United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600^-93/093
May 1993
f/EPA
The Forest Sector Carbon
Budget of the United States:
Carbon Pools and Flux Under
Alternative Policy Options
-------
THE FOREST SECTOR CARBON BUDGET OF THE UNITED
STATES: CARBON POOLS AND FLUX UNDER ALTERNATIVE
POLICY OPTIONS
Edited by
David P. Turner
ManTech Environmental Technology, Inc.
US EPA Environmental Research Laboratory - Corvallis
Jeffrey J. Lee
US Environmental Protection Agency
US EPA Environmental Research Laboratory - Corvallis
Greg J. Koerper
ManTech Environmental Technology, Inc.
US EPA Environmental Research Laboratory - Corvallis
Jerry R. Barker
ManTech Environmental Technology, Inc.
US EPA Environmental Research Laboratory - Corvallis
Contributors
R. J. Alig, J. R. Barker, H. Gucinski, M. E. Harmon, R. W. Haynes, J. S.
Kern, G. A. King, G. J. Koerper, J. J. Lee, C. E. Peterson, P. E.
Schroeder, D. P. Turner
May 1993
EPA/600/A-93/093
US Environmental Protection Agency
Environmental Research Laboratory
200 Southwest 35th Street
Corvallis, Oregon 97333 USA
-------
DISCLAIMER
The research described in this report has been funded by the US Environmental
Protection Agency. It has been subjected to the Agency’s peer and administrative
,eview and it has been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
Table of Contents
Acknowledgments . . iii
Tables iv
Figures vii
Boxes ix
Unit Conversions sj....
Executive Summary xii
SECTION 1. INTRODUCTION 1
A. Scope and objectives 1
B. Overview of modeling approach 4
C. Overview of this report 8
SECTION 2. BACKGROUND: THE TERRESTRIAL CARBON BUDGET 11
A. Introduction 11
B. The global carbon cycle 11
C. Regional carbon budgets 14
D. Vegetation and the global carbon cycle 16
E. Mitigation of rising carbon dioxide levels 17
F. Forest management concerns, constraints, and approaches 17
G. Conclusions 19
SECTION 3. METHODS 21
A. Base year carbon pools 21
1. Construction of stand-level carbon budgets 21
2. Forest inventory databases 27
B. Base year carbon flux 30
1. Biologically-driven carbon flux 30
2. Harvest-driven carbon flux 30
3. Other changes in carbon pools 31
C. Projected US forest sector carbon pools and flux 31
1. Use of the TAMM and ATLAS models for private timberlands 32
2. Methods used in the analysis and treatment of public lands 37
3. Fate and disposition of carbon after timber harvest 39
4. The forest-sector scenarios 40
Scenario 1. The 1989 RPA Assessment Projection 40
Scenario 2. Current Forest Plans 41
Scenario 3. Reduced National Forest Harvest 51
Scenarios 4 and 5. Moderate and High Increased Recycling 52
Scenarios 6, 7, 8, 9. Afforestation 53
Scenario 10, 11, 12. Combinations 55
-------
SECTION 4. RESULTS AND DISCUSSION 57
A. Current pools 57
B. Current flux 57
1. Base year biologically driven carbon flux 57
2. Base year harvest-driven carbon flux 65
3. Complete base year flux analysis 68
C. Forest policy scenarios 70
1. The 1989 RPA Assessment scenario 70
2. The Current Forest Plans scenario 76
3. Reduced National Forest harvest scenario 78
4. Increased Recycling scenarios 81
5. Afforestation scenarios 81
6. Combination scenarios 84
7. Scenario comparisons 84
8. Results of HARVCARB analysis 91
D. Discussion 94
E. Presentation of results for all scenarios 96
SECTION 5. POTENTIAL IMPACTS OF CLIM A!E CHANGE 107
SECTION 6. FUTURE RESEARCH DIRECTIONS 115
A. Refine the existing modeling framework and current inventory 115
B. Integrate national-level studies with global scale analyses 116
C. Evaluate the potential effects of climate change 116
SECTION 7. CONCLUSIONS .. 119
SECTION 8. LITERATURE CITED
SECTION 9. SUPPORTING DOCUMENTS ... 133
A. Uncertainty in Country-wide Forest Biomass Estimates 133
B. Biomass and Carbon for Woodlands in the. Wës rn US 143
C. Woody Debris Budgets for Selected Forest Types in the US 151
D. Net Primary Productivity of Non-forested Lands 179
E. Glossary 195
F. Metric and English Equivalents .? 201
G. Acronyms ..— 202
II
-------
Acknowledgments
The editors gratefully acknowledge all persons wriose contributions led to the
completion of this document. This especially includes the following:
1.0 Writers, Database Design, and Technical Advisors:
P. Alig, USDA Forest Service
B. Baker. Computer Services Corporation
J. Barker. ManTech Environmental Technology. Inc.
R. Birdsey, USDA Forest Service
G. Baumgardner, ManTech Environmental Technology. Inc.
G. Carroll, USDA Forest Service
D. Chojnacky, USDA Forest Service
H. Gucinski 1 , ManTech Environmental Technology, Inc.
M. Harmon, Oregon State University
R. Haynes, USDA Forest Service
D. Jacobson. Computer Services Corporation
J. Kincaid. University of Washington.
G. Koerper, ManTech Environmental Technology, Inc.
J. Lane, Lane Technical Writiç g and Editing Service
J. Lee, US Environmental Protection Agónc ’y
B. Mead, USDA Forest Service
J. Mills, USDA Forest Service
E. Moore, USDA Forest Service
C. Peterson 2 , ManTech Environmental Technology, Inc.
D. Turner, ManTech Environmental Technology. Inc.
D. Van Hooser, USDA Forest Service
K. Waddell. USDA Forest Service
S. Woudenberg, USDA Forest Service
2.0 DocumenfPrdduction:
G. Baumgardner, ManTech.Environmental Technology, Inc.
.T. Miller, ManTecl Ej,yironrnental Technplogy, Inc.
B. Rosenbaum, ManTech Environmental,Technology, Inc.
3.0 Reviewers:
Internal
D. Coffey, ManTecti Environmental Technology, Inc.
A. Dixon, US Environmental •Protection Agency
J. Weber, US Environmental Protection Agency
External
L. Heath. USDA Forest Service
E. Hughes. Electric Power Research Institute
W. Kurz, Environmental and Social Systems Analysts
A. LeBlanc, Environmental Defense Fund
P. Maclaren, Forest Research Institute
J. Perez-Garcia, University of Washington
N. Trexler, Trexier and Associates
t Currently with the USDA Forest Service, Forestry Sciences Laboratory, Corvallis, OR.
2 Currently with the USDA Forest Service, Pacitic NW Research Station, Portland, OR.
-------
Tab’es
Table 1.1; Specifications for the policy options. See Table 1 (Executive
Summary) and the text for additional characterization 6
Table 3.1. Factors to convert tree volume (cubic feet) to carbon (kg).
Conversion factors from Birdsey (1992a) 23
Table 3.2. Projected harvests (as growing stock) from National Forests for
RPA, Current Forest Plans, and Reduced National Forest Harvest
Scenarios 42
Table 3.3. Lumber shares (%) by region for 1952, 1992, and projected 43
Table 3.4. Deflated price indexes for selèctedtimber products in the US, by
softwoods and hardwoods 48
Ta, le 3.5. Harvest and inventory projections by region and public/private
timberland owners 50
Table 3.6. Projected percent of total fiber froni recycled wastepaper in RPA,
Moderate Recycling, and High Recycling scenarios 54
Table 3.7. Projected consumption of sawlogs and pulpwood in RPA and
Moderate Recycling Scenarios 54
Table 3 8. Enrollment schedules of forestation of marginal lands b region for
two funding levels and two studies 55
Table 4.1. Base year timberland carbon budget 68
Table 4.2. . Summary of policy scenario results . . 90
Table 4.3. Cumulative disposition of carbon (Tg-C) for the Current Forest
Plans scenano 92
Table 4.4 Difference in carbon disposition (T C) between th RPA scenario
,an ttie Current Forest Plans scer ario
Table 4.5. Difference in carbon disposition (Tg-C) between the High
Recycling scenario and the Current Forest Plans scenario 93
Table 4.6. The 50-year average change in flux to product, landfill, and energy-
offset pools for the four alternative policy scenarios, compared to
the Current Forest Plans scenario. A plus (+) sign means an
additional carbon sink 93
Table 4.7. Undiscounted and discounted carbon benefit for eleven forest
policy scenarios 98
iv
-------
Table 5.1. Summary of predicted changes in global mean temperature and
mean annual precipitation as simulated by four General Circulation
Models 108
Table 5.2. Percent future forest area relative to current forest area using
several different vegetation models and climate scenarios. BIOME
and Holdridge results reflect global change 110
Table 9.B.1. Computation of carbon for woodlands in the western United States 144
Table 9.C.1. Regression statistics for the relationship between diameter at
breast height, stump height, and stump diameter 152
Table 9.C.2. Rate constants of dead wood input for US forest types based upon
the ratio of growing stock mortality to growing stock volume 159
Table 9.C.3. The proportion of live tree car on left on site after timber harvest as
woody detritus for US forest types 162
Table 9.C.4. Decomposition rate constants for US forest types as estimated from
the literature 163
Table 9.C.5. Estimated steady-state stpres of woody detritus for US forests 166
Table 9.D.1. Regions of the United States as defined by USDA (1989) 180
Table 9.D.2. Summary table of nonforested land productivity 18 4
Table 9.D.3. Estimates of net primary productivity for rangelands, croplands and
pastlirelands by region 184
Table 9.D.4. Estimate of net primary productivity for federal and non-federal
rangelands in the conterminous US 190
Table 9.D.5. Estimate of’ net primary productivity for croplands in the
conterminous US :190
Table 9.D.6. Estimate of net primary productivity for pasturelands in the
contermipous United States 1 . :1
Table ’9.D.7. Estimate’ ‘of net primary productivity for wetlands in the
conterminous US ......._ 19t.
V
-------
Figures
Figure 1.1. The global carbon cycle, including major pools and annual flux of
carbon 2
Figure 1.2. Modeling framework for forecasting carbon pools 7
Figure 3.1. Examples of stand-level carbon budgets for three forest types
using the FCM 26
Figure 3.2. Definitions of regions used in assessment of the US forest sector
carbon budget 28
Figure 3.3. Base year distribution of available and reserved timberland by
region and ownership 29
Figure 3.4. General structure of TAMM 33
Figure 3.5. Lumber and structural panel products consumption, 1950-1988
with projections to 2040 44
Figure 3.6. Softwood sawtimber stumpage prices from years 1952 to 1986,
with projections through 1990 to 2040 47
Fig’ure 4.1. Relative contribution of carbon pools to total carbon for US tim-
berland 58
Figure’4.2. Base year carbon pools by region, ownership, and component for
UStimberland 59
Figure 4.3. Base year carbon storage per unit area by region and component 62
Figure 4.4. Net ecosystem productivity of selected forest types 63
Figure 4.5. Base year biologically-driven carbonfkix by region, ownership,
and component for US timberland 64
Figure 4.6. Base year biologically-driven carbon flux per unit area by region
and component ..- 66
Figure 4.7. Base year growing stock reductions due to harvesting by own- -
ership and component 67
Figure 4.8. Projections of net carbon accumulation or lossfor US timberlands
(sum of available and reserved) for the RPA Assessment
scenario 71
Figure 4.9. Private timberland land base 72
Figure 4.10. Projected annual growing stock volume reductions from US
timberlands for the RPA Assessment scenario 73
vi
-------
Figure 4.11. Projections of carbon pools on available public lands for the
PNWW, PNWE and PSW regions for the RPA Assessment
scenario 75
Figure 4.12. Projections of net carbon accumulation or loss for US timberlands
(sum of available and reserved) for the Current Forest Plans
scenario 77
Figure 4.13. Projected annual growing stock volume reductions from US
timberlands under Current Plans scenario 79
Figure 4.14. Change in carbon transfer to atmosphere over time from areas not
harvested in the Pacific Northwest West Region 80
Figure 4.15. Annual carbon accumulation for recycling scenarios relative to
Current Forest Plans projections 82
Figure 4.16. Comparison of projected total forest carbon pool for the High and
Moderate Recycling and Current Forest Plan scenarios 83
Figure 4.17. Annual carbon accumulation difference for afforestation scenarios
relative to Current Forest Plans projections .84
Figure 4.18. Comparison of projected total forest carbon pool for afforestation
scenarios 86
Figure 4.19. Annual carbon accumulation difference for combination scenarios
relative to Current Forest Plans projections
Figure 4.20. Comparison of projected total forest carbon pool for the combi-
nation scenarios and the Current Forest Plans scenario ,88
Figure 4.21. Scenario comparison based on projected total forest carbon pool 97
Figure 4.22. Scenario comparison based on difference from the Current Forest
Plans scenario in projected total forest carbon pool 98
Figure 4.23. Eflects of assumed discount rate on the carbon sequestration•
benefit 100
Figure 4.24. Scenario comparison based on the difference from the Current
Forest Plans scenario in carbon accumulation rate .... 101
Figure 4.25 Scenario comparison based on carbon accumulation rate ...1 02.
Figure 4.26. Resutts for all scenarios in the total forest carbon pool format 103
Figure 4.27. Resutts for all scenarios in terms of differences from the Current
Forest Plans scenario in the total forest carbon pool 104
v i ’
-------
Figure 4.28. Results of all scenarios in terms of the differences from the Current
Forest Plans scenario in carbon accumulation rate 105
Figure 4.29. Results for all scenarios in terms of the carbon accumulation rate 106
Figure 9.A.1. Percent error associated with state level and sub-state level
timberland volume estimates the Eastern US 136
Figure 9.A.2. Percent error associated with state level and sub-state level
timberland volume estimates in the Western US 137
Figure 9.A.3. Percent error associated with timberland volume estimates
within astate 138
Figure 9.A.4. Percent error associated with timberland area estimates stratified
by ownership and type within a state 139
Figure 9.C.1. Observed versus estimated steady-state mass of woody detritus
for US mature to old-growth forests 160
Figure 9.C.2. Carbon increment associated with tree growth and woody
detritus for Pacific northwest Douglas-fir forests 165
Figure 9.C.3. Observed versus estimated mass inputs of woody detritus for US
mature to old-growth forests 170
Figure 9.C.4. Carbon increment associated with tree growth and woody
detritus for southeastern bottom land hardwood forests 171
Figure 9.D.1. Relative distribution of nonforested land in the conterminous
United States, summarized by cover type 185
Figure 9.D.2. Relative distribution of nonforest net primary production,
summarized by cover type 186
Figure 9.D.3. Relative distribution of rangeland, cropland, and pastureland
area in the United States, summarized by region 187
Figure 9.D.4. Relative distribution of net primary production of rangeland,
cropland, and pastureland in the United States, summarized by
region 189
vi”
-------
Boxes
Box 1. UNCED Forestry Principles .3
Box 2. What is a Forest-sector Carbon Budget’) 5
Box 3. Changes in the Size of the US Forest Land Base 8
Box 4. Global View of Carbon Conservation/Sequestration 12
Box 5. Defin ion of Timberland and Woodland 22
Box 6. Literature-based Estimate of Timberland and Woodland Biomass Pools 60
Box 7. Net Accumulation in Product and Landfill Pools 69
ix
-------
Unit Conversions
Documents treating the relationship of policy issues to the global carbon cycle
typically use units of megatons (Mt) and gigatons (Gt). We have used these units in
the executive summary but throughout the remainder’of this report we use primarily
SI units. Some of the relevant conversions are as follows:
Mt = megaton = million metric tons = 1012 g = teragram = Tg,
Gt = gigaton = billion metric tons = 1 1 5 g = petagram = Pg.
The pools and flux of biomass or carbon dioxide are expressed as carbon
throughout the report.
x
-------
Executive Summary
Rising levels of atmospheric 002 and
other radiatively important trace gases
(e.g., CH4, N20) from anthropogenic
sources are likely to induce changes in
the earth’s climate over the coming
decades (Houghton et al. 1991). While
there remains considerable uncertainty
about the rate and magnitude of the
possible climate change, there is an
emerging consensus that policies rele-
vant to stabilizing or reducing the
emission of 002 and other “green-
house gases” should be explored (NAS
1991, Rubin et al. 1992).
Terrestrial ecosystems have a major
role in the global cycling of carbon.
Vegetation and soil contain about three
times as much carbon as is in the atmo-
sphere. Also, each year, they exchange
about 20 times as much carbon with the
atmosphere as is emitted from fossil
fuel use. For this reason, one proposed
approach for slowing the increase in
atmospheric C02 is to manage terres-
trial ecosystems to conserve and
sequester carbon by changing the bal-
ance between the uptake of carbon
dioxide from the atmosphere through
photosynthesis and the release of car-
bon dioxide through respiration and
biomass burning (Schroeder and Ladd
1991; Sedjo and Solomon 1989;
Sampson and Hair, in press). Forest
ecosystems are particularly important in
consideration of carbon conservation
and sequestration opportunities
because of their large accumulations of
woody biomass.
The utility of managing terrestrial
ecosystems to conserve or sequester
atmospheric carbon must be measured
against the policy objectives for altering
carbon emissions. Carbon emissions
from the use of fossil fuels in the US
were about 1,300 Mt-C per year (Mt =
million metric ton = 1012 g) in 1988
and, from 1980 to 1988, emissions
increased at an average rate of 6.4
Mt-C per year (NAS 1991), with the
largest increase in the later years.
Emission increases were greater
toward the end of this period, reflecting
increased economic activity. Thus, a
policy objective of stabilizing carbon
emissions would require emissions
reductions, or implementation of
strategies for conserving/sequestering
carbon, amounting to an additional 6 to
10 Mt-C per year each year. A policy
objective of reducing emissions by to
to 20% would require emission reduc-
tions or increased carbon conserva-
tion/sequestration amounting to 130 to
260 Mt-C per year.
This document presents a model of the
current and future carbon budget asso-
ciated with the forest ecosystems of the
conterminous US. The focus is on
effects of economic and environmental
policy changes for the period of 1990-
2040. In this study, the concept of “for-
est ecosystem” has been expanded to
“forest sector” by including biomass that
has been physically removed
(harvested) for human use. The poten-
tial effects of climate change have not
been incorporated; however, current
research is directed at addressing this
question. The specific objectives of this
report are to:
1. Develop a methodology for quantify-
ing the current and future forest
sector carbon budget for the US.
2. Apply the methodology to quantify
the current status of carbon pools
xi
-------
Executive Summary
and flux within the US forest sector.
3. Apply the methodology to several
scenarios based on alternative pol-
icy options.
4. Identify gaps and
data and in
approaches that
effort.
The specific policy options considered
in the report were assembled by the
EPA Office of Policy, Planning, and
Evaluation (OPPE) (Table 1). They
function as a preliminary look at some
of the types of options available. They
do not represent the full spectrum of
options that could be considered (NAS
1991, Rubin et al. 1992). Furthermore,
they may not necessarily be the “best’
options. The intent was to develop a
tool that can be used to give broad
technical guidance in evaluating vari-
ous policy options for conserving or
sequestering carbon in the forest sec-
to r.
The basic approach to modeling the
current pools and net flux of carbon in
US forests has been to develop a
database which links inventory data on
volume and area of different forest
types by age class, with stand-level
carbon budgets which quantify the
pools of carbon associated with each
age class within those forest types.
Changes in carbon storage in response
to various forest management policy
scenarios (Table 1) were then evalu-
ated by examining changes in forest
inventories from 1990 to 2040, as well
as changes in the post-harvest product
and landfill pools. The HARVCARB
model (Row and Phelps 1990) was
used to estimate changes in the amount
of carbon in forest products and in
landfills from material harvested from
1990 to 2040.
Changes in forest inventory on private
timberlands were estimated by linking
an economic model (TAMM), which
estimated tree harvest volume by
region (Haynes and Adams 1985), to a
forest inventory model (ATLAS), which
tracked forest age-class distribution by
forest type and productivity class (Mills
and Kincaid 1992). TAMM insures that
the economic consequences of imple-
menting the various policy options were
allowed to impact the carbon budget.
The inventory at any point in time was
run through a forest carbon model
(FCM) to estimate total forest carbon.
The change in total forest carbon over a
10-year interval, divided by 10, gave
the average annual flux for that interval.
For public lands, there are no stand-
alone inventory models. Furthermore,
inventory assumptions are embedded
in a variety of harvest planning models.
It was thus necessary to estimate future
carbon pools by using data from the
planning efforts at the level of individual
National Forests and other public
agencies.
Over half of the total timberland carbon
for the base year was in the mineral
soil. Tree carbon, which included
coarse roots, was the next largest com-
ponent at 31%, followed by woody
debris (11%), forest floor (6%), and
understory (1%). Total carbon storage
in living trees on timberland in the US
was estimated at 11.1 Gt-C (Gt = billion
metric tons = 1015 g). This quantity
conforms closely with the estimate
based on biomass factors derived from
Cost et al. (1990) and Koch (1989), and
to the estimates of Birdsey (1 992a). The
addition of woodlands carbon brought
the total for all forests in the US to 11.6
Gt-C.
For the base year, the net effect of the
biological and harvest-related transfers
deficiencies in the
the modeling
require further
XII
-------
Executive Summary
was a carbon accumulation of 91 Mt on
US timberland. The majority of the car-
bon accumulation was in the tree (52
Mt) and woody debris (36 Mt) pools.
The forest floor showed a small gain (5
M l) and the understory a small loss (-2
Mt).
Harvested removals amounted to 139
Mt. If the harvested carbon entered into
permanent storage, the forest product
pool would constitute an additional sink
of that amount for the base year.
However, determtnation of the net car-
bon transfer required consideration of
carbon emissions associated with the
decay or burning of these and previ-
ously produced forest products. Our
estimated of the net change in the
products pool was 36 Mt-C per year
gain.
Annual carbon emissions from wildfire
on US forests vary widely between
years. Based on examination of USDA
Forest Service statistics on fire fre-
quency and extent over the last several
decades (A. Auclair, Science & Policy
Associates, personal communication)
an estimate of 20 Tg-C was used in the
present analysis. Approximately one-
third of the total emissions are associ-
ated with woodlands, which tends to
maintain a long-term carbon equilib-
rium on those lands. Direct emissions
from burning of timberlands were thus
estimated at 13 Tg-C.
Summation of the carbon flux pro-
cesses described here suggests that
the forest sector in the conterminous
US was a sink for about 114
(91 +36-13) Mt-C per year of carbon for
the base year 1990. By way of compari-
son, carbon dioxide emissions (as car-
bon) from fossil fuel and cement manu-
facture in the 1990 was about 1,300
Mt-C per year. Thus the forest sector
offset about 8% of US carbon emis-
sions in 1990.
It is worth noting that, in principle, forest
products which are used to produce
energy may act to offset carbon emis-
sions from fossil fuel combustion in the
energy sector. This quantity has not
been specified in the current analysis.
However, the development of biomass
energy schemes offers a potential
opportunity for using the forest sector to
reduce carbon emissions.
The scenarios ranged from the Current
Forest Plans, which is the base sce-
nario and reflects current best estimates
of future harvest levels on public lands,
to scenarios that called for reduced
harvests on public lands, increased
recycling, or increased afforestation.
The 1989 Resource Planning Act (RPA)
Assessment scenario was included
because it represents the most recent
well documented scenario produced by
the USDA Forest Service. Differences
between the RPA and the Current
Forest Plans scenarios are a function of
policies already in place.
The outcomes of the scenarios associ-
ated with the various forest policy
options predict that US forests will con-
tinue to accumulate carbon during each
decade of the 50-year period (Table 2).
The rate of carbon accumulation will,
however, decline over time. Each sce-
nario predicts that the private forests, as
simulated by the TAMM economics
model coupled with the ATLAS timber
inventory projection model, will change
from accumulating carbon to losing
carbon between 2010 and 2040.
Contributing factors include rising har-
vest rates, shorter average rotation
lengths, and the conversion of the forest
land base to other uses such as urban-
ization. Less rigorous data and models
were available to project public forest
inventories. Both available and
x’u
-------
Executive Summary
reserved public timberlands appear to
act as carbon sinks throughout the
simulation period and more than bal-
ance carbon losses from private timber-
lands in later decades.
Though this trend was universal among
the scenarios, there were important dis-
tinctions among them which contributed
to differences in carbon pool totals and
in temporal variations in carbon flux.
The 1990 total carbon pool of US
forests was estimated to be 36.2 Gt.
Under the now obsolete 1989 RPA
scenario, this figure was predicted to
rise to 37.6 Gt in 2040. Increased car-
bon accumulation in public forests
accounted for all of this difference.
Although private timberlands had
increased total carbon at the midpoint
of the simulation, the total declined to
the 1990 level by the simulation’s end.
The Current Forest Plans marked a
significant departure from the RPA
scenario. A substantial portion of the
public forest land base was shifted from
the available category, i.e. forests sub-
ject to harvest, into long-term preserves.
As a result, carbon accumulation in
public forests increased, with carbon
uptake rates 7 to 14 Mt-C per year
higher than the RPA scenario over the
course of the simulation. Harvest rates
on the private forests increased slightly
in response to the change in public for-
est management policies. From 1990 to
2010, carbon uptake was 2 Mt-C per
year less on private lands than in the
RPA scenario. However, this increased
rate of harvest was not sustained and
by the 2030’s carbon uptake was 1
Mt-C per year in excess of the RPA
scenario. For 2040, the total US forest
carbon pool is estimated at 38.0 Gt.
Again, the carbon total for private tim-
berlands changed less than 0.1 Gt.
In this preliminary analysis, there was
little effect of the addit;onal harvest
reduction (0.36 billion cubic feet per
year reduction in 2000) on National
Forests. The growth and standing
inventory for the available timberland in
National Forests over the 50-year
period was nearly the same for the
Current Forest Plans and Reduced
National Forest Harvest scenarios. The
bulk of the additional harvest reduction
occurs in the Rocky Mountain Region,
yet the projected inventory there was
virtually unchanged. As with the Current
Plans scenario, there was not a sus-
tained compensatory harvest increase
on private lands. Because the estimates
of future growth and growing stock vol-
ume were based on reports from indi-
vidual National Forests. the detailed
analysis of age class distributions
needed to explain the insensitivity of
the growth and growing stock inventory
to harvest reductions was unavailable.
Any firm conclusion regarding effects of
these additional harvest reductions
must therefore await more detailed
treatment of long term shifts in stand
age class distribution.
Both the recycling scenarios and the
afforestation scenarios further
enhanced carbon accumulation over
the 50-year simulation period. The
afforestation scenarios had greater
impact on the 2040 carbon pool total:
values ranged from 38.3 to 38.7 Gt ver-
sus 38.3 to 38.4 for the recycling poli-
cies. However, the High Recycling
scenario manifested the largest imme-
diate increase in carbon uptake. For the
period 1990 to 2010. carbon accumula-
tion rates were an average of 13 Mt-C
per year higher than for the Current
Forest Plans. By comparison, the most
effective afforestation policy over this
same period, the $220 million per year
Parks and Hardie scenario, achieved a
8 Mt-C per year increase. The major
benefits of afforestation policies are
xiv
-------
Executive Summary
delayed until the latter half of the simu-
lation, as the growth rates of the young
forests begin to substantially increase.
For the period 2010 to 2040, the aver-
age carbon accumulation rate of the
$220 miUion Parks and Hardie scenario
is 19 Mt-C per year greater than the
Current Forest Plans uptake rate. The
High Recycling scenario achieved only
a 6 Mt-C per year advantage over
Current Forest Plans over this same
period.
In the afforestation scenarios consid-
ered here, the maximum increase in the
forest land base was 5 x 106 ha (12.3 x
i06 acres). This resulted in carbon sink
increase of nearly 16 Mt-C per year
over the 50-year simulation. Nationally,
however, there are approximately 47
million ha (116 x 106 acres) of
marginally productive, privately-owned
crop and pasture land that is biologi-
cally capable of supporting tree growth
(Parks 1992). This land presents an im-
portant opportunity for expanding the
area of forest land in the US, and
thereby increasing the rate of carbon
sequestration.
The result of the first combination sce-
nario was essentially the same as the
Moderate Recycling scenario, in part
due to the apparently negligible effect
of lower public forest harvest rates on
the forest inventory (combination sce-
narios assume public lands manage-
ment as described in the Reduced
National Forest Harvest). However, as
noted, there is considerable uncertainty
about the data for public forests. The
effect of increasing exports in the sec-
ond combination was to significantly
reduce the benefits of implementing the
Moderate Recycling scenario. The third
combination, combining the Moderate
Recycling assumption with doubled
export levels and the $110 million per
year Moulton and Richards scenario,
yielded a benefit significantly greater
than the unaltered afforestation sce-
nario. The increased timber supply
associated with paper recycling and
afforestation was greater than the har-
vest increases associated with the
doubling of timber exports. The combi-
nation scenarios emphasized the signif-
icant impact which otherwise
unconsidered factors, in this case
timber export trends, may have upon
the simulations described in this report.
The analyses of carbon budgets using
the TAMM/ATLAS/FCM modeling
framework reported here do not include
possible consequences of climate
change. Projected climate changes
could significantly alter terrestrial car-
bon sources or sinks. Of particular con-
cern is the carbon source which would
be created by extensive forest dieback
at the warmer, drier end of each tree
species’ range. Even where particular
forest types would still be maintained,
effects of climate change may include
significant changes in forest productivity
and, hence, on the forest sector carbon
budget.
The major research issues related to
advancing the modeling framework
developed in this analysis involve
improving the forest inventory for public
lands, incorporating potential effects of
climate change on forest distribution
and productivity, and integration of the
US forest sector model with compara-
ble models being developed in other
countries. Satellite remote sensing will
provide a means for developing a
simplified but functional forest inventory
across all land ownership classes.
Coupling of vegetation/climate models,
process-based ecosystem models, and
the existing TAMM/ATLAS/FCM frame-
work will permit projections of climate
change effects on the forest sector car-
bon budget. The agreement on global
xv
-------
Executive Summary
warming, recently signed by many
counties at the UNCED Conference,
will foster development of comparable
country-level carbon budgets.
Interaction with the research groups
formulating these budgets, and with the
research community using atmospheric
C02 concentrations to model terrestrial
carbon sources and sinks, will promote
improved understanding of anthro-
pogenic influences on the global car-
bon cycle.
Considered as a whole, there are sig-
nificant opportunities for conservation
and sequestration of carbon in the
forestry sector of the US and globally
(Sampson and Hair 1992). Obviously,
forest sector policies alone will not
compensate for the fossil carbon emis-
sion increases projected over the com-
ing decades (Lashof and Tirpak 1990,
IPPC 1992a). However, they can con-
tribute to moderating the overall trend
towards an increasing atmospheric
C02 concentration.
xvi
-------
Executive Summary
Table 1. Summary of policy Scenarios.
1989 RPA Assessment
Based on 1989 estimates of future forest timber supply,
demand, and harvest. Harvest (growing stock volume) on
National Forests increases from 2.16 billion cubic feet in
1986 to 2.17 in 2000 and to 2.53 billion cubic feet in 2040.
Population increases from 242 million to 333 million, and
GNP grows by 2 to 3% per year, from 1990 to 2040.
Wastepaper recycling increases from 20.9% in 1986 to
21.0% in 2000 and 28.0 % in 2040.
Current Forest Plans
(base scenario)
Like the RPA Assessment, except harvest on National
Forests decreases from 2.16 billion cubic feet in 1986 to
1.72 billion cubic feet in 2000, and then increases to 2.01
billion cubic feet in 2040.
Reduced National Forest
Harvest
Like the Current Forest Plans scenario, except harvest on
National Forests, decreases from 2.16 bilUon cubic feet in
1986 to 1.36 bilkon cubic feet in 2000 and then increases
to 1.59 billion cubic feet in 2040.
Moderate Recycling
Like the Current Forest Plans, except wastepaper recy-
cling, increases from 20.9% in 1986 to 28.5% in 2000 and
41.5% in 2040.
High Recycling
Like the Current Forest Plans, except wastepaper recy-
cling, increases from 20.9% in 1986 to 45.0% in 2000 and
remains constant through 2040.
Afforestation
Moulton and Richards
$110 million
$1 10 million per year for 10 years is used in afforestation
of pasturelands and marginal croplands (2.3 x 106 ha).
Assumptions of annual costs and forest growth rates are
from Moulton and Richards (1990). Otherwise follows
Current Forest Plans.
Atforestation
Moulton and Richards
$220 million
$220 million per year for 10 years is used in afforestation
of primarily pasturelands (4.0 x 106 ha). Assumptions of
annual costs and forest growth rates are from Moulton and
Richards (1990). Otherwise follows Current Forest Plans.
Atforestation
Parks and Hardie
$1 10 million
$110 million per year for 10 years is used in atforestation
of primarily pasturelands (2.8 x 106 ha). Assumptions of
annual costs and forest growth rates are from Parks and
Hardie (1992). Otherwise follows Current Forest Plans.
Afforestation
Parks and Hardie
$220 million
$220 million per year for 10 years is used in afforestation
of primarily pasturelands (5.0 x 106 ha). Assumptions of
annual costs and forest growth rates are from Parks and
Hardie (1992). Otherwise follows Current Forest Plans.
Combination 1
Includes the assumptions of the Reduced National Forest
Harvest scenario and the Moderate Recycling scenario.
Combination 2
Includes the assumptions of the first combination scenario
and an assumption of a doubling in timber exports.
Combination 3
Includes the assumptions of second combination scenario
and the Low Moultan and Richards afforestation scenario.
xvii
-------
Executive Summary
Table 2. Summary of policy scenario results.
1989 RPA Assessment Forest land accumulates carbon at a rate of 66 Mt-C per
year in the 1990’s after harvest removals are accounted
for. The accumulation rate declines over time to 5 Mt-C per
year because of a shrinking private forest land base,
increased harvest levels and a shift to younger age
classes.
Current Forest Plans
(base scenario)
.
Public lands accumulate an additional 10 Mt-C per year,
bringing the national sink to 74 Mt-C per year in the
1990’s. The downward trend over time in the carbon
accumulation rate is similar to the RPA Assessment
scenario.
Reduced National Forest
Harvest
Harvest reductions are likely to retain an additional sev-
eral Mt-C per year on public lands.
Moderate Recycling
Carbon uptake, relative to Current Forest Plans, gradually
increases to a peak of 13 Mt-C per year greater in 2020’s.
Carbon pool increase by 2040 is 75% of that achieved by
High Recycling.
High Recycling
Reduced harvests on private lands lead to a carbon
accumulation rate of 8 Mt-C per year greater than the
Current Forest Plans in the 1990’s, and 17 Mt per year
greater in the 2000’s. The difference declines in the later
decades, and the same trend of reduction in the carbon
accumulation rate over time is observed. Carbon pool in
2040 is 0.4 Gt greater than Current Forest Plans estimate.
Afforestation
Moulton and Richards
$110 million
Afforestation in 1990’s enhances carbon accumulation
relative to Current Forest Plans by 13 Mt-C per year in
2010’s. Carbon pool value in 2040 is 0.4 Gt greater.
Aflorestation
Moulton and Richards
$220 million
Carbon accumulation gains over Current Forest Plans are
about 150% of the gain from the $110 million per year
scenario.
Afforestation
Parks and Hardie
$110 million
Essentially the same result as the Moulton and Richards
$110 million scenario.
Afforestation
Parks and Hardie
$220 million
Most effective of the afforestation scenarios. Carbon pool
in 2040 is 0.8 Gt greater than Current Forest Plans.
Combination I
Very similar to Moderate Recycling.
Combination 2
Reduces the carbon accumulation rates of Combination 1,
in excess of Current Forest Plans, by 50% from 2010 to
2040.
Combination 3
Increases the carbon accumulation rate from Combination
2. Carbon pool in the year 2040 is 0.5 Gt greater than the
Current Forest Plans scenario.
xviii
-------
Executive Summary
REFERENCES
Auclair, A. N. D. and K. J. Andrasko. In
press. Net C02 flux in temperate
and boreal forests in response to
climate change this century: a case
study. In: K. M. Peterson (ed.).
Proceedings of the NATO Advanced
Workshop on the Biological
Implications of Global Climate
Change, Springer-Verlag. Berlin,
FAG.
Birdsey, A. A. 1992. Carbon storage
and accumulation in United States
forest ecosystems. USDA Forest
Service General Technical Report
WO-57.
Cost, N. D., J. 0. Howard, B. Mead, W.
H. McWilliams, W. B. Smith, D. D.
VanHooser, and E.H. Warton. 1990.
The biomass resource of the United
States. USDA Forest Service
General Technical Report WO-57.
Haynes, R. W. and D. M. Adams. 1985.
Simulations of the effects of alterna-
tive assumptions on demand-supply
determinants on the timber situation
in the United States. USDA Forest
Service Forest Resources
Economics Research, US
Government Printing Office,
Washington, DC.
Houghton, R. A., J. Unruh, and P. A.
Lefebvre. 1991. Current land use in
the tropics and its potential for se-
questering carbon. In: D. Howlett
and C. Sargent (eds.). Proceedings
of a technical workshop to explore
options for global forest manage-
ment, Bangkok, Thailand, April 24,
1991. International Institute for
Environment and Development,
London.
IPCC (Intergovernmental Panel on
Climate Change). 1992. IPCC 1992
Science Assessment. Report to
IPCC from Working Group 1, pre-
pared by the IPCC Group at the
Meteorological Office, Bracknell, UK.
Koch, P. 1989. Estimates by species
group and region in the USA of
below-ground root weight as a per-
centage of ovendry complete tree
weight and carbon content of tree
portions. Consulting Report to the
USDA Forest Service.
Lashof and Tirpak (eds.). 1990. Policy
options for stabilizing global
change: Report to Congress. US
EPA Environmental Protection
Agency, Office of Policy, Planning,
and Evaluation, Climate Change
Division, Washington, DC.
Mills, J. A. and J. C. Kincaid. 1992. The
aggregate timberland assessment
system--ATLAS: a comprehensive
timber projection model. USDA
Forest Service General Technical
Report PNW-GTR-281.
Moulton, R. J. and K. R. Richards. 1990.
Costs of sequestering carbon
through tree planting and forest
management in the United States.
USDA Forest Service Genera!
Technical Report WO-58.
NAS (National Academy of Science).
1991. Policy implications of green-
house warming. Report of the
Mitigation Panel. National Academy
Press, Washington, DC.
Parks, P. J. 1992. Opportunities to
increase forest area and timber
growth on marginal crop and pas-
xix
-------
Executive Summary
ture land. In: R. N. Sampson, and D.
Hair (eds.). Forests and Global
Warming. American Forestry
Association, Washington, DC.
Row, C. and R. B. Phelps. 1990. Tracing
the flow of carbon through the US
forest products sector. Proceedings
of the 19th World Congress of the
International Union of Forest
Research Organizations, August 5,
1990, Montreal, Canada,.
Rubin,E. S., R. N. Cooper, A. A. Frosch,
T. H. Lee, G. Marland, A. H.
Rosenfeld, and D. D. Stine. 1992.
Realistic mitigation options for
global warming. Science 257:148-
149, 261-266.
Sampson, R. N. and D. Hair (eds.).
1992. Forests and Global Change.
Volume One: Opportunities for
Increasing Forest Cover. American
Forestry Association, Washington,
DC.
Schroeder, P. E. and L. Ladd. 1991.
Slowing the increase of atmospheric
carbon dioxide: a biological
approach. Climatic Change 19:283-
290.
Sedjo, A. A. and A. M. Solomon. 1989.
Climate and forests. In: N. J.
Rosenberg, W. E. Easterling, Ill, P.
R. Crosson, and J. Darmstadter
(eds.). Proceedings of a Workshop,
Washington DC, June, 1988.
Resources for the Future, Inc.,
Washington, DC.
xx
-------
SECTION 1. INTRODUCTION
A. Scope and objectives
Rising levels of atmospheric C02 and
other radiatively important trace gases
(e.g., CH4, N20) from anthropogenic
sources are likely to induce changes in
the earth’s climate over the coming
decades (IPCC 1990). While there
remains considerable uncertainty about
the rate and magnitude of the possible
climate change, there is an emerging
consensus that policies relevant to
stabilizing or reducing the emission of
C02 and other “greenhouse gases’
should be explored (NAS 1991, Rubin
et al. 1992).
Terrestrial ecosystems have a major
role in the global cycling of carbon
(Figure 1.1). Vegetation and soil con-
tain about three times as much carbon
as the atmosphere. Also, each year,
they exchange about 20 times as much
carbon with the atmosphere as is emit-
ted from fossil fuel use. For this reason,
one proposed approach for slowing the
increase in atmospheric C02 is to
manage terrestrial ecosystems to con-
serve or sequester carbon (Schroeder
and Ladd 1991, Sedjo and Solomon
1989, Sampson and Hair 1992) by
changing the balance between the
uptake of carbon dioxide from the
atmosphere through photosynthesis
and the release of carbon dioxide
through respiration and biomass burn-
ing. Ecosystems in which release of
C02 exceeds uptake are net sources”
of C02; if uptake from the atmosphere
exceeds release, then they are Nnet
sinks” of C02 from the atmosphere. In
this report, “conservation” means that
sources are being managed to
decrease the net release of C02.
Examples of conservation include
decreasing deforestation rates or
decreasing loss of soil carbon from
disturbed systems. “Sequestration”
refers to management of ecosystems to
establish or augment sinks of C02.
Reforestation, afforestation, and reha-
bilitation of degraded land are exam-
pies of sequestration.
Forest ecosystems are particularly
important in consideration of carbon
conservation and sequestration oppor-
tunities because of their large accumu-
lation of woody biomass over time.
Globally, forests contain about 60% of
terrestrial carbon (Waring and
Schlesinger 1985). In 1989, the
Noordwijk Ministerial Conference rec-
ognized the importance of forests in
global carbon cycling and established a
provisional goal of increasing world
forests at a rate of 12 million ha yr 1
(Noordwijk Conference Report 1989).
Research on national-level forest car-
bon budgets is proceeding in several
countries around the world, notably
Canada (Apps and Kurz 1991), Russia
(Kolchugina and Vinson, in press),
Sweden (Eriksson 1991), and New
Zealand (Hollinger and Hunt 1992). In
Joensuu, Finland, in May 1992 a
meeting, sponsored by the
Intergovernmental Panel on Climate
Change (IPCC), focused on appropriate
approaches and methodologies for
modeling the forest sector carbon bud-
get (IPCC, in press). This meeting fol-
lowed earlier IPCC meetings (Bangkok,
Thailand 1991; Canberra, Australia
1992) on the potential role of forests in
mitigating global climate change
I
-------
Section 1
Figure 1.1. The global carbon cycle, including major pools and annual flux of carbon
(Schneider 1989b).
(Howlett and Sargent 1991; IPCC
1992b). A global assessment of
promising management practices con-
cluded that conservation! sequestration
of carbon in forests was highly competi-
tive with other options to mitigate
greenhouse gas emissions (Dixon et al.
1991a).
The Climate Change Convention
Framework promulgated by the US and
other countries at the June 1992 United
Nations Conference on Environment
and Development (UNCED) calls for
completion of comprehensive national-
level carbon budgets for all countries.
Biologically-based emission sources
such as deforestation make a
significant contribution to global C02
emissions. Biologically-based sinks,
such as growing forests, may also have
a global scale impact. Thus, forest land
carbon budgets are an essential
component of the national analyses.
UNCED also enumerated several
forestry principles which must be
considered when evaluating options for
managing forests to conserve/
sequester atmospheric carbon (Box 1).
The utility of managing terrestrial
ecosystems to conserve or sequester
atmospheric carbon must be measured
against the policy objectives for altering
carbon emissions. Currently, carbon
emissions from the use of fossil fuels in
the US are about 1,300 Tg-C yr 1 (Tg =
i012 g 106 tons — megaton). From
1980 to 1988, emissions increased at
an average rate of 6.4 Tg-C yr 1 (NAS
1991). Emission increases were greater
toward the end of this period because
of increased economic activity. Thus, a
2
-------
Section 1
Box 1. UNCED Forestry Principles
Proposals to manage forests to conserve and
sequester atmospheric carbon must be con-
sidered in the context of many potentially
conflicting or complimentary management
options. Demographic and environmental
pressures in many parts of the world are
sIrair ing the capacity of forests to continue
to supply basic goods and services (Mather
1990). As forests are exploited, serious
impacts include loss of biodiversity, soil
erosion, and reduced biomass productivity.
These impacts and others affect the eco-
nomic status of nearly 2 billion people
wor1dv ide whose livelihoods depend on
forest systems. Multiple stresses resulting
from global climate change projected for the
next half century could further contribute to
adverse forest impacts (Smith and Tirpak
1989).
In 1990. the G-7 nations (Canada, France,
Germany, Great Britain, Italy, Japan, and
the United States) agreed to support the con-
cept of developing an international agree-
ment on the use and management of forest
resources (Maini 1991). The United Nations
Conference on Environment and
Development (UN CED), recently concluded
in Rio de Janeiro, addressed this topic
specifically. The result of the UNCED
negotiations was a “non-legally binding
authoritative statement of principles for a
global consensus on the management, con-
servation, and sustainable development of
all types of forests” (UNCED 1992).
The statement of principles contains 51
paragraphs covering a wide range of
topics. Paragraphs addressing major themes
affecung forest management include:
• Forests are related to economic develop-
ment and any evaluation of forest policy
should take place within the broader con-
text of national development.
• Countries have sovereign rights to sus-
tainably use their forest resources as
needed. These rights include all biological
resources, including genetic material.
• All population groups should be
respected and included in the formulation
of forest policy. This includes local popu-
lations, indigenous people, and women.
• International cooperation is critical for
financing conservation, cost sharing,
redressing external debt of developing
countries, technical assistance, and tech-
nology transfer.
• Efforts should be undertaken toward
greening the world, including reforesta-
tion, afforestation, and forest conserva-
tion.
• Research, institutional, and educational
capabilities should be strengthened and
there should be a free international
exchange of research results.
• Free trade in forest products should be
facilitated and trade barriers to higher
value-added products should be
reduced or removed.
3
-------
Section 1
policy objective of stabilizing carbon
emissions would require emissions
reductions, or implementation of
strategies for conserving/sequestering
carbon, amounting to an additional 6 to
10 Tg-C yr 1 each year. A policy objec-
tive of reducing emissions by 10 to 20%
would require emission reductions or
increased carbon conservation/
sequestration amounting to 130 to 260
Tg-C yr 1 .
This document presents a model of the
current and future carbon budget asso-
ciated with the forest ecosystems of the
conterminous US. The focus is on
effects of economic and environmental
policy changes for the period of 1990-
2040. In this study, the concept of “for-
est ecosystem” has been expanded to
“forest sector” by including biomass that
has been physically removed for
human use (Box 2). The potential
effects of climate change have not been
incorporated; however, current
research is directed at addressing this
question. The specific objectives of this
report are to:
1. Develop a methodology for
quantifying the current and pos-
sible future forest sector carbon
budget for the US.
2. Apply the methodology to quan-
tify the current status of carbon
pools and flux within the US for-
est sector.
3. Apply the methodology to sev-
eral scenarios based on
afternative policy options.
4. Identify gaps and deficiencies in
the data and in the modeling
approaches that require further
effort.
Although the specific policy options
considered in the report are realistic
(Table 1.1), they do not represent the
lull spectrum of options that could be
considered (NAS 1991; Rubin et al.
1992). Furthermore, they may not nec-
essarily be the “best” options. The intent
is to develop a tool that can be used to
give broad technical guidance in eval-
uating various policy options for con-
serving or sequestering carbon in the
forest sector.
B. Overview of modeling ap-
proach
The basic approach to modeling the
current pools and net flux of carbon in
US forests (see Figure 1.2) has been to
develop a database which links inven-
tory data on the areal extent of different
forest types and age classes, with
stand-level carbon budgets which
quantify the pools of carbon associated
with each age class within those forest
types (Sections 3.A, 3.B). The inventory
itself then provides the basis for an
estimate of the carbon pool sizes and
the age-specific increments; that is, dif-
ferences between pool sizes in suc-
cessive age classes provide the basis
for a flux estimate. The forest inventory
data and the stand-level carbon bud-
gets described in this report are the
primary inputs to the Forest Carbon
Model (FCM) which calculates total
carbon pools. Additional information on
harvesting, wildfire, and change in the
pool of forest products still in use and in
landfills was used to estimate a base
year carbon flux.
The inventory of forest carbon and the
forest land base itself can change sub-
stantially as a result of economic and
policy drivers (Box 3). In this study,
changes in carbon storage in response
to various forest sector policy scenarios
(Table 1.1) were evaluated by examin-
ing changes in forest inventories from
4
-------
Section 1
Box 2. What is a Forest-sector Carbon Budget?
A carbon budget is a bookkeeping system
for tracking the amount of carbon in vari-
ous reservoirs (“pools”), and the amount of
carbon transferred among the reservoirs
and between the reservoirs and the atmo-
sphere (“flux”) (see figure below). The for-
est sector carbon budget, as referred to in
this report, has three major pools: the
organic matter of forest ecosystems, the
products derived from forests, and
products discarded into landfills. The major
pools are, in turn, made up of subsidiary
pools. For example, important pools within
forests are trees, other vegetation, soil, the
forest floor, and woody debris. “Net flux”
is the difference between total uptake into a
pool and total output from the pool, and is
equal to the change in the pool size during
some time interval.
The main uptake of carbon into forests is
through photosynthesis, the fixation of at-
mospheric C02 by green plants (“primary
producers”). Carbon loss from forests is
mainly through respiration by green plants
and other organisms, through burning, and
through harvest of v ood by humans. The
first two processes result in the direct
release of carbon to the atmosphere. The
difference between photosynthesis and
respiration by green plants is Net Primary
Production (NPP). The difference between
photosynthesis and respiration by all
organisms, including decomposers (i.e.,
Net Ecosystem Production, NEP), is the
total change in the carbon content of forests
due to biological processes. Thus, NEP is
also the net flux of carbon from the
atmosphere to the forest from biological
processes. The “net accumulation” by
forests is NEP minus carbon removed by
harvest. NPP, NEP, and the flux among
the carbon pools within the forest (i.e.,
internal forest dynamics), are determined
by stand characteristics such as tree
species, age class, and site productivity.
Regional and national estimates are ob-
tained by combining forest stand NEP es-
timates with forest inventory data. In
contrast, the harvest and reforestation rates
are determined by external economic and
policy factors. This report describes the
methods and results of analyzing internal
forest dynamics to estimate regional and
national NEP, and the coupling of forest
dynamics to the external economic and
policy effects.
Unlike the forests, there are no systematic
inventories of the product and landfill pools
or of the net annual transfer into or out of
these pools. The net annual change in
carbon stored in the forest products pool
can be estimated as the difference between
published national values for the production
of forest products and the transport of
forest products to landfills. The national ne
change in the pool of forest products in
landfills can be estimated as the difference
between the amount of forest products
transported to landfills and the carbon
emissions to the atmosphere from landfills
due to forest products estimated proportion
of emission of carbon from landfills to the
atmosphere due to forest products.
•LarvlSrlw
5
-------
Section 1
Table 1.1. Specifications for the policy options. See Table 1 (Executive Summary)
and the text for additional characterization.
National Forests
Scenarios Harvest
(x it 3 )
Recycling Intensity
(%)
Afforestation
(x 106 ha)
2000 2040 2000 2040
1989 RPA Assessment 2.17 2.53 21.0 28.0
Current Forest Plans 1.72 2.01 21.0 28.0
(base year)
Reduced National Forest 1.36 1.59 21.0 28.0
Harvest
Moderate Recycling 1.72 2.01 28.5 41.5
High Recycling 1.72 2.01 45.0 45.0
Afforestation 1.72 2.01 21.0 26.0 2.3
Moulton and Richards
$110 million
Aflorestation 1.72 2.01 21.0 28.0 4.0
Moulton and Richards
$220 million
Aftorestation 1.72 2.01 21.0 28.0 2.8
Parks and Hardie
$110 million
Atforestation 1.72 2.01 21.0 28.0 5.0
Parks and Hardie
$220 million
Combination 1 1.36 1.59 28.5 41.5
Combination 2 1.36 1.59 28.5 41.5
Combination 3 1.36 1.59 28.5 41.5 2.3
Additional assumption of a doubling of timber exports.
6
-------
Section 1
GNP and
atio H
TAMM forest Harvest by
sector model region
Volume by
inventory class
Forest Carbon
Model
UPDATED FOREST
CARBON POOLS
ATLAS forest
inventory model
Harvest by
ntory class
HARVCARB forest
product model
FOREST PRODUCTS
CARBON POOLS
Figure 1.2. Modeling framework for forecasting carbon pools.
(
7
-------
Section 1
1990 to 2040, as well as changes in the
post-harvest product and landfill pools,
as projected by models of various types
(Figure 1.2). Changes in forest
inventory on private timberlands were
estimated by linking an economic
model (TAMM), which estimates tree
harvest volume by region (1-faynes and
Adams 1985), to a forest inventory
model (ATLAS), which tracks forest
age-class distribution by forest type and
productivity class (Mills and Kincaid
1992). TAMM insures that the economic
consequences of implementing the
various policy options are allowed to
impact the carbon budget. The
inventory at any point in time is run
through FCM (Section 3.A) to estimate
total forest carbon. The change in total
forest carbon over a 10-year interval,
divided by 10. gave the average annual
flux for that interval. The model
HARVCARB (Row and Phelps 1990)
was used to estimate changes in the
amount of carbon in forest products and
in landfills from material harvested from
1990 to 2040.
For public lands, there are no stand-
alone inventory models. Furthermore,
inventory assumptions are embedded
in a variety of harvest planning models.
It was thus necessary to estimate future
carbon pools by using data from the
planning efforts at the level of individual
National Forests and other public
agencies.
C. Overview of this report
This document examines the role of
biogeochemical processes and
harvesting in the current forest land
carbon budget of the US and evaluates
several forest policy scenarios in terms
of their impact on carbon conservation
and sequestration. We begin with back-
ground material on the global carbon
cycle, and on other efforts to model
national forest sectors (Section 2). We
then describe methods used in evaluat-
ing (1) the current status of carbon
Box 3. Changes in the Size of the US Forest Land Base
About 25% of the area of the United States is
occupied by closed forests, 226 million ha.
This represents 8% of the world’s closed
forests. Approximately 45%, or 102 million
ha, of US closed forests are managed. This is
10% of the world’s managed forest area.
Other wooded areas cover 71 million ha, or
8% of the country. The US accounts for
17% of annual global reforestation, i.e..
replanting after harvest, about 1.8 million ha
yr 1 (WRI 1992).
sents a decline of over 45% which was
caused primarily by clearing for agriculture
and by industrial felling and lumbering.
Some recovery associated with abandonment
of farmland and its reversion to forest, was
evident through the 1960’s but more recently
timberland area has began to decrease again
(Aug c i a]. 1990).
From the arrival of European settlers around
1630 until the 1920 ’s, the area of commercial
forest land in the United States decreased
from 344 million ha to about 200 million ha
(Claw son 1979, Williams 1988). This repre-
The future changes in forest land area in the
US are highly dependent on economic and
political factors. The study of Aug et al.
(1990) suggested a slow decrease in the area
of private timberland over the next 50 years.
Alternatively, large areas of public and pri-
vate land might be planted.
8
-------
Section 1
stored in various parts (i.e., “pools”) in
forests and other terrestrial ecosystems
in the US, (2) the amounts of carbon
exchanged with the atmosphere or
removed through harvests each year
(i.e., net flux), and (3) the projected
changes in carbon pools and flux
through the year 2040 as to several
policy options that might alter these
pools and flux (Section 3). This is done
in the context of anticipated economic
demand for forest products for the
period 1990-2040.
Results of these analyses are pre-
sented in Section 4 along with a dis-
cussion of the policy considerations.
Section 5 introduces the potential ef-
fects that climate change might have on
forest carbon pools and flux, and
Section 6 discusses research needed
to begin assessing the effects of climate
change on national carbon budgets.
