to the 90+ year class.
Any portion, from 0 to 100%, of a stratum can be harvested at a time. The harvested acres
then flow back into a pool from which they can be allocated for several different modes of
regeneration as new timber stands or shifted to agricultural use. FASOM allows for newly
regenerated stands, whether on timberland or agricultural land, to be subject to several
different levels of management intensity. Although management intensity shifts cannot occur
after a stand has been regenerated (as can occur within ATLAS), this is not thought to be a
problem, given that the model employs perfect foresight in allocating land to competing
activities
FASOM simulates the growth of existing and regenerated stands by means of timber yield
tables, which give the net wood volume per acre in unharvested stands for strata (e.g., owner,
species) by age cohort Relative density adjustment mechanisms (Mills and Kincaid, 1992)
were used to adjust tree volume stocking levels in deriving yields for existing timberland and
for any timberland regenerated into the low timber management class. Timber yields for
plantations on agricultural lands are based on the most recent, reconciled estimates by
Moulton and Richards (1990) and Birdsey (1992) 6
The above growth and inventory structure is used only for private timberland.7 Two
additional categories of forest land also need to be addressed, public timberlands and
nontimber, forested land. Private timberland constitutes about 70% of total timberland and
about 50% of the total forest land in the United States. Timberland in various public
6 Timber yields contained in Moulton and Richards (1990) were derived from estimates for
plantations from Risbrudt and Ellefson (1983). In some cases such as for the Rocky Mountains region,
these estimates have been the subject of some debate because they are fairly high relative to yields on
commercial timberland. Estimates of timber yields used by Birdsey (1992), based on yield tables in
ATLAS and used for the RPA, are much lower
7 Under Forest Service definitions (adopted here), forest land is any land with at least 10% tree
cover (as would be identifiable in an aerial photograph) Timberland is forest land that has the capability
to grow at least 20 cubic feet per acre per year of commercial timber products.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-8
ownerships—including federal, state, and local owners—represents about 30% of the
timberland and about 20% of the forest land in the United States. Data from which to
sufficiently characterize the site quality and age structure of the public inventory in the East
were not available when parallel private timber data sets were assembled for the FASOM
study. In the West, which contains the majority of public timberland, the USDA Forest
Service is in the process of collecting the necessary data for public timberlands to augment
the private inventory data. Currently, the data that do exist for public timberlands in the West
are relatively sparse and difficult to organize, and the earliest date that a full reliable set may
possibly be available is 1997 for use in the next RPA Assessment. Because of the current
unavailability of data for key regions and the complexity and cost associated with developing
an inventory of public timberlands, FASOM does not model their inventory. Harvest on these
lands is taken as exogenous. This is consistent with current public policy and the expected
direction of future policies that will most likely place increasing weight on the management
of public lands for nonmarket benefits.
Nontimber, forested land constitutes about 30% of the forest land in the United States. This
includes transition zones such as areas between forested and nonforested lands that are
stocked with at least 10% of forest trees It also includes forest areas adjacent to urban and
developed lands (e.g., Montgomery County, Maryland) and some pifSon-juniper and chaparral
areas of the West. Although this area is large, the data to characterize the site class and
inventory structure of this inventory are much poorer than for public lands. Also, the fact that
this land is not very productive and is widely dispersed among private owners with a variety
of management objectives makes it a very difficult "target" for either regulatory or incentive-
based forest management programs In light of these difficulties, harvest on this land is taken
as exogenous and changes in inventory volumes or structure are not accounted for in the
model.
4.2.3 Production Technology, Costs, and Capacity Adjustment
Harvest of an acre of timberland involves the simultaneous production of some mix of
softwood and hardwood timber volume In FASOM, this is translated into hardwood and
softwood products (sawlogs, pulp wood, and fuelwood) in proportions that are assumed to be
fixed. The product mix changes over time as the stand ages and between rotations if the
management regime (intensity) changes. Downward substitution (use of a log "normally"
destined for a higher valued product in a "normally" lower valued application) is allowed
when the price spread between pairs of products is eliminated. Sawlogs can be substituted for
pulpwood and pulpwood can be substituted for fuelwood, provided that the prices of sawlogs
and pulpwood, respectively, fall low enough (or the substitutes increase high enough) to
become competitive substitutes for pulpwood and fuelwood, respectively. This "down
grading" or interproduct substitution is technically realistic and prevents the price of the
pulpwood from rising above that of sawlogs and the price of fuelwood from rising above that
of pulpwood.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-9
Strata in the inventory have specific management (planting and tending) costs that vary by
inventory characteristics and type of management. These costs were derived from a variety of
sources, including Moulton and Richards (1990) and those used in the 1989 RPA Timber
Assessment (Alig et al., 1992).8 Each product, in turn, has specific harvesting and hauling
costs (hauling in this instance relates to the movement of logs from the woods to a regional
concentration or delivery point) These costs were derived from the TAMM data base and
cost projections used in the Forest Service's 1993 RPA Timber Assessment Update
(Haynes et al., 1994a).
Consumption of sawlogs and pulpwood in any given time period is restricted by available
processing capacity in the industries that use these inputs. FASOM solves for investment in
additional capacity by allowing purchases of capacity to occur incrementally at externally
specified prices. This raises the maximum capacity in both current and future periods. It also
reduces producers' surplus by the cost of the new capacity acquisition. Over time, capacity
declines by an externally specified depreciation rate. Capacity increments in any period are
also limited by preset bounds. Since capacity may be added but not fully depreciated before
the end of the projection, a standard accounting practice is employed. Specifically, the
objective function is augmented by a term giving the current market value of the
undepreciated stock.
4.2.4 The FASOM Tableau: Forestry Component
A tabular overview of the forestry component of FASOM for a single region is given in
Table 4-1 The formulation in the table is consistent with the equations in Appendix C
governing production of stumpage, capacity, terminal inventory, and the conservation of land
It also includes the land balances shared by the forestry and agricultural sectors. The columns
in this table represent variables, and the rows represent equations. The tableau has been
simplified so that it can be presented on a single page and still convey the basic structure and
features of the model The tableau does not portray external product trade (import/export)
activities or constraints nor does it show the data computations that are made within the
GAMS code The carbon sector is also omitted
4.3 Approach, Inputs, and Limitations of Analysis
This section outlines the approach used to simulate the economic effects of changes in forest
productivity. It identifies the specific yield changes used in the analysis by region and species
for each scenario. Finally, it addresses the impact of model limitations on the analysis
8 See Appendix C for a more detailed description of the timber growth and yield, management
costs, and assumptions about trends in nonagricultural uses of forested lands.
RCG/Hagler Bailly
Final Report
-------
o
$ o
§ £
¦8 "
2 CD
Table 4-1
Sample Tableau for Overall FASOM Model Emphasizing Forestry Component
(Note: The coefficients give the signs/values of the parameters in the model)
Equation!
1
Harvest EiUtlng
Land
Now +10 +20 Never
2
Reforest In
Reforest Now +10 for
for Harvest Harvest in
+10 +20 Never +20 Never
3
Transfer
Forest Land to
Ag Use
Now +10 +20
3
Transfer Ag
Land to Forest
Use
Now +10 +20
4
Ag Land Use
Now +10 +20
S
Transport
Forest
Products
Now +10 +20
6
Sell Forest
Products
Now +10 +20
7
Build Forest
Process
Capacity
Now +10 +20
9
Terminal
Value
Timb Cap
RHS
Objective
Fn
+/- +/- +/-
-
+ + +
...
+ +
Existing
Forest
+ 1 +1 +1 +1
S +
Forest
Products
Supply
Now
+ 10
+20
-
-
+ 1
+1
+1
£ +
S, +
i +
Forest
Products
Demand
Now
+ 10
+20
-1
-1
-1
+1
+1
+ 1
£ 0
£ 0
£ 0
Forest
Products
Capacity
Now
+ 10
+20
+1
+1
+ 1
- -
<. +
i +
S +
Reforested
Lands
Balance
Now
+10
+20
-1
-1
-1
+ 1 +1 +1
-1 +1 +1
-1 +1
+ 1
+ 1
+ 1
-1
-1
-1
£ -
£ -
Ag
Lands
Balance
Now
+ 10
+20
-1
-1 -1
-1 -1 -1
+ 1
+ 1 +1
+1 +1 +1
+ 1
+ 1
+ 1
£ +
£ +
i +
Max Land
Transfer
to Ag
to For
+ 1 +1 +1
+1 +1 +1
S +
i +
Terminal
Balance
Timber
Capacity
-1
-1 -1
. . .
+ 1
+ 1
£ 0
S 0
~Tj
>
C/J
O
s
2
o
u
P
¦u
I
-------
FASOM Model and Economic Analysis Methods ~4-11
4.3.1 Approach
In Chapters 2 and 3 we developed yield changes for commercially important hardwood and
softwood forest types in the FASOM regions. These scenarios were based on information
from a wide number of published studies and the authors' judgment about the effects of
changes in climate and C02 fertilization on forest ecosystems. Four scenarios were developed,
based both on the extent of climate change (2.5°C and 4°C) and a sensitivity range (low and
high) 2 5 low, 2.5 high, 4.0 low and 4.5 high The sensitivity range was intended to reflect
uncertainty associated with the magnitude of the warming and the fertilization effects of C02
from the published literature and variation in the literature.
There are at least two different approaches that one can take to simulate the impacts of
changes in climate and elevated C02 concentrations on forests. The first is to try to simulate
the effects of changes in climate and C02 on forests slowly, as they occur over time. For
convenience, and to be consistent with the science, we can call this a "transient" approach.
The second is to move forward in time and evaluate the economic impacts of changes in
forest productivity once the changes in forest productivity have reached a significant level
This is the so-called "equilibrium" approach, so named because it assumes that the forest
ecosystem has fully adjusted to a fixed change in climate and C02, usually associated with a
doubling
FASOM is capable of employing either approach, yields can either be adjusted slowly over
time in the model or all at once However, there are a number of other complications that
arise in selecting between these two approaches One that is very important is that changes in
forest productivity, as simulated by all the different types of models we looked at, take a very
long time. Once a C02 doubling is achieved, which may take 35 to 75 or more years to reach,
it takes on the order of centuries for plant communities to fully adjust to these changes. This
feature of C02 buildup, climate change, and community ecosystem response creates several
interesting problems for economic modelers
First, some economic models, and FASOM is one of these, simply cannot be run over these
long time spans because of computational limitations. In that case, a transient analysis that is
truncated in time may give a distorted picture of the economic implications of climate change
if the changes in yields that are outside of the model's time domain are large in magnitude
A second problem, which is an issue for both the transient and equilibrium approaches, is
related to the need to forecast future changes in economic variables that are exogenous to the
economic model In FASOM, these variables either determine the future demand for stumpage
products, for example, population, housing starts, and Gross National Product, or else
determine the level of technology used to harvest trees and then convert stumpage into wood
products Not only are these variables very difficult to forecast 20 to 100 or more years in the
future, in many cases they are not even exogenous, but can be expected to change as society
adapts to climate change over time.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods *4-12
Any errors or biases that are in the forecasts one makes about these variables will propagate
through an economic model, and unless the model is classically linear, will result in even
larger errors and possibly biases in the model forecasts of endogenous variables of interests.
