United States Office of Research and Development EPA 3/R-95/044
Environmental Protection Environmental Research Laboratory January 1
Agency Gorvallis OR 97333
&EPA The Contribution of Forest
Land Use To Total National
Carbon Flux: Case Studies
in the Former Soviet Union,
United States, Mexico, and
Brazil
-------
The Contribution of Forest Land
Use to Total National Carbon
Flux: Case Studies in the Former
Soviet Union, United States,
Mexico and Brazil
Edited by
Michael A. Cairns
U.S. Environmental Protection Agency
Environmental Research Laboratory—Corvallis
Tatyana P. Kolohugina
Department of Civil Engineering
Oregon State University
David P. Turner
ManTech Environmental Research Services Corporation
Environmental Research Laboratory—Corvallis
Jack K. Winjum
National Council for Air and Stream Improvement
Environmental Research Laboratory—Corvallis
January 1995
EPA/600/R-95/044
U.S. Environmental Protection Agency
Environmental Research Laboratory
200 Southwest 35th Street
Corvallis, Oregon 97333
USA
-------
Key Words
Carbon, Forest, Land use, Former Soviet Union, United States,
Mexico, Brazil
Preferred Citation
Cairns, M.A., T.P. Kolchugina, D.P. Turner, and J.K. Winjum
(Eds.). 1995. The contribution of forest land use to total
national carbon flux: Case studies in the former Soviet
Union, United States, Mexico and Brazil. EPA/600/R-95/044.
U.S. Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR.
Disclaimer
The information in this document has been funded wholly by the
U.S. Environmental Protection Agency. It has been subjected to
the Agency's peer and administrative review, 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.
ii
-------
Table of Contents
Key Words ii
Preferred Citation ii
i •
Disclaimer ii
Table of Contents iii
Figures viii
Tables ix
Boxes xii
Acknowledgments xiii
Terminology, Units and Acronyms xiv
Executive Summary xvi
I. Introduction 1
A. Global Change and the International Response . . . . 1
B. Framework Convention on Climate Change Mandates
for National Greenhouse Gas Emissions Inventories . 2
C. U.S. Government, Environmental Protection Agency
and Environmental Research Laboratory-Corvallis
Roles 3
D. Purpose, Scope and Organization . . , 4
II. Background 6
A. The Importance of Forests in the Global Carbon Cycle 6
B. Approach to Estimating Carbon Flux at the
National Scale 8
1. Critical Processes 8
a. Forest Land Base 9
b. Products and Landfills 12
c. Anthropogenic Emissions 12
iii
-------
2. Special Issues for Carbon Budgets in the
Three Biogeographic Zones 13
a. Boreal Zone 13
b. Temperate Zone 14
c. Tropical Zone 16
3. River Transport of Carbon 19
III. Former Soviet Union Case Study 22
A. Objectives 22
B. Approaches 22
C. Carbon Pools and Net Primary Production 23
1. Methods 24
a. Areal Estimates 24
b. Carbon Pools 26
c. Net Primary Production 32
2. Results and Discussion 34
D. Net Carbon Flux 35
1. Methods 35
a. Non-Forest Ecosystems 39
b. Forest Ecosystems 39
c. Disturbances 43
2. Results and Discussion 47
a. Non-Forest Ecosystems 4 7
b. Forests 47
E. Forest Management Options for Mitigating
Carbon Emissions '56
1. Reforestation 56
2. Increased Fire Control and Prevention .... 56
iv
-------
3. Enhanced Forest Productivity ... 56
4. Improved Disease and Pest Control 57
5. Reconstruction of the Forestry Industrial
Sector > 57
6. Increased Rotation Length . 57
IV. United States Case Study 61
A. Introduction 61
B. Methods 62
1. Forestland 62
a. The Forest Inventory 63
b. Construction of Stand-Level Carbon
Budgets 63
c. Estimation of Carbon Flux 64
d. Biologically-Driven Carbon Flux 64
e. Harvest-Driven Carbon Flux 66
f. Fire Emissions 68
2. Woodlands 68
3. Rangeland, Pastureland, Cropland and
Other Land . . 69
C. Results 70
1. Forestland 75
a. Carbon Pools 75
b. Biologically-Driven Flux 75
c. Harvest-Driven Carbon Flux ....... 76
d. Complete Forestland Flux Analysis .... 76
2. Woodland 78
3. Grassland, Rangeland, Pastureland, Cropland
and Other Land 78
v
-------
D. , Discussion 79
V. Mexico Case Study 85
A. Introduction . . 85
B. Methods 85
1. Land Cover 87
2. Carbon Density Estimates 92
3. Transition Rates 98
4. Carbon Flux Estimates 98
C.. Results 101
D. Discussion 106
E. Conclusions 110
VI. Brazil Case Study 116
A. Introduction 116
B. Methods 119
1. Conceptual Model 119
2. Vegetation and Land-Use Types 119
3. Brazil's Carbon Pools 124
4. Brazil's Biotic Carbon Flux 127
C. Results 136
1. Carbon Pools 136
2. Carbon Flux 137
D. Discussion 139
1. Comparisons to Other Studies 144
2. Base Map Used for Study 145
3. The Role of Secondary Forests:
A Sensitivity Analysis 147
vi
-------
4. Changes in Combustion Efficiency 148
5. Wood Products and Fossil Fuels 150
6. Unquantified Flux Processes 153
7. Forest Management for C02 Mitigation 155
VII. Discussion: Comparison of the C Budgets for Four
Nations 159
A. Carbon Pools 162
B. Carbon Flux 165
C. Policy Perspective and Prospects for
Stabilization of Emissions . 171
VIII. Uncertainties and Research Needs 175
A. Global 175
B. Boreal Zone 178
C. Temperate Zone 178
D. Tropical Zone 179
E. Carbon Dynamics Through Modeled Projections . . . 181
1. Individual Tree Growth Models 181
2. Forest Gap Models . 182
3. Biogeographic Models 182
4. Ecosystem Process Models 183
5. Modeling Terrestrial Carbon Dynamics 184
6. Land-Use Change Models 185
7. Spreadsheet Modeling of Carbon Dynamics . . . 185
8. Modeling of Carbon Dynamics: In Conclusion . . 186
IX. Literature Cited 187
vii
-------
Figures
Figure III.l USSR vegetation/soil map 25
Figure IV.1 United States land-cover map 60
Figure IV.2 Pacific Northwest Region standard carbon
budget for medium productivity of Douglas-fir 65
Figure V.l Forest map of Mexico 84
Figure VI. 1 Vegetation map of Brazil 117
Figure VI.2 Conceptual model of biotic net carbon
balance 120
Figure VI.3 Annual net arbon balance for Brazil using
the Olson (1983, 1985) carbon density values . 140
Figure VI.4 Annual net carbon balance for Brazil using
the Fearnside (1992) carbon density values . .141
Figure VI.5 Annual net carbon balance for Brazil
during 1990 using Brown and Lugo (1992)
carbon density values 142
viii
-------
Tables
Table II. 1 Carbon Fluxes and Assumed Pools 10
Table III.l Areas of Major Ecosystems and Land-Use
Types of the FSU 27
Table III.2 Densities of Phytomass, FSU 28
Table III.3 Densities of Coarse Woody Debris (CWD), FSU . 29
Table III.4 Accumulation of Phytomass and Coarse Woody
Debris in FSU Forests 31
Table III.5 Densities of Litter, Soil Organic Matter
(SOM), and Net Primary Productivity (NPP), FSU 33
Table III.6 Phytomass, FSU 36
Table III.7 Coarse Woody Debris, FSU 37
Table III.8 Litter, Soil Organic Matter (SOM), and Net
Primary Productivity (NPP), FSU 38
Table III.9 Net Accumulation of Phytomass (NAPh), Coarse
Woody Debris (NACWD), and Soil Organic
Matter Accumulation (NASOM), FSU 49
Table III.10 Densities of Net Accumulation in Phytomass
(NAPh), Coarse Woody Debris (NACWD), and Soil
Organic Matter Accumulation (NASOM), FSU ... 50
Table III.11 Summary: Carbon Pools, Fluxes, Net
Sequestration, and Emissions, FSU 54
Table IV.1 Area and Carbon Density by Land-Use Category,
United States 71
Table IV.2 Total Carbon Pools for the Conterminous
United States 72
Table IV.3 Total Net Primary Production for the
United States 73
Table IV.4 Carbon Balance on the Forest Land Base,
United States 74
Table IV.5 Net Carbon Flux for the Conterminous
United States 80
Table V.1 Areas of Major Land-Cover Types in Mexico . . 86
ix
-------
Table V.2 Estimates of Land Cover and Annual
Deforestation Rates in Mexico 88
Table V.3 Phytomass Carbon Density (above and below-
ground) in Mexico by Land-Cover Type 93
Table V.4 Total Area, Flux Parameters and Estimates
Used to Compute Net Carbon Balance for Mexico 94
Table V.5 Carbon in Mexican Soils by Land-Cover
Type 99
Table V.6 Carbon Pools of Mexico by Land-Cover Type,
1990 102
Table V.7 Area-Weighted Mean Carbon Densities by Major
Land-Cover Type in Mexico 104
Table V.8 Carbon Flux for Mexico 105
Table VI.1 Vegetation and Land-Use Types, Brazil, 1994 . 118
Table VI.2 Carbon Density in Phytomass by Vegetation
and Land-Use Type, Brazil 12
Table VI.3 Root/Shoot Ratios and Factors Used in the
Brazilian Study 12 9
Table VI.4 Carbon Density in Phytomass Components
and Soil by Vegetation and Land-Use Type,
Brazil 130
Table VI.5 Total Area, Area Cleared, Flux Parameters
per Hectare, and Estimates Used to Compute
Net Carbon Balance, Brazil 131
Table VI.6 Total Carbon Pools in Brazil by Vegetation and
Land-Use Types Calculated with a Variety of
Carbon Density Values 133
Table VI.7 Net Carbon Flux by Vegetation and Land-Use
Types, Brazil 139
Table VI.8 Sensitivity Analysis of the Effect of the
Area of Secondary Forest on Brazil's Net C
Balance 150
Table VI.9 Total Net Annual Carbon Flux from Brazil,
All Sources 153
Table VII.1 Land Area of Major Vegetation Types by Nation 161
x
-------
Table VII.2 Comparison of Carbon Densities Among Nations 162
Table VII.3 Estimates of Carbon Pools for the Vegetation
Types Common to All Four Nations, 1990 .... 164
Table VII.4 Land Base Carbon Flux for the Four Case
Studies 167
Table VII.5 Net Carbon Flux for the Four Case Studies . . 169
Table VII.6 Greenhouse Warming Potentials . . 173
xi
-------
Boxes
Box 1: Former Soviet Union 22
Box 2: United States 60
Box 3: Mexico 84
Box 4: Brazil 115
xii
-------
Acknowledgments
The editors gratefully acknowledge all persons whose contribu-
tions led to the completion of this report. In particular, we
thank the following individuals:
Writers and Technical Advisors:
J. Barker, ManTech Environmental Research Services Corp. (MERSC),
contributed the discussion of modeling in Section VII. G.
Baumgardner, Computer Services Corporation (CSC), contributed
data base management for Section IV. M. Cairns, U.S. Environmen-
tal Protection Agency (EPA), was primary author of Sections I, II
and VII. T. Kolchugina, Oregon State University (OSU), was
primary author of Section III on the former Soviet Union and
contributed to Sections II and VIII. J. Lee, U.S. Environmental
Protection Agency (EPA), contributed policy perspective to
Section VII. R. Riley, National Council for Air and Stream
Improvement (NCASI), was the primary author of Section V on
Mexico. P. Schroeder, ManTech Environmental Research Services
Corp. (MERSC), contributed to Section VI. D. Turner, ManTech
Environmental Research Services Corp. (MERSC), was the primary
author of Section IV on the conterminous United States, contrib-
uted to Section II and VIII, and provided the comparison of
national carbon fluxes in Section VII. T. Vinson, Oregon State
University (OSU), contributed to Section III. J. Winjum,
National Council for Air and Stream Improvement (NCASI), was the
primary author of Section VI on Brazil, and Section VII and
contributed to Sections II and VIII.
Document Production:
C. Chapman, ManTech Environmental Research Services Corp. (MERSC)
K. Gundersen, Senior Environmental Employment, Program (SEEP)
M. Schuft, ManTech Environmental Research Services Corp. (MERSC)
S. Volk, Senior Environmental Employment Program (SEEP)
Reviewers t
Jnterr^;
S. Brown, U.S. Environmental Protection Agency (EPA)
J. Kern, ManTech Environmental Research Services Corp. (MERSC)
G. King, ManTech Environmental Research Services Corp. (MERSC)
External
L. Heath, U.S. Department of Agriculture, Forest Service
W. Makundi, University of California
0. Masera, Universidad Nacional Autonoma de Mexico
R. Meganck, United Nations Environment Programme
L. Molion, Universidade Federal de Alagoas
R. Victoria, universidade de Sao Paulo
D. von Hippel, Eugene OR, Associate of Stockholm Environment
Institute - Boston Center
xiii
-------
Terminology, Units and Acronyms
In this document:
Phytomass refers to live plant material, both above- and
belowground (i.e., live roots).