Other research needed to refine and
extend the Forest Carbon Model is also
discussed. A set of supporting docu-
ments is presented in Section 9. These
include a discussion of uncertainties in
the inventory data, an estimate of the
carbon in woodlands of the US, a doc-
umentation of the approach used in es-
timating woody debris pools, an esti-
mate of the carbon flux associated with
nonforest lands, and a glossary.
9
-------
Section 1
10
-------
SECTION 2.
A. Introduction
BACKGROUND: THE TERRESTRIAL CARBON
BUDGET
The CO 2 concentration in the atmo-
sphere is increasing at a rate of approx-
imately 1.5 ppm per year as a result of
fossil fuel combustion and deforestation
(IPCC 1990). As atmospheric CO 2 1ev-
els rise, global temperatures may
increase through the greenhouse effect
(Hansen et al. 1988, Schlesinger and
Mitchell 1985). Predictions are that
average global temperatures may
increase by 1.5 to 4.5 C in the first half
of the next century.
The potential consequences of increas-
ing atmospheric carbon levels are
important enough that governments are
considering programs to stabi-
lize/reduce anthropogenic emissions
and increase forest-carbon sequestra-
tion (Lashof and Tirpak 1990, Downing
and Cataldo 1992). However, to
develop effective policy options, an
understanding of the global carbon
cycle, and how anthropogenic carbon
emissions may affect it, is necessary.
Our understanding of the global carbon
cycle (i.e., sources and sinks) suggests
that managing terrestrial ecosystems,
especially forests, to sequester and
conserve carbon may contribute to
moderating the rise in atmospheric C02
(Box 4; Sampson and Hair 1992).
However, better knowledge of carbon
pools and transfer rates is needed.
Such information will be crucial in for-
mulating national and international
policy to reduce the risks of altering the
composition of the atmosphere.
B. The global carbon cycle
The global carbon cycle is the move-
ment of carbon among the atmosphere,
terrestrial biosphere, and ocean pools
(Schneider 1989a, Post et al. 1990,
Dixon and Turner 1991, Simpson and
Botkin 1992). Although the general
nature of the global carbon cycle is well
known, relatively large uncertainties
exist in the estimation of the magnitude
of the global carbon pools and the flux
among them. On a global scale, the
amount of carbon in the different pools
has been estimated by a variety of
means (Whittaker and Likens 1975,
Ajtay et al. 1979, Bolin et al. 1979,
Olson et al. 1983, Schlesinger 1984.
Mooney et al. 1987, Schneider 1989,
Schneider 1989a, Downing and
Cataldo 1992, Simpson and Botkin
1992). Current estimates suggest that
oceans store by far the largest amount
of carbon at 38,500 Gt-C (Gt = tons
= Gigaton = g). The second
largest reservoir is fossil fuels with
estimates ranging from 5,000 to 10,000
Pg-C. Soil is the next largest pool at
1170 to 1740 Pg-C. The vegetation and
atmospheric carbon reservoirs are
about equal and are the smallest in
comparison to the others. Estimates of
carbon storage in vegetation ranges
from 560 to 830 Pg-C. The atmospheric
pool contains approximately 740 Pg-C
of carbon.
Even though the terrestrial biosphere,
vegetation and soils, store much less
carbon than does the oceans, the
amount of carbon that cycles annually
11
-------
Section 2
Bo ,. 4. (Ilobal View of Carbon Conservation/Sequestration
In 1991. a global assessment was
undertaken of the potential of forest
management practices to store atmospheric
carbon (Dixon et al. 199]a; Winjum et al.
1992; Dixon et al. , in press; Winjum et
al, in press; Schroeder et al., in press).
The assessment was based on information
on the rates of carbon storage per hectare
for many practices, their implementation
costs, and estimates of the amounts of land
suitable for forest management.
Information was compiled through a survey
of current published technical literature for
forested nations representing boreal.
temperate and tropical regions of the world.
Key findings of the assessment are
highlighted in the fo1lo ing paragraphs.
Because of the geographic scope of the
1991 assessment 1 it was necessary to use
generalized representations of ecoregions
and data. Thus, the results are broadly
applicable for comparing ecoregions and
countries. However, the)’ lack the spatially
detailed basis needed for within-country
analyses such as that presented for the US
in this document. Also, the data for the
1991 assessment did not consider change
over time so that ii was not possible to
estimate carbon flux.
Carbon Storage
The assessment indicated that the most
promising forest management practices to
sequester carbon in the terrestrial biosphere
include: reforestation in the temperate and
tropical latitudes, afforestation in the
temperate regions, and agroforestry and
natural reforestation in the tropics. Least
promising from a carbon storage standpoint
were the application of silviculcural
practices, such as thinning, fertilization and
other stand improvement treatments, at all
latitudes.
The potential carbon storage ranges of
forestation and silvicultural practices by
major latitudinal biomes were:
Forestation
Silvicuiture
iC/ha
Boreal
15-40
3-10
Temperate
30-180
10-45
Tropical
30-130
14-70
Costs of Storing Carbon
The median cost efficiency for all
management practices in terms of
establishment costs was about $5/i-C with
an interquartile range (middle 50’ of
observations) of $1 to $19/t-C. The most
cost-efficient forestry and agroforestry
practices, based on establishment costs.
within zones of latitude are shown in the
table below.
Total costs per ton of carbon sequestered.
which include land rental, would be
considerably higher. The magnitude of the
total costs is difficult to estimate because of
economic uncertainties regarding factors
such as land rental.
Suitable Land Area
Preliminary estimates indicated that there
are large amounts of land which are
technically suitable for expanding the area
of managed forest and agroforestry
systems. Approximate areas of suitable
land by latitudinal regions appear to be:
boreal, 100 to 200 million ha (Volz et al.
1991); temperate, 200 to 300 million ha
(Moulton and Richards 1990, Volz et al.
1991); and tropics, 420 million (Grainger
1991) to almost 2 billion ha (Houghton et
al. 1991). These estimates of technically-
suitable lands must be viewed with caution.
The amount of land actually available from
12
-------
Section 2
a social and economic standpoint remains
uncertain, but it is likely to be much less, in
most cases, than the land area that is
technically suitable (Trexler 1991a). The
fallow stage of a shifting agricultural cycle
serves as an illustration. Technically, such
land is suitable for natural vegetative regen-
eration and the accompanying accum-
ulation of biomass and carbon. in reality,
such land may be a part of an important
mosaic of land use patterns that is required
to support the agricultural requirements of
local populations. Therefore, more
complete data on the amount of land
actually available for expanding forest and
agroforestry systems are clearly needed,
and research efforts are underway to
achieve this objective.
Establishment costs of cost-efficient practices.
Boreal:
Natural Regeneration
Reforestation
Temperate:
Natural Regeneration
Afforestation
Reforestation
Tropical:
Natural Regeneration
Agroforesiry
Reforestation
Median S!tC
5
(4-11)
8
(3-27)
I
(<1-1)
2
(< 1-5)
6
(3-29)
(<1-2)
5
(2-11)
7
(3-26)
Median s/ha
93
(83-126)
324
(127-455)
9
(9-10)
259
(41-444)
357
(257-911)
178
(106-238)
454
(255-699)
450
(303-1183)
*J.nterquartile ranges are shown in parentheses
13
-------
Section 2
between the terrestrial biosphere and
the atmosphere is about equal to that
which cycles between the oceans and
the atmosphere. Their combined
annual flux is about 30% of the total
atmospheric carbon pool.
A key concern, from the perspective of
possible global climate change, is that
the flux of CO 2 in and out of the atmo-
sphere is not balanced. That is, more
carbon, mainly in the form of C0 2 , is
accumulating in the atmosphere than is
being removed through plant photosyn-
thesis and abiotic processes. It is esti-
mated that the atmosphere is gaining
about 3 Pg of carbon per year
(Schneider 1989a, Tans et al. 1990).
The two principal sources of the atmo-
spheric CO 2 are fossil fuels combustion
(Marland 1989, Schneider 1989b) and
deforestation (Houghton et al. 1983,
Woodwell et al 1983, Houghton et al
1985, Palm et al. 1986, Detwiler and
Hall 1988). The annual contributions of
these two anthropogenic carbon
sources total 6 10 8 Pg with 1-2 Pg
derived from deforestation. Atmospheric
CO 2 contributions from deforestation
and fossil fuel combustion are likely to
increase (IPCC 1990).
One of the critical tasks in assessing
recent changes in the carbon cycle is
determining where the anthropogenic
carbon emissions are accumulating,
since only about half remains in the
atmospheric pool. The two possibilities
are the oceans and the terrestrial bio-
sphere. Some recent analyses suggest
that temperate and tropical forest
ecosystems are sequestering more
carbon than previously reported (Sedjo
1992, Sedjo and Solomon 1989).
However, large uncertainties remain
and although the potential for unman-
aged and managed forest ecosystems
to sequester and conserve atmospheric
carbon appears to be large the current
sequestration rate and the potential for
additional sequestration needs to be
better quantified.
A recent workshop on the global carbon
cycle evaluated the role of terrestrial
carbon sinks (Lugo and Wisniewski
1992). Some of the major conclusions
and recommendations of the workshop
were:
• The global carbon cycle is important
but poorly understood.
• The exchange of CO 2 between
biota and the atmosphere has not
been in steady slate for several
centuries. Many terrestrial ecosys-
tems are currently accumulating
carbon.
• Natural plant succession after dis-
turbance (e.g., deforestation, forest
tree die back) may sequester large
amounts of atmospheric carbon. It is
important 10 understand the role that
the various seral stages may serve
as sources or sinks of carbon.
• Increasing levels of atmospheric
CO 2 may increase plant productivity
and may increase carbon storage in
terrestrial ecosystems.
• The carbon cycle is
linked to nutrient and
cycles.
• Proactive terrestrial ecosystem
management can enhance carbon
sequestration and conservation and
is compatible with biodiversity con-
servation, sustainable land use, and
economic development.
C. Regional carbon budgets
The global carbon budget quantifies the
carbon cycle by describing the transfer
inextricably
hydrological
14
-------
Section 2
of carbon among the various reservoirs
(BoIin at al 1979, Post et al. 1990). In
recent attempts to quantify the global
carbon cycle a discrepancy has arisen
between results of models which focus
on the time course in the change in the
atmospheric C02 concentration and
models which reconstruct the time
course of deforestation and land use
change.
Reconciliation of these modeling
approaches will require critical analysis
of assumptions about long-term
changes in terrestrial sources and sinks
of carbon (Lugo 1992). Attention to
possibly oversimplified approaches for
representing terrestrial systems is also
needed (Lugo 1992, Downing and
Cataldo 1992). For example, a recent
IPCC report (Melillo et al. 1990) used
only two biomes to represent all of the
tropics. Such reductive models lump
too much variety into a few biome cate-
gories. This results in poor estimates of
carbon pools and flux because sources
and sinks are averaged without proper
weighting.
Lugo and Brown (1991) argued that
each tropical forest life zone has its own
unique biota, history of land-use man-
agement, and set of carbon fluxes and
sinks, and therefore, should be evalu-
ated separately to quantify carbon
pools and transfer rates. This same
reasoning should also extend to tem-
perate and boreal forests and to other
biomes (e.g., grassland, desert) as well.
These considerations point towards the
need for more refined carbon budgets
at the regional or national scale.
Circumboreal countries such as Finland
have characteristic sinks for carbon in
the form of peatlands (Lame and
Pàivänen 1992). More highly forested
countries may have greatly differing
rates of tree 9rowth, harvest and distur-
bance, all of which strongly impact car-
bon flux. Data on rates of deforestation,
reforestation, forest stand classification
and condition, and fossil-fuel combus-
lion are usually specific to a
region/country and much more accu-
rately determined on geo-referenced
basis than using a global approach.
Region- or country-level carbon budget
analyses are important since they can
add to or show the inconsistencies
within the global-scale analyses (Sedjo
1992). Such geographically specific
data bases may verify the model-based
global carbon budgets, or they can
raise questions as to their credibility
and provide direction for additional
research. Also, country-level studies
will provide scientific data to help delib-
erations on policy issues regarding
CO 2 emissions, deforestation, refor-
estation, and pending global climate
change.
Some examples of region- and country-
level, forest-sector carbon budget anal-
yses are the tropical forests (Makundi
and Sathaye 1992), temperate forests
(Sedjo 1992), European forests
(Kauppi et al. 1992), Canada (Kurz et
al. 1992), Finland (Karjalainen and
Kellomäki, in press), Sweden (Eriksson
1991), Netherlands (Hell et al., in
press), Poland (Galinski 1992), the for-
mer Soviet Union (Kolchugina and
Vinson 1992), and New Zealand
(Maclaren and Wakelin 1991).
The value of these country level analy-
ses can be illustrated by evaluating the
carbon budget for New Zealand. The
amount of carbon currently in the
forests of New Zealand is approxi-
mately 1.13 Pg-C. An additional 0.22
Pg-C are stored in forest products such
as building lumber and paper. New
Zealand forests currently sequester
approximately 3.5 Tg-C yr 1 . Carbon
15
-------
Section 2
dioxide releases into the atmosphere
are estimated to be 6.0 Tg yr-i.
Obviously, New Zealand is a net source
for CO 2 . To achieve a favorable C02
balance, 50,000 ha yr 1 of reforestation
is necessary to offset the current emis-
sions. Reforestation at this rate would
increase New Zealand’s forest estate to
4.5 million ha by 2050, nearly 4 times
its present size. Such information can.
be used by the New Zealand govern-
ment in developing policy issues with
regards to CO 2 emissions and relor-
estatior .
D. Vegetation and the global
carbon cycle
Terrestrial vegetation is an important
component of the carbon cycle (Goreou
1990, Simpson and Botkin 1992). Plant
photosynthesis removes C02 from the
atmosphere while plant and microbial
respiration returns CO 2 to the atmo-
sphere. The carbon removed from the
atmosphere through photosynthesis
sustains plant growth and development.
Carbon is then transferred to other
organisms through a complex food
web. The impact of terrestrial vegetation
on atmospheric CO 2 levels is illustrated
by its seasonal fluctuation (Bolin et al
1979, King et al. 1987, Keeling et al.
1989). In the northern hemisphere,
atmospheric CO 2 is lowest during the
summer growing season when photo-
synthesis is at a maximum, and highest
during winter months when photosyn-
thesis is minimal. Bolin et al. (1979)
have estimated a 10% seasonal fluc-
tuation in atmospheric CO 2 levels.
In addition to seasonal variations, vege-
tation can influence long-term atmo-
spheric CO 2 levels through carbon
sequestration (Simpson and Botkin
1992). Not all forms of vegetation have
an equal potential to affect atmospheric
CO 2 . For example, trees and shrubs
can store carbon for hundreds of years
in woody plant parts. The carbon stored
is only returned to the atmosphere
through decomposition or burning. On
the other hand, the carbon stored in
non-woody plant parts and herbaceous
plants is often returned to the atmo-
sphere within a year of sequestration.
Thus, over broad spatial scales vegeta-
tion can be either a carbon source or
sink depending on many factors such
as past land-use history, degree of dis-
turbance, successional stage, stand
age, management practices, and vege-
tation type (Detwiter and Hall 1988).
Deforestation and biomass burning are
other mechanisms by which vegetation
affects the global carbon cycle.
Estimates of net global CO 2 emission
from land-use change such as
deforestation for 1980 are
approximately 10 to 30% of annual
anthropogenic emissions (Lashof and
Tirpak 1990).
Vegetation also influences the carbon
cycle indirectly through its effects on
climate (Simpson and Botkin 1992).
Energy and water balance simulations
of tropical rain forests suggest that
deforestation results in elevated
albedo, decreased evapotranspiration,
and increased soil temperatures; large-
scale deforestation may affect regional
precipitational patterns (Dickinson
1991). Such changes in climate can
reduce the potential for plant growth
and the rate of carbon sequestration.
The strong influence of vegetation on
the global carbon cycle suggests that
accurate assessments on the extent of
major ecosystems and vegetation
biomass densities with changes over
time are needed (Simpson and Botkin
1992).
16
-------
Section 2
E. Mitigation of rising carbon
dioxide levels
Atmospheric CO 2 has been increasing
since the industrial revolution in
response to anthropogenic emissions
and deforestation (Keeling et al. 1989).
Various policy options to moderate the
rising CO 2 concentration have been
presented that, if put into action, would
decrease the risks of global climate
change (Lashof and Tirpak 1990).
Maintaining CO 2 emissions at current
rates will not immediately reverse the
trend of increasing atmospheric levels.
Even if emissions remained at 1985
rates, the greenhouse effect would con-
tinue to intensify with CO 2 reaching 500
ppm by 2100 (Lashof and Tirpak 1990).
A 50 to 80% reduction in CO 2 emis-
sions would be necessary to stabilize
atmospheric CO 2 at the current level of
about 355 ppm. This is because CO 2
can have a long atmospheric residency
time and maintaining current emissions
rates will still result in rising levels.
Approximately 50 to 100 years would
be required following a drastic reduc-
tion in CO 2 emissions for the oceans
and terrestrial biosphere sinks to sub-
stantially reduce atmospheric CO 2 .
However, action can be taken to help
moderate the current trend of rising
atmospheric CO 2 and reduce the risks
of global climate change (Lashof and
Tirpak 1990). Action to reduce CO 2
emissions could include increased
automobile fuel efficiency, increased
use of alternative motor fuels or elec-
tricity, greater use of solar and nuclear
energy, improved efficiency in industrial
and residential energy use, and
decreased deforestation.
Proactive reforestation and afforestatio n
programs could also help to
counterbalance the rise in atmospheric
CO 2 levels (Dixon et al. 1991a). The
goal of reforestation and afforestation
programs would be to improve the
carbon sink potential of terrestrial
ecosystems. For this to occur, vast
areas of suitable land need to be
reforested. In the policy scenario
analyses by Lashof and Tirpak (1990),
reforestation was considered a
significant policy option in decreasing
the global climate change potential.
F. Forest management concerns,
constraints, and approaches
Although forests occupy approximately
36% of the world’s land surface, only
about 10% receive any type of man-
agement (Allan and Lanly 1991,
Goodland et al. 1990, Poore et al. 1990,
WRI 1990). Obviously, natural
resources management could play a
much more important role in the global
forest sector. In addition, the expansion
or intensification of management prac-
tices could improve the goods and ser-
vices obtained from global forests. One
critical service derived from the forest
sector, that needs more management
focus within the context of pending
global climate change, is the ability of
trees to sequester and conserve atmo-
spheric carbon (McG Tegart 1992).
Management strategies that would
have immediate benefits in mitigating
the atmospheric CO 2 concentration are
those that promote a decreased rate of
global deforestation and an increased
of afforestation or reforestation of
degraded lands capable of sustaining
forest tree growth (Dixon and Andrasko
1992). Global deforestation has
resulted not only in a reduction of
forested land, but has also decreased
plant and animal biodiversity, increased
desertification including loss of plant
cover and increased soil erosion, and
released vast amounts of CO 2 into the
17
-------
Section 2
atmosphere (Maini 1991, NAS 1991).
Given the overall decline in forested
areas resulting from tree harvesting and
land use change, interest in technically
managing forests for sustainability of
goods and services has gained
momentum among natural resource
managers and policy makers. Slowing
the rate of deforestation and expanding
reforestation projects throughout the
world have been proposed as cost-
effective means to sequester carbon,
protect soils and watersheds, and pro-
vide a continual source of chemicals,
food, fiber and fuel (Dixon et al. 1991b,
Lugo and Wisniewski 1992).
For sustainable forest management to
be successful in mitigating the buildup
of atmospheric CO 2 through carbon
sequestration, it is essential that it meet
the social and economic needs of the
local people (Gregerson et al. 1989).
Forests, especially those located in
tropical regions, have long been of
economic importance and thus influ-
enced greatly by anthropogenic use
and pressure (Gregerson et al. 1989.
Schneider 1989). Millions of hectares of
forest land around the world are
impacted annually by various activities
such as lumber production, livestock
grazing, or fuel wood harvesting (WRI
1990). However, intact, undisturbed
forests may provide more value to the
local economy in terms of food, fiber,
chemicals, and biodiversity than
degraded forest land (Peters et al.
1989). Economics may provide the
strongest argument for sustainable for-
est management practices, at least at
the local level, where benefits are
directly realized by the community. This
approach of sustainable forest man-
agement could lessen the pressure for
further global deforestation (Sanchez
1990).
Forest scientists have proposed various
strategies for using forest management
techniques in adaptation to and
mitigation of potential climate change
(Andrasko 1990a, McG Tegart 1992,
Trexler 1991b). Examples of such
recommendations are as follows:
• Create financial and policy incen-
tives to reduce rates of forest con-
version for unsustainable agricul-
ture, grazing, and forest harvesting.
• Reduce the frequency, interval,
scale, and amount of forest and
savannah consumed by biomass
burning which is intended to create
or maintain pasture and grassland.
• Introduce forest management sys-
tems which utilize low impact har-
vesting methods, replacing destruc-
tive high impact logging practices.
• Conserve standing primary and old-
growth forests as stocks of biomass,
offering a stream of economic ben-
efits.
• Substitute technically sustainable
technologies for slash-and-burn
agriculture that requires deforesta-
tion.
• Decrease consumption of forests
and trees for cash crops and devel-
opment projects through environ-
mental planning and management.
• Promote more complete utilization of
forest products.
• Promote the recycling of forest
products.
• Discourage the production of dis-
posable forest products by substitut-
ing durable wood or other goods.
• Prevent loss of soil carbon by slow-
18
-------
Section 2
ing erosion in forest systems caused
by harvesting and livestock grazing.
• Improve forest productivity on exist-
ing land through management and
genetic manipulation.
• Establish plantations on surplus
cropland and urban lands.
• Restore degraded forest and savan-
nah ecosystems through natural
regeneration, reforestation, and
protection.
• Promote opportunities to increase
soil carbon storage by leaving slash
from tree harvesting.
• Improve the efficiency of biomass
(fuel wood) combustion in cooking
and heating stoves and industrial
uses.
G. Conclusions
The general nature of the global carbon
cycle is known. However, uncertainties
exist in estimating carbon pools and the
transfers among them in relationship
with anthropogenic CO 2 emissions.
The region- or country-level approach
to identify carbon sources and sinks will
aid in validating the global-scale anal-
ysis and provide the basis for policy
decisions regarding anthropogenic
carbon emissions and carbon seques-
tration. Until the time when the relation-
ships among anthropogenic CO 2
emissions and biological carbon sinks
are understood, there will be unre-
solved scientific issues that will hinder
the policy making process.
19
-------
Section 2
20
-------
SECTION 3.
A. Base year carbon pools
A comprehensive estimation of the
magnitude of carbon pools on US tim-
berland (Box 5) was conducted using
yield tables associated with the
Aggregate TimberLand Analysis
System (ATLAS) and existing forest
inventory databases. ATLAS is a timber
inventory model developed for national
timber resource assessment by the
USDA Forest Service (Section 3.B,
Mills and Kincaid 1992). ATLAS con-
siders only growing stock volume. The
yield tables in ATLAS were used as the
basis for constructing stand-level car-
bon budgets which treated all major
carbon pools. Carbon pool size per unit
area per age class was thus estimated
for the living tree biomass, woody
debris, uriderstory vegetation, forest
floor, and soil components of the forest.
Separate stand-level carbon budgets
were prepared for each of 422 yield
tables supplied by the USDA Forest
Service as input to ATLAS. Carbon
storage at the regional and national
level could then be determined by
coupling these stand-level carbon bud-
gets to forest inventory data on the
areal extent and stocking volume of
each age class within each inventory
type. The set of stand-level carbon
budgets provides the foundation for
what is referred to in this document as
the Forest Carbon Model (FCM).
1. Construction of stand-level carbon
budgets
Living tree biomass
The construction of a stand-level car-
METHODS
bon budget began with an ATLAS yield
table which related stand age to the
volume of growing stock (i.e., mer-
chantable wood) at normal stocking
levels. Classification of the ATLAS
tables is by region, forest type, site pro-
ductivity, ownership and management
scheme. Age class increments in the
tables are 5 or 10 years, depending on
the region. The maximum stand age
may reach 90 to 175 years, depending
on region, forest type, and management
practices.
Since the ATLAS yield tables include
only growing stock volume, an adjust-
ment factor was applied to account for
noncommercial species (1.01 for soft-
woods, 1.14 for hardwoods; see
Section 9.C). Noncommercial species
include those of “small size, poor form,
or inferior quality which normally do not
develop into trees suitable for industrial
wood products” (Waddell et al. 19B9).
Adjusted merchantable bole volume
was converted to bole carbon based on
the relative proportion of hardwood and
softwood volume by age class, and for-
est-type-specific volume to carbon
ratios for softwoods and hardwoods
(Table 3.1, Birdsey 1992c). The subse-
quent conversion from merchantable
bole carbon to total tree carbon was
based on the analysis of Harmon
(Section 9.C). His analysis used har-
vest statistics and allometry to account
for the contribution of roots, stumps,
branches, and tops (Column 2, Table
9.C.3).
Sapling carbon (trees less than 12.5 cm
dbh) was included in the tree carbon
pool estimates reported here. Carbon
contributed by saplings was derived
21
-------
Section 3
Box 5. Definition of Timberland and Woodland
A primary functional classification of forest
land in the US differentiates between timber-
land. reserved timberland, and other forest.
Waddell et al. (1989) define timberland as:
Forest land other than timberland and
reserved timberland. It includes available
and reserved unproductive forest land.
which is incapable of producing annually
20 cubic feet per acre per year of industrial
wood under natural conditions because of
adverse site conditions such as sterile soils,
dry climate, poor drainage, high elevation,
steepness. or rockiness.
Forest land that is producing or is capable of
producing crops of industrial wood and not
withdrawn from timber utilization by statute
or administrative regulation. (Note: Areas
qualifying as timberland are capable of pro-
ducing in excess of2O cubic fret per acre per
year [ 1.4 m 3 ha 1 yr lj of industrial wood in
natural stands. Currently inaccessible and
inoperable areas are included.)
Throughout this document, the term
“woodland has been used to refer to other
forest land.
Reserved timberland refers to land having the
productive capacity of timberland but which
is withdrawn from timber utilization by
statute or administrative regulation.
Other foresi land includes:
The analyses reported in this document are
mostly confined to timberland forest,
including reserved timberland. These lands
account for the vast majority of the forest
biomass and forest productivity of the con.
lerminous US. Woodland acreage and
biomass values are described in Box 6.
22
-------
Section 3
Table 3.1. Factors to convert tree volume (cubic feet) to carbon (kg). Conversion
factors from Birdsey (1992a).
Spec ic Grav y 1
9 (wood)/cc (wood)
Percent Carbon 2
g-C/g (wood)
Factor 3
kg-Cltt 3 (wood)
Region Forest Type Softwood Hardwood
Softwood Hardwood
Softwood Hardwood
Southeast Pines .510 .639 .531 .497 7.66 9.01
and Oak-hickory .536 .639 .531 .479 8.07 901
South Central Oak-pine .523 .639 .531 .497 7.88 9.01
Bottornland .460 .580 .531 .497 6.93 8.18
hardwoods
Northeast Pines .378 .543 .521 .498 5.59 7.67
and Spruce-Fu .369 .525 .521 .498 5.45 7.41
Mid Atlantic Oak-hickory .374 .636 .521 .498 5.53 8.98
Maple-beech- .384 .600 .521 .498 5.67 8,48
birch
Bot lomland .460 .580 .52 1 .498 6.80 8 18
hardwoods
North Central Pines .421 .530 .521 .498 6.22 7.49
and Spruce-Fir .351 .480 .521 .498 5.19 6.78
Central Oak-hickory .416 .632 .521 .498 6.15 8.93
Maple-beech .372 .576 .521 .498 5.50 B 14
Aspen-birch .370 .465 .521 .498 5.47 6.57
Bottomland .460 .580 .521 .498 6.80 8 18
hardwoods
Rocky Douglasfir .473 .380 .512 .496 6.87 5.35
Mountain Ponderosa pine .416 .380 .512 .496 604 535
and Fir-spruce .349 .380 .512 .496 4.45 4 85
Pacific Coast Hemtock-sftka .434 .433 512 .496 5.53 5.53
sp.
Lodgepole pine .423 .380 .512 .496 5.39 4 85
Larch .508 .433 .512 .496 6.48 5.53
Redwoods .416 .580 .512 .496 5.31 7,40
Hardwoods .424 .384 .512 .496 5.41 4.89
Weighted average specWic gravity of the three most common (in terms of volume) softwood or hard-
wood species within the forest type.
2 FromKoch(1989).
3 Factor • specific gravity times the volume of a bic foot (28 36 x 1O cc) times percent carbon
23
-------
Section 3
from the ratio of bole carbon to sapling
carbon. These factors, specific to indi-
vidual forest types, were based on
regionaf relationships described by
Cost at al. (1990) and were dependent
on the predominant stand fiber type.
FCM treats the ratio for total tree carbon
to merchantable bole carbon as a con-
stant by forest type; in fact, the ratio
varies with stand age. The ratio is high-
est early in stand development.
Quantitative analyses in forest planta-
•tions suggest that after canopy closure
the ratio is relatively stable. As a result,
FCM underestimates total tree carbon
in young stands. But because the abso-
lute volume in young forests is low, the
net effect of not accounting for the
change in this ratio is small. In old
stands, use of the fixed ratio would tend
to overestimate tree biomass because
the mass of branches tends to remain
constant while the stem volume contin-
ues to increase. However, checks of
biomass estimates for older age
classes against biomass studies
reported in the literature (e.g., Cannell
1982) did not reveal a strong bias. In
future development of the FCM, this
ratio and the wood density may be
allowed to vary with stand age.
Woody debris
The woody debris pool consists of
standing dead trunks, dead coarse (>2
mm diameter) roots, and dead woody
material of greater than 2 cm diameter
lying on the forest floor. Because there
has been no consistent treatment of
woody debris in the USDA Forest
Service Forest Inventory and Analysis
(FIA) studies, we used an approach
based on an analysis by Harmon
(Section 9.C) to estimate woody debris
pool sizes. The woody debris pool for
each age-class was modeled as the
woody debris pool at age 0 plus an
increment due to mortality (Table
9.C.2), and minus a decrement due to
decay (Table 9.C.4). The woody debris
pool at age 0 was estimated differently
for western and eastern forests.
Western forests have undergone rela-
tively few rotations; thus, we assumed
that the woody debris from the primary
forest was still present. The woody
debris pool at age 0 was estimated as
the steady state pool for a mature forest
(Table 9.C.5), plus an increment due to
residue from live trees left on-site after
harvest (Table 9.C.3.).
Because many eastern forests have
been harvested several times, we
assumed that much of the debris from
the primary forest has decayed. Woody
debris at age 0 was estimated by an
iterative process. The steady state
value for mature forest (Table 9.C.5)
was used as an initial seed ”. The
debris values for successive age
classes were estimated by increment-
ing and decrementing this value for
mortality and decay, assuming current
stocking levels. The value at age 0 was
then recalculated as the estimated
debris value at rotation age plus an
addition due to harvest residue. This
procedure was iterated until woody
debris at age 0 converged to a stable
value.
Understory carbon
The understory carbon pool consists of
herbs and shrubs. The pool sizes were
estimated on a forest-type-specific
basis and were based on relatively few
examples in the literature (Plantinga
and Birdsey, in press). The pool size
rises rapidly after harvest or distur-
bance but falls as the canopy closes. It
eventually begins rising again as
stands mature and gaps occur in the
canopy. Understory carbon represents
only about 1% of total forest carbon, as
described in Section 4.A. Though there
24
-------
Section 3
are only lirn ted data to support the
biomass trends applied here, related
errors will have very little impact on
stand-level analyses.
Forest floor carbon
Estimates of initial forest floor carbon
and the age-specific increases in its
pool size are based primarily on Vogt et
al. (1986). Values rise gradually as the
forest stand ages. The estimates for for-
est floor carbon of fully-stocked stands
are also the same as used in Plantinga
and Birdsey (In press). However, in
FCM the pool size is adjusted propor-
tional to the stocking level in the current
inventory.
Soil carbon
To arrive at representative soil carbon
pools for each forest type, geographic
information systems technology was
used to overlay the spatial distribution
of each forest type with a map of soil
carbon. Estimates of initial soil carbon
within each region were based on the
soil carbon map developed by Kern
and Johnson (1991). In that study, data
from over 3,000 pedons on record at
the US Soil Conservation Service
Lincoln Laboratory were used to esti-
mate mean soil carbon at the great
group laxonomic level. The geographic
distribution of the great groups was
derived from the 1982 National
Resource Inventory (SCS 1987).
Forest-type distributions were taken
from a digitized version of the Society of
American Foresters map of the major
forest types (Eyre 1980). The map was
developed from indicators of potential
vegetation and USDA Forest Service
inventory data. The spatial resolution of
the forest type and soils maps is rela-
tively coarse. Nevertheless, the esti-
mates of soil carbon pools for each for-
est type appear to be consistent with
the geographic extent of species, with
regard to climate, and earlier studies of
soil carbon distribution in relation to
climatic factors (Post et al. 1982).
The changes in soil carbon to be
expected after stand initiation depend
in part on stand origin. In the case of
conversion from cropland or pasture to
forest, there is likely to be a slow
increase in soil carbon storage. The
vast majority of current stands in the US
originated from some form of forest
disturbance, especially harvesting.
Earlier carbon models designed for
analysis of land use change over large
geographical areas (e.g., Moore et al.
1981) have used a 20% or more
decrease in the “soil” carbon pool after
harvest. However, the soil pool was
actually composed of soil, litter, and
woody debris in those models. Recent
studies in temperate forests do not indi-
cate a significant decrease in soil
organic carbon after tree harvest
(Johnson 1992a, 1992b). Thus, in the
FCM stand-level carbon budgets soil
carbon remains constant. The net effect
of our approach is comparable to
Moore el al. (1981) in that the sum of
carbon in soil, litter, and woody debris
decreases significantly after stand
initiation.
Examples of the stand-level carbon
budgets which were developed using
the above approach are given in Figure
3.1. Immediately after harvesting,
woody debris is the largest pool. After
one to two decades woody debris has
declined and the tree carbon pool has
surpassed it. The most rapid accumula-
tion of tree carbon occurs in the
Southeast region but the maximum tree
carbon storage is lower there than in
other regions, especially the Pacific
Northwest.
25
-------
so
j’ pu. carbon
iO
so
2 0
I D — — — — — — — 2 0d74,b,1.
floor
S — • -
4
13
Is
I’
I D
F’
4
PHW Doug si—F r
Carboo PooLs (kg ro 2 )
a
1:
!4o
0
0
Jso
Section 3
NE M,ple—Seeeh—Birch
Corbon Pools (kg 3_Z)
rs. carbon
I0
.0
S
0
I
to
1.)
__..... . . ————i.r.,t IIOOT
. .-—-:_ - V.od dubni
-
C 50 75 tOO 125 150 175
• 25 50 P5 tOO 125 150 175
SE Plan .d—Ptne
Carbon Pools (kg m 2 )
irs, cOzboD
3
‘ ‘I
‘I __ _ Toreit floor
I / — — dubni
Underitory
0 so so so so
Stand Age (Ys.r)
Figure 3.1. Examples of stand-level carbon budgets toT three forest types using the
FCM. Abbreviations are as follows: PNW = Pacific Northwest, NE = Northeast, and SE
= Southeast.
26
-------
Section 3
2. Forest inventory databases
Private lands
A dataset for the forest inventory on pri-
vate lands in 1990 (the base year) was
generated by the coupled
TAMM/ATLAS model described in
Section 3.C. The survey data used to
generate the forest inventory input to
ATLAS was the same survey data used
to compile the 1989 RPA Assessment
(Mills 1988, Haynes 1990) and includes
information on forest type, stand age,
ownership, productivity level, and tim-
ber management intensity. There is a
direct linkage of ATLAS inventory cate-
gories to specific stand-level carbon
budgets based on ATLAS yield tables.
ATLAS also carries information on cur-
rent growing stock volumes.
For the eastern US, the RPA survey
data was collected by the USDA Forest
Service FIA work units at state and
county levels for all ownerships. In con-
trast, data collected by the FIA work
units at state and county levels in the
western US have been limited to non-
federal lands. A double sampling
scheme (Cochran 1977) is used in the
F1A surveys, whereby owner group,
land class, and forest condition class
are identified on photo points and the
photo sample is subsampled with a grid
of field plots. The field plots are the
basis for information on timber volume,
growth and mortality, and attributes
such as forest type, site class, and
stand-size class. Uncertainty in these
surveys related to sampling intensity
and frequency are discussed in Section
9.A.
Public lands
The inventory data developed for
ATLAS contained age class information
for private lands only. To estimate car-
bon pools on public timberland, it was
necessary to generate age class spe-
cific area and volume data which could
be processed with the stand-level car-
bon budgets previously described.
Information on the areal extent of forest
types on public lands and the total
growing stock volume was available
from the 1 987 RPA database. Data on
the age class distributions on public
lands within each forest type exist for
some particular areas, but have not
been compiled on a national scale.
Additional inventory work and database
augmentation would improve future
analyses. Satellite remote sensing has
begun to be applied for this purpose
(e.g., Tepley and Green 1991); how-
ever, there is as yet no published sys-
tematic treatment of public lands.
Lacking a national inventory for age-
class distributions on public lands, sev-
eral assumptions were employed for
the purposes of this initial carbon bud-
get development. Within the Southeast,
South Central, Northeast and North
Central regions (Figure 3.2), the area of
public timberland is relatively small
(Figure 3.3). Age class distribution for
public timberlands in these regions was
assumed to be equivalent to the age
class distribution on private timberland
for the corresponding forest type and
region. Total volume of public lands
reported in the 1989 RPA Assessment
by forest type and region (Waddell et al.
1989) was allocated among age
classes in direct proportion to the
ATLAS inventory data for private lands.
For each forest type represented on
public lands, the corresponding stand-
level carbon budget from ATLAS was
applied regionally.
In the Rocky Mountain, Pacific
Northwest East, and Pacific Southwest
regions, the area of public timberland is
quite large; however, the average
27
-------
Section 3
PACWIC
w T
PA:rIc
$OuThwE3i
PA:IP.C
NO TWWES1
Figure 3.2. Definitions of regions used in assessment of the US forest sector carbon
budget. Abbreviations are as follows: PNWW = Pacific Northwest West, PNWE —
Pacific Northwest East, PSW — Pacific Southwest, RM — Rocky Mountain, SC — South
Central, SE South Southeast, NC North Central, and NE — North Northeast.
ROCKY MOUP(TAIN
wopTh NT AL
NO
$3L W
28
-------
Section 3
PNTV
PN.n
Pub I
p ’
Pub I
P,t
Pub
M Pub
•C Pub
Nt
U Pub
Nt
NC Pub
Nt
NE Pub
0
Figure 3.3.
ownership.
ownership,
I I
10 20 50 40 00
Forest Ares (he x 10’)
Aveii.ble Reserved
Base year distribution of available and reserved timberland by region and
Abbreviations for regions are the same as in Figure 3.2, Pub z public
Pvt private ownership.
29
-------
Section 3
volume per unit area is low relative to
other regions. Lacking any consistent or
comprehensive age class data, the
assumption of equivalent age class
distributions for public and private lands
of identical forest type was again
invoked by region. Stand-level carbon
budgets were similarly chosen.
In the Pacific Northwest West region,
recent state level studies collated age
class data for national forests and other
public lands (Oregon, Sessions 1991;
Washington, Adams et al. 1992,
MacLean et al. 1991). This information
was used to estimate age class distri-
butions and stocking levels on public
lands and, as before, the stand-level
carbon budgets were used for estima-
tion of carbon pools. Total carbon data
for the oldest age class in the stand-
level carbon budgets of this region are
comparable to those reported in studies
of old growth forests (e.g., Grier and
Logan 1977), so areas of age classes
greater than 175 years were assigned
to the 175-year age class.
B. Base year carbon flux
The reference state in this analysis
must be the net accumulation or loss of
carbon from the forest land base over a
one year period due to factors associ-
ated with the forest sector. The term flux
refers to carbon transfer between the
forest and the atmosphere in either
direction or transfer out of the forest for
human use. In an effort to isolate the
effects of harvesting, we have distin-
guished between biologically-driven
carbon flux and harvest-driven carbon
flux. We have also estimated the
change in carbon stored in forest prod-
ucts which are still in use and still in
landfills.
1. Biologically-driven carbon flux
The biologically-driven carbon flux is an
estimate of the change in total carbon
storage expected over a 1-year period
assuming there had been no harvest.
This flux for individual forest compo-
nents, and total flux which reflects net
changes in all the pools, was based on
the inventories and stand-level carbon
budgets described earlier. The estimate
of net carbon uptake or loss on an
annual basis for a given age class was
simply the change in the carbon pools
over a specified interval, divided by the
number of years in the interval (e.g., 5
years in the Southeast region and 10
years in the Pacific Northwest). Initial
stocking levels and expected adjust-
ments to the stocking levels, which are
prescribed by ATLAS, were allowed to
impact uptake rates. Within each forest
type, the area within each age class of
each relevant stand type was then mul-
tiplied by the associated carbon flux
rate per unit area. The estimate of
annual carbon uptake over a region
was an aggregate of the uptake for all
forest types within that region.
The approach used here does not
account for woody debris decomposi-
tion on lands which have been defor-
ested and are no longer in the forest
inventory. This factor may not be large
in the US which has a relatively stable
forest land base. However, in regions or
countries of rapid deforestation, the for-
est inventory should, in principle, carry
lands deforested in previous decades
and include this source of C02.
2. Harvest-driven carbon flux
In addition to the biologically-driven
carbon flux, the flux resulting from har-
vesting and commercial thinning was
also considered. For the areas har-
vested in the base year, an accounting
30
-------
Section 3
was made for the carbon in all pools.
The volume of growing stock and non-
growing stock removals on private and
public lands for the base year was
derived from the 1989 RPA Assessment
and is comparable to the removals
reported in Waddell et at. (1989) and
Haynes (1990). Conversions from vol-
ume to carbon were made using the
factors in Table 3.1. The harvested car-
bon was removed from the forest inven-
tory and represented a large offset to
the biologically driven flux. The residue
from harvest which is left on site was
modeled as a transfer of carbon from
the living tree biomass to the woody
debris pool. Thus, there was no effect
on total carbon storage. In practice
some of this carbon might be burned as
slash; however, rather than attempt to
account for rates of slash burning, we
have let the residue decay as part of the
woody debris pool. This approach pro-
vides assurance that this carbon is
consistently accounted for.
The change in carbon in the understory
and the forest floor of harvested stands
was treated as instantaneous since it
Consists of rapidly decomposing mate-
rial. Thus there was a small carbon
pulse from the forest land base associ-
ated with the reduction of understory
and forest floor in the harvested areas.
In order to estimate this carbon pulse,
the distribution of the harvest among
forest types and age classes was
needed. On private lands the harvest
volume was distributed across forest
types and age classes by ATLAS
(described in Section 3.A). On public
lands, the RPA Assessment (Haynes
1990) provided harvest volume by
region but no information on the distri-
bution among forest types and age
classes.
For this analysis, we assumed an
equivalence in the distribution of har-
vest across forest types and age
classes between public and private
lands, except for Pacific Northwest
forests west of the Cascades. In that
region, we used data from Sessions
(1991) on the distribution of the harvest
among age classes on public lands.
3. Other changes in carbon pools
If none of the harvested carbon
returned to the atmosphere, it could be
considered as a carbon sink in addition
to whatever net change occurred on the
forest land base. However, this carbon
is returned to the atmosphere eventu-
ally, much of ii rather rapidly. The actual
sink associated with the forest harvests
is only the change in the pool com-
posed of products still in use and prod-
ucts in landfills but not yet decomposed
or burned.
The change in the size of the two prod-
ucts pools is a function of annual inputs
and outputs. The annual change in
these pools was evaluated by reference
to literature-based estimates of the pro-
portion of the harvest going into prod-
ucts, the annual transfer of wood prod-
ucts into landfills, and the annual car-
bon flux out of landfills.
C. Projected US forest sector
carbon pools and flux
The approach to estimating future car-
bon pools and flux relies on inventory
projections provided by models which
focused on forest economics and inven-
tories. The inventory at any point in time
is run through the Forest Carbon Model
to estimate total forest carbon. The
change in total forest carbon over a 10-
year interval, divided by 10, gives the
average annual flux for thai interval.
31
-------
Section 3
1. Use of the TAMM and ATLAS
models for private timberlands
Timber is a unique commodity in that its
production period requires more long-
term planning than virtually any other
renewable resource. Informal assess-
ments of the timber situation date from
1866 in the US. These were formalized
by the Forest and Rangeland
Renewable Resources Planning Act of
1974 (RPA) which calls for periodic
long-range assessments of timber
supply and demand to provide informa-
tion for stewardship and management
decisions in both the public and private
sectors. Starting in the 1960’s, the
USDA Forest Service began develop-
ing more formal planning frameworks
for timber assessments.
The basic approach was a trend analy-
sis of both demands for forest products
and availability of timber resources
(USDA Forest Service 1965, 1974, and
1982). Policies and emerging trends
were discussed in the context of the
gap between the trajectories of the
demands for forest products (expressed
in roundwood equivalents) and the
prospective availability of timber. In
1977, the USDA Forest Service under-
took the development of an assessment
model designed to explicitly explain
regional stumpage price behavior. This
model was built around a spatial equi-
librium algorithm. 1 The initial purpose
was to develop regional harvest and
1 The emphasis on regional markets meant that
the model had to deal with the spatial
juxtaposition of stumpage markets. Samuelson
(1952) is generally credited with defining the
problem, but until the last decade its solution has
always involved complex computational
methods. Previous experience with linear
programming models, both in agriculture and
forestry, led model builders in the US to select a
computationally simpler algorithm (a Reactive
Programming algorithm developed by King and
Ho (19721) to solve for spatial equilibrium.
price trends for the 1979 RPA Timber
Assessment. The resulting model,
called TAMM, exceeded expectations
(Adams and Haynes 1980, Haynes and
Adams 1985). It produced regional
price trends in both product and factor
(stumpage) markets that recognized
simultaneous market interactions and
differences in regional timber
resources. This model built upon early
econometric studies and linear pro-
gramming efforts that attempted to solve
spatial market concerns.
TAMM provides an integrated structure
for considering the behavior of regional
prices, consumption, and production in
both stumpage and product markets. It
is designed to provide long-term pro-
jections of price, consumption, and pro-
duction trends and to simulate the
effects of alternative forest policies and
programs. To a far greater extent than
was possible in the past, the use of
TAMM has focused attention on the
dependence of projections on
(exogenous) input assumptions. In
Forest Service assessments these
assumptions concern the major deter-
minants of the supply and the demand
for various forest products.
This section provides broad details
about how TAMM operates. Details
about the various input assumptions
used in TAMM are described briefly in
the next section and in more detail in
the 1989 RPA Timber Assessment
(Haynes 1990).
The general structure of TAMM is
shown in Figure 3.4. It is composed of
ten major parts with exogenous inputs
from four other models (Pulp, Trade,
Fuelwood, and Area Change). Imports
are explicitly treated by including
Canada in the model. Briefly, product
demand, such as softwood lumber, is
obtained by multiplying the ratio of
32
-------
Section 3
HARDWOOD SAWTIMBER MARKET
SOFTWOOD DEMAND APPROXIMATION
CAPACITY ADJUSTMENT
w PRODUCT/STUMPAGE SUPPLY EQUATIONS
PULP FIBER REQUIREMENTS *— PULP MODEL
MARKET SOLUTION 4 TRADE MODEL
z
REMOVALS COMPUTATION 4— FUELWOOD MODEL
GROWTH-DRAIN INVENTORY COMPUTATION
AT 5-10 YEAR INTERVALS TRANSFER
CUMULATIVE HARVEST TO ATLAS
ATLAS - 4 AREA CHANGE MODEL
I
Figure 3.4. General structure of TAMM.
33
-------
Section 3
product use per unit of activity (such as
the number of housing starts) times the
number of units and summing these
results over all the various end-uses for
the product. The activity measures are
exogenous and are generally taken
from long-term macro forecasts pre-
pared by Wharton Econometrics
Forecasting Associates (1987). An
example demand function (ignoring the
time subscript) would be given by the
expression:
C = F 1 (P DSI.u •7
Pu i U!
U
where:
C 1 is the consumption for product i,
(1)
F u represents a system of equa-
tions whose solutions give the
end use (u) factors as a function
of their several arguments.
P, the price of product i,
is a vector of prices of substi-
tute materials,
Z is a vector of other shifters, and
J is a measure of activity in end
use ii.
The demand relationships are regional-
ized by accounting for differences in
regional per capita consumption and
the ratio of regional prices to national
prices. Hardwood lumber is treated on
the demand side in about the same end
use detail as softwood lumber but con-
sumption and production are set equal
and price is determined as a function of
softwood lumber prices. l’iardwood
sawtimber 2 stumpage prices are a
2 Sawhmber stumpage Includes live trees of
function of hardwood lumber prices.
The product supply equations are esti-
mated in the form:
$ = (p - avc) k
where
(2)
$ is the supply volume,
p is the product price,
avc is the average variable cost, and
k is the installed capacity.
Average variable costs are composed
of wood and nonwood elements that
have been converted to a unit output
basis by means of product recovery and
productivity factors that reflect the cur-
rent (or projected) technology of the
industry. Each product supply function
includes installed capacity as an inde-
pendent variable. Shifts in installed
capacity are modeled as a function of
anticipated changes in relaVve regional
profitability or rate of return. Timber
supply is modeled as a function of
stumpage price and timber inventory.
Finally, the timber demand functions
are derived from product market
demand and supply functions. The
pulp fiber requirements are input vari-
ables from a pulp model developed by
Ince, Durbak, and Howard (in press).
Trade and fuelwood projections are
also input variables 3 . The trade projec-
tions reflect a future where the US
remains a net importer of softwood for-
est resources.
The remaining steps in the annual cycle
commercial species containing at least one 12-
foot sawlog or two noncontiguous 8-foot logs.
and meeting regional specifications for freedom
from defect. Softwood trees must be at least
90-inches dbh, and hardwood trees must be at
least 11.0•inches dbh.
3 See the 1989 RPA Timber Assessment for
further details (Haynes 1990).
34
-------
Section 3
of TAMM involve timber supply and an
inventory projection system (ATLAS). 4
The methods used to project timber
supplies require a number of assump-
tions that include the form of the inven-
tory projection system, projections of
timberland area change, assumptions
linking harvest to removals from inven-
tory, and projected harvest flows from
public timberlands. The basic economic
representation of timber supply
describes the supply of timber at any
point in time as being a function of the
private timber inventory levels,
stumpage prices, and the amount of
public harvest available at that time.
The Aggregate Timberland Assessment
System (ATLAS) was used to make the
inventory projections on all private tim-
berland in the US. It is an age-based,
yield-table system that projects acres by
detailed strata for periods consistent
with the inventory stand-age classes. It
evolved from earlier work by Beuter et
a!. (1976) an Tedder et a!. (1987). In
the model, the inventory is represented
by acreage cells classified by region,
ownership, management type, man-
agement intensity, and age class. In the
South and the Pacific Northwest West,
the strata were also identified by site
productivity class and stocking class.
4 This inventory projection system considers
only the timber inventory located on timberland
that is producing or is capable of producing
crops of industrial wood and not withdrawn from
timber utilization by statute or administrative
regulation Currently, areas qualifying as
timberland have the capability 01 producing in
excess ci 20 cubic feet per acre per year of
industrial wood in natural stands and may be
inaccessible and/or inoperable. Further, these
models include only live trees of commercial
species meeting specified standards of quality or
vigor. Cull trees are excluded. When associated
with volume, these growing stock inventories
include only trees 5.0-inches dbh and larger.
The ATLAS model was developed by Mills and
Kincaid (1992) for the 1989 RPA Timber
Assessment.
This study used 10-year age classes in
all regions except the South where 5-
year age classes were used.
The ATLAS model starts with data
arrayed in the following fashion for
each relevant combination of region,
forest type, ownership, site productivity,
level and management intensity:
Age
Class
Acres
Volume/
Acre
Yield
Table
1
S
S
n
ad
•
S
ac
CFA 1
S
S
CFA
VT 1
S
S
VT 1 ,
where:
ac is
class k
the inventory acreage of age
CFAk is the average cubic feet of
growing stock per acre (volume)
inventoried for age class k
YTiç is the cubic feet growing stock
per acre of a fully-stocked stand
of age class k.
The basic time interval in ATLAS is set
by how age classes are specified. If, for
example, 10-year age classes are
specified then the time dimension in
ATLAS is in 10-year periods. That is, t
is year 1 but t+1 is year 10. In the equa-
tions that follow, period 1 would be the
first 10-years (except in the South
where it would be the first 5 years).
ATLAS uses the basic growth-drain
identity:
lflVt+1 =lnvt+gwtht- hVtt
where:
(3)•
35
-------
Section 3
IflVt+1 is the inventory volume in
period t+1 ,
Invi is the inventory volume in
period t
gwth is the growth in period t, and
hv’tt is the harvest in period t (This is
an input variable to ATLAS)
Timber inventory as used here is the
growing stock volume, defined as the
cubic volume of all live trees of com-
mercial species, meeting specified
standards of quality or vigor, which are
at least 5.0 inches in diameter (breast
height).
The various variables are computed in
a given period (t) as follows:
lnvk=ack CFAk
SRk = CFAkIYTk
SRk+1 = a + b SRk
ACHVTk = (hark. fMt)I(CFA c’ GHk)
gwth k Ct’Tk’ SRk)- (VTk.1 ‘SRk.l)
where:
SRk is the stocking ratio for the k
age class,
a,b are coefficients for stocking
adjustments.
ACHVT are the number of acres
harvested,
hark is the harvest proportion for the
k age class, and
GHk is the growth-on-harvest pro-
portion for the k age class.
Equation 6 adjusts the stocking per-
centages for stands over time, recogniz-
ing that stand density changes with
time. This adjustment is called
“approach-to-normality” and has been
used in conjunction with growth predic-
tions from normal (fully stocked) yield
tables since the 1940’s (Husch 1963).
In each simulation period, inventory
change is the result of growth, area
change, and timber harvest. Growth is
the result of acreage following upward
sloping net yield functions that are cal-
culated with approach-to-normal
assumptions. Each cell in the starting
inventory may have an independent
yield function. A major attribute of the
model is that it can simulate shifts in
management intensities, and the con-
sequent changes in yields based upon
alternative assumptions about the
future. The inventory model is not, in
principle, an even-age model since it
can simulate growth and removal pro-
cesses across all age classes. ATLAS
can also account for both partial har-
vests and commercial thinning. Volume
can be removed at harvest from either
the oldest age class or spread across
several age classes. Final-harvested
acres can be assumed to be regener-
ated in any management type or may
leave the timberland base entirely.
Partial cutting affects a proportion of
acres in a cell and the treated acres are
assumed to return in the cell growth
function to a younger age class.
Inputs to the inventory model include
estimates of harvest, acreage shifts,
and growth parameters. The levets of
harvest are derived through interaction
with TAMM. Area change information is
supplied by a modified version of the
Southern Area Model (Alig 1985). Yield
tables and approach-to-normal
parameters were derived from timber-
land inventory plot data collected by the
various USDA Forest Service FIA units
and from previous studies (the inven-
tory data inputs are summarized by
Mills 198B]).
(4)
(5)
(6)
(7)
(8)
36
-------
Section 3
The solution of TAMM represents a
spatial equilibrium in the markets mod-
eled for each year of the projection
period. These solutions do not repre-
sent, nor can the basic market solution
algorithm 5 be readily used to find,
intertemporal production or consump-
tion strategies that are in some sense
optimal. The production, consumption,
and price-time paths are only estimates
of outcomes of contemporaneous inter-
actions in freely competitive markets.
Details about Ihe various input
assumptions used in TAMM are
described in the 1989 RPA Timber
Assessment (Haynes 1990).
2. Methods used in the analysis and
treatment of public lands
As with information for the base year
analysis, data representing projected
changes in the public timberland inven-
tory by age class are limited. For
National Forests, process records from
some individual National Forest plans
contain projections by age class.