Given the long time periods with which one must work in climate change analyses and the
widely acknowledged problems associated with forecasting economic variables that far into
the future, the errors and biases added by these forecasts could arguably swamp the errors and
biases inherent in climate and forest ecosystem models. As an example, one can probably
eliminate much of the impact of climate change on many natural resource sectors using
straight line projections of technological change over the last 100 years. However, it is far
from clear that historical rates of technological change in yields and processing technology
can be maintained far into the future.
As noted above, the problem of forecasung exogenous economic variables is shared by both
the transient and equilibrium approaches Furthermore, a true equilibrium approach would also
involve having to forecast the forest inventory from the present well into the future once a
doubling of C02 was achieved, and then simulating the adjustment of that inventory to the
equilibrium adjustments of the various plant communities, a formidable undertaking.
Given all of the above problems, we decided to take an approach first suggested in the
climate change context by Callaway et al. (1982) In his study of methods for estimating the
economic effects of climate change on the agricultural sector, he suggested that an alternative
approach would be to impose an instantaneous equilibrium change in productivity on a
renewable resource sector under current (i.e., base case) economic conditions, and then trace
the economic effects of these changes in productivity over time on the sector using the same
values for exogenous economic variables as in the base case.
This approach has both limitations and strengths The major limitation is that this approach
overstates the value of damages (or benefits) in a way that is indirectly related to the rate of
change in physical impacts Another drawback of this approach is that it does not take into
account advancements in technology that might offset the negative impacts, or enhance the
positive impacts, of climate change However, economists have not been very successful in
projecting technological change over long periods of time, and instead have been forced to
make educated, "exogenous" guesses about what future technological change might look like
The major benefit of this approach is that it holds all inputs to the analysis constant at base
case values and focuses entirely on the effects of changes in physical productivity on the
relevant renewable resource sector. These effects are not confounded by the errors in one's
assumptions about future economic conditions.
This approach provides a relevant policy context for evaluating the economic effects of
climate change on renewable natural resource sectors. Estimates of annualized changes in
consumer and producer surplus can be viewed as the long run welfare impacts of an
equilibrium change in climate The impacts on production levels, product prices and
inventories can, in turn, be can also be interpreted as long run adjustments to an equilibrium
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods * 4-13
climate change. However, because endogenous technological change is not (and perhaps can
never be) factored into the analysis, adverse economic impacts will be somewhat overstated.
This is because the application of specific technological advancements to mitigate climate
change would not be adopted in the future unless they made society better off.
4.3.2 Inputs: Yield Changes by Region and Species
The yield changes we developed in Chapters 2 and 3 for commercially important hardwood
and softwood forest types in the FASOM regions are shown in Table 4-2. They were
translated into scalar factors, expressed as 1 plus the fractional change in yields for all species
in all regions There are two sets of yield tables in FASOM: one for the existing inventory
and one for the inventory that replaces it. These yield tables relate the amount of growing
stock timber volume on an acre of timberland to the age of the cohort on those acres through
yield coefficients. For each scenario, all of the yield coefficients for a given region and
species, except those in the base year (1990), were multiplied by the scalar yield change
coefficient associated with that region and species.
In summary, this approach is consistent with the assumption that the productivity of
timberland in the United States moves from its state in 1990 to a state dominated by climate
change by 2000 and remains constant thereafter. This sudden change in yields, while it is not
realistic, not only allows one to define a worst case transition, but also provides a long period
of adjustment to the lower yields. An alternative is to gradually alter the yields in the model
over time, and look at the economic adjustment to these more gradual shifts In this case,
economic adjustments will also take place more gradually and welfare losses will be smaller.
The problems with this transient approach are that, in most cases, we do not have good
information about transient yield changes, and it is difficult to run FASOM for more than
100-120 years, making it difficult to look at the transition in a realistic time frame
There are at least three points that need to be made about these yield changes that have
important ramifications for the economic analysis that follows in the next chapter.
Specifically, these points are
» The decreases in softwood yields in the South (Southeast and South Central regions)
should be regarded as especially severe in all cases This is not only because the South
produces more softwood products than any other region, but also because rotation
lengths are relatively short compared to other regions, so that under normal conditions
"gaps" in the inventory can be filled more quickly by reforestation in the South than
elsewhere.
~ The 2.5 scenario without C02 has softwood yield decreases that are more severe in the
South than in the 4.0 sensitivity scenario, with C02.
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-14
Table 4-2
Percent Changes in Yields by Region and Species
for Four Climate Scenarios
Region
Scenarios
2.5 with
C02 Change
2.5 without
COz Change
4.0 with
COz Change
4.0 without
C02 Change
Softwood
Pacific Northwest-West
20
-10
-32
-53
Pacific Northwest-East
20
-10
-32
-53
Northeast
-40
-55
-68
-78
Lakes States
-51
-64
-66
-76
Corn Belt
-51
-64
-66
-76
Southeast
-56
-83
-76
-100
South Central
-75
-87
-76
-100
Rocky Mountains
70
30
79
33
Pacific Southwest
20
-10
-32
-53
Hardwood
Pacific Northwest-West
0
0
0
0
Pacific Northwest-East
0
0
0
0
Northeast
46
14
71
36
Lakes States
39
11
53
21
Corn Belt
38
11
53
21
Southeast
20
-10
-8
-30
South Central
0
-20
-44
-60
Rocky Mountains
0
0
0
0
Pacific Southwest
0
0
0
0
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-15
~ In all scenarios, hardwood yields do not decrease, or else they increase, in at least
seven of the nine regions. This can potentially create opportunities for substituting
hardwoods for softwoods. As it turns out, this is a complicated issue that is explored
at the end of Chapter 6.
4.3.3 Limitations of the Economic Analysis
No economic model or analytical approach is perfect. Both will suffer from limitations, and it
is always important to identify these limitations and point out how they might affect the
economic analysis. We have identified three important limitations of the economic analysis in
this report. Although they have already been discussed in connection with the analysis, it is
helpful to present them together.
~ Agricultural and forest sectors are not linked. The linked, two-sector model was not
used in this analysis because of the lack of consistent physical effects scenarios across
the two sectors in existing studies. Linkages between the two sectors are important
when climate change differentially affects the opportunity cost of land across sectors,
causing land to move to better opportunities. There is no current scientific basis for
determining whether this will be the case. We simulated hypothetical yield changes in
both sectors using the linked model However, the results did not show large land
shifts between the two sectors relative to the base case.
~ Climate impacts on forests in other regions were not included in the analysis. This
type of analysis was not done because of the lack of information about effects of
climate change on the export demand and import supply curves in the FASOM model
Climate change will influence forest productivity in other regions of the world These
changes have the potential to indirectly shift the comparative that various regions have
in the production of stumpage and wood products made from stumpage relative to
other goods In that case, changes in trade patterns in wood products would be
accompanied by changes in stumpage and product prices, harvest levels, and inventory
stocks, and the welfare of U.S consumers and producers. If forest productivity
increases in other countries relative to the United States, then America could
potentially import more logs and primary products and, thus, offset some of the
economic losses that would occur otherwise. This possibility is particularly relevant
when it comes to Canada.
*- The model does not include specific MICs or technological advances to adapt to
the effects of climate change. While the model has a wide range of management
intensity classes to deal with current conditions, none were specifically tailored to the
direct effects of climate change. This limitation was not thought to be serious in light
of the management opportunities already in the model
RCG/Hagler Bailly
Final Report
-------
FASOM Model and Economic Analysis Methods ~ 4-16
4.4 References
Alig, RJ., J.M. Vasievich, and K.J. Lee. 1992. "Economic Opportunities to Increase Timber
Growth on Timberland." In A. Qureshi, ed. Proceedings of the Conference: Forests in a
Changing Climate. Climate Institute, Washington, DC. pp. 115-125.
Birdsey, R. 1992. Forests and Global Change: Vol. I—Opportunities for Increasing Forest
Cover. In R.N. Sampson and D. Hair, eds. American Forestry Association. 285 pp.
Callaway, J.M., F. Cronin, J.W. Currie, and J. Tawill. 1982. An Analysis of Methods and
Models for Assessing the Direct and indirect Economic Impacts of COj-Induced Climate
Changes in the Agricultural Sector of the U.S. Economy. PNL-4384. Pacific Northwest
Laboratory, Richland, WA.
Haynes, R., D. Adams, and J. Mills. 1994a The 1993 RPA Timber Assessment Update. USDA
Forest Service, PNW Station. Draft Report.
Haynes, R., R. Alig, and E Moore. 1994b. "Alternative Simulations of Forestry Scenarios
Involving Carbon Sequestration Options* Investigation of Impacts on Regional and National
Timber Markets " USDA Forest Service, PNW Station, Research Paper—in press.
Ince, P.J 1994 Recycling and Long-Range Timber Outlook. USDA, Forest Service, Rocky
Mountain Forest and Range Experiment Station, General Technical Report RM-242. Fort
Collins, CO.
Mills, J. and J. Kincaid 1992 The Aggregate Timberland Assessment System—ATLAS: A
Comprehensive Timber Projection Model USDA Forest Service General Technical Report
PNW-GTR-281. PNW Station, Portland, OR 160 pp.
Moulton, R. and K. Richards. 1990 Costs of Sequestering Carbon Through Tree Planting and
Forest Management in the United States. USDA Forest Service, GTR WO-58. Washington,
DC.
Powell, D., J Faulkner, D. Darr, et al. 1993. Forest Resources of the United States, 1992.
USDA Forest Service General Technical Report RM-GTR-234 Rocky Mountain, Ft. Collins,
CO. 132 pp.
Risbrudt, C and P. Ellefson. 1983. An Economic Evaluation of the 1979 Forestry Incentives
Program. Station Bulletin 550. Agricultural Experiment Station, University of Minnesota, St.