Total ecosystem carbon (TEC) in is the sum of the carbon in
Mg on a site which is contained in the phytomass, the dead
organic matter (i.e., litter as well as the coarse woody
debris), and the soil organic C.
Pools and flux of biomass or carbon dioxide are expressed
as carbon (C).
Carbon flux values in the text, tables, and figures are
designated with a plus sign (+) for C sequestration or
uptake and a minus sign (-) for emissions from the bio-
sphere.
Units of measure are metric by The International System of
Units (SI) or acceptable equivalents. Frequently used
units in the report and equivalents are:
Pg (petagram) = 1015 g
= Gt (gigaton)
= billion metric tons
Tg (teragram) = 1012 g
= Mt (megaton)
= million metric tons
Mg (megagram) = 106 g
= 103 kg
= one metric ton
ha (hectare) = lO4 m2
= 10~2 km2
Acronyms
CCAP:
Climate Change Action Plan (Clinton and Gore 1993)
CFC:
chloroflourocarbon
CH<:
methane
CNIF:
Camara Nacional de la Industria Forestal, Mexico
C02:
carbon dioxide
CWD:
coarse woody debris
ERL-C:
Environmental Research Laboratory-Corvallis (U.S.
Environmental Protection Agency)
FAO:
Food and Agriculture Organization of the United Nations
FCCC:
Framework Convention on Climate Change
FIA:
Forest Inventory and Analysis
FSU:
former Soviet Union
GCRP:
Global Change Research Program
xiv
-------
GIS: geographic information system
GPP: gross primary productivity
GHG: greenhouse gases
GS: growing stock
IBDF: Instituto Brasileiro de Desenvolvimento Florestal,
Brazil
IBGE: Instituto Brasileiro de Geografia e Estatistica, Brazil
INPE: Instituto Nacional de Pesquisas Espaciais, Brazil
IPCC: Intergovernmental Panel on Climate Change
N20: nitrous oxide
NAPh: net accumulation of phytomass
NACWD: net accumulation of coarse woody debris
NASOM: net accumulation of soil organic matter
NEP: net ecosystem productivity
NIW: net annual accumulation of stem wood
NPP: net primary productivity
NRI: National Resources Inventory
Ra: respiration of autotropic organisms (or autotrophic
respiration)
Rg: heterotrohic respiration
SARH: Secretaria de Agricultura y Recursos Hidraulicos, Mexico
SCS: Soil Conservation Service, U.S. Department of
Agriculture
SOM: soil organic matter
TEC: Total ecosystem carbon
UNAM: Universidad Nacional Autonoma de Mexico
UNCED: United Nations Conference on Environment and Development
UNESCO: United Nations Educational, Scientific, and Cultural
Organization
USDA: U.S. Department of Agriculture
WRI: World Resources Institute
xv
-------
Executive Summary
Various human activities are causing an increase in emissions to
the atmosphere of carbon dioxide C02 and other greenhouse gases,
for example CHA, N20, and CFCs. The annual increase in
atmospheric C02 concentration approximates 1.5 ppm (Houghton et
al. 1990). Although the dynamics and feedbacks are poorly
understood (Houghton et al. 1992; Sundquist 1993), scientific
consensus maintains that the increased atmospheric concentrations
will cause an increase in global mean temperature and alterations
in related components of the Earth's biogeochemical cycles.
Researchers predict the global mean temperature will rise between
0.2 and 0.5cC per decade (Houghton et al. 1990), or from 1.5 to
4.5#C in the first half of the 21st century (Schlessinger and
Mitchell 1985; Hansen et al. 1988; Mitchell 1989; Manabe et al.
1991) .
Fossil fuel burning, the single largest source of C02, is esti-
mated to produce 5.4 ± 0.5 PgC/yr, with an additional net contri-
bution of 0.9 ± 0.4 PgC per year from deforestation, vegetation
growth, and other land-use changes (Dixon et al. 1994). Oceans
absorb 2.0 ± 0.8 PgC/yr and 3.2 ± 0.1 PgC/yr remains in the
atmosphere, leaving unaccounted approximately 1.1 ± 1.0 Pg/yr
(Dixon et al. 1994). It is theorized that the world's forests
are the "missing C sink" (Tans et al. 1990; Lugo and Brown 1992;
Taylor and Lloyd 1992).
The United Nations Conference on Environment and Development
(UNCED), held in Rio de Janeiro in June 1992, agreed to a Frame-
work Convention on Climate Change (FCCC). The objective of the
FCCC is to
.achieve...stabilization of greenhouse gas concentra-
tions in the atmosphere at a level that would prevent
dangerous anthropogenic interference with the climate
xv i
-------
system. Such a level should be achieved within a time
frame sufficient to allow ecosystems to adapt naturally to
climate change, to ensure that food production Is not
threatened, and to enable economic development to proceed
in a sustainable manner." (UNCED 1992)
The Convention, which went into force in March, 1994, requires
that each nation prepare an inventory of greenhouse gases.
This project assessed forest ecosystem land-use and land-cover
change and their impacts on C dynamics. At the national level,
the former Soviet Union (FSU), the United States (48 conterminous
states), Mexico and Brazil were examined in detail to estimate C
pools and flux rates. Carbon storage and emissions of the four
nations were compared to identify trends and patterns in relation
to biogeography and level of economic activity. The prominent
role of global forests in the total terrestrial C pool and the
large annual quantities of C moving through global biogeochemical
cycles are reported. The report describes approaches to
estimating C dynamics at the national scale.
The biogenic C cycle consists of a combination of pools and
fluxes. Pools are C stores in soil and vegetation, including
living vegetation (i.e., phytomass), coarse woody debris (CWD,
aboveground and belowground), soil, and litter. Processes
associated with formation of new organic matter in soil and
vegetation (i.e., humification and NPP) represent C influxes.
Efflux is associated with C emissions resulting from plant
respiration and decomposition of organic matter. Combustion of
organic matter associated with wildfires, slash burning, and peat
utilization represents the other major biogenic C efflux.
Forest ecosystem land-use change is estimated to contribute
annual emissions totaling from 0.5 to 1.3 PgC (Dixon et al.
1994). The four case study countries in this report accounted
for approximately 46% of the world's total forest area (Makundi
xv ii
-------
et al. 1992; FAO 1993; Kolchugina and Vinson 1993a; Turner et al.
1993a). Regrowth of formerly degraded or disturbed forests may
largely balance the flux associated with land-use change (Kauppi
et al. 1992). In addition, the assumption that forests that have
not been cleared within the past 100 years or so are in C equi-
librium (Houghton et al. 1983) is not currently supported (Lugo
and Brown 1992).
This assessment of net C exchange between the atmosphere and the
biosphere employed examination of biological processes in forest
ecosystems, harvest of trees for wood and paper products, and
direct C emission from fires. Critical processes considered
include those associated with the forest land base, forest
products and landfills, and anthropogenic emissions. Quantified
land-base processes were changes in the intensity and extent of
forest use, such as growth, mortality, decomposition, harvest and
fire. Outside the forests, agricultural practices, peat accumu-
lation, cement production, landfills, and fossil fuel production
and use were also considered.
Methods of modeling biogenic C dynamics varied in this study
because of inherent differences among boreal, temperate and
tropical biogeographic zones. Examples of such differences
include peat accumulation (boreal), increasingly intensified
forest management (temperate), and paucity of forest inventory
information (tropical). Assessments varied among countries,
because of both quality and quantity of data available in the
following order: U.S. > FSU > Brazil > Mexico.
Land area of the FSU, U.S., Mexico, and Brazil total 3.92 X 109
ha, about 30% of the Earth's total land area. These countries
contained 35% of the world's closed forests, 26% of the wood-
lands, and 38% of the total croplands. Carbon densities in these
common vegetation types ranged from 184 to 271 Mg/ha in closed
xviii
-------
forests, from 88 to 241 Mg/ha in woodlands, and from 80 to 156
Mg/ha in croplands.
The terrestrial ecosystems of the four countries contained
approximately 866 PgC, or about 44-52% of the world's total, with
the FSU alone having one-third of the global total. Of this
amount, forest C pools of the four countries were as follows (in
PgC): 138 (FSU), 109 (Brazil), 37 (U.S.), and 14 (Mexico).
Average forest phytomass C densities decreased from south to
north as follows (in MgC/ha): 152 (Brazil), 64 (Mexico), 63
(U.S.), and 70 (FSU). While densities were much lower in
woodlands and croplands, the increasing north-to-south trends
followed for phytomass C densities. Average litter and CWD C
densities of forest/woodlands ranged from highs of 30 Mg/ha in
U.S. forests and 16 Mg/ha in FSU woodlands to lows of 7 Mg/ha in
Mexican forests and 2 Mg/ha in both Mexican and Brazilian wood-
lands. Average soil C densities in all three vegetation types
generally increased from south to north, ranging from 123 Mg/ha
in FSU woodlands to 60 Mg/ha in Brazilian woodlands.
For the C flux comparison, land-base C flux was separated from
the set of all C flux terms. The land-base flux was divided into
a net biological term and a land-use/land-cover change term. The
net biological flux refers to the expected change in C storage on
the land base if no further anthropogenic impacts were imposed.
Thus it included "inherited" emissions from previous disturbanc-
es. The components of the net biological flux were the net
changes in the pools of phytomass, CWD, and soil organic matter.
The net change in the phytomass pool-a balance of woody phytomass
growth and mortality—was positive in all four countries and
relatively high in the FSU, the U.S. and Brazil because large
areas of young forest stands promoted C accumulation. The total
net biological flux (in Tg/yr) was positive in the FSU (631) and
the U.S. (332), but negative in Mexico (37) and Brazil (16).
xix
-------
Besides net biological flux, the other main land-base flux was
associated with anthropogenic land uses. Components included (1)
removal of wood during commercial harvest, (2) direct emissions
from burning (deforestation in the tropics and timber harvest
slash burns in temperate and boreal countries), (3) soil c
emissions from tillage of agricultural soils, and (4) peat
combustion. Total land-use flux was negative (i.e., were emis-
sions) in all four countries (in TgC/yr): FSU (-342), U.S.
(-243), Mexico (-35), and Brazil (-235).
The total net effect of the biological and anthropogenic factors
on the land base was a C sink in the FSU and the U.S. and a C
source in both Brazil and Mexico.
In the complete country level C budgets, four additional flux
components were included. The accumulation of C in forest
products still in use ranged from 36 and 33 Tg/yr in the U.S. and
FSU, respectively, to 3 and 1 Tg/yr in Brazil and Mexico, respec-
tively. Accumulation in landfills was greatest in the U.S. and
FSU (10 TgC/yr), but this was a relatively small flux term in all
four budgets. Emissions from energy production and use was the
dominant term in the U.S. (1296 TgC/yr), FSU (1020 TgC/yr), and
Mexico (74 TgC/yr) budgets, whereas in Brazil the net land-base
flux (-251 TgC/yr) was dominant.
In all cases, C sources exceeded sinks, i.e. all countries were
net emitters of C. The U.S. is the largest emitter because of
its high per capita industrial emissions and large population.
The FSU had a similar distribution of sources and sinks and, like
the U.S., was a large.net emitter. The estimated land-base sink
in the FSU accounted for a much larger proportion (41%) of the
total C source than was the case in the U.S. (8%). Brazil was
notable because the land-base source was larger than the
industrial source. In Mexico, industrial emissions were of about
the same magnitude as the land-base source. The summed net
-------
emissions from these four countries (2347 TgC/yr) represented
about 36% of the estimated global net C02-C emission.
To improve assessments of C flux due to land-use and land-cover
change, data are required for (1) classifying land cover and its
relative change over time due to land-use change, (2) quantifying
C mass per unit area—C densities—of the various land-cover
classes, and (3) quantifying C flux per unit area within the
specific land-cover types. Research needs specific to the three
biogeographic zones include C accumulation in CWD/soils/peat and
frequency/extent/intensity of fires (boreal zone); improved
forest inventories and soil/CWD C dynamics following harvest
(temperate zone); and secondary forest regrowth, possible mature
forest c accretion vs. equilibrium, and lotic c export (tropical
zone).
xxi
-------
I• Introduction
A. Global Change and the International Response
Human activities are causing an increase in emissions to the
atmosphere of C02 and other greenhouse gases, for example CH4,
N20, and CFCs. The annual increase in atmospheric C02 concentra-
tion approximates 1.5 ppm (Houghton et al. 1990). Scientific
consensus maintains that the increased atmospheric concentrations
may cause an increase in global mean temperature and alterations
in related components of the Earth's biogeochemical cycles. The
dynamics and feedbacks are, however, poorly understood (Houghton
et al. 1992; Sundquist 1993). Current predictions indicate that
the global mean temperature will rise between 0.2 and 0.5°C per
decade (Houghton et al. 1990), or 1.5 to 4.5°C in the first half
of the 21st century (Schlessinger and Mitchell 1985; Hansen et
al. 1988; Mitchell 1989; Manabe et al. 1991; Houghton et al.
1992).