Process records are part of the back-
ground information maintained by
National Forests, as documentation
required by planning regulations, and
are used to monitor plan implementa-
tion. Many National Forest plans do not
display complete sets of projections of
harvest or timber inventory by age
class, and in some cases the informa-
tion is not readily extractable from
models used in developing plans. In
particular, different versions of the
Forest Planning Model (FORPLAN)
have been used across National
5 A revised reactive programming algorithm is
used (Brooks and Kincaud 1987). Briefly,
reactive programming is a method for solving
continuous demand and supply function by
successive ad ustmen1 of quantities produced
and their distribution to demand regions, so as to
maximize producer profits net of transport costs
in each supply region.
Forests, which complicates the type of
resource-related information provided
in planning efforts. An additional com-
plication is that National Forest plan-
ning and associated adjustments are
an ongoing process. With significant
developments such as the recent judi-
cial rulings on spotted owl set-asrdes,
harvest projections (and linked
attributes such as the inventory level on
timberland) for the National Forests are
in a state of flux.
In the present study, the most recent
projections of future removal rates and
forest inventories on National Forests,
as assembled by USDA Forest Service
staff at the Pacific Northwest Station,
were used for the scenarios described
below. These were produced on a
growing stock basis (m 3 ). The projec-
tions of timber inventories and harvest
levels for Other Public lands (i.e. non
National Forests) are the same for all
scenarios and are taken from the RPA
Assessment (Haynes 1990). Since only
information on growing stock volume
was available for the forest manage-
ment scenarios, volume was used to
estimate total carbon storage.
Unlike private lands, the conversion
from growing stock volume to total car-
bon inventory on public lands was
made on a regional basis by reference
to the ratios of growing stock volume to
total non-soil carbon (i.e., tree biomass,
understory, forest floor, and woody
debris). These ratios were taken from
the base year analysis (Section 4.A).
Soil carbon values were treated as
constant in the scenarios because of
the large uncertainties concerning vari-
ability of soil carbon in forests over time.
The projected timber volumes and total
carbon on reserved lands were most
problematical because inventory mod-
els do not exist for these lands. For this
37
-------
Section 3
analysis, the inventory on the unavail-
able public area for the base year was
allowed to grow in 10-year increments
and new carbon pools were calculated
by running the inventory through the
Forest Carbon Model.
The assumption of a fixed ratio of vol-
ume to total carbon has several limita-
tions with regard to assessing the
effects of harvest reductions. First, it
does not account for age class distribu-
tion effects. The ratios of volume to
woody debris and volume to forest floor
carbon vary with stand age. A second
problem concerns the treatment of
changes in carbon storage in old
growth stands. Because the yield tables
which provide the foundation of the
inventory approach to estimating car-
bon pools and flux reach a limit at 175
years, no increment occurs in these
older stands. Thus, no change in car-
bon storage is reflected over a decadal
period from all lands which remain in
the oldest class. As discussed in
Section 9.C, it appears that carbon
does in fact accumulate in both living
biomass and woody debris in these
stands.
Because of the limitations in the prelim-
inary USDA Forest Service inventory
projections and in using ratios to esti-
mate carbon from volume, an additional
approach was taken to evaluate the
effects of the Reduced National Forest
Harvest Scenario described below
(Section 3.C.4, Scenario 3). In that sce-
nario, a greater than 50% reduction in
the harvest was implemented for
National Forests in the Pacific
Northwest West region. The approach
for assessing the consequences of this
change was to determine the volume
not harvested, to estimate from that vol-
ume the area not harvested, and to
compare changes in carbon flux over
time for this area on the basis of harvest
and no-harvest scenarios.
Three potential differences in carbon
flux were considered in this analysis.
The first flux was the loss of the carbon
sink which would have been main-
tained if the trees had not been har-
vested. The magnitude of this sink
tends to drop after the rotation age. But
even 50 years after the rotation age, it is
still over half of the maximum rate
achieved early in the rotation for Pacific
Northwest West Douglas-fir stands. The
second flux was the on-site carbon
source created by the harvest. As dis-
cussed in Section 3.B, harvested
stands are carbon sources for about the
first 20 years in the Pacific Northwest
West because the decay of woody
debris dominates the carbon uptake
associated with tree growth. During the
next 30 years these harvested areas
begin to accumulate carbon. Lastly, the
off-site carbon source which would
have been created by the decay of
relatively short lived wood products
derived from the harvest was estimated.
The carbon which would have accumu-
lated in forest products in and Out of
landfills is assumed to be balanced by
carbon emissions from landfills.
Changes in volumes harvested were
taken from the TAMM/ATLAS run, with
the 1989 RPA Assessment as the refer-
ence harvest level. The initial year of
the analysis was 1995, when the major
harvest reductions were in place. The
land area harvested per unit volume
harvested was based on a ratio derived
for National Forest lands in Washington
and Oregon. Data were available from
several sources for harvest by forest
type and age class in these states
(Sessions 1991, MacLean et al. 1991).
The area not harvested was then
assigned to a stand-level carbon bud-
get associated with Douglas-fir on a
medium productivity site, as described
38
-------
Section 3
in Sections 3.A and 3.B.
The carbon sink associated with the
unharvesled area was modeled
through the year 2045. During each
year of the simulation, the new unhar-
vested acres were added to an inven-
tory, with an age class corresponding to
the rotation age (80 years). Then the
time course for the total carbon flux from
that time forward was used to estimate
the carbon sink associated with that
area, assuming it had not been har-
vested. Thus, over the 50-year simula-
tion, the carbon increments for the age
classes from 80 years to t30 years
were used.
To estimate the on-site carbon source
that was avoided by not harvesting, an
area equivaleni to the unharvested
area was added to a second inventory
with an age class corresponding to the
youngest age class, thus simulating
harvest. The first 50 years of the same
Douglas-fir stand-level carbon budget
was used to estimate the biologically-
based carbon flux which would have
been generated after the harvest.
The off-site carbon source associated
with decay of short-lived harvested
wood products, which would be
avoided, was estimated at 58% of the
volume harvested. This is an approxi-
mation of the proportion of the harvest
in the Pacific Northwest which returns
relatively quickly to the atmosphere
(Harmon et al. 1990). There may be
compensatory increases in harvest that
might occur on private lands but that
has not been considered here.
The net effect of not harvesting was cal-
culated as the sum of the on-site sink
which was maintained, the on-site
source which was avoided, and the off-
site forest product source which was
avoided. Each year over the 50-year
simulation, the net effect for all areas
not harvested relative to the RPA
Scenario was summed for the total net
effect.
3. Fate and disposition of carbon after
timber harvest
During processing, use, and eventual
disposal, all the carbon harvested from
forests is eventually returned to the
atmosphere as carbon dioxide or
methane by either burning or decay.
However, a large proportion of har-
vested carbon may remain in use in
products for varying lengths of time, or
accumulate in landfills where decom-
position rates are slow. To the degree
that the products and landfill pools are
increasing or decreasing, a sink or
source of carbon is created. It is impor-
tant to account for these carbon cycle
flows because their impact on the
overall US forest-sector carbon budget
could be significant.
The model HARVCARB (Row and
Phelps 1990) was used to examine
distribution and transfer of post-harvest
carbon quantities predicted by the
TAMM/ATLAS analyses of policy alter-
natives. The model traces the flow of
forest harvest carbon through five
phases of processing, use, and dis-
posal:
• From trees to roundwood logs taken
from the forest.
• Through processing into forest
products like lumber and paper.
• Into uses in construction, manufac-
turing, and paper products.
• To discard products (Construction
debris and municipal solid waste).
• Finally, to landfills, incinerators, and
ultimately, back to the atmosphere.
The first three phases occur within a
39
-------
Section 3
year or two. Once the products are
manufactured, some might remain
intact for years or decades, and might
be replaced or abandoned at a slow
rate, such as construction timber.
Others, such as most paper products,
may be used and discarded promptly.
After being discarded, the materials
might reside in landfills and debris
dumps for long periods.
The results of the HARVCARB analysis
perlain only to carbon harvested during
the analysis period beginning in 1980
and not to any carbon that was already
in products and landfills. It is difficult to
assess the actual incremental changes
in these pools because not only are
their pre-1980 sizes unknown, but the
fluxes of pre-1980 carbon between
products, landfills, and atmosphere are
also unknown. Some indication of the
impact of any given scenario can be
obtained by comparing the differences
between it and the reference scenario
in the carbon stored in products and
landfills.
4. The forest-sector scenarios
In this report we consider 12 forest
sector scenarios (Table 1.1). The sce-
narios ranged from the Cureent Forest
Plans, which is the base scenario and
reflects current best estimates of future
harvest levels on public lands, to sce-
narios that called for reduced harvests
on public lands, increased recycling, or
increased afforestation. The 1989
Resources Planning Act (RPA) scenario
was included because it represents the
most recent well-documented scenario
produced by the USDA Forest Service.
Differences between the RPA and the
Current Forest Plans scenarios are a
function of policies already in place.
The last three scenarios consist of
combinations of the assumptions asso-
ciated with some of the other nine sce-
narios. Each scenario was run using
the combined TAMM/ATLAS/FCM
modeling framework (Figure 1.2).
Scenario 1. The 1989 APA Assessment
Projection
The assumptions and projections of the
1989 RPA Assessment are discussed in
detail by Haynes (1990). For private
lands these assumptions are basically
the same as in the Current Forest Plans
Scenario, described subsequently.
The background assumptions for any
TAMM model run include projections of
future forest products markets (based
on forecasts of the major determinants
of product demand, and of timber and
product supply). On the demand side,
the major determinants are future
growth in aggregate economic activity,
as measured by gross national product
(GNP), and on certain key end-uses for
forest products, such as new residential
construction/repair and alteration of
existing dwellings. In the RPA Scenario,
the population of the United States is
expected to grow from 242 million in
1986 to 333 million by 2040, and con-
siderable improvement is expected in
worker productivity. The result is pro-
jected GNP growth rates of 2 to 3% over
the next 50 years. This portrays a strong
and resilient future economy, with an
increasingly affluent populalion. It sug-
gests continued strength in demand for
forest products, such as paper and
paperboard, that are used in a wide
range of sectors in the economy.
Residential construction particularly
consumes solid wood products.
Projections of the factors that determine
long-term demands for new housing
units (household formations, replace-
ment of units lost from the housing
stock, and the inventory of vacant units)
indicate continued high levels in the
early 1990’s with about 2.0 million
40
-------
Section 3
starts of all types of housing per year,
decreasing gradually after 2000 to
about 1.5 million starts per year. About
20% of the lumber and structural panel
producis and 15% of nonsiructural
panel products are used each year for
the upkeep and improvement of exist-
ing units. Residential upkeep and repair
have grown rapidly during the 1980’s
and are expected to continue growing
as the nation’s housing inventory and
the average age of its component units
increase. Key supply determinants
include levels of public harvest,
changes in manufacturing and logging
technology, and trends in pnvate tim-
berlands.
Scenario 2. Current Forest Plans
This scenario reflects current (1992)
knowledge about future timber supply,
demand and harvest on private and
public lands. As discussed below, it
includes some reductions in harvest on
public lands due to habitat preservation
and transfer of lands from available to
reserve classification. It is currently
considered to be the “base scenario” for
USDA Forest Service planning.
Because it differs from the RPA
Assessment Scenario, it is described in
detail below.
In the last two decades, it has become
increasingly clear that the future of tim-
ber production on the National Forests
depends in part on (1) success in find-
ing suitable ways to integrate timber
production with other uses of forest
land, and (2) the need to protect and
maintain the forest environment, includ-
ing endangered and threatened
species. The controversy surrounding
habitat protection for old growth species
such as the spotted owl illustrates the
increasing constraints on timber pro-
duction on the National Foresis. The
Current Forest Plans and the Reduced
National Forest Harvest scenarios are
presented to address these two issues.
The total National Forest harvest levels
for each case are shown in Table 3.2.
The 1989 RPA Assessment harvest
projection essential(y represents a con-
tinuation of the late 19B0’s harvest lev-
els. The increases after 2000 come
from modest increases of planned har-
vests in the Rocky Mountains. The
National Forest harvest projections for
the Current Forest Plans scenario pro-
jection assume the adoption of the
Interagency Scientific Strategy for
Protection of the Northern Spotted Owl
(ISC 1990) and the most recent
National Forest Plans (on file with each
National Forest). Additional factors
which are considered include a more
prolonged recession and changes in
log export restrictions for Western
states.
For the Current Forest Plans scenario,
National Forest softwood harvest is
projected to fall from the recent levels of
2.2 billion cubic feet to about 1.7 billion
cubic feet by 2000 and then rise to 2.0
billion by 2040. All of the downward
adjustment in the three Pacific Coast
slates results from the adoption of both
final forest plans and the conservation
strategy for the northern spotted owl
(1SC 1990). The combination of these
two changes, for example, is to reduce
National Forest harvest in the Douglas-
fir region by 51%. National Forest
hardwood harvest remains relatively
stable at about a biflion board feet per
year. Harvests from other public timber-
lands are projected to increase as
those inventories increase.
Projections of technological trends for
the Current Forest Plans scenario are
described by Skog et al. (1991) and
changes in the private timberland base
are described by AUg and Wear (1992).
In spite of the numerous changes, the
41
-------
Section 3
Table 3.2. Projected harvests (as growing stock) from National Forests for RPA,
Current Forest Plans, and Reduced National Forest Harvest scenarios.
Year
RPA Current Reduced
Assessment Forest Harvest National Forest Harvest
Projection Plans
billion cubic feet
2.16 2.16 2.16
1986
2000
2.17 1.72 1.36
2010
2.32 1.82 1.42
2020
2.40 1.89 1.47
2030
2.47 1.96 1.54
2040
2.53 2.01 1.59
general outlook remains roughly the
same suggesting that the 1989 RPA
Assessment projections are robust both
as a base for policy actions and as a
view of the future.
Assumptions about markets and prices
for manufactured products were as fol-
lows:
Lumber. Total lumber consumption in
1986 was 8.84 billion cubic feet. 6
Consumption of lumber is projected to
rise, reaching 10.58 billion cubic feet in
2040, as the use of softwoods in con-
struction (particularly upkeep and
alteration) and hardwoods in manufac-
turing and shipping continue to
increase (Figure 3.5). After 2020,
declining consumption in housing is
more than offset by continued growth in
manufacturing and nonresidential uses
for both hardwoods and softwoods.
National Forest harvest reductions lead
6 A 1 1 volumes reported for manufactured
products in this section are in roundwood
equivalent.
to increases in the level of Canadian
lumber imports, but in the longer term
rising delivered wood costs in Canada
lead to a decline in imports 10 1.98 bil-
lion cubic feet by 2040. Hardwood con-
sumption also increases steadily and
the sector continues to be a net
exporter.
Lumber production shares by region for
1952, 1992, and projected years are
listed in Table 3.3.
Softwood lumber production has shifted
among US regions during the last four
decades and is expected to continue to
shift. In ¶992, the South and the Pacific
Coast states accounted for 80% of the
softwood lumber production (down from
86% in 1952). During the 1950’s and
1960’s, the three Pacific Coast slates
were the major lumber producing
region, but starting in the 1970’s the
lead transferred to the South. Softwood
lumber production is projected to shift
from the Pacific Coast regions to the
South over the next five decades, due
to continued growth in stumpage costs
42
-------
Section 3
Table 3.3. Lumber shares (%) by region for 1952, 1992, and projected.
1952 1992 2000 2040
Percent
Pacific Coast
57 46 44 37
Rockies
8 14 15 15
North
6 6 6 8
South
29 34 35 41
in the Douglas-fir and Pacific
Southwest regions. Softwood lumber
production in the northern regions and
in the Rocky Mountains rises through-
out the projection period. The expan-
sion in production in the Rocky
Mountains results from increased re-
gional National Forest harvest levels
which lead to decreased stumpage
costs relative to other regions.
Structural Panel Products. Structural
pariets (softwood plywood and oriented
strand board, OSB) consumption rises
to 2.56 billion cubic feet in 2040, about
50% above 1986 consumption. Most of
the increase is due to continued growth
in OSB, which is projected to reach 1.0
billion cubic feet by 2040, more than 3.6
times its use in 1986. After slowly de-
clining through 2010, softwood plywood
consumption increases to 1.6 billion
cubic feet in 2040. By 2040. OSB will
compose over 40% of total structural
panel consumption, up sharply from
16.6% in 1986. Consumption of panels
is expected to increase across all end-
uses except for new housing.
Regional production of structural panel
products has undergone major
changes in the past. Southern softwood
plywood output has grown to 51% of
total production and the North has
emerged as a major producer of OSB
structural panels. Continued regional
shifts are projected in the next 20 years.
By 2000, softwood plywood production
will decline both in the South Central
and Douglas-fir subregions. OSB pro-
duction will nearly double in the North
during the same period. Between 2000
and 2040, panel production for both
softwood plywood and OSB expands in
all regions except for the Rocky
Mountains and regional production
shares stabilize.
Nonstructural Panel Products.
Consumption of nonstructural panels,
including hardwood plywood, insulating
board, hardboard, and particleboard
was 0.77 billion cubic feet in 1986. It is
projected to rise to 1.13 billion cubic
feet by 2040. Different trends in de-
mand are projected for the various
products. Little growth in insulating
board, whose major market is residen-
tial construction, is expected. Hardwood
plywood, used in manufacturing as well
43
-------
Section 3
60
50
C,
C,
C D
030
.0
.2
20
10
0
1950 1960 1970 1980
25
20
a,
a,
15’
10 0
5
0
1990 2000 2010 2020 2030 2040
Year
Figure 3.5. Lumber and structural panel products consumption, 1950-1988 with
projections to 2040.
44
-------
Section 3
as construction, increases slowly
through 2040, while particleboard rises
until 2010, but shows little growth after-
wards. Hardboard is the only nonstruc-
tural panel product to show a steady in-
crease throughout the projection pe-
riod.
Fiber Products. Consumption and pro-
duction of paper and board is projected
to increase, although not as strongly as
in the past. Total consumption will ex-
ceed 100 million tons in 2000 and 173
million tons by 2040. Imports exceed
exports over the projection period, with
production approaching 166 million
tons by 2040. Growth in per capita con-
sumption of paper and board slows, re-
flecting substitution of non-paper for
paper materials in packaging and con-
struction products and less rapid ex-
pansion in the consumption of white
paper.
Woodpulp (the major input for paper
and board manufacture) is projected to
increase 93% between 1986 and 2040
to 116 million tons. Imports of woodpulp
are projected to continue to exceed ex-
ports. At the same time, consumption of
wastepaper is projected to triple,
reaching 46 million ions per year (about
29% of total fiber consumption). The
projected expansion in wastepaper use
is sizable by historical standards in the
US. but the 2040 level is stiH quite
modest compared to other industrial-
ized countries.
Pulpwood consumption grows less
rapidly than use of woodpulp due to
higher yields from pulpwood and
changing species mix. Hardwood
pulpwood consumption increases from
2.2 billion cubic feet in 1986 to 5.1 bil-
lion cubic feet by 2040, an increase of
more than 130%. Softwood pulpwood
consumption is projected to increase by
50% from 1986 to 2040. The higher rate
of increase for hardwood pulpwood re-
flects a gradual shift in the industry to-
ward high-yield mechanical pulps.
Miscellaneous Timber Products. A va-
riety of other industrial timber products,
including poles, piling, posts, log ex-
ports, round mine timbers, bolts used
for shingles, handles, and woodturn-
ings, and chemical wood, are produced
or consumed in the US. Consumption
of roundwood in these products will rise
slowly from 0.8 billion cubic feet in 1986
to 0.9 billion cubic feet by 2040. The
softwood log export component of this
category (mostly derived from
Washington, Oregon, and California) is
projected to decline to roughly 0.5 bil-
lion cubic feet by 2000 and remain at
that level through 2040.
Fuelwood. Total fuetwood consumption,
which in 1986 was an estimated 3.1 bil-
lion cubic feet, is projected to increase
to 5.4 billion cubic feet in 2040 with
most of the increase occurring before
2010. This is a result of projected cost
advantages during this period for fuel-
wood relative to nonwood fuels for in-
dustrial, commercial and residential
uses. After 2020, with moderating
growth in nonwood fuel costs, fuelwood
use in residences declines.
Product Price Projections 7 . In the
Current Forest Plans projection, soft-
wood lumber prices continue to grow
for the next two decades at about the
long-term average (1.4% per year), re-
flecting declining sawtimber 8 in-
7 A1 1 prices are measured in 1982 dollars. This
excludes tIle effects of general price inflation or
deflation. The increases shown, therefore,
measure change relative to the general prices of
most competing materials.
8 1-Iere we are using a more general definition of
sawtimber that includes the part of harvest being
used in the manufacture of lumber, plywood.
and miscellaneous products, and as log exports
45
-------
Section 3
ventories on private timberlands in both
the West and South (Figure 3.6). After
2010, however, major increases in
softwood sawtimber inventories roughly
stabilize product prices (see Table 3.4).
Over the next five decades, average
softwood price growth is 0.7% per year.
Hardwood lumber prices are expected
to rise at an average annual rate of
about 0.9% per year, reflecting contin-
ued growth in major end-uses and a
steady decline in the availability of
larger timber for higher quality lumber
grades. Projections (Table 3.4) also
show rising real prices for plywood in
the near-term but roughly constant
prices for OSB. Efficiency improve-
ments lead to little change in pulp and
paper prices. We expect that future
price trends will mirror the experience
of the period 1962-1976, when cost
reductions due to technical changes
allowed stable or declining real prices.
Projections of Timber Markets and
Inventories. Given the projections of
total product demand and net trade,
harvest from US timberlands will
increase sharply over the next five
decades, rising some 50%, from 18.0
billion cubic feet in 1986 to 27.1 billion
cubic feet in 2040. Hardwood harvest is
expected to rise more rapidly than soft-
woods. Between 1966 and 2040, hard-
wood Cut is projected to rise about 79%
to 11.3 billion cubic feet, while softwood
cut grows by some 35% to 15.8 billion
cubic feet.
Projections of sawtimber softwood
stumpage prices are shown in Figure
3.6 for selected regions. Softwood
prices rise substantially in aH regions.
Projected growth rates over the 1986-
2040 period are: 2.0% in the Pacific
Coast region, 3.2% in the Rockies,
1.5% in the South, and 2.5% in the
North. In all regions, price increases are
most rapid in the period up to 2010 with
a clear slowing in growth (or even
decline in some regions) in the last
three decades of the projection. Growth
in eastern hardwood prices averages
1.4% per year.
The time patterns of timber prices
shown in Figure 3.6 reflect both the
underlying assumptions regarding pub-
lic harvests and projected changes in
the volumes and ages of timber inven-
tories on private lands. For softwoods,
the rapid price growth to 2010 results
from limitations on timber supply in both
the Pacific Coast and Southern regions
over the next two decades. Between
1990 and 2000, though public cut is
assumed relatively stable, industrial
private harvest in the Pacific Coast
region falls in the face of dwindling vol-
umes in merchantable ages and sizes.
This forces timber, and ultimately solid-
wood product, prices upward and leads
to some retrenchment in the forest
products industry in the region. Harvest
in the South continues to grow during
this period, but by 2000 inventory
restrictions are encountered in that
region as well. As a result of limited
regeneration activities on nonindustrial
lands during the period from 1965 -
1985, timber of merchantable ages
declines sharply in the period 2000 to
2010 and harvest falls. After 2010,
however, large areas of young-growth
on industrial lands in the Pacific Coast
region reach merchantable size, as
does timber recently planted on nonin-
dustrial lands in the South under
various public and private programs. As
a result, harvest rises rapidly enough
through the remainder of the projection
to stabilize timber prices.
Hardwood stumpage prices are
expected to grow relatively slowly over
the next 15 years as hardwood inven-
tories continue to expand. After 2000,
hardwood stumpage prices increase
46
-------
Section 3
350
300
.! 250
U
I d )
o 200
-J
LI.
1
4, 100
150
50
0
1950 1960 1970 1980 1990 2000 2010 2020 2030 2040
Figure 3.6. Softwood sawlimber stumpage prices from years 1952 to 1986, with
projections through 1990 to 2040.
Year
47
-------
Section 3
Table 3.4. Deflated price indexes for selected timber products in the US, by sottwoods
and hardwoods 1952, 1962, 1970, 1976, and 1986, with projections to 2040.1
Observations (previous years ) Projections (future years)
1952 1962 1970 1976 1986 2000 2010 2020 2030 2040
PRODUCTS. UNIT &
SPECIES GROUP Index of price per unit - 1982 • 100
LUMBER
Softwoods 99.8 883 953 1260 108.0 138.6 161.8 177.5 171.8 168.3
Hardwoods 104.7 103.7 1186 109.9 118.3 108.9 120.0 133.8 149.2 164.3
STRUCTURAL
PANELS
Plywood 172.0 119.0 109.2 143.6 109.3 120.4 136.8 152.8 1488 1543
OSB-waferboard 163.0 95.0 95.2 88 4 95 4 94.5 94.6
NON-SIR UCTURAL
PANELS
Hardwood 184.5 1745 1534 1106 90.7 884 867 85.0 833 81.6
Plywood
Otherpanels 2 151.4 1150 92.1 107.1 107.1 107.1 107.1 107.1 107.1
PAPER&BOARD 100.2 1053 101.5 101.5 115.8 107.6 105.7 102.9 100.3 989
1 Prices are measured in constant (1982) dollars and are net of inflation or deflation They measure price changes
r&alive to the general price level and most competing materials
2 Hardboard, particleboard, and fiberboard producis
Projections USDA Forest Service
48
-------
Section 3
because of declining hardwood inven-
tories and increased demand, espe-
ciafly for pulpwood.
Projections of harvest by region and
public/private ownerships are given in
Table 3.5. Nearly all of the near-term
increase in both softwood and hard-
wood harvest comes in the Northeast
and Southeast. With only a limited
basis for expanding harvest on private
lands, the harvest share of the Pacific
Coast region declines steadily through
the projection. At the national level the
ownership pattern of harvest differs
between the hardwood and softwood
sectors. Increased softwood harvest
during the period to 2040 comes from
both forest industry and nonindustrial
timberlands located in the East.
Increments in hardwood harvest, in
contrast, come mostly from nonindus-
trial lands. This ownership accounted
for 75% of harvest in 1986, and by 2040
the share rises to 83%. This increase
reflects declining harvests from indus-
trial lands in the East as mixed species
and some hardwood sites are con-
verted to softwood plantations.
For the US as a whole, softwood inven-
tories remain relatively constant until
2000, then rise slowly back to their mid-
1970’s levels in the ensuing 40 years.
Net softwood growth falls by 2000 but
rises rapidly thereafter as the area of
young stands increases on private
lands in the West and South. Hardwood
inventories, in contrast, are expected to
increase in total until 2010, as losses in
the South are more than offset by
growth in the North and West. In later
years, type conversions and loss of
land to non-forest uses in the South
dominates and aggregate inventory
falls. Net annual growth for hardwoods,
which was constant between 1976 and
1986, falls steadily through the projec-
tion period, due in large part to type
conversions in the South.
Over the next 50 years the timber inven-
tory in most regions will decline in both
average size and age. For example, the
average diameter of softwood timber
harvested for all private timberlands in
the South declines from 25.1 to 22.4 cm
(9.9 to 8.8 inches) during the next five
decades. The largest changes, how-
ever, are expected for softwoods on the
Pacific Coast, where the average
diameter of harvest falls 20% to 38.6 cm
(15.2 inches). At the same time, soft-
wood inventories on industrial timber-
lands in some parts of the country will
be approaching something like the
forester’s fully regulated state, with
growth and harvest in balance and a
rotation near the minimum age of mer-
chantability (roughly 25 years in the
South and 45 years in the Douglas-fir
subregion). These conditions are likely
to exist after 2010 on industrial timber-
lands in the Douglas-fir subregion and
the South.
Key Conclusions for Current Forest
Plans scenario
Earlier assessments (e.g., USDA Forest
Service 1974, 1982) have consistently
projected a future with continued
growth in consumption of forest prod-
ucts, less rapid growth in timber inven-
tories, and rising real prices for
stumpage and products. The Current
Forest Plans analysis yields a similar
view until about 2020, when declining
growth in consumption and increasing
timber inventories reduce and/or stabi-
lize the price increases. From the
Current Forest Plans case and related
work we conclude that:
1. Over the next five decades, the
primary source of growth in forest
products consumption in the US will be
in paper and paperboard. Thus,
49
-------
Section 3
Table 3.5. Harvest and inventory projections by region and public/private timberland
owners.
Region Harvest Inventory
Owner 1986 2000 2040 1986 2000 2040
billion cubic feet
Soltwoods
North
Public .1 .2 .2 11.3 13.9 18.8
Private .6 1.0 1.6 36.1 42.1 41.2
South
Public .4 .4 .5 14.1 14.8 20.2
Private 5.3 5.6 7.5 89.7 85.4 85.6
Rockies
Public .5 .6 .7 81.9 63.7 89.3
Private .3 .5 .6 18.4 18.5 12.7
Pacific coast
PubFic 1.8 1.3 1.3 106.0 95.1 92.7
Private 2.1 2.3 2.9 56.3 46.2 51.1
Total US 11.2 11.8 15.4 413.8 399.7 411.6
Hardwoods
North
Public .2 .3 .4 29.6 37.3 55.6
Private 1.8 3.8 4.9 110.1 136.8 145.0
South
Public .1 .2 .3 15.2 14.5 14.4
Private 2.8 5.3 5.3 119.0 119.8 71.5
West
Public .1 - - - - 9.4 4.4 4.6
Private .1 .3 .4 17.3 17.7 20.7
Total US 5.1 9.9 11.3 300.6 330.5 311.8
50
-------
Section 3
uncertainties in the outlook related
to this sector, such as rates of
wastepaper recycling and use, are
particularly critical.
2. Price increases in solidwood
products and sawtimber until 2010
are nearly inevitable unless there is
a significant reduction in timber
demand, as might be possible with
higher levels of recycling. More
intense forest management and
increased planting will not cure
short-term price increases or harvest
shortfalls, though their long-term
impacts can not be disputed.
3. Major plantation areas in the
South (both on industrial and non-
industrial timberlands) and large
areas of young growth in the
Douglas-fir subregion, which are
already in place, will cause slower
softwood price growth after 2010.
4. The South will be the major
source of any expansion in softwood
supply over the next 50 years. If high
planting rates in the South continue
into the 1990’s, as assumed in the
Current Forest Plans case, flat or
declining product and timber prices
seem likely.
5. Hardwoods will increase in impor-
tance relative to softwoods in the
total US harvest due to increased
use in fiber products. In this expan-
sion, the North has the potential to
match the South in terms of contri-
butions to incremental fiber output.
Hardwood area and inventory will
drop, however, if past trends in soft-
wood plantation establishment in
the South continue in the future.
6. Reductions in public timber har-
vest in the West will elicit little inter-
owner substitution in western
stumpage markets owing to limited
merchantable private inventory, but
national impacts will be reduced as
a result of significant interregional
substitution, including increased
lumber imports from Canada.
Western regions will continue to
lose market share (in all products) to
eastern regions because of rising
relative wood costs. This trend will
accelerate if public harvest falls, if
higher rates of wastepaper recycling
are realized (recycling favors the
cost-effectiveness of the South
compared to the West), or if major
public programs of tree planting are
undertaken on nonindustrial lands.
7. The potential impacts of higher
wastepaper recycling appear to be
dramatic especially in the East
where harvest reductions would be
concentrated. These impacts cumu-
late over time as lower harvest lev-
els lead to more rapid inventory
growth.
Scenario 3. Reduced National Forest
Harvest
The Reduced National Forest Harvest
scenario includes harvest reductions
caused by elimination of harvest of old
growth volumes in the Pacific
Northwest, protection of spotted owl
habitat in Washington, Oregon and
California, protection of the red cock-
aded woodpecker in the South, elimi-
nation of below cost timber sales, and
elimination of harvesting in existing
roadless areas. Overall reduction in
National Forest harvest (as growing
stock) for the year 2000 was 20% below
that in the Current Forest Plans
scenario.
Past analyses have revealed that
decreases in National Forest harvests
are partly offset by changes in harvests
of other owners or in other regions. In
regions where there are sufficient pri-
51
-------
Section 3
vate timber supplies, decreases in
National Forest harvest lead to higher
stumpage prices that, in turn, increase
timber harvests from private timber-
lands. For example, the National Forest
harvest in the Douglas-fir subregion is
reduced by 96 million cubic feet per
year (see Chapter 8, in the 1989 RPA
Timber Assessment; Haynes 1990).
Total harvest, however, is reduced by
only 40 million cubic feet by 2000, pri-
vate harvests having increased by 56
million cubic feet per year. In the
Douglas-fir subregion, these offsetting
changes cannot be sustained after
2000 because of a worsening timber
inventory situation. In other regions,
such as the Rocky Mountains, the
reduction in National Forest harvest is
partly offset throughout the projection
period.
In this type of future, declines in timber
inventories are reflected in intensified
competition for the available timber and
higher prices for softwood stumpage
prices. The prices in the Pacific
Northwest, for example, are 17% above
the Current Forest Plans base by 2040.
Finally, softwood lumber prices are
2.4% higher in 2040 than in the 1989
Assessment projection. Because of the
lumber price increases, total lumber
consumption is down 1.2% and lumber
imports from Canada are up 21% by
2040. The increase in lumber imports
comes progressively after 2000
because domestic production is
reduced as a consequence of the lower
timber inventories and the associated
higher prices. By 2040, domestic lum-
ber production is 5.2% less than the
1989 RPA Assessment projection.
There are different impacts among
regions. The largest impacts are in the
western states, particularly the Pacific
Northwest with its large National Forest
resources.
There are no significant impacts on the
hardwood resource associated with this
scenario, further illustrating the small
role of National Forests in the hard-
wood sector.
There were no large differences in the
projected timberland inventories
between this scenario and the Current
Forest Plans scenario, based on
TAMM/ATLAS for private lands and
inventory projections from the USDA
Forest Service for public lands. As indi-
cated earlier, for public lands there are
also limitations in the approach to esti-
mating future carbon flux based on the
ratio of timber volume to total carbon. In
this study we have therefore focused on
the effects of harvest reductions on
National Forest land in the region with
the greatest differences, the Pacific
Northwest West region.
Scenarios 4 and 5. Moderate and High
Increased Recycling
These scenarios examine the impacts
of two levels of additional increases in
recycling in the forest sector. There is a
growing interest in the increased use of
wastepaper as raw material for paper
and board production. Producers in
other developed countries (e.g., Japan
and European countries) use about
twice as much wastepaper as raw ma-
terial for the production of paper and
board as US producers. In the US this
recent interest seems to stem largely
from concerns about waste disposal
rather than concerns about raw material
availability. In the Moderate scenario,
the impacts of increasing wastepaper
use to 41 % of total fiber furnished were
examined. The High Recycling
assumption is derived from suggestions
made by the American Paper Institute
(API) and differs from the first primarily
in how soon the higher levels of recy-
cling are achieved. The API proposal is
that recycling levels reach 45% by
52
-------
Section 3
1995. Projected rates of fiber recycling
for the RPA and both increased recy-
cling scenarios are shown in Table 3.6.
The Moderate Recycling Assumption
was analyzed as an alternative future in
the 1989 PPA Timber Assessment
(Haynes 1990). Increased recycling led
to a 3.7% reduction in total demand for
forest products by 2040. Consumption
of sawtimber and pulpwood is shown in
Table 3.7. Also, by 2040 pulpwood use
was 17.6% less than for the RPA
scenario due to increased use of
wastepaper. Some of the savings in
wood that would have been used as
pulp are being used for the manufac-
ture of other products, especially lum-
ber. US softwood lumber consumption
rises 2.8%, imports of softwood lumber
from Canada drop by 48.2%, and US
softwood lumber production rises by
12.2%. The largest harvest reductions
are for softwoods, particularly in the
Pacific Coast states where harvest falls,
both because of lower pulpwood use
and because of competition from
increased lumber and plywood produc-
tion in the South. Reductions in harvest
volumes in the South are larger for
hardwoods than for softwoods.
Another way to look at this relationship
is in terms of reductions in area har-
vested due to increased recycling. As
the use of wastepaper increases, fewer
acres are harvested each year
(440,000 acres less in 2010). The bulk
of this unharvested area is in the South,
and is evenly split between hardwoods
and softwoods. Changes in wood
prices are another way to gauge the
impact of increased use of wastepaper.
In the South, the reductions in harvest
have fairly substantial price impacts. By
2010, sawtimber stumpage prices are
expected to be reduced by 19.7%. This
makes the region more competitive
compared to Canada, for example, and
raises lumber production by 26% by
2040. Hardwood sawtimber stumpage
prices, on the other hand, are reduced
by only 0.5%, reflecting Ihe relalively
abundant supplies of hardwood pulp-
wood and the limited interaction
between hardwood pulpwood and
sawtimber markets.
Scenarios 6, 7, 8, 9. Afforestation
Several studies have examined the
potential land available for afforestation
in the US, as well as the associated
costs of forest stand establishment
(Moulton and Richards 1990, Parks and
Hardie 1992). These studies provide
the basis for estimating areas, by forest
type and region, which could be
planted given a specified investment.
For the present study, four scenarios
have been developed based on
investments of $110 and $220 Million
per year for 10 years applied to
afforestation schemes described by
Moulton and Richards (1990) and Parks
and Hardie (1992). The assumption has
been made that after 10 years of annual
payments, the land owner will maintain
the land as forest for the course of the
rotation and beyond.
The primary source of land for
alforestation was marginal pasture-
lands, the majority occurring in the
South Central region (Table 3.8).
Allocations among regions, and among
forest types within regions, were made
based on the distributions in Moulton
and Richards (1990) and Parks and
Hardie (1992). The yield tables associ-
ated with each forest type, which are
required by the ATLAS forest inventory
model, were likewise derived from the
yield estimates in the original docu-
ments. Stand level carbon budgets
were built as described in Section 3.A
with two exceptions. First was the
absence of a carry over in the woody
53
-------
Section 3
Table 3.6. Projected percent of total fiber from recycled wastepaper in RPA, Moderate
Recycling, and High Recycling scenarios.
Year
RPA Assumption
Moderate Recycling
Assumption
High Recycling
Assumption
1986
20.9
.209
.209
2000
21.0
.285
.450
2040
28.0
.415
.450
Table 3.7. Projected consumption of sawlogs and pulpwood in RPA and Moderate
Recycling scenarios.
Sawloos Pulpwood
Year
RPA
RPA
billion cubic feet
1986
9.0
9.0
5.8
5.8
2000
8.2
8.3
7.4
7.1
2020
9.5
9.6
9.6
8.4
2040
9.6
9.9
10.8
8.9
debris and forest floor pools, since
there was no previous accumulation,
nor inputs from debris associated with
timber harvest. Second was an
allowance for increases in the soil car-
bon pool during stand development.
The basis for the accumulation of soil
carbon is field studies suggesting sig-
nificant reductions in soil carbon when
land use changes from native vegeta-
tion to agricultural production or grazing
(e.g., Mann 1986), and a reversal of that
process in cases of reforestation (e.g.,
Jenkinson 1971). The reduction is
driven by reduced inputs of organic
residues, and increased rates of
decomposition, associated with greater
aeration and higher temperature of the
soil. Typically, losses are on the order
of 10-50% in the temperate zone
(Schlesinger 1986, Johnson 1992a,
1992b). The soil carbon accumulation
after reforestation is associated with
humic materials derived from decom-
position of leaf litter and dead roots.
Such gains are on the order of 30-50 g
rn 2 yr 1 in the temperate zone
(Jenkinson 1971).
For this analysis, we have set initial soil
carbon for the planted areas at 20%
below the average for the relevant for-
est type and region (Section 3.A). Since
most of the converted land is marginal
paslureland, this intermediate value
Moderate
Recycli flQ
Moderate
A ecyclinq
54
-------
Section 3
Table 3.8. Enrollment schedules of forestation of marginal lands by region for two
funding levels and two studies.
Funding = $1 10 M per year for
- 10 years -
Funding = $220 M per year for
10 years
was deemed appropriate. Converted
croplands would start with a lower soil
carbon level. Soil carbon was then
assumed to return to the average level
over the course of one rotation. These
assumptions result in soil carbon
accumulation rates ranging from 30-50
g m 2 yr- 1 in the southern regions and
15-25 g m 2 yr 1 elsewhere.
The accumulated wood volumes on the
atforested lands were permitted to be
harvested within the TAMM modeling
framework. Harvested areas were
assumed to be replanted in the same
forest type and assigned a yield table
associated with high productivity but
low management intensity.
Scenario 10, 11, 12. Combinations
The individual forest policy scenarios
previously described include effects
that tend to either reduce supply of for-
est products (e.g., reductions in
National Forest harvest levels) or
increase supply (e.g., recycling and
afforestation). In order to evaluate the
potential for combining these options in
ways that minimize price impacts and
avoid sharp changes in social welfare,
an additional suite of combination sce-
narios was examined.
The first combination scenario paired
the Reduced National Forest Harvest
scenario with the assumptions of
Moderate Recycling. The blend of these
scenarios was anticipated to have
opposing effects on timber supply, thus
dampening price fluctuations. The
second combination included all the
assumptions of the first and added a
doubling of timber exports. A third sce-
nario combined the $110 million per
year afforestation scenario of Mouttan
and Richards with the conditions of the
second scenario.
TAMM/ATLAS
Region
Parks/Hardie
hax iø
Moulton/Richards
hax
Parks/Hardie
hax 10
Moulton/Richards
hax
SC
2607
2335
3453
3257
SE
0
0
153
0
NC
0
0
998
0
NE
33
0
94
0
RM
0
0
0
736
PSW
0
0
26
0
PNW
192
0
244
0
TOTAL
2832
2335
4968
3993
55
-------
Section 3
56
-------
SECTiON 4.
RESULTS AND DiSCUSSION
A. Current pools
Over half of the total timberland carbon
is in the mineral soil (Figure 4.1). Tree
carbon, which includes coarse roots, is
the next largest component at 3 %,
followed by woody debris (11%), forest
floor (6%), and understory (1%). Total
carbon storage in living trees on timber-
land in the US was estimated at 11.1
Pg-C. This quantity conforms with the
estimate identified in Box 6, which is
based on biomass factors derived from
Cost et a). (1990) and Koch (1989). The
comparable estimates of Birdsey
(1992a) was 11.4 Pg-C. The addition of
woodlands carbon, as described in
Section 9.B, brings the total for all
forests in the US to 11.6 Pg-C.
Privately-owned forests contain 64% of
timberland carbon. In terms of the
regional distribution, the Northeast and
South Central regions have the largest
absolute quantities of carbon (Figure
4.2). A diflerent pattern is seen for the
average quantity of carbon per unit
area, with the highest average found in
the Pacific Northwest West and the
lowest in the South Central region
(Figure 4.3).
The relatively large pool of woody
debris demonstrates that consideration
of this pool is necessary for a compre-
hensive analysis of forest carbon.
Recently harvested stands may have
large quantities of woody debris in the
form of dead roots, slumps, and logging
residue. Relatively old stands contain
large quantities of woody debris
because of the accumulation of snags
and fallen logs. Because of the limited
availability of inventory data on woody
debris, we have relied on a modeling
approach in the present study to esti-
mate woody debris pools (Section
3.A.1). Better inventory data on variabil-
ity in the woody debris pool among for-
est types and age classes is needed
and would allow for more refined anal-
yses.
The uncertainties in the pool estimates
for private timberlands have been dis-
cussed (Section 3.A, 9.A) and relate in
part to the intensity and frequency of
sampling by the FIA surveys. For public
lands, we have depended on area and
volume inventories, primarily provided
by the USDA Forest Service, which
vary widely in their reliability. The rela-
tively limited availability of data regard-
ing forest stand age class distributions
on public lands suggests the need for
additional study. Satellite remote
sensing holds considerable promise for
both vegetation classification and eval-
uation of forest structure (Iverson et al.
1989: Loveland et a). 1991; Turner et
a)., in press). Research efforts in this
area are underway at the US EPA
Environmental Research Laboratory,
Corvallis, Oregon, and may provide
new evaluations of forest condition on
public lands in the western US as well
as verification of USDA Forest Service
inventory data.
B. Current flux
1. Base year biologically driven carbon
flux
The estimates of carbon flux derived
from the stand-level carbon budgets
indicate a common sequence over the
time course of the analysis (Figure 4.4).
57
-------
Section 4
SOIL
18.3
51%
0.5
1%
LIVING TREE
11.1
31%
3.8 6%
11%
Figure 4.1. Relative contribution of carbon pools (Pg-C) to total carbon for US tim-
be rland.
Early in stand development the net
ecosystem productivity (NEP), or net
change in the total carbon pool, is
negative because of carbon loss to the
atmosphere associated with decom-
position of the woody debris pool. Near
the time of canopy closure, the rate of
carbon accumulation associated with
tree growth begins to exceed the rate of
carbon emissions from woody debris,
and the system as a whole is a carbon
sink. For older age classes the NEP
decreases again because greater
maintenance respiration costs and
other physiological or anatomical con-
straints decrease growth in the living
tree pool. Also, microbial respiration
from decay of woody debris increases
in old stands. The time course of the
fluctuations in NEP varies with forest
type (Figure 4.4). NEP in the Southeast
planted pine forest type peaks much
earlier than in the Pacific Northwest
West Douglas-fir and Northeast Maple-
Beech-Birch forest types.
The overall net flux of carbon driven by
biological processes was 286 Tg-C yr 1
moving from the atmosphere into forest
stands. Private timberlands accounted
for 78% of the total US uptake. The two
southern regions had the highest abso-
lute uptake (Figure 4.5) followed by the
two northern and four western regions.
The net uptake per unit area (Figure
4.6) was also highest in the South East
and South Central regions (0.18 to 0.21
kg-C rn 2 yr 1 ), followed by the Pacific
Northwest West, the Northeast, and the
Pacific Southwest (0.14 kg-C rn 2 yri).
UNDERSTORY
FOREST
WOODY
FLOOR
DEBRIS
2.3
58
-------
SectIon 4
Figure 4.2. Base year carbon pools (Pg-C) by region, ownership, and component for
US timberland. Region abbreviations are same as In Figure 3.2. Pub = public
ownership, Pvt = private ownership.
S
S
4
n
I
Pub P t Pub P Pub P’t Pub P’t Pub P,t Pub Pvt Pub P t Pub P,t
PNVE P 1 1 PC fl NC NI
C*rbon pool SOU TOrSlt tIOOT
U d.r,t.ory Woody dsbrta
Ves
-------
Section 4
Box 6. Literature-Based Estimate of Timberland and Woodland Biomass Pools
The Forest Carbon Model (FCM) uses
inventory data in combination with a set of
stand level carbon budgets to estimate current
carbon reservoirs. As an effort to validate the
estimate for biomass. an analysis was made of
the current biomass in timberland and
woodlands using existing literature data.
Cost et al. (1990) was the primary source for
estimates of aboveground wood, bark and fo-
liage on commercial timberlands. In that study,
fl.A plots, Bureau of Land Management inven-
tory, and stand volume information from
National Forest plots were all used for volume
estimates. These estimates were then converted
to biomass by Cost et al. using existing infor-
mation from the literature, the extent of which
varied by region (there tends to be much more
detailed county by county information and
biomass information in the East than in the
West). This approach estimated above-ground
biomass of li e sound and cull c ees, including
tops. foliage, and biomass of seedlings and
saplings. In their report. Cost et al. give
ovendry weight. VbhiCh has the advantage over
green weight in that the latter contains inher-
enLly larger differences betv ’een softvboods and
hardv oods ( ater content affecting specific
gravity) and even among species within these
fiber type categories. An example of detailed
state/county information which is available for
some Eastern states cart be found for Alabama
(FIA Staff 1985).
Belowground estimates foT biomass were
adapted from Koch (1989) as a function of the
ratio of aboveground to belowground biomass.
Region Softwood T 8 Hardwood T B
Western
Eastern
North
South
%TB=percent root biomass of total tree
biomass, summarized from Koch (1989).
Koch’s estimates are for trees > 12.7
centimeter (5 inches). The above averages ex-
clude white spruce and alpine fir, where TB
was approximately 21 7 . For both timberlands
and woodlands, total biomass for unreserved
or available lands on a per hectare basis was
applied to reserved area for which no biomass
estimates are available.
Softwood and hardwood biomass estimates
were converted to carbon using the conversion
factors (0.491 for hardwoods and 0.521 for
sofiwoods) in Koch (1989).
Woodlands
Since biomass estimates from Cost et al.
(1990) were limited to commercial unreserved
timberlands, volume and biomass for
woodlands were derived from current resource
area and state inventory reports. Major
emphasis was placed on the Western US,
because the vast proportion of woodlands are
in the West. The data requested from FIA
groups for woodlands by forest type within
state. included reserved and unreserved area.
total volume, dead volume, and number of
vees by diameter class. Whenever possible, the
diameter class information was used in
conjunction with equations to estimate top and
branch weights. Much of the data came from
volume and growth analyses in progress, with
a s ong effort by FIA to provide us with at
least preliminary results in some cases or
advice as to applying average growth data to
woodland area for which biomass information
has not been summarized.
The average kg ha 1 of biomass compiled for
the West was then assumed to hold for the
relatively small area classified as “Woodland in
the East.” References corresponding to
biornass estimates for given states and forest
types or species are given in Section 9.B. All
data used are on file with the US EPA
Environmental Research Laboratory in
Corvallis, Oregon.
14.6
17.Oc
16.3c
15.5w
15.5
19.79k
60
-------
Section 4
Estimates of Tree Carbon Pools
The estimates for tree carbon pools are
summarized by region in the table below. We
estimated aboveground carbon storage of US
forests to be 10 Pg-C, of which 9.6 Pg-C is in
timberlands, with woodlands accounting for
the remainder (Section 9.B). The literature-
based estimate, exclusively for the tree carbon
in forests in the conterminous 48 states and
including roots, is 11.6 Pg-C.
Tree Carbon Pools for Forests in the Conterminous US 1
Tree Carbon Pool 3 (Pg-C) 4
l See text for data sources
2 West comprises Washington. Oregon. and California. North Central includes Great Plains states, remainder foUovb
Waddell ci al 1989
Includes estimates for saplings, stumps and foliage on timberlands
I Petagram (Pg.Cj = I biUion mernc tons
Forest Area (Mha)
Timberland Woodbnd
Total
Region 2
West Coast
Intermountain
North Central
Northeast
Southeast
South Central
Total
Timberland
Above Below
25.6
28.2
32.9
34.0
35.1
45.3
Woodland
10.6
27.8
1.4
0.5
0.5
1.6
36.2
55.9
34.2
34.5
35.5
46.9
1.63
0.97
1.31
1.73
1.75
2.23
9.62
201.1 42.4 243.2
0.25
0.15
0.20
0.29
0.26
0.36
1.50
Total
2.00
1.44
1.53
2.02
2.03
2.60
11.62
0.12
0.32
0.02
0.01
0.01
0.02
0.50
3
4
61
-------
Section 4
Figure 4.3.
component.
Base year carbon storage (kg-C m 2 ) per unit area by region and
Region abbreviations are the same as In Figure 3.2.
40
30
10
I0
PNTT PIV IC 3 ! NC N!
R.gioD
C&rbo pool Soil Torut fooT
tJ d.rvtory goody d•brLs
62
-------
Section 4
Figure 4.4. Net ecosystem productivity (g-C rn 2 yrl) of selected forest types. Positive
values are carbon transfer from the atmosphere to the land base. Negative values are
carbon transfer to the atmosphere. Region abbreviations are the same as in Figure
3.2.
00
100
oo
PNWW Douglas-Fir
00
100
500
U
e
0.
E
V 0
I
—100
U
Z —200
— — — — — — NE Maple—Beech—B Irch
‘S — —
S. — — —
SE Planted—Pine
—300
0
26 76 100 226 160
Stand Age (Year)
63
-------
Section 4
Figure 4.5. Base year biologically-driven carbon flux (Tg-C yrl) by region, ownership,
and component for US timberland. Positive values are carbon transfer from the
atmosphere to the land base. Negative values are carbon transfer to the atmosphere
from the land base. Region abbreviations are the same as in Figure 3.2. Pub = public
ownership, Pvt = private ownership.
so
70
l0
60
40
so
*0
5
10
0
—10
LI
LI
Li
Pub P,t Pub P,t
PNYV
Pub P,t
Pub PVt
P S ’
Pub Pvt
C.rbon flux For.•t floor Und.ritory
Woody d.bris tZ
Sc
Pub P,t Pub P t Pub Pvt
NC
Mt
64
-------
Section 4
The Rocky Mountains had the lowest
uptake per unit area (0.05 kg-C rn 2
yr 1 ), reflecting relatively dry and colder
climates.
In the Pacific Northwest West, where
the age class distribution on public
lands was taken into consideration, pri-
vate lands accounted for 62% of the
total uptake but only 45% of the total
timberland area. Thai difference is due
to the greater productivity of the
younger stands which characterize pri-
vate lands in this region. Sessions
(1991) reported that 40% of the total
area of public timberland in Oregon
was greater than 150 years of age,
while the comparable value for
Douglas-fir stands on forest industry
lands (based on the RPA database)
was about 5%.
The only pool to consistently show a net
carbon emission was the woody debris
pool. This pattern is driven by the decay
of woody debris in the early to mid
stages of stand development. During
this period, the quantities of carbon
emitted from woody debris created at
the time of stand origin (due to harvest,
wildfire, etc.) and from woody debris
remnant from the previous stand, far
exceeds inputs to the compartment due
to tree mortality. Woody debris decom-
position provided the greatest offset to
tree growth in the Pacific Northwest
West and had a minor impact in the
eastern regions (Figure 4.6).
Limited opportunities exist for validation
of the base year flux estimates. Waddell
et al. (1989) estimate timber volume
growth on US available timberland at
22.3 x io9 ft 3 for 1986. The comparable
value in the present analysis is 21.0 x
1 9 ft 3 , a difference of 6%. The differ-
ence in the growth estimate occurs pre-
dominately on public lands where there
is the greatest uncertainty about inven-
tories and growth rates.
2. Base year harvest-driven carbon flux
The national growing stock reduction
associated with harvest amounted to
130 Tg-C yr 1 for the base year, with
the largest quantities occurring in the
South Central and Southeast regions
(Figure 4.7). To convert growing stock
reduction data to actual carbon removal
from forests, several factors must be
considered. Not alt growing stock
reduction results in carbon removal--
portions are left on site as logging
residues. But in addition to growing
stock, other biomass (non-growing
stock) is removed in the form of tree
tops, boles of trees less than 12.5 cm
dbh, branches and other material. In the
past decade, considerable quantities of
hardwood have been salvaged from
US forests for use as fuelwood (Haynes
1990). Using the ratios of growing stock
reductions to total removals by Haynes
(1 990), total removal was estimated to
be 139 Tg-C yr 1 .
As noted in Section 3.B, the effect of
timber harvest is manifested in several
ways. Most significant is the transfer of
a large quantity of carbon out of the tree
biomass pool. In the US, this removal is
a large proportion of the annual carbon
accumulation from tree growth. The
difference between annual growth and
harvest removals is greatest in the
Northeast, Southeast, and Southcen-
tral regions.
Besides the direct removal of carbon by
harvesting, there is the transfer of car-
bon from the living tree pool into the
woody debris pool. The difference
between total tree carbon and mer-
chantable wood carbon, for the har-
vested area, should show up in the
woody debris pool. However, compari-
son of the woody debris pool in the
65
-------
Section 4
Carbon flux
— Und.rvtory
Laying tru
Figure 4.6. Base year biologically-driven carbon flux (kg-C rn 2 yrl) per unit area by
region and component. Positive values are carbon transfer from the atmosphere to the
land base. Negative values are carbon transfer to the atmosphere from the land base.
Region abbreviations are the same as In Figure 3.2. Values are the sum of public and
private ownership.
0.1
I
—0.1
—0.2
PNTW PH P 2W SC SE NC NE
RegioD
Forest floor
Woody debris
66
-------
Section 4
4° T
30
SO
____ _ EL
PNVV PNWE P5W SC U NC NE
Region
Ownership Privet. Public
Figure 4.7. Base year growing stock reductions due to harvesting (Tg-C) by ownership
and component. Region abbreviations are the same as In Figure 3.2.