Paul, MN.
RCG/Hagler Bailly
Final Report
-------
Chapter 5
Results of the Economic Analysis
In this chapter we present the major results for the base case and the climate change
scenarios. We focus on the time paths of effects of the simulated yield changes on
~ the magnitude and composition of consumer and producer surpluses
*¦ the magnitude and regional distribution of producer surpluses
» aggregate product prices
*¦ aggregate production (i.e., harvest) levels
~ inventory levels
5.1 The Base Case Summarized
The results for the base case are presented and explained in detail in Appendix F of Adams et
al. (1994) Here we provide a summary of the trends in the base case that we believe are
directly relevant to the climate change scenarios. These trends are the following:
*¦ Demand shifts. Product demands in FASOM increase (i.e., the demand functions shift
out) from 1990 to 2030 in response to exogenous forecasts for key driver variables
such as GNP and housing starts. Demand function parameters are held fixed after
2030. No explicit assumptions about technological change are assumed in these
forecasts.
+ Supply shifts. As indicated previously, no specific technological advancements are
included other than those that are reflected in current management practices and
genetic stock.
*¦ Product prices. For the most part, aggregate product prices did not change much over
time, rising from about $1 35/cu ft in 1990 to $1.43/cu ft in 2060.
~ Production levels. Aggregate production (i.e., harvest) levels for hardwoods and
softwoods (combined) increased over the period 1990-2030 from about 170 billion
cu ft to 230 billion cu ft and then leveled off as demands stabilized in 2030.
* Inventory volumes. Projected inventory volumes for hardwoods and softwoods
(combined) rose from about 440 billion cu ft in 1990 to 490 billion cu ft in 2030 and
then dropped back to 475 billion cu ft by 2060
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis ~ 5-2
* Inventory structure. The major structural change in the inventory during this period
was an increase in softwood inventory volumes and a decrease in hardwood inventory
volumes, due to conversion of hardwood to softwood acres and downgrading
~ Management intensity. The combined features of increased production and relatively
stable product prices can be explained in part by the conversion of acreage from
hardwoods to softwoods and the redistribution of the inventory from medium and low
management intensities into the high and passive management (low-low) management
intensity categories, respectively.
5.2 Climate Change Scenarios
In this chapter, we present the results of the climate change scenarios grouped by topic (i.e.,
welfare, prices, production, and inventory levels) rather than by scenario to make it easier for
the reader to compare the results across scenarios.
5.2.1 Welfare
Table 5-1 summarizes the effects of the yield changes on the magnitude and composition of
welfare components in the forest sector. Consumer surplus includes the surplus of domestic
buyers of hardwood and softwood stumpage, primarily firms. Producer surplus is the surplus
of private timberland owners in the United States. Foreign trade surplus is the surplus of
offshore buyers of U.S. stumpage. The terminal inventory surplus is measured by the market
sale value of the inventory at the end of the period (2080). The public cut surplus is the
producer surplus associated with the exogenous harvest of public lands. In theory, these rents
could be captured by the federal government; however, in practice, a substantial portion of
these rents accrue to private owners who contract with the government to harvest the timber
on public lands.
Decreases in the annualized value of total surplus range from about 4% in the 2.5 with C02
scenario to about 19% in the 4 0 without C02 scenario. This translates into fairly substantial
annualized1 losses: $2.5 billion/yr and $12 billion/yr, respectively. These losses are near the
lower end of the range of losses estimated by Callaway et al. (1986)—$3.4 to $5.0 billion/yr
—and more recently by Cline (1992)—$3.3 billion/yr. In general, the surplus losses, as well
as the changes in the individual surplus components, increase with both the magnitude of the
temperature change and the C02 fertilization effect The one possible exception to this is the
2.5 without C02 scenario in which the decrease in total welfare and the changes in the
individual welfare components are higher in magnitude (in both directions) than in the 4.0
1 The annualization factor is (r) * [1 - (1 + r)'0]"1 = 0.0412 for n = 90 and r = 0 04.
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-3
Table 5-1
Annualized Values1 of Welfare Components
for the Base Case and Percent Changes from the Base Case
for Four Climate Change Scenarios
Base
2.5
IS
4.0
4.0
Welfare Component
($ Millions)
with C03
without COz
with COj
without C02
Consumer Surplus
56,748
-17.42%
-33.47%
-28.34%
-42 09%
Producer Surplus
3,871
145.67%
237.40%
194.50%
207.69%
Foreign Trade
324
-21.79%
-39.34%
-35.62%
-50.10%
Surplus
Terminal Inventory
1,805
-6.13%
-13.89%
-11.33%
-33.13%
Public Cut
2,023
79.56%
159.84%
137.05%
222.87%
Total Surplus
64,771
-4.35%
-10.73%
-9.42%
-18.68%
1 Annualized values calculated for 1990-2080 using a discount rate of 4%
with C02 scenario As suggested previously, this is because the softwood yield decreases in
the South are higher in the 2 5 without C02 scenario than in the 4.0 with C02 scenario.
The magnitude of the gains and losses in the individual welfare components also varies
considerably over the four scenarios. For example, consumer surplus is reduced by 17 42%
($9 9 billion/yr) in the 2.5 with C02 scenario and by 42.09% ($24 billion/yr) in the 4 0
without C02 scenario. Producer surplus, on the other hand, increases in all scenarios. The
welfare gains for this group range from 145.67% ($5.6 billion/yr) in the 2.5 with C02
scenario to 207 69% ($8 billion/yr) in the 4 0 without C02 scenario. The difference in the
direction of the changes in consumer and producer surpluses is explained by the extremely
inelastic nature of the product demands in U.S stumpage markets. Reductions in forest yields
shift the supply curves for stumpage to the left in each period, causing prices to increase
along these inelastic demand curves Consumer surplus decreases (and producer surplus
increases) because landowners can pass most of the price increases directly on to buyers of
stumpage. This is a classic result, observed in previous studies of the forest sector's response
to reduced yields by Callaway et al (1986) and Callaway (1991).
Foreign buyers of stumpage experience larger welfare losses as yields decrease for the same
reasons as domestic stumpage buyers: higher prices along inelastic demand functions The
value of the terminal inventory falls as the severity of the yield reductions increase. This is
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-4
because, even though the "prices" associated with units of the terminal inventory increase as
the inventory shrinks, the quantity demanded along the terminal inventory function declines as
prices increase. Thus, the product of the two must decrease. In FASOM, the public cut is
exogenous since this, historically, has been a public policy decision. The same exogenous
forecast for the public cut was used in all cases. This forecast sees the public cut dropping to
2.06 billion cu ft in 2000, rising slowly to 2.2 billion cu ft by 2030, and constant thereafter.
(This is consistent with current USDA Forest Service policies to eliminate clear cutting and
protect endangered species.)
Estimates of the future (i.e., undiscounted) value of the regional producer surpluses are
displayed in Figures 5-1 to 5-3 for 2000, 2030, and 20602 for the base case and the four
climate change scenarios. The North (Figure 5-1) consists of the Northeast, Great Lakes
States, and Corn Belt. The South (Figure 5-2) includes the Southeast and the South Central,
and the West (Figure 5-3) includes all of the Pacific Northwest, the Rocky Mountains, and
the Pacific Southwest Several important results emerge from these figures Generally
speaking, the aggregate results seen in Table 5-1 are mirrored in the regional figures: reduced
yields tend to increase producer surplus. This is particularly true in the North where the future
values of producer surplus increase consistently, over time, with the severity of the yield
reductions. This also true, but to a lesser extent, in the West, where the increases in producer
surplus tend to stabilize after 2030. The pattern is most uneven in the South, where the
producer surpluses initially fall relative to the base case in the two with C02 scenarios and
increase in the two without C02 scenarios. Also, the yield losses in the 4.0 without C02
scenario are so extreme that, unlike other regions, the producer surplus gains are smaller than
in all the other climate change scenarios for 2030 and 2060.
Finally, the severity of the yield losses also affects the ordering of the producer surpluses
across regions. In the base case, the South has the highest producer surplus, followed in order
by the West and the North. However, this order is reversed somewhat in 2000, with the West
taking over as the leader. The sudden decline of the South in this regard is explained not only
by the very large yield decreases in the region, but also by the fact that the rotations are
shorter and that it is the major producing region of stumpage in 1990. Thus, it is initially
affected more adversely than all the other regions. In later years, the South also falls behind
the West in terms of the magnitude of producer surplus in the 2 5 without C02 (in 2030) and
4 0 (in 2030 and 2040) without C02 sensitivity cases. When yield losses are very high, the
direct effect of reduced output on revenues outweighs the indirect effects of inelastic demands
on prices. Therefore, profits fall
2 Although the projection period is 1990-2080, information for FASOM is truncated at 2060 in all
figures in case terminal conditions influence the 2070 and 2080 projections.