Fossil fuel burning, the single largest source of C02, is esti-
mated to produce 5.4 i 0.5 Pg carbon (C) per year, with an
additional net contribution of 0.9 ± 0.4 PgC/yr from deforesta-
tion, vegetation growth, and other land-use changes (Dixon et al.
1994). Oceans absorb 2.0 ± 0.8 PgC/yr and 3.2 ± 0.1 PgC/yr
remains in the atmosphere, leaving unaccounted approximately 1.1
± 1.0 PgC/yr (Dixon et al. 1994). It is theorized that the
world's forests are repositories of this "missing C sink" (Tans
et al. 1990; Taylor and Lloyd 1992).
The Intergovernmental Panel on Climate Change (IPCC) was estab-
lished in 1988 by the World Meteorological Organization and the
United Nations Environment Program. This panel has sought to
draw together scientific research on global warming, assess
impacts on human and natural systems, and recommend appropriate
1
-------
actions to mitigate or adapt to the impacts. The IPCC began
discussions and analyses that led to an international convention
on climate change in 1992 (Parson et al. 1992).
B. Framework Convention on Climate Change Mandates
National Greenhouse Gas Emissions Inventories
The United Nations Conference on Environment and Development
(UNCED), held in Rio de Janeiro in June 1992, drafted a Framework
Convention on Climate Change (UNFCCC), which was signed by 161
nations. It was ratified by the minimum 50 countries and went
into force on March 21, 1994. As of July 13, 80 countries had
ratified and are now legally bound by the terms of the UNFCCC.
Developed countries were to submit national action plans for
reducing GHG emissions by September 21, 1994. Developing coun-
tries have three additional years. The first session of the
conference of the signatories opens on March 28, 1995 in Berlin.
The objective of the UNFCCC is to
"achieve...stabilization of greenhouse gas concentrations
in the atmosphere at a level that would prevent dangerous
anthropogenic interference with the climate system. Such
a level should be achieved within a time frame sufficient
to allow ecosystems to adapt naturally to climate change,
to ensure that food production is not threatened, and to
enable economic development to proceed in a sustainable
manner." (UNCED 1992)
The Convention requires each nation to prepare an inventory of
anthropogenic greenhouse gases. Such inventories will enable
nations to (1) set priorities for controlling emissions from
different sources and economic sectors and (2) verify progress
toward achieving emission reduction targets.
2
-------
The United States recently published the national response
strategy to reduce emissions to 1990 levels by 2000 (Clinton and
Gore 1993). The U.S. Climate Change Action Plan (CCAP) commits
the U.S. to a target of 1,500 Tg C02-C or equivalents, requiring
a net annual reduction in projected emissions of 106 TgC by 2000.
Implementation of the CCAP calls for establishment of
public-private partnerships with key industries to reduce green-
house gas emissions across many sectors of the U.S. economy. The
CCAP lists 50 separate actions selected for effectiveness, rapid
implementation, and low cost. Included in the plan are three
actions relating specifically to the forest sector, designed to
provide a net reduction of 9.5 TgC annually or approximately 9%
of the total specified emission reductions. The forest sector
actions are
• Accelerated source reduction, pollution prevention, and
recycling
• Assistance for improved timber management on nonindustrial
private forest lands
• Accelerated tree planting on nonindustrial private forest
lands
C. U.S. Government, Environmental Protection
Agency and Environmental Research
Laboratory-Corvallis Roles
The U.S. government, through its interagency Committee of the
Environment and Natural Resources of the National Science and
Technology Council, developed the U.S. Global Change Research
Program (USGCRP). The program, designated as a National Research
Program, aims to provide a sustained, long-term effort for global
change research within the federal government. The USGCRP seeks
to produce a predictive understanding of the earth system and its
biogeochemical cycles to support national and international
3
-------
policy decisions across a range of environmental issues. Four
working groups of the USGCRP address these broad activities:
Observations and data management
Process research
• Integrated modeling and prediction
Assessments
Research at the U.S. EPA Environmental Research Laboratory -
Corvallis (ERL-C) is conducted in each activity group, within the
EPA's GCRP. Process research at ERL-C is described in the Global
Processes and Effects Program (1993); integrated modeling and
prediction research is described in U.S. EPA (1993). The current
report was an assessment-oriented research product. Other recent
ERL-C GCRP assessment projects are described in Dixon et al.
(1991) and Turner et al. (1993a).
D. Purpose, Scope and Organization
This study assessed forest ecosystem land-use change and its
impact on C dynamics. At the national level, the largely boreal
zone FSU, the temperate zone U.S. (conterminous states), and the
tropical zone nations of Mexico and Brazil were examined in
detail to estimate C pools and flux. Carbon storage and emis-
sions of these four countries were compared to identify trends
and patterns in relation to biogeography and level of economic
activity. The report discusses the prominent role of global
forests in the total terrestrial C pool and the large annual
quantities of C moving through global biogeochemical cycles.
The report also describes approaches to estimate C dynamics at
the national scale by quantifying processes including changes in
the intensity and extent of forest use, including growth, mortal-
ity, decomposition, harvest and fire. Two additional processes,
4
-------
while not primary foci of this report—the fate of harvested
forest products and anthropogenic C emissions from other sources-
are discussed in the context of national and global C flux. A
review of C budget models puts the modeling approach of this
report in context.
For each country, background information supplies the context for
current C pools of the various land-cover/land-use types, current
overall C flux, including fossil fuel emissions, and possible
forest management options for conserving terrestrial C and
sequestering atmospheric C.
5
-------
II. Background
The C cycle consists of a combination of pools and fluxes. The
pools are C stored in soil and vegetation, including living
vegetation (i.e., phytomass), coarse woody debris (CWD, above-
ground and belowground), soil, and litter (Kolchugina and Vinson
1991; Vinson and Kolchugina 1993). The term litter includes fine
woody debris and leaves that are not completely decomposed. The
processes associated with formation of new organic matter in soil
and vegetation (i.e. humification and NPP) represent C influxes.
Efflux is associated with C emissions resulting from plant
respiration and decomposition of organic matter. The NPP equals
the difference between GPP and RA. The RA amounts to 44% to 52%
(48% on average) (Kobak 1988) of GPP. NEP is NPP minus hetero-
trophic respiration; it can be measured as the net increase or
decrease of C in the soil, phytomass, CWD, and litter after C is
expended for autotrophic and heterotrophic respiration.
The NEP may be positive, zero or negative at the beginning of
secondary succession depending on the combination of NPP soil
respiration and previous land use. The NEP increases with time
reaching a maximum at some intermediate age, depending on forest
type, e.g., 20 to 80 years for Russian forests (Vorobyov 1986;
Kolchugina and Vinson 1993d). In a climax ecosystem, it is
hypothesized that the rate of autotrophic and heterotrophic
respiration is high, and NEP may approach zero.
A. The Importance of Forests in
the Global Carbon Cycle
Land-use change currently contributes an amount of C equivalent
to 17% of the net emissions of fossil fuel burning (Dixon et al.
1994). This proportion represents a decrease from 30% in the
1980s (Houghton et al. 1992; Sarmiento and Sundquist 1992).
6
-------
Because their total ecosystem carbon (TEC) is large, the world's
forests are the primary ecosystems whose alteration contributes
to these land-use based C emissions (Solomon et al. 1993).
The main pools of the global carbon cycle are oceans, containing
approximately 38,000 PgC, followed by fossil fuel deposits (4,000
PgC), litter and soil organic matter (1,100-1,600 PgC), and the
atmosphere (750 PgC) (Post et al. 1990; Smith et al. 1993;
Sundquist 1993). Global vegetation contains 420-830 PgC, the
range in values indicating the uncertainty in ecosystem
classification, as well as area and carbon density of each
vegetation class (Post et al. 1990). However, plants exert a
large influence on C dynamics due to the processes of photosyn-
thesis, respiration and decomposition (Simpson and Botkin 1992;
Woodwell 1992). It is estimated that the entire atmospheric C02
content cycles through the terrestrial biosphere every seven
years, with approximately 70% of that C exchange passing through
the Earth's forest ecosystems (Waring and Schlesinger 1985).
Forests occupy approximately 4.1 X 109 ha, or 31% of the Earth's
surface (Allan and Lanly 1991; WRI 1992). Of this total, the
boreal biome currently contains 33%, temperate regions 25%, and
the tropics 42% of the world's forest area (Dixon et al. 1994).
The nations of the FSU, U.S., Mexico and Brazil account for
approximately 46% of the total forest area (Makundi et al. 1992;
FAO 1993; Kolchugina and Vinson 1993a; Turner et al. 1993a).
Forests of the world store 40% of the total belowground and
nearly 80% of the total aboveground terrestrial C (Olson et al.
1983; Schlesinger 1984; Zinke et al. 1984; Waring and Schlesinger
1985; Allan and Lanly 1991; Dixon and Turner 1991; WRI 1992;
Schindler and Bayley 1993). Forests currently contain 75% of the
world's living C (Houghton et al. 1993).
Employing a 1980 base year, forest ecosystem land-use change has
been estimated to contribute annual emissions totaling from 0.4
7
-------
to 2.5 PgC (Houghton et al. 1983, 1987; Detwiler and Hall 1988;
Houghton and Skole 1990; Flint and Richards 1994). A more recent
estimate has reduced this uncertainty to a global net flux of 0.5
to 1.3 PgC (Dixon et al. 1994). Regrowth of formerly degraded or
disturbed forests may largely balance the flux associated with
land-use change (Kauppi et al. 1992). In addition, the assump-
tion that forests that have not been cleared within the past
century are in C equilibrium (Houghton et al. 1983) is not
currently supported (Lugo and Brown 1992).
B. Approach to Estimating Carbon Flux
at the National Scale
An assessment of net C exchange between the atmosphere and the
biosphere must involve examination of biological processes in
forest ecosystems, conversion of forests to alternate land
covers, harvest of trees for wood and paper products, and direct
C emission from fires. Outside the forests, agricultural prac-
tices (particularly tillage intensity and emissions from rice and
livestock production), peat accumulation (in boreal zones),
cement production, landfills, and fossil fuel production and use
must also be considered. Although this report focused on land-
use, other greenhouse gas sources and sinks are also discussed.
1. Critical Processes
Flux is the rate of movement of C from one pool to another. The
principal forest vegetation C pools include both live and dead
matter. Live aboveground pools include stemwood, bark, branches,
foliage, shrubs and herbs; the primary live belowground pool is
root phytomass. The aboveground dead C pools are standing and
fallen logs, and litter; the primary dead belowground C pools are
soil and dead roots. Other C pools (e.g., animals) exist
although they are usually ignored in C budgets because they are a
8
-------
relatively small portion of the ecosystem totals. Carbon pools
are frequently aggregated due to the difficulty of measuring
individual pools or the small contribution of particular pools to
the total ecosystem C content.
Biological processes that result in changes in forest C pools
include primary production, litterfall, mortality, grazing, and
autotrophic (RA) and heterotrophic (Rg) respiration. Ideally C
movement between all pools would be measured (Table II.1).
Because some processes are relatively unimportant or very
difficult to measure, net changes are usually quantified rather
than the processes themselves. The flux components typically
quantified are net tree growth, net change in litter and
understory, net change in dead vegetation components (input from
mortality and output from decomposition), and net change in soil
C (including peat and agricultural tillage) over a specified time
period (e.g., one year). Net flux in an ecosystem can be
represented as
NEP = NPP - (heterotrophic respiration - loss from leaching),
where
NPP = (change in live above- and belowground phytomass) +
(litterfall + wood mortality + grazing + root mortality).
a. Forest Land Base
Natural and human-induced mortality of trees causes succession,
which influences C flux at various stages. Knowledge of the
forest age structure is valuable in assessing C flux in forested
nations although age structure is not readily determined in the
tropics. NEP depends upon age structure of forest ecosystems,
approaching zero near the start of successional processes,
gradually increasing and reaching its maximum near the mid-point
of succession, a time dependent upon the particular biogeographic
zone. In most climax ecosystems, NEP again approaches zero
9
-------
Table II.1. Carbon Fluxes and Assumed Pools*
Flux
Litterfall
Conifer needles
Hardwood leaves
Woody
Reprod. parts
Other
Herb turnover
Tree mortality
Grazing
Autotrophic respiration
Aboveground
Belowground
Heterotrophic respiration
O horizon
Standing & fallen logs
Mineral soil
Gross primary production
Aboveground
Belowground
Net primary production
Aboveground
Belowground
Leaching loss
Net ecosystem production
Aboveground
Belowground
Donor pool
Tree foliage
Tree and shrub foliage
Tree wood parts, except
stemwood
no pool
no pool (epiphytes, grass,
insects and unidentifiable)
Herb
Tree woody parts
Coarse roots
Foliage from all vegetation
strata
Aboveground tree components
Coarse & fine roots
O horizon
Standing & fallen logs
Mineral soil
Atmosphere
Atmosphere
Atmosphere
Atmosphere
no pool
Atmosphere
Atmosphere
Receptor pool
Litter
Litter
Litter
Litter
Litter
Litter
Standing dead
Atmosphere
Atmosphere
Atmosphere
Atmosphere
Atmosphere
Atmosphere
Aboveground vegetation
Belowground vegetation
Aboveground vegetation
Belowground vegetation
Out of ecosystem
Aboveground vegetation
Belowground vegetation
Modified from Grier and Logan (1977).