67
-------
Section 4
youngest age class arid at rotation age,
for the area harvested and replanted,
did not account for all of it. Part of this
deficit is related to loss of woody debris
associated with landuse change. The
fate of this carbon is unclear but it is
needed to balance the woody debris
budget in the base year analysis. Part
of the deficit in woody debris accumula-
tion may also be a function of an
underestimate of the woody debris pool
in the youngest age classes. If so, then
inputs to the woody debris pool may be
greater than is simulated by the stand
level carbon budgets (Section 3.A).
The rapid change in forest floor and
understory carbon associated with or
immediately following harvesting
amounted to a 33.0 Tg-C yr 1 loss of
carbon. Total area harvested was about
2.85 x 106 ha.
3. Complete base year flux analysis
Considering both the biological and
harvest-related transfers, the net carbon
accumulation on the land base was 91
Tg-C (Table 4.1). The majority of the
carbon accumulation was in the tree
(52 Tg-C) and woody debris (36 Tg-C)
pools. The forest floor showed a small
gain (5 Tg-C) and the understory a
small loss (-2 Tg-C).
If the harvested carbon entered into
permanent storage, the forest product
pool would constitute an additional sink
of 139 Tg-C yr- 1 for the base year. But
determination of the net carbon transfer
requires consideration of carbon emis-
sions associated with the decay or
burning of these and previously pro-
duced forest products. Based on the
analysis in Box 8, we estimate a net
change in the products pool of only 36
Tg-C yr 1 .
Annual carbon emissions from wildfire
on US forests vary widely between
years. Based on examination of USDA
Forest Service statistics on fire fre-
quency and extent over the last several
decades (A. Auclair, Science & Policy
Associates, personal communication)
an estimate of 20 Tg-C was used in the
present analysis. Approximately one-
third of the total emissions are associ-
ated with woodlands, which tends to
maintain a long-term carbon equilib-
rium on those lands. Direct emissions
from burning of timberlands were thus
estimated at l3Tg-C.
Table 4.1. Base year timberland carbon budget. Positive values (Tg-C) are net
accumulations and negative values are net reductions.
Biological
Harvest
Combined
Tree
310
.2481
52
Woody Debris
-60
96
36
Forest Floor
31
-26
5
Understory
5
.7
-2
TOTAL
91
1 Includes harvest
removals
(139), woody
debris
formation
(96)
and
loss
to land
use
change (13).
66
-------
Section 4
Box 7. Net Accumulation in Product and Landfill Pools
Unlike the forests, there are no systematic
inventories of the pToduct or landfill car-
bon pools. Neither the tot.al amount of
carbon in these pools, nor the distribution
among constituent pools, is known. The
model HARVCARB (Row and Phelps
1990; Section 9.E of this document) can
be used to estimate the flux to products
and landfills from newly harvested
biomass. However, because the initial
pool sizes are not known, HARVCARB
cannot calculate flux from existing stocks
of forest products in use OT in landfills.
Thus, to obtain an estimate of the net flux
associated with the product and landfill
pools, it was necessary to use published
national values for the production of for-
est products, for the transport of forest
products to landfills, and for the release
of carbon from landfills. Based on analy-
ses described here, the net flux of carbon
into the product and landfill pools was
estimated to be 36 Tg. .C yr 1 in 1990 (20
Tg-C yr 1 stored as products. 16 Tg-C
yr 1 accumulated in landfiils).
Product pool. Paper production in the
US was 85.2 Tg in 1990. Of this, 63.7
Tg was from US sources of pulp and
21.5 Tg from imported sources of pulp
(US D0C 1991). In the same year, 209
Tg of waste were transported to landfills.
Paper and paperboard account for 41% of
waste transported to landfills, or 85.7 Tg
(EPA 1991). Thus, the pool of carbon in
paper is approximately in steady state;
that is, the net flux is close to zero.
In 1990, 21.8 million m 3 (9.2 billion
board-feet) of hardwood lumber was
produced in the US. At a density of 392
kg m 3 (VanHooser and Chojnacky
1983), hardwood production was 8.5 Tg.
Softwood lumber consumption was 106
million m 3 (45.0 billion board-feet),
which included 78 million m 3 (32.9 bil-
lion board-feet) produced in the US and
28 million m 3 (12.1 billion board-feet)
imported into the US (US DoC 1991).
Using a density of 498 kg m 3 for soft-
wood .Jumber (Van Hooser and
Chojnacky 1983), 38.6 Tg of softwood
lumber was produced from US sources
and 14.2 Tg was imported. Thus, total
production of lumber from US sources
was 47.1 Tg out of a total consumption
of 61.3 Tg. Wood accounts for 3.7% of
the annual waste stream to landfills or 7.7
Tg (EPA 1991). If the fraction of US-
produced wood is the same in the waste
stream as it is in the production-stream,
then 5.9 Tg of US-produced wood was
transported to landfills. Thus, the US-
produced wood-product pool increased
by 41.2 Tg. Assuming 50% of this was
carbon, then the US-produced wood-
product carbon pool increased by about
20 Tg in 1990. This is an overestimate
because there are losses from the product
pool in addition to transport to landfills.
Landfill pool. A total of 93.4 Tg of
paper and wood products (85.7 + 7.7 Tg)
was transported to landfills in 1990.
Assuming 50% carbon content, the
amount of carbon in this flux was about
47 Tg. Emission of methane (CH4) from
US landfills is about 18 Tg yr 1 (EPA
1991). Because methane is 75% carbon,
this flux contains 13.5 Tg of carbon.
Assuming that an equal amount of C02-C
was emitted (EPA 1991), total carbon
flux from US landfills is about 27 Tg
yr- 1 . If all of the emitted carbon came
from decomposition of forest products,
then the pool of forest products in
landfills increased by about 20 Tg of
carbon in 1990. Considering that 75.6%
of paper and wood products are produced
from US sources, then the increase in
wood products in landfills from US
sources was about 16 Tg. This is clearly
an underestimate because some of the
emitted carbon came from other sources,
such as food and yard wastes.
69
-------
Section 4
Summation of the carbon flux pro-
cesses described here suggests that
the forest sector in the conterminous
US was a sink for about 114 (91+36-
13) Tg-C yr 1 of carbon for the base
year 1990. By way of comparison, car-
bon dioxide emissions (as carbon) from
fossil fuel and cement manufacture in
the 1980’s averaged about 1,300 Tg-C
yr 1 (Boden et at. 1992). Thus the forest
sector offset about 8% of US carbon
emissions in 1990.
Some of the uncertainties regarding
specific components of the carbon
pools and flux analysis have been dis-
cussed earlier. Considering the general
agreement that forest growing stock
volume is increasing in the US
(Waddell et al. 1989, Birdsey 1992a), it
seems unlikely that the sign of the esti-
mated net flux is wrong, i.e. US forests
are currently a net sink for carbon. The
estimate for the absolute magnitude of
the current sink is less certain.
Alternative assumptions about post-
harvest losses and gains in soil carbon
(Birdsey 1992c) would probably
increase the estimated sink. Using the
net growth of growing stock on public
lands in Waddell et al. (1989) would
also increase the sink. An outline of
future research aimed at reducing the
uncertainties in the present analysis is
given in Section 6.
C. Forest polIcy scenarios
During the period 1980-1988, US
emissions from fossil fuel use increased
annually at a rate of 6.5 Tg-C yr 1 ’ or
about 6% of the sink-strength of the for-
est sector. Increase in emission rates
were greater than this toward the end of
the period because of increased eco-
nomic activity. A sustained increase in
the sink-strength of the forest sector
could offset some or all of the expected
increase in emissions from fossil fuel
use. In this section, we describe several
forest policy scenarios and their pro-
jected influence on the forest sector
carbon budget.
1. The 1989 RPA Assessment scenario
Private Lands
The total non-soil forest carbon budget
is impacted over time by both changes
in the land base and changes in the
age class distribution of the forests and,
to a lesser extent, by modification of
forest management techniques. For pri-
vate timberland, the trend in total non-
soil carbon storage is upward in the
1990’s. Thus the net accumulation for
private lands is a carbon storage
increase of 46 Tg-C yr 1 for the period
1990-2000 (Figure 4.8). After 2010,
total carbon on private forest
decreases. Some of the carbon lost
from private forests will be emitted to
the atmosphere and some will be found
in products and landfills.
The fact that over the course of the 50-
year scenario the acreage of private
land forests is reduced 7% (Figure 4.9)
accounts for some of this carbon loss
but is not the dominant factor. The more
important factors are the continuing
increase in the harvest level (Figure
4.10) and a trend towards more inten-
sively managed forests which are char-
acterized by younger average age
classes and lower average carbon
storage (e.g., Cooper 1983, Harmon et
al. 1990).
Early in the scenario, increases in the
live tree carbon component of forests,
particularly in the Northeast region,
more than offset the loss of carbon due
to land use change and management
intensification. Private land in the
Pacific Northwest West also accumu-
lates carbon in the 1990’s as the age
70
-------
Section 4
I
I
20
Owri.rsbip — PubUc Prt.sta Total
Figure 4.8. Projections of net carbon accumulation or loss (Tg.C yrl) for US
timberlands (sum of available and reserved) for the RPA Assessment scenario.
Positive values are net carbon accumulations. Negative values are carbon losses.
•0
6o
40
30
1S90— 1.09 *000—2009 solo—aol. 2020—2020 *030—2050
71
-------
Section 4
150
345
V 40
I 25
I I I
1980 1990 2000 2010 2020 2030 2040
Year
Figure 4.9. Forest land base (ha x 106) for private timberlands.
72
-------
c
I
I
•1
S
=
Section 4
Figure 4.10. Projected annual growing stock volume reductions (Tg-C yrl) from US
timberlands for the RPA Assessment scenario.
73
1900—1999
1000—2000 1010—1019 2020—2029
Own.rehip Pvbllc Prtv.ta Total
-------
Section 4
class distribution shifts to slightly older
trees. After the year 2000, total non-soil
carbon begins to drop, however,
affected by both decreases in the land
base and decline in average tree car-
bon per unit area. These trends are dis-
cussed further in Haynes (1990).
Public lands
In contrast to private holdings, the forest
acreage on public lands does not
change substantially over time in the
RPA Assessment scenario. On lands
subject to timber harvest, carbon accu-
mulates at the rate of 9 to 23 Tg-C yr 1
throughout the 50-year scenario for the
nation as a whole (Figure 4.8). Carbon
reductions due to harvesting never
exceed annual carbon accretion by
growth. However, in the Pacific
Northwest West, Pacific Northwest East
and Pacific Southwest, carbon invento-
ries decline in the first few decades
(Figure 4.11). These trends reflect a
policy of prioritizing harvest of high-vol-
ume old growth stands and including
this acreage in future timber production
rotation management (Haynes 1990).
Since harvesting is disallowed on
reserved public lands, carbon accumu-
lates at a much faster rate per unit area
than on lands available for timber har-
vest. Stands greater than 175 years of
age (90 years in southern regions) do
not change in volume, due to limitations
of the FCM stand-level carbon budget
data. The accumulation of carbon thus
tends to be underestimated. That ten-
dency is balanced to some extent by
ignoring disturbances, such as catas-
trophic wildfire, which would create
young stands. The net accumulation for
all public reserved lands was about 9 to
11 Tg-C yr 1 throughout the scenario.
The dominant regions were the
Northeast and Rocky Mountains.
National Trends
The forest land carbon accumulation
rate for the conterminous US is about
65 Tg-C yr 1 in the 1990s (Figure 4.8).
Despite continuing carbon accumula-
tion on available and reserved public
lands, the magnitude of the carbon
accumulation rate declines continu-
ously. By the decade of the 2030s, the
rate is predicted to decline to 5 Tg-C
yr-i. A long term carbon accumulation
would, in principle, not be expected on
reserved lands because natural distur-
bances insure a mix of early and late
successional stages. Carbon storage
on lands managed for timber produc-
tion would likewise not be expected to
increase because of continuous harvest
removals and cycling through stand
development.
Growing stock reductions driven by
harvesting increase from 130 Tg-C yr
in 1990-1999 to about 175 Tg-C yr 1 in
2030-2039 (Figure 4.10), with private
lands accounting for most of the
change. This more intensive harvest
regime contributes to a declining ratio
of growth to removals.
The annual carbon accumulations
above do not include the possible
change in carbon contained in forest
products still in use or in landfills.
Assuming the same ratio of carbon
accumulation in forest products and
landfills to total removals as was esti-
mated for the base year (23%). this sink
grows from 36 Tg-C yr 1 to 50 Tg-C yr-i
during the 50-year period.
As with the base year analysis, it should
be noted that the current approach,
which simply takes the difference in
woody debris pools at two points in time
for all forested lands, does not account
for decomposition of woody debris on
deforested lands. The same limitations
74
-------
Section 4
3.,
1.•
a..
3.4
1.3
a..
1.1
3.0
Figure 4.11. Projections of carbon pools (Pg-C) on available public lands for the
PNWW, PNWE and PSW regions for the RPA Assessment scenario. Region
abbreviations are the same as In Figure 3.2.
—a .oo--I )—acoo-1 l—aoio— —aozo--I I—a o--l I—ao4o-1
•g1on — PNWW P5W P1 (51
75
-------
Section 4
as found in the base year analysis in
accounting for all woody debris from
harvest also apply. An effort will be
made in future analyses to carry previ-
ously forested lands in the inventory in
order to account for that C02 source.
2. The Current Forest Plans scenario
The primary effect of the Current
National Forest Plans scenario was an
increased accumulation of carbon on
public reserved lands compared with
the RPA Assessment scenario. At the
national level the net effect of this
change was a slightly higher rate of
carbon accumulation in the 1990’s and
a moderation of the long-term decline in
forest carbon accumulation rates calcu-
lated for the RPA scenario (Figure 4.6
and 4.12).
Pr,vate Lands
Despite a large reduction in public
available timberland acreage, timber
harvest removal from public forests is
only 10 to 15 % less for any year of the
projection than in the RPA scenario.
Since removals from private lands far
outweigh those from the public sector,
the overall impact on private forest
pools and flux is minimal relative to the
RPA Assessment scenario previously
described (Figures 4.8, 4.12).
The net accumulation of carbon in the
forests for the period 1990-1999 is an
increase of 44 Tg-C yr 1 , within 5% of
the RPA scenario value of 46 Tg-C yr 1
for this period (Figure 4.12). After 2010,
the 10-year mean changes in storage
are negative, i.e., a net loss of carbon
occurs. The rates of carbon loss are
essentially the same for the two scenar-
ios. At the end of the projection in year
2040, the total carbon storage of private
forest lands is 22,393 Tg-C compared
to 22,427 Tg-C for the RPA scenario, a
difference of less than 1%.
Public Lands
Compared with the RPA Assessment
scenario, current National Forest plans
call for a substantial reduction in the
acreage subject to timber harvest rota-
tion (from 19% reserved in the RPA
scenario to 41% reserved in the Current
Forest Plans scenario). In the most
extreme case, the Pacific Northwest
West, this area reduction is approxi-
matefy 57%; i.e., the total government-
owned acreage in timber production
falls from 51,194 ha in the RPA sce-
nario to 37,289 ha. Forest acreage
classified as reserved, i.e., deemed
“unsuitable” for management harvest,
more than doubles.
The reduction of acreage open to tree
harvesting increases the net storage of
carbon on public lands, particularly in
the period 1990-2010. For example,
both the RPA and Current Forest Plans
scenarios express a 9 Tg-C yr-i net
accumulation of carbon from 1990 to
2000 for lands subject to harvest. The
remainder of the accumulation is due to
maturation of the reserved forest lands.
With more than twice as much land in
this category, the Current Forest Plans
scenario has a total net carbon
accumulation of 30 Tg-C yr 1 (Figure
4.12) during the 1990’s, versus the 19
Tg-C yr 1 for the RPA scenario (Figure
4.8).
The difference in net accumulation
between the scenarios is reduced over
the course of the projection as the
reserved forest stands mature and gen-
erally move into slower growing age
classes. In year 2040, there is a 43
Tg-C yr 1 net accumulation associated
with public forests, compared with a 32
Tg-C yr 1 accumulation for the RPA
scenario. At this same time, total carbon
76
-------
Section 4
g
a
E
I
ao
10
Own.rsbp Public Pvt ata Tctl
Figure 4.12. Projections of net carbon accumulation or loss (Tg-C yrl) for US
timberlands (sum of available and reserved) for the Current Forest Plans scenario.
Positive values are carbon accumulations. Negative values are carbon losses.
I C ,
r
40
30
—10
1,10—1999 1000—1009 *010— 3 019 1020—20*9 1030—1039
77
-------
Section 4
storage is estimated at 14,980 Tg-C
versus 14,310 Tg-C, for the RPA sce-
nario.
National Trends
The larger acreage of National Forests
set aside from timber product manage-
ment increases the overall net accumu-
lation values associated with this sce-
nario. For the period 1990-1999, the net
accumulation of carbon is estimated at
75 Tg-C yr 1 (Figure 4.12), about 15%
greater than for the RPA scenario. This
accumulation declines over the projec-
tion period to 17 Tg-C yr 1 accumula-
tion in 2040. The accumulation is
always 9 to 14 Tg-C yr 1 greater than
the RPA scenario (Figure 4.8).
Since the harvest levels for the nation
as a whole do not differ greatly between
the RPA Assessment and Current Plans
scenarios (Figure 4.11, 4.13) there is
unlikely to be a large difference in the
pool of forest products still in use and in
landfills. The analysis in Section 4.C.8
reports a difference between these
scenarios of 130 Tg-C in this pool in the
year 2040. The average difference in
the sink strength over 50 years would
thus be about 2.6 Tg-C yr 1 .
3. Reduced National Forest harvest
scenario
In this preliminary analysis, there was
little effect of the additional harvest
reduction (0.36 billion cubic feet per
year reduction in 2000) on National
Forests. The growth and standing
inventory for the available timberland in
national forests over the 50-year period
was nearly the same for the Current
Forest Plans and Reduced National
Forest Harvest scenarios. The bulk of
the additional harvest reduction occurs
in the Rocky Mountain Region, yet the
projected inventory there was virtually
unchanged. As with the Current Forest
Plans scenario, there was not a sus-
tained compensatory harvest increase
on private lands.
A harvest reduction on public lands
might be expected to result in higher
standing biomass retained on that land
base as indicted in the analysis of
Harmon et al. (1990). The harvest
reduction itself represents about 2.4 Tg-
C yr 1 not removed from the public
forests and, as described below,
release of carbon via decomposition of
stumps and woody debris is also
avoided.
Because the estimates of future growth
and growing stock volume were based
on reports from individual National
Forests, the detailed analysis of age
class distributions needed to explain
the insensitivity of the growth and
growing stock inventory to harvest
reductions was unavailable. Any firm
conclusion regarding effects of these
additional harvest reductions must
therefore await more detailed treatment
of long term shifts in stand age class
distribution.
The case study analysis in the Pacific
Northwest allowed an examination of
the various factors which contribute to
the net effect of harvest reductions. The
initial net effect of reduced harvesting of
timber in the Pacific Northwest was an
increased accumulation in the forest of
1.1 Tg-C yr 1 per year. By year 2045
this effect had risen to 3.3 Tg-C yr 1
(Figure 4.14). The source from product
decay dominated early in the scenario,
but its contribution dropped to half of
the total within 10 years. The harvested
areas remained a source of carbon until
year 2033 when the carbon sink from
the area harvested early in the scenario
began to exceed the new carbon
source being created by harvesting.
78
-------
Own.rsbip — Public Prlvsta Total
Section 4
Figure 4.13. Projected annual growing stock volume reductions (TgC yr. 1 ) from US
timberlands under Current Plans scenario.
•00
l eo
3oo
I
ieeo- 1000 2000—2009 *010—2019 2020—2020 *030—2039
79
-------
Section 4
4
Net effect
— — — —
3
/ _9_00
,
/ • Not harvested
2
I
I - Harvested area
oducteo ce
1293 2000 2010 2020 2030 2040
Yea?
Figure 4.14. Change in carbon transfer (Tg-C yr-I) to atmosphere over time from
areas not harvested in the Pacific Northwest West Region. Values are based on
differences between harvest in this region under the RPA scenario and the Reduced
National Forest Harvest (0.293 x 1 cubic feet difference in 2000). Negative values
are carbon sources created by harvesting. Positive values are carbon sinks
maintained by not harvesting. The net effect is the total benefit of not harvesting in
terms of carbon not emitted to the atmosphere.
80
-------
Section 4
The carbon sink which was eliminated
by harvesting increased over time and
became the largest factor in the net
effect at year 2045.
This case study of the Pacific Northwest
was intended to provide a framework
for assessing the effects of harvest
reduction on carbon flux between the
atmosphere and land surface. The
results suggest that savings in carbon
emissions increase over time from har-
vest re uc1ion and that over a 50-year
period they remain relatively high.
The effects of harvest reductions in
other regions would be different than in
the Pacific Northwest because of differ-
ences in the harvest volume per unit
area and in the time course of the total
carbon increment for different forest
types. Comparable studies in other
regions would be of interest in evaluat-
ing a national policy with regard to har-
vest reduction on public lands.
4. Increased Recycling scenarios
The two recycling scenarios considered
in these analyses assume a public
lands management policy identical to
that of the Current Forest Plans sce-
nario; differences amongst these three
scenarios are derived solely from the
state of private timberlands. The High
Recycling scenario assumes a rapid
increase in the rate of wastepaper
recycling between 1986 and 2000, and
this is reflected in a rate of carbon
accumulation on private timberlands of
8 Tg-C yr 1 greater than the Current
Plans scenario in the 1990’s (Figure
4.15). This difference peaks in the
2000’s at 17 Tg-C yr 1 . but then
declines nearly linearly to only 1 Tg-C
yr 1 in the 2030’s. The absolute carbon
accumulation rate decreases continu-
ously over the course of the simulation,
from 82 Tg-C yr 1 in the 1990’s to 15
Tg -C yr-I in the 2030’s.
The Moderate Recycling scenario
models a more conservative increase in
the recycling rate. In the 1990’s its car-
bon accumulation rate is equivalent to
the Current Forest Plans scenario. As
the recycling rate increases over the
next three decades, the carbon accu-
mulation rate also rises relative to
Current Forest Plans (Figure 4.15). By
the 2020’s the accumulation rate ex-
ceeds that of the Current Forest Plans
by l3Tg-C yr 1 , and also surpasses the
accumulation rate of the High Recycling
scenario in this and the succeeding
decade. Here, too, the absolute accu-
mulation rate declines with each
decade from 74 Tg-C yr- 1 in the 1990’s
to 25 Tg-C yr 1 in the 2030’s.
The carbon sink benefits derived from
the recycling scenarios are directly
related to the reductions in harvest
rates on the private timberland resulting
from reuse of paper fibers. Although the
rate of recycling assumed in the High
Recycling scenario is always greater
than that of the Moderate Recycling
scenario (45% versus 28.5 to 41.5%),
economic factors applied by the TAMM
model, particularly projected stumpage
prices and levels of lumber imports from
Canada, dictate higher levels of timber
harvest for the High Recycling scenario
in the 2030’s and 2040’s. This deter-
mines the reversal in relative sink
strength of these scenarios displayed in
Figure 4.15. Integrated over the 50-year
simulation, these accumulation rates
produce 2040 forest carbon pools 0.3
Pg and 0.4 Pg higher than the Current
Forest Plans scenario for the Moderate
and High Recycling scenarios, respec-
tively (Figure 4.16).
5. Afforestation scenarios
As in the case of the recycling
81
-------
Section 4
SCENARIO °S HI R
• OD R
Figure 4.15. Annual carbon accumulation (Tg.C yr-I) for recycling scenarios relative
to Current Forest Plans projections. Scenario abbreviations are as follows: HI R =
high recycling, MOD A = moderate recycling.
a
C
C
I
I
I
1991—2000 20C 1—2010 #011—2020 2021—2030
Yeez
2032—2040
82
-------
Section 4
S
3?0
2040
fr MOD R
Figure 4.16. Comparison of projected total forest carbon pool (Pg-C) for the High and
Moderate Recycling and Current Forest Plan scenarios. Scenario abbreviations are
the same as in Figure 4.15. Scenario abbreviations are as follows: C Plan = Current
Forest Plans, HI R = High Recycling, MOD R = Moderate Recycling.
3990 2000 2030 2020 2030
390
380
580
380
Yeoz
SCENARIO °° C PLAN HI R
83
-------
Section 4
scenarios public lands management
under the afforestation scenarios con-
forms with the Current Forest Plans
scenario, and differences arise only
due to operations on private timber-
lands. Carbon accumulation is
enhanced relative to the Current Plans
scenario 1 reaching peak differences of
13 to 26 Tg-C yr 1 in the 201 0’s (Figure
4.17). The trend of the accumulation
difference parallels growth rates of
South Central lobiolly pine plantations:
the latter dominate the aflorestation
acreage allotments of all four scenarios.
The accumulation rate in excess of the
Current Forest Plans scenario declines
over 50% over the next two decades, to
5 to 12 Tg-C yr 1 . As for all scenarios,
absolute accumulation rates decline
continuously over the course of the
simulation, from 75 to 77 Tg-C yr-I in
the 1990’s to 19 to 27 Tg-C yr 1 in the
2030s.
At the $110 million per year (for 10
years) funding level, the impacts of the
Moulton and Richards and the Parks
and Hardie scenarios are indistin-
guishable. At $220 million per year
funding. the additional carbon accumu-
lation rates achieved by the Parks and
Hardie scenario are nearly double
those of the Moulton and Richards
approach (Figure 4.17). This result may
reflect both the larger area planted
(Table 3.8) and the regional distribution
of the afforestation effort. The afforesta-
tion acreage in the South Central
region is nearly equal in both
scenarios. All other afforestation
acreage in the Moulton and Richards
scenario is located in the Rocky
Mountains region. Additional acreage
01 the Parks and Hardie scenario is
more widely distributed but occurs pri-
marily in the North Central region,
which would be expected to exhibit
relatively higher growth rates.
Carbon pools at the end of the simula-
tion (2040) are the highest predicted,
ranging from 0.4 Pg ($110 million per
year Moulton and Richards) to 0.8 Pg
($220 million per year Parks and
Hardie) in excess of the Current Forest
Plans scenario (Figure 4.18).
6. Combination scenarios
Unlike the recycling and afforestation
scenarios, the combination scenarios
assume public lands management
identical to the Reduced National
Forest Harvest scenario. The impact of
the latter, in terms of carbon pools and
accumulation rates, was nearly equiva-
lent to the Current Forest .Plans sce-
nario. Therefore, the results of the first
combination, joining Moderate
Recycling with the Reduced National
Forest Harvest, are very similar to the
Moderate Recycling scenario previ-
ously described. The second combina-
tion, which adds the assumption of
doubling of export levels, reduces the
carbon sink effectiveness of moderate
recycling, relative to the Current Forest
Plans scenario, by 50% for decades in
the period 2010 - 2040 (Figure 4.19
and 4.20).
Adding a $110 million per year (total
$1.1 billion) afforestation effort more
than tripled the carbon accumulation
difference exhibited by the second
combination relative to the Current
Forest Plans. Here the increased paper
recycling rate appears to more than
compensate for increased exports. The
absolute carbon accumulation rate is
42% above the afforestation scenario in
the absence of these factors.
7. Scenario comparisons
The outcomes of these scenarios pre-
dict that US forests will accumulate car-
bon during each decade of the 50-year
84
-------
SCENARIO 000 MRI1O MR22O PHI1O D O PH220
Section 4
Figure 4.17. Annual carbon accumulation difference (Tg.C yr-i) for afforestation
scenarios relative to Current Foil rest Plans projections. Scenario abbreviations are
as follows: MR11O = Moulton and Richards, $110 million; MR220 = Moulton and
Richards, $220 million; PHI1O = Parks and Hardie, $110 million; PH220 = Parks and
Hardie, $220 million.
30
20
25
10
I
8
—5
—10
1 91—2000 2001—2C 0 2011—2020 2021—20 0 20 1—204O
Yeer
85
-------
Section 4
Comparison of Total Projected Carbon Pool (Pg) in the
Conterminous U.S. for Different Forest Policy Scenarios
1990 2000 2010 2020 2030
* MRi1O
o 0 0
L MR22O
Figure 4.18. Comparison of projected total forest carbon pool (Pg-C) for afforestation
scenarios. C Plan = Current Forest Plans ‘scenario and abbreviations for the
afforestation scenarios are the same as in Figures 4.16 and 4.17.
39
38.5
38.0
- 3.7.5
—
0
37.0
0
I-
36 5
36.0
Year
0-pp
SCENARIO C PLAN
PHi 10
2040
86
-------
Section 4
30
25
20
10
0
C
—.
:6
E
C.,
C.,
0
.0
-5
U
—10
—15
1a91—20 0 0 20332040
SCENARiO COMB 1 COMB2 COMB3
Figure 4.19. Annual carbon accumulation difference (Tg-C yr-i) for combination
scenarios relative to Current Forest Plans projections. Scenario abbreviations are as
follows: COMB1 = Combination 1, COMB2 = Combination 2, and COM.B3 =
Combination 3.
2001—2010 203 —2O20 2021—2030
Year
87
-------
Section 4
. 37.5.
0
0
1
C
U
2040
SCENARIO • : c PLAN * CO} 1 -• CO} 2 DO•O CO} 3
Figure 4.20. Comparison of projected total forest carbon pool (Pg-C) for the combi-
nation scenarios and the Current Forest Plans scenario. Scenario abbreviations are
the same as in Figures 418 and 4.19.
1 0 2000 2010 2020 2030
350
37.0
3e.5•
Ye sr
88
-------
Section 4
period (Table 4.2). The rate of carbon
accumulation will, however, decline
dramatically over this same period.
Each scenario predicts that the private
forests, as simulated by the TAMM eco-
nomics model coupled with the ATLAS
timber inventory projection model, will
change from accumulating carbon to
losing carbon between 2010 and 2040
(Figure 4.12). As noted earlier, con-
tributing factors include rising harvest
rates, shorter average rotation lengths,
and the conversion of the land base to
other uses such as urbanization. Less
rigorous data and models were avail-
able to project public forest inventories.
Both available and reserved public
timberlands appear to act as carbon
sinks throughout the simulation period
and more than balance carbon losses
from private timberlands in later
decades (Figure 4.12).
Although these trends are universal
among the scenarios, there are impor-
tant distinctions which contribute to dif-
ferences in carbon pool totals and in
temporal variations in carbon accumu-
lation or loss. The 1990 total carbon
pool of US forests is estimated to be
36.2 Pg-C. Under the now obsolete
RPA Scenario, this figure is predicted to
rise to 37.6 Pg-c in 2040 (Figure 4.16).
Increased carbon storage in public
forests accounts for all of this difference.
Private timberlands have increased
total carbon at the midpoint 0 f the simu-
lation, but the total proceeds to decline
to the 1990 level by the simulation’s
end.
The Current Forest Plans mark a signif-
icant departure from the RPA Scenario.
A substantial portion of the public forest
land base is shifted from the available
category. i.e. forests subject to harvest,
into long-term preserves. As a result,
carbon accumulation in public forests is
increased, with rates 7 to 14 Tg-C yr 1
higher than the RPA Scenario over the
course of the simulation.
Harvest rates on private forests
increased only slightly in response to
the change in public forest manage-
ment policies. From 1990-2010, carbon
accumulation rate on private lands is
only 2 Tg-C yr-i less than the RPA
Scenario. However, this increased rate
of harvest is not sustained, and by the
2030’s the carbon accumulation rate is
1 Tg-C yr 1 in excess of the RPA sce-
nario. For 2040, the total US forest car-
bon pool is estimated a 38.0 Pg. The
carbon total for private timberlands
changes less than 0.1 Pg during the
period 1990-2040.
Both the recycling scerarios and the
afforestation scenarios fL rther enhance
carbon accumulation over the 50-year
simulation period. The afforestation
scenarios have greater ‘mpact on the
2040 carbon pool total: values range
from 38.3 to 38.7 Pg-C versus 38.3 to
38.4 for the recycling poicies (Figures
4.16, 4.18). Howeve, the High
Recycling scenario rr:anifests the
largest immediate increase in carbon
accumulation. For the period 1990 to
2010, carbon accumulation rates are an
average of 13 Tg-C yr 1 higher than for
the Current Forest Plans (Figure 4.15).
By comparison, the most effective
afforestation policy over this same
period, the $220 million per year Parks
and Hardie scenario, achieves a 8 Tg-C
yr-I increase (Figure 4.17). The major
benefits of atforestation policies are
delayed until the latter half of the simu-
lation as the growth rates of the young
forests begin to substantially increase.
For the period 2010 to 2040, the aver-
age accumulation rate of the $220 mil-
lion Parks and Hardie scenario is 19
Tg-C yr 1 greater than the Current
Forest Plans accumulation rate (Figure
4.17). The High Recycling scenario
89
-------
Section 4
Table 4.2. Summary of policy scenario results.
1989 RPA Assessment Forest land accumulales carbon at a rate of 66 Mt-C per
year in the 1990’s after harvest removals are accounted
for. The accumulation rate declines over time to 5 Mt-C per
year because of a shrinking private forest land base,
increased harvest levels and a shift to younger age
classes.
Current Forest Plans
(base scenario)
Public lands accumulate an additional 10 Mt-C per year,
bringing the national sink to 74 Mt-C per year in the
1990’s. The downward trend over time in the carbon
accumulation rate is similar to the APA Assessment
Scenario.
Reduced National Forest
Harvest
Harvest reductions are likely to retain an additional sev-
eral Mt-C per year on public lands.
Moderate Recycling
Carbon uptake, relative to Current Forest Plans, gradually
increases to a peak of 13 Mt-C per year greater in 2020’s.
Carbon pool increase by 2040 is 75% of that achieved by
High Recycling.
High Recycling
Reduced harvests on private lands lead to a carbon
accumulation rate of 8 Mt-C per year greater than the
Current Forest Plans in the 1990’s, and 17 Mt per year
greater in the 2000’s. The difference declines in the later
decades, and the same trend of reduction in the carbon
accumulation rate over time is observed. Carbon pool in
2040 is 0.4 Gt greater than Current Forest Plans estimate.
Afforestation
Moulton and Richards
$110 million
Afforestation in 1990’s enhances carbon accumulation
relative to Current Forest Plans by 13 Mt-C per year in
2010’s. Carbon pool value in 2040 is 0.4 Gt greater.
Afforestation
Moulton and Richards
$220 million
Carbon accumulation gains over Current Forest Plans are
about 150% of the gain from the $110 million per year
scenario.
Afforestation
Parks and Hardie
$110 million
Essentially the same result as the Moulton and Richards
$110 million scenario.
Aflorestation
Parks and Hardie
$220 million
Most effective of the atforestation scenarios. Carbon pool
in 2040 is 0.6 Gt greater than Current Forest Plans.
Combination 1
Very similar to Moderate Recycling.
Combination 2
Reduces the carbon accumulation rates of Combination 1,
in excess of Current Forest Plans, by 50% from 2010 to
2040.
Combination 3
Increases the carbon accumulation rate from Combination
2. Carbon pool in the year 2040 is 0.5 Gt greater than the
Current Forest Plans scenario.
90
-------
Section 4
achieves only a 6 TgC yr-i advantage
over the Current Forest Plans scenario
during this same period (Figure 4.15).
The combination scenarios add rela-
tively little information. As has been
mentioned, the first combination is
essentially the same as the Moderate
Recycling scenario, in part due to the
apparently negligible effect of lower
public forest harvest rates on the forest
inventory (combination scenarios
assume public lands management
described as the Reduced National
Forest Harvest). The third combination,
combining the Moderate Recycling
assumption with doubled export levels
and the $110 million per year Moulton
and Richards Scenario, yielded a
benefit significantly greater than the
unaltered afforestation scenario. The
increased timber supply associated
with paper recycling and afforestation
was greater than the harvest increases
associated with the doubling of timber
exports. The combination scenarios
emphasized the significant impact
otherwise unconsidered factors, in this
case timber export trends, may have
upon the simulations described in this
report.
8. Results of HARVCARB analysis
In the Current Forest Plans scenario,
HARVCARB allocated 81% c i each
years harvest during the first decade to
storage in products or landfills (Table
4.3). The balance of harvested carbon
returns to the atmosphere immediately,
primarily from use for energy. To the
extent that it substitutes for energy from
fossil fuel use, the use of forest-
products for energy can be considered
to be a permanent carbon sink or an
energy offset.
In succeeding years, carbon from cur-
rent harvests continues to be allocated
largely to products. However, there is
also movement of carbon from the
products pool to both landfills and the
atmosphere as older products are
retired. By the year 2040, 64% of the
carbon harvested to that point is still in
products or landfills.
Because of a lack of inventory data on
the amount of forest carbon currently in
products and landfills, the absolute
magnitude of the current and future
products pool cannot be known with
much certainty. However, the relative
effect of different forest policy scenarios
initiated at the current time can be eval-
uated using HARVCARB. Tables 4.4
and 4.5 indicate the differences in the
forest-products, landfill, and energy-on-
set pools under two of the forest policy
scenarios compared to the Current
Forest Plans scenario. Unlike all other
scenarios, the RPA scenario results in
more carbon stored in products and
landfills because projected harvests are
higher.
The effect of the a ernative scenarios.
compared to the Current Forest Plans
scenario, can be summarized as the
50-year average annual change in
effective carbon storage (Table 4.6).
This value was calculated as the sum of
the difference in the product, landfill,
and energy-offset pools in 2040, minus
the difference in the non-energy related
release to the atmosphere, divided by
50. For example, the High Recycling
scenario forecasts a change in the
product, landfill, and energy-offset
pools for 2040 of -58.6, -44.7 and -26.8
Tg, respectively. Thus, the total change
in these pools is -130 Tg. This is
partially offset by a change of 33.9 Tg-C
in the amount of carbon released to the
atmosphere and not used for energy.
Thus, the net average change in effec-
live carbon storage is (-130 - (-33.9))/
50, or -1.9 Tg-C yr 1 .
91
-------
Section 4
Table 4.3. Cumulative disposition of carbon (Tg-C) for the Current Forest Plans
scenario.
Products still
Still in landfills
To atmosphere
Used for
Not used for
Year
in -use
and dumps
energy
energy
1980
0.0
0.0
0.0
0.0
1985
499.4
168.0
101.2
4.3
1990
959.3
375.7
255.1
60.6
1995
1360.9
588.3
420.3
205.7
2000
1764.9
621.8
598.8
381.5
2005
2174.0
1078.4
789.2
573.8
2010
2585.5
1360.9
990.1
781.1
2015
2988.4
1673.7
1205.3
1002.4
2020
3380.0
2011.8
1433.8
1236.3
2025
3770.7
2373.7
1672.9
1483.1
2030
4150.7
2756.0
1920.2
1741.8
2035
4516.8
3157.8
2173.7
2011.2
2040
4861.7
3577.6
2431.9
2288.4
Table 4.4. Difference in carbon disposition (Tg-C) between the RPA scenario and the
Current Foresi Plans scenario. Positive numbers indicate more allocation of carbon
than for the Current Forest Plans scenario.
Products still
Still in
landfills
To atmosphere
Used for
Not used for
Year
in-use
and
dumps
energy
energy
1980
0.0
0.0
0.0
0.0
1985
0.2
0.1
0.0
0.0
1990
2.3
0.7
0.5
0.0
1995
4.9
1.9
1.2
0.1
2000
11.7
5.4
3.9
0.5
2005
18.2
9.1
7.0
2.8
2010
27.8
14.6
11.3
5.8
2015
39.6
21.3
16.5
9.7
2020
47.0
27.3
21.1
14.6
2025
53.3
33.3
25.5
20.3
2030
59.3
39.7
30.4
25.6
2035
68.6
47.7
36.0
30.7
2040
75.0
55.5
41.4
36.3
92
-------
Section 4
Table 4.5. Difference in carbon disposition (Tg-C) between the High Recycling
scenario and the Current Forest Plans scenario. Negative numbers indicate less
allocation 0 f carbon than for the Current Forest Plans scenario.
Products still Still in landfills
To atmosphere
Used for
Not used for
Year
in-use
and dumps
energy
energy
1980
0.0
0.0
0.0
0.0
1985
-0.4
0.0
0.0
0.0
1990
-2.8
-0.9
-0.7
0.0
1995
-3.1
-0.5
+0.2
-0.3
2000
-21.9
-5.7
-1.9
-0.6
2005
-49.8
-15.7
-7.9
.2.3
2010
-67.3
-24.2
-14.0
-7.9
2015
-76.7
-31.0
-19.4
-15.4
2020
-81.6
- 36.5
-23.8
-22.0
2025
-82.6
-40.9
-26.9
-27.3
2030
-80.3
-44.4
-28.5
-30.9
2035
-71.0
-45.5
-28.4
-33.3
2040
-58.6
- 44.7
-26.8
-33.9
Table 4.6. The 50-year average change in flux to product, landfill, and energy-offset
pools for the four alternative policy scenarios, compared to the Current Forest Plans
scenario. A plus (+) sign means an additional carbon sink.
Policy Scenario
Average
Change in Sink (Tg-C
yr-i)
Total Carbon
CH4 Corrected
RPA
+2.7
-0.2
Reduced National Forest Harvest
-1.2
+2.4
Moderate Recycling
-0.7
-0.5
High Recycling
-1.9
+0.8
A l l orestation
Moulton & Richards $110 million
+1.7
-0.2
Moulton & Richards $220 million
+2.4
0
Parks and Hardie $110 million
+1.7
-0.2
Parks and Hardie $220 million
+2.4
0
Combination 1
-1.1
-1.1
Combination 2
-0.9
+0.6
Combination 3
-1.6
+0.8
93
-------
Section 4
The changes in effective carbon stor-
age generally followed changes in har-
vest levels. The Afforestation scenarios
had somewhat higher harvests than the
Current Forest Plans scenario and have
increased carbon sequestered in
products and fandsfills.
From a strict carbon-accounting point-
of-view, the decreases in storage in the
product, landfill, and energy-offset
pools associated with other scenarios
partially cancel the increases in the
carbon sink on the forest land base; i.e.,
to a certain extent, these scenarios shift
carbon from the products, landfills, and
energy-offset pools to the forests. It
should be noted, however, that about
half the carbon released from landfills is
in the form of methane (EPA 1991),
which, molecule for molecule, is about
25 limes more effective than carbon
dioxide as a greenhouse gas (NAS
1991). The difference in the global
warming potential of each emitted
molecule of a greenhouse gas also
depends on its atmospheric lifetime.
Integrated over 100 years the global
warming potential of emitted methane is
about 10 times greater than emitted
C02 (IPCC 1990).
The effect of methane emissions can be
put into perspective by giving half the
non-energy emissions a weight of 10.
For example, for the High Recycling
scenario, the non-energy release would
change from -33.9 Tg-C to -(16.9 10),
or -169 Tg. Using this value, the net
average change in flux becomes (-130 -
(-169))/50, or 0.8 Tg-C yr 1 . In other
words, from a climate change point-of-
view, the High Recycling scenario
creates an effective carbon sink of 0.8
Tg-C yr-i by reducing methane
emissions. This should be added to the
forest carbon sink to estimate the total
impact of this scenario compared to the
Current Forest Plans scenario.
Correcting for methane has similar
effects on the average flux associated
with the other scenarios (Table 4.6).
When the decreased landfill methane
emissions associated with the
alternative scenarios are viewed in this
way, the benefits in terms of reduced
creation of warming potential are gen-
erally greater than the loss of a carbon
sink in the forest products and landfill
pools. These results emphasize the
importance of accounting for the post-
harvest disposition of forest products,
both in terms of carbon and in terms of
potential climate impact.
D. DiscussIon
The dramatic influence of human activ-
ity on the global carbon cycle has been
well documented (IPCC 1990). As
understanding increases concerning
the mechanisms involved in anthro-
pogenic perturbation of the carbon
cycle, opportunities will develop to
apply policies for managing these
human influences. The basis for policy
decisions must in part be quantification
of potential benefits of particular policy
options, and the present study is a step
in that direction.
The analyses presented here suggest a
current potential rate of carbon accumu-
lation (considering only plant growth
and decay) on timberlands in the US of
286 Tg-C yr 1 . This rate of potential
accumulation is possible only because
of the continuous removal of carbon by
harvesting and the resetting of the
stand development cycle. Under
unmanaged conditions the rate of car-
bon accumulation on the land surface
would be lower or even negative
because fire and increased rates of
woody debris decomposition would
balance uptake. Nevertheless, the
value of 286 Tg-C yr-i indicates the
current short-term carbon sink in the
94
-------
Section 4
absence of wood harvesting.
When the effects of harvesting and fire
are considered, the net carbon accumu-
lation on the forested land surface
under current conditions is about 78
Tg-C yr-i. The corresponding net car-
bon accumulation in forest products
and landfills is at least 36 Tg-C yr 1 . By
way of comparison, the forest sector in
Canada is also thought to be a small
carbon sink (25 Tg-C yr 1 on the forest
land base and 21 19-C yr-i in prod-
ucts). Interestingly, the tree biomass
pool is decreasing and the carbon
accumulation on the land base appears
to be primarily in the pool of woody
debris and soil carbon (Kurz et al.,
1992).
Under the assumptions of the 1989
RPA Assessment, which called for
maintaining a relatively high level of
harvesting on the National Forests but
showed a long-term reduction in the
land base for private timberland, the net
carbon accumulation rate for US forests
decreases over time. There appears to
be a close relationship between the
increase in harvest (about 50 Tg.C yr 1
from 1990-2040) and the decrease in
the accumulation rate (decrease of
about 60 Tg-C yr 1 from 1990-2040).
Thus, forest sector policies which tend
to reduce harvests, such as increased
paper recycling and direct reductions in
National Forest harvest levels, are likely
to increase the rate of carbon accumu-
lation on the forest land base. Note that
the carbon accumulation rate in forest
products still in use and in landfills
would decrease with harvest reduction.
However, the magnitude of that
decrease is small relative to the benefit
achieved in terms of increased carbon
accumulation on the forest land base.
When methane emissions from landfills
are accounted for, decreased accumu-
lation in landfills is a net benefit with
respect to potential global warming.
Changes in the forest land base repre-
sent the other major option for
increased carbon sequestration cov-
ered in this analysis. In the scenarios
considered here, the maximum
increase in the forest land base was 5
million ha, and that level of afforestation
resulted in carbon sink averaging
nearly 15 TgC yr 1 over the 50-year
simulation. Nationally, however, there
are approximately 47 million ha of
marginally productive, privately-owned
crop and pasture land that is biologi-
cally capable of supporting tree growth
(Parks 1992). Thus, the maximum
afforestation scenario considered in this
report represents only 10% of land
potentially available. This land presents
an important opportunity for expanding
the area of forest land in the US and
thereby increasing the rate of carbon
sequestration.
Other strategies for increasing carbon
storage by forests include: (1) clearing
and regenerating poorly stocked forest
land if current productivity is well below
average; (2) applying intermediate
stand treatments if the current stand is
overstocked to the point of stagnation;
(3) carefully evaluating harvest and
regeneration of mature forests using the
longest possible time frame; (4) and
managing for longer rotation lengths
(Birdsey 1992b). In the case of longer
rotation lengths, an upper limit on living
forest biomass is eventually reached
and additional tree growth is largely off-
set by mortality. Thereafter, the net
additions to the terrestrial carbon stor-
age become marginal, except in those
systems where woody debris accumu-
lation continues.
As a more permanent approach to off-
setting carbon dioxide buildup in the
95
-------
Section 4
atmosphere, woody biomass crops
might be used as substitutes for fossil
fuels in energy production. Carbon is
cycled between the biosphere and the
atmosphere in that scheme, rather than
being emitted in a one way flux from
fossil fuel combustion. Assuming that
optimal conversion and production
technologies are applied, and given a
conservative estimate of the available
US land base, biomass energy could
displace as much as 6% of current
annual carbon emissions (Wright et al.
1992). Technological advancements in
the future could result in displacement
of as much as 20% of fossil fuel carbon
emissions (Wright et al. 1992). There
would also be a short term sink during
establishment of energy plantations.
Substitution of wood for materials such
as cement, aluminum, and plastics
which use considerable energy in their
manufacture, or emit C02 as a byprod-
uci of manufacture, is another approach
to reducing net C02 emissions
(CORRIM 1976). The degree to which
such substitution is made depends in
part on the cost of wood. Under the
recycling and afforestalion scenarios
projected stumpage price in the South
are driven down considerably. This
trend would promote wood substitution
and reduce C02 emissions. We have
not attempted to evaluate the
reciprocial situation, where reductions
in harvest promote use of energy
intensive wood subsitutes. Additional
carbon sinks related to management of
the biosphere include reconversion of
farmland to wetland and modified agri-
cultural practices to reduce the loss of
soil carbon (Johnson and Kern 1991).
Considered as a whole, there are sig-
nilicant opportunities for conservation
and sequestration of carbon in the
forestry sector of the US and globally
(Samuelson and Hair 1992). Obviously,
forest sector policies alone will not
compensate for the fossil carbon emis-
sion increases projected over the com-
ing decades (Lashof and Tirpak 1990,
IPPC 1992a). However, they can con-
tribute to moderating the overall trend
towards increasing C02 conentration.
E. Presentation of results for all
scenarios
Under a business as usual scenario,
US fossil emissions are expected to
increase at a rate of about 6-10 Tg-C
yr-i. The resulting emission rate in the
year 2000 would thus be nearly 1,400
Tg-C yr 1 . Assuming a goal of stabiliz-
ing US emissions at the 1990 level by
the year 2000, fossil emissions would
need to begin decreasing within the
next few years and return to the 1,300
Tg-C yr-i by the year 2000. As
described in this study, the forest sector
presents a variety of options by which to
increase carbon sequestration in the
US and thus offset fossil carbon
emissions. The carbon benefits might
be credited in a variety of ways, and the
most appropriate format for results of
the scenarios depends on how the
resutts will be used.
The first format for presentation of the
scenario comparisons is simply the total
forest carbon pool over time, including
vegetation, litter, woody debris, and soil
carbon. Results from several of the
scenarios are presented in this format
in Figure 4.21. Care must be taken with
this approach to avoid artificial varia-
tions in the soil carbon associated with
changes in the forest land base. The.
data in Figure 4.21 allow only for soil
carbon gains associated with afforesta-
tion.
A second format presents the time
course for the change in the carbon
pool over the 50-year simulation
96
-------
Section 4
990
Comparison oI Total Projected Carbon Pool (Pg) in the
Conterminous U.S. for Different Forest Policy Scenarios
SCENARJO Cure t NeUooe.1 Forest Ple.ns
‘ • M gh Recycle
Arorestet on (Pe.rks and } a .rdy. S 11I1on becte.res)
19e9 USDA Forest Service RPA
Figure 4.21. Scenario comparison based on projected total forest carbon pool (Pg-C).
Abbreviations are as follows: C Plan = Current Forest Plans, HI R = High Recycling,
PH220 = Afforestation (Parks and Hardie $220 Million), RPA = 1989 RPA Assessment.
3g.o
38.5
38.0
0
37.5
2 37.0
1
C.)
36,5
36 0
2000 2010 2020 2030
Ye e.r
2040
97
-------
Section 4
Table 4.7. Unthscounted and discounted carbon benefit for eleven forest policy
scenarios. Values are the increased carbon storage achieved over the 50 year simu-
lation compared to the Current Forest Plans scenario. Units are million tons (Mt = Tg)
of carbon. Abbreviations: MR11O and MR220 are Moulton and Richards $110 million
and $220 million afforestation scenarios; PH11O and PH220 are Parks and Hardie
afforestation scenarios.
Scenario
Undiscounted
Discount
Rate (%)
1
2
5
10
RPA
-474
-454
-358
-261
-187
Reduced NF Harvest
-17
-15
-8
-2
1
Moderate Recycling
328
310
203
105
43
High Recycling
429
411
354
279
208
MA 11O
369
350
259
164
91
MR220
556
527
390
246
137
PH11O
404
383
283
117
97
PH220
766
727
529
327
176
Combination 1
319
302
192
95
36
Combination 2
143
135
82
34
5
Combination 3
525
497
348
200
96
(Figure 4.22). The value at each time is
the thfference in the total carbon pool
between a reference scenario and the
specified policy scenario. The Current
Forest Plans scenario is used as the
reference scenario because it reflects
the most up-to-date assumptions about
harvests on federal lands. As an
example of its utility, this format reveals
the comparatively rapid accumulation of
benefits from the High Recycling sce-
nario compared to the afforestation
scenarios.
A third approach compares the scenar-
ios in terms of the total benefit over the
course of the 50 year simulation. The
carbon benefit could be applied to off-
set emissions in excess of the stabiliza-
tion level either before or after the year
2000. To reflect the increasing uncer-
tainty over time associated with model
results, a sensitivity analysis has been
made using annually compounded dis-
Count rates. For the present study, we
have used discount rates beginning in
the year 2000, of 10, 5, 2 and 1%
(Table 4.7, Figure 4.23). There is an
interaction of the timing of the benefits
and the magnitude of the discount rate
such that ranking of the scenarios may
change in some cases.
A fourth presentation of the resutts is in
terms of differences in the annual car-
bon sink over time (Figure 4.24). This
format again indicates the benefits
achieved by the policy scenarios with-
out regard to the overall source or sink
of the forest sector. This approach
allows for ready comparison to rates of
fossil emissions.
A fifth approach to presenting the
98
-------
Seótion 4
‘I
-200
—300
—400
—aoo
2040
SCENARIO 9 e Current National Forest Plans
H h Recycle
&fr Affore8t Lcn (Parks az d Hardy. 5 ithe bectare
Figure 422. Scenario comparison based on difference from the Current Forest Plans
scenario in projected total forest carbon pool (Tg-C). The abbreviations are as follows:
HI A = High Recycling, PH220 = Afforestation (Parks and Hardie, $220 Million),
COMB3 = Combination 3.
Comparison of Total Projected Carbon Pool (Pg) in the
Conterminous U.S. for Different Forest Policy Scenarios
19Q3 2000 2010 2020 2030
Year
99
-------
Section 4
400
300
C
200
100
0
Figure 4 23. Effects of assumed discount rate on the carbon sequestration benefit
(Tg-C over 50 years). Abbreviations are the same as in Figure 4.22.
600
0.00 001 0.02 0.03 0.04 0.05 0,08 0.0w 0.08 Q.0 0.10
Axuiual Discount Rate (beginning year 2000)
SCENAR 0 C 33 E H I R
100
-------
SCENARIO COMB3 *4H1R
Section 4
Figure 4.24. Scenario comparison based on the difference from the Current Forest
Plans scenario in carbon accumulation rate (Tg-C yrl). Abbreviations are the same
as in Figure 4.22.
30
25
20
15
10
—6
I-
C
0
0
.0
I-
C
—10
—15
lQ 1—2OOC
2001—2010 2011—2020 2021—2030
Year
2031—2040
fr88 PH220
101
-------
Section 4
SCENARIO COM 3 HI R
2031—2040
Figure 4.25. Scenario comparison based on carbon accumulation rate (Tg-C yr-i).
Abbreviations are same as in Figure 4,22.
results focuses on the time course of
the total net flux (i.e. net uptake per
year) for the forest land base (Figure
4.25). This perspective emphasizes the
magnitude of the responses to policy
options compared to the general trend
in the forest sink assuming no policy
changes.
A complete set of figures showing all
scenarios in each of the four formats
follows (Figures 4.26 - 4.29).
0
C
E
V
C.,
‘C
a
C
1091—20C0 200 —2010 2011—2020 2021—2030
YeAr
PH2 O
102
-------
3’ 3
U 3 0
Y.sr
C PUJ I4 1JV
Section 4
Figure 4.26. Results for aU scenarios in the total forest carbon pool (Pg-C) format.