RCG/Hagler Bailly
Final Report
-------
I Base
~ 2.5 with C02
CU2.5 without C02
04 with C02
04 without C02
-------
Billions
$140
$120
$100 -
$80 -
$60 -
$40 -
$20 -
Figure 5-2
Projected Producer Surplus for the South
Base Case and 4 Climate Change Scenarios
2000, 2030, and 2060
I Base
02.5 with C02
D2.5 without C02
^4 with C02
04 without C02
-------
Billions
Figure 5-3
Projected Producer Surplus for the West
Base Case and 4 Climate Change Scenarios
2000, 2030, and 2060
$120
^ i
a £
ft O
O CO
H
$100
$80
$60 -
$40-
$20 -
$0
m
2000
I Base
02.5 with C02
D2.5 without C02
S3 4 with C02
~ 4 without C02
B
C/5
§
to
S
B
1
O
«<
to
S3
-------
Results of the Economic Analysis * 5-8
5.2.2 Product Prices
FASOM projections of aggregate product price indices, as measured by the Tornquist-Theil
price index, are shown for softwood in Figure 5-4, for hardwood in Figure 5-5, and for the
two species combined in Figure 5-6. The index in all scenarios is based (i.e., it is equal to
unity) at 1990 price levels for softwood and hardwood products (pulp, sawtimber, and
fuelwood) in the base case.3 As can be seen in Figure 5-4, softwood prices increase
dramatically relative to the base case in all of the scenarios. These dramatic price increases
are due to reductions in timber supplies in each period as a result of the yield reductions in
the in the base case, climate change scenarios. The average annual price increases for 1990 to
2060 range from about 1%/yr for the 2.5 with C02 scenario to about 1.8%/yr for the 4.0
without C02 scenario. This is approximately consistent with a doubling and quadrupling of
softwood product prices, relative to base case prices in 1990. And it is these high prices that
explain the substantial losses in consumer surplus and increases in producer surplus, shown
earlier in Table 5-1
Predictably, the direction of the changes in hardwood prices in Figure 5-5 is largely (but not
entirely) negative, both in absolute terms and relative to the base case. The pattern of price
changes in all scenarios is the same, falling from 1990 to 2030 and then rising from 2030 to
2060 Hardwood prices in the two with C02 scenarios are consistently lower than in the base
case. Hardwood prices in the two without C02 scenarios are lower than in the base case until
2040 to 2060, when they rise The fact that hardwood prices stay fairly low is because
hardwood yields in all scenarios increase in the North, are constant in the West, and fall only
in the South. The upward trend in hardwood prices from 2030 onward reflects a projected
increase in the management intensity of hardwoods, as producers try to make up for reduced
softwood productivity. The large disparity between the growth in softwood and hardwood
price indices indicates that FASOM finds it unprofitable to convert hardwood growing stock
acreage to softwoods, thereby easing the price pressure on softwood products. We will further
explore this issue at the end of this chapter
Finally, the combined price indices in Figure 5-6 reflect the importance of softwood
production. The price indices for all the climate change scenarios are slightly lower in each
year than the corresponding softwood price indices
3 The Tomquist-Theil price index (PJ is calculated from i = 1, . . ., n goods for penod t = 0, . . ., T
from the following: ln(P,) = £, 0 5(s„ + s,„) ln(Plt/Pl0), where su = the expenditure share of product i in
period t and Plt = the price of product i in period t. In our calculations, all indices are evaluated using
1990 base case prices and shares for and s,o Thus the base period index = 1.00 in 1990.
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-9
Figure 5-4
Projected Tornquist-Theil Softwood Price Index
for Softwood Forest Products Prices
Base Case and 4 Climate Change Scenarios, 1990 to 2060
1.0 Units
4 -
3 -
2040
2050
2060
1990
2000
2010
2020
2030
Year
Base 2.5 with C02 ® 2.5 without C02 4 with C02 4 without C02
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis »• 5-10
Figure 5-5
Projected Tornquist-Theil Hardwood Price Index
for Hardwood Forest Products Prices
Base Case and 4 Climate Change Scenarios, 1990 to 2060
1.0 Units
5 -i
4 -
3 -
1990 2000 2010 2020 2030 2040 2050 2060
— Base * 2.5 with C02 ® 2.5 without C02 4 with C02 4 without C02
0
Year
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis *>5-11
Figure 5-6
Projected Tornquist-Theil Price Index
for Forest Products Prices
Base Case and 4 Climate Change Scenarios, 1990 to 2060
1.0 Units
4 -
3 -
2 -
1990
2000
2010
2020
2030
2040
2050
2060
Year
— Base * 2.5 with C02 «¦ 2.5 without C02 + 4 with C02 4 without C02
RCG/Hagler BaiUy
Final Report
-------
Results of the Economic Analysis ~ 5-12
5.2.3 Production (Harvests)
FASOM projections of the production of pulpwood, sawtimber, and fuelwood (combined) are
shown for softwoods in Figure 5-7, for hardwoods in Figure 5-8, and for the two species
combined in Figure 5-9. In the base case, softwood production is initially flat at about
110 billion cu ft/yr because of limited supplies in the South and Pacific Northwest in the
merchantable age classes. Thereafter, production increases, as demands shift out, until 2030,
when product demands stabilize and a more or less "steady state" level of softwood
production is achieved at about 150 billion cu ft/yr. The impact of the yield reductions on
production follows a common pattern in almost all of the scenarios. Production in all
scenarios drops sharply from 1990 to 2000 because of the shortage of timber in the
merchantable age classes. After that, production increases slightly up until 2030 in all of the
scenarios, except for the 4.0 without C02 scenario in which it is more or less constant from
2000-2010 to 2060. For the remaining scenarios, production decreases slightly after 2030.
Overall, production levels in 2030 are about 25% (2.5 with C02) to 65% (4 0 without C02)
below those in the base case, depending on the severity of the yield reduction. This spread is
only slightly larger by 2060.
Hardwood production (Figure 5-8) in the base case follows the same general pattern as
softwoods, although the increase in production through 1990 to 2030 is more or less constant
in response to increasing demands for hardwood products. After 2030, hardwood product
demand curves, like their softwood counterparts, cease to shift out in FASOM, and production
stabilizes at about 80 billion cu ft/year. Hardwood production in the four climate change
scenarios follows this same pattern However, hardwood production in all of the climate
change scenarios increases more rapidly than in the base case, such that in 2030 hardwood
production levels in all these scenarios are slightly higher than in the base case This is
possible due to substantial increases in hardwood inventory acreage in regions where yields
are increased, i.e, in the North (and in the Southeast in the 2.5 with COz scenario) After
2030, production dips back down in three of the four climate change scenarios However, in
the 2 5 with C02 scenario, increased hardwood supplies in the North and Southeast make it
possible to sustain production levels somewhat above those in the base case (and at lower
prices), even after 2030.
Combined hardwood and softwood production levels (Figure 5-9), like the aggregate price
indices, reflect the importance of softwood in total production. Combined production levels in
the climate change scenarios in 2030 range from about 15% (2.5 with C02) to 40% (4.0
without COz) below base case values, depending on the severity of the yield losses. Decreases
in production by 2060 are somewhat larger.
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis *5-13
Figure 5-7
Projected Softwood Production
Base Case and 4 Climate Change Scenarios
1990 to 2060
Cu. Ft. (Billions)
150 -
130-
110
90 -
70 -
50-
2040
1990
2000
2010
2020
2030
2050
2060
Year
Base "*"2.5 with C02 "B" 2.5 without C02 "*"4 with C02 ~*~4 without C02
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis * 5-14
Figure 5-8
Projected Hardwood Production
Base Case and 4 Climate Change Scenarios
1990 to 2060
Cu. Ft. (Billions)
100 -
90 -
80 -
70 -
50
2050
1990
2000
2040
2060
2010
2020
2030
Year
Base 2.5 with C02 2.5 without C02 4 with C02 4 without C02
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-15
Figure 5-9
Projected Total Production
Base Case and 4 Climate Change Scenarios
1990 to 2060
Cu. Ft. (Billions)
240
220-
200 -
180 -
160
140 -
120 -
100
2000
1990
2010
2030
2040
2020
2050
2060
Year
Base 2.5 with C02 "®" 2 5 without C02 4 with C02 4 without C02
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis ~ 5-16
5.2.4 Inventory Levels
FASOM projections of inventory volumes over time are shown for softwoods in Figure 5-10,
for hardwoods in Figure 5-11, and for the two forest types combined in Figure 5-12. The base
case softwood inventory level (Figure 5-10) drops from 1990 to 2000 because of the lack of
timber in the merchantable age classes in the South and Pacific Northwest. However, from
2000 onward, until 2060, the softwood inventory in FASOM increases, from 178 billion cu ft
in 2000 to 240 billion cu ft in 2060. The climate change scenarios share with the base case an
initial drop in softwood inventory stocks, although the magnitude of these decreases increases
with the severity of the yield decrease. The inventory drop in the 2.5 with C02 scenario is
just about equal to that in the base case. However, the softwood inventory in the 4.0 without
C02 scenario is about 40% below the base case inventory in 2000. After 2000, the inventory
levels in all of the climate change scenarios decrease very slightly. By 2030, the softwood
inventory in the 2.5 with C02 scenario is about 22% below the base case inventory, whereas
the inventory level in the 4.0 without C02 scenario is fully 70% below the base inventory in
that same year
As expected, the hardwood inventory levels shown in Figure 5-11 mirror the changes in
hardwood prices and production levels. Inventory levels in the 2.5 with C02 and 4.0 with C02
scenarios increase initially, faster than in the base case, and remain higher than the base
inventory level until 2030 After 2030, inventory levels in both these scenarios decline. By
2060, the hardwood inventory in the 2.5 with C02 scenario is about the same as in the base
case, while the inventory in the 4.0 with C02 scenario is only about 5% lower than in the
base case Hardwood inventory levels in the two without C02 cases are consistently lower
than those in the base case. By 2060, the inventory levels in these two scenarios converge
together at a point about 25% lower than the hardwood inventory in the base case.
The hardwood inventory is consistently larger than the softwood inventory in all the scenarios
(except the base case). However, the species composition of the combined inventory varies
considerably from scenario to scenario and from year to year. This is reflected in the pattern
of inventory changes shown in Figure 5-12. For example, in the 2.5 with C02 scenario the
inventory expansion from 1990 to 2020-2030 and the decline thereafter largely reflect the
movement of the hardwood inventory. For the other climate change scenarios, the expansion
of the hardwood inventory from 1990 to 2030 helps to moderate the decrease in total
inventory levels; thereafter the steady decline in the combined inventory levels also tends to
reflect the decrease in hardwood inventories. By 2060, the difference between the combined
inventory volume in the base case and the four climate change scenario ranges from about
15% (2.5 with C02) to about 45% (4.0 without C02)
It should be noted that inventory volumes cannot decline permanently and still sustain
production. The figures in this section do not show the inventory volumes in 2070 and 2080,
which are still lower than those observed in 2060. This raises the very serious issues of the
sustainability of the U S forest sector in the face of the types of yield declines used in this
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-17
Figure 5-10
Projected Softwood Inventory Volume
Base Case and 4 Climate Change Scenarios
1990 to 2060
Cu. Ft. (Billions)
250
200 -
150 -
100 -
1990
2000
2010
2020
2030
2040
2050
2060
Year
Base 2.5 with C02 "®" 2.5 without C02 4 with C02 4 without C02
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis * 5-18
Figure 5-11
Projected Hardwood Inventory Volume
Base Case and 4 Climate Change Scenarios
1990 to 2060
Cu. Ft. (Billions)
350
300-
200 -
150
1990
2000
2050
2060
2020
2040
2010
2030
Year
Base "*"2.5 with C02 "®"2.5 without C02 "*"4 with C02 "~"4 without C02
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis ~ 5-19
Figure 5-12
Projected Total Inventory Volume
Base Case and 4 Climate Change Scenarios
1990 to 2060
Cu. Ft. (Billions)
550
500 -
450 -
400 -
350 -
300 -
250 -
200
1990
2000
2010
2030
2050
2020
2040
2060
Year
Base "*"2.5 with C02 "®" 2.5 without C02 "*"4 with C02 "~"4 without C02
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-20
study. Unfortunately, we did not run the model for a period longer than 90 years to further
examine this problem. Having said that, we note that FASOM does not include the Canadian
stumpage market. Softwood inventory volumes in Canada are substantial and, because of
Canada's location, global warming could actually enhance timber yields in most of that
country In that case, the Canadian timber supplies would act as a "safety valve" and relieve
some of the pressure on the U S. domestic inventory. Depending on the magnitude of the
yield increases in Canada, adequate timber supplies could well be maintained to keep product
prices from rising above the $3-$4/cu ft level. This would lower consumer surplus losses, but
it would also greatly reduce, and perhaps reverse, the producer surplus gains observed in this
study. Thus, the domestic winners and losers from global climate change could be reversed if
forest yields in Canada increase substantially.