10
-------
because of the high rate of autotrophic and heterotrophic
respiration.
Carbon is removed from forest ecosystems when vegetation is
removed for wood and paper products, for use as fuelwood, for
charcoal, and for non-wood forest products (FAO 1993). National
imports and exports of both long-lived wood products (e.g.,
structural lumber and furniture) and short-lived products (e.g.,
paper and paper board) are economic processes involving C from
forests. These processes need to be factored into national C
budgets. Current roundwood production—including industrial
roundwood, fuelwood and charcoal—in the four nations that are
the focus of this report equals 1.139 X 109 m3, or 33% of the
world's total. Fuelwood and charcoal consumption in these coun-
tries totals 0.373 X 109 m3, or 20% of the world's total (FAO
1993). It should be noted, however, that international negotia-
tions have yet to reach agreement on credit and responsibility
for C emissions due to import and export of forest products.
Carbon is lost to the atmosphere when phytomass is removed from
forests as products and from harvest-associated processes,
including slash burning of understory as well as increased
formation of litter and coarse woody debris that subsequently
undergo oxidation to C02. In addition to current direct impacts
on C flux in forest systems, harvest may cause indirect, future
effects on primary production and heterotrophic respiration that
impact C flux.
Direct C02 emissions occur during wildfire, forest management
practices (e.g., slash burning and prescribed fire suppression
bums), agricultural burning (e.g., crop residues and grass-
lands) , and shifting cultivation burning. The largest source of
C emissions from the Earth's forest land base, however, is due to
conversion of forest to pasture or croplands, when trees are
cleared and burned to create bare land or grasslands. These
11
-------
sources have proven to be difficult to isolate and quantify
(Levine 1991).
b. Products and Landfills
Wood and paper products have particular half-lives. Structural
and furniture wood may have turnover times of many decades,
whereas paper products have lives ranging from weeks or months to
a few years. In order to assess flux from product decay in any
base year, products from trees harvested both currently and in
the past should be counted. A detailed C accounting is typically
used to quantify fluxes from all wood products and landfills.
Landfills are repositories for forest products and act as both
sinks and sources of greenhouse gases (Subak et al. 1993). Wood
and paper products deposited in landfills must be accounted.
Models continue to be developed (e.g., Row and Phelps 1990) to
account for transport of forest products C to landfills and for
the subsequent release of C from landfills. These models must
account for both current release of C from past wood product
landfilling and future C release from current product land-
filling.
c. Anthropogenic Emissions
Carbon emissions from anthropogenic sources, principally fossil
fuel production and combustion and cement manufacture, must be
parts of any national C budget. Because these statistics are
more regularly maintained, the accounting is less difficult than
for processes associated with land-use change. In 1988, the four
countries of this report emitted 2.440 PgC, or 44% of the world's
total emissions from the energy sector. During the same year,
the four nations emitted 35.08 MgC, or 23% of the world's total C
from cement production (Subak et al. 1993).
12
-------
Another form of C emissions, CHA results from energy production
and use, landfills, phytomass burning, decomposition following
timber harvest, livestock production and rice cultivation.
Carbon monoxide, released during phytomass burning and fossil
fuel production and use, is likewise a C source, exerting its
influence by being oxidized to C02 in the atmosphere (Houghton et
al. 1992). Other greenhouse gases include N20, produced by
biotic processes and accelerated by fertilizer use and the
numerous halocarbons used in various industrial and household
applications.
2. Special Issues for Carbon Budgets in
the Three Biogeographic Zones
Certain characteristics typify each biogeographic zone and
influence C budgets. Temperature, climate, topography, soils and
decomposition rates are among the characteristics. The zones
represented in this study were boreal, temperate and tropical.
a. Boreal Zone
The FSU contains boreal ecosystems characterized by a significant
accumulation of CWD and litter due to relatively cold tempera-
tures and low decomposition rates compared to temperate and
tropical ecosystems. The CWD and litter are thus important
components of the C budget of boreal forests. The CWD must,
therefore, be explicitly accounted for in the NEP estimates.
Peatlands, which are wetlands where peat is accumulating, store
significant amounts of carbon. The organic soil carbon content
of boreal peatlands may reach 2,000 MgC/ha (Bohn 1982). Wetlands
are a source of CH4 released to the atmosphere. Although the
atmospheric concentration of CH4 is much lower than the
13
-------
concentration of C02, CH4 is 20 times more effective (per
molecule) than C02 as a greenhouse gas (Blake and Rowland 1988) .
Development of boreal forests is influenced by disturbances, such
as fires, logging and insect attacks. Disturbances may partly or
completely kill the vegetative component of the ecosystem.
Climatic constraints e.g., low temperatures and the presence of
permafrost, do not allow rapid regeneration, consequently,
definite age structure characterizes boreal forests; a signifi-
cant presence of young stands is a result of the influence of
disturbances and slow regeneration. Knowledge of the forest age
structure is important in the assessment of the C flux within
forest ecosystem regions.
b. Temperate Zone
In the absence of human disturbance, deciduous forests of the
temperate zone tend to be in C equilibrium at the stand level;
gaps created by large tree falls maintain a shifting mosaic
steady state (Bormann and Likens 1979; Runkle 1985). Patches
dominated by young trees where carbon is accumulating alternate
with patches where woody debris decomposition creates a carbon
source. Temperate zone coniferous forests may likewise be in C
equilibrium although a somewhat larger spatial context must be
considered; large disturbance events such as catastrophic fires
tend to create a shifting mosaic of stands at different ages
(Franklin and Hemstrom 1981). In recent history, the natural
disturbance regimes have been augmented by disturbances associ-
ated with human activities. At its extreme, this human activity
has meant the loss of large areas of forests due to land-use
change or degradation with an associated large release of C. In
the conterminous U.S., for example, current forest cover is about
55% of potential forest cover (Turner et al. 1993b).
14
-------
Anthropogenic influences have also meant an altered species
composition and age class distribution.
Rates of deforestation in developed temperate zone countries have
tended to decrease in recent decades (e.g., Alig et al. 1990) and
in some cases significant reforestation has occurred (Kauppi et
al. 1992). At the same time, the age-class distribution has been
shifted towards relatively young stands with high rates of C
uptake. Thus, although harvest rates have tended to increase
over time, the forest land-base in many cases is a C sink (Sedjo
1992). The forests have, in general, come under intensified
management. Several aspects of intensified forest management are
essential to consider in a C budget.
First is the C transfer at the time of harvest. Only about half
of the C in trees that are harvested is removed from the land-
base (Harmon et al. 1990). The remaining c is either quickly
returned to the atmosphere via slash burns or slowly returned via
decomposition of woody debris.
Second, the disposition of the C which has been converted to
forest products is important. Estimates of approximately 40%
have been used for the proportion of forest products which go
into long-term storage (Harmon et al. 1990; Birdsey et al. 1993).
In principle, a C budget is needed which accounts for historical
trends in harvest levels, the allocation of the harvest among
product types, and product turnover times.
Third, a long-term effect of extensive harvesting followed by
reforestation is to increase national-level tree growth. Rates
of tree growth in late stand development tend to fall because of
various physiological and nutrient cycling constraints (Sprugel
1985; Ryan and Waring 1992), so a shift to a younger age class
distribution will generally induce greater growth. Note,
however, that national carbon pools may be decreasing even as net
15
-------
annual growth is increasing. A younger age-class distribution
tends to contain less tree C density than an older age-class
distribution in the same area. The input of carbon to the woody
debris pool via natural mortality also tends to decrease as the
age class distribution is shifted downward. In the absence of
management intervention to promote woody debris formation, the
effect of the mortality decrease may be a general decrease in the
pool of woody debris.
In order to account for these various considerations associated
with a harvest-dominated forest land base, the combination of an
inventory and a set of yield tables and associated stand level C
budgets is desirable. Although the degree to which woody debris
and soil C are treated varies, this general approach has been
applied in the U.S. (Plantinga and Birdsey 1993; Turner et al.
1994a), Canada (Kurz et al. 1992), and FSU (Kolchugina and Vinson
1993b).
A related approach to estimating forest C flux retrospectively
employs inventory data from repeated forest surveys, i.e. the
difference in the inventory across two points in time provides
the basis for a change in C storage and hence a C flux. This
approach has only been employed to estimate changes in tree c
storage (Birdsey 1992; Kauppi et al. 1992); changes in forest
floor, understory, woody debris and soil have not been treated.
Other problems with the approach include the difficulty and cost
of performing a complete resurvey of the inventory on a regular
basis as well as possible differences in inventory approaches at
different points in time.
c. Tropical Zone
For 1990, FAO (1993) reports that the tropical forest cover in
the world totaled 1,756 X 106 ha. This forest area is about 37%
16
-------
of all land in the tropics and is close to 45% of the world's
forest cover. Tropical forests are located around the world
roughly between latitudes 23°30'N and 23°30'S and are distributed
among three broad regions (X 106 ha): Africa with 528? Asia &
Pacific with 311; and Latin America & Caribbean with 918 (FA'O
1993).
Tropical forests are contained, for the most part, in 90 nations
which have been classified within the group of 140 nations in the
world with "low" levels of income per capita, i.e., the "develop-
ing" or win-transition" nations (Keating 1993). The 90 nations
contain over 4 billion people (FAO 1993) or approximately 75% of
the world's total population (Sitarz 1993). About 25% of the
people have incomes equivalent to less than U.S.$l/day (Keating
1993). This context affects C flux determinations for tropical
nations in two important ways.
First, most developing nations do not have accurate data on the C
inventory, phytomass growth, or land-use changes (e.g., defores-
tation) in their forested areas. Pressing economic, social, and
political issues have generally precluded intensive surveys of
national forest resources. The 1990 assessment by FAO of the
forest resources of the tropics (FAO 1993) is the most compre-
hensive yet produced. For the first time, FAO used GIS and
remote sensing technology in such an undertaking. Despite this
more technical effort, the 1990 inventory is still an estimate
based on multi-date, sub-national inventory data and models.
Increasing the precision in calculating the C flux for tropical
nations, therefore, is limited until forest inventories become
more widespread and accurate.
Second, in an attempt to improve their national economies and the
conditions of their people, developing nations through history
have been prone to rapid and unplanned exploitation of natural
resources. This often results in increased rates of land-use
17
-------
change. In recent years for forested nations of the tropics,
development has led to high levels of deforestation to provide
land for agricultural crops, pasturelands, dams, roads, settle-
ments, or harvest of tropical hardwood trees. FAO (1993)
estimates that the average annual deforestation rate in the
tropics during the last decade was 15.4 X 106 ha. This defores-
tation rate is significant because, on a per ha basis, phytomass
density levels for moist tropical forests are in the high
category among the world's terrestrial ecosystems. Levels often
approach 200 MgC/ha (Brown and Lugo 1982; Waring and Schlesinger
1985; Fearnside 1992). Thus, within individual tropical nations,
land-use changes like deforestation may be the dominant source of
emissions of C to the atmosphere. For example, in 1990,
emissions from 1.3 X 106 ha of deforestation in Brazil were
estimated to range from 69 to 97 TgC (Schroeder and Winjum 1995a
and 1995b). Emissions from this one source were the largest for
that nation and thus were about three times the 50 TgC from
burning fossil fuels estimated for 1990.
Two other processes potentially important to C flux determina-
tions for tropical forests bear mention: possible C uptake by so-
called "mature" forest and the extent and growth rate of
secondary forests. At present, there is not sufficient data
available to confidently estimate the size of either phenomenon,
but their possible significance warrants a brief description.
Despite deforestation, the tropical zone still has 1756 X 106 ha
of mature forest cover (FAO 1993). Ecologists have often assumed
that so-called "undisturbed" tropical forests not undergoing
land-use changes are in C equilibrium, meaning a steady state
without net C gains or losses. There may be, however, reason to
question the steady state assumption. Lugo and Brown (1992)
argue that the phytomass and phytomass accumulation rates of
tropical forests change over time due to past disturbances by
people and natural catastrophes. Citing plot level field data
18
-------
from Venezuela and Puerto Rico, they show nature forests
accumulating phytomass C at rates of 1-2 MgC/ha/yr. Data do not
exist to quantify such increases worldwide, but Lugo and Brown
(1992) surmise that if the mature tropical forests of the world
in 1980 averaged as little as 0.25 MgC/ha/yr in net uptake by
photosynthesis, they would remove annually about 1 PgC from the
atmosphere (see Section VI.D.6).
Secondary forests regenerating on many of the recently deforested
lands in the tropical zone now appear to sequester significant
amounts of C. In Brazil, a large C flux in secondary forests
results from the combination of phytomass growth and C accumula-
tion in litter-CWD and soil. The current estimate is 260 TgC/yr
based on 5 MgC/ha/yr accumulating in phytomass, litter and CWD,
and soil (an average across all age classes) on 52 X 106 ha
(Schroeder and Winjum 1995b). Generally however, knowledge is
limited about the tropical zone on the extent and phytomass
productivity of secondary forests (Brown and Lugo 1990). Uptake
of C in total for these forests is likely a significant flux
whose determination awaits better data (see Section VI.D.3).