Abbreviations are as follows: C Plan = Current Forest Plans, Red Han, = Reduced
Harvest scenario, RPA — RPA scenario, HI R = High Recycling scenario, MOD R =
Moderate Recycling scenario, MR 110 — Moulton and Richards ($110 million)
scenario, MR 220 Moulton and Richards ($220 million) scenario, PH 110 = Parks
and Hardie ($110 million) scenario, PH 220 Parks and Hardie ($220 million)
scenario, COMB 1 = Combination 1 scenario, COMB 2 = Combination 2 scenario,
COMB 3 = Combination 3 scenario.
e o.
2DCC 2 o : 2030 2040 1110 2300 2310
2=0 2020 20 0
1CtM RiO I0I 0D2
310•
— “ 0.
g
‘ I ’
‘see
2010 . l e0
V. . :
0 P0III0 P1 C
20= 20lC
Y..r
— CVV20 l
103
-------
Section 4
I I
I000•
I
U
I.
2
b
S
U
I I
-100
::
400
103C 10 O
0010 0 0040
Y.si
II — C S tI 10 ,1 C
11001
19C 00CC 10 0 0020 10 C 0040
Ye..,.
SCgPIAR2O C IC 1 VOD I
1 *00 !
*100 1000 10 O 2000 1030 1040 —
*A 0 ee — .ee ...
Figure 4.27. Results for all scenarios in terms of differences from the Current Forest
Plans scenario in the total forest carbon pool (Tg-C). Abbreviations are the same as in
Figure 4.26.
104
lee: 2 zcc icz:
Yeir
D AR1: RHAI’ —IP*
‘ IC.
s0
t. 400
0
J 1001
I IX
-------
•1o
Is
g to
1:
. : 2 C .ZC3C Oct 1 —ZCZO 1 I— 3C 2 I—2O4C
Yiar
gRIO
Section 4
tPS — OC Ot—1O Ntt—10 O— t— ’.C
Y.ar
& o oiio _ flto _pg
Figure 4.28. Resu s of afl scenarios in terms of the differences from the Current Forest
Plans scenario in carbon accumulation rate (Tg-C yr 1 ). Abbreviations are the same
as in Figure 4.26.
C
0
E
U
U
C
I
< 5
— I C
I.
0.
a
I..
C
a
E
U
U
C
2
U
g
a
E
C
U
U
I
0)SVt0 1ClI—a fl—COX
tsar
ICSIIALO O5 OOU
x.
* 0 1
tISt— IC CI—SVlO II•pCIC t-1O O
Ysir
105
-------
Section 4
I
j;i
It
lN — OC OI—WIO OtI—SC O IO I— 3O I l—$t4O
‘II ,
DaXO ttS P1W0
I0
IW$. C ISCl— IO II— 1&— t I— IC
Ysir
4A O I X I
Figure 4.29. Results for atl scenarios in terms of the carbon accumulation rate (Tg-C
yrl). Abbreviations are the same as in Figures 4.26.
IN —loOC *OOl—ZtIO OII— ZQ IX2— O — IO
“I :
Pt.4J K&)V
INI -ItX LO&— tO 1l—l t X1I. L.IOIO
T..r
ICTWa.1JO KI 10t1
S
sO
E
p
U
s u,
106
-------
SECTION 5.
POTENTIAL IMPACTS OF CLIMATE CHANGE
Projected climate change could lead to
widespread vegetation redistribution.
The analyses of carbon budgets earlier
in this document ignored possible con-
sequences of climate change. It will be
important in future assessments to
begin considering climate change
impacts. Because of the importance of
vegetation in the global carbon cycle,
climate changes could reduce or aug-
ment terrestrial carbon sources or sinks.
These changes could in turn lead to
significant feedbacks to the climate
system. Here we summarize the poten-
tial magnitude of future climate change,
the potential response of vegetation to
climate change, and the implications Cf
vegetation change for the global carbon
cycle.
Climate Change
The effect of a doubling of CO 2 concen-
trations on global climate has been
simulated using General Circulation
Models (GCMs). These models are
complex mathematical representations
of the important processes in the earth
system that affect atmospheric circula-
tion and thus climate. Currently the
models have significant limitations in
their ability to simulate the present cli-
mate, especially regional climate pat-
terns (Dickinson 1986, Grotch 1988,
Dickinson 1989). Several of the most
significant limitations include: (1) the
inadequate simulation of cloud pro-
cesses; (2) the inadequate coupling of
atmospheric and oceanic processes;
and (3) coarse spatial resolution (Gates
1985, Schlesinger and Mitchell 1985,
Dickinson 1986). Nevertheless, GCMs
can be used to indicate the potential
range of changes in global climate.
ranging from 1.5 to 4.50 C at a rate of
0.1 to 0.5 C per decade (Table 5.1,
Houghton et al. 1990). Global precipita-
tion is simulated to increase from 3 to
15% over current amounts.
Vegetation Change
Several different types of vegetation
models have been used to simulate the
response of vegetation to climate
change. At local scales, forest gap
models can be used to simulate
changes in tree species composition in
response to climate change (Botkin et
al. 1972, Shugart and West 1977,
Shugart 1984). Forest gap models
simulate forest succession after an
opening is formed in the forest canopy,
caused, for example, by the death of an
overstory tree. Currently, the models
are primarily based on empirical obser-
vations and data on the relative com-
petitive ability of tree species. This limits
their usefulness for evaluating the
effects of climate change.
To overcome the limitations, efforts are
underway to integrate physiological
responses into the gap models, for
instance, by coupling them with individ-
ual tree physiology models such as
TREGRO (Weinstein and Beloin 1990,
Weinstein et at. 1991). Several different
gap models have been used to simu-
late climate change impacts on forest
composition at specific geographic
locations (e.g., Urban and Shugart
1989; Botkin et al. 1989; Prentice et al.,
in press). Solomon (1986) used the gap
model FORENA to simulate climate
All GCMs simulate increases
mean temperature under
atmospheric CO 2 levels, with
in glDbal
doubled
increases
107
-------
Section 5
Table 5.1. Summary of predicted changes in global mean temperature and mean
annual precipitation as simulated by four General Circulation Models.
Global Circulation Model
Temperature
(°C)
Precipitation
(%)
Oregon State University 1 (OSU)
+ 2.8
+ 8
Geophysical Fluid Dynamics Laboratory 2
(GFDL) (R15 and R30 runs)
+ 4.0
+ 8
Goddard Institute for Space Studies 3 (GISS)
+ 4.2
+ 11
United Kingdom Meteorological Office 4 (UKMO)
+ 5.2
+ 15
1 Schlesinger and Zhao 1989
2 Manabe and Wetherald 1987
3 Hansen et al 1988
Wilson and Mitchell 1987
change impacts at 21 sites in eastern
North America. His research predicted
significant changes in vegetation.
indicative of a northward movement of
tree species. Subsequent modeling
studies in the same region have
supported these results (Urban and
Shugart 1989, Botkin et al. 1989).
Ecosystem process models that operate
at local to regional scales have been
developed for grassland and forest
systems. Examples include the Forest-
BGC model (Running and Coughian
1988), which simulates ecosystem pro-
ductivity and hydrologic processes, and
CENTURY (Pa t ton et al. 1988), which
simulates biogeochemical cycles in
grassland ecosystems. These models
can be used to simulate changes in
ecosystem plant productivity under
changing climate conditions. However,
these models can not predict changes
in the distribution of plant lifeforri,s or
species.
Several different vegetation models
have been developed to simulate con-
tinental to global scale changes in the
distribution of natural vegetation. One
class of models is based primarily on
correlations of biomes with several cli-
mate variables such as temperature
and precipitation (Holdridge 1947,
1967; Prentice 1990). A second class of
models is based on correlations of the
distribution of plant life forms with a
suite of climate variables that describe
the seasonality of climate (Box 1981).
Another class of broad-scale models,
including BIOME (Prentice et al. 1992)
the Canadian Climate Vegetation
Model (CCVM; Lenihan, Oregon State
University, personal communication)
and Mapped Atmosphere Plant Soil
System (MAPSS; Neilson, USDA
108
-------
Section 5
Forest Service, personal communica-
tion) are based on a mechanistic
understanding of the climate factors that
have been shown to limit the distribu-
tion of plant life forms, such as tempera-
ture and soil water. MAPSS is the most
sophisticated model in that it includes a
mechanistic calculation of water bal-
ance in which vegetation cover interac-
tively affects soil moisture (Neilson,
USDA Forest Service, personal com-
munication). The model also incorpo-
rates a stomatal conductance function
which can be modified to simulate the
direct effects of enhanced CO 2 concen-
trations on plant growth.
Before results of the simulation of vege-
tation redistribution by these different
broad-scale models are summarized, it
is important to include the direct effects
of C02 in simulations of future vegeta-
tion response. Consider that potential
evapotranspiration (PET) would
increase significantly under warmer
climate conditions (Marks 1990).
Increased drought-stress caused by
increased evapotrarispiratiori could be
the primary cause of vegetation change
(Neilson et al. 1969; Neilson 1990).
Even though precipitation is predicted
to increase globally, the increases
would not be sufficient in many regions
to offset the increase in evaporative
stress on plants caused by the increase
in PET. Hence, plants would be sub-
jected to increased drought stress.
However, enhanced concentrations of
CO 2 could mitigate the impact of
increased drought stress by improving
water-use efficiency (Jarvis and
McNaughton 1986, Allen 1990, Eamus
1991, Mooney et aL 1991). Water-use
efficiency can be defined as the amount
of carbon fixed per molecule of water
transpired. Under elevated CO 2 con-
centrations, plant stomates would not
need to remain open as long for a
specific amount of CO 2 to diffuse into
the plant and be fixed, and thus transpi-
rational losses would be reduced. This
hypothesized increase in plant water-
use efficiency under higher CO 2 con-
centrations has been demonstrated in
greenhouse and field experiments (e.g.,
Strain and Cure 1985, Drake et al.
1987, Allen 1990). Also, higher CO 2
concentrations have decreased dark
respiration in some experiments, which
would further increase water-use effi-
ciency (Norby et al. 1992).
Overall, enhanced concentrations of
CO 2 increased plant growth rates in
many experimental studies (Mooney et
al. 1991). However, some short- and
long-term studies have demonstrated
that the short-term positive growth
response may be lost over a period of
time, even in the same growing season
(Tinus 1972, Tolley and Strain 1984,
Surano et al. 1986, Tissue and Oechel
1987, Houpis et al. 1988). Additional,
long-term exposure experiments are
needed to improve our understanding
of the physiological response of plants
to higher CO 2 concentrations.
With this background information, sce-
narios of future vegetation redistribution
can be discussed. All the broad-scale
models simulate major shifts in vegeta-
tion (Table 5.2, 5.3; Neilson et al.,
USDA Forest Service, personal com-
munication). For instance, the
Holdridge life zone model suggests that
16 to 56% of the earths land surface
would change from one vegetation type
to another as a result of a doubled-CO 2
climate change (Smith et al. 1992a).
Temperate and boreal vegetation tend
to shift polewards in these simulations
and tropical forests expand.
Predictions of total forest
doubled-C02 climate
area under
conditions
109
-------
Section 5
Table 5.2. Percent future forest area relative to current forest area using several
different vegetation models and climate scenarios. B 1OME and Holdridge results
reflect global change. CCVM + MAPSS results are for combined areas of US and
Canada only.
Vegetation Model
OSU
Global Circulation Model
GISS GFDL UKMO
BIOME 1
117
120 119 125
Holdridge 2
100
101 85 95
CCVM + MAPSS 3
(Current Water Use Efficiency)
47
CCVM + MAPSS 3
(Increased Water-Use
Ethciency) 4
104
Prentice el al 1992 Neilson. USDA Forest Service, personal com-
munication
2 I1oldridge 1947. 1967 Srnfth et al 1992a Stomatal conductance reduced by 50%.
depend on the vegetation model and
climate change scenario used in the
simulation (Table 5.2). For the OSU and
GISS climate scenarios, the Hoidridge
life zone model simulates no net
change in global forest area, but under
the warmer GFDL and UKMO scenarios
it simulates a decrease in forested area.
The BIOME model simulates a 17 to
25% increase in global forest area.
However, the more process-based
models CCVM and MAPSS, when used
in conjunction, simulate a decrease in
forest area in Canada and the US
under the relatively mild GISS scenario,
assuming no change in current water-
use efficiency of ecosystems (Neilson et
al., USDA Forest Service, personal
communication). if water-use efficiency
increases by 50%, CCVM and MAPSS
simulate a slight increase in forest area.
In the US, MAPSS simulations suggest
that future forest area could range from
2 to 98% of current forest area, depend-
ing on the climate scenario, wind speed
assumptions, and plant water-use effi-
ciency (Table 5.3). The most extreme
decrease in forest area occurs under
future doublad-C02 climate conditions
as simulated by the GFDL model
(including simulated future wind
speeds) and with current water-use
efficiency. Wind speeds are simulated
by GCMs to approximately double
under doubled-CO 2 climate conditions,
increasing PET and thus drought stress.
In this scenario, closed forests are
restricted to northern Maine and
northwest Washington. The least
extreme scenario occurs under
increased water-use efficiency
(stornatal conductance reduced by
110
-------
Section 5
Table 5.3. Percent future forest area in the conterminous US relative to current forest
area, using MAPSS and several different climate and waler-use efficiency scenarios
(Neilson, USDA Forest Service, personal communication).
Climate Scenario/MAPSS Simulations
GFDL
GISS
Current Winds and Current WUE’
30
5
Current Winds and Increased WUE 2
98
60
Doubled-C02 Winds 3 and Current WUE
11
2
Doubled-C02 Winds 3 and Increased WUE 2
63
16
1 WUE Water -use-efliciency.
2 Stomatal conductance reduced by 50%.
3 Wind speed conditions for dcubled-C02 atmosphere pie
dicted by GCM.
50%), current wind
GISS simulation of
precipitation.
In summary, these model results sug-
gest that there coutd be major shifts in
the location of forests and other vegeta-
tion types under doubled-CO2 climate
conditions. However, the wide range in
vegetation change scenarios demon-
strates the critical need to: (1) increase
our understanding of the direct eflects
of C02 on plant physiology; (2) improve
the biophysical foundation of broad-
scale vegetation models; and (3)
improve the climate models used to
sirnulale doubled-C02 climate condi-
tions. The MAPSS simulations demon-
strate the importance of including the
direct effects of CO 2 in simulating future
vegetation response. Lowered rates of
stomatal conductance have the effect of
reducing drought stress and moderat-
ing the resulting vegetation change.
One additional point is that all the
above broad-scale models are equilib-
rium approaches for simulating vegeta-
tion response to climate change. That
is, they predict the occurrence of vege-
lation after sufficient time has elapsed
10 permit its full establishment under the
set of new climate conditions. They do
not address: (1) transition dynamics,
such as seedling dispersion, propaga-
lion and survival; (2) the processes of
dieback of the previously adapted
vegetation: nor (3) the influence of soil
properties and nutrient availability.
These limitations also need to be
addressed in the future development of
broadscale vegetation models.
Effects on Carbon Flux
A change in the distribution of vegeta-
lion resulting from climate change could
have significant effects on the global
carbon cycle, which in turn could affect
the rate of climate change. As climate
changes, carbon storage in the terres-
trial biosphere would change primarily
in response to changes in the area of
the world’s forests and in the age class
speeds, and the
temperature and
111
-------
Section 5
distribution of the stands within different
forest types (Section 4). If the climate
changes, vegetation types character-
ized by high levels of carbon storage
may be replaced by those with lower
carbon storage (or vice versa). Hence,
a net transfer of carbon between the ter-
restrial biosphere and atmosphere
could occur.
Prentice and Fung (1990) estimated a
mass transfer from the atmosphere into
above- and below-ground pools of the
biosphere of 235 Pg-C of carbon after
the world’s vegetation has responded
to a doubled-CO 2 climate change.
Their estimate was based on a simula-
tion of future vegetation distribution
using a modified Hoidridge model
(Prentice 1990) and the GISS GCM run
for doubled-CO 2 climate conditions
(Table 5.1). The GISS model predicts a
warmer world with a relatively large
increase in precipitation. Much of the
simulated increase in terrestrial carbon
storage occurs due to expansion of
tropical rain forests. Smith et al. (1992b,
in press), using similar techniques,
projected increases in terrestrial carbon
storage ranging from 8.5 Pg-C to 180
Pg-C. Increases in terrestrial carbon
storage of these magnitudes would
constitute a significant negative feed-
back to climate change if the carbon
sequestering occurred over a century or
shorter time-scale.
Several factors mitigate against the
scenario of a strong biospheric nega-
tive feedback to climate change. First,
human impacts on land use may pre-
vent much expansion of forests, particu-
larly in tropical regions. Second, faster
decomposition rates associated with
climate warming might cause significant
losses of soil carbon (Dixon and Turner
1991). Third, there may be a transient
pulse of carbon emissions to the atmo-
sphere generated by dynamics of vege-
tation redistribution, as described
below.
Neilson (1990), King- and Neilson
(1992), and Neilson et al. (USDA Forest
Service, personal communication)
developed an approach for exploring
the effect of transient vegetation dynam-
ics on carbon emissions. Under dou-
bled-CO 2 conditions, boreal and tem-
perate forests are expected to shift
poleward with their warmer and more
arid margins dying back and being
replaced either by different forest types
or by nonforest vegetation. In the inter-
vening areas where lifeform and
species associations remain fairly sta-
ble, moisture regimes may be expected
to vary with changes in PET and pre-
cipitation. Carbon would be
sequestered from the atmosphere in
areas of forest expansion and lost in
areas of dieback. In the intervening
forested areas where drought stress
increases, biomass and carbon loss
could occur gradually if the rate of
increase of drought stress is slow, lithe
rate of increase is fast, biomass and
carbon loss could occur rapidly. In sum,
if vegetation change is slow and grad-
ual there may be minimal effects on ter-
restrial carbon storage other than a
gradual change towards carbon levels
simulated for steady state doubled-CO2
conditions. If vegetation change is rapid
and if extensive dieback occurs, there
could arise a transitional carbon flux
(pulse) to the atmosphere serving as a
transient positive feedback to climate
change.
Neilson (1990), King and Neilson
(1992), and Neilson et at. (USDA Forest
Service, personal communication)
developed and revised a simple model
to provide global estimates of potential
carbon fluxes caused by vegetation
redistribution. The model, which was
constructed using the concepts dis-
112
-------
Section 5
cussed above, estimates potential
yearly carbon fluxes to the atmosphere
caused by forest redistribution. It em-
ploys several different estimates of
global biogeographic shills and severaJ
different assumptions on the rates and
sequences of vegetation change to
doubled-C02 conditions. The predicted
carbon fluxes are primarily dependent
on the magnitude of climate change
and the resulting vegetation redistribu-
tion, the assumed carbon density of
existing forests (5 kg rn 2 or 10 kg rn- 2 ),
and the time required for the climate to
change to the simulated doubled-C02
conditions (50 years or 100 years). The
model does not factor in changes in
tropical forests and only considers
changes in above ground carbon stor-
age. The key results of the Neilson et al.
(USDA Forest Service, personal com-
munication) study are summarized next.
Consider first the scenario that
assumes that forest dieback in
response to a changing climate does
not commence until 25 years afler the
climatically-induced expansion of
forests into non-forested regions (e.g.,
the tundra) begins. In this case, carbon
would be sequestered at a rate of 0.2 to
1.6 Pg-C yr-i for 25 years, depending
upon the rate of climate warming and
the assumed above ground biomass
density (1 Pg = 1 Ct = 1015 g). The
faster the rate of climate change (i.e., 50
years) and the higher biomass density
(i.e., 10 kg rn- 2 ), the greater the seques-
tration rate. When forest dieback
begins, the terrestrial biosphere
changes from a net sink to a net source
of C02. Net emissions of carbon to the
atmosphere range from 0.8 to 4.6 Pg-C
yr 1 for up to 35 years, depending upon
the magnitude of the climate-induced
vegetation redistribution and the den-
sity of carbon. The greater the vegeta-
tion change and the higher the above
greater the
If forest dieback is assumed to begin at
the same time as forest expansion, the
terrestrial biosphere acts as a source of
carbon to the atmosphere, with emis-
sion rates in the same range as
reported above but for a shorter period
of time.
In total, forest dieback could result in
the emission of 20 to 160 Pg-C of car-
bon to the atmosphere over 35 years Of
more. In all the simulations, forest
dieback produced a net flux of carbor
to the atmosphere. To put these result5
in context, currently global fossil tue
emissions are 6.0 Pg-C yr 1 (Marland e
al. 1989, Boden et aI. 1992). Vegetation
redistribution could produce additional
carbon emissions to the atmosphere of
over half that amount. However, in
almost all the simulations, the carbon
released to the atmosphere during
dieback was reassimilated by the later
regrowth of forests. Thus, the additional
carbon added to the atmosphere by
forest dieback is transitory: forest
regrowlh in areas of previous dieback
removes the carbon.
Another important point is that these
results also show that if forest expan-
sion into non-forested regions occurs
before forest dieback begins, the terres-
trial biosphere could be a short term but
significant sink of carbon, at a rate of
0.2 to 1.6 Pg-C yr 1 . This could stow the
increase in radiative forcing, but only
until climate had changed enough to
begin forest dieback. Thus, climate
change is likely to cause biospheric
changes that could be first a negative
feedback to the climate system and
then change to a positive Ieedba k to
the climate system.
These estimates suggest that the feed-
ground
emission
density, the
rate.
113
-------
Section 5
back potential of biospheric processes
are not trivial. Furthermore, biospheric
feedbacks are not limited to direct
effects of carbon exchange. Other
model studies have attempted to delin-
eate the effect of changing vegetative
cover on evapotranspiratiofl, ground-
level temperature, precipitation, and
rates of direct and latent heat flux
(Robinson et al. 1983, Rind 1984, Salati
1986, Dickinson et al. 1991). A model
analysis of the effect of replacing the
forests of the Amazon basin by grass-
lands showed significant decreases in
both evapotranspiration and precipita-
tion (Shukta et al. 1990), suggesting
that the hydrologic cycle would undergo
significant alteration. Similarly, if the
boreal forest were replaced by non-for-
est vegetation or bare ground, northern
latitude temperatures could decrease
just through the resulting changes in
surface albedo (Bonan et al. 1992).
In summary, the projected changes in
global climate and associated changes
in the distribution of forests are large
enough that climate change must be
considered in planning mitigation
strategies. For example, if widespread
forest plantations are planned, it would
be wise to consider where future cli-
mate change would permit continued
forest growth and where there would be
minimal increases in drought stress.
Future climate conditions should also
be considered in selecting species for
planting. For example, it may be appro-
priate to choose species adapted to
warmer and drier conditions than
presently occur at a site. Also, since
high latitude areas are expected to
warm more than low latitude areas and
tree density is low in current taiga
regions, it may be worthwhile consider-
ing forest technologies and manage-
ment practices that would enable the
establishment of forest plantations in
these high latitude areas.
For the US, climate change and the re-
sulting vegetation redistribution could
have significant effects on the potential
for forest management strategies to
partially offset US fossil fuel emissions.
The most glaring dilemma is posed by
the most severe MAPSS vegetation
change scenarios in which large frac-
tions of US forests are simulated to
dieback during a doubled-C02 climate
change. The potential exists for forest
dieback to change the forest sector from
a net sink to a net source of carbon over
the next century. Forest management
options for increasing carbon seques-
tration in the US should be developed
with this considerable uncertainty in
mind.
114
-------
SECTION 6.
FUTURE RESEARCH DIRECTIONS
A. Refine the existing modeling
framework and current Inven-
tory
Management of public lands played a
significant role in three of the policy op-
tions explored in this analysis, yet the
treatment of the current age class distri-
bution on these lands was highly sim-
plified. Several approaches may yield
improved inventory data for input to the
TAMM/ATLASIFCM framework. These
approaches include better accumula-
tion of existing inventory data, incorpo-
ration of the resu s of on-going surveys
using satellite remote sensing, and ini-
tiation of new remote sensing based
studies.
The USDA Forest Service FIA units
cover public lands in some cases, no-
tably in the Northeast and North Central
regions. These data are not readily
available but could be collected. There
is also a considerable amount of public
lands inventory data assembled for in-
dividual National Forests as part of their
required planning efforts. Such infor-
mation might be acquired on a piece-
meal basis for areas of particular con-
cern.
Several recently initiated surveys using
satellite remote sensing have produced
age class information on public lands
over broad geographic areas. The
USDA Forest Service contracted with a
private consulting firm to quantify the
amount of old growth forest in the
National Foresis of the Pacific
Northwest Douglas-fir region using data
from the Thematic Mapper sensor
(spatial resolution 30 m) (Tepley and
Green 1991). A parallel study was
made by the Wilderness Society.
Analysis of the same region across all
incorporate ownerships classes and
with an effort to incorporate additional
age classes is being made by Cohen et
aI. (in press).
These studies will not yield an inventory
with 10-year age classes. However,
they are likely to provide enough differ-
entiation among age classes to make
possible an improved assessment of
the regional carbon flux. Additional ar-
eas in which satellite remote sensing
might be of particular value are Alaska,
the Northern Rocky Mountains, and the
Central Rocky Mountains. Ultimately, it
is desirable to have a national inventory
based on satellite remote sensing or
other methodology, and to carry such
information in inventory models such as
ATLAS.
A second major improvement to the cur-
rent analysis would be an integration of
the public lands inventory data into the
TAMM/ATLAS/FCM framework. The
current version of TAMM assumes that
constraints on public land harvests are
by forest policy rather than available in-
ventory. As the proportion of public land
available for harvest is increasingly re-
stricted, it will become more important
to incorporate information on age class
distributions, stocking levels, and man-
agement intensity on public lands in the
TAMM projections. These inventory
projections can then be run through
FCM to estimate changes in carbon
storage.
A more sophisticated treatment of
emerging management strategies on
public lands (Swanson and Franklin
115
-------
Section 6
1992), such as selective cuts and
maintenance of tree species diversity
would be desirable as public lands
were brought into the TAMM/ATLAS/
FCM framework. Extensive structural
modification of ATLAS might not be re-
quired to accommodate these changes.
Additional work is also needed on the
dynamics of the woody debris pool and
the pool of carbon in forest products
and landfills. Better inventory data on
woody debris would permit more com-
prehensive calibration and validation of
the modeling approach used for woody
debris dynamics in FCM. More infor-
mation on the fate of harvest residue,
particularly on lands converted to non-
forest use, would also contribute to de-
velopment of a balanced woody debris
budget. The most important require-
ment for a thorough treatment of forest
products is a time course for the harvest
level and a breakout of the harvest into
pools with differing residence times as
products still in use or still in landfills.
A last area of anticipated improvement
is in the user interface for the TAMM/
ATLAS/FCM database. Currently, any
of more than 400 stand-level carbon
budgets can be readily compared using
a menu-driven user interface. A second
interface allows comparisons of the age
class distribution and harvested area by
age class within any combination of
forest type, region, management in-
tensity policy scenario, or time period. A
planned interface will allow an equally
flexible examination of the carbon pools
and flux by region, time period, and
policy scenario.
B. Integrate national-level studies
with global scale analyses
As noted in Section 2.A, considerable
controversy remains over the question
of whether the terrestrial biosphere as a
whole is currently a source or sink of
carbon, and over the geographic distri-
bution of possible regional sources and
sinks. Relatively detailed analyses,
such as the present one, may be
needed in many countries in order to
resolve these questions since land use
and management are such strong influ-
ences on regional carbon budgets.
The framework for an Agreement on
Global Warming, recently signed at
UNCED, requires all signers to formu-
late a national inventory of greenhouse
gas emissions For many countries, the
forest sector will be a significant com-
ponent of the national carbon budget.
There is a need to combine information
from these studies into a common
framework, and to compare their results
with estimates of current carbon flux (at
continental to global scales) derived
from alternative modeling strategies.
Results of analyses using global carbon
cycle tracer models, which estimate
carbon flux based on observed spatial
and temporal variation in the atmo-
spheric CO 2 concentration, are of par-
ticular interest (Keeling et al. 1989,
Tans et al. 1990).
C. Evaluate the potential effects
of climate change
The analyses in this document, treating
how policy changes might impact car-
bon pools and flux over time periods of
decades, are based on the assumption
of constancy in the geographic distribu-
tion of forest types, in forest productivity
and in forest disturbance regimes. To
the degree that climate changes in re-
sponse to increasing concentrations of
trace gases, these assumptions may
break down. The role of increasing CO 2
per se on plant productivity and on
drought tolerance is also likely to im-
pact forest distribution and productivity
patterns. The net effect of projected
116
-------
Section 6
changes in climate and CO 2 concen-
tration might be to increase or decrease
carbon storage on forest lands.
Changes in silvicultural practices in re-
sponse to climate change might also be
expected to impact carbon dynamics.
In order to evaluate potential impacts of
climate change on the forest sector car-
bon budget, climate/vegetation models
which simulate vegetation response to
climate are needed. At the US EPA
Environmental Research Laboratory in
Corvallis, Oregon, research over the
coming years will include development
and application of several such models
as well as the spatially distributed sur-
faces for the climate variables which
are needed to drive them. These
modeling analyses will be made in
conjunction with on-going ecophysio-
logical studies which will aid in parame-
terizing plant responses to multiple en-
vironmental variables.
The initial approach to evaluating cli-
mate change effects on forest carbon
flux will be a regional scale case study
of the Columbia River Basin in the
Pacific Northwest. Two cli-
mate/vegetation models will be spatially
distributed over the basin and applied
using the current climate and future
(e.g., doubled-C02) climate scenarios.
The first model, Mapped Atmosphere-
Plant Soil System (MAPSS), uses cli-
mate drivers and site water balance to
estimale potential leaf area index (LAI)
and vegetation type (Neilson, personal
communication; King and Neilson
1992). MAPSS requires information
about soil water holding capacity, and
mean monthly estimates of precipitation
and potential evapotranspiration.
Vegetation type is estimated for each
grid cell based on current correlations
in the distribution of vegetation and cli-
mate variables, particularly the sea-
sonality of precipitation. Potential LAI is
likewise based on calibration using ob-
served patterns in LAI and site water
balance. MAPSS does not treat photo-
synthesis, respiration, and nutrient cy-
cling, so a second, more process based
biogeochemistry model, will be used to
evaluate potential effects of climate
change on carbon balance.
One biogeochemistry model being
considered is Forest BGC (Running and
Coughlan 1988) which was designed
for analysis of the carbon, nitrogen, and
water cycles at the regional scale. It is a
“big leaf” model, meaning that a few
simple compartments (including foliage,
stems, and roots) are used rather than
representations of individual trees. The
BGC model estimates photosynthesis,
evapotrarispiration, and runoff at a daily
time step. Carbon allocation to foliage,
roots and boles, as well as decomposi-
tion of litter and soil is estimated an-
nually. Daily climatic inputs include
precipitation, maximum arid minimum
temperature, dew point temperature
and total short wave solar radiation.
The site variables include soil water
holding capacity, LAI, and the carbon
and nitrogen in foliage, stems, roots.
litter, and soil. The focus in BGC is on
incorporating basic plant physiology in
as general a fashion as possible in or-
der to simulate effects of climatic gradi-
ents or climatic change on ecosystem
mass flux. Comparison of BGC outputs
for above-ground net primary produc-
tivity and site water balance with ob-
served patterns of carbon and water
flux has indicated reasonable agree-
ment for a variety of locations (Running
and Coughlan 1988, Running and
Nemani 1988, Hunt et al. 1991).
To the degree possible, satellite remote
sensing will be used for initializing.
calibrating and validating the cli-
mate/vegetation models. As an exam-
117
-------
Section 6
pie, the recent land cover map based
on imagery from Ihe Advanced Very
High Resolution Radiometer (AVHRR)
sensor (Loveland et al. 1991) provides
high spatial resolution (1 km) informa-
tion on vegetation classification.
Additional land surface features poten-
tially detectable by satellite imagery in-
clude snow cover, surface temperature,
foliar biomass and vegetation structure.
The development and deployment of
new sensors, on the planned EOS
satellites, with higher spatial and spec-
tral resolution will expand the possibili-
ties for coupling models of terrestrial
processes to satellite imagery.
118
-------
SECTION 7.
Terrestrial vegetation is one of the
dominant components of the global
carbon cycle. An accurate assessment
of changes in terrestrial sources and
sinks of carbon, particularly as
impacted by anthropogenic factors, is
necessary to predict future changes in
the atmospheric C02 concentration.
Forest are of particular importance
among terrestrial ecosystems because
c i their great capacity to accumulate
carbon. Because stand age class distri-
butions and forest harvest rates deter-
mine the net forest carbon flux,
national-level analyses of forest sector
carbon budgets, based on forest inven-
tories, are necessary to evaluate terres-
trial carbon flux over large areas. Forest
sector carbon models also permit eval-
uation of forest policy options for miti-
gating the rising level of atmospheric
CO 2 .
The US forest sector currently (—1990)
is a significant carbon sink. Nearly 70%
of the accumulation is on the forest land
base and the remainder is in forest
products still in use or in landfills. The
pattern of carbon accumulation and
loss among regions, forest types, and
ownerships reflects a long-term
increase of forest volumes on non-
industrial private lands, a relatively
young age class distribution on industry
iands, and continued transition on pub-
lic lands from primary forest to a man-
aged condition.
CONCLUSIONS
Projections of the forest sector carbon
budget indicate a Continuous increase
in stored forest carbon but a steady
decline in the national accumulation
rate. The positive rate of sequestration
at the national level is the sum of con-
tinuous net gains on the public forest
land, and a trend of decreasing rates of
accumulation on private timberlands.
The latter are projected to show net
losses ci carbon by 2010.
The various policy options treated in
this analysis result in reduced harvest
levels or increases in the forest land
base, which moderate the trend
towards lower carbon accumulation
rates. The magnitude of the additional
carbon sequestered under these
options is significant relative to the goal
of stabilizing US carbon emission over
the next decade.
The effects of projected changes in cli-
mate over the coming decades on the
US forest sector carbon budget have
not been considered in this analysis.
Significant alteration of temperature
and precipitation regimes may be
expected to modify future forest distri-
bution and growth rates, as well as
occurrences of wildfire and other dis-
turbances. Coupling of spatially dis-
tributed models, which simulate the
distribution of vegetation types under
different climate scenarios, and pro-
cess-based ecosystem models, which
quantify climatic effects on forest pro-
ductivity, will be necessary to evaluate
these potential climate change effects.
119
-------
Section 7
120
-------
SECTION 8. LITERATURE CITED
Adams, D M and A W. Haynes 1980. The
1980 Softwood Timber Assessment Market
Model: Structure. Projections, and Policy
Simulations Forest Science Monograph No.
22.
Adams, D. M. and R W Haynes. 1990 Public
policies, private resources, and the future of
the Douglas-fir region forest economy.
Western Journal of Applied Forestry 5:64-
69
Adams. D M, P J. Aug. and D K. Anderson
1992 Future prospects for western
Washington’s timber supply University of
Washington. College of Forest Resources.
Ajtay. G L, P Ketner. and P. Duvigneaud.
1979 Terrestrial primary production and
phytomass In. S. Kempe and P. Ketner
(eds), SCOPE 13 The Global Carbon
Cycle. John Wiley & Sons. New York, NY.
Alig. R J 1985 Modeling area changes in forest
ownerships and cover types in the
Southeast USDA Forest Service Research
Paper RM-260, Rocky Mountain Forest and
Mange Experiment Station, Ft. Collins, CO.
Aug. R J W. G Hoheristein, B C. Murray. and
R G Haight 1990. Changes in area of
timberland in the United States. 1952-2040,
by ownership, forest type, region, and state
USDA Forest Service General Technical
Report SE-64, Southeastern Forest
Experiment Station
Atug. P J. and D. N. Wear. 1992. Changes in
Private Timberlands Statistics and
Projections for 1952 to 2040 Journal of
Forestry 90:31-36.
Allan, T. and J. P. Lanly. 1991. Overview of
status and trends of world forests. In:
Proceedings 01 a Tropical Workshop to
Explore Options for Global Forest
Management, Paper No. 31, Bangkok,
Thailand.
Allen, 1. H. J. 1990. Plant responses to rising
carbon dioxide and potential interactions
with air pollutanis. Journal of Environmental
Quality 19.15-34
Andrasko, K. 1990. Climate change and global
forests: Current knowledge of potential
effects, adaptation, and mitigation options
Forestry Dept., UN Food & Agriculture
Organization, No. FO:MISC19O/7, Rome,
Italy.
Apps. M. J. and W A. Kurz 1991. Assessing the
role ot Canadian forests and forest sector
activ ies in the global carbon balance. World
Resources Reviews 3.333-344
Auclair, A N D and K. J Andrasko. In press
Net C02 flux in temperate and boreal forests
in response to climate change this century: a
case study. In. K. M. Peterson (ed).
Proceedings of the NATO Advanced
Workshop on the Biological Implications of
Global Climate Change, Springer-Verlag.
Berlin, FRG
Beuter, J. N., K. N. Johnson, and H L
Scheurman 1976. Timber for Oregon’s
tomorrow, an analysis of reasonable possible
occurrences. Forest Research Laboratory.
Oregon State University Research Bulletin
19.
Birdsey, R. A. 1990. Inventory of carbon storage
and accumulation in US forest ecosystems
In. H.E. Burkhart et al. (ed.). Research in
forest inventory, monitoring, growth, and
yield. Proceedings of the IUFRO World
Congress , Montreal, Canada, August 5-11,
1990. School of Forest and Wildlife
Resources, Publication FWS-3-90 Virginia
Polytechnic Institute and State University.
Blacksburg. VA.
Birdsey. R. A. 1992. Carbon storage and
accumulation in United States forest
ecosystems. USDA Forest Service General
Technical Report WO-57.
Birdsey, P. A. 1992b Prospective changes in
forest carbon storage from increasing forest
area and timber growth, USDA Forest
Service, Washington, DC.
Birdsey, R. A. 1992c Carbon storage in trees
and forests In: R N. Sampson and D. Hair
(eds.). Forest and Global Change
121
-------
Section 6
Opportunities for Increasing Forest Cover.
American Forests, Washington, DC.
Boden, T A, R J Sepanski, F W. Stoss (eds.).
1992 Trends ‘91. A compendium of data on
global change - highlights. Carbon Dioxide
Information Analysts Center. Oak Ridge
National Laboratory. Oak Ridge. Tennessee
Bohn. B.. E. T. Degens. S Kempe, and P.
Ketner (eds) 1979 The Global Carbon
Cycle. SCOPE 13. John Wiley & Sons. New
York. NY.
Borian, G. B., D. Pollard, and S. L. Thompson
1992 Effects of boreal forest vegetation on
global climate Nature 359:716718.
Botkiri. D. 8., J F Janak. and J. R. WaHis. 1972.
Some ecological consequences of a
computer model of forest growth. Journal of
Ecology 60849-873
Botkin. D. B. R. A Nisbet. and 1. E. Reynolds.
1989. Effects of climate change on forests
of the Great Lakes States In. J B Smith and
D Tirpak (eds) Potential Etlecis of Global
Climate Change on the United Slates. US
Environmental Protection Agency.
Washington. DC.
Box. E 0. 1981 Macroclimate and Plant Forms
An Introduction to Predictive Modeling in
Phytogeography Dr W. Junk, The Hague.
Box. E 0 1988 Estimating the seasonal carbon
source-sink geography of a natural, steady-
state terrestrial biosphere. Journal of
Applied Meteorology 27.1109-1124
Brooks, D. J and J. Kincaid 1987. REACTT: An
algorithm for solving spatial equilibrium
problems. USDA Forest Service General
Technical Report PNW-GTR-204.
Cannell. M. G. R. 1982. World Forest Biomass
and Primary Production Data. Academic
Press London.
Clawson, M. 1979. Forests in the long sweep of
American history. Science 204:1168-204.
Cochran, W. G. 1977. Sampling Techniques
(third edition). John Wiley & Sons. New York,
NY.
Cohen, W B., D. 0. Wallin, M. E. Harmon, P.
Sollins. C. Daly, and W. K. Ferrell. In press
Modeling the effects of land use on carbon
storage in the forests of the Pacific
Northwest To appear in the IGARSS 1992
Proceedings of the International
Geoscience and Remote Sensing
Symposium. Publishing Services. Institute
for Electrical and Electronic Engineers. New
York, NY.
Cooper, C. F. 1983. Carbon storage in managed
forests. Canadian Journal of Forest
Research 13:155-166.
CORRIM. 1976. Renewable resources for
industrial materials Report of the Committee
on Renewable Resources for Industrial
Materials, Board of Agriculture and
Renewable Resources, Commission on
Natural Resources. National Research
Council. National Academy of Sciences.
Washington, DC.
Cost, N D., J. 0. Howard, B Mead, W. H
McWilliams, W. B Smith, 0. 0. VanHooser,
and E. H. Warton. 1990. The biomass
resource of the United States USDA Forest
Service General Technical Report WO-57.
Detwiler. R. P. and C A S Hall 1988. Tropical
forests and the global carbon cycle. Science
239 42-46
Dickinson. R. E 1986. How will climate change
The climate system and modeling of future
climate In: B. R. Bolin, J J. Doos, and R A
Warrick (eds). Scope 29. The Greenhouse
Effect, Climatic Change and Ecosystems
John Wiley. New York. NY.
Dickinson, R. E. 1989. Uncertainties of
estimates ol climatic change: A review
Climatic Change 15.5-13
Dickinson, R. E. 1991. Global change and
terrestrial hydrology--a review. Tellus
43A 8:176-181.
Dickinson, R. E., R. M. Emco, F. Giorgi, and G. T.
Bates. 1989. A regional climate model for
the western US. Climate Change 15:383-
422.
Dickinson, R. E., A. Henderson-Sellers. C.
Rosenzweig, and P. J. Sellers. 1991.
Evapotranspiration models with canopy
resistance for use in climate models, a
review. Agricultural and Forest Meteorology
54.373-388.
122
-------
Section 8
Dixon, R. K and 0 P. Turner. 1991. The global
carbon cycle and climate change.
Responses and feedbacks from below-
ground systems. Environmental Pollution
73.245-262.
Dixon, R. K., P. E. Schroeder, and J. K. Winjum
1991a. Assessment of Promising Forest
Management Practices and Technologies
for Enhancing the Conservation and
Sequestration of Atmospheric Carbon and
Their Costs at the Site Level, US
Environmental Protection Agency. Office of
Research and Development. EPA/600/3-
91/67, Washington. DC.
Dixon, R. K. J. K. Winjum, 0. N Krankina
1991b Forestation and forest management
options and their costs at the site level In D.
Howlett and C Sargent (eds.). Proceedings
Technical Workshop to Explore Global
Forest Management Options lIED, London.
UK
Dixon, R. K. and K Andrasko 1992 Integrated
systems. Assessment of promising
alternative landuse practices to enhance
carbon conservation and sequestration. In:
Climate Change Proceedings of the
Intergovernmental Panel on Climate Change
(IPCC) Workshop on Assessing
Technologies & Management Systems for
Agriculture & Forestry in Relation to Global
Climate Change. January 20-23, 1992.
Canberra. Australia
Dixon. R K., J. K. Winjum and P. E. Schroeder.
In press Conservation and sequestration of
carbon the potential of forest and agroforest
management practices. Global
Environmental Change.
Downing. J. P. and D A. Cataldo 1992. Natural
sinks of C02: Technical synthesis from
Palmas Del Mar workshop In: J. Wisniewski
and A. E. Lugo (eds.). Natural Sinks of CO 2
Kluwer Academic Publishers. Dordrecht.
Netherlands.
Drake, B. G., P. S. Curtis, W. J. Arp, P. W.
Leadly, J. Johnson, and D. Whigham. 1987.
Effects of elevated carbon dioxide on
Chesapeake Bay wetlands. Ill. Ecosystem
and whole plant responses in the first year
of exposure. Carbon Dioxide Research
Division Report 044, Office of Energy
Research.
Eamus. D. 1991. The interaction of rising C02
and temperatures w h waler use efficiency
Plant, Cell and Environment 14:843-852
EPA. 1991. Air Emissions from Municipal Solid
Waste Landfills-Background for Proposed
Standards and Guidelines EPAI45O/3-
90/Olla. US Environmental Protection
Agency. Office of Air and Radiation.
Research Triangle Park. NC.
Eriksson, H. 1991. Sources and sinks of carbon
dioxide in Sweden. Ambio 20:146-150.
Eyre. F. H (ed). 1980. Forest cover types of the
United States and Canada. Society of
Amencan Foresters. Washington. DC.
FAO/Unesco (United Nations Food and
Agriculture Organization/Educational.
Scientific, and Cu uraI Organization) 1974-
78. Soil Map of the World. Volumes IX, FAO.
Paris.
Forest Inventory and Analysis Research Work
Unfl (FIA Staff). 1985. Forest Statistics for
Alabama Counties in 1982. Resource
Bulletin SO-97. USDA Forest Service.
Southern Forest Experiment Station, New
Orleans, LA.
Franklin, J. F., F. J. Swanson, M E. Harmon, D.
A. Perry. T A Spies, V. H. Dale, A. McKee.
W. K Ferrell, J. E. Means, S. V. Gregory. J
D. Laltin. T. D. Schowalter, and 0. Larsen
1991. Effects of Global Climate Change on
Forests in Northwestern North America The
Northwest Environmental Journal 7:233-
254.
Galinski, W. 1992. Polish forest ecosystem C02
balance as influenced by changing
economical system of the country In:
Climate Change. Proceedings of the
Intergovernmental Panel on Climate Change
(IPCC) Workshop on Assessing
Technologies & Management Systems for
Agriculture & Forestry in Relation to Global
Climate Change. January 20.23, 1992,
Canberra. Australia.
Gates, W. L. 1985. Modeling as a means of
studying the climate system. In: N. C.
MacCracken and F. M. Luther (eds)
Projecting the Climatic Effects of Increasing
Carbon Dioxide. DOE/ER .0237. US
Department of Energy, Washington. DC.
123
-------
Section 8
Goodland, R. E Asibey, J Post. and M Dyson.
1990. Tropical moist forest hardwoods’ The
urgent transition to sustainability. The
International Society for Ecological
Economics Conference. The World Bank.
Washington. DC.
Gorean, T. J 1990. Balancing atmospheric
carbon dioxide Amb,o 19.230-236.
Grairiger, A. 1991. Overcoming constraints on
assessing feasible aflorestalion and
revegetation rates to combat global climate
change In D. Howlelt, and C. Sargent
(eds.) Proceedings of a technical workshop
to explore options for global forest
management. Bangkok, Thailand, April 24,
1991. International Institute for Environment
and Development. London, UK.
Gregerson, H. S Draper, and 0 Elz. 1989
People and trees The role of social forestry
in sustainable development. EDI Seminar
Series The World Bank, Washington. DC.
Grier. C. C and R S Logan 1977. Old-growth
Pseudotsuga menziesii communities of a
western Oregon watershed: Biomass
distribution and production budgets
Ecological Monographs 47.373-400.
Grotch. S. L 1988 Regional Intercomparisons
of General Circulation Model Predictions
and Historical Climate Data DOEJNBB-0084.
Hansen, J. I Fung, A Lacis, C Rind. S.
Lebedeft, R. Ruedy. and G. Russell. 1988
Global climate changes as forecast by
Goddard Institute for Space Studies three-
dimensional model. Journal of Geophysical
Research 93.934 1 -9364
Harmon, M. E., W. K. Ferrell, and J. F. Franklin
1990. Effects on carbon storage of
conversion of old-growth forests to young
forests. Science 247:699-702.
Haynes, R. W. 1990. An analysis of the timber
situation in the United Stales: 1989-2040.
USDA Forest Service General Technical
Report RM-199.
Haynes, R. W. and D. M. Adams. 1985.
Simulations of the effects of alternative
assumptions on demand-supply
determinants on the timber situation in the
United States. USDA Forest Service Forest
Resources Economics Research, US
Government Printing Office, Washington.
DC.
Heil, G W M. Gerrits, L. H. J. M. Janssen and J.
B Weenink. In press A simulation model to
evaluate CO 2 sink enhancement-modelling
the carbon storage and the carbon balance
of forested ecosystems in the Netherlands.
In: Carbon Balance of the World’s Forested
Ecosystems: Towards a Global Assessment.
Proceedings of the IPCC Workshop, May
11-15, 1992, Joensuu, Finland.
Hoidridge, L. P. 1947. Determination of world
formulations froni simple climatic data
Science 105.367-368.
Holdridge, L. P. 1967 Life zone ecology. Rev.
ed. Tropical Science Center. San Jose,
Costa Rica
Hollinger. 0. V. and J. Hunt. 1992. Global
change’ Impacts on agricufture and forestry.
Science House, Wellington. New Zealand.
August 13, 1991, Royal Society of New
Zealand.
Houghton, R. A, J. E. Hobbie. J M. Melillo. B
Moore, B J. Peterson, G. R. Shaver. and G
M Woodwell. 1983. Changes in the carbon
content of terrestrial biota and soils between
1860 and 1980. Net releases 01 CO to the
atmosphere. Ecological Monographs
53:235-263.
Houghton. R A., R 0. Boone, J. M Melillo, C A
Palm, G M. Woodwell, N Myers, B. Moore III,
and 0 K. Skole 1985 Net flux of carbon
dioxide from tropical forests in 1980 Nature
316:617-620.
Houghton, P. A., R. 0. Boone, J P. Fruci, J. E.
Hobbie, J M Melillo, C. A. Palm. B J.
Peterson, G. R. Shaver, and G. M
Woodwell. 1987. The flux of carbon from
terrestrial ecosystems to the atmosphere in
1980 due to changes in land use:
Geographic distribution of the global flux.
Tellus 39B.122-139.
Houghton, P. A., J. Unruh, and P. A. Lefebvre.
1991. Current land use in the tropics and its
potential for sequestering carbon. In: D.
Howlett and C. Sargent (eds,). Proceedings
of a technical workshop to explore options
for global forest management, Bangkok.
Thailand, April 24, 1991. International
124
-------
Section 8
Institute for Environment arid Development,
London.
Houpis, J. L J K A Surario. S Crowles, and J.
H Shinn 1988 Chlorophyll and carotenoid
concentrations in two varieties of Par us
ponderosa seedlings subjected to long-term
elevated carbon dioxide. Tree Physiology
4:187-193.
Howleti, D and C Sargent 1991. Proceedings
of a technical workshop to explore options
for global forestry management. Bangkok,
Thailand. April 1991. International Institute
for Environment and Development, London
Hunt, E P., Jr. F. C. Martin, and S W. Running
1991. Simulating the effects of climatic
variation on stem carbon accumulation of a
ponderosa pine stand Comparison with
annual growth increment data. Tree
Physiology 9 161-171.
Husch. B 1963 Forest Mensuration and
Statistics Ronald Press, New York, NY.
Ince, P. J I Durbak, and J Howard. In press.
The FPL Pulpwood Model. data,
assumptions, and projections to the year
2040. USDA Forest Service.
IPCC (Intergovernmental Panel on Climate
Change) 1990 Climate change, The IPCC
Scientific Assessment. Cambridge
University Press Cambridge, UK
IPCC (Intergovernmental Panel on Climate
Change) 1992a. IPCC 1992 Science
Assessment. Report to IPCC from Working
Group 1, prepared by the IPCC Group at the
Meteorological Office, Bracknell, UK.
IPCC (Intergovernmental Panel on Climate
Change). 1992b. Climate Change.
Proceedings of a workshop on assessing
technologies and management systems for
agriculture and forestry in relation to global
climate change. January 20 23, 1992.
Australian Government Publishing Service,
Canberra, Australia,.
1PCC (Intergovernmental Panel on Climate
Change). In Press. Preserving tropical
forests and climate: the role of trees in
greenhouse gas emissions trading. In:
Carbon Balance of the World’s Forested
Ecosystems. Toward a Global Assessment.
Proceedings of the IPCC workshop, held
May 11-15, 1992, Joensuu, Finland.
ISC (Interagency Scientific Commitlee). 1990 A
Conservation Strategy for the Northern
Spotted Owl. USDA Forest Service.
lverson, L. P.. R. L. Graham, and E A. Cook.
1989. Applications of satellite remote
sensing to forested ecosystems. Landscape
Ecology 3:131-143.
Jarvis. P. G. and K G. McNaughton. 1986
Stomatal control of transpiration: scaling up
from leaf to region. Advances in Ecology
Research 15:1 -49.
Jenkinson, D. S. 1971. The accumulation of
organic matter in soil left uncultivated
Rathamsted Experimental Station Report for
1970. Part 2.
Johnson, D. W. 1992a Effects of forest
management on soil carbon storage
National Council of the Paper Industry for Air
and Stream Improvement. Technical Bulletin
No. 628
Johnson, D W. 1992b. Effects of
management on soil carbon storage
Air, and Soil Pollution 64.83-120.
Johnson, M G and J. S. Kern. 1991
Sequestering carbon in soils. A workshop to
explore the potential for mitigating global
climate change. EPA/60013-91/031.
Johnson, W. C. and D. M Sharp. 1984 The ratio
of total to merchantable forest biomass and
its application to the global carbon budget
Canadian Journal of Forest Research
13:337-383.
Karjalainen, T. arid S. Kellomáki In press
Carbon storage of forest ecosystems in
Finland. In’ Carbon Balance of the World’s
Forested Ecosystems: Towards a Global
Assessment. Proceedings of the 1PCC
Workshop, May 11-15, 1992, Joensuu,
Finland.
Kauppi, P. E., K. K. K. Mielikãinen, and K.
Kuusela. 1992. Biomass and carbon budget
of European forests, 1971 to 1990 Science
256:70-74.
Keeling, C D., P B Bacastow, A. F. Carter. S. C
Piper, 1. P. Whorl, M Heimann, W. G Mook,
forest
Waler,
125
-------
and H Roeloft2en 1989 A three-
dimensional model of atmospheric C02
transport based on observed winds:
Analysis of observational data Geophysical
Monograph 55 165-236 American
Geophysical Union, Washington, DC.
Kern, .1 & and M G Johnson 1991 . The impact
of conservation tillage use on soil and
atmospheric carbon in the contiguous
Unfted Slates US Environmental Protection
Agency. Environmental Research
Laboratory. EPAJ600I3-91/056. Corvallis,
OR.
King. A W.. D L. DeAnqealis. and W. M Post.
1937. The seasonal exchange of carbon
dioxide between the atmosphere and the
terrestrial biosphere extrapolation from sIte-
specific models to regional models
Publication 2988. Environmental Sciences
Division. Oak Ridge National Laboratory, Oak
Ridge. TN.
King. G A and R P Neilson. 1992. The
transient response c i vegetation to climate
change a potential source of CC 2 to the
atmosphere Water. Air and Soil Pollution
64 365-383
King. R A. and F. S. Ho 1972. Reaclive
programming a market simulation spatial
equilibrium Department of Economics,
Economic Research Report 21
Koch. P 1989. Estimates by species group and
region in the USA 01 below-ground root
weight as a percentage 01 ovendry complete
tree weight and carbon content of tree
portions Consulting Report to the USDA
Forest Service
Kolchugirta. T. P. and T S Vinsort. in press
Equilibrium analysis of carbon pools and
fluxes of forest biomes in the Soviet Union.
Canadian Journal ol Forest Research.
Kuri, W. A. P. J. McNamee. and 1. PA. Webb.
¶990. The annual carbon budget 01 the
Canadian forest sector Phase I: Assessment
of the current net Balance. ESSA
Environmental and Social Systems Analysts
Ltd. Vancouver. BC.
Kurt W. A.. M. J. Appe, T. M. Webb, and P. J.
McNamee. 1992 The carbon budget of the
Canadian forest sector: Phase I. Informative
Report NOR-X-326 Forestry Canada.
Lastiof and Tirpak (edsj. 1990 Policy options
br stabil zing globat change: Report to
Congress. US EPA Environmental
Protection Agency. Office of Policy.
Planning. and Evaluation. Climate Change
Division, Washington. DC.
Loveland, T. R., J. W. Merchant. D. 0. Ohien.
and J. F. Brown. 1991. Development of a
land-cover characte stics database for the
conterminous 115. Photograrnetric Engi-
neering and Remote Sensing 57:1453-
1463.