5.3 Substitution Issues
One of the important features of the climate change scenarios was the difference in the yield
changes across the two species. As shown in Table 2-2, hardwood yield reductions were
never as severe as softwood reductions in the South, and hardwood yields increased in all of
the scenarios in the North and remained constant in the West. As a result, we saw that while
softwood prices increased dramatically in all the scenarios, hardwood prices were fairly flat;
that hardwood production did not deviate greatly from the base case; and that hardwood
inventory volumes in some cases were higher than base case levels. This portrays a picture of
a "two-tier" timber economy with either limited room for substitution among species or one in
which whatever substitution opportunities exist have already been exploited.
From a conceptual standpoint, there are three different points at which substitution between
species can occur in the timber economy First, landowners can harvest their timber of one
species and reforest that land in another species Second, the stumpage of one species can be
processed at the mill as a product from another species. An example of this would be the
ability to substitute hardwood pulpwood for softwood pulpwood. Finally, end use demands
can shift such that the product demand for one species increases while the demand for another
species falls. As the relative price of softwood increased, one might become willing to
substitute hardwood lumber for softwood lumber 4 However, this assumes that such
substitution possibilities exist. At the current time, this is not the case and there are only
limited opportunities for this to occur because of the price-induced evolution of the end-use
species match.
FASOM is currently designed, conceptually, to "handle" all types of substitution between
products. However, extreme substitutions between hardwood and softwood at higher market
4 Substitution away from wood to other materials is governed by exogenous shifts in the demand
functions as market conditions change over time.
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-21
levels are not reflected in the observed data for stumpage supplies and demand Thus, these
substitutions cannot be reflected very well in the estimated parameters of the stumpage
demand equations. In FASOM, land can be converted from hardwoods (or softwoods) to
softwoods (or hardwoods) after harvest. This conversion must involve planting the new
species, so a positive cost is associated with this type of land substitution. These costs are in
the range of $25 to $85 per acre. Currently, FASOM allows downward, interproduct
substitution from sawtimber to pulp for the same species, and from hardwood and softwood
pulp to softwood fuelwood. The model does not allow interspecies substitution for either
pulpwood or sawtimber. Interspecies substitution for sawtimber is technically costly and, for
some products, simply infeasible. However, interspecies substitution for pulpwood is currently
being practiced and is feasible at very low costs in some regions.
We decided to explore the impact of increasing both land substitutability and interspecies
product substitution on the climate change scenarios. Because this was an exploratory
investigation, we limited the scope of the analysis to a single scenario, the 4.0 without C02
case This scenario was selected because the very large reductions in softwoods yields in this
scenario, especially in the South, should make hardwood substitution more valuable to the
model.
We performed two sets of experiments on this scenario. First, we reduced the cost of
converting land from hardwood to softwood growing stock systematically by 25% and 50%
We selected these percentages because they were indicative in past practice of the proportion
of establishment cost that the federal government has historically been willing to subsidize in
its reforestation programs Second, we reset the conversion costs to their original (base case)
levels and then added the possibility of substituting hardwood pulp for softwood pulp at the
mill We added this possibility because it is currently feasible in some regions For both of
these experimental cases, we ran FASOM with and without the yield changes in the 4 0
without C02 scenario.
The results of the land substitution experiment were clear cut. There was virtually no effect
on the species composition of the inventory, on price and production levels over time, or on
welfare This is because the large softwood yield reductions in the 4.0 without C02 scenario
make it economically infeasible to convert hardwood acreage to softwood acreage at even
50% of the current conversion cost. This type of substitution might become profitable at
lower conversion costs, but we did not explore this possibility, since there is no reason to
believe that such cost reductions would be achieved or that the government would be willing
to cost share a larger amount.
The second experiment was more successful at producing substitution However, before
looking at the results of this experiment, it is helpful to look at the projected price paths for
softwood and hardwood pulpwood shown in Figures 5-13 and 5-14. As can be seen, the
prices diverge sharply in all the scenarios. For the 4.0 without C02 scenario, softwood prices
in 2030 and 2060 are about $6.00/cu ft and $6.10/cu ft, respectively, and the corresponding
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis ~ 5-22
Figure 5-13
Projected Price Path
Softwood Pulpwood
1990 to 2060
$6 -
$5 -
$4-
$3 -
$2 -
$1 "
1990
2000
2050
2010
2020
2030
2040
2060
Year
Base 2.5 with C02 ® 2.5 without C02 4 with C02 4 without C02
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis »> 5-23
Figure 5-14
Projected Price Path
Hardwood Pulpwood
1990 to 2060
$6 -
$5 -
$4 -
$3 -
$2 -
1990
2000
2010
2020
2030
2040
2050
2060
Year
Base "*"2.5 with C02 "®"2.5 without C02 "*" 4 with C02 "*"4 without C02
RCG/Hagler Bailly
Final Report
-------
RESULTS OF THE ECONOMIC ANALYSIS ~ 5-24
hardwood pulpwood prices are in the neighborhood of $1.10 to $1.20. The $5.00 or so price
difference between these products suggests that there are ample opportunities for substitution
The results of the second substitution experiment (hardwood pulp for softwood pulp) can be
seen partially in Figures 5-15 through 5-17. Figure 5-15 shows the price paths of softwood
pulpwood for the original base case, for the revised base case (with interspecies substitution),
for the 4.0 without C02 scenario, and for the revised 4.0 without C02 scenario The important
feature about this figure is that the price path for pulpwood in the revised case increases at a
much more gradual rate than in the original case. In 2030, the price of pulpwood in the
revised 4.0 without C02 scenario is about $2.50/cu ft, as opposed to about $6.00/cu ft in the
original 4.0 without C02 scenario. By 2060, the difference has narrowed somewhat, but it is
still substantial; pulpwood sells for about $3.55/cu ft in the revised 4.0 without C02 scenario
and about $6.10/cu ft in the original 4.0 without C02 experiment.
Figure 5-16 shows the price paths of hardwood pulpwood for the original base case, for the
revised base case (with interspecies substitution), for the 4.0 without COz scenario, and for
the revised 4.0 without C02 scenario Since softwood and hardwood pulpwood are perfectly
substitutable, the price paths for softwood pulpwood and hardwood in the revised base case
and revised 4 0 without C02 scenario track each other perfectly, separated only by a $.10
processing cost differential. Thus, in the case of hardwood pulpwood, the introduction of
substitution at the mill leads to much sharper increases in the price of hardwood pulpwood
than in the original scenarios.
Finally, Figure 5-17 presents the Tornquist-Theil price indices for the two base cases and two
4 0 without C02 scenarios. With substitution, the price index in the revised 4.0 without C02
falls below that in the 4.0 without C02 scenario until 2040, when the two indices cross, and
then the aggregate price in the revised case is somewhat higher. Recall that the index is based
on production in 1990 for the original base case. If the index were re-based in 1990 for the
revised base case, then the aggregate prices for the revised 4 0 without C02 scenario would
be consistently lower.
Table 5-2 compares the welfare estimates for the original and revised base case and 4 0
without C02 scenarios. Enhancing the substitution possibilities in this experiment substantially
reduces welfare losses. Overall, the percentage reduction in the annualized value of total
welfare in the 4.0 without C02 case was 18.7% or a loss of about $52.7 billion, whereas in
the revised 4.0 without C02 scenario this was reduced to a welfare loss of 14.4%, or about
$55.6 billion, which translates into about $9.3 billion/yr when annualized Thus, the
introduction of technological change in this experiment reduced net welfare losses by
approximately $3 billion/yr
Enhancing the substitution possibilities in the model also affected the distribution of welfare
gains and losses. In the revised 4.0 scenario, both producer and consumer surpluses are
increased relative to the original 4 0 without C02 scenario Relative to the revised base, the
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis •> 5-25
Figure 5-15
Projected Softwood Pulpwood Price Path
for Base Case, 4 Hi Scenario, Rev. Base Case, Rev. 4 Hi Scenario
1990 to 2060
$6 -
$5 -
$4-
$3 -
$2 -
SI -
1990
2000
2010
2020
2030
2040
2050
2060
Year
Base ~*"4 wihtout C02 ""Rev. Base "~"Rev. 4 without C02
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-26
Figure 5-16
Projected Hardwood Pulpwood Price Path
for Base Case, 4 Hi Scenario, Rev. Base Case, Rev. 4 Hi Scenario
1990 to 2060
~
1990 2000 2010 2020 2030 2040 2050 2060
Year
Base "~"4 without C02 """Rev. Base '~'Rev. 4 without C02
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis > 5-27
Figure 5-17
Projected Tornquist-Theil Price Index
for Forest Products Prices
Base Case, 4 Hi Scenario, Rev. Base Case, Rev. 4 Hi Scenario, 1990 to 2060
1.0 Units
4 -
2 -
1990
2000
2010
2020
2030
2040
2050
2060
Year
Base 4 wihtout C02 Rev. Base ~ Rev. 4 without C02
RCG/Hagler Bailly
Final Report
-------
Results of the Economic Analysis ~ 5-28
Table 5-2
Annualized Values1 of Welfare Components
for the Base Case and Revised Base Case
Compared with the Original and Revised 4.0 without C02 Scenarios
(S Millions)
Welfare Components
Base
4.0
without C02
Revised
Base
Revised 4.0
without C02
Consumer Surplus
56,748
32,862 (-42.1%)
56,675
33,297 (-41.2%)
Producer Surplus
3,871
11,910 (207.7%)
4,180
15,039 (259 8%)
Foreign Trade Surplus
324
162 (-50.1%)
320
161 (-49.7%)
Terminal Inventory
1,805
1,207 (-33.1%)
1,795
1,179 (-34.3%)
Public Cut
2,023
6,532 (222.9%)
1,997
5,961 (198.5%)
Total Surplus
64,771
52,673 (-18.7%)
64,967
55,637 (-14 4%)
Annualized values calculated for 1990-2080 using a discount rate of 4%.
decrease in consumer surplus in the revised 4 0 without C02 scenario was slightly less severe
than in the original 4.0 without C02 scenario (-41.5% vs -42.09%). At the same time, the
relative increase in producer surplus in the revised 4 0 without C02 scenario was quite a bit
larger than in the original 4.0 without C02 scenario (259 79% vs. 207 69%)
The two substitution experiments showed that while climate change may make it
economically infeasible to replant softwoods on hardwood acreage, welfare losses would be
reduced in all the climate scenarios by substitution of hardwood for softwood pulpwood,
where feasible It is also quite possible that large future differences between softwood and
hardwood products would not persist, because these would create long-run substitution
opportunities to replace softwood sawtimber products with hardwood sawtimber through
technological change. At the same time, these price differences might be expected to induce
further substitutions in consumption of hardwood for softwood products further up the product
chain All of these substitutions have the long-run potential to further reduce the welfare
losses simulated with FASOM in this study.