3. River Transport of Carbon
Rivers contain C, which originates via runoff from terrestrial
ecosystems, and transport it to the oceans. Globally, about 55%
of this riverine C was recently in the atmosphere and fixed by
photosynthesis; the remaining 45% is from weathering of rocks
(Meybeck 1993). The global estimate of the C fixed from the
atmosphere and transported by rivers is 540 TgC annually, of
which 46% travels through the rivers of the humid tropics, i.e.,
about 250 TgC/yr (Meybeck 1993). The Amazon River alone
transports an estimated 68 TgC/yr to the ocean (Kempe 1988), and
66% of the Amazon basin lies in one nation, Brazil. River basins
in ecosystems within regions outside the tropics contribute the
19
-------
remaining half of the global total, i.e., temperate forests and
grasslands, 31%; boreal forests, 14%; savannah and sub-arid
regions, 5%; and tundra, 4% (Meybeck 1993).
The fate of the riverine C is not totally understood. A portion
is oxidized by heterotrophic respiration and thus returns to the
atmosphere either directly from the river or estuaries or after
entering the oceans. The C not oxidized goes into long-term
storage in the sediments of the ocean bottom (although worldwide,
rivers transport about 10% of this C into lakes [Meybeck 1993]).
The amount of riverine C deposited in the ocean bottom as a sink
is unknown, and therefore, is not accounted for in the C flux
analyses for terrestrial ecosystems (Meybeck 1993).
If a significant portion of the 540 TgC/yr transported by the
world's rivers is eventually found to be deposited in ocean
sediments, it could explain up to 30% of the so-called "missing
C" currently estimated to be 1800 Tg/yr (Sundquist 1993).
Indeed, over 10% might be accounted for if the same results are
found for just the 250 TgC/yr transported by tropical rivers.
Additionally, it is not clear if disturbances in terrestrial
ecosystems significantly contribute to the transport of riverine
C to ocean sinks, particularly from deforestation in the tropics.
(J. Richey, Univ. Washington, pers. comm., 1993). However, since
river drainage from the tropical ecosystems (largely forested
ecosystems) is the highest among all global zones, estimates of
the ocean sink for this riverine C could prove important for many
national C flux estimates for the tropics (see Section VI.D.6).
Though the estimated levels of river transported C are less for
the temperate and boreal regions as noted above, the caveat
applies there as well until additional data is available (see
Maybeck 1993).
20
-------
Box It FOnaer
Before its dissolution, the former
Soviet Union (PSU) was the largest
country in the world. The FSU
occupied one-sixth of the land
surface of the Earth. The total .
land area of the FSU was 2240 X 106
ha. In 1983, the population of the
FSU was 271.4 million. The former
republics of the Soviet Union in-
cluded Russia. Belorussia, Ukraine,
Moldova. Baltic states (i.e..
Lithuania, Latvia, and Estonia),
Armenia, Georgia, Azerbaijan,
Tajikistan, Turkmenistan,
Uzbekistan, Kazakhstan, and
Kyrgystan. Industrial carbon
emissions iri the FSU were between
0.998 PgC/ha (1988) (Subak et al.
1993) to 1.02 PaC/yr (1990) (Makarov
and Bashmakov 1990).
The territory of the FSU is repre-
sented by a variety of climate
conditions. The major part of the
territory is in the boreal and
temperate biogeographic zones. The
climate in the FSU changes from
arctic and subarctic in the north to
warm temperate and desert in the
south. From west to east, the
climate makes a transition from
maritime to continental to monsoon.
Matthews (1983) identified eight
principal types of vegetation in a
global data base: forest, woodland,
shrubland, grassland, tundra,
desert, peatlands, and cultivated
land. The vegetation of the FSU in-
cludes all of these major types
(Kolchugina and Vinson 1991).
Arctic deserts and tundra formations
are found in the northern regions of
the FSU; deserts and semi-deserts
occur in the southern regions. A
vast area is occupied by forests.
The total area.of land under State
forest management in 1983 was 1,259
X 10° ha (Vorobyov 1985) which is
approximately 56.5% of the total
area of the FSU. The FSU possesses
814 106 ha of forested lana (Alimov
et al. 1989) with most of this land
in closed forest, or 28% of the area
of closed forests of the world. A
quarter of the world's growing stock
of timber can be found in FSU.
Seventy five percent of all FSU
forests is coniferous boreal forests
(i.e., needle-leaf or pine, larch,
fir, spruce); 21% is deciduous soft-
wood forests (e.g., aspen, poplar,
linden, birch, except stony birchj;
the rest of tne area is occupied by
deciduous hardwood species (e.g.,
maple, beech, oak, stony birch).
The terms deciduous softwood and
Soviet Union
deciduous hardwood species are
commonly used in FSU to distinguish
between deciduous species with wood
of different density.
About 95% of the FSU forested area
in the FSU is in Russia. Russia
began to experience a high demand in
timber in tne 17th and loth cen-
turies; from the 1600s to 1814, 70
Mha of forest were cut in European
Russia (Kolchugina and Vinson 1995).
In the 19th century large scale
logging began in Siberia.
Immediately after the revolution in
1917, Russian forests were not under
centralized management. The overall
political and economical instability
of the period resulted in an especi-
ally destructive use of forests.
During the 1930s and 1940s, Siberian
and northern European forests were
extensively logged; these activities
continued through the 1950s and
1960s (Kolchugina and Vinson 1995).
After World War II, reforestation
efforts were initiated throughout
the FSU. Forest inventory data from
1966 to 1988 indicate that the area
of forested land under State forest
management in the FSU has remained
at a stable level. At present, FSU
forests are intensively exploited.
The annual area of principal log-
ging is approximately 2.0 106 ha.
The FSU has the greatest expanse of
peatlands in the world (Tyuremnov
1976), estimated at 164.8 Mha (Botch
et al. 1995; Kolchugina and Vinson
1994a).
A significant area of the FSU is
underlain by continuous permafrost
zone, estimated at 824.2 Mha
i Kolchugina and Vinson 1993b,c).
he continuous permafrost zone is
approximately 4o% of the FSU terri-
tory or 5% or the land surface area
of the world.
Vast areas of the FSU are covered by
grasslands, an important component
of the terrestrial C cycle. Despite
the fact that grasslands do not
accumulate large quantities of plant
mass compared with forest
ecosystems, they exhibit high net
primary, productivity (NPP)
(Bazilevlch 1986) and, therefore,
may influence the terrestrial c
cycle.
21
-------
III. Former Soviet Union Case Study
A. Objectives
The objective of the research presented herein was to provide an
evaluation of the C cycle of terrestrial ecosystems of the FSU in
the mid-late 1980s with special consideration given to forest
ecosystems. The scope of work also included the estimate of C
pools and fluxes of FSU non-forest ecosystems, including
peatlands and agroecosystems. The C pools and fluxes of non-
forest ecosystems were based on thematic maps and statistical
data on peat and agricultural production. Forest statistical
data were used to estimate net ecosystem productivity (NEP) of
forest ecosystems and C flux associated with logging and forest
fires.
B. Approaches
The C budget components of terrestrial ecosystems may be estimat-
ed using different approaches. For example, a number of ecoregi-
ons (i.e. relatively homogeneous regions with respect to phytoma-
ss, CWD, litter, NPP, and SOM densities per ha) may be isolated
with thematic maps. When C values of phytomass, CWD, litter,
NPP, and SOM are assigned to these ecoregions and the ecoregional
estimates are aggregated, a calculation of the C budget compo-
nents may be performed at the regional or national scale
(Kolchugina and Vinson 1991, 1993a; Vinson and Kolchugina 1993).
As another example, countries with significant timber resources
typically have extensive forest statistical databases. Forest
statistical data can be used to estimate the forest component of
a C budget for the country; for instance, it can be related to
phytomass and CWD estimates (Kolchugina and Vinson 1993d, 1993e).
22
-------
These approaches have their advantages and disadvantages. The
approach based upon thematic naps has the advantage of describing
ecosystem types and geographical location of the ecosystem.
Also, FSU scientists at more than a dozen leading institutes have
accumulated data on C in soil and vegetation for several decades
(Kazimirov and Morozova 1973; Aleksandrova 1977; Tytlyanova 1977;
Vatkovskyi 1976; Dylis and Nosova 1977; Kazantseva 1980;
Bazilevich 1986; Bazilevich et al. 1986; and others). These data
combined with the areal extent of specific ecosystems identified
with the help of thematic maps may serve as a basis for carbon
pool and flux estimates. The main disadvantage is that most of
these data do not describe specifics of carbon accumulation
depending upon ecosystem age. Also, a limited number of site
measurements is provided; hence, it is difficult to assess the
spatial variation of carbon accumulation parameters.
An approach using forest statistical data has the advantage of
containing an accurate assessment of commercial timber volume
based on a large number of measurements, but no information is
given with respect to ecosystem components (i.e., branches,
roots, understory, shrubs) other than stem volume. Further,
forest statistical data allow one to incorporate the age-class
distribution of forest stands and estimated NEP. Use of all
available data in a combination approach would undoubtedly result
in the most comprehensive and possibly reliable estimate of the C
budget. A combination approach is a long-term goal for C budget
determination for the FSU.
C. Carbon Pools and Net Primary Production
The assessment of C pools of non-forest, natural (i.e., non-
arable land) ecosystems and NPP of all natural ecosystems was
based on the areal extent and density (per ha) of C pools and NPP
of a specific ecosystem type. The assessment of C pools of
23
-------
forest ecosystems and C pools and NPP of arable land was based on
the statistical data.
1. Methods
Thematic maps were used to identify the area of the major biomes
(complex of ecosystems within specific climatic belt or subbelt)
in the FSU. The area of forest and non-forest ecosystems within
the forest biomes was identified with the help of data on percent
forest cover and forest statistical data on the areal extent of
coniferous and deciduous forests.
a. Areal Estimates
About 95% of the territory of the FSU, including Russia, Ukraine,
Belorussia, Moldova, Kazakhstan, and the Baltic states, was
categorized with the help of a map showing soil-vegetation
associations, i.e., ecosystems (Ryabchikov 1988) (Figure III.l).
The main ecosystem types identified with the help of this map
were the polar desert, tundra, forest-tundra/sparse taiga, taiga,
mixed-deciduous forest, forest-steppe, and warm temperate wood-
lands, steppe, and desert-semidesert.
Because the FSU has a vast longitudinal extent, ecosystems with
the same vegetation cover and soils may differ in carbon pools
and productivity (e.g., densities of phytomass, CWD, litter, and
NPP) depending on their geographical location. Consequently,
eight georegions which reflect the geoclimatical dependence of C
accumulation parameters were isolated in the FSU: West, Central,
and Eastern European Parts; West, Central, and Eastern Siberia;
Near-Ocean region; and Kazakhstan. The maps were superimposed
using a Geographic Information System (GIS) to identify eco-
regions (i.e., ecosystems presented on soil-vegetation map
isolated within specific georegions) related to different soil-
24
-------
fttptttd lot
0 $ J.I. CAIBOX JUDCff MOIICT
U.S. fatfiMMatil Proiegaioi
ltd J. Vfnion, Project Ve»tar
hrbtr* I. tosenbtum. CIS SpeeUlhl
Dioifl/«d Itodi
'Geographic Belli tad Zona1 frtet ol Undicaoei
01 f*e world', ittlt UU.U9.999. I9U.
Ir *t»l a.w. InocMI«» w/ift co-iviho'j.