Lugo. A , E. 1992. The search for carbon sinks in
the tropics. In: J. Wisniewski and A. E. Lugo
(eds). Natural Sinks of C D 2 . K luwer
Academic Publishers. Dordrecht.
Nethertands.
Lugo and Brown. 1991. Comparing tropical and
temperate forests. In: J. Cole. 6 Lovett. and
S. Findlay (eds.). Comparative analyses of
ecosystems’ patterns mechanisms, arid
theories. Spnnger-Verlag. New York
Lugo. A. E. and J. Wisniewski (eds). 1992
Conclusions. in. Natural Sinks of CO 2
K luwer Academic Publishers. Dordrecht.
Netherlands.
Mactaren. J. P. and S. J. Wakehn. 1991 Forestry
and Forest Products as a Carbon Sink in
New Zealand. FRI Bulletin No. 162. tin Bryce
Printers, Rotorua, New 2ealand.
MacLean, C. P.. .1. L. Ohmann, and P. M
Bassett. 1991. Preliminary timber resource
statistics for southwest Washington
Resource Bulletin PNW- Re. 177. USDA
Forest Service, Pacific Northwest Research
Station, Portland, OR.
Maini, J. S. 1991. Towards an international
instrument of forest sustainable
development and the Canadian forest
sector. In: Proceedings of a Tropical
Workshop to Explore Options for Global
Forest Management. Paper No. 4.2.
Bangkok, Thailand.
Makundi. W. R. and J. A. Sathaye 1992. Carbon
omission and sequestration in tropical
forests: A comparative analysis using
country studies Draft Summary Report of
the Research Network on Tropical Forestry
and Global Climate Change. Lawrence
Berkeley Laboralory. Berkeley. CA
126
-------
Section 8
Manabe. S and R. T. Wetherald. 1987. Large
scale changes in soil wetness induced by an
increase in carbon dioxide. Journal of
Atmospheric Science 44 1211-1235.
Mann, L. K. 1986 Changes in soil carbon
storage after cultivation. Soil Science
142.279-288.
Marks, D. 1990 A continental-scale simulation of
potential evapotrarispiration for historical and
projected doubled-C02 climate conditions.
In H. Gucinski, D. Marks, and D. P. Turner
(eds) Biospheric Feedbacks to Climate
Change The Sensitivity 01 Regional Trace
Gas Emissions, Evapotranspiratiori, and
Energy Balance to Vegetation
Redistribution Status of Ongoing Research.
US Environmental Protection Agency.
EPA/600/3-901078.
Marland. G 1989 The role of forest in
addressing the CO 2 greenhouse In: J. C.
White (ed) Global climate change linkages
acid rain, air quality and siratosphenc ozone
Elsevier Science Publishers, Ltd., The
Netherlands.
Marland, G, T. A Boden, R. C Griffin, S. F.
Huang. P Kanciruk, and T. R. Nelson. 1989.
Estimates of C02 emissions from fossil fuel
burning and cement manufacturing, based
on the United Nations energy statistics and
the US Bureau of Mines cement manu-
lacturing data ORNLICDIAC-25, NDP-030.
Mather, A. S. 1990 Global forest resources
Timber Press, Portland. OR.
McG Tegart. W J. 1992. The relationship
between impacts of climate change and
response strategies (with reference to
forestry and agriculture) In. Climate Change
Proceedings of the Intergovernmental Panel
on Climate Change (IPCC) Workshop on
Assessing Technologies & Management
Systems for Agriculture & Forestry in
Relation to Global Climate Change. January
20-23, 1992, Canberra, Australia.
Melillo, J. M., T. V. Callaghan, F. I. Woodward. E.
Salati, and S. K. Sinha. 1990. Effects on
ecosystems. In: J. T. Houghton, G. J.
Jenkins, and J. J. Ephraums (eds.). Climate
Change. the IPCC Scientific Assessment.
Cambridge University Press, Cambridge,
UK.
Mills, J. R. 19B8 Inventory, Growth, and
Management Assumptions for the 1989
RPA Timber Assessment, Part I-The South
and Part Il-The North and West. Reports on
file at the Pacilic Northwest Research
Station, Portland, OR.
Mills, J. R. and J. C. Kiricaid. 1992. The
aggregate timberland assessment system--
ATLAS. a comprehensive timber projection
model. USDA Forest Service General
Technical Report PNW-GTR .281.
Mooney, H. A., P. M. Vitousek, and P. A.
Matson. 1987. Exchange of materials
between terrestrial ecosystems and the
atmosphere. Science 238:926-932.
Mooney, H. A., B. G Drake. R. J. Luxmoore, W
C. Oechel. and L. Pitelka. 1991. Predicting
ecosystem responses to elevated C02
concentration. BioScience 41(2): 96.104
Moore. B., R. D. Boone, J. E. Hobbie, R. A.
Houghton. J M Melillo, B. J. Peterson. G. R
Shaver. C. J. Vorosmarly, and G. M.
Woodwell 1981. A simple model for analysis
of the role of terrestrial ecosystems in the
global carbon cycle. In: B. Botin (ed
Modeling the Global Carbon Budge. SCOPE
16. John Wiley and Sons, New York, NY.
Moulton, R. J and K. R. Richards. 1990. Costs
of sequestering carbon through tree
planting and forest management in the
United States. USDA Forest Service Genera!
Technical Report WO-58
NAS (National Academy of Science). 1991.
Policy implications of greenhouse warming
Report of the Mitigation Panel. National
Academy Press, Washington, DC.
Neilson, R. P. 1990. Biosphere feedback during
climate change. In: G. A. King, J. K. Winjum,
R. K. Dixon, and L. Y. Arnaut (eds.)
Response and Feedbacks of Forest
Systems to Global Climate Change. US
Environmental Protection Agency, Corvallis,
OR. USA.
Neilson, R. P., G. A. King, R. 1.. DeVelice, J.
Lenihan, D. Marks, J. Do h, W. G. Campbell,
and 3. Glick. 1989. Sensitwity of ecological
landscapes and regions to global climate
change. EPA/600/3 ’89/073. NTIS No
P590 120 072/AS. Washington. DC.
National Technical Information Service.
127
-------
Section 8
Noordwijk Conference Report. 1989 The
Noordwijk Declaration on Climate Change
Ministerial Conference on Atmospheric
Pollution and Climate Change, Noordwijk,
The Netherlands
Norby, R. J., C. A Gunderson, S D.
Wullschleger, E. G. O’Neill, and M. K.
McCracken. 1992. Productivity and
compensatory responses of yellow-poplar
trees in elevated C02. Nature 357:322324.
Office of Technology Assessment (OTA). 1991.
Changing by degrees: steps to reduce
greenhouse gases US Congress Office of
Technology Assessment.
Olson. J S.. J A. Watts. and L J. Allison 1983
Carbon in live vegetation of major world
ecosystems. ORNL-5862. Oak Ridge
National Laboratory, Oak Ridge. TN.
Overpeck, J T., D Rind, and R. Goldberg 1990.
Climate-induced changes in forest
disturbance and vegetation. Nature 343:51-
53.
Palm. C A, A A Houghion. J M Melillo, and 0.
L. Skole 1986 Atmospheric carbon dioxide
from deforestation in southeast Asia
Biotropica 18.177-188
Parks, P. J 1992 Opportunities to increase
forest area and timber growth on marginal
crop and pasture land. In. R. N Sampson.
and D Hair (eds) Forests and Global
Warming American Forestry Association,
Washington. DC.
Parks. P J and Hardie, I. W 1992. Least-cost
forest carbon reserves: cost-effective
subsidies to convert marginal agricultural
land to forests Report to American Forestry
Association, Washington, DC.
Partori, W. J.. J. W. B. Stewart, and C. V. Cole.
1988 Dynamics of C, N. P and S in
grassland soils. A model. B,ogeochemislry
5109-131.
Peters. C. M., A. H. Gentry. and R. 0.
Mendelsohn. 1989. Valuation of an
Amazonian rainlorest Nature 339:655-656.
Plantinga, A. J. and A. A. Birdsey. tn press.
Carbon fluxes resuiting from US private
timberland management. Climatic Change.
Poore, M.. P. Burgess, J Palmer. S. Rietbergen,
and 1. Synotl. 1990 No Timber without
trees: Sustainability in the tropical forest.
Earthscan Press, London, UK.
Post, W. M., W R Emanuel, P. J Zinke, and A
G. Stangenberger. 1982. Soil carbon pools
and world l e zones. Nature 298 156-159.
Post, W. M., T-H Peng. W. R. Emanuel, A W
King. V. H. Dale. and 0 L. DeAngelis. 1990.
The global carbon cycle. American Scientist
78:3 1 0-326.
Prentice, K. C. 1990. Bioclimatic distribution of
vegetation for GCM studies Journal of
Geophysical Research 95.811-830.
Prentice. K. C., and I. V. Fung. 1990. Bioclimatic
simulations test the sensitivity of terrestrial
carbon storage in perturbed climates Nature
346.48-51.
Prentice. I. C., W. Cramer. S P. Harrison, A.
Leen,arts. A. A. Monserud. and A. Solomon.
1992 A global biome model based on plant
physiology and dominance, soil properties
and climate. Journal of Biogeography
19 117-134.
Prentice. I. C., M. 1. Sytes, and W. Cramer In
press. A simulation model for the transient
effects of climate change on forest
landscape Ecological Modeling 65:51.70.
Rind, 0. 1984. The influence of vegetation on
the hydrologic cycle in a global climate
model In. J. E. Hansen, and T. Takahashi
(eds). Climate Processes and Climate
Sensitivity Geophysical Monograph Series.
American Geophysical Union, Washington.
DC
Robinson. 0. A., M. Wilson, G. Kukia. and A
Henderson-Sellers. 1983. Observation of
surface albedo and its variation for climate
models. In: Proceedings of the 17th
International Symposium on Remote
Sensing of Environment. ERIM, Ann Arbor,
Ml.
Row, C. and R. B. Phelps. 1990. Tracing the flow
of carbon through the US forest products
sector. Proceedings of the 19th World
Congress of the International Union of
Forest Research Organizations. August 5.
1990, Montreal, Canada,.
128
-------
Section B
Rubiri. E S. R N Cooper. R A Frosch, T. H.
Lee. G Marland, A H Rosenfeld, and D. D.
Stine. 1992 Realistic mitigation options for
global warming Science 257.148-149, 261-
266.
Running. S W and J G Coughlan. 1988. A
general model of forest ecosystem
processes for regional application.
Ecological Modeling 42 125-154.
Running. S W and R. R. Nemani 1988
Relating seasonal patterns of AVHRR
vegetation index io simulated
photosynthesis and transpiration of forests
in different climates Remote Sensing of the
Environment 24 347-351.
Running, S W and E. A. Hunt, Jr. In press
Generalization 01 a forest ecosystem
process model for other biomes. biome-
BGC, and an application for global-scale
models In: J. A Ehleringer and C. Field
(eds.) Scaling Processes between Leaf and
Landscape Levels.
Salati, E. 1986 Amazon forest and the
hydrological cycle. In. C Rosenzweig. and
R Dickinson (eds). Climate-Vegetation
Interactions. UCAR and OIES, Boulder, CO.
Sampson. P N and D. Hair (eds). 1992. Forests
and Global Change Volume One
Opportunities for Increasing Forest Cover
American Forestry Association. Washington.
DC
Samuelson. P. A 1952 Spatial price equilibrium
and linear programming American Economic
Review 42.283-303.
Sanchez. P. A. 1990 Deforestation reduction
initiative An imperative for world
sustainability in the twenty-first century In.
A. F. Bouwnian (ed). Soils and the
Greenhouse Effect. Ed Wiley, New York,
NY.
Schlesinger, M. E. and J. F. B. Mitchell. 1985.
Model projections of the equilibrium climatic
response to increased carbon dioxide. In: M.
D. McCracken and F. M. Luther (ads.).
Protecting the Climatic Effects of Increasing
Carbon Dioxide. DOEIER-0237, US
Department of Energy, Washington. DC.
Schlesinger. M. E. and 2 C Zhao. 1989
Seasonal climate change introduced by
doubled C02 as simulated by the OSU
atmospheric GCM/mixed layer model
Journal of Climatology 2.429-495.
Schlesinger, W H. 1984. Soil organic matter A
source of atmospheric C02. In: G. M.
Woodwell (ad.). The Role of Terrestrial
Vegetation in the Global Carbon Cycle
Measurement by Remote Sensing. John
Wiley & Sons, New York, NY.
Schlesinger. W H. 1986 Changes in soil carbon
storage and associated properties with
disturbance and recovery In: J. R. Trabalka
and 0. E. Reichle (ads) The Changing
Carbon Cycle: A Global Analysis Springer-
Velag. New ‘York. NY.
Schlesinger, W H. 1990. Evidence from
chronosequence studies for a low carbon-
storage potential of soils Nature 348232-
234.
Schneider, S. H. t989a. The changing climate
Scientific American 261.70-79
Schneider, S H. 1989b Global Warming Are
We Entering the Greenhouse Century
Sierra Club, San Francisco
Schroeder, P. E. and L. Ladd. 1991. Slowing
the increase of atmospheric carbon dioxide
a biological approach. Climatic Change
19.283-290.
Schroeder. P. E., R. K. Dixon and J K Winjum
In press. An assessment of forest
management and agroforestry to sequester
and conserve atmospheric CO 2 ljnasylva
Soil Conservation Service (SCS) 1987 Basic
statistics, 1982 National Resources
Inventory Statistical Bulletin Number 756.
Iowa State University, Ames, IA
Sed o, R. A. 1989. Forests a tool to moderate
global warming? Environment 31.14-20.
Temperate forest
global carbon cycle.
Sedjo, R. A. and A. M. Solomon. 1989. Climate
and forests. In: N. J. Rosenberg. W. E
Easterling, III, P. A. Crosson. and J.
Darmstadter (eds.). Proceedings of a
Workshop. Washington DC, June, 1988
Sedjo, P. A. 1992.
ecosystems in the
Ambao 21:274-277.
129
-------
Resources for the Future, Inc., Washington.
DC
Sessions, J (coordinator). 1991 Timber for
Oregon’s tomorrow The 1989 update.
Forest Research Laboratory, College of
Forestry. Oregon Stale University. Corvallis,
OR.
Shugart, H. H. 1984 A Theory 01 Forest
Dynamics Springer-Verlag. New York, NY.
Shugarl, H. H. and D. C. West. 1977.
Development of an Appalachian deciduous
forest succession model and its application
to assessment of the impact of the chestnut
blight. Journal of Environmental
Management 5161-170.
Shukla, J., C Nobre, and P. Sellers. 1990
Amazon deforestation and climate change
Science 247 1322-1325
Simpson, L G and D. B. Bollcin 1992
Vegetation, the global carbon cycle, and
global measures In D A Dunnette and R.
J. O’Brien (eds.). The Science ol Global
Change The Impact of Human Activities on
the Environment American Chemical
Society. Washington, DC
Skog. K, I. Durbak, .3. Howard, H Speller, D.
Adams, and R Haynes. 1991 Potential tim•
bet market impact of current wood utilization
research Forest Products Journal 41 44-52.
Smrth. T. M., H H. Shugart, G. B Bonan, and . 3
B Smith 1992a Modeling the potential
response of vegetation to global climate
change. Advances in Ecological Research
22.93-116.
Smith, J.B and D. Tirpak (eds.). 1989. The
potential effects of climate change on the
US. EPA-230•05-89-050, US Environmental
Protection Agency, Washington, DC.
Smith, T. M., R R. Leemans, and H. H. Shugart.
1992a Sensitivity of terrestrial carbon stor-
age to CO 2 induced climate change: com-
parison of four scenarios based on general
circulation models. Climate Change 21:367-
384.
Smith, 1’. M., J. F. Weishampel, H. H. Shugart,
and G. 8 Bonan. 1992b. The response of
terrestrial C storage to climate change:
Modeling C dynamics at varying temporal
and spatial scales. Water, Air, and Soil
Pollution 64:307-326.
Solliris, P. and G. Spycher. 1983 Processes 01
soil organic-matter accretion at a mudilow
chronosequence, Mt Shasta, California.
Ecology 64:1273-1 282.
Solomon, A. M. 1986. Transient response of
forest to C02-induced climate change.
simulation modeling experiments in eastern
North America. Oecologsa 68:567-579.
Strain, B. R. arid J. D. Cure. 1985. Direct effects
of increasing carbon dioxide on vegetation
US Department of Energy report DOEIER-
0238.
Surano, K. A., P. F. Daley, J. L. .3. Houpis, J H.
Shim, .3. A. Helms, R. J. Palassou, and M P.
Costella. 1986. Growth and physiological
responses of Pinus poriderosa Douglas ex
P. Laws. to long-term elevated C02
concentrations Tree Physiology 2.243-259.
Tans, P. P., I. Y. Fung. and T. Takahasi. 1990.
Observational constraints on the global
atmospheric C02 budget Science
247.1431-1438.
Tedder, P. L., R. N. LaMont, and J C. Kincaid
1987. The timber resource inventory model
(TRIM)’ A projection model for timber supply
and policy analysis. USDA Forest Service,
Pacitic NW Research Station, Portland. OR.
Tepley. J. and K. Green. 1991. Old growth
forest. How much remains? Geographic
Information Systems 1 .23-31.
Tinus, R. W. 1972 C02 enriched atmosphere
speeds growth of ponderosa pine and blue
spruce seedlings. Tree Planters’ Notes
23.1 2-18.
Tissue, 0. T. and W. C. Oechel. 1987
Response to Eniophorum vaginatum to
elevated C02 and temperature in the
Alaskan tussock tundra. Ecology 68(2)401-
410.
Tolley, L. C. and B. R. Strain. 1984. Effects of
C02 enrichment and water stress on growth
of Liquidamber sryraciflua and Pinus taeda
seedlings. Canadian Journal of Botany
62:2135-2139.
130
-------
Section 8
Trexier. M. C 1991a Minding the carbon store’
Weighing US forestry strategies to slow
global warming World Resource Institute,
Washington. DC
Trexter. M. 1991b Proceedings of a technical
workshop to explore options for global forest
management In D. Howlett and C. Sargent
(eds) Estimating tropical biomass futures. A
tentative scenario. Bangkok. Thailand, April
24. 1991. International Institute for
Environment and Development. London.
UK.
Tucker. C. J. I V. Fung, C. 0. Keeling, and R. H
Gammon. 1986 Relationship between
atmospheric C02 variations and a satellite-
derived vegetation index. Nature 319:195-
199
Turner, D P., G J Koerper, H Gucinski, C. E
Peterson, and R K. Dixon In press.
Monitoring global change Comparison 01
forest cover estimates using remote sensing
and inventory approaches Environmental
Monitoring and Assessment
UNCED (United Nations Conference on
Environment and Development 1992
Adoption of agreements on environment
and development Agenda item 97 pp
Urban, D L and H H Shugart 1989 Forest
response to climatic change a simulation
study for southeastern forests In: J. B.
Smith and D Tirpak (eds) The Potential
Eflects of Global Climate Change on the
United States US Environmental
Protection Agency, Washington. DC.
USDA Forest Service. 1965. Timber Trends in
the United States Forest Resources Report
17.
USDA Forest Service. 1974. The outlook for
timber in the United Stales. Forest
Resources Report 20.
USDA Forest Service. 1982. An analysis of the
timber situation in the United States 1952-
2030. Forest Resources Report 23.
US Department of Commerce (DoC). 1991.
Statistical Abstract of the United States,
1992.
Van Hooser, 0 D. and D. C. Chojnacky 1983.
Whole tree volume estimates for the Rocky
Mountain states. Bulletin INT.29, US
Department of Agriculture. Forest Service.
Intermountain Forest and Range Experiment
Station, Odgen. UT.
Vogt, K. A, C. C. Grier, and D. .1. Vogt 1986
Production, turnover, and nutrient dynamics
of above- and below-ground detritus of
world forests. Advances in Ecological
Research 15.303-377.
VOIZ, H.. W. U Kriebitzsch, and T. W. Schneider.
1991. Assessment of potential, feasibility
and costs of forest options in the temperate
and boreal zones. In: 0. Howlett and C.
Sargent (eds.). Proceedings of a technical
workshop to explore options for global forest
management Bangkok. Thailand. April 24.
1991. International Institute for Environment
and Development, London, UK.
Waddell, K L., D. 0. Oswald, and D. S Powell
1989 Forest Statistics of the United States.
Pacific Northwest Research Station. USDA
Resource Bulletin PNW-RB-168.
Waring. P H. and W. H. Schlesinger. 1985
Forest Ecosystems Concepts and
Management. Academic Press, Inc.
Orlando, FL.
WEFA (Wharton Econometrics Forecasting
Associates). 1987. Long-term alternative
scenarios and 25-year extension Annual
model Bala Cynwyd, PA.
Weinstein, D. A.. and R. Beloin. 1990
Evaluating effects of pollutants on
integrated tree processes: A model of
carbon, water and nulnent balances. In A K
Dixon, P. S. Meldahi, G. A Ruark, and W G
Warren (eds.). Process Modeling of Forest
Growth Responses to Environmental Stress
Timber Press, Portland, OR
Weinstein. 0. A., P M Beloin, and R. D. Yani
1991. Modeling changes in red spruce
carbon balance and allocation in response to
interacting ozone and nutrient stress Tree
Physiblogy 9:127-146.
Whittaker, R. H. and G. E. Ukens. 1975. The
biosphere and man. In: Primary Productivity
01 the Biosphere. Ecol. Study No. 14.
Springer-Verlag. Berlin, Germany.
Williams. M. 1988. The death and rebirth of the
American forest’ clearing and reversion in
131
-------
Section 8
the United Slates. 1900.1980 In: J. F.
Richards and R. P. Tucker (eds.). World
Deforestation in the Twentieth Century.
Duke University Press, Durham, NC.
Wilson, C A and J. F B Mitchell 1987. A
doubled C02 climate sensitivity experiment
with a global climate model including a simple
ocean. Journal of Geophysical Research
9213315-13343.
Winjum, J. K , R K Dixon and P. E. Schroeder.
1992. Estimating the global potential of
forest and agroforest management practices
to sequester carbon. Water. Air and Soil
Pollution 64 213-227.
Winjum. J. K.. R K. Dixon and P. E. Schroeder.
In press An assessment of forest
management practices for sequestration of
carbon I Estimating the global potential
Journal of World Forest Resource
Management.
Woodwell. G. M., J. E. Hobbie. R A. Houghton,
J. M. Melillo, B Moore. B. .J. Peterson. and
G. R. Shaver. 1983. Global deforestation
Contribution to atmospheric carbon dioxide.
Science 222.1081-1088.
WRI (World Resources Institute). 1990. World
resources 1990-91. Oxford University
Press, Oxford, UK.
WRI (World Resource Institute). 1992. World
Resources 1992 - 93. , Oxford University
Press, New Yod . NY.
Wright, L. L., R. L. Graham, A. F. Tumhollow, and
B. C. English. 1992. Opportunities for short
rotation woody crops and the potential
impacts on carbon conservation. In: R. N.
Sampson and 0. Hair (eds.). Forests and
Global Warming American Forestry
Association, Washington, DC.
132
-------
SECTION 9. SUPPORTING DOCUMENTS
A. Uncertainty In Country-wide Forest Blomass Estimates 1
Charles E. Peterson 2 . 3 and David P. Turner 2
ABSTRACT : Country-wide estimates of forest biomass are the major driver for
estimating and understanding carbon pools and flux, a critical component of global
change research. Important determinants in making these estimates include the areal
extent of forested lands and the associated biomass. Estimates for these parameters
may be derived from surface-based data, photo interpretation or satellite remote
sensing, with varying degrees of uncertainty. Ground data is typically aggregated by
forest type, stand age, productivity level, and ownership. Survey priority is usually
given to regions and forest types with timber of commercial value, such that
information on understory biomass and forested lands of low commercial value is
either absent or of limited reliability. Furthermore, information on below ground
biomass, which is costly and time-consuming, is not generally collected. Typically,
uncertainty in survey slatistics increases as the level of post-stratification increases
because of reduced sample size. Likewise, literature-based expansion factors also
add to the uncertainty of a final estimate because of the often unknown spatial
inference for those factors. Estimates based on modeled processes may provide
relatively limited information on uncertainty. The discussions in this paper are based
on research funded by the US Environmental Protection Agency, and data provided by
the US Forest Service.
Keywords: reliability, inference, sampling error, sensitivity
INTRODUCTION
The major variables requiring consideration in the assessment of uncertainty
associated with country-wide carbon budgets include forest land area, the carbon from
biomass associated with that area, and the soil carbon. Of these three major areas,
we are concentrating here on reliability of biomass estimates insofar as they affect
‘Paper for presentation at IPCC AFOS-workshop on carbon balance of world’s
forested ecosystems: towards a global assessment, held May 11-15, 1992 at the
University of Joensuu, Joensuu, Finland.
2 ManTech Environmental Technology, Inc., US EPA Environmental Research
Laboratory, 200 SW 35th Street, Corvallis, Oregon 97333, USA.
3 Currently with the USDA Forest Service, Pacific Northwest Research Station, P0 Box
3890, Portland, Oregon 97208.
133
-------
Section 9.A
carbon pools. It is essential to understand the effectiveness or reliability of the
(biomass) estimation procedures if modelers are to manage the expectations that
others may have from their interim and final products. Thus, our goal is not to provide
absolute values for error in a particular case, but to use some empirical examples to
illustrate how one may arrive at estimates of reliability for biomass as the major
component in determining carbon pools and flux. Although biomass is the example
used for this workshop, the ideas are easily extended to other components of the
carbon budgeting process.
As part of the US Environmental Protection Agency’s Climate Change Program, we
are addressing the area within the USA that is bounded by the conterminous “lower
48” states. In each state, thousands of photo points from current aerial photographs
are used to stratify and compute expansion factors for the field (i.e., located on the
ground) sample data (e.g., see Bassett and Oswald 1983). The effectiveness of the
estimation procedure for area and volume, that is, how representative the estimates
are for the real population, is primarily a function of two important measures: precision
(sampling error) and the confidence level. An additional component in variation called
nonsampling errors (e.g., mistakes in choosing a design, data processing, analytical
mistakes, etc.) are usually not known, so we rely on the sampling precision and
associated confidence (probability) level to indicate reliability for the survey estimates
(HusCh et al. 1982). We also recognize that there are many other sources of
uncertainty, such as photo/map interpretations, biomass equations, volume equations,
and measurement error. Because most of these kinds of uncertainty are not quantified
along with their published results, we decided to select the examples used in this
paper.
For a US carbon budget, we are using volume and biomass estimates from USDA
Forest Service inventory and resource data that are summarized regionally and are
derived from a statistically-based survey. Because a major objective is to provide
carbon budget information by state, and by types/classes according to ownership.
species, and productivity, it is necessary to examine the possible tradeoffs in reliability
of estimates when stratifying for those kinds of information. Volume is used instead of
biomass as an example, because (i) volume is inventoried nationwide (biomass is not)
and (ii) problems of reliability are equally applicable to both measures.
APPROACH
Precision refers to the variation among repeated sample estimates (e.g.. the clustering
of observations about the average), and is usually expressed as a standard error of
the mean (average) of those estimates. Confidence intervals (CIs), expressed by the
sample mean ± one or two standard errors, are computed for state resource reports
using data sampled from US forested timberlands. Although the confidence intervals
vary according to the magnitude of the estimate and variation in the attribute being
measured, they provide some approximate reliability for the reported statistics. In this
paper, we otter examples of volume and area estimates accompanied by 67% CIs,
where the odds are two out of three (67% probability) that the true value (timberland
area or volume) will fall inside the range described by the sample mean ± 1 standard
error. Although confidence intervals vary with both the size of the estimate and the
variance of the item being sampled, we have included examples from data published
134
-------
Section 9.A
in numerous state reports to illustrate how the width of the confidence band changes
with the size of the estimate.
In the procedure for estimating biomass over large regions, we generally have data or
information that come from three primary sources:
(i) Survey data -. usually statistically based (i.e., has sampling error) and therefore
provides some estimate of regional precision:
(ii) Research plots -- results summarized in the literature and hopefully providing
some biological understanding but probably no estimates of regional precision;
(iii) Modeling -- may be statistical with some measure of precision, or for
understanding or incorporating processes having no precision estimates and
perhaps limited generality.
It is important to keep these three sources of information in mind, since they are
generally used in combination with each other. As you move from survey information
to literature and finally to modeling, you generally lose the ability to track or estimate
precision. As an example, our biomass estimates for timberlands are a combination of
survey (aboveground biomass), literature values (below-ground biomass, understory,
and forest floor), and modeling efforts (e.g., woody debris).
RESULTS
For the examples of reliability presented here, we are using percent sampling error,
defined as one standard error of the mean divided by the average (mean) value. In
the Eastern US, the percent sampling errors associated with individual state
timberland volume estimates are quite low (Figure 9.A.1). The situation in the West is
quite similar (Figure 9.A.2), although both the averages and ranges of percent
sampling error associated with timberland volume were greater than in the East. In
general, for most state reports that displayed 67% CIs, the percent sampling error was
around ±5% for areas by major owner groups and when timber volumes were lumped
as softwoods or hardwoods (e.g., Lloyd et al., 1986). However, a state-level sampling
error of ±1.4% might increase up to ±50% when volume is stratified by forest type (FIA
Staff 1985).
Further examples of percent sampling errors taken from data reported in Gedney et al.
(1989) and Bassett and Oswald (1983), are given in Figures 9.A.3 and 9.A.4, to
illustrate how reliability changes with the size of the estimate. Note that smaller
volume estimates are associated with greater uncertainty (Figure 9.A.3), and in the
case of area (Figure 9.A.4), the reliability differs as to whether the data is partitioned by
owner or by type, with a trend of increased sampling error by type when estimating
strata of smaller forest land area.
135
-------
Section 9.A
PERCENT ERROR FOR TIMBERLAND VOLUME OF EASTERN STATES
0.09
0.08
CO CO CO (0 CO N. N. C’) C )
C.,
(fl () C) C) . CO C)
w 5 p U) C’) CO C J
( 0 CO r- I C ) ‘ C)
(0 (0 (0 CO ü CO CO C’)
N. C) O-C) P-C) C ) C)
C’) r- C) ir ) N. W C’)
It) C) N. C’) CO (0 C ’) (0
(Q C N. C’ J C’)
N. (0 CO CO C’) C’) C
Timberland Volume (l000s Cu’ meters)
Figure 9.A.1. Percent error (1 s.e. divided by the average value) associated with state
level and sub-state level timberland volume estimates the Eastern US.
I
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
iiiiiiuiiiiiii iliI.,
, I I I I I I I I I I I I I I
I I
I I
136
-------
Section 9.A
PERCENT ERROR FOR TIMBERLAND VOLUME OF WESTERN STATES
0.16
P fl IA ______
0
L i i 0.12
0 )
,E 0.10
Timberland Volume (l000s cu. meters)
. ‘ 0 0
r t (b
o t - in
N in
in in in
Figure 9.A.2. Percent error (1 s.e. divided by the average value) associated with state
level and sub-state level timberland volume estimates in the Western US.
137
-------
Section 9.A
RELIABILITY OF TIMBERLAND VOLUME ESTIMATES
0.50
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
o 0 0 0 10
N I D
Timberland Volume (millions cu. m.)
Figure 9.A.3. Percent error (1 se. divided by the average value) associated with
timberland volume estimates within a state.
0
I .
w
C)
0.
E
C l )
C
C .,
I . .
0.
HH.
C ) N
138
-------
Section 9.A
SAMPLING ERROR AS F (AREA, OWNERSHIP, TYPE)
0 60
o 50
0 40
0 30
020
o 10
000 r .t II_11_III
600 400 320 240 160 80 10
Timberland Area
(l000s hectares)
Figure 9.A.4. Percent error (1 s.e. divided by the average value) associated with
timberland area estimates stratified by ownership and type within a state.
I-
0
I.
I-
L i i
0
L.
C)
O Owner
Type
I:
-I -
139
-------
Section 9.A
DISCUSSION
In the above examples, we have illustrated how the sampling errors associated with
forest area and volume estimates from statistically-based surveys are relatively small
(i.e., high reliability) for large areas. However, uncertainty increases substantially if
volume or area are stratified by forest type (the same result is expected for stratifying
by productivity class). Because stratification results in a reduced sampling intensity for
some combinations of species type or class, an accompanying decreased confidence
in the final estimate is expected.
In addition to biomass estimates, most approaches to estimating country-wide carbon
budgets will also incorporate results of site-specific studies from the literature and
model outputs, neither of which can be characterized with the same kind of reliability
that might be expected from a statistically-based survey. A good sensitivity analysis of
the models over a range of input values for a variety of scenarios should be employed
to assess the reliability of the estimation framework. In order for the end results to be
assigned any degree of reliability, even if it is only qualitative, the data origins (i.e.,
measured or simulated), all assumptions or judgments, and sensitivity analyses, must
be fully documented.
Finally, a significant portion of aggregated forest data may not have known reliability.
One is then constrained to assuming reliability for that stratum and thus assuming
reliability for the aggregate total. In the US for example, volume by age class for all
ownership strata has been summarized for the Eastern US. However, similar
estimates for public lands in the Western US are not yet summarized, and may require
some assumptions about their age-class distributions.
RECOMMENDATIONS
It is important to have some idea as to how our estimation procedures might mask or
influence our ability to detect an incremental change in the carbon budget. To that
end, we offer the following considerations for reducing uncertainty in country-wide
forest biomass estimates:
• Define (with glossary) all components identified with the estimation process.
• Examine the proportional contributions of these components, as well as others
(e.g., soils) to the carbon budget.
• Provide some quantitative ranking (importance) of the carbon components
according to their magnitude or direction, and how that ranking might differ
under climate change scenarios.
• Assign some reliability to the estimates associated with each component,
drawing a clear distinction between sources such as statistical sampling error,
quantitative sensitivity analyses, and “expert opinion.”
• Assess the change in reliability when different spatial scales or finer
classifications are employed.
140
-------
Section 9.A
BIBLIOGRAPHY
Bassett, Patricia M. and Daniel D. Oswald. 1983. Timber resource statistics for eastern
Washington. Resource Bulletin PNW-104. Portland, OR: US Department of
Agriculture, Forest Service, Pacific Northwest Research Station. 32 pp.
Forest Inventory and Analysis Research Work Unit (FIA Staff). 1985. Forest Statistics
for Alabama Counties in 1982. Resource Bulletin SO-97. New Orleans, LA: USDA
Forest Service, Southern Forest Experiment Station. 31 pp.
Gedney, D. R.. P. M. Bassett, and M. A. Mel. 1989. Timber resource statistics for all
forest land, except National Forests, in eastern Oregon. Resource Bulletin PNW-
RB-i 64. Portland, OR: US Department of Agriculture, Forest Service, Pacific
Northwest Research Station. 25 pp.
Husch, B., C. I. Miller, and T. W. Beers. 1982. Forest Mensuration, 3rd edn. John Wiley
& Sons. 402 pp.
Lloyd, J. D., J. Moen, and C. L. Bolsir ger. 1986. Timber resource statistics of the
Sacramento resource area of California. Resource Bulletin PNW-134. Portland,
OR: US Department of Agriculture, Forest Service, Pacific Northwest Research
Station. 32 pp.
141
-------
Section 9.A
142
-------
B. Biomass and Carbon for Woodlands In the Western US
Charles E. Peterson 12
The following table represents a compilation of total tree biomass for woodlands
(growth < 1.4 m 3 ha 1 yr 1 ) from most of the western conterminous United States.
References corresponding to the citation numbers given in the table represent USDA
Forest Service state inventories and resource reports that were used to compile the
data, The number corresponds to all computations in the box for a given state. All
data are on file at the US EPA Environmental Research Laboratory in Corvallis,
Oregon.
In the western conterminous states , excluding Washington and Oregon (no data avail-
able), the average above-ground carbon pool is estimated at 1.15 kg rn- 2 (11.5 tons
per ha). Below-ground estimates (from Koch 1989) raise those estimates to about 1.3
kg rn”’ 2 . The average carbon density was quite variable, ranging from 0.5 to 2.1 kg m 2
for Juniper in Wyoming and Colorado respectively, and estimates ranging from 1.0 to
4.0 kg m 2 (average of 2.5 kg m ’ 2 ) for chaparral in California. Average carbon for
hardwoods within California, ranged from 0.7 to 1.9 kg rn” 2 (average is 1.6 kg rn- 2 ).
Although forest woodlands comprise nearly half of the forested lands in the contermi-
nous Western US, they account for less than 5% of the above-ground carbon pool for
ALL forest lands of the conterminous US. While this information may indicate that non-
merchantable forests play a small role in assessing carbon pools, these levels of car-
bon might be a benchmark or indication of the final carbon pools that might be
expected if timberlands, under climate change, were to transition to carbon storage
and productivity of ihose lands now classified as woodlands.
Another exception to the contribution of woodlands to the total carbon pool is in
Alaska, where by virtue of the extremely large area of Jand cJass fied as woodland, the
total carbon above-ground is greater for these forest lands (about 0.57 Pg-C) than the
total carbon pool from Alaskan timberlands (about 0.50 Pg-C).
If the average carbon storage for the woodlands of the weslern US is applied to the
area of woodlands in the conterminous US, then total woodland tree carbon storage is
0.49 Pg-C.
The following references correspond to the citation reference numbers given in Table
9.6.1. These references represent USDA Forest Service state inventories and
resource reports that were used to compile total tree biomass for woodlands (growth <
1.4 rn- 3 ha yrl) in the western conterminous US. The number corresponds to all
computations in the box for a given state. All data are on file at the US EPA
Environmental Research Laboratory in Corvallis, Oregon.
1 ManTech Environmental Technology, Inc., US EPA Environmental Research Laboratory. 200 Sw 35th
Street, Corvallis, Oregon 97333, USA.
2 CurTentIy wfth the USDA Forest Service, Pacific Northwest Research Station, P.O. Box 3890, Portland,
Oregon 97208.
143
-------
TaMp 9 I Compulaluon of carbon lar woodan& bi ie *eslp ,n US
Whim. ToW bole walghi • biccn ci I v. bol, vokm... daad bole voluma and wood duinuty facto,
ToW carbon Is average carbon density (kg ha from unreserved areas nijttuplied times tolal woo ,Jand area (racarvuid plus unrursmryed
7007
5126
0977
27080
6574
46 1
14404
1006
4899
2077 0
078
051
o co
279
066
78
51
90
27 1
66
ID P.J
SUBTOT
611
207
018 0
9407 41957 2.1 3911 21880
9123 12724 I 1721 6256
54680 563 28136
015 S e
075 75
085 85
S bi Fomsl
Type
WY P4
.1
Oak
SIJBTOT
Ajea-lew
A,.a-
Total
Total
ma
Area
Luvuibola
Daadeola
Dansuly 1
bola
bOos
boOs
bOOs
I0 0 0c
Ifl(lOs
Faclor
Weuqhl
acres
acres
l lSC ffi%
cubic ft
cubic ft
kgcl 1
kqtOEO6
Topitwanc l ith
Weight Uni C
kqlOEO6 kqtOEO6
2. 143 2e 031 6 2k 5060 2. 1091 9401 867 24 002 452
691 1 34 ! 291 216127 31292 III 27464 182 14334
Vi I I 16582 1658 1123 2048 093 1006
447 II 76073 12737 1123 9973 036 4099
7771 37 396 313850 46781 10352 533 20690
Total Total Total
Unr C Carbon Cvbon Carbon
kaha t kgIOEO6 k rjm 7 tonslia t
3723
366
28
635
162
02
40
68
10
5913
00
00
00
00
00
00
00
00
00
00
443075
18484
1386
3155 1
1965 5
9815
481 2
743 1
1204
000 538214
217
*1 2982001
I 1133001
5849 21880
7468 6256
331
411500
0
8499 28136
NV
P4
J
oI ler
71561 4. 2896
16841 682
145 25 69
34988811
560305
58954
220065
34080
7613
1407
I II
1123
523256
65917
7415
2348
3252
120
272739 9415 272739
34543 5069 34543
3901 6649 4574
094
051
066
94
51
66
SUBTOT
8985 25 3646
4918140
26*158
596100
5729
311183 8558 311856
086
86
—.
UT
P4
8846
354
92399
461635
124
124
Aspesi
Coflnwd
900
675
35
15
5075
5075
18556
1417
051
051
51
51
OSIPIHOW
15362
32
5075
32208
051
51
FIR-SP It
3917
91
12399
2011 I
124
124
OF
1956
6
12399
10116
124
124
PP
959
99
12399
4907
124
124
LP
1481
193
12399
8400
124
124
othe SFT
24
94
12399
1275
124
124
SIJBTOT
12205 4287
0
0
10097 558629
1 09
109
CO
P-J
6. 46516] IS 1859 1955
Sb 383321
47532
1407
60677
441
31636 4692 91060
047
47
j
450 5 16 183
581993
75445
199
73553
8*0
38367 21615 39629
298
216
Oak
7052 218 294
116460
11769
1123
14400
365
7085 2614 7691
026
26
Other
135 28 7
11471
1068
1123
1407
009
691 13135 867
139
131
SUBTOT
5821 I 212 I 2442
1098445
136202
150037
1705
77783 6896 140047
069
57
1 .2
P4
J
7S 9070 Tb 6726 3943
2109 0 854
7c 5273209
688145
7d
9062555
140948
9407
111
891484 7.
92911
1087
3706
464520 92653 495969
48607 5695 48607
127
051
927
57
0 . 9
009 282 310
397200
58264
1123
51149
675
25147 6990 25887
070
70
I I acqJ I s ,
126* 91 517
149481
28235
1123
19958
140
9006 1911 9877
019
19
0 59w, ,
154 13 63
95755
4919
9923
2232
032
1097 1769 1107
018
I I
SUBTOT
13190 7093 5746
6523793
1302421
1057140
5119
519170 10059 581447
101
102
MT
PJ
j
Oak
Oitmr
S. 0 Sb 82 3
7135 I 33
0 I 022 0
0 L 124 1
SC [
30206
0
0
2372
III
00 IS
3616
00
00
000
053
000
000
00 00
1807 6534 2117
00 00
00 00
065
65
SIJBTOT
7135 2063 37
30206
2372
3696
053
1887 6534 2177
065
56
-------
NM P.J
J
OThN
SUBTOT
CA ch pd 1
W Juniper I
P.J
N ce l hdwd
N mm , tided
Sacmito Med
C coact tided
S tided
Total
Total tidede (GS1 45
sUØTOT
6845 506
766000
347000
776000
1694000
I 208000
4188000
10 11 69426O0
7944000
1311215
42299
10346
29
I393 38
15423
66452
10074
2464
14361
489705 454 451
16159 067 67
17951 110 ItO
63 022 22
523879 144 144
739977 246 2445
23010 046 46
46276 lOS 105
43167 494 494
20391 069 69
44794 066 66
406662 447 447
70657 069 69
285656 095 96
414202 138 139
108067 I 1 57 157
Stibts
A K Aiasha
AZ Aniona
CA Calulornia
c0 Colotacto
m
MT Montana
NM N ,uwMpxlco
NV Nevada
OR Oiegon
UT Utah
WA Wasbtaglon
WY Wyoning
Specla
Co1b oJ oollonaood
OF Douglac-Ir
FIR SPA Spwce
Ndwd herdwoode
J jtai lpec
LP lodgepole pine
o0 edl1WS other ha,dwoods
oIIwrSFT other softwoode
P J PSOYOI t juniper
PP ontburor.a pm.
Ama-maw Ljnmaerv.d ama
Area mc Rimcruvcud area
Un , C Carbon I cr imieserved tende
N coacth#ad No .1 1, coast herdwocde
N tetcu, tided North .,tenc , hardwoods
Sacrndo hded Sacraneurto hardwoods
C toad tided Central coadt hardwoods
S hdwd Southern PIerIMOOde
78652 lb 1362 3238 lc [ 5253741
5961 32 243 232594
4020 23 164 264540
35 37 3 4 1 16
88667 US) 3647 5751316
3044
16c 520
I 440
3604 3844
1255 30
1029 545
104 545 12
667 45
1600 75
1570 226
2135 448
403800
597600
225
296
601
727
$045
2975
1407
923699
2432
481374
I II
30542
3364
16072
1123
34240
2453
17854
1123
59
988509
005
6081
30
515330
III
00
14922
358066
23352
24550
4598
1O
1407
94092
43807
10520
Ii? )
86022
42237
19150
1423
38068
19433
6982
4123
81145
42788
6575
1123
189809
9324$
14675
1123
435659
66609
6759
0
537692
000
264007
9530
0
0
779654
908558
000
000
382810
808039
13849
15681
12733 4438 6949
TOTAi.S 83.7679 5.9827 29.228 285.0436 11.045 325.961 1
(Al slates ascept
WA. OR. and AM)
Average 115 115
c i ’
SUBTOT Subtotal
-------
Section 9 .B
REFERENCES for Table 9.B.1
la. Van Hooser, Dwane D., Chojnacky, David C. Whole tree volume estirnates.for
the Rocky Mountain States. Resource Bulletin INT-29, Table 1. Ogden, UT:
USDA Forest Service, lntermountain Forest and Range Experiment Station;
1983. 69 pp.
1 b. Van Hooser, Dwarle D., Chojnacky, David C. Whole tree volume estimates for
the Rocky Mountain States. Resource Bulletin INT-29, Equations (2)-(6),
Ogden, UT: USDA Forest Service, Intermountain Forest and Range Experiment
Station; 1983. 69 pp.
2a. Green, Alan W.: Conner, Roger C. 1989. Forests in Wyoming. Resour. Bult.
1NT-61, Table 12. Ogden, UT: USDA Forest Service, lntermountain Research
Station. 91 pp.
2b. Green, Alan W.; Conner, Roger C. 1989. Forests in Wyoming. Resour. Bull.
INT-61, Table 13. Ogden, UT: USDA Forest Service, Intermountain Research
Station 91 pp.
2c. Green, Alan W.; Conner, Roger C. 1989. Forests in Wyoming. Resour. BuU.
INT-61, Table 57. Ogden, UT: USDA Forest Service, Intermountain Research
Station. 91 pp.
2d. Green, Alan W.: Conner, Roger C. 1989. Forests in Wyoming. Resour. Bull.
INT-61, Table 47. Ogden, UT: USDA Forest Service, Intermountain Research
Station. 91 pp., plus equations in reference #1 above.
3a. Benson, Robert E.; Green, Alan W.; Van Hooser, Dwane D. Idaho’s forest
resources. 1987. Resource Bulletin INT-39, Table 67. Ogden, UT: USDA
Forest Service, lntermounttain Research Station. 114 pp.
3b. Benson. Robert E.; Green, Alan W.; Van Hooser, Dwane D. Idaho’s forest
resources. 1987. Resource Bulletin INT-39, Table 68. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 114 pp.
3c. Tables on Numbers of live trees (circa 1980), received April, 1991 from Sharon
Woudenberg plus equations in reference #1 above.
4a. Nevada Report. Preliminary. In Press. Resource Bulletin Table 5. Ogden, UT:
USDA Forest Service, Intermountain Research Station.
4b. Nevada Report. Preliminary. In Press. Resource Bulletin. Table 9. Ogden, UT:
USDA Forest Service, Intermountairi Research Station.
4c. Nevada Report. Preliminary. In Press. Resource Bulletin Table 16. Ogden, UT:
USDA Forest Service, ntermountain Research Station.
146
-------
Section 9.B
4d. Nevada Report. Preliminary. In Press. Resource Bulletin Table 8. Ogden, UT:
US Department of Agriculture, Forest Service, Intermountain Research Station;
plus equations in reference #1 above.
5a. Van Kooser, Dwane D.; Green, Alan W. 1983. Utah’s forest resources, 1978.
Resource Bulletin INT-30, Table 15. Ogden, UT: USDA Forest Service,
Intermountain Forest and Range Experiment Station. 58 pp.
5b. Carbon in kg ha for UT are based on average of AZ, NM, and NV rates of
kg had, by forest type. Remaining estimates are based on that estimate.
6a. Benson, Robert E.; Green, Alan W. 1987. Colorado’s timber resources.
Resource Bulletin INT-48, Table 3. Ogden, UT: USDA Forest Service,
Intermountain Research Station. 53 pp.
6b. Conner, Roger C.; Green, Alan W. 1988. Colorado’s woodland resources on
state and private land. Resource Bulletin INT-SO, Table 18. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 53 pp.
6c. Conner, Roger C.; Green, Alan W. 1988. Co!orados woodland resources on
state and private land. Resource Bulletin INT-50, Table 20. Ogden, UT: USDA
Forest Service, lntermountain Research Station. 53 pp.
6d. Conner, Roger C.: Green, Alan W. 1988. Colorado’s woodland resources on
state and private land. Resource Bulletin INT-50, Table 17. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 53 p.: plus equations in
reference #1 above.
7a. Conner, Roger C.; Born, J. David; Green, Alan W.; O’Brien, Renee A. 1990.
Forest Resources of Arizona. Res. Bull. INT-69, Table 43. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 92 pp.
7b. Woudenberg, Sharon. Personal communication (April, 1991). Table of Area of
reserved Arizona woodland, excluding NFS.
7c. Conner, Roger C.; Born, J. David; Green, Alan W.; O’Brien, Renee A. 1990.
Forest Resources of Arizona. Res. Bull. INT-69, Table 49. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 92 pp.
7d. Conner, Roger C.; Born, J. David; Green, Alan W.; O’Brien, Renee A. 1990.
Forest Resources of Arizona. Res. Bull. INT-69, Table 53. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 92 pp.
7e. Conner, Roger C.; Born, i. David; Green, Alan W.; O’Brien, Renee A. 1990.
Forest Resources of Arizona. Res. Bull. INT•69, Table 46. Ogden, UT: USDA
Forest Service, Intermountain Research Station. 92 p.; plus equations in
reference #1 above.
147
-------
Section 9.B
Ba. Woudenberg, Sharon. Personal communication (April, 1991). Table of Area of
nonreserved Montana woodland, excluding NFS (circa 1989).
Bb. Woudenberg, Sharon. Personal communication (April, 1991). Table of Area of
reserved Montana woodland, excluding NFS (circa 1989).
8c. Woudenberg, Sharon. Personal communication (April, 1991). Table of Volume
on nonreserved Montana woodland, excluding NFS (circa 1989).
Bd. Woudenberg, Sharon. Personal communication (April, 1991). Table of Dead
Volume on nonreserved Montana woodland, excluding NFS (circa 1989).
8e. Woudenberg, Sharon. Personal communication (April, 1991). Table of Numbers
of live trees on nonreserved Montana woodland, excluding NFS (circa 1989):
plus equations in reference #1 above.
9a. Woudenberg, Sharon. Personal communication (April, 1991). Table of Area of
nonreserved New Mexico woodland, including NFS (circa 1987).
9b. Woudenberg. Sharon. Personal communication (April. 1991). Table of Area of
reserved New Mexico woodland, excluding NFS (circa 1987).
9c. Woudenberg, Sharon. Personal communication (April, 1991). Table of Volume
on nonreserved New Mexico woodland, including NFS (circa 1987).
9d Woudenberg, Sharon. Personal communication (April, 1991). Table of Dead
Volume on nonreserved New Mexico woodland, including NFS (circa 1987).
9e. Woudenberg. Sharon. Personal communication (April, 1991). Table of Numbers
of live trees on nonreserved New Mexico woodland, including NFS (circa
1987); plus equations in reference #1 above.
lOa. Bolsinger, Charles L. 1988. Shrubs 0 f California’s chaparral, timberland, and
woodland: area, ownership, and stand characteristics. Resource Bulletin PNW-
RB-160, Table 3. Portland, OR: USDA Forest Service, Pacific Northwest
Research Station. 50 pp.
lOb. Bolsinger, Charles L. 1989. California’s western juniper and pinyon-juniper
woodlands: area, stand characteristics, wood volume, and fenceposts.
Resource Bulletin PNW-RB-166, Table 1. Portland, OR: USDA Forest Service,
Pacific Northwest Research Station. 37 pp.
lOc. Bolsinger, Charles L. 1989. California’s western juniper and pinyon-juniper
woodlands: area, stand characteristics, wood volume, and fenceposts.
Resource Bulletin PNW-RB-166, Table 7. Portland, OR: USDA Forest Service,
Pacific Northwest Research Station. 37 pp.
1 Od. Bolsinger, Charles L. 1988. The hardwoods of California’s timberlands,
woodlands, and savannas. Resource Bulletin PNW-RB-148, Tables 9-13.
148
-------
Section 9.9
Portland, OR: USDA Forest Service, Pacific Northwest Research Station. 148
pp.
lOe. Bolsinger, Charles L. 1988. The hardwoods of California’s timberlands,
woodlands, and savannas. Resource Bulletin PNW-R8-148, Tables 21-25
(growing stock volume). Portland, OR: USDA Forest Service, Pacific Northwest
Research Station. 148 pp.
lOf. Bolsrnger, Charles L. 1988. The hardwoods of California’s timberlands,
woodlands, and savannas. Resource Bulletin PNW-RB-l 48, growing stock
volume (Tables 21-25) expanded by factor of 1.45 (woodland volume ratio of
total tree to growing stock) from Table 31. Portland, OR: USDA Forest Service,
Pacific Northwest Research Station. 148 pp.
log. Assumes 50 tons ha 1 from: Riggan, Philip J. and Paul H. Dunn. 1982.
Harvesting chaparral biomass for energy -- an environniental assessment. Gen.
Tech. Rep. PSW-58. Berkeley, CA. Pacific Southwest Forest and Range
Research Station. Forest Service. USDA Forest Service.
149
-------
Section 9.B
150
-------
C. Woody Debris Budgets for Selected Forest Types In the US
Mark Harmon’
INTRODUCTION
One of the most important questions facing ecologists today is the role of the biota in
the global carbon cycle. The fact that 1 to 3 Pg yrl of fossil fuel carbon can not be
documented (Post et al. 1990) raises serious questions concerning our current
understanding of this cycle. Recent efforts to decrease uncertainty of global carbon
cycling processes have focused upon improved estimates of land-use change in the
tropics (Dale et al. 1991), sequestration in soil humus (Schlesinger 1977, 1990),
increased production associated with C02 fertilization (Keeling et al. 1989), and
sequestration by temperate forests recovering from past harvest (Tans et al. 1990).
While these efforts have been crucial to increase understanding, there may be another
fundamental problem with current carbon budgets: dead trees and associated woody
detritus are not considered (Kobak 1988, Houghton and Woodwell 1989, Post et al.
1990). This may seem like an unbelievable statement given that >80% of the globes
living biomass (boles, branches, and coarse roots) is woody (Woodwell et al. 1978),
yet a quick examination of estimated above-ground detrital fluxes and stores confirms
this major oversight (Harmon and Hua 1991). There are many explanations of this
development (Harmon et al. 1986). First, decay of woody detritus is a long-term
process and, as with other long-term ecological processes (i.e., forest succession),
there are few long-term observational studies (Franklin et al. 1990). Second,
inappropriate methodology has been used to study dead wood dynamics. For
example, litler production is often estimated from a sample area <10 m 2 for one to two
years, and although this is suitable for fine litter, it is inappropriate for woody detritus
which requires many hectare-years for reliable estimates (Sollins 1982). Finally, there
are few estimates of dead tree stores because they have little commercial value and
until recently were associated with “unhealthy” forests or bad management practices
(Franklin 1989). Thus 1 little empirical data exists upon which to base estimates or
ground assumptions. The time has come to critically evaluate these commonly held
assumptions and to begin to include woody detritus in carbon budgeting efforts.
A tacit assumption in many carbon budgets appears to be that tree mortality is a minor
flux. This may stem from a common misunderstanding that above-ground litterfall
measurements include measurements of tree mortality. While this may seem like a
valid assumption, only 40 of the 204 litter production studies cited by Vogt et al. (1986)
included fine woody production, and <10 included tree mortality. In the few mature
ecosystems that tree mortality has been examined, it is a major pathway of
above-ground detrital input, accounting for 30% of the input in old-growth deciduous
forests (Tritlon 1980), 55% in old-growth conifer forests (Sollins 1982), and 30 to 50%
in humid tropical forests (Brown and Lugo, in review). Recent global estimates of tree
1 USDA Forest Service, Forestry Sciences Laboratory, 3200 SW Jefferson Way, Corvallis, OR 97333.
151
-------
Section 9.C
Table 9.C.1. Regression statistics for the relationship between diameter at breast
height, stump height, and stump diameter.