RCG/Hagler Bailly
Final Report
-------
results of the Economic Analysis ~ 5-29
5.4 References
Adams, D.M., R. Alig, J.M. Callaway, and B.A. McCarl. 1994. Forest and Agricultural
Sector Optimization Model Final Report, prepared for Mitigation Branch, Climate Change
Division, U.S EPA on Contract No. 68-W2-0018. RCG/Hagler Bailly, Boulder, CO.
Callaway, J.M. 1991. "Economics," Section 4.9 in 1990 Integrated Assessment Report, The
NAPAP Office of the Director, 722 Jackson Place, NW, Washington, DC, pp. 392-398.
Callaway, J.M., R.F. Darwin, and R.J. Nesse. 1986. "Economics Effects of Hypothetical
Reductions in Tree Growth in the Northeastern and Southeastern United States." PNL-5939.
Battelle Pacific Northwest Laboratory, Richland, WA.
Cline, W. 1992. The Economics of Global Warming Institute for International Economics,
Washington, DC.
RCG/Hagler Bailly
Final Report
-------
Chapter 6
Conclusions and Suggestions for Further Research
In this report, we reviewed published literature in which the authors estimated the effects of
climate change on forest productivity in the United States. We used the results from these
studies and other published literature about C02 fertilization to develop our own estimates of
regional and national changes in hardwood and softwood yields. The yield changes were
estimated based on results from the literature using the so-called "gap models." These models
provide regional estimates of how yields of specific species may change over time in response
to climate change. Four climate change scenarios were developed, which use different levels
of temperature change and the effects of elevated C02 on yields. In general, the results from
the literature are that climate change would lead to a decline in softwood yields across most
of the East, Central, and South, with mixed results in the Northwest and Mountain states. The
decline in softwoods is estimated to be particularly severe in the South. On the other hand,
hardwood yields are generally estimated to increase in northern states and have mixed results
in southern states. These differences in response between hardwood and softwood species and
between regions had important consequences for the economic results.
A multi-regional timber supply model of the United States was used to simulate the economic
effects of these forest yield changes on welfare, stumpage harvests and prices, and forest
inventory stocks over time. The results from this analysis indicated economic losses to
consumers and producers ranging from $2.8 billion/yr to $12 billion/yr, depending on the
severity of the changes in climate. Average product prices increased steadily in all scenarios
By 2060, these prices had risen about 100 to 200% in response to climate-induced supply
shortages Total production of stumpage products decreased in response to these higher prices.
Projected total production in 2060 was about 25 to 65% below base case levels As might be
expected, forest inventory stocks declined steadily over the period. By 2060, forest inventory
volumes were from 15 to 45% below stocks in the base case. After 2060, projected inventory
stocks continued to decline in the model, raising questions about the sustainability of the
forest inventory.
The FASOM results largely reflect the importance of the softwood resource in the South in
the overall forest economy. All of the changes in softwood prices, softwood production, and
softwood inventories were larger in absolute magnitude (i.e., worse) than for hardwoods.
Simulated increases in hardwood yields, introduced into FASOM, helped moderate welfare
losses. In general, hardwood price increases were very modest in comparison to softwood
price changes, and the effects of the scenarios on hardwood production and inventory volumes
were also quite small by comparison.
RCG/Hagler Bailly
Final Report
-------
Conclusions and Suggestions for Further Research ~ 6-2
Because of the large quantitative differences in the impacts on hardwoods and softwoods, this
study also explored the effects of lowering the costs of converting land from hardwoods to
softwoods and increasing the substitutability of hardwood pulp wood for softwood pulpwood
We found that reducing land conversion costs did not lead to hardwood-softwood land
conversion because the return on softwood acreage was still too low to justify the cost.
However, increasing the substitution possibilities in the model did help to mitigate welfare
losses to some extent. Large expected differences in softwood and hardwood prices due to
climate change would create additional incentives to expand substitution possibilities through
technological change, further mitigating the effects of climate change on welfare in the forest
sector.
6.1 Conclusions
The major conclusions of this study are as follows-
~ Studies using gap models indicate there could be significant changes in yield of U.S.
forests. Softwood yields are estimated to decline outside of the northwestern quadrant
of the country. In the deep South, softwoods are estimated to be virtually eliminated
Yields could fall significantly in the Northeast quadrant of the country. In parts of the
Northwest quadrant, softwood yields could increase.
*¦ In contrast, the literature shows that hardwood yields could increase In the North,
hardwood yields could more than double. In the South, whether yields increase or
decrease depends on the magnitude of temperature change and the C02 fertilization
effect.
~ Studies of climate change impacts on vegetation using biogeography and
biogeochemical models tend to show a wider range of variation in the potential
changes in vegetation than studies using the gap models. Some of these studies
conclude that there could be significant increases in biomass in the United States,
while others conclude there could be significant decreases. Based on a review of all of
the literature, it is uncertain whether biomass in the United States will increase or
decrease in response to climate change.
~ The FASOM analysis revealed important differences in the magnitude of the welfare
losses due both to the magnitude of the simulated climate changes and the sensitivity
of yields. The total welfare losses in the two 4.0° scenarios (-9.42% and -18.68%)
were almost twice as high as the losses in the 2.5° scenarios (-4.35% and -10.73%).
The differences in the welfare losses across the sensitivity levels (high versus low)
were almost as great ( -4.35% and -9.42% versus -10.73% and -18.68%).
RCG/Hagler Bailly
Final Report
-------
Conclusions and Suggestions for Further Research ~ 6-3
*¦ The distribution of total welfare losses due to climate change could be very uneven In
all cases, buyers of stumpage were much worse off because of the yield reductions,
while timberland owners were made much better off by the same yield reductions.
This phenomenon is due largely to the highly price-inelastic nature of stumpage
demands in the United States.
~ All of the scenarios produced long-term large declines in the softwood inventory.
These decreases called into question the sustainability of the softwood resource in the
United States in the face of yield changes predicted in the literature.
*¦ The study showed that expanded opportunities for substituting hardwoods for
softwoods at the mill and in consumption would, in the long run, help offset the
welfare losses reported in this study.
~ The study did not include Canada because of model limitations. Increases in softwood
yields in Canada could offset much of the "damage" to U.S. stumpage markets in that
increases in Canadian supplies would help to offset domestic supply losses. This
would improve the welfare condition of U.S. buyers of stumpage, but it would also
result in lower profits for U.S. timberland owners, relative to the results reported in
this study.
6.2 Suggestions for Further Research
Additional research should be conducted to improve our understanding of the biophysical
impacts of climate change on forests We need to improve our knowledge of the transient
response of forests to climate change In particular, we need a better understanding of how
forests would respond to a warmer and wetter climate and the extent to which C02
fertilization would mitigate negative climate change effects and enhance any positive climate
change effects
We believe that there are at least four areas in which additional research would further
enhance our ability to understand how the U S. forest sector might respond to changes in
climate.
The first involves developing scenarios with significant increases in timber yields in the
United States The VEMAP study (see Chapter 2) found that a number of vegetation models,
run under similar climate change scenarios as those used in studies on which we relied for
this report, estimated increases in vegetation cover and productivity over the United States. To
reflect the potential for increased yields, the results from these models should be transformed
into regional changes in hardwood and softwood yields and run through FASOM.
RCG/Hagler Bailly
Final Report
-------
Conclusions and Suggestions for Further Research ~ 6-4
The second area involves mnning FASOM with climate change scenarios included in both the
agricultural sector and the forest sector. This study was performed using the stand-alone forest
sector version of FASOM. As a result, the land base available to the forest sector was
exogenous. However, since the simulated impacts of the yield changes on the forest
inventories were fairly substantial, it would make sense to allow land to move back and forth
between the two sectors in response to endogenously determined changes in land rents in the
two sectors. This would require developing consistent yield change scenarios and assumptions
across the two sectors.
The third area of research that would improve this study involves expanding the FASOM
forest (and possibly agricultural) sectors to include Canada. The softwood resource base in
Canada is currently quite large. Canadian stum page does not compete with U.S. supplies in
our domestic market because of trade barriers, other opportunities in overseas markets, and
higher marketing costs. However, since the relative impacts of climate change on forest yields
will probably favor Canada, the economic situation could change and with it would come
incentives to reduce trade barriers Thus, as we have suggested above, the omission of Canada
from the analysis could seriously bias the impacts, causing welfare losses to be higher than
they would be if Canada were included in the analysis. One potentially troublesome aspect of
including Canada in FASOM is that it would involve the development of data in FASOM for
the Canadian softwood inventory, which is poorly documented. However, it would be possible
to develop the Canadian inventory in much less detail than applies to the U.S. inventoiy. One
could then conduct some preliminary analyses with the modified version of the model to
determine if the inclusion of Canada produces enough uncertainty in the results to warrant
further data development.