School oI 6eogrtphr. Waico* Stale Ualtrriilr
Yegrlillon I Soi/s
1 let
1 Stony Itntai I r: Tundra • polar deiert
I Arcta-toadra I dry crylc arctic laili I r: Sbrib • polar deieri
I Men* Meadow I put laili I r. Shrib - tundra
I Grau-moii-ikrub tandra ' |l«|lc,tvri tall* I t: Open woodland tondri
4 iKben-aon lundfi • gleyic toili, padbvr'
1 Sbrub-moti.ibrvb lunorj ' p««i podiollted toili,'podbun
I Sharb-meado*.mall-leal open «oodlan«t coarie hvnic-lurl toll*
I Shrub. coniferous woodland* . gleilc-ferrohuaic podiol podburt
II Cooilrrout .tbrab woodland ' ferro-baaiic podzal.'podbvM
I r: laigi - aeido* ¦ tondri
11 Coaifeioui woodlindi ferro-bataic podrolv'pod&un
I r: Taiga ¦ undra
12 Grass herb ihiub mridowt ' peat-lori toili
I i: Coalferooi loreil ¦ tundra
11 Weado» tallgriii.iowll leal ope* woodland- larf softs. 'podbun
I r: tihed loreil • conlferon lir»il - tondri
14 Coaiferoat buoid taiga .filarial hamo-lentc podioft.'padbur»
I »: llied foreil • conlferon J»re»t - «krub • taeadow
ISa Conlferooi Moderate!* hiraid taiga • ilfa»lal ferro-»(rimc.cr)f< podioft
1Sb Canilrroti aaderatel* hoaiid taiga ' torl-podiollc toils.lerrobuitii podiois
15c CoRilerooi moderate!) bataid Ulga . padtollt tallt. lerlr podznli
Ha (onlferoit moderate!* horaid uiga ferra-hamit padroii.'podburt
egetation/Soil Map
16b Coaileroat ¦ode«tflr huaiif taiga I turl-podrolic toil^
I4i Conileroat aoderilel) kuald Uiga I coarse hunlc podtoii pt-r * oil*
t/j Conileroat tatgi <»r»e taiga, pale toili
17b Conileroat laiga luil-podrolw tollt
II Detiduooi-<:o(ilerout tutiil«elt huald lorett add (aab>toit.podiollr <
I r: Iroadleal forest - conlferooi loreil ¦ alpine meadow
19 Coalferout-broadfea' humid l«re»t ¦ Irrrit podroit torl-podroltr toil.
I r: f aresl-ilcppt ¦ conlfeioui rore«l - meadow ¦ lundra I alpine mcado*
10 Iroadleal coalferooi horaid forest carabisolt.luil-podiollt iolt»
I r: Steppe ¦ conlleroat loreil tandra
?1 Iroadleal-coalleroat noderalelr bonid loretl turl-pediolic to>i«
I r Steppe ¦ conlferooi foreit - alpine raeado*
22 Soall-leaf co»ll»io»i loreti> labtngai ' lad podrolit imlt, grtr laviiolt
I r: Steppe • nliel faietl • meada-
2) Iroadleaf hunid loreil ' larl-podtolic toili. ortfeK rarabiiolt
I r: Steppe ¦ caalleroai form ¦ alpiat taeadow I meadow
24 Iroadleal noderitelt laoid foreili'tart-podrolK toilt gre> luvitolv ado
carablioli I r: Sealdettrl - conlferooi lorett ilplar meadow 1 meidon
26 Swall-leal-coaileroai lor«tl-«itppe>podzol>ird
-------
vegetation associations. Then, ecoregions were combined with
site specific data on carbon accumulation parameters (Kolchugina
and Vinson 1991; Vinson and Kolchugina 1993; Kolchugina and
Vinson 1993a). Forest inventory data on percent forest cover
(Vorobyov 1985; U.S.S.R. State Forestry Committee 1990) were used
to estimate the areal extent of non-forest ecosystems within the
forest biomes (Kolchugina and Vinson 1993d).
The exact area of coniferous and deciduous forests in FSU forest
statistical data is provided for only 654 X 106 ha of FSU forests
under the direct management of FSU. The ratio between the areas
of coniferous and deciduous (separately hardwood and softwood)
forests (reported for 654 X 10* ha) was applied to the areal
extent of forest ecosystems (814 X 10( ha) estimated with the
help of maps and statistical data on percent forest cover.
The estimate of the areal extent of peatlands (Table III.l) was
based on the recent assessment by Botch et al. (1995 which
considered the peatland type distribution, and data from FSU peat
statistics. The area of agricultural land (cropland, pasture-
lands, and haylands) was obtained from the World Resource Insti-
tute report (WRI 1992) and Kolchugina and Vinson (1994b).
b. Carbon Pools
Densities (per ha) of phytomass (Table III.2) and CWD (Table
III.3) of the polar desert and tundra biomes, meadows and grass-
land ecosystems were obtained by correlating the data base on
site-specific densities of total phytomass (aboveground and
belowgroiind) and CWD (aboveground and belowground) compiled by
N.I. Bazilevich (1986) to the areal extent of ecoregions isolated
with the 6IS analyses. The data base on phytomass, CWD, litter,
and NPP resulted from studies of approximately 1,500 vegetation
complexes (sites) in the FSU. The data base is a comprehensive
26
-------
Table III.l. Areas of Major Ecosystems and Land-Use Types of the FSU
Area
Ecosystem/Land-Use Type* (X 106 ha)
Polar desert 8.it
Tundra (cold grasslands/small shrubs) 226.0*
Coniferous (needleleaf) forest (including sparse
northern extreme) 610.7*
Deciduous hardwood forest (oak, stony birch, etc.) 35.1*
Deciduous softwood forest (birch, poplar, aspen,
linden) 168.5*
Cold meadows (partially tundra) 112.1*
Boreal and temperate meadows (moderately moist) 297.9*
Dry temperate grasslands 113.0*
Desert/semidesert 169.o'
PeatlandB 164.8s
Arable land 227.0"
Subtotal - nonforest ecosystems 1317.9
Subtotal - forest ecosystems 814.3
Total 2132.2
*Polar desert, tundra, grasslands, and desert from Ryabchinov 1988.
Coniferous forest, deciduous hardwood forest, deciduous
softwood forest, cold meadows, as well as boreal and temperate
meadows from Ryabchikov 1988; Vorobyov 1985; and U.S.S.R. State
Forestry Committee 1990. PeatlandB from Botch et al. 1995.
Arable land from Kolchugina and Vinson 1994b and WRI 1992.
'Based on Kolchugina and Vinson 1991.
'Based on Kolchugina and Vinson 1993d; U.S.S.R. State Forestry
Committee 1990; Ryabchikov 1988; Vorobyov 1985.
'Based on Botch et al. 1995.
'Kolchugina and Vinson- 1994b; WRI 1992.
27
-------
Table III.2.
Densities of Phytomass, FSU
Phytomass (MgC/ha)
Ecosystem/Land-Use Type*
Above- Below-
Total ground ground
Polar desert 0.3* 0.2* 0.1*
Tundra (cold grasslands/
small shrubs) 8.7* 2.9' 5.8*
Coniferous (needleleaf) forest
(including sparse north-
ern extreme) 64.3* 48.2* 16.1*
Deciduous hardwood forest (oak,
stony birch, etc.) 63.6* 50.9* 12.7*
DeciduouB softwood forest
(birch, poplar, aspen,
linden) 48.5* 38.8* 9.7*
Cold meadows (partially tundra) 11.2* 2.1* 9.1*
Boreal and temperate meadows
(moderately moist) 6.7* 1.3* 5.4*
Dry temperate grasslands 6.0* 1.1* 4.9*
Desert/semidesert 5.1* 1.2* 3.9*
Peatlands 20.0* 10.0* 10.0*
Arable land 2.9* 0.5* 2.3*
*Polar desert, tundra, grasslands, and desert from Ryabchikov 1988. Coniferous
forest, deciduous hardwood forest, deciduous softwood forest, cold
meadows, as well as boreal and temperate meadows from Ryabchikov 1988;
Vorobyov 1985; and U.S.S.R. State Forestry Committee 1990. Peatlands from
Botch et al. 1995. Arable land from Kolchugina and Vinson 1994b and WRI
1992.
*Based on Bazilevich 1986 and Ryabchikov 1988. Density was weighted by area of
different plant communities within each ecosystem type.
'Obtained by dividing totals presented in Table III.6 by the ecosystems area
(Table III.l).
'Based on Kolchugina and Vinson 1994b.
"No data were available on aboveground and belowground allocation; phytomass
was distributed equally above and below ground.
28
-------
Table III.3. Densities of Coarse Woody Debris (CWD), FSU
CWD (MgC/ha)
Ecosystem/Land-Use Type*
Total
Polar desert
Tundra (cold grasslands/
small shrubs)
Coniferous (needleleaf)
forest (including
sparse northern
extreme)
Deciduous hardwood forest
(oak, stony birch,
etc.)
Deciduous softwood forest
(birch, poplar,
aspen, linden)
Cold meadows (partially
tundra)
Boreal and temperate
meadows (moderately
moist)
Dry temperate grasslands
Desert/semidesert
Peatlands
Arable lands
0.4*
12.4*
14.8*
14.6*
11.2*
15. 5*
6.2*
6.2*
4.8*
NA
0J
Above-
ground
0.3*
4.1*
11.1*
11.7*
9.0*
2.9*
0.4*
1.2*
1.2*
NA
01
Below-
ground
0.1*
8.3*
3.7*
2.9*
2.2*
12.6*
5.8*
5.0*
3.6*
NA
0s
*Polar desert, tundra, grasslands, and desert from Ryabchikov 1988. Coniferous
forest, deciduous hardwood forest, deciduous softwood forest, cold
meadows, as well as boreal and temperate meadows from Ryabchikov 1988;
Vorobyov 1985; and U.S.S.R. State Forestry Committee 1990. Peatlands
from Botch et al. 1995. Arable land from Kolchugina and Vinson 1994b and
WRI 1992.
*Based on Bazilevich 1986 and Ryabchikov 1988. Density was weighted by area
of different plant communities within each ecosystem type.
*Based on Kolchugina and Vinson 1993d, Yermolenko and Yermolenko 1982
and Bazilevich 1986.
'Based on Kolchugina and Vinson 1994b.
NA No data available.
29
-------
source of information on common vegetation types in the FSU.
Data on forest ecosystems given in Bazilevich (1986) do not
reflect the forest age and, therefore, cannot be used directly
for the assessment of the vegetation C pools, NPP, and NEP of FSU
forests. The C content of plant mass was assumed at 50% (Kobak
1988).
Phytomass of FSU arable land was estimated from data on
economical yields of major FSU crops applying harvest indexes
(Gaston and Kolchugina 1994).
The estimate of phytomass and CWD of coniferous and deciduous
forests was based on the forest statistical data (data on the
total growing stock [GS]) (Alimov et al. 1989; U.S.S.R. State
Forestry Committee 1990). Three major problems must be resolved
when converting the volume of commercial timber to total forest
plant mass. Specifically, one should account for (1) parts of
trees other than stems, (2) ecosystem components other than
trees, and (3) CWD accumulation. Sampson (1992) suggests the use
of a multiplier of 0.53 to convert the volume of commercial
timber (m3) to total forest phytomass (MgC) (including roots,
branches, and understory). Isayev et al. (1993) developed
conversion factors specifically for FSU forests depending upon
tree genera and forest age (these factors will be provided later
in Table III.4). These conversion factors do not account for
other ecosystem components than trees. Conversion factors
developed by Isayev et al. (1993) were used herein with an
addition of 3% for other-than-trees ecosystem components and
assuming 50% of C in the plant mass. Density of phytomass was
obtained by dividing the total phytomass of the coniferous and
deciduous forests by their area. Density of CWD (Table III.3)
was estimated based on the ratio of stem volume to aboveground
CWD derived from FSU yield tables as described in Kolchugina and
Vinson (1993e). The distribution of phytomass and CWD by above-
30
-------
Table III.4. Accumulation of Phytomass and Coarse Woody Debris in FSU Forests
Conversion
Factor (m3
timber to CMg
total forest
Net
Net
Growing
Net Increment
phytomass)
Accumulation
Accumulation
Average
Stock*
of Stem Hood
Assuming
of Phytomass
of CWD*
Ager
Area*
(GS) ,
(NIW)*
50% Carbon in
(NAPh)
(NACWD)
Class*
SDecies*
(vra)
(X 106 ha)
(X 10 ml
(X 106m3/vr)
Plant Mass*
(TqC/vr)
ITaC/vrl
Young I
Coniferous
10
48.0
675.1
67.51
0.44
29.77
1.75
Hardwood
10
1.6
32.7
3.27
0.62
2.03
0.38
Softwood
10
11.8
114.0
11.4
0.39
4.44
1.04
Young II
Coniferous
30
41.8
2401.8
90.70
0.44
39.99
6.45
Hardwood
30
2.3
148.2
5.06
0.62
3.14
0.92
Softwood
30
12.5
417.8
14.85
0.39
5.78
1.07
Middle Age
Coniferous
50
101.0
11837.0
301.68
0.37
112.67
33.23
Hardwood
50
8.1
928.2
20.31
0.51
10.26
2.77
Softwood
40
38.0
3641.3
118.56
0.39
45.86
6.58
Premature
Coniferous
70
49.7
7652.9
91.41
0.37
33.50
16.52
Hardwood
70
3.2
359.1
-0.38
0.49
-0.19
0.71
Softwood
60
13.0
1730.2
24.22
0.39
9.36
1.71
Mature/
Coniferous
174
277.9
37999.4
218.56
0.38
82.93
16.15
Overmature
Hardwood
138
8.3
1048.7
7.62
0.51
3.89
1.06
Softwood
67
37.1
5900.0
87.91
0.39
33.92
4.53
Subtotals
Coniferous
518.4
60566.2
769.85
298.87
74.10
Hardwood
23.5
2516.9
35.88
19.14
5.84
Softwood
112.4
11803.3
256.95
99.35
14.93
Total
654.3*
74886.4
1062.68
417.36
94.87
'Based on U.S.S.R. State Forestry Committee 1990 and Aliroov et al. 1989.
Species average age: coniferous, 113 yrs; hardwood deciduous, 79 yrs; softwood deciduous, 47 yrs (U.S.S.R. State
Forestry Committee 1990).
*See methodology (Section III.D.6, first approach).
*Based on Isayev et al. 1993; Bazilevich 1986; Kobak 1988.
'Based on the decay rate applied to declining mortmass pool (i.e., amount decayed = [stock of deadwood]*[l - decay
rate ] "*•).
'The complete report is provided for 654.3; the total FSU forested area is 814.3. Estimates were extrapolated to the
entire forested area.