Diameter
Species
Ba
r 2
N
Rangeb
(cm)
Abies amabalis
0.204
0.926
62
25-102
Acer macrophy/la
0.167
0.959
34
25-63
Alnusrubra
0.168
0.854
27
25-51
Pinus monticola
0.217
0.935
45
25-76
Piceasitchensis
0.323
0.968
81
25-140
Pseudotsuga mensiezii
0.183
0.982
89
25-140
Thujaplicata
0.290
0.949
85
25-140
Tsuga heterophylla
0.195
0.935
84
25-140
a The regression equation was in the form of D.BDBHSB where D is the difference between diameter at breast
height and stump diameter. DBH is the diameter at breast height. SB is the diflerence beiween stump height and
breast height.
b Diameter of stumps was determined for each 12.5 cm at 7 stump heights of 0.15, 0 30, 0 46, 0 61. 0.76. 0 91.
1.06, 1.21, and 137 m.
mortality indicate 3 to 15 Pg-C yr 1 of woody detritus is produced within intact stands,
while natural catastrophic disturbances add another 1 to 2 Pg yr_i globally (Harmon et
al. 1991: Harmon et al., in press). To correctly estimate global above-ground detrital
production, total tree mortality input of 4 to 17 Pg-C yr 1 must be added to the 28 Pg-C
yr-i input by fine litterfall (Meentemeyer et al. 1982). Therefore, a flux at least as large
as fossil fuel burning has been excluded from global carbon budgets.
If the decay dynamics of woody and leafy detritus were similar, then a 10 to 40%
underestimate of above-ground detritus production might lead to a similar sized
underestimate of above-ground detrital stores. In reality, decay rates of woody detritus
are an order of magnitude less than leafy litter (Harmon et al. 1986), and given a
global input of 4 to 17 Pg-C , the result would be a potential steady-state store of
85 to 295 Pg of woody detritus carbon (including dead boles, branches, and coarse
roots) (Harmon et al, in press). Therefore, a detrital pool at least as large as 70-84 Pg
of fine litter carbon (Kobak 1988, Post et al. 1990) has been ignored.
Another common assumption justifying the exclusion of woody detritus is that this
detrital pool is in steady-state. From the small amount of data on this component, this
appears to be a completely untenable assumption. Given an order of magnitude
152
-------
Section 9.C
difference in decay rate, it implies that woody litter will take an order of magnitude
longer to reach steady-state than fine liner. Yet many carbon budgets include
successional changes in fine litter stores but exclude woody litter.
In addition to slower decay rates, the period required to reach maximum input rates
following disturbance is considerably longer for woody than leafy litter. Fine litterfall
peaks at crown closure, which may occur in five years in deciduous forests (Bormann
and Likens 1979) and <50 years in coniferous forests (Gessell and Turner 1976). In
contrast, the highest rates of woody detrital inputs are associated with disturbances,
a hough in many successional studies this is overlooked. Disturbances are followed
by a long period in which woody litter inputs remain low. Input rates then increase as
the later stages of succession are reached, a stage which may take >100 years
(Harcombe et al. 1990). This implies that stores of woody detritus will not reach
steady-state levels for at least a century following disturbance. While many other
ecosystem components reach a minimum immediately following disturbance, woody
detrital stores are usually highest following natural disturbance, comprising the
majority of above-ground mass (85%) and reaching levels as high as 1300 Mg ha 1
(Agee and Huff 1987). Woody detritus production rates are usually lowest
immediately following disturbance, and increase gradually until the later stages of
succession are reached (Tritton 1980, Harcombe et al. 1990). The pulse of woody
detritus input associated with disturbance, followed by temporal lag of woody detritus
produced by the new stand, typically causes a U- or J-shaped pattern of woody detrital
stores during succession (Harmon et al. 1986, Spies et al. 1988, Gore and Patterson
1986). This often means that woody detrital stores are lowest during the middle stages
of succession, a-time when many other ecosystem parameters are approaching
steady-state.
In natural forests, the large mass of carbon and nutrients in woody detritus contributes
to ecosystem stability by building soils, providing habitat, and storing nutrients
(Harmon et al. 1986). In old-growth conifer forests, woody detritus comprises up to
22% of the above-ground mass and 81% of the above-ground detritus (Soflins et al.
1980). In old-growth deciduous forests, dead wood can also comprise a substantial
fraction of above-ground mass (14%) and detrilal mass (48%) (Tritton 1980). Even
some tropical ecosystems can store a substantial amount of organic matter in the form
of woody detritus (20-40 Mg ha- 1 , Kauffman et al. 1988, UhI and Kauffman 1990,
Whigham et al. 1991).
Woody detritus mass remains quite high throughout succession in natural forests, but
in managed forests it is dramatically reduced (Spies et al. 1988). In managed forests,
woody detritus is removed by timber harvest, salvage operations, and firewood
gathering (Cramer 1974). This may represent one of the most dramatic, long-term and
widespread departures from the natural forest resulting from management (Harmon et
al. 1990). This is because other ecosystem components recover more quickly than
woody detritus, which does not increase until large trees are present in stands (Spies
et al. 1988). Given this long recovery time, short-term rotations can reduce ecosystem
carbon storage of woody detritus even when the living portion of the ecosystem has
recovered (Harmon et al. 1990).
153
-------
Section 9.C
Excluding dead trees has several profound impacts upon understanding forest carbon
dynamics. First, the size of the terrestrial detrital pool has been underestimated
(Harmon and Hua 1991, Harmon et al., in press). If woody detritus was in steady-state
this might not be a major concern; however, land-use changes have placed woody
detritus in disequilibrium. In the Pacific Northwest, for example, reduction of woody
detritus accounted for 33% of the carbon flux to the atmosphere associated with the
conversion of old-growth forests to plantations (Harmon et at. 1990). Is this result
unique to this region, or does it apply to other forests in the boreal, temperate, and
tropical regions? If this is a general rule, then past calculations of carbon flux from
forest clearing have been underestimated in past studies (i.e., Melillo et al. 1988). It
also implies that a large, unaccounted carbon sink may be present in forests
recovering from past harvest (Harmon and Hua 1991). Second, excluding dead trees
will lead to an unrealistic assessment of the response of detritat stores to projected
climate change.
Analysis of transient responses to projected climate change indicate a large pulse of
carbon may be injected into the atmosphere as forests are disturbed or migrate
(Neilson and King 1991). To a large degree, the temporal dynamics of this transient
pulse will be controlled by the decomposition of woody plants killed by catastrophic
disturbances (i.e., fire) or increased stress. Yet our understanding of decomposition
and mortality rates is so poor that rates have to be guessed for major forest regions
(e.g., taiga forest). Third, the estimated effect of intensive forest management on
carbon sequestration (Sedjo and Solomon 1991) would probably be reconsidered if
woody detritus could be taken into account. Removal of dead wood from forest
ecosystems by current sustained, intensive silvicuttural practices may represent one of
the most dramatic, long-term and widespread departures from the natural forest which
result from management (Harmon et al. 1990).
In this report I provide preliminary estimates of the input rate, decay rate, and
steady-state stores for major forest types in the United States. For the purposes of this
analysis I have included all large and small dead wood that occurs above- and
below-ground. Thus my estimates include dead tree boles, stumps, tops, branches,
and coarse roots. Unfortunately, ihere is not a wea h of data to draw upon. However,
enough is known to place reasonable limits upon these rates and stores. At this point,
the error made in excluding woody detritus from budgets is probably greater than the
errors associated with poor parameter estimates.
METHODOLOGY
Mortality Turnover Rates
Mortality turnover rates of woody detritus were estimated from USDA Forest Service
forest survey data. Resource Bulletins for states containing a given forest type were
used to derive a minimum and maximum ratio of growing stock mortality to growing
stock volume. While an extensive computer database exists, I only had access to the
published summary reports. This meant a number of assumptions had to be made to
make estimates of woody inputs from the forest survey reports. First, I assumed that
most woody inputs are associated with tree death. As there are few reliable estimates
of branch or coarse root mortality available, there was no other alternative to this
154
-------
Section 9.C
assumption. Therefore 1 branch inputs and dead coarse root inputs are likely to be
underestimated. Second, mortality data were reported for all forests, regardless of
age. Therefore I assumed the estimated turnover rates would apply to all ages of
forest of a given type. In reality mortality turnover rates (as opposed to absolute input
rates) are likely to decrease with forest age. Third, the mortality and volume data are
reported for species and not by forest type. I therefore assumed that the majority of
each species grew in its respective forest type. This third assumption would probably
have little effeci on the overall estimates of turnover.
Finally, I assumed that the turnover rate of non-growing stock was similar to growing
stock. This is also likely to lead to an underestimate of the overall turnover rate since
smaller suppressed trees, or those with extensive defects, are more likely to die than
trees considered growing stock. However, this error may not be particularly large for
some forest types. For example, in the Southeastern Region merchantable volume
statistics are given for both growing stock and total live trees (Bechtold et al. 1987a,
1 987b, Thompson 1989). For southeastern conifer forests the mean ratio of total living
merchantable volume and growing stock merchantable volume was 1.01 (standard
error = 0.02, N=16), indicating non-growing stock volume is a minor component of
stands. For southeastern hardwood forests the proportion of non-growing stock
merchantable volume increases, with a mean total live to growing stock ratio of 1.14
(standard error = 0.02, N=18), indicating the possibility of a larger underestimate of
stand level mortality turnover rates. Given that three of the four assumptions made
concerning mortality will cause underestimates of the turnover rates, the mortality
turnover estimates should be considered conservative.
The fraction of the tree bole left on the site in tops decreases with increasing tree and
decreasing top diameter (Keays 1971). To estimate the volume left in conifer tops, I
assumed conifers were cones, therefore the fraction of bole volume left in tops (Tf) was
approximately:
Tf=(Dtop/DBH) 3 [ 4]
where DBH is the diameter at breast height, and Dtop is the top diameter. For the
purposes of this analysis I assumed the top diameter was 10 cm for all conifer forests.
In the case of hardwood species, I assumed that top mass would be included in the
estimates of branch mass.
The amount of decay in stems is a function of species, location (Boyce 1932), tree age
and size (Boyce 1932, Englerth 1942). As trees become older and larger the
proportion of the stem with rot increases. These complex relationships were not
modeled, instead I used standard volume deductions (Boyce 1932, British Columbia
Forest Service 1966, Englerth 1942, Hepting 1971) to estimate the bole mass left as
decayed or cull wood.
Breakage of boles during felling also can convert merchantable bole volume into
pieces too small to be yarded. Breakage differs widely between species (British
Columbia Forest Service 1966), and increases with tree size and roughness of
topography (Boyce 1932). I increased breakage with diameter, as observed for
Douglas-fir (Boyce 1932). to estimate the mass left in broken conifer boles. For
155
-------
Section 9.C
hardwoods I assumed the majority of breakage would occur in the branches and
therefore made no breakage deductions for these species.
Non-merchantable woody portions of trees left after harvest were derived from
published and unpublished biomass equations (Grier and Logan 1977, Sanantonio et
al. 1977, Tritton and Hornbeck 1982). All branches were added to the woody detrital
pool, whereas roots were only added if the dominant tree species in a forest were not
able to sprout. All coniferous forests, except those dominated by redwoods, were
assumed to have complete root mortality. In the case of redwood forests I assumed
that 20% of the root systems would die. In the case of oak-hickory, maple-beech, and
bottom land hardwoods, I assumed the root systems would all survive harvest. In the
case of oak-pine, I assumed that 50% of the roots would die following harvest. For
hardwood forests containing birch, I assumed 33% of the roots would die following
harvest.
To estimate the amount of woody detritus left on the site after harvest, I applied these
relationships to the published diameter distributions of major US forest types.
Diameter distributions were taken from stand tables for Douglas-fir (McArdle and
Meyer 1930), western hemlock-sitka spruce (Meyer 1938), ponderosa pine (Meyer
1937), lodgepole pine (Haig 1932; Moir et al., in review), northern hardwoods (Reed et
al. 1986), southern pines (Anonymous 1929), spruce-fir (Moir et al., in review), and
yellow poplar (McGee and Della-Bianca 1967). For each forest type, I estimated the
woody mass left on the site at rotation age under three scenarios: (1) all above-ground
mass removed, (2) all boles removed, and (3) only Sound, unbroken boles removed.
In forest types without readily available stand tables, I used the proportions from the
most similar forest type as a substitute.
Decay Rates
Estimates of woody decay rates were derived from the literature and unpublished data.
As the majority of literature reports values for woody debris (>10 cm diameter), and the
majority of woody material is in the coarse fraction, I used woody debris decay rates as
the starting point to determine overall decay rates. These values were then increased
based on data from studies with branch, and/or root decay rates (Abbot and Crossely
1982, Cline et al. 1981, Erickson et al. 1985, Fahey et al. 1988, Harris et al 1972, Miller
1983, Yavitt and Fahey 1982). To make this adjustment, I assumed that 15 and 20% of
the total dead wood input was in the form of branches and roots, respectively. I then
calculated a weighted average decay rate for the forests in which the decay rates of
root, branch and coarse wood rates were known. In the case where only root or
branch decay rates were known, I assumed branch and root decay rates were the
same.
As there were a number of forest types and regions for which no data on woody decay
rates existed, it was necessary to extrapolate from known forest types and regions.
The effect of climate upon woody decay rates is poorly understood. Increasing
temperature may lead to an increase in decay rate, although much of this effect
appears to be an interaction with moisture content. For coarse woody detritus, excess
moisture tends to limit decomposition more that excessive drying. Increasing
temperatures increase the rate of drying, which in turn increases the rate of decay.
156
-------
Section 9.C
Rather than try to model these rather complex interactions, I increased the decay rate
33% if the extrapolation was southward and decreased the decay rate 33% if the
extrapolation was northward. In the case of mixed species stands, I used the rates for
the fastest decaying and slowest decaying species to estimate the range of decay
rates.
Steady-State Stores
The median, minimum, and maximum steady-state solutions for dead wood pools in
old-growth forests were calculated for each forest type. These calculations follow the
equations presented in Olson (1963) in which the steady-state mass (Xss) is:
Xss= I/k
where I is the input rate in Mg ha 1 yrl at steady-slate, and k is the annual decay rate.
The minimum steady-state stores were calculated using the maximum decay rate and
the minimum mortality turnover rate. The maximum steady-state stores were
calculated using the minimum decay rate and the maximum mortality turnover rate.
The median steady-state stores were calculated using the median decay rate and the
median input turnover rate.
To judge if the calculations were reasonable, I compared the estimated steady-state
stores to published and unpublished values. In order to make these comparisons it
was necessary to make some adjustments of the published data. One difficulty is that
many coarse wood studies onty measured downed wood (i.e., logs) and did not
include snags. In sites where only log mass was reported, I set the snag mass equal
to 0.75 log mass, which is the mean snag/log ratio from Harmon et al. (1986). In the
case of redwood forests, where there are probably few snags, I computed a range of
total woody detrital mass by correcting for roots and branches and by correcting for
roots, branches and snags. Another problem is that no published studies of woody
debris includes dead roots or branches leading to an underestimate of total dead
wood mass. In order to estimate dead roots and branches from coarse wood
estimates, I multiplied by 1.2, assuming that although branches and coarse roots form
approximately 35% of the tree mass, the decay rates of these forms of detritus are
slightly higher than the coarse wood fraction.
As there are few biomass values to compare with estimates of steady-state stores, I
also used the ratio dead wood mass to live wood mass as another check. Harmon
and Hua (1991) found that ratios of 0.2 to 0.3 are common for temperate conifer
forests. Muller and Liu (1991) found that this ratio for northern hardwoods ranged
between 0.13 and 0.19, whereas for southern hardwood forests ii ranged between
0.05 and 0.17.
157
-------
Section 9.C
PARAMETER ESTIMATES
Input Rates
The published mortality turnover rates of growing stock differed greatly among forest
types, from a low of 0.0003 yr 1 for redwoods to a high of 0.027 yr 1 for North Central
spruce/fir (Table 9.C.2). As might be expected, forest types with extremely long-lived
species (e.g., redwoods arid Douglas-fir) had the lowest turnover rates, whereas
forests with short-lived species (e.g., aspen/birch, southeastern pines) had high
turnover rates. Forests with insect outbreaks (e.g., lodgepole pine, spruce/fir) or
disease outbreaks (such as bottom land hardwoods in the North Central region which
have Dutch elm disease) also had high turnover rates.
Comparison of estimated and observed mass input rate of woody detritus indicates
considerable agreement for most forest types with detailed ecological studies
(Figure 9.C.1). A region by region comparison indicates that
1. In the Pacific Northwest estimated inputs of 130 to 308 and 160 to 477 g-C
m 2 yr 1 for Douglas-fir and western hemlock/sitka spruce, respectively, are similar
to the 120 to 330 (Solltns 1982) and 210 to 380 g-C m 2 yr 1 values (Sollins 1982,
Harcombe et al. 1990) observed for these types.
2. In the Rocky Mountains, woody inputs were 20 g-C m 2 yr 1 in an Arizona
ponderosa pine forest (Avery et at. 1976), as compared to an estimated input of 16
to 38 g-C m 2 yr 1 . In contrast, estimated inputs for Rocky Mountain spruce/fir and
lodgepole pine of 100 to 230 and 60 to 170 g-C m 2 yr 1 , respectively, are
considerably higher than the 110 and 25 g-C m 2 yr 1 observed in these types
(Moir et at., in review). Therefore, it is likely that mortality turnover rates, mass
inputs, and steady-state stores have been overeslimated for these two types (see
below). A more reasonable estimate of mass inputs for lodgepole forests results if
the mortality turnover observed in the Pacific Northwest is substituted.
3. In the Southeast forest types, estimated inputs of 70 to 103 and 130 to 160 g-C
m- 2 yr- 1 for the oak/hickory and natural pine forest types, respectively, were close to
the observed values of 120 and 150 g-C m 2 yr l (Harris et at. 1973).
4. In the Northeast/North Central regions, the estimates for spruce/fir inputs of 78 to
151 g-C m 2 yr 1 bracket the observed input of 110 g-C m 2 yr 1 (Frank and Blum
1978). For white/red pine forests, the estimated range of 86 to 235 g-C rn 2 yr 1 is
close to the BOto 180 g-C rn 2 yr 1 observed (Jensen and Zasada 1977, Peet
1984). In contrast, the estimated range of inputs for maple/beech/birch forests of 50
to 60 g-C rn 2 yr 1 is on the low end of the observed range of 80 to 180 g-C m 2
yr 1 (Eyre and Longwood 1951, Gosz et al. 1973). This might lead to an
underestimate of steady-state stores for this forest type. Finally, the estimate of 30
to 80 g-C m 2 yr- 1 for oak/hickory spans the measured value of 60 g-C rn 2 yr 1
(McMilIan 1 9B1).
158
-------
Section 9.C
Table 9.C.2. Rate constants of dead wood input for US forest types based upon the
ratio of growing stock mortality to growing stock volume.
Region Mortality constants (yrl)
Ecosystem minimum maximum
Pacific Northwest
Douglas-fir 4 . 9 0.0030 0.0070
Spruce/fir 0.0030 0.0070
Hemlock/spruce 48 0.0030 0.0090
Lodgepole pine 1 . 5 0.0020 0.0070
Hardwoods 4 . 12 0.0040 0.0080
Ponderosa pine 15 0.0040 0.0060
Redwoods 12 0.0013 0.0010
adjusted 15 0.0020 0.0030
Rocky Mountains
Douglas-fir 36 0.0013 0.0033
Ponderosa pine 3 0.0014 0.0034
Spruce/fir 3 . 7 0.0086 0.0200
adjusted 15 0.0050 0.0075
Larch2,6 0.0034 0.0042
Lodgepole pine6.7 0.0047 0.01 31
adjusted 15 0.0020 0.0070
South/Southeast
Oakihickory 1014 0.0062 0.0087
Oakipin&O . 14 0.0086 0.0108
Bottom land 1014 0.0080 0.0082
Natural pine 10 . 14 0.0110 0.0129
Planted pine 10 • 14 0.0110 0.0129
Northeast
White/red pine 0.0031 0.0085
Spruce/fir 0.0070 0.01 30
Oak/hickory 0.0040 0.0110
Maple/beech/birch 0.0060 0.0072
North Central
White/red pine 11 . 13 0.0031 0.0085
Spruce/fir 11 . 13 0.0140 0.0270
adjusted 0.0070 0.0130
Maple/beech 11 . 13 0.0035 0.0037
adjusted l5 0.0060 0.0720
Aspen/birch 11 . 13 0.0146 0.0149
Bottom land 1113 0.0128 0.0159
0ak/hickory 1 . 13 0.0020 0.0055
1. Bassett and Oswald 1983.2 Bensen at al 1988. 3 Conner at al. 1990.4 Gedney at al 1986, 5 Gedney at al
1989. 6. Green at al 1985.7. Green and Conner 1989. 8 Harcombe at al 1990. 9 Harmon and Hua 1991, 10 Kelley
and Sims 1990, 11. Kingsley 1991, 12. Lloyd el al 1986, 13. Spencer at al 1989, 14 Vissage and Duncan. 1990
15. Value recommended to be used in order to match estimated and observed steady-stale stores.
159
-------
Section 9.C
Es rnated minimum ( Mg/ha year )
— 1:1 ratio
— 1:1 ratIo
A
60 80 100
minimum (Mg/ha year)
0 estimated minimum
B
0 estImated maximum
Figure 9.C.1. Observed versus estimated sleady•state mass of woody detrilus for US
mature to old-growth forests. The arrows indicate estimates where one of the
parameters was adjusted. A. minimum steady-state mass, B. maximum steady-state
mass.
0
o0
20 40
Observed
140
120
100
80
60
40
20
0•
0
300
250
200
150
100
50
0
0
120
140
50 tOO 150
Observed maximum (Mg/ha year)
200
160
-------
Section 9.C
Harvest Input Rates
The estimates of woody detritus left on the site after harvest (including boles, tops,
stumps, branches, and coarse roots) did not appear to be highly sensitive to forest
age. For example, I examined the proportion of woody detritus left on site as a function
of stand age for a Douglas-fir forest with a site index of 3 (McArdle and Myer 1930). At
age 60 years, the amount of tree mass left on site for whole tree, whole bole and only
merchantable harvests was 13, 34 and 62%, respectively. For a 160-year old stand
the eslimated residual material for the three harvest levels was exactly the same. This
indicates that the intensity of harvest is far more important in determining the mass of
woody material left on site than stand age.
In contiast to forest age, the amount of woody material left after harvest was strongly a
function of forest type (Table 9.C.3). For a whole tree harvest system, which removes
all parts except the roots, the lowest input was for oak-hickory (0% of total biomass)
and the highest was for eastern pine forests (22% of total biomass). Total removal of
the boles was estimated to leave between 20% (redwood forest) and 54%
(aspen-birch) of the total live mass on the site after logging. In a harvest system that
does not remove tops, broken or decayed boles, a substantial fraction of the formerly
live biomass is left on the site following harvest. These estimates range between 30
and 65% of the total biomass that is left on the site, with the later proportion being
typical for non-sprouting forests. The percentage of biomass left on the site after
harvest is up to twice the 33% estimated by Houghton et al. (1983) for temperate
forests. The rational for the estimates by Houghton et al. were not explained, although
they are likely to be best guesses. As my approach demonstrates, it is unlikely that
hardwood and conifer forests would have the same proportions, given the ability of the
former to root sprout.
The absolute mass added by timber harvest strongly depends upon forest age at the
time of harvest. In a Douglas-fir forest, for example, leaving breakage and decayed
boles would result in 12.4, 23.2, and 29.7 kg-C m 2 at age 55, 105 and 155 years,
respectively. Since the stores of previously dead wood in 55 to 155-year old
Douglas-fir forests range from 5 to 10 kg-C rn 2 , a substantial increase will occur
during harvest. Moreover, accounting for the decomposition of this detritus indicates a
harvested Douglas-fir forest is a net source of carbon to the atmosphere for at least 25
years. This is a period when decomposition of wood detritus exceeds the estimated
uptake of tree and understory carbon.
Decay Rates
Decay rates of coarse wood are highly variable among forest types and can vary by an
order of magnitude (Table 9.C.4). The lowest decay rate constant was estimated to be
0.005 yr 1 for redwood and the highest, 0.08 yr- 1 , was measured in bottom land
hardwood forests of Tennessee (Onega and Eickmeier 1991). While the former value
was estimated, it was set equal to a value observed for western redcedar (Harmon,
unpublished data), a species that is probably less resistant to decay than redwood. As
expected, hardwood forest types generally have higher decay rate constants than
conifer forests (Harmon et al. 1986).
161
-------
Section 9.C
Table 9.C.3. The proportion of live tree carbon left on site after timber harvest as
woody detritus for US forest types.
Region
Ecosystem
Proportion of Live Biomass Left
— Harvest System
Whole Tree Whole Bole Merchantable Bole
PacIfic Northwest
Douglas-fir 3 0.18 0.35 0.56
Spruce/fir 0.18 0.41 0.60
Hemlock/spruce 5 0.16 0.48 0.66
Lodgepole pine 0.18 0.41 0.60
Hardwoods 0.06 0.52 0.56
Ponderosa pine 6 0.20 0.45 0.65
Redwoods 0.05 0.20 0.45
Rocky Mountains
Douglas-fir 0.18 0.35 0.56
Ponderosa pine 0.20 0.45 0.65
Spruce/fir 7 0.18 0.41 0.60
Larch 2 0.18 0.35 0.56
Lodgepole pine 7 0.18 0.41 0.60
South/Southeast
Oak/hickory 4 ’ 8 0.00 0.28 0.36
Oak/pine 0.11 0.32 0.45
Bottom lartd 4 0.00 0.24 0.30
Natural pine 1 0.22 0.38 0.59
Planted pine 1 0.22 0.38 0.59
Northeast
White/red pine 0.22 0.38 0.59
Spruce/fir 0.18 0.41 0.60
Oak/hickory 0.00 0.28 0.36
Maple/beech/birch 8 0.06 0.52 0.56
North Central
White/red pine 0.22 0.38 0.59
Spruce/fir 0.18 0.41 0.60
Maple/beech 8 0.00 0.46 0.50
Apen/birch 0.08 0.54 0.58
Bottom land 0.00 0.24 0.30
Oak/hickory 0.00 0.28 0.36
1. Anonymous 1929. 2. Haig 1932:3. Mcardle and Meyer 1930; 4. McGee and DeIIa .Bianca 1967:5. Meyer 1937; 6.
Meyer 1938, 7. Mor at at, in review; 8. Reed at al. 1986
162
-------
Section 9.C
Table 9.C.4. Decomposition rate constants for US forest types as estimated from the
literature.
Region
Decay rate c
Coarse wood
onstants (yr 1 )
Adjusted
Overall
Ecosystem
minimum
maximum
minimum
maximum
Pacific Northwest
Douglas-fir 5 0.011 0.016 0.015 0.022
Spruce/fir 0.015 0.020 0.021 0.028
Hemlock/spruce 3 ’ 4 0.009 0.010 0.013 0.014
adjusted 14 0.020 0.022 0.028 0.031
Lodgepole pine 1 0.015 0.020 0.031 0.041
Hardwoods 0.050 0.080 0.052 0.082
Ponderosa pine 6 0.010 0.012 0.014 0.017
Redwoods 0.005 0.007 0.010 0.014
Rocky Mountains
Douglas-fir 0.011 0.016 0.015 0.022
Ponderosa pine 0.010 0.012 0.014 0.017
Spruce/fir 11 0.009 0.011 0.013 0.014
Larch 0.011 0.016 0.015 0.022
Lodgepole pine 1 ’ 11 0.010 0.011 0.021 0.023
South/southest
Oak/hickory 0.040 0.050 0.060 0.075
Oak/pine 13 0.030 0.040 0.045 0.060
Bottom land 12 0.060 0.080 0.084 0.112
Natural pine 13 0.020 0.040 0.028 0.056
Planted pine 0.030 0.040 0.042 0.056
Northeast
White/red pine 0.020 0.030 0.028 0.042
Spruce/fir 2 ’ 8 0.020 0.030 0.028 0.042
Oak/hickory 0.040 0.050 0.060 0.075
Maple/beech/birch 0.050 0.080 0.052 0.062
North Central
White/red pine 0.020 0.030 0.028 0.042
Spruce/fir 0.020 0.030 0.028 0.042
Maple/beech 0.060 0.080 0.062 0.082
Aspen/birch 1 ° 0.050 0.080 0.052 0.082
Bottom land 0.060 0.080 0.084 0.112
Oak/hickory 9 0.030 0.040 0.045 0.060
1. Fahey 1983, 2. Foster and Lang 1982, 3. Graham and Cromack 1982. 4. Gnat 1978, 5. Harmon and Hua 1991. 6
Harmon unpublished data, 7. Harmon 1982, 6. Lambert at al 1980. 9 McMilhan 1988, 10 Miller 1983. 11. Motr at al in
review, 12. Onega and Eickmeier 1991, 13. Smith and Boring 1990, 14. Recommended changes to make estimated
and observed steady-state mass match
163
-------
Section 9.C
Calculations of the factor needed to correct coarse woody decay rates to include all
forms of woody detritus indicated a dependence upon forest type. The rate for all
forms of woody detritus was 1.40, 2.05, 1.52, 1.41, and 1.03 times the coarse wood
rate for Douglas-fir (Cline et a!. 1981: Erickson et al. 1985: Harmon, unpublished):
lodgepole pine, (Yavitt and Fahey 1982) oak/hickory (Abbot and Crossley 1982),
botlom land hardwoods (Harris et a!. 1972), and maple/beech/birch (Fahey et al. 1988,
Miller 1983), respectively. Given the lack of studies in which coarse wood, roots and
branch decay have been compared, these correction factors are extremely tentative.
Nonetheless, the overall trend appears reasonable, in that the decay rate of coarse
wood is largely determined by the amount of decay resistant heartwood present ( and
which is largely absent in coarse roots and branchwood). Therefore, one would
expect that forest types with trees possessing decay resistant heartwood (Douglas-fir,
pine, oak) would have a larger correction factor than those with less resistant trees
(i.e., maple, birch).
Using the estimated overall wood decay rate constant indicates that, given a steady
input of woody detritus, a steady-state would be reached in 27 to 300 years depenthng
upon the forest type. Using rate constants of 0.02 and 0.08 yr 1 (which are more
typical of the majority of forest type ) would give a steady-state mass within 37 to 150
years. Coupled with the lag in reaching old-growth levels of woody detritus input, this
reinforces the notion that woody detrital stores will not be in steady-state unless
harvest rotations greatly exceed current practices.
Steady-State Stores
An examination of the original steady-state solutions indicated a number of forests
differed widely from reported steady-state stores (Figure 9.C .2, Table 9.C.5).
Moreover, in many cases the ratio of dead wood to living biomass was beyond the
ranges expected. There were three possible sources of error in these estimates: (1)
the decay rate constant, (2) mortality turnover rates, and (3) steady-state live biomass.
The third factor seems to explain, in part, why the original estimate of steady-state in
redwood forest was low. In most cases, however, the lack of fit was probably caused
by a poor estimate of the mortality turnover rate. In the following text each forest type
with field data is compared to the estimates and required adjustments are discussed.
Pacific Northwest Forests . The steady-state stores for Douglas-fir forest are within the
observed range of 9,000 to 33,000 g-C rn 2 for this type (Spies et al. 1988, Agee and
Huff 1987). The latler value is somewhat unusual for the Douglas-fir region as a whole
because it occurred on a very cool, moist site. This would greatly reduce the decay
rate and result in a high mass of dead wood. A more typical upper mass would be
12,000 g-C rn 2 (Harmon et al. 1986). In western hemlocklsitka spruce forest, Grier
(1978) reported 13,200 g-C rn- 2 of woody detritus which is much lower than the
estimated minimum. My own unpublished data from mature hemlock-spruce stands
indicating an average of 8,520 g-C m 2 for this type, indicated that a maximum
steady-state estimate is unrealistically high. The input rates for this forest type are
quite reasonable, suggesting that the decay rate constants have been underestimated
in past field studies. I therefore recommend that decay rate constants of 0.028 to
0.031 yr 1 be used in modeling dead wood dynamics in this forest type. For redwood
164
-------
Section 9.0
Increment (MgCfha/year)
— Tree C
A
— Dead Wood
B
—— TotaiC
0 20 40 60 80 100 120 140
Time (years)
160
— Tree C
—— Total C
Figure 9.C.2. Carbon increment associated with tree growth and woody detritus for
Pacific northwest Douglas-fir forests. A. minimum case and B. maximum case.
0 5 15 25 35 45 55 65 75. 85 95 105 115 125 135 145 155
Time (years)
Increment (Mg C/ha year)
8
6
4
2
0
-2
-4
-6
-8
— Dead Wood C
165
-------
Section 9.C
Table 9.C.5. Estimated steady-state stores of woody detritus for US forests.
Region Steady-state stores Dead/Live
Ecosystem (g-C rn- 2 ) Ratio
minimum maximum median
Pacific Northwest
Douglas-fir 5902 20030 11658 0.26
Spruce/fir 2143 6667 4082 0.20
Hemlockispruce 3785 17031 9083 0.17
Lodgepole pine 439 2049 1129 0.13
Hardwoods 952 3047 1758 0.09
Ponderosa pine 3802 6844 5185 0.32
Redwoods 12544 26341 18293 0.20
Rocky Mountains
Douglas-fir 897 3312 1881 0.12
Ponderosa pine 938 2732 1753 0.16
Spruce/fir 3673 6733 5050 0.45
Larch 3757 6750 4976 0.20
Lodgepole pine 1142 4397 2692 0.21
South ISo Ut he a St
Oak/hickory 975 1710 1302 0.11
Oak/pine 1318 2207 1699 0.18
Bottom land 828 1131 958 0.08
Natural pine 1527 3925 2326 0.19
Planted pine 2443 3819 3033 0.24
Northeast
White/red pine 903 3716 2028 0.17
Spruce/fir 1862 5187 3192 0.29
Oak/hickory 947 1775 1315 0.07
Maple/beech/birch 1456 2096 1747 0.12
North Central
White/red pine 2039 8387 4578 0.17
Spruce/fir 2076 6007 3648 0.59
Maple/beech 590 903 753 0.05
Aspen/birch 2163 3533 2690 0.22
Bottom land 2801 4639 3569 0.15
Oak/hickory 987 3618 2114 0.14
166
-------
Section 9.C
forests, the minimum and the maximum steady-state estimates appear to be far below
field measurements for this type. Bingham and Sawyer (1988) found 10,000 g-C m 2
of logs; inclusion of snags, roots and branches would indicate that a steady-state value
of 24,000 g-C m 2 could be expected. A major problem with estimating steady-state
stores for this type appears to be an extremely low mortality rate, giving a species
longevity of over 10,000 years. Setting the mortality turnover rate to be 0.002 to 0.003
yr 1 gives reasonable mass inputs and brings the steady-state value closer to the
expected number. Another problem is that the maximum biomass reported by Birdsey
(1990) for this type is probably less than half the reported maximum. When this
difference is also taken into account the steady-state stores are similar to the estimated
stores.
Rocky Mountain Forests . Brown and See (1981) and Sacket (1979) measured 1,150
and 1,800 g-C rn 2 of logs in ponderosa pine forest of Montana and Arizona,
respectively. Including snags, coarse roots, and branches, a value of 1,782 to 2,415
g-C rn 2 would seem reasonable for the total dead wood stores in ponderosa pine
forests. This would indicate that the estimated steady-state stores are somewhat low
for this type, although the estimated decay rate constants, mortality turnover and
mature forest biomass all seem reasonable for this type. The maximum estimate of
17,956 g-C r r 2 in spruce-fir is considerably higher than the 3,600 g-C m 2 of logs and
snags measured by Moir et al. (in review) in a 300-year old spruce-fir forest in
Colorado, even if an additional 720 g-C rn 2 is added to correct for total wood mass. In
Montana, Brown and See (1981) found 2,650 g-C m 2 in logs for spruce-fir forests,
while correcting for snags, roots and branches would indicate a total wood store as
high as 5,565 g-C m 2 would be reasonable for this type. The mortality turnover rate
appears to be too high for this forest type (see above) by a factor of two. Reducing the
mortality turnover rate for Rocky Mountain spruce-fir forests to 0.005 to 0.0075 yr 1
gives more realistic steady-state stores of 3,673 to 6,733 g-C m 2 . The estimated
median steady-slate mass of 4,976 g-C rn- 2 for larch forests is higher than the
adjusted total of 4,200 g-C rn 2 (Brown and See 1981). However, the overall
dynamics of the forest appear reasonable. In lodgepole pine forests, both the
minimum and maximum steady-state estimates are too high. Fahey (1983) measured
2,100 in logs and snags; Moir et al. (in review) measured 2,000 g-C m 2 in logs and
snags; and Brown arid See (1981) found 1,600 g-C rn 2 in logs within undisturbed old
lodgepole pine forests. Correcting for the missing dead wood components indicates a
range of 2,400 to 3,360 g-C rn 2 would be reasonable for this forest type. The most
likely problem for lodgepole pine forests is an overestimate of the mortality turnover
rate, which gives mass inputs larger than observed. Reducing the input rates to values
observed in lodgepole pine forests of the Pacific Northwest gives a more reasonable
median steady-slate estimate of 2,692 g-C rn 2 .
Southeast/South . Oak/hickory forests have been reported to have 1,080 g-C rn 2 in
logs and snags (Muller and Liu 1991), and between 1,050 g-C m 2 (Harmon et al.
1986) and 1,200 g-C rn 2 (Lang and Forman 1978) in logs. Adjusting to total dead
wood stores would indicate a range of 1,300 to 2,500 g-C m 2 , which is slightly higher
than the estimated range for this forest type c i 975 to 1,710 g-C rn 2 . The dynamics of
167
-------
Section 9.C
this forest type may be correct, however, as field studies included the dead woody
mass killed by the chestnut blight. In bottom land hardwoods, Onega and Eickmeier
(1991) measured 1,000 g-C rn 2 in logs and snags. Correcting for dead roots and
branches indicates the measured mass of 1,200 g-C m 2 , which falls slightly above the
estimated minimum and maximum of 826 to 1,131 g-C m 2 . In southeastern natural
pine forests, a range of 551 to 890 g-C rn 2 of logs has been measured (Harris et al.
1973, Harmon et al 1986). Correcting to total dead wood stores indicates a range of
1,323 to 2,136 g-C rn- 2 , which is lower than the estimated range of 1,527 to 3,925
g-Cm 2 .
Northeast . In maple/beech/birch forests the estimated maximum and minimum is
much lower than the observed range. Tritton and Hcrnbeck (1982) measured 2,500
g-C m 2 in logs and snags, while Gore and Patterson (1986) found 1,700102.000
g-C m 2 in logs. Adjusling these to the dead wood total indicates a range of 3,000 to
4,200 g-C m 2 , which is almost twice the estimated median steady-state store. The
mostly likely problem is a poor underestimate of the input rate since New England
Forest Surveys do not report mortality rates. However, increasing these rates gives
mass input values that exceed observed values. It is also possible that decay rate
constants were overestimated, although again there is little data to indicate it this is the
case.
North Central . In oak hickory forests McMillan (1981) measured 800 g-C m 2 in logs,
while adjusting for other types of dead wood indicates a value of 1,680 g-C rn- 2 (which
is bracketed by the estimated range of 493 to 1609 g-C m 2 ). The initial steady-state
estimates for spruce-fir forest in this region are roughly twice that observed in other
regions with this forest type. As the decay rate is probably relatively constant for this
type, it is most likely that the mortality turnover for spruce-fir in the North Central region
is too high. I therefore recommend using a range of 0.007 to 0.013 yr 1 .
MODELING WOODY DETR1TUS
The parameters presented above can be used to model the woody detrital dynamics of
a wide range of forest types. In this section I will illustrate these dynamics by
examining the behavior of two forest types, Douglas-fir in the Pacific Northwest and
bottom land hardwoods of the Southeast. These two forest types were selected
because they represent the extremes of the parameter estimates (Tables 9.C.2 thru
9.C.5). For each of these two types, I simulated a minimum and maximum case of
dead wood storage. For the minimum, I used the minimum mortality turnover rate,
steady-state stores, and harvest input, as well as the maximum decay rate. For the
maximum case, I used the maximum mortality turnover rate, steady-state stores, and
harvest input, as well as the minimum decay rate. Thus, the four simulations represent
the extremes of behavior both among forest types and within forest types.
In all cases, except the bottom land hardwood minimum, the addition of dead wood
dynamics greatly altered the carbon flux dynamics of harvested forests (Figures 9.C.3
and 9.C.4). For Douglas-fir forests, the addition of woody detritus makes the harvested
forest a net source of carbon for approximately 25 years (Figure 9.C.3). Moreover,
168
-------
Section 9.C
these cumulative losses are not balanced by the regrowing trees until the forest is 35
to 60-years old. The addition of woody detritus also influenced the behavior of bottom
land hardwood forests, but to a lesser extent than Douglas-fir forests (Figure 9.C.4).
Nonetheless, in the case of the bottom land hardwood maximum, addition of dead
wood dynamics made the harvested system a net source of carbon for the first 15
years, and these cumulative losses were not oflset by new growth until the forest was
27-years old.
In addition to increasing the net source immediately following harvest, the addition of
woody detritus increased the net sink in the later stages of succession. These
increases ranged from approximately 10% in the bottom land hardwood minimum
case to 20% in the Douglas-fir maximum case. While these increases may not seem
important, these low rates of increased uptake extend over many decades and may
greatly increase the overall sink of older forests.
169
-------
Section 9.C
Estimated minimum (M9/ha year)
2.5
2
1.5
I
0.5
0
5
A
0 0.5 1 1.5 2 2.5
Observed minimum (Mg/ha year)
— 1:1 ratio estimated minimum * adjusted
B
Estimated maximum (MgTha year)
4-
3-
2
I
0•
0 1 2 3
Observed maximum (Mg/ha year)
C
4
0 stimated maximum
• adjusted value
Figure 9.C.3. Observed versus estimated mass inputs of woody detritus for US mature
to ofd-growth forests. The arrows indicate mortality turnover rates that were adjusted
A. minimum mass inputs, B. maximum mass inputs.
0
0
C
C
— 1:1 mVo
170
-------
Section 9.C
5
4
3
2
1
0
Increment (Mg C/ha year)
A
o 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Time (years)
— Tree C — Dead Wood C Total C
6
4
2
0
-2
-4
B
0 10 20 30 40 60 60 70 80
Time (years) S
— Tree C Dead Wood C —— Total C
Figure 9.C.4. Carbon increment associated with tree growth and woody detritus for
southeastern bottom land hardwood forests. A. minimum case and B. maximum case.
Increment (Mg C/ha year)
171
-------
Section 9.C
LITERATURE CITED
Abbott, D. T. and D. A. Crossely. 1982. Woody litter decomposition following
clear-cutting. Ecology 63:35-42.
Agee J. K. and M. H. Huff. 1987. Fuel succession in a western hemlock/Douglas-fir
forest. Canadian Journal of Forest Research 17:697-704.
Anonymous. 1929. Volume, yield, and stand tables for second-growth southern pines.
USDA Miscellaneous Publication No. 50.
Avery. C. C., F. A. Larson, and G. H. Schubert. 1976. Fifty-year records of virgin stand
development in southwestern ponderosa pine. USDA Forest Service General
Technical Report RM-22.
Bassett, P. M. and D. D. Oswald. 1983. Timber resource statistics for eastern
Washington. USDA Forest Service Resource Bulletin PNW-104.
Bechtold, W. A., M. J. Brown, and J. B. Tansey. 1987a. Virginia’s forests. USDA
Forest Service Resource Bulletin SE-95.
Bechtold, W. A., M. J. Brown, and R. M. Sheffield. 1987b. Florida’s forests, 1987.
USDA Forest Service Resource Bulletin SE-lb.
Bensen, R E., A. W. Green, and D. D. Van Kooser. 1988. Idaho’s timber resources.
USDA Forest Service Resource Bulletin INT-39.
Bingham B. B. and J. U. Sawyer. 1988. Volume and mass of decaying logs in upland
old-growth redwood forest. Canadian Journal of Forest Research ¶8:1649-1651.
Bormann, F. H. and G. E. Likens. 1979. Pattern and Process in a Forested Ecosystem.
Springer-Verlag, Inc., New York.
Boyce, J. S. 1932. Decay and other losses in Douglas-fir in western Oregon and
Washington. USDA Technical Bulletin No. 286. 60 pp.
British Columbia Forest Service. 1966. Net volume loss factors. Forest Survey Notes
No. 8, British Columbia Forest Service, Victoria, British Columbia. 80 pp.
Brown, J. K. and T. E. See. 1981. Downed dead woody fuel and biomass in the
Northern Rocky Mountains. USDA Forest Service General Technical Report INT-1 17.
Brown, S. and A. E. Lugo. In review. Production of coarse woody debris and its role in
organic matter budgets in tropical forests of Venezu&a. Ecology.
Cline, S. P., P. Soflins, T. Verhoeven, and K. Cromack Jr. 1981. Comparative
decomposition of coarse woody debris in western Oregon: Snags, fallen boles,
stumps, and roots. Bulletin of the Ecological Society of America 62:113.
172
-------
Section 9.C
Conner. P. C., A. W. Green, J. D. Born, and R. A. O’Brien. 1990. Forest resources of
Arizona. USDA Forest Service Resource Bulletin INT-69.
Cramer, 0. P. (ed]. 1974. Environmental effects on forest residues management in the
Pacific Northwest. USDA Forest Service General Technical Report PNW-24. 416 pp.
Dale, V. H., A. H. Houghton, and C. A. S. Hall. 1991. Estimating the effects of land-use
change on global atmospheric C02 concentrations. Canadian Journal of Forest
Research 21:87-90.
Englerth, G. H. 1942. Decay of western hemlock in western Oregon and Washington.
Yale School of Forestry Bulletin No. 50.
Erickson, H. E., R. L. Edmonds, and C. E. Peterson. 1985. Decomposition of logging
residues in Douglas-fir, western hemlock, Pacific silver fir, and ponderosa pine
ecosystems. Canadian JoLirnal of Forest Research 15:914-921.
Eyer , F. H. and F. A. Longwood. 1951. Reducing mortality in old-growth northern
hardwoods through partial cutting. USDA Forest Service Station Paper 25, Lake
States Experimental Forest Station.
Fahey, T. J. 1983. Nutrient dynamics of above-ground detritus in lodgepole pine
(Pinus contorta sp. latifolia) ecosystems, Southeastern Wyoming. Ecological
Monographs 53: 5 1-42.
Fahey, T. J., J. W. Hughes, M. Pu, and M. A. Arthur. 1988. Root decomposition and
nutrient flux following whole-tree harvest in northern hardwood forests. Forest
Science 34:744-768.
Foster, J. R. and G. E. Lang. 1982. Decomposition of red spruce and balsam fir boles
in the White Mountains of New Hampshire. Canadian Journal of Forest Research 12:
617-626.
Frank, R. M. and B. M. Blum. 1978. The selection system of silviculture in spruce-fir
stands--procedures, early results, and comparison with unmanaged stands. USDA
Forest Service Research Paper NE-425.
Franklin, J. F. 1989. Toward a new forestry. American Forests 85:37-44.
Franklin. J. F., C. S. Bledsoe, J. T. Callahan. 1990. The Long-Term Ecological
Research Program: a developing network. Bioscience 40:509-523.
Gedney, D. R., P. M. Bassett, and Mary A. Mie. 1986. Timber resource statistics for the
non-Federal Forest land on Northwest Oregon. USDA Forest Service Resource
Bulletin PNW-140.
Gedney, D. R., P. M. Bassett, and M. A. Mel. 1989. Timber resource statistics for all
forest land, except National Forests, in eastern Oregon. USDA Forest Service
Resource Bulletin PNW-164.
173
-------
Section 9.C
Gessell, S. P. and J. Turner. 1976. Litter production in western Washington
Douglas-fir stands. Forestry 49 :63-72.
Gore, J. A. and W. A. Patterson II I. 1986. Mass of downed wood in northern hardwood
forests in New Hampshire: potential effects in forest management. Canadian Journal
of Forest Research 16:335-339.
Gosz, J. R., G. E. Likens, F. H. Bormann. 1973. Nutrient release from decomposing
leaf and branch litter in the Hubbard Brook Forest, New Hampshire. Ecological
Monographs 43:173-191.
Graham, R. L. and K. Cromack Jr. 1982. Mass, nutrient content, and decay rate of
dead boles in rain forests of Olympic National Park. Canadian Journal of Forest
Research 12:511-521.
Green, A. W. and A. C. Conner. 1989. Forests in Wyoming. USDA Forest Service
Resource Bulletin INT-61.
Green, A. W., P. A. O’Brien, and J. C. Schaefer. 1985. Montana’s forests. USDA
Forest Service Resource BuHetin INT-38.
Grier, C. C. 1978. A Tsuga heterophylla-Picea sitchensis ecosystem of coastal
Oregon: Decomposition and nutrient balance of fallen logs. Canadian Journal of
Forest Research 8:198-206.
Grier, C. C. and R. S. Logan. 1977. Old-growth Pseudotsuga menziesii communities
of a western Oregon watershed: Biomass distribution and production budgets.
Ecological Monographs 47: 373-400.
Haig, I. T. 1932. Second-growth yield, stand, and volume tables for the western white
pine type. USDA Forest Service Technical Bulletin No. 323.
Harcombe, P. A. , M. E. Harmon, and S. E. Greene. 1990. Changes in biomass and
production over 53 years in a coastal Picea sitchensis.Tsuga heterophylla forest
approaching maturity. Canadian Journal of Forest Research 20: 1602-1610.
Harmon, M. E. 1982. Decomposition of standing dead trees in the Southern
Appalachian Mountains. Oecologia 52: 10-14.
Harmon, M. E. and Chen Hua. 1991. Coarse woody debris dynamics in two
old-growth ecosystems. Bioscience 41:604-610.
Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N. H.
Anderson, S. P. Cline, N. G. Aumen, J. R. Sedell, 3. W. Lienkaemper, K., Cromack Jr.,
and K. W. Cummins. 1986. Ecology of coarse woody debris in temperate ecosystems.
Advances in Ecological Research 15:133-302.
174
-------
Section 9.C
Harmon, M. E., W. K. Ferrell, and J. F. Franklin. 1990. Effects of carbon storage of
conversion of old-growth forests to young forests. Science 247: 699-702.
Harmon, M. E., S. Brown, and S. T. Gower. 1991. From gaps to the globe: Ecosystem
consequences of tree death. Bulletin of the Ecological Society of America 72:135.
Harmon, M. E., S. Brown, and S. T. Gower. In press. Consequences of tree mortality
to the global carbon cycle. In: Carbon Cycling in Boreal and Sub-arctic Ecosystems.
Corvallis, OR.
Harris, W. F., G. S. Henderson, and D. E. Todd. 1972. Measurement of turnover of
biomass and nutrient elements from the woody component of forest litter on Walker
Branch Watershed. Eastern Deciduous Forest Biome Memo Report #72-146, Oak
Ridge. TN.
Harris, W. F., R. A. Goldstein, and G. S. Henderson. 1973. Analysis of forest biomass
pools, annual primary production, and turnover of biomass for a mixed deciduous
forest watershed. In: JUFRO Biomass Studies (H. E. Young editor), University of Maine
Press, Orono.
Hepting, C. H. 1971. Diseases of forest trees of the United States. USDA Agricuftural
Handbook No. 386.
Houghton, R. A. and G. M. Woodwell. 1989. Global climate change. Scientific
American 260: 36-44.
Houghton, R. A., J. E. Hobbie, J. M. Melillo, B. Moore, B. J. Peterson, C. A. Shaver, and
G M. Woodwell. 1983. Changes in the carbon content of terrestrial biota and soils
between 1860 and 1980: a net release of C02 to the atmosphere. Ecological
Monographs 53:235-262.
Jensen, R. A. and Z. A. Zasada. 1977. Growth and mortality in an old-age jack pine
stand. Minnesota Forestry Note 261.
Kauffman, J. B., C. UhI, and D. L. Cummings. 1988. Fire in the Venezuelan Amazon 1:
Fuel biomass and fire chemistry in the evergreen rain forest of Venezuela. Qikos
53:167-175.
Keays, J. L. 1971. Complete-tree utilization: an analysis of the literature. Part 1.
Unmerchantable top of bole. Forest Products Laboratory, Canadian Forest Service
Information Report VP-X-69, Vancouver, British Columbia.
Keeling, C. D., R. B. Bacastow, A. F. Carter, S. C. Piper, T. P. Whort, M. Heimann, W. G.
Mook, and H. J. Roelotfzen. 1989. A three dimensional model of atmospheric C02
transport based on observed winds: I. Analysis of observalional data. Geophysical
Monographs 55: 165-236.
Kelley, J. F. and M. Sims. 1990. Forest resources of Mississippi. USDA Forest
Service Resource Bulletin SO-147.
175
-------
Section 9.C
Kingsley, N. P. 1991. Forest statistics for Minnesotas aspen-birch unit. USDA Forest
Service Resource Bulletin NC-i 28.
Kobak. K. 1. 1988. Biotical Compounds of Carbon Cycle. Leningrad,
Gydrometecizdat (in Russian).
Lambert, R. C., G. E. Lang, and W. A. Reiners. 1980. Loss of mass and chemical
change in decaying boles of a sub-alpine balsam fir forest. Ecology 61:1460-1473.
Lang, G. E. and R. T. T. Forman. 1978. Detritus dynamics in a mature oak forest:
Hutchenson Memorial Forest, New Jersey. Ecology 59: 580-595.
Lloyd, J. D. Jr., J. Moen, and C. L. Bolsinger. 1986. Timber resource statistics for the
North Coast resource area of California. USDA Forest Service Resource Bulletin
PNW-131.
McArdle, R. E. and W. H. Meyer. 1930. The yield of Douglas-fir in the Pacific
Northwest. USDA Technical Bulletin 201.
McGee, C. E. and L. Della-Bianca. 1967. Diameter distribution in natural
yellow-poplar stands. USDA Forest Service Research Paper 25.
McMillan, P. C. 1981. Log decomposition in Donaldson’s Woods, Spring Mill State
Park, Indiana. American Midland Naturalist 106:335-344.
McMillan, P. 1986. Decomposition of coarse woody debris in an old-growth Indiana
forest. Canadian Journal of Forest Research 18:1353-1362.
Meentemeyer, V., E. 0. Box and R. Thompson. 1982. World patterns and amounts of
terrestrial plant litter production. Bioscience 32: 125-128.
Melillo, .1. M., J. R. Fruci, and H. A. Houghton. 1988. Land-use change in the Soviet
Union between 1850 and 1980: causes of net release of C02 to the atmosphere.
Tel/us 40B:1 16-128.
Meyer, W. H. 1937. Yield of even-aged stands of sitka spruce and western hemlock.
USDA Forest Service Technical Bulletin 554.
Meyer, W. H. 1938. Yield of even-aged stands of ponderosa pine. USDA Forest
Service Technical Bulletin 630.
Miller, W. E. 1983. Decomposition rates of aspen bole and branch litter. Forest
Science 29:35 1 -356.
Moir, W., M. E. Harmon, F. G. Halksworth, R. Brainerd, and J. F. Franklin. In review.
Fifty years of succession in old-growth lodgepole pine and spruce-fir forests in
Colorado. Ecology.
176
-------
Section 9.C
Muller R. N. and Y. Liu. 1991. Coarse woody debris in an old-growth deciduous forest
on the Cumberland Plateau, southeastern Kentucky. Canadian Journal of Forest
Research 21:1567-1572.
Neilson, R. P. and G. A. King. 1991. Continental scale biome responses to climatic
change. In: Ecological Indicators. Proceedings of International Symposium, Ft.