Alternatively, and perhaps more generally, it may be possible to incorporate the impacts of
climate change effects in other regions on exports and imports in FASOM. This could
possibly be done using the global forest model currently maintained by the Center for
International Trade in Forest Resources (CINTRAFOR). By running climate change scenarios
and physical effects scenarios through this model, it would be possible to develop information
that could be used to parameterize the import supply and export demand functions in FASOM
to reflect climate change impacts
Fourth, we believe it is important to extend the analysis of interproduct and interspecies
substitution to determine the extent to which further substitution possibilities can reduce
welfare losses The most obvious initial effort would involve extending the pulp substitution
analysis to all the scenarios, including also any costs that might be required to make these
substitutions technically feasible Further efforts could be focused on delimiting and including
feasible substitutions of hardwood sawtimber for softwood sawtimber and on exploring
interproduct substitution in demand
RCG/Hagler Bailly
Final Report
-------
Conclusions and Suggestions for Further Research * 6-5
Finally, it should be recalled that the yield shifts used in the analysis were introduced in an
equilibrium adjustment framework. For any given set of yields, this type of analysis produces
fairly severe negative economic impacts. However, one can be reasonably certain that, given
time to adjust to these impacts, the actual impacts of yield changes similar to the ones used in
this study would be smaller. How much smaller is difficult to determine because of the
problems associated with forecasting adaptive technological change. However, given the
magnitude of the impacts seen here, it may be worthwhile to look at how the impacts of yield
changes will affect stumpage markets, if these changes are introduced slowly, for example,
over the period from 1990 to 2050 or so.
Including all of these refinements into the analysis would greatly improve our understanding
of the long-run responsiveness of the forest sector in the United States to climate change.
RCG/Hagler Bailly
Final Report
-------
Appendix A
-------
Appendix A
Details on Estimating Changes in Growing Stock of
Softwoods in Pacific Northwest, Pacific Southwest, and
Rocky Mountain Regions
A.l Pacific Northwest and Pacific Southwest
As mentioned in Chapter 3, to estimate changes in yields of softwood forests in the West, we
first had to estimate changes in habitable area as elevation shifts. Fred Swanson (Forestry
Science Laboratory, Corvallis, Oregon) provided a profile of a typical, 7000 feet high
mountain in the western Cascades We assumed this profile was representative of mountains
in the Pacific Northwest and Pacific Southwest. We further assumed that softwood forests,
and Douglas fir in particular, are currently found at 500 to 1000 meter elevations (Urban et
al., 1993). Using shifts in elevation zones reported by Urban et al. and King and Tingey, we
assume that a 2.5°C warming would shift elevation zones up 500 meters and a 4°C warming
would shift elevation zones 1000 meters upslope. Using the mountain profile provided by
Swanson, we calculated the change in area at these elevations. The area between 1000 and
1500 meters is 60% smaller than the area between 500 and 1000 meters, while the area
between 1500 and 2000 meters is 83% smaller than the area between 500 and 1000 meters
We then developed estimates of how yields at different elevations would change As stated
above, we used Urban et al.'s results for current elevations For the 2.5°C scenarios We
assumed that yields in the 1000 to 1500 meter zone are the same as in the 500 to 1000 meter
zone under current climate For the 4°C scenario, we assumed that yields in the 1500 to 2000
meter zone are the same as in the 500 to 1000 meter zone under current climate and yields in
the 1000 to 1500 meter zone are equal to yields in the 500 to 1000 meter zone under the
2 5°C scenarios. Adjustments for C02 fertilization were made in the manner described in
Chapter 2
For the 2 5°C high scenario, we assumed:
~ At 500-1000 m, 50% yield decline according to OSU
~ At 1000-1500 m, yields same as currently at 500-1000 m.
For the 2 5°C low scenario, we adjusted yield estimates to account for C02:
* At 500-1000 m, 30% decline in yields
~ At 1000-1500 m, 25% increase in yields.
RCG/Hagler Bailly
Final Report
-------
Appendix A ~ A-2
For the 4°C high scenario, we assumed:
*¦ At 500-1000 m, 90% yield decline according to GISS
~ At 1000-1500 m, yields decline according to 2.5°C high scenario (50%)
~ At 1500-2000 m, yields same as currently at 500-1000 m.
For the 4°C low scenario, we adjusted yield estimates to account for C02:
~ At 500-1000 m, 80% decrease in yields
~ At 1000-1500 m, 33% decrease in yields
~ At 1500-2000 m, 25% increase in yields.
A.2 Rocky Mountains
For this region, we assumed the same shifts in elevation as were assumed for Pacific
Northwest. The procedure for estimating changes in yields is similar to that used in the
Pacific Northwest/Southwest:
For the 2.5°C high scenario, we assumed:
~ At 500-1000 m, 10% yield decline according to OSU
~ At 1000-1500 m, yields same as currently at 500-1000 m.
For the 2 5°C low scenario, we adjusted yield estimates to account for C02:
~ At 500-1000 m, 20% yield increase
~ At 1000-1500 m, 25% yield increase.
For the 4°C high scenario, we assumed:
»• At 500-1000 m, 20% yield decline according to GISS
~ At 1000-1500 m, 10% yield decline similar to 500-1000 m under 2.5°C high
scenario
~ At 1500-2000 m, yields same as currently at 500-1000 m.
RCG/Hagler Bailly
Final Report
-------
Appendix A »• A-3
For the 4°C low scenario, we adjusted yield estimates to account for C02:
~ At 500-1000 m, 10% increase
~ At 1000-1500 m, 20% increase (similar to 500-1000 m under 2.5°C low
scenario)
~ At 1500-2000 m, 25% increase.
RCG/Hagler Bailly
Final Report
-------
Appendix B
-------
Appendix B
Tables
This appendix contains the tabular information from the Figures in Chapter 6. For
consistency, the titles of the tables in this appendix refer to their corresponding figure in
Chapter 6.
RCG/Hagler Bailly
Final Report
-------
Appendix B ~ B-2
Table B
-1 (Figures 6
-1 thru 6-3)
Future Value of Producer Surplus by Region
($ Millions)
REGION
BASE
2.5 LOW
2.5 HIGH
4.0 LOW
4.0 HIGH
2000
North
7,125
8,462
13,071
8,129
16,762
South
15,481
6,804
18,233
5,271
16,488
West
13.058
56.757
76.276
77.265
80,387
2030
North
6,758
15,867
33,776
32,522
47,528
South
22,639
81,172
89,517
95,902
47,295
West
21,013
73.465
102.004
84,541
95,012
2060
North
4,382
26,135
41,178
33,853
61,522
South
24,412
110,476
109,292
121,881
42,179
West
9.212
66.020
91.715
73.149
99.614
RCG/Hagler Bailly
Final Report
-------
Appendix B ~ B-3
Table B
-2 (Figures 6-
4 thru 6-6)
Tornquist-
-Theil Division Price Indices for Product Prices, by Species
BASE
2.5 LOW 2.5 HIGH
4.0 LOW
4.0 HIGH
Combined Hardwood & Softwood
1990
1
1.08
1.22
1.1
1.12
2000
1.05
1.32
1.71
1.64
2.11
2010
1.03
1.46
1.99
1.85
2.44
2020
1.02
1.62
2.25
2.08
2.77
2030
0.99
1.71
2.44
2.24
3.03
2040
0.97
1.79
2.53
2.33
3.14
2050
0.98
1.88
2.63
2.40
3.16
2060
0.98
1.97
2.76
2.50
3.31
Softwood
1990
1.00
1.15
1.28
1.15
1.14
2000
1.06
1.48
2.01
1.96
2.60
2010
1.03
1.69
2.42
2.25
3.14
2020
1.01
1.92
2.79
2.57
3.67
2030
0.97
2.07
3.01
2.79
4.05
2040
0.95
2.13
3.10
2.87
4.13
2050
0.95
2.27
3.16
2.94
3.98
2060
0.97
2.36
3.28
3.06
4.10
Hardwood
1990
1.00
0.86
1.04
0.92
1.07
2000
1.00
0.84
0.94
0.81
0.94
2010
1.01
0.83
0.91
0.83
0.93
2020
1.03
0.80
0.93
0.82
0.96
2030
1.07
0.80
1.00
0.83
1.03
2040
1.05
0.81
1.12
0.94
1.21
2050
1.05
0.81
1.29
1.02
1.45
2060
1.03
0.88
1.49
1.09
1.66
RCG/Hagler Bailly
Final Report
-------
YEAR
1990
2000
2010
2020
2030
2040
2050
2060
1990
2000
2010
2020
2030
2040
2050
2060
1990
2000
2010
2020
2030
2040
2050
2060
HIGH
06458
59073
53987
53332
53976
50616
55938
50036
55229
69282
72376
78186
78941
76857
74002
73443
161687
128355
I26363
131518
132917
I27473
I29940
I23479I
APPENDIX B ~ B-4
Table B-3 (Figures 6-7 thru 6-9)
Forest Products Production (MMCF) by Species
BASE
2.5 LOW
2.5 HIGH
4.0 LOW
Softwood
109260
109688
122080
136014
148727
149260
149301
148346
105619
95574
98290
104086
112901
111931
106725
103744
102017
77413
73476
77100
82232
79829
78280
73880
106095
79129
80623
85317
89951
87695
86171
81901
Hardwood
60181
60104
56967
58393
66012
69490
67938
70328
70093
74664
72050
73703
76096
82961
77693
78946
77211
84103
78445
80113
78587
81565
77140
79832
80127
83412
75791
78550
80029
83111
73552
78686
Total
169441
175700
192173
212110
225938
227847
229428
228375
165723
165064
172954
187047
197004
193496
190137
186855
158984
145351
145526
154793
160677
156969
154071
147432
164488
149457
154326
164263
170064
167527
164721
160587
RCG/Hagler Bailly
Final Report
-------
Appendix B * B-5
Table B-4 (Figures 6—10 thru 6-12)
Inventory Volume (MMCF) by Species
YEAR
BASE
2.5 LOW
2.5 HIGH
4.0 LOW
4.0 HIGH
Softwood
1990
193802
193802
193802
193802
193802