31
-------
and belowground parts was made according to Bazilevich (1986) and
Yermolenko and Yermolenko (1982).
Average densities of litter and soil organic natter (SOM) of the
polar desert, tundra, forest biomes, and grasslands (Kobak 1988)
were weighted by the area of soil types (Ryabchikov 1988;
Kolchugina and Vinson 1993a; Kolchugina et al. 1993) (Table
III.5). Kobak (1988) provides data on over a hundred soil
profiles (In depth) conpiled from the review of approximately 70
published sources. Litter and SOM (i.e., peat) density of
peatlands were obtained fron Bazilevich (1986) and Botch et al.
(1995. An entire peat thickness was considered. Area weighted
average SOM C content of the arable land was calculated based on
the distribution of soil types (Ryabchikov 1988) within FSU
arable land (Cherdantsev 1961; Kolchugina et al. 1993). Mann
(1986) developed an equation to predict current carbon content of
cultivated soil as a function of the initial carbon content. The
overall loss of 24% of the initial C content was estinated for
current conditions of the arable land of the FSU (Gaston et al.
1993). The estimate was obtained by applying the equation of
Mann to major soil types in the FSU.
c. Net Primary Production
The NPP densities of meadows, grasslands, and peatlands were
obtained directly from Bazilevich (1986) (Table III.4). The NPP
of croplands was assumed to be equal to their phytomass (Gaston
and Kolchugina 1994).
The NPP density of coniferous, deciduous hardwood, and deciduous
softwood forests was estimated with the help of Bazilevich (1986)
data adjusted for the presence of young forest ecosystems using
data on the age-class distribution reported in FSU forest
statistical data (Kolchugina and Vinson 1993d). Bazilevich data
32
-------
Table III.5. Densities of Litter, Soil Organic Matter (SOM), and Net
Primary Productivity (NPP), FSU
Ecosystem/Land-Use Type
Litter
(MgC/ha)
SOM
(MgC/ha)
NPP
(MgC/ha/yr)
Polar desert
Tundra (cold grasslands/small
Bhrubs)
Coniferous (needleleaf) forest
(including sparse north-
ern extreme)
Deciduous hardwood forest
(oak, stony birch, etc.)
Deciduous softwood forest
(birch, poplar, aspen,
linden)
Cold meadows (partially
tundra)
Boreal and temperate meadows
(moderately moist)
Dry temperate grasslands
Desert/semidesert
Peatlands
Arable land
0.16*
4.70*
15.30*
7.20'
7.20*
0. 60*
0.70*
1.30*
0. 50*
16.90*
0
50.0*
200.0*
120.0'
160.0'
120.0'
150.0*
185.0*
292.0*
45.0*
1299.0s
153.55
0.06*
1.54*
2.70'
4.60'
4.40'
4.40*
7.50*
6.40*
4.00*
3.40*
2.90"
*For 1 m of mineral soils and for this entire thickness of peat.
*Based on Kobak 1988 and Ryabchikov 1988. Density was weighted by area of
different plant communities within each ecosystem type.
'Based on Kolchugina and Vinson 1993d.
'Based on Botch et al. 1995
^Based on Kolchugina and Vinson 1994b.
*Based on Bazilevich 1986 and Ryabchikov 1988.
33
-------
relate to established ecosystems (middle-age to overmature),
Based on the information presented in forest statistical data
(U.S.S.R. State Forestry Committee 1990), it was estimated that
the ratio of established, forest ecosystems (i.e., middle-age to
overmature) to young forest ecosystems in the FSU is 4.5 to 1.
Thus, for a total of 814 X 106 ha, an estimated 666 X 106 ha
would be occupied by established forest ecosystems and 148 X 106
ha would be occupied by young forest ecosystems. Within the
young forests, initial area-weighted average NPPs were decreased
by a factor of four (this factor was derived based on the
examination of FSU yield tables).
2. Results and Discussion
The largest area of the FSU was occupied by forest ecosystems and
boreal/temperate herbaceous ecosystems (i.e., 38 and 32% of the
total area of FSU terrestrial ecosystems, respectively) (Table
III.l). Peatlands and polar desert/tundra ecosystems occupied 8
and 11% of the area, respectively. Desert/semidesert ecosystems
accounted for 8% of the area. Arable land occupied 11% of the
territory.
The estimate of the forested area was based on the information
provided by the U.S.S.R. State Forestry Committee which is the
best available source of information related to FSU forests.
Kolchugina and Vinson (1993d), using the Ryabchikov (1988) map
and data on percent forest cover, arrived at approximately the
same result. It may be possible, however, that some of the low-
productive and low-stocked northern forest ecosystems (especially
in the forest-tundra zone) are omitted in the State forestry
inventory. The area of peatlands reported herein (Table III.l)
was twice the area estimated by Kolchugina and Vinson (1991) who
based their assessment on the map (Isachenko 1988) which under-
represents peatlands of the tundra zone and the Far East.
34
-------
The total phytomass C pool of FSU terrestrial ecosystems was
estimated at approximately 60.4 PgC with 41.1 PgC and 19.2 PgC
above- and belowground, respectively (Table III.6). The total
CWD C pool of the FSU was estimated at 19.3 PgC (10.4 and 8.9 PgC
above- and belowground) or 30% of the total phytomass C pool
(Table III.7). In polar desert and tundra, CWD exceeded that of
phytomass. The total litter pool was estimated at 15.2 PgC,
i.e., it was almost the same as the pool of CWD (Table III.8).
The SOM C pool was estimated at 506 PgC. Major contribution was
provided by peatlands (214.1 PgC).
Terrestrial ecosystems of the FSU contained 601.1 PgC.
(Disagreement between the above and the value in Table VII.3 is
due to rounding.) Phytomass (Table III.6), CWD (Table III.7),
and litter C pools were 10.0% and 3.2%, and 2.5% of the total,
respectively. The soil C pool accounted for 84.2% of the total C
stored in FSU terrestrial ecosystems. FSU forests (almost all
are in Russia) (49.6 PgC) accounted for 82.1% of the total
terrestrial phytomass C pool and 59.1% (i.e., 11.4 PgC) of the
total C pool of CWD. The NPP of FSU terrestrial ecosystems was
estimated at 8.2 PgC.
D. Net Carbon Flux
The C flux between FSU forests and the atmosphere is determined
by the difference between NEP, and C returned to the atmosphere
through disturbances, i.e. forest fires, logging, and influence
of pollutants, pests, and diseases.
1. Methods
The rate of C sequestration in undisturbed forest ecosystems was
assumed to be equal to their NEP. The NEP equals net
35
-------
Table III.6.
Phytomass, FStJ
PhytomasB (TgC)
Bcosystem/Land-Use Type*
Total*
Polar desert
Tundra (cold grasslands/
small shrubs)
Coniferous (needleleaf)
forest (including
sparse northern
extreme)
Deciduous hardwood forest
(oak, stony birch,
etc.)
Deciduous softwood forest
(birch, poplar,
aspen, linden)
Cold meadows (partially
tundra)
Boreal and temperate
meadows (moderately
moist)
Dry temperate grasslands
Desert/semidesert
Peatlands
Arable land
2.4*
1966.2f
39268.0*
2232.4*
8172.3*
1255.5'
1996.0*
678.0*
861.9*
3296.0*
635.6*
Above-
ground*
1.6*
655.4*
29435.7*
1786.6*
6537.8*
235.4*
387.3*
124.3*
202.8*
1648.0*
113.5*
Below-
ground*
0.8*
1310.8*
9832.3*
445.8*
1634.5*
1020.1*
1608.7*
553.7*
659.1*
1648.01
522.1*
Total
60364.3
41128.4
19235.9
Polar desert, tundra, grasslands, and desert from Ryabchikov
1988. Coniferous forest, deciduous hardwood forest, deciduous
softwood forest, cold meadows, as well as boreal and temperate
meadows from Ryabchikov 1988; Vorobyov 1985; and U.S.S.R. State
Forestry Committee 1990. Peatlands from Botch et al. 1995
Arable land from Kolchugina and Vinson 1994b and WRI 1992.
*Obtained from the multiplication across data presented in Table
IIZ.l and Table.III.2.
*Based on U.S.S.R. State Forestry Committee 1990, Zsayev et al.
1993 (Table III.4), Bazilevich 1986, and Yermolenko and
Xermolenko 1992.
'No data were available on aboveground and belowground allocation;
phytomass was distributed equally above and below ground.
36
-------
Table III.7. Coarse Woody Debris, FSU
CWD (TgC)
Ecosystem/Land Use Type*
Totalt Above- Below-
ground^ ground*
Polar desert
Tundra (cold grasslands/small
shrubs)
Coniferous (needleleaf) forest
(including sparse northern
extreme)
Deciduous hardwood forest (oak,
stony birch, etc.)
Deciduous softwood forest
(birch, poplar, aspen,
linden)
Cold meadows (partially tundra)
Boreal and temperate meadows
(moderately moist)
Dry temperate grasslands
Desert/semidesert
Peatlands
Arable land
3.2
2802.4
9038.4
512.5
1887.2
1737.6
1847.0
700.6
811.2
NA
0
2.4
926.6
6778.8
410.7
1516.5
325.1
119.2
135.6
202.8
NA
0
0.8
1875.8
2259.6
101.8
370.7
1412.5
1727.8
565.0
608.4
NA
0
Total
19340.1
10417.7
8922.4
*Polar desert, tundra, grasslands, and desert from Ryabchikov 1988.
Coniferous foreBt, deciduous hardwood forest, deciduous softwood
foreBt, cold meadows, as well as boreal and temperate meadows from
Ryabchikov 1988; Vorobyov 1985; and U.S.S.R. State Forestry Committee
1990. Peatlands from Botch et al. 1995. Arable land from Kolchugina
and Vinson 1994b and WRI 1992.
'obtained from the multiplication across data presented in Table III.l
and Table III.3.
NA No data available.
37
-------
Table III.8.
Litter, Soil Organic Matter (SOM), and Net Primary Produc-
tivity (NPP), PSU
Ecosystem/Land-Use Type*
Polar desert
Tundra (cold grasslands/small
shrubs)
Coniferous (needleleaf) forest
(including sparse north-
ern extreme)
Deciduous hardwood forest
(oak, stony birch, etc.)
Deciduous softwood forest
(birch, poplar, aspen,
linden)
Cold meadows (partially
tundra)
Boreal and temperate meadows
(moderately moist)
Dry temperate grasslands
Desert/semideeert
Peatlands
Arable land
Litter*
(TgC)
1.3
1062.2
9343.7
252.7
1213.2
67.3
208.5
146.9
84.5
2785.1
0
SOM*
(TgC)
405.0
45200.0
73284.0
5616.0
20220.0
16815.0
55111.5
32996.0
7605.0
214075.2
34844.5
NPP*
(TgC/yr)
0.5
348.0
1648.9
161.5
741.4
493.2
2234.3
723.2
676.0
560.3
658.3
Total
15165.4
506172.2
8245.6
*Polar desert, tundra, grasslands, and desert from Ryabchikov 1988.
Coniferous forest, deciduous hardwood forest, deciduous softwood
forest, cold meadows, as well as boreal and temperate meadows from
Ryabchikov 1988; Vorobyov 1985; and U.S.S.R. State Forestry Committee
1990. Peatlands from Botch et al. 1995. Arable land from Kolchugina
and Vinson 1994b and WRI 1992.
^Obtained from the multiplication across data presented in Table IXI.l
and Table III.5.
38
-------
accumulation of phytomass (NAPh), net accumulation of CWD
(NACWD), net accumulation of SOM (NASOM).
a. Non-Forest Ecosystems
The NAPh and NACWD were assumed to be zero for non-forest
ecosystems of the FSU. These C pools were considered to be in
equilibrium with atmospheric C.
The NASOM was represented by the rate of stable SOM formation.
It was assumed that the accumulation of labile SOM is balanced by
decompositional processes. Density of stable SOM accumulation
was derived from data reported by Kobak (1988). For each
vegetation type shown in Table III.l, a representative soil type
was defined based on the Ryabchikov (1988) soil-vegetation map.
The rate of peat accumulation in FSU peatlands was based on
estimates made by Botch et al. (1995). The research considered
the rate of peat accumulation and the areal extent of every
peatland zone of the FSU.
b. Forest Ecosystems
The presence of significant areas of "growing" forests (i.e.
forests which have not reached the age when C accumulation is
balanced by C emissions due to respiratory processes) is crucial
for the existence of significant NEP. Forest age-structure is
closely connected with the history of disturbances.
Disturbances to FSU forests were especially intensive at the end
of the 19th century and continued through the mid-20th century
(Kolchugina and Vinson 1994a). At present (1988), there is
approximately 323 X 106 ha of young to premature forests reported
for 654 X 106 ha of the forested area (Table III.4). However,
39
-------
the total forested area of FSU is 814 X 106 ha; the total
forested area of Russia is 771 X 106 ha (95%) (1988 Forest
Inventory [U.S.S.R. State Forestry Committee 1990]). The mature
(generally 80 to 100 yr, but may be younger) class refers to
forests that have accumulated maximum timber and are eligible for
commercial harvesting (Vorobyov 1985). The age of overmature
forests may be greater than 160 yrs. FSU forests continue to
seguester carbon beyond the age of technical maturity, e.g.,
larch forests accumulate wood for 100-130 yrs beyond their age of
technical maturity (Kolchugina and Vinson 1993d).