Lauderdale, FL, October 16-19, 1990.
Olson, J. S. 1963. Energy storage and the balance between producers and
decomposers in ecosystems. Ecology 44: 322-331.
Onega, T. L. and W. G. Eickmeier. 1991. Woody detritus inputs and decomposition
kinetics in a Southern temperate deciduous forest. Bulletin of the Torrey Botanical
Club 118:52-57.
Peet, R. K. 1984. Twenty-six years of change in a Pinus Strobus - Acer Saccha rum
forest, Lake Itasca, Minnesota, Bulletin of the Torresy Botanical Club 111:61 -68.
Post, W. M., T. Peng, W. R. Emanuel, A. W. King, V. H. Date, and D. L. DeAngelis.
1990. The global carbon cycle. American Scientist 78: 31 0-326.
Reed, D. D., M. J. Holmes, and J. A. Johnson. 1986. A 22-year study of stand
development and financial return in northern hardwoods. Northern Journal of Applied
Forestry 3:35-40.
Sackett, S. S. 1979. Natural fuel loadings in ponderosa pine and mixed conifer
forests of the Southwest. USDA Forest Service Research Paper RM-21 3.
Sanantonio, D., R. K. Hermanr and W. S. Overton. 1977. Root biomass studies in
forest ecosystems. Pedobiologica 17:1-31.
Schlesinger, W. H. 1977. Carbon balance in terrestrial detritus. Annual Review of
Ecology and Systematics 8:51-81.
Schlesinger, W. H. 1990. Evidence from chronosequence studies for a low carbon
storage potential of soils. Nature: 232-234.
Schneider, S. H. 1989. The changing climate. Scientific American 261 (3):70-79.
Sedjo, R. A. and A. M. Solomon. 1991. Climate and forests. In: Greenhouse Warming:
Abatement and Adaptation. Proceedings for the Workshop, Washington, DC. pp.
105-119.
Smith, R. N. and L. R. Boring. 1990. Pinus rigida coarse woody debris inputs and
decomposition in pine beetle gaps of the southern Appalachians. Ecological Society
of America 71:329.
Sollins, P. 1982. Input and decay of coarse woody debris in coniferous stands in
western Oregon and Washington. Canadian Journal of Forest Research 12: 18-28.
177
-------
ec ion
Sollins, P., C. C. Grier, F. M. McCorison, K. Cromack, Jr., R. FogeL and R. L.
Fredricksen. 1980. The internal element cycles of an old-growth Douglas-fir
ecosystem in western Oregon. Ecological Monographs 50: 261-285.
SoHins. P., S. P. Cline, T. Verhoeven, D. Sachs, and G. Spycher. 1982. Input and
decay of coarse woody debris in coniferous stands in Western Oregon and
Washington. Canadian Journal of Forest Research 12:18-28.
Spencer, J. S., Jr., W. B. Smith, J. T. Hahn, and G. K. Raile. 1989. Wisconsin’s fourth
forest inventory, 1983. USDA Forest Service Resource Bulletin NC-I 07.
Spies, T. A., J. F. Franklin, and T. B. Thomas 1988. Coarse woody debris in
Douglas-fir forests 01 western Oregon and Washington. Ecology 69:1 689-1 702.
Tans, P. P., I. V. Fung, and T. Takahashi. 1990. Observational constraints on the
global atmospheric C02 budget. Science 247:1431-1438.
Thompson, M. T. 1989. Forest statistics for Georgia, 1989. USDA Forest Service
Resource Bulletin SE-109.
Tritton, L. M. 1980. Dead wood in the northern hardwood forest ecosystem. Ph.D.
dissertation, Yale University, New Haven, CT.
Tritton, L. M. and J. W. Hombeck. 1982. Biomass equations for major tree species on
the northeast. USDA Forest Service General Technical Report NE-69.
UhI, C. and J. B. Kauffman. 1990. Deforestation, fire susceptibility, and potent aI tree
responses to fire in Eastern Amazon. Ecology 71:437-449.
Vissage, J. S. and K. L. Duncan. 1990. Forest statistics for Tennessee counties-1989.
USDA Forest Service Resource Bulletin SO-148.
Vogt, K. A., C. C. Grier, and D. J. Vogt. 1986. Production, turnover, and nutrient
dynamics of above- and below-ground detritus of world forests. Advances in
Ecological Research 15: 302-377.
Whigham, D. F., I. Olmsted, E. C. Cano, and M. E. Harmon. 1991. The impact of
Hurricane Gilbert on trees, litterfall, and woody debris in a dry tropical forest in the
northeastern Yucatan Peninsula. Biotropica 23:434-441.
Woodwell, G. M., R. H. Whittaker, W. A. Reiners, G. E. Likens, C. C. Deiwiche, and D. B.
Botkin. 1978. The biota and the world carbon budget. Science 199:141-146.
Vavitt, J. B. and T. J. Fahey. 1982. Loss of mass and nutrient changes of decaying
woody roots in lodgepole pine forests, Southeastern Wyoming. Canadian Journal of
Forest Research 12:745-752.
178
-------
0. Net Primary Productivity of Non.forested Lands
George A. King 1
Carbon flux on forested lands in the United States is summarized in Section 4 of this
report. In this section, net primary productivity (NPP) of nonforested lands in the
conterminous US is estimated in order to complete a first approximation of NPP in the
US. The treatment of the land base in Section 4 and this section is consistent with the
data developed by the USDA for the 1989 Resource Protection Act Assessment
(USDA 1989). Estimates of NPP are made for the three major nonforesi cover types in
the US as identified by the Soil Conservation Service (SCS): rangeland, cropland,
and pastureland (USDA, SCS 1987). Estimates of NPP are also made for the
remaining nonforested land not included in these three land use types, here termed
“other land. This category is described in more detail later in this summary. Data on
wetland productivity are also presented, although wetland is not a primary land cover
type used by the SCS (USDA, SCS 1987). Wetlands are included primarily under
forested lands (for forested wetlands) and in the “all other land’ cover type. In addition,
some wetlands are also considered to be open water and not described as wetlands
by SCS, so the total wetland area presented here is an underestimate of the US total.
Methods
Two generalS steps were used in est mating nonforest NPP. The areal extent of each
land cover type in each state was first estimated, and then the NPP per unit area of
those cover types was estimated using averages by state of site specific
measurements of NPP. Total annual NPP was calculated by multiplying the area
estimates and NPP per unit area estimates together.
All nonfederal land cover estimates are based on the 1982 Natural Resources
Inventory (NRI) database (USDA, SCS 1987). The 1982 NRI is a statistically-based
sampling of land use on nonfederal lands in Ihe US conducted by the SCS. The
sampling strategy was designed to provide natural resource data at a substate level,
specifically for Major Land Resource Areas (USDA, SCS 1981). In the analysis
presented here, the NRI data are summarized by state, and then aggregated into
regions of the US (Table 9.D.1) as defined by the USDA (1989). Since the 1982 NRI
did not include federal land, estimates of federal rangeland by region were obtained
from the USDA (1989). The area of federally owned pastureland and cropland was
assumed to be insignificant.
Different methods were used to estimate the productivity of each land cover type
depending on the type and spatial density of the available estimates of site specific
NPP, as detailed below. In all cases, dry organic matter weights were divided by two
to convert to their carbon equivalent weight.
1 ManTech Environmental Technology, US EPA Environmental Research Laboratory. 200 SW 35th
Street, Corvallis, OR 97333. USA.
179
-------
Section 9.D
Table 9.D.1. Regions of the United States as defined by USDA (1989).
Northeast Southeast Great Plains
Connecticut Florida Kansas
Deraware Georgia Nebraska
Maine North Carolina North Dakota
Maryland South Carolina South Dakota
Massachusetts Virginia Pacific Northwest
New Hampshire South Central Alaska t
New Jersey Alabama Oregon
New York Arkansas Washington
Pennsylvania Kentucky Pacific Southwest
Rhode Island Louisiana California
Vermont Mississippi Hawaii
West Virginia Oklahoma
North Central Tennessee
Illinois Texas
Indiana Rocky Mountains
Iowa Arizona
Michigan Colorado
Minnesota Idaho
Missouri Montana
Ohio Nevada
Wisconsin New Mexico
Utah
Wyoming
Not included in this study
Rar geland Productivity
Rangeland is defined as “land on which the climax vegetation (potential natural plant
community) is predominantly grasses, grass-like plants, lorbs or shrubs suitable for
grazing and browsing” (USDA SCS 1967). Thus rangelands include the tall, mixed,
and short grass prairies of the central US, as well as shrublands in the Great Basin,
deserts in the Southwest and many wetlands.
Several steps were taken to estimate rangeland NPP. First, the amount of above-
ground production for each NRI rangeland sample point was estimated using the
Soils-5 interpretation record number and its corresponding estimate of above-ground
NPP (ANPP; USDA, SCS 1983). ANPP estimates for favorable, normal, and
unfavorable years are made for each Soils-5 interpretation record by the SCS.
Favorable, normal and unfavorable are defined based on total annual precipitation.
The wettest 10% of the years in the historical record for a site are termed favorable,
while the driest 10% of the years are termed unfavorable (USDA. SCS 1976).
180
-------
Section 9.D
Estimates of ANPP for normal years are used in this analysis. A total of 8,272 Soils-5
production estimates were used in estimating rangeland ANPP. Each ANPP estimate
was multiplied by the area represented by the NAt sample point to estimate the total
annual ANPP for the rangeland represented by that sample point. Mean ANPP per
hectare of each USDA region (USDA 1989) was calculated as an area-weighted
average of the ANPP values of all the NR( sampfe points in the region.
The next step was to estimate total NPP from the estimates of ANPP. Few studies
have estimated both ANPP and below-ground net primary productivity (BNPP) for
grasslands (Coupland 1979). ANPP and BNPP data at nine central grassland sites
were averaged together to estimate the proportion of total plant production that occurs
underground. At these sites, BNPP is twice as great as ANPP, so the ANPP estimates
for rangetands were tripled to estimate total plant NPP.
NAt data and the corresponding Soils-5 interpretation database are only available for
non-federal lands. In order to estimate total NPP for all US rangelands, it was
assumed that rangeland productivity was the same for federal and nonfederal lands in
each region of the US. Thus, estimates of total production on federal rangeland were
made by multiplying the total federal rangeland area in each region by the mean NPP
of the nonfederal rangeland in the same region.
Crooland Productivity
Cropland productivity was estimated by converting crop yield into total plant
production for each state. Crop yields per unit area were obtained for 19B4 from
USDA (1986) as compiled by Tom Moser (ManTech Environmental Technology,
Corvallis, OR, personal communication). Actual yield data for each crop were used for
each of the top ten producing states in the US. The mean crop yield per unit area of
the top ten states was used as an estimate of the mean yield in the remaining states.
Some of the NAt crop cover types are not crop specific, such as “alt other row crops”.
The average NPP of crops for which yield information was available was used to
estimate the NPP of these general cover types.
Crop yield was converted to total plant yield using the conversion factors listed in
Sharp et al. (1976) or by using the harvest indices from Donald and Hamblin (1976).
The conversion factors in Sharp el at. (1976) account for below-ground as well as
above-ground production. They do not account for moisture content of the crop. The
total plant production was therefore reduced by the percent moisture content of the
crop, as estimated by Sharp et at. (1976) or in the American Society of Agticultural
Engineers Standards Handbook (ASAE 1991). If harvest index values were used to
estimate total plant production, they were modified to estimate below-ground
production. After a mean NPP was estimated for each crop, the total primary
production for each state was calculated by multiplying the mean NPP by the crop
acreage from the 1982 NAt.
Pastureland Productivity
Pastureland is defined by the SCS to include “(and used primarily for production of
introduced or native forage plants for livestock grazing. Pastureland may consist of a
single species in a pure stand, a grass mixture, or a grass-legume mixture” (USDA,
181
-------
Section S.D
SCS 1987). Pastureland can be intensively managed, including fertilization
treatments and weed control.
Pastureland productivity for each state was assumed to be equal to that of haylands in
the same state. Hayland yield was obtained from USDA (1986), and converted to total
plant yield assuming total NPP is twice the above-ground productivity. No data on
below-ground productivity in pasturelands or haylands was found for this study, so
doubling above-ground productivity to estimate total NPP is viewed as conservative.
The productivity data were corrected for moisture content, assuming a moisture
content of 14% (Sharp et al. 1976). The areal estimates of pastureland obtained from
the 1982 NRI database were multiplied by these productivity estimates to calculate the
total primary production for each state.
Other Land Productivity
This category includes the remainder of land cover types recognized in the 1982 NRI
database, including “other cropland, other land in farms, barren land, urban land, and
rural transportation.” This category constitutes about 9.7% of the surface land area of
the US. About 25% of this land is unvegetated and thus has no NPP. NPP per unit
area of the remainder of this category was assumed to be equal to the combined mean
NPP (per unit area) of rangelands, croplands, and pasturelands (3838 kg ha’ 1 yr-i).
Total primary production was estimated by multiplying this mean NPP by the total area
of other land covered by vegetation.
Wetland Productivity
The area of wetlands in each state in the conterminous US was estimated using the
NRI dataset (USDA, SCS 1987). As previously mentioned, this is not an independent
land use category in the NPI database, but is a descriptor of one of the other land
cover types. About 52% of the wetland NRI samples points are classified as forests,
23% are in the “all other land” category, and 7% are included in the rangeland cover
type. Thus, it is inappropriate to include much of the wetland area in the overall
estimate of nonforest productivity. Furthermore, the NRI classifies some wetlands as
water, so the NRI underestimates wetland area in the US. Nevertheless, an estimate
of the productivity of the sampled wetlands in the US was made in order to ascertain
the relative importance of wetlands in the overall carbon balance of the United States.
Few estimates of net primary productivity in wetlands are available, and estimates of
below-ground production In forested wetlands are especially rare. Productivity
estimates were obtained from the summaries by Brinson et al. (1981) and Bradbury
and Grace (1983). Estimates were available from 45 different sites, and these data
were first averaged by state and then by region. Relative below-ground productivity
was estimated using all the data available for the US where below-ground productivity
was measured. This proportion was used to estimate total production at sites where
below-ground productivity was not measured. Wetland productivity for a state was
estimated either using the average productivity for the state as calculated from the data
summaries, or using the regional mean productivity if no state data were available.
182
-------
Section 9.D
Sensitivity analysis of relative below-ground productivity
One of the greatest uncertainties in this data analysis is the relative proportion of total
plant productivity contributed by below-ground portions of plants. Below-ground
productivity can contribute up to 40 to 90% of total plant primary production and is
especially important in grasslands (Stanton 1988). Unfortunately, measurements of
below-ground productivity are rare, In this analysis, the few data available on below-
ground productivity were used to convert estimates of above-ground productivity into
total plant productivity. Because these estimates of below-ground productivity are few
and imprecise, the effect of changing the relative proportion of below-ground
productivity on NPP in rangelands, pasturelands, and croplands was investigated. The
sensitivity analysis consisted of calculating total productivity, assuming the below-
ground productivity was twice that of above-ground productivity, equal to above-
ground productivity, or half above-ground productivity.
Results
Nonforested land composes about 70% of the land surface in the US (USDA, SCS
1987). The NPP of these lands is estimated as 2.03 Pg-C yr 1 or 3703 kg-C ha 1 yr 1
(Table 9.D.2). Rangeland is the dominant land cover type, accounting for about half of
the nonforested land (Table 9.D.2, Figure 9.D.1). Cropland is the second largest land
cover type, accounting for about a quarter of nonforested land. However, rangelands
have the lowest NPP per unit area, averaging less than half the productivity of
cropiands (Table 9.D.2). Consequently, croplands account for more of the total
nonforest NPP than rangelands (Fig. 9.0.2). Pasturelands are intermediate between
croplands and rangelands in their NPP per unit area. Other land accounts for less
than 15% of the nonforested land area and productivity.
Land cover and productivity data for pasturelands, rangelands, and croplands has
been combined by region (Table 9.D.3, Figs. 9.D.3-4). The Rocky Mountain region
contains about one third of the above three land cover types, with the South Central,
Great Plains, and North Central regions containing over 10% of these land cover
types. Only 2% of these land cover types are found in the Northeast.
In terms of total production by region, the South Central region accounts for the
highest total NPP. The Rocky Mountains contributes relatively little NPP, considering
that it has the largest total area of rangeland, cropland, and pastureland. Mean total
NPP for the Rocky Mountains is about half that of the national average for rangelands,
cropands, and pasturelands (Table 9.D.3). NPP per unit area is about the same for
the easternmost regions (including the Great Plains), but is substantially less for the
western regions. Recall though that these values are for nonforest land use types.
Forested areas in the West occur at higher elevations with greater rainfall than the
lowland areas where the nonfotest cover types are located. Thus, forested NPP can
be expected to be greater, increasing the NPP per unit area of the western regions. In
the East, the climate of nonforested areas does not differ significantly from forested
areas, so including forest NPP in the eastern estimate may not change the NPP per
unit area values as much as it would in the West.
183
-------
Section .
Table 9.D.2. Summary table of nonforested land productivity
Land Use Type
Area 2
(x 106
Mean
Net Primary
Productivity
(kq-C ha 1 yr t )
Total Net Pnmary
Production
(Tg-C)
Rangelands
264.8
2427
642.6
Pasturelands
54.0
4763
257.1
Croplands
Other Land 1
154.7
74.9
5924
2867
916.3
214.7
US Total
548.4
3703
2030.7
Other Land = Cropland other (summer fallow. aquacuPture, other cropland not planted, grassland in
rotation). other land in farms, barren land, other lands, urban and bui up land, rural transportation
(SCS 1987).
2 Based on ‘1982 NRI data (SCS 1987) for non-federal land and USDA (1989) for federal rangeland
Calculated by dividing the total nontorest net primary production in the US by the total nonforested
area in the US
Table 9.D.3. Estimates of net primary productivity for rangelands, croplands and
pasturelands by region.
Region 1
(x
Area
106 ha)
Mean
Net Primary Productivity
(kg-C ha ’ 1yr 1 )
Total Net Pnmary
Production
(Tg .C)
Northeast
11.5
5439
62.5
North Central
68.6
5682
390 0
Southeast
15.2
5415
82.2
South Central
98.1
4840
474.6
GreatPlains
66.7
5211
347.6
RockyMountains
158.6
1986
315 1
Pacific Northwest
34.3
1979
67.8
Pacific Southwest
20.5
3715
76.1
USTotal
473.5
38352
1815.9
1 See Table 9 .D.1. for list of states in each region (USDA 1989).
2 Calculated by dividing the total net primary production on rangelands, croplands, and pasturelands by
the total land area of these cove’ types in the US.
184
-------
Section 9.D
Relative Distribution of Nonforest Area by Cover Type
Rangelands (48.3%)
Croplands
(28.2%)
Pasturelands
(9.8%)
Other Lands
(13.7%)
Figure 9.D.1. Relative distribution of nonforested land in the conterminous United
States, summarized by cover type. See Table 9.D.2. for absolute estimates of areal
coverage of these four principal cover types.
I O
-------
Section 9.0
Relative Distribution of Nonforest Net Primary Production by
Cover Type
Rangelands (31.6%)
:. her Lands
(10.6%)
‘ -I
<4’. i., .. :;
S
•‘
Pastur&ands
I I fJ• LI!
Croplands
(45.1%) 1
Figure 9.0.2. Relative distribution of nonforest net primary production, summarized by
cover type. See Table 9.0.2. for absolute estimates of NPP by cover type.
186
-------
Section 9.D
Relative Distribution of
North Central
(14.5%)
Northeast (2.4%)
Pacific Southwest
(4.3%)
Pacific Northwest
(7.2%)
Rangeland, Cropland, and Pastureland
Area by Region
Southeast
(3.2%)
South Central
(20.7%)
Great Plains
(14.1%)
Mountains
(33.5%)
Figure 9.0.3. Relative distribution of rangeland, cTopland, and pasturelana area in me
United States, summarized by region. See Table 9.D.3. for absolute estimates of areal
coverage by region.
Rocky
187
-------
The NPP data have also been summarized for each land cover type in each region
(Tables 9.D.4-6), Over halt of the rangeland area occurs in the Rocky Mountains
(Table 9.0.4). Most rangeland NPP occurs in the Great Plains, Rocky Mountains, and
South Central regions. On a per unit area basis, the highest productivity occurs in the
North Central and Southeast regions, where the NPP per unit area is about three
times higher than the national average (Table 9.D.4).
About 80% of the cropland in the US occurs in the North Central, South Central, South
East, and Great Plains regions (Table 9.D.5). These regions also contribute about
80% of the total crop productivity in the US. NPP per unit area is about the same in all
regions, except for the Rocky Mountains being relatively low and the Pacific Southwest
(California) being relatively high. The California NPP is high probably because of the
intensive use of irrigation.
Almost three quarters of the pastureland in the US occurs in the North Central,
Southeast and South Central regions (Table 9.D.6). These three regions account for
about the same proportion of total primary production. All regions have about the
same NPP per unit area, except for the Pacific Southwest where Ihe estimated NPP is
about twice the national average, again, probably because of the intensive use of
irrigation.
Wetlands constitute about 4% of the land in the US (Table 9.D.7) as sampled by the
1982 NRI. Total estimated NPP for wetlands is about 11% of the total estimated for
nonforested land. Most of the wetland area and production is concentrated in the
North Central, South Central, and Southeast regions. Recall, however, that these
values are underestimates, because some wetland areas are considered open water
in the NRI.
Discussion
Our estimate of nonforest NPP can be combined with information on forest NPP from
Section 4 10 estimate total NPP for the US. The total annual wood production from the
analysis in Section 4.B is 337 Tg-C yr-I. For a very rough approximation that number
could be tripled to account for foliage and fine root production. This would bring the
total national NPP (forest plus nonforest) estimate to 3.04 Pg-C yr 1 .
This estimate can be compared with modeled values of NPP for potential natural
vegetation in North America recently completed by McGuire et al. (1992). They
estimated a total NPP of 7.03 Pg-C yr 1 for all of North America. The conterminous US
comprises about 37.8% of North America, and assuming that NPP is uniform across
North America, the US should account for 2.65 Pg-C yr 1 of the modeled NPP. Our
eslimate of 2.84 Pg-C yr 1 NPP in the US is consistent with this estimate. Furthermore,
McGuire et at. (1992) simulated a NPP per unit area of 3,387 kg-C ha 1 yr 1 . Our
eslimate of nonforest NPP is 3, 703 kg-C ha 1 yr 1 1 which faUs within 10% of the
modeled estimate for North America. NPP per unit area is highest in the eastern US in
the model sirnutations, as it is in the estimates made here. Rangeland is the least
disturbed nonforest land cover type, and thus is probably most appropriate to compare
with equivalent land cover types in the potential natural vegetation classification used
188
-------
Section 9.D
Relative Distribution of Rangeland, Cropland, and Pastureland
Net Primary Production by Region
Southeast
(4.5%)
North Central
(21.5%)
Northeast (3.4%)
Pacific Southwest
(4.2%)
Pacific Northwest
(3.7%)
South Central
(26.1%)
Great Plains
(19.1%)
Mountains
(17.4%)
Figure 9.D.4. Relative distribution of net primary production of rangeland, cropland,
and pastureland in the United States, summarized by region. See Table 9.D.3. for
absolute estimates of NPP by region.
Rocky
189
-------
Table 9.D.4. Estimate of net primary produclivity for federaf and non-federal
rangelands in the conterminous US.
Region
Area 1
(x 106 ha)
Mean
Net Primary
Productivity
(kg-C ha 1 yr 1 )
Total Net
Primary
Production 2
(Tg-C)
Northeast
0
0
0
North Cenlral
0.2
7947
1.2
Southeast
0.6
7878
4.8
South Central
45.2
4053
183.1
Great Plains
31.2
4406
137.4
Rocky Mountains
142.5
1663
237.0
Pacific Northwest
29.2
1381
40.3
Pacific Southwest
160.0
2428
38.8
US Total
264.8
2427 642.6
1 Based on 1982 NRI data (USDA, SCS 1987) for non-federal rangelands and USDA (1989) for federal
rangelands
2 Based on productiv y estimates from Soils 5 (USDA. SCS 1983) interpretation database (SCS 1983).
3 Calculated by dividing the total net primary production on rangelands by total rangeland area in the
US.
Table 9.D.5. Estimate of net primary productivity for croplands in the conterminous
US.
Region
Area 1
(x 106 ha)
Total Net
Pnmary
Production 2
(Tg•C)
Northeast
7.1
5816
41.4
North Central
54 1
5862
317.0
Southeast
8.9
5538
49.1
South Central
31.9
6209
197.8
Great Plains
32.1
6105
195.8
Rocky Mountains
13.2
4815
63.0
Pacific Northwest
3.7
5609
20.8
Pacific Southwest
3.9
8043
31.5
US Total
154.7
5924
916.3
1 Based on 1982 NRI data (USDA. SCS 1987) for non-federal land.
2 Based on crop yields form USDA (1986) and conversion I actois listed in Sharp et al. (1976) and
Donald and Harnblin (1976).
3 Calculated by dividing total net primary production on croplands by total cropland area in the US.
190
-------
Section 9.D
Table 9.D.6. Estimate of net primary productivity for pasturelands in the conterminous
United States.
Mean Net
Total Net
Region
Area 1
x i06 ha
Pnmary
Productivity 2
(kg-C ha 1 yr 1 )
Pnmary
Productivity
(Tg-C)
Northeast
4.4
4825
21.1
NorthCentral
14.4
4983
71.9
Southeast
5.7
4962
28.4
Soulh Central
21.0
4457
93.7
Great P$ains
3.4
4171
14.4
Rocky Mountains
3.1
4962
15.1
PachicNorthwest
1.4
4962
6.7
Pac ic Southwest
0.6
9916
5.8
LiSTotat
54.0
4763
257.1
1 Based on 1982 NRI data (USDA. SCS 1987) for non-federal lands.
2 Based on hayland productivity given in USDA (1986)
Calculated by dividing total primary production on pasturelands by total pastureland area in the US.
Table 9.D.7. Estimate of net primary productivity for wetlands in the conterminous US.
Region
Area 1
(x 106 ha)
Mean Net Primary
Productivity
(kg-C ha 1 yrl)
Total Net Primary
Productivity 2
(Tg.C)
Northeast
3.2
8264
264
NorthCeritral
7.6
8088
61.5
Southeast
8.2
6141
50 4
South Central
7.3
8663
63.2
Great Plains
2.1
2795
5.9
Rocky Mountains
1.4
11949
16.1
Pac ic Northwest
0.5
2618
1-3
Pacthc Southwest
0.6
2618
1.5
US Total
30.8
7346
226.3
1 Based on 1982 NRI data (USDA, SCS 1987) for non-federal land.
2 Based on productivfty estimates from Bronson et al. (1981) and Bradbury and Grace (1983) database
3 Calculated by dividing total net primary production on wetlands by the total wetland area in the US
191
-------
by McGuire et at. (1992). Our estimate of rangeland NPP falls between the maximum
and minimum values of NPP estimated for short and tall grasslands and arid
shrublands by McGuire et al. (1992)
The rangeland estimates in the Great Plains can also be compared to those of Sala et
at. (1988), who used a similar NPP database based on the SCS data to estimate
ANPP (Joyce et at. 1986). The mean ANPP of the Great Plains region calculated here
(1,470 kg-C ha 1 equivalent to 294 g rn 2 of biomass) is consistent with the location of
the 300 9 rn 2 isopleth in Sala et al. (1988).
That the estimates presented here are comparable to the modeled values is
encouraging but not that surprising given that the Terrestrial Ecosystem Model (TEM)
model of McGuire et at. (1992) was partially calibrated using actual measured
estimates of NPP at specific locations in North America. The more significant question
is how well do the measured values of NPP correspond with actual NPP, which is
discussed next.
Productivity estimates of grasses, herbs and crops are primarily based on the change
in plant biomass during the growing season (e.g., following methods developed by the
lnternalional Biological Program [ IBP); Cooper 1975, Coupland 1979). Long et at.
(1989) demonstrated that this method can underestimate NPP up to fivefold because
it does not account for biornass turnover between measurements. This is most critical
in regions that are relatively aseasorial (Long and Hutchin 1991). The revised
techniques of Long et at. (1989) have not been used to measure NPP in Great Plains
grasslands as of yet, so the degree to which the IBP methods underestimate NPP in
the US is uncertain.
The relative allocation of NPP above and below-ground is the most uncertain aspect of
these analyses because there are relatively few measurements of below-ground
production. Changing the relative partitioning of above and below-ground production
has a large effect on mean NPP per unit area. For instance, if one assumes that
below-ground production is 50% of above ground production, mean NPP in
rangelands is 1,220 kg-C ha t yr 1 . However, if one assumes that below-ground
production is twice that of above-ground production, than mean NPP becomes 2,430
kg-C ha l yr 1 1 double the previous estimate. Data from IBP studies in the prairie
suggest that BNPP is twice that of ANPP (Coupland 1979). Consequently, this value
was used to convert the rangeland ANPP data to estimates of NPP. Our use of a high
proportion of BNPP relative to NPP in grasslands is supported by Stanton (1988).
However, recent isotopic measurements suggest that grassland BNPP is more likely
equal to ANPP than double ANPP (Milchunas and Lauenrolh 1992). Additional
measurements and evaluation of methodologies are required to improve BNPP
estimates, and thus estimates of total plant NPP.
The estimates of cropland production given here are probably the mosi precise of all
four land types because of the detailed yield data available on a slate by state basis.
Still, the estimate could be greatly improved by using region-specific conversion
factors from crop yield to plant production. Also, betler estimates are needed of
moisture content of specific crops. The rangeland ANPP data is also relatively precise,
because of the large number of ANPP estimates available from the Soils-5 database.
Pastureland and other land production estimates are the least precise. No direct
192
-------
Section 9.D
estimales of pastureland NPP were used in this analysis. The other land I4PP
estimate was made assuming that vegetated land in that category has the same NPP
per unit area as nonforesi land in the US. For instance, this estimate could be
improved by obtaining data on urban land NPP.
The wetland estimate is only a first approximation since the land cover data are
incomplete for wetlands. More wetland productivity data are also required.
These nonforesi NPP estimates are derived from a mixture of productivity estimates
based on average climate conditions (i.e., rangeland data) and from estimates made
for a specilic year (i.e., cropland data). Sala et al. (1988) has shown that the difference
in NPP between favorable and unfavorable years can be 30 to 90% of the average
NPP. Consequently, annual NPP can be expected to vary significantly year to year,
depending on the differences in seasonal weather conditions. tnterannual variability
of NPP could be estimated using currently available data, and would lead to a fuller
understanding of biospheric carbon dynamics.
References
American Society of Agricultural Engineers (ASAE). 1991. Standards 1991.
Standards, Engineering Practices, and Data. 38th edition. American Society of
Agricultural Engineers. St. Joseph, Ml.
Bradbury. I. K. and J. Grace. 1983. Primary production in wetlands. pp. 285-310 in,
Gore, A. J. P. (editor). Mires: Swamp, Bog, Fen and Moor. General Studies.
Ecosystems of the World 4A. Elsevier Scientific Publishing Company. Amsterdam
Brinson, M. M., A. E. Lugo. and S. 8rown. 19B1. Primary productivity, decomposiLion
and consumer activity in freshwater wetlands. Annual Review Ecological Systems
12:123-161.
Cooper, J. P. 1975. Photosynthesis and productivity in different environments.
Cambridge University Press, Cambridge, England.
Coupland, R. T. 1979. Conclusion. pp. 335-355 in Coupland, R. T. (editor) Grassland
ecosystems of the world: analysis of grasslands and their uses. Cambridge University
Press, 401 pp.
Donald, C. M. and J. Hamblin. 1976. The biological yield and harvest index of cereals
as agronomic and plant breeding criteria. Advanced Agronom(cs 28:361-405.
Joyce, L. A., D. E. Chalk, and A. Vigil. 1956. Range forage data base for 20 Great
Plains, Southern and Western states. USDA Forest Service General Technical Report
RM-1 33.
Long, S. P. and P. R. Hutchin. 1991. Primary production in grasslands and con fèrous
forests with climate change: an overview. Ecological Applications 1:139-156.
193
-------
becilon
Long, S. P., E. Garcia Moya, S K. Imbamba, A. Kamna lrut, M. T. F. Piedade, J. M. 0.
Scurlock, V. K. Shen, and 0. 0. HaIl. 1989. Primary productivity of natural grass
ecosystems of the tropics: A reappraisal. Plant and Soil 115:155-166.
McGuire, A. D., J. M. Melillo, L. A. Joyce, 0. W. Kicklighter, A. L. Grace, B. Moore Ill, and
C. J. Vorosmarty. 1992. Interactions between carbon and nitrogen dynamics in
estimating net primary productivity for potential vegetation in North America. Global
Biogeochemical cycles. 6:101-124.
Milchunas 0. G. and W. K. Lauenroth. 1992. Carbon dynamics and estimates of
primary production by harvest, 14 C dilution, and 14 C turnover. Ecology 73: 593-607.
Moser, Tom. Personal communication.
Sala, 0. E., W. J. Parton, L. A. Joyce, and W. K. Lauenroth. 1988. Primary production
of the central grassland region of the United States. Ecology 69: 40-45.
Sharp, D. D., H. Leith, G. R. Noggle, and I-I. D. Gross. 1976. Agricultural and forest
primary productivity in North Carolina 1972-1973. North Carolina Agricultural
Experiment Station, Technical Bulletin No. 241.
Stanton, N. L. 1988. The underground in grasslands. Annual Review Ecological
Systems 19:573-589.
USDA. 1986. Agricultural statistics. US Government Printing Office, Washington, DC.
USDA Forest Service. 1989. An analysis of the land base situation in the United
States: 1989-2040. General Technical Report RM-181. Rocky Mountain Forest and
Range Experiment Station, Fort Collins, CO. 76 pp.
USDA, Soil Conservation Service (SCS). 1976. National range handbook.
Washington, DC, USA.
USDA, Soil Conservation Service (SCS). 1981. Land resource regions and major
land resource areas of the United States. Agriculture Handbook 296. Washington,
DC. USA.
USDA, Soil Conservation Service (SCS). 1983. National Soils and Agricultural
Handbook Number 430. US Government Printing Office, Washington, DC.
USDA, Soil Conservation Service (SCS). 1987. Basic Statistics, 1982 National
Resources -Inventory. Iowa State University, Statistical Laboratory, Statistical Bulletin
Number 756, 153 pp.
194
-------
E. Glossary
Age class 1 . The midpoint of an interval of
stand age, where stand age is based on trees in
the stand sampled at breast height.
Annual mortality 2 : The volume of sound
wood in trees that died from natural causes
during a specified year.
Annual removals 2 : The net volume of trees
removed from the inventory during a specified
year by harvesting. cu uraI operations such as
timber stand improvement, or land clearing.
ATLAS: A USDA Forest Service forest
inventory model ATLAS allocales prescribed
harvest levels among forest types and age
classes and updates forest inventory (growing
stock volumes) over time At present only lands
under private ownership are treated
Base year The reference year for which
carbon pools and flux are estimated
Breast height. Point on tree approximately
1 37 rn (4.1/2 feet) above ground
Bureau of Land Management (BLM) 2 . An
ownership class 01 Federal lands administered
by the Bureau of Land Management, US
Department of the Interior.
Coarse material 2 : Wood residues suitable for
chipping. such as slabs, edgings, and trimmings.
Commercial species 2 : Tree species suitable
for industrial wood products.
Cull tree 2 : A live tree, 12.5 cm (5.0 inches) in
diameter at breast height (dbh) or larger, that is
unmerchantable for sawlogs now or
prospectively because of rot, roughness. or
species. (See definitions for rotten and rough
trees.)
Diameter class 2 : A classification of trees
based on diameter outside bark measured at
breast height Dbh is the common abbreviation
for “diameter at breast height.” With 5 cm (2 inch)
diameter classes, the 18 cm (6 inch) class, for
example, includes trees 12.5 through 17.2 cm
(5.0 through 6.9 inches) dbh, inclusive.
Douglas.flr subregion 2 : The area in Oregon
and Washington that is west 01 the crest of tte
Cascade Range.
Farmer 2 : An ownership class of private lands
owned by a person who operates a farm, either
personally doing the work or directly supervising
the work.
Federal 2 : An ownership class of public lands
owned by the U$.Government.
Fiber product 2 Products derived from wood
and bark residues. such as pulp. composition
board products. and wood chips for export
Fine materials 2 : Wood residues not suitable
for chipping. such as planer shavings and
sawdust
Forest floor. All dead. organ c matter above thern.
mineral soil horizon exceot woody debriS.
Forest industry 2 : An b ners ’hip . class df
private lands owned by companies or individuals
operating wood-using plants
Forest Inventory and Analysis (FIA).
USDA Forest Service Inventory work group
Forest land 2 : Land at least 10% (16 7% in
southeast Alaska) stocked by forest trees of any
size, including land that formerly had such tree
cover, and that will be naturally or artificially
regenerated. Forest land .includes..transdion
zones, such as areas between heavily forested
and nonforested lands that are at least 1 %
stocked with forest trees, and forest are as
adjacent to urban and built-up lands. At o
included are pinyonjuniper and chaparral areas
In the West and afforested areas. The minimum
area for classification of forest land is one acre.
Roadside, streamside, and shelterbelt strips of
limber must have a crown width at least 36 m
(120 feet) to qualify as forest land. Unimproved
roads and trails, streams, and clearings in forest
areas are classified as forest if less than 36 m
(120 feet) wide.
195
-------
Section 9.E
Forest type 2 : A classification of forest land
based on the species presently forming a
plurality of the live-tree stocking.
Fuelwood 2 : Wood used for conversion to
some form of energy, primanly residential use.
Growing stock 2 A classification of timber
inventory that includes live trees of commercial
species meeting specified standards of quality or
vigor. Cull trees are excluded When associated
with volume, includes only trees 12.5 cm (5.0
inches) dbh nd larger.
Hardwood 2 : A dicotyledonous tree, usually
broad-leafed and deciduous.
industrial wood 2 : All commercial roundwood
products -except fuelwood.
Land area 2 (a) Bureau of the Census: The
area of dry land and land temporarily or partly
covered by water, such as marshes, swamps,
and river food plains; streams, sloughs.
estuaries, and canals less than, 200 m
(one-eighth statute mile) wide: and lakes,
reservoirs, and ponds less than 40 acres (16 ha)
in area (b) ’F.orest Inventory and Analysis: Same
as (a) except that the minimum width of streams,
etc , is 36 ii’ (120 feet). and the minimum size of
lakes. etc., is 0 4 ha (1 acre). This latter definition
is the one used in t Iis publical ion
Litter: Recently fallen and decomposing
organic mailer above the mineral soil Does not
include wo.ody material greater than 2 cm
diameter.
LoggIng resIdues 2 : Downed and dead wood
volume left on the ground after trees have been
cut on timberland.
Merchantable: Trees of diameter greater than
12.5 cm except cull trees.
Merchantable bole .volume. Merchantable
wood v lume in the boles. Includes commercial
and noncommercial species.
Merchantable bole carbon: Tree
merchantable bole volume expressed as carbon.
The proportions of hardwood and softwood and
the relevant specific gravity and percent carbon
are used 10 make the conversion.
National Forest (NF) 2 : An ownership class of
Federal lands, designated by Executive Order or
statute as National Forests or purchase units.
and other lands under the administration of the
Forest Service including experimental areas and
Bankhead-Jones Titte I II lands.
Net annual growth 2 . The net increase in the
volume of trees during a specified year.
Components include the increment in net
volume of trees at the beginning of the specific
year surviving to ifs end, plus the net volume of
trees reaching the minimum size class during the
year, minus the volume of trees that died during
the year, and minus the net volume of trees that
became cull trees during the year.
Net cubic volume 2 : The gross volume in
meters (cubic feet) less deductions for rot,
roughness, and poor form. Volume is computed
for the central stem from a 30 cm (1 foot) stump
to a minimum 10 cm (4.0 inch) top oiameter
outside bark, or to the point where the central
stem breaks into limbs.
Non-Federal (Private)- Lands owned by
private individuals (e.g., farmers). private
industries, and Native and other corporations
Non commercial species 2 : Tree species of
typically small size, poor form, or inferior quality.
which normally do not develop into trees suitable
for industrial wood products.
Nonforest land 2 : Land that has never
supported forests and lands formerly forested
where use of timber management is precluded
by development for other uses (Note: Includes
area used for crops, improved pasture.
residential areas, city parks, improved roads of
any width and adjoining clearings, powerline
clearings of any width, and 0.4 to 16 ha (1- to
40-acre) areas of water classified by the Bureau
of the Census as land. If intermingled in forest
areas, unimproved roads and nonforest strips
must be more than 36 m (120 feet) wide, and
clearings, etc., more than 0.4 ha (1 acre), to
qualify as nonfo rest land.)
Nonstocked areas 2 : Timberland less than
10% (16.7% in southeast Alaska) stocked with
growing-stock trees.
Old Growth stand: A stand originating from
natural disturbance, with a stand age greater
than those specified in the APA yield tables.
Other Federal 2 : An ownership class of
196
-------
Section 9.E
Federal lands other than those administered by
the Forest Service or the Bureau of Land
ManagemenL
Other forest land’: Forest land other than
timberland’ and reserved timberland. It includes
available and rreserved unproductive forest land,
which Is incapable of producing annually 1.4
cubic meters per ha per year (20 cubic feet per
acre) of ind striaI wood under natural conditions
because of adverse site conditions such as
sterile soiI dry climate. pooi.drainage, high
elevation, steepness. or rockiness.
Other land 2 : Nonforest land less the area in
streams, sloughs, estuaries, md canals between
36 and 198 m (120 and 660 feet) wide arid lakes.
reservoirs, and ponds between 0.4 and 16 ha (1
and 40 acres) in area (that is. nonforest land less
non-Census water area)
Other private 2 : An ownership :class of private
lands that are riot owned by’lorest industry or
farmers.
Other products : A miscellaneous category
of roundwood products that includes such items
as cooperage, pilings, poles, posts. shakes,
shingles, board mills, charcoal, and export logs.
Other public 2 An ownership class that
includes all public lands except National Forest.
Other red oaks 2 A group of species in the
genus Ouercus that includes scarlet oak,
northern-pin oak, southern red oak, bear oak.
shingle oak, laurel oak, blackjack oak, water oak,
pin oak, willow oak, and black oak.
Other removals 2 : Unutilized wood volume
from cut or otherwise killed growing stock, from
non-growing stock sources on timberland (for
example, precommercial thinnings), or from
timberland clearing. Does not include volume
removed from inventory through reclassification
of timberland to reserved timberland.
Other sources 2 : Sources of roundwood
products that are non- growing stock. These
include salvable dead trees, rough and rotten
trees, trees of noncommercial species. trees
less than 12.5 cm (5.0 inches) dbh, tops, and
roundwood harvesled from nontorest land (e.g.,
fence rows).
Other white oaks 2 . A group of species in the
genus Ouercus that includes overcup oak,
chestnut oak, and post oak.
Ovendry ton: .One dry ton, 907 kg (2000
pounds) of biomass.
Ownership 2 : The property owned by one
ownership unit, including all parcels of land in the
United States.
Ownership unit2: A classification of
ownership encompassing all types of legal
entities having an ownership interest in land.
regardless of the number of people involved. A
unit may be -an individuaL.a -combination of
persons; a legal entity such as a corporation,
partnership, club, or trust: or a public agency. An
ownership unit has control of a parcel or group of
parcels of land.
Pg Petagram — i 015 g billion metric tons
gigaton
Ponderosa pine subregion 2 : The area in
Oregon and Washington that is east of the crest
of the Cascade Range.
Private. An ownership class which includes
forest industry, farmers, and other private lands
Productivity class 2 : A classificafton of forest
land in terms of potential annual cubic meters per
ha (cubic-loot volume growth per acre) at
culmination of mean annual increment in fully
stocked natural stands.
Public: An ownership class which includes
National Forest and other public lands.
Pulpwood 2 : Roundwood. whole-tree chips, or
wood residues that are used for the production
of wood pulp.
Reserved timberland 2 : Forest land that
would otherwise be classified as timberland
except that it is withdrawn from timber utilization
by statute or administrative regulation.
Roots: Tree carbon below the ground level
Residues 2 : Bark and woody materials that are
generated in primary wood-using mills when
roundwood products are converted to other
products. Examples are slabs, edgings.
tnmmings, miscuts, sawdust, shavings, veneer
cores and clippings, and pulp screenings.
197
-------
Section 9.E
Includes bark residues and wood residues (both
coarse and fine materials) but excludes logging
residues
Rotten tree 2 . A live tree of commercial
species that does not contain a sewlog now or
prospectively primarily because of rot (that is.
when rot accounts for more than 50% of the total
cull volume)
Roundwood products 2 : Logs, bolts, and
other round timber generated from’ harvesting
trees for industrial or consumer use.
Saplings Live trees between 0 4 cm and 2.3
cm in diameter at breast height
Sawtog 2 A log meeting minimum standards of
diameter, length, and defect, including logs at
least 2 4 m (8 feet) long sound and siraight. and
w h a minimum diameter inside bark of 15 cm (6
inches) for sotiwoods and 20 cm (8 inches) for
hardwoods, or meeting other combinations 01
size and detect specified by regional standards
Sawllmber 2 : A classification of timber
inventory that is composed of sawlimber trees
Sawtlmber trees 2 . Live trees of commercial
species containing at least one 3.6 m (12 fool)
sawlog or two noncontiguous 2 4 m (8 foot) logs,
and meeting regional specifications for freedom
from detect Softwood trees must be at least
225cm (90 inches) 127.5 cm (11.0 inches) in
Alaskal dbh. and hardwood trees must be at
least 27.5 cm (11.0 inches) dbh
Second Growth stands Originating from
harvest of old growth stands
Softwood 2 : A coniferous tree, usually
evergreen, having needles or scalelike leaves.
SoIl organIc carbon: Absolute weight of
carbon to a one meter depth. Includes all carbon
in the less than 2 mm fraction organic.
State 2 : An ownership class of public lands
owned by states or lands leased by states for
more than 50 years.
StockIng 2 : The degree of occupancy of land
by trees, measured by basal area or number of
trees by size and spacing, or both, compared to
a stocking standard: that is. the basal area or
number of trees, or both, required to fully utilize
the growth potential of the land.
Stumps Woody biomass between the roots
and the level of the cut.
Tg Teragram 1012 g= million metric tons =
megaton
ThIrd Growth stands: Originating’ from
harvest of second groWth stands
TImberland 2 . Forest land that is prodi)cjng: br
is capable of producing, crops of industriarwo&d
and not withdrawn from timber utilization by
statute or administrative regulation (Nd’te Areas
qualifying as timberland are capable of produ ing C
in excess 011.4 cubic meters per ha (20 cubic
feet per acre) pe r. year of industrial wood in
natural stands. Currently inaccessible and
inoperable areas are included)
Tops 2 : The’ wdod of a tree above
merchantable height (or above the point on the
stem 10 cm (4.Ornches) diameter outsidebark It
includes the usable material in the uppermost
stem.
Total blomass carbon: Total tree c Pbon plus
sapling carbon.
Total tree carbon:’ lJ livihg -tree carbon
including roots, tops. and branches and cull
trees Does not include saplings
Tree carbon. All liviñQ tree catbon In 1uding’
roots, tops. and branches and cuH trees i
Unreserved forest land 2 : Forest land that is
not withdrawn from use by st lute or.
administrative regulation. Also termed available.
Understory: All live vegetation except that
defined as merchantable tree cai ’bon ‘and
saplings.
Veneer iog2: A roundwoád product 1(om
which veneer is sliced orsawn and thãt’usually
meets certain standardS of minimum diarfiètér’
and length and maximum defect.
WeIght 2 : The weight of wood arid bark,
oven-dry basis (approximately 12% ‘rnois ure
content). .
Woodlands Forest land where timber species
make less than 10% stoc1 ing or fOreStlandtl at
198
-------
Section 9.E
is incapable of producing crops of industrial
wood under natural conditions, because of
adverse site conditions.
Woody debris. All dead organic matter above
the mineral soil with a diameter greater than two
centimeters Includes standing dead.
Yield table 2 : A tabular presentation of volume
per unit area and other stand characteristics of
even-aged stands by age classes, site classes.
species. and density.
Malor Eastern forest.tvoe oroups 3 -
White-red-Jack pine: Forests in which
eastern white pine, red pine, or jack pine, singly
or in combination, comprise a plurality of the
stocking. Common associates inclijde hemlock.
aspen. birch, and mapii
Spruce-fir: Forests in which spruce or true firs.
singly or in combination, comprise a plurality of
the stocking Common associates include white
cedar, tamarack, maple, birch, and hemlock.
Longleat-slash pine: Forests in which
longleaf or slash pine, singly or in combination,
comprise a plurality of the stocking Common
associates include other southern pines, oak,
and gum.
Lobiolly-shortleaf pine: Forests in which
loblolly pine, shortleaf pine, or southern yellow
pines, except longleaf or slas,h_ pine, singly or in
combination, comprise a plutaiity of the stocking
Common associates include oak, hickory, and
gum.
Oak-pine. Forests in which hardwoods (usually
upland oaks) comprise a plurality of the slocking.
but in which pine or eastern redcedar comprises
25 to 50% o! the stocking Cqrrimon associates
Include gbm, hickory, and yeliow-poplar:
Oak.hickory: Forests in which upland oaks or
hickory, singly or in combination, comprise a
plurality of the stocking except where pines
comprise 25 to 50%. in which case the stand is
classified as oak-pine. Common associates
include yellow-poplar, elm, maple, and black
walnut.
Oak .gum cypress: Bottomland forests jh
which tu élo blackgism, sweetgum, oaks,
southern cypress, singly or in combination,
comprise a plurality of the stocking except where
pines comprise 25 to 50%, in which case the
stand is classified as oak-pine. Common
associates include cottonwood, willow, ash, elm,
hackberry, and.maple.
Elm-ash-coftonwood: Forests in which elm,
ash, or cottonwood, singly, or in combination,
comprise a plurality of the’ stocking. Common
associates include willow, sycamore, beech, and
maple.
Maple-beech-birch: Forests in which maple,
beech, or yellow birch, singly or in combination.
comprise a plurality of tM stocking. Common
associates ’ include hemlock, elm, basswood, and
white pine.
Aspen-birch: Forests in which aspen, balsam
poplar, paper biréti, or gray birch, singly or in
combination, comprise a plurality of the stocking
Common associates include maple and balsam
fir.
Major Western forest-type aroups
Doug las.fIr: Forests in which Douglas fir
comprises a plurality of the stocking Common
associates .inciude-:western hemlock, western,
redcedar, the true firs, redwood, ponderosa
pine, and larch.
Hemlock.Sltka spruce: Forests in which
western hemlock and!or Sutka spruce comprise a
plurality of the stocking Common associates
include Douglas-fir. silver fir, and western
redcedar.
Redwood: Forests in which redwood
comprises a plurality of the stocking Common
associates include Douglas-fir, grand fir, and tan
oak.
Ponderosa pine: ‘Forests in which ponderosa
pine comprises p u , raiity of the stocking’.:’
Common associates include Jeffrey pine, sugar,
pine, limber iñe, ArliOna pine, Apache pine,’
Chihuahua pine, Douglas-fir, incense-cedar, and..
white fir.
Western white pine: Forests in which
western white pine comprises a plurality of the’
stocking. Common associates include western
redcedar, larch, white fir, Douglas-fir. lodgepole
pine, and.Engelrflann spruce.
Lodgepole pine: Forests in which lodgepole’
pine comprises a plurality of the stocking
199
-------
Section 9.E
Common associates include alpine fir, western
white pine. Engelmann spruce, as.pen, and
larch.
Larch: Forests in which wè tern larch
comprises a pluralfly of the stocking Common
associates include Douglas•fir, grand fir, western
redcedar, and western white p ine.
Fir—spruce: Forests in which true firs,
Engelmann spruce or Colorado blue spruce,
singly or in combination, comprise a plurality of
the stocking Common associates include
mountain hemlock and lodgepole pine.
Western hardwoods: Forests in which
aspen, red alder, or other western hardwoods.
singly or in combination, comprise aplurality of
the stocking.
Chaparral: Forests of héâvily branâhèd’
dwarfed trees or shrubs, usually evergreëh, Ihe
crown canopy of which at maturity covers more
than 50% 01 the ground and whose primary
value is watershed protection. The r ore
common chaparral constituents are species of,
Quercus, Cercocarpus., Garrya, Ceanothus,
Arctostaphylos, and Adeno tomaj Types
dominated by such shrubs as Artemisia,
Chrysothamnus. Purshia, -Gutierrazia, Qr
semidesert species. are not commonly
considered chaparral. -
Plnyon.Junlperi Fpresjs in which pinyDp_pine
or juniper, or both, •comprise a plurality ot the
stocking.
1 Husch, B., C. I. Miller and T. W. Beers. 1982. Forest Mensuration, Third Edifion, John Wiley & Sons. 402
pp
2 Extracted, w h metric inserts, from. Waddell, K L., 0 0. Oswald and’D. S..Powell. 1989. Forest ,$j4tistgcs
of the United States USDA Resource Bulletin PNW -RB-168 Pacific NørthwestResearch Station,
Portland, Oregon
3 AS Defined in Waddell, K. L., 0 0. Oswald. and 0. S. Powell 1989 Forest Statistics of the United
States USDA Resource Bulletin PNW-RB-168 Pacific Northwest Research .Stat on, Portland, OreQpn.
200
-------
Section 9.1
F. Metric an4 English Equivalents
Symbol Metric unit = English unit Symbol
• Length
—
cm 1 centimeter = 0.3937 inch in
m 1 meter = 3.28 feet ft
m 1609 meter = 1 mile mi
Area -
cm 2 1 centimeter 2 = 0.155 inch 2 in 2
m 2 1 meter 2 = 10.764 feet 2 ft 2
km 2 1 kilometer 2 - = 0.386 mile 2 m1 2
ha I hectare = 2.471 acre ac
— Seed.
Mass (weight)
g 1 gram 0.0353 ounce oz
kg 1 kilogram = 2.2046 pound lb
t 1 metric ton = 2204.6 pound lb
Volume
—
m 3 1 meter 3 = 35.31 feet 3 ft 3
Examples of Mass or Volume per untt area
—
1 m 3 ha 1 14.3 ft 3 ac- 1
1 m 2 ha 1 4.35 ft 2 ac- 1
Taken from: Husch, B., C. I. Miller, and T. W. Beers. 1982. Forest Mensuration, 3rd
edn. John Wiley & Sons. 402 p.
201
-------
Section 9.G
G. Acronyms
API Americarc Pa er Institute
ATLAS Aggregate Tjmberland Assessmept System
BGC Bipgeochernistr.y model
CENTURY Soil carbon dyr amics model
DBH Diameter at Breast Height
DoC US Departm t or Commerce
EPA Et vironmenta1 Pràtection Agency
E RL-C EnvironmentaI èsearch Laboratory - Corvallis
EROS Earth Resource Observation System, USGS’
FCM Forest Carbon ? odel
FIA Forest Inventory and Analysis, USDA’ Fô ’èst Ser ice
FORCARB FORest CARBon model
FORENA Population-based Forest Dynamics mp eI
FORPLAN FORest PLANning model
GCM Global Circulation Model
GISS Goddardlnsti ute for Space Studies
GI 1 FP Gross National Product
IPCC Intergovernmental Panel on Climate Change
LAI Leaf Area Index
MAPSS MappedAtmosphere Plant Soil System
NPP Net Primary Production or Net Primary Prodücti ity
NRI National Resource Inventory
OSB Oriented Strand Board
OSU Oregon State University
PET Potential Evapotranspi ration
RPA Forest id Rangeland Renewable Resou e 1an ii g c of1974
TAMM Timb r Ass èssment Market Model r
T EGRO Phy IogicaIly-based Tree Growth Model
UKMO United Kingdom Meteorological bflice
US United States
US DA United Sta e Department of Agriculture
USGS United States Geological Survey
LJNCED ,, United Nations Conference on Environment an De teloprnent
202
------- |