2000
176961
175359
141457
139992
98079
2010
198259
176676
130740
135381
85229
2020
207459
175798
130986
133813
81406
2030
225762
178395
131789
134009
79492
2040
221907
172595
123785
129951
76235
2050
227360
164664
119234
125225
76695
2060
242332
170463
117410
126926
71673
Hardwood
1990
246249
246249
246249
246249
246249
2000
268496
313822
255871
287064
231308
2010
272085
323559
260951
289291
229485
2020
278372
322932
255994
281690
225456
2030
267467
305407
239168
262180
215893
2040
258065
267257
223272
247109
204939
2050
244313
251500
202080
231876
191786
2060
232465
230867
184340
219330
183871
Total
1990
440051
440051
440051
440051
440051
2000
445457
489181
397328
427056
329387
2010
470344
500235
391691
424672
314714
2020
485831
498730
386980
415503
306862
2030
493229
483802
370957
396189
295385
2040
479972
439852
347057
377060
281174
2050
471673
416164
321314
357101
268481
2060
474797
401330
301750
346256
255544
RCG/Hagler Bailly
Final Report
-------
Appendix B ~ B-6
Table B—5 (Figures 6—13 thru 6-16)
Forest Product Prices for Hardwood-Softwood Pulpwood Experiment
PRODUCT YEAR BASE 2.S LOW 2.5 HIGH 4.0 LOW 4.0 HIGH REV. BASE REV. 4.0 HIGH
Softwood
Fuelwood
1990
0.01
0.01
0.01
0.01
0.01
0.01
044
Fuelwood
2000
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Fuelwood
2010
0.01
0.01
0.01
0.01
0.3
0.01
0.01
Fuelwood
2020
0.01
0.01
0.08
0.09
0.61
0.01
0.01
Fuelwood
2030
0.01
0.01
0.01
0.06
0.82
0.01
0.01
Fuelwood
2040
0.01
0.01
0.01
0.01
0.51
0.01
0.01
Fuelwood
2050
0.01
0.01
0.01
0.01
0.51
0.01
0.01
Fuelwood
2060
0.01
0.03
0.01
0.01
0.01
0.01
0.01
Pulpwood
1990
1.73
1.56
1.56
1.35
1.4
127
1.27
Pulpwood
2000
1.77
2.53
3.39
3.32
3.95
1.28
1.7
Pulpwood
2010
1.14
2.73
3.99
3.39
4.37
1.16
1.89
Pulpwood
2020
1.23
3.22
4.42
3.79
5.09
1.25
2.14
Pulpwood
2030
1.68
3.51
5.06
4.62
6.06
1.39
2.51
Pulpwood
2040
1.65
3.6
5.2
4.73
6.21
1.42
2.98
Pulpwood
2050
1.57
3.83
5.16
4.61
5.9
1.48
3.36
Pulpwood
2060
1.61
3.86
5.35
4.77
6.16
1.53
3.58
Sawtimber
1990
1.63
2.02
2.32
2.12
2.07
1.68
2.03
Sawtimber
2000
1.76
2.43
3.29
322
4.4
1.78
4.48
Sawtimber
2010
1.93
2.82
4
3.83
5.38
1.9
5.39
Sawtimber
2020
1.87
3.17
4.67
4.42
6.28
1.84
6.31
Sawtimber
2030
1.58
3.41
4.96
4.61
6.72
1.73
6.64
Sawtimber
2040
1.55
3.5
5.1
4.76
6.92
1.68
6.63
Sawtimber
2050
1.59
3.73
5.27
4.98
6.71
1.69
6.68
Sawtimber
2060
1.61
3.94
5.45
5.18
7
1.71
6.81
Hardwood
Fuelwood
1990
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Fuelwood
2000
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Fuelwood
2010
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Fuelwood
2020
0.01
0.01
0 01
0.01
0.01
0.01
0.01
Fuelwood
2030
0.01
0 01
0.01
0.01
0.01
0.01
0.01
Fuelwood
2040
0.01
0.01
0 01
0.01
0.01
0.01
0.01
Fuelwood
2050
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Fuelwood
2060
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Pulpwood
1990
1.09
0.95
1.13
1.01
1.16
1.17
1.31
Pulpwood
2000
1.09
0.92
1.03
0.89
1.03
1.18
1.6
Pulpwood
2010
1.1
0.91
1
0.91
1.02
1.19
1.79
Pulpwood
2020
1.12
0 88
1.02
0.91
1.05
1.23
2.04
Pulpwood
2030
1.16
0.88
1.09
032
1.12
1.29
2.41
Pulpwood
2040
1.14
0.9
1.21
1.03
129
1.32
2.88
Pulpwood
2050
1.14
0.89
1.39
1.11
1.43
1.38
3.26
Pulpwood
2060
1.12
0.97
1.6
1.18
1.4
1.45
3.48
Sawtimber
1990
0.99
0.85
1.03
0.91
1.06
1.07
1.21
Sawtimber
2000
0.99
0.82
0.93
0.79
0.93
1.08
1.5
Sawtimber
2010
1
0.81
0.9
0.81
0.92
1.09
1.69
Sawtimber
2020
1.02
0.78
0.92
0.81
0.95
1.13
1.94
Sawtimber
2030
1.06
0.78
0.99
0.82
1.02
1.19
2.31
Sawtimber
2040
1.04
0.6
1.11
0.93
1.23
1.22
2.78
Sawtimber
2050
1.04
0.79
1.29
1.01
157
1.28
3.2
Sawtimber
2060
1.02
0.87
1.5
1.08
1.99
1.35
3.59
RCG/Hagler Bailly
Final Report
-------
Appendix B ~ B-7
Table B-6 (Figure 6-17)
Tomquist-
Theil Division Price Indices
For Hardwood-
Softwood Pulpwood Experiment
YEAR
BASE
4.0 HIGH REV. BASE
REV. 4.0 HIGH
1990
1.00
1.12
0.98
1.14
2000
1.05
2.11
1.02
2.04
2010
1.03
2.44
1.04
2.38
2020
1.02
2.77
1.04
2.69
2030
0.99
3.03
1.04
2.95
2040
0.97
3.14
1.03
3.17
2050
0.98
3.16
1.05
3.36
2060
0.98
3.31
1.08
3.53
RCG/Hagler Bailly
Final Report
-------
Appendix C
-------
Appendix C
Harvest and Management Decision Model
The purpose of this appendix is to present the equations that constitute the forest sector
portion of FASOM. When this sector is run in a stand-alone capacity, the total amount of
land in the forest inventory in a given year is fixed and net-migration of land is treated as
exogenous. Otherwise, the sector is modeled in a manner exactly identical to that in the dual
sector version of FASOM.
Subscripts (owner, land suitability class, and site class are omitted to reduce the complexity
of notation)-
r region
t projection period (1, 2, ..., T-l, T) measured in decades
c age (cohort) of an aggregate in existence at start of problem (1, 2, . ., N-l
where N-l is the oldest recognized age class measured in decades)
h age of aggregate at harvest (1, 2, ..., N-l, N where N indicates "never
harvested" or not harvested during projection period)
d date of aggregate harvest (decade midpoints 1,2, ., T and N as defined above)
m management intensity class (MIC = high, medium, low, and no or "passive"
management)
s species (order) class (softwood followed by softwood, softwood followed by
hardwood, hardwood followed by hardwood, hardwood followed by softwood)
g product (softwood and hardwood sawtimber, pulp wood, and fuelwood)
Variables (activities).
Xt,c,(Mm acres harvested of an existing aggregate in region r, from cohort c, d periods
after start of problem, in MIC m, and species s
Nr,t,m. acres regenerated in region r, period t, and harvested at age h, from MIC m and
species s
RCG/Hagler Bailly
Final Report
-------
Appendix C ~ C-2
total volume harvested of product g in region r, period t
volume of product g consumed domestically in period t
volume of product g exported from region r in period t
volume of product g imported to region r, period t
volume of product gl "downgraded" to the next product class in region r,
period t (sawtdmber can substitute for pulp wood and pulpwood can substitute
for fuelwood within a species and softwood fuelwood can substitute for
hardwood fuelwood)
volume of log processing (consumption) capacity added for product g, region r,
period t
terminal (perpetual) harvest volume of product g in region r (beginning at the
end of period T)
Exogenous variables and functions:
DT^ per unit volume domestic transport costs for product g from region r, period t
PCrrn,t planting cost per acre in region r for MIC m and species class s, period t
TT^ trade transport costs for product g in region r, period t
XYgj.c+t-i.m^ Per acre yields of existing aggregates of product g, in region r, at age c+t-1,
MIC class m, and species class s
NYert,m« per acre yields of aggregates originating during the projection period of product
g, in region r, at age h, MIC m, and species class s
G^r^s.t Per acre aggregate growing (or tending) costs between origination and harvest
in region r, MIC m, species s, in period t
KCg^t unit capacity costs for product g, region r, period t
PD&„ PE^ PMg^., national domestic demand (PD), regional export demand (PE), and
regional import supply (PM) functions solved to give own price as a
function of quantity for product g, region r, and period t
RCG/Hagler Bailly
Final Report
H^t
DDfrt
DM^,
Ugi,,t
TH^
-------
Appendix C ~ C-3
PN&t sum of domestic and all regional export demand functions solved to give own
price as a function of quantity for product g and period t
UCg^ unit cost of substituting (downgrading) product g to the next "lower" product
class in region r
R^. approximate optimal rotation for aggregates in region r, MIC m, and species s
(derived from harvest ages observed in model projections)
NLC,„.t net timberland area change due to shift to or from other uses in region r, MIC
m, species s, in period t
HCg^ harvesting cost per unit volume removed of product g, in region r, period t
starting inventory of area in existing aggregates in region r, cohort c, MIC m,
and species s
5g rate of capacity depreciation for product g
i discount rate
, volume of log processing capacity for product g, region r, at start of projection
period
HG^ harvest from public lands of product g, region r, period t
Equations
Objective function maximize (for the set of activities shown above)
r DD»
E <£ / «>,, W* - Y.dts„ (h,„
f-1 g 0 T
EEE
m m h g r
-EE(
g r
RCG/Hagler Bailly
Final Report
-------
Appendix C ~ C-4
EEEt p a ~ o-"" -
-------
Appendix C ~ C-5
Demand-supply balance constraints:
DD„ - E <»,,, " DEw ~ DM,,, ) + ¦£ W, " E W, (C^>
r r r
for all g and where gl(g) is the index for the product that can be substituted for g and g2(g)
is the index of the product for which g can be substituted.
Regional timber harvest:
// , = Y,Y(Y. xrrt_. xy + y. NrB, NYort ) + HGsrJ (C-5)
if,t ^ Tff-ppi* gjf-pjrL? ' gjrj
mm e p*t
for all g, r, and t.
Capacity constraints:
DD.» * k*,/1 -«,)'»Ed (c-6)
k-1
for all g, r, and t
Terminal (perpetual) log supply volume:
™t,"2 EE —
^2 nj + 52 Nrj>J-p^f ^gsJ-pjnj )
P*T
SS
+ HGg/j + DMgjj (C-7)
for all g and r.
RCG/Hagler Bailly
F/wa/ Report
------- |