The NAPh was estimated from the net annual increment of stem wood
(NIW). The NIW of the young, middle-age, and premature (i.e.,
the category between middle-age and mature) forests was estimated
as the difference of the GS densities (per ha) of a given and
preceding age-class divided by the difference in age and
multiplied by the area of a given age-class. The NAPh was
estimated using multipliers given by Isayev et al. (1993) to
convert m3 of stem wood to Mg of total tree phytomass. Three
percent was added to account for content C for other than main
story ecosystem components. Carbon of plant mass was assumed at
50% (Kobak 1988) (Table III.4).
The NACWD was estimated from data on (1) the ratio of accumulated
CWD in coniferous, hardwood, and softwood deciduous forests to
their GS (FSU yield tables [Kozlovski and Pavlov 1967]), and the
ratio is dependent on the site quality and age of the forest; (2)
the age-class distribution of growing stock (GS); (3) the
distribution of forests by site-quality (1988 Forest Inventory
[Vorobyov 1986; U.S.S..R. Statfe Forestry Committee 1990]); and (4)
the application of a two percent per year decomposition rate for
coniferous and hardwood deciduous forests and six percent
decomposition rate for softwood deciduous forests (M. Harmon,
pers. comm., 1993). This approach did not account for
decomposition of residue from previous disturbances.
40
-------
Because the difference in the average age between the
mature/overmature and premature forests may be very large,
another method was used to estimate NIW for these forests. The
mature/overmature category includes all forests which are older
than the age at which forests are eligible for principal cutting
(i.e., industrial cuttings other than thinnings and sanitary
cuttings) (as reported by the U.S.S.R. State Forestry Committee
1990; Alimov et al. 1989). The mature/overmature forests include
a variety of forest stands: (l) the even-age forests which are
approximately 80-100 years and older (for the softwood deciduous
forests the age of principal cutting may be designated at a much
earlier date, e.g., 60 years and earlier) and (2) the uneven-age
very old primary forest ecosystems ("old" is a relative term and
refers to the fact that these forests have reached the
successional stage when they can be described as "climax" forests
[Odum 1953]) .
The GS density of the old even-age forests and uneven-age forests
may be significantly lower than the GS density of the premature
forests because of the natural environmental conditions (climate,
soils, elevation, etc.) under which these stands were developing.
For example, in Siberia there is a vast expanse of the uneven-age
forests which are included into the category of the mature/
overmature forests with a relative low GS density. Unfortu-
nately, at the present time, there are no available data on the
exact areal extent and the average age of these two groups within
the mature/overmature forests. Two different approaches were
used to estimate NAPh and NACWD of the mature/overmature forests.
Under Approach I, the average age of the mature/overmature
forests was estimated from the data on the areal extent of forest
age-classes, the average age of every age-class, and the average
age of FSU forests (U.S.S.R. State Forestry Committee 1990)
(Table III.8). Accumulation of timber continues at the average
age estimated for the mature/overmature forests (Kozlovskyi and
41
-------
Pavlov 1967). The NAPh of the mature/overmature forests was
estimated by dividing their GS by the average age and applying
conversion factors. This procedure may overestimate the current
rate of NAPh for old forests (Kozlovsky and Pavlov 1967). The
NACWD was estimated using the same methodology applied for other
age-classes.
The first approach used to estimate NAPh and NACWD of the
mature/overmature forests (Approach I) was based on the assess-
ment of the average age of this forest category, which was
established from data on the average age and the age-class
distribution of FSU forests. However, it is difficult to assess
precisely the average age of FSU forests for the reasons
presented above (i.e., a presence of the uneven-age forests). A
second approach (Approach II) was used to assess NPAh and NACWD
of the mature/overmature forests, and it did not rely on the data
on dynamics of the GS of the mature/overmature forests in 1983
and 1988 reported by the U.S.S.R. State Forestry Committee
(Alimov et al. 1989).
The data reported by Alimov et al (1989) allow one to estimate
the difference in the GS of the mature/overmature forests between
1983 and 1988 (i.e., a 5 year period). It is important to
remember that principal cuttings take place specifically in the
mature/overmature category of forests (by definition of this
forest age category). Therefore, the difference in the GS of the
mature/overmature forests between 1983 and 1988 reflects the
dynamics of timber growth, principal cuttings, and other
disturbances which took place during the period considered. This
approach assumes inventory procedures are consistent across time
and that equivalent forest areas are treated at each point in
time.
Considering (1) the difference in the GS of the mature/ over-
mature forests between 1983 and 1988, (2) the rate of principal
42
-------
cuttings during this period, and (3) the forested areas included
in the 1983 and 1988 inventories, the rate of phytomass accumula-
tion/loss (from other reasons than principal cuttings) was
estimated (separately for the coniferous, hardwood deciduous, and
softwood deciduous forests). It should also be noted, that the
U.S.S.R. State Forestry Committee (Alimov et al. 1989) reports
data for only 654 X 106 ha (the area under the direct management
of the Federal Government). The final results must be adjusted
to the total forested area of 814 X 106 ha.
The NASOM was assessed separately for the young-to-premature
and the mature/overmature forests. In mature/overmature
forests, NASOM was considered to be represented by the rate of
stable soil organic matter formation and was estimated as
previously described. In young—to—premature forest ecosystems,
an additional accumulation of labile soil organic matter was
assumed. The total accumulation of SOM of forest ecosystems
(labile and stable) was estimated at approximately 0.1 MgC/
ha/yr (Kolchugina and Vinson 1993a).
c. Disturbances
Approximately 2.5 X 106 ha of forests are burned and reported
annually in areas of active monitoring, which amount to 64% of
the total area under the State forest management. The total
burned area is larger (i.e., 3.5 X 106 ha in 1991 and includes
non-forest ecosystems (Korovin 1993). A larger burned area of
10.0 X 106 ha/yr for Central Siberia alone was obtained from
satellite system remote sensing data (Stocks 1991). However,
these data may misrepresent forest fires in the observed area
because the State Forest Inventory reports only 1.67 X 106 ha of
forest fires in Siberia for the same period (Izrael and
Rovinskyi 1990).
43
-------
Forest fires cause an immediate C efflux to the atmosphere from
burned plant mass, litter (forest floor), and SOM. Later,
mortality, soil respiration, and decomposition of killed (but
not burned) trees may cause additional C emissions (Auclair
1985; Agafonov 1989). The C loss in an ecosystem depends on
the type of site, location, and the intensity of forest fire
(Phalaleyev 1985; Kurbatskyi and Tsykalov 1988). An intense
fire may deplete the volatile nutrients in the soil (Raison et
al. 1985) and inhibit growth of vegetation in the post-fire
period, however, in some instances a decrease in C content may
occur in the uppermost soil layer with no significant change in
the lower mineral horizons.
A pyrogenic enhancement of forest growth is often observed
after fires of low intensity, owing to mineralization of
organic matter, mobilization of nutrients in soils, thermal
amelioration (especially in areas with severe climatic condi-
tions, e.g., the permafrost zone), and natural selection of
woody plants with high growth rates. In light-crown coniferous
forests (larch and pine), fire opens the crown, allowing shade-
intolerant young growth to reach the canopy. The bark of these
trees helps to resist fires of low intensity (Phalaleyev 1985;
Johnson 1992) .
Even though there is no strong scientific evidence that the
extent of forest fires in the FSU is significantly greater than
officially reported, this possibility should not be excluded.
Light-crown coniferous forests accumulate fuel that is avail-
able for combustion every 30 to 40 yr (Phalaleyev 1985), so a
40-year return interval was assumed.
The areas of pine and larch forests are 120 X 106 ha (GS density
125 m3/ha) and 258 X 106 ha (GS density 100 m3/ha), respectively
(Vorobyov 1986). The forest plant mass densities are 55 MgC/ha
and 44 Mg C/ha, respectively (using multipliers of 0.37 (Table
44
-------
III.4) to convert m3 of timber to ecosystem Mg C and 1.19 to
account for mortality (phytomass to CWD ratios, see Tables
III.2 and III.3, coniferous forests). It was assumed that
during low-intensity understory fire with a 40-yr frequency,
10% of the total plant mass is burned. This corresponds to 45
TgC/yr for the total fire area of 9.5 X 106 ha. The probability
of post-fire emissions is low, since little damage occurred; on
the contrary, fire could enhance plant mass accumulation by
opening space for young trees and mobilizing nutrients.
Fire monitoring may register only large crown fires which are
predominant in dark-crown coniferous forests (e.g., "cedar"-
pine [Pinus sibirica, P. koraiensis] and also fir and spruce).
The total stock of phytomass, CWD, and SOM available for
combustion would be approximately 31 TgC/yr, 6 TgC/yr, and 20
TgC/yr, respectively. These values were calculated consider-
ing (1) the GS of dark-crown coniferous forests of 138 m3 ha,
(2) conversion factors presented above, (3) average soil carbon
density 136 MgC/ha (Kolchugina and Vinson 1993a), (4) 25% of
SOM would be lost (the selection of this value was based on
literature review [Kurbatskyi and Tsykalov 1988; Phalalevev
1985; Nikodimov 1988; Furyaev 1989; Johnson 1992]), and (5) the
fact that 0.6 X 106 ha of large forest fires are annually
registered in FSU official data. The total direct C emissions
were estimated at 57 TgC/yr. In this case, a post-fire effect
is possible, but very little plant mass is available for
additional decomposition under an assumption made herein that
fire consumes the entire stock of phytomass and CWD.
In exposed soils with a 10°C temperature increase, decomposi-
tion of belowground CWD and SOM may be 2.4 times greater (Raich
and Schlesinger 1992). The emissions of C from the soil
surface associated with decomposition of belowground CWD,
litter, and SOM in FSU taiga forest ecosystems under normal
temperatures was estimated at 2 MgC/ha/yr on average
45
-------
(Kolchugina et al. 1993). To estimate the upper level of post-
fire emissions, it was assumed that no regrovth of vegetation
occurs and rates of decomposition in exposed soil are higher
during the 10-yr period after a fire within 0.6 X 106 ha
subjected to intense crown fires. The total post-fire emis-
sions were estimated to be 3 TgC/yr. This is a very rough
approximation because the exposure to sun may inhibit decompo-
sition processes by increasing temperatures and decreasing
moisture content to an extent that limits microbial growth.
From the 1960s through the 1980s, timber production in the FSU
remained at a constant level of 380.8—411.5 Mm3/yr (Vorobyov
1986, 1985). It is reasonable to assume that carbon was
accumulating in long-term wood products. From 389 Mm3/yr of
wood harvested in 1987-1989 (WRI 1992), 84 Mm3/yr was burned,
and 305 Mm3/yr was used as sawnwood, panels, and paper products.
This represents 148 TgC/yr, considering a conversion factor of
0.38 (for the prevailing coniferous and soft deciduous stands,
Table III.4) to convert m3 of timber to Mgc of ecosystem
phytomass, including slash carbon. Wood products accounted for
79 TgC/yr (conversion factor 0.26 to convert m3 of timber to
MgC). Melillo et al. (1988) estimated carbon effluxes from the
decay of 10- to 100-yr living products at 46 TgC/yr; the net
accumulation of wood products, therefore, was estimated at 33
TgC/yr. This amount was subtracted from the estimate of total
carbon associated with phytomass of ecosystems subjected to
logging. Net carbon emissions from forest logging were
estimated at 115 TgC/yr.
Industrial pollution in the FSU is localized; in 1989, 0.107 X
106 ha of the FSU forests were killed by different pollutants
(U.S.S.R. Committee for Protection of Nature, 1989). Approxi-
mately 0.039 X 106 ha/yr of forests were killed by pests,
diseases, and wild animals. The total area severely affected,
but not killed by pollutants, pests, and diseases is 3.0 X 106
46
-------
ha; an additional 0.14 X 106 ha of forests were cleared for
purposes other than commercial timber production (U.S.S.R.
Committee for Protection of Nature 1989; Izrael and Rovinskyi
1990). The decrease in productivity of affected forest
ecosystems was ignored herein because this area is only 0.4% of
Russian forests. Besides, forests affected by pollutants,
pests, and diseases are the primary target for fires. There-
fore, it may be assumed that the productivity loss in forests
affected by pollution, pests, and diseases may be accounted for
in forest fire emissions. Loss in productivity may also be
accounted for in forest statistical data.
2. Results and Discussion
In the mid-to-late 1980s, non-forest terrestrial ecosystems
were a source of C to the atmosphere, while forest ecosystems
acted as a net sink for atmospheric C in the FSU.
a. Non-Forest Ecosystems
The total NASOM of non-forest ecosystems was estimated at 5.3
TgC/yr (Table III.9). Considering peat accumulation (52.7
TgC/yr), and burning (200 Tg of peat as dry organic matter, or
100 TgC) the total loss from peat accumulation/burning was 47.3
TgC/yr (C efflux).
At present, the arable land of the FSU is losing 94.2 TgC/yr
due to mineralization of SOM and soil erosion, respectively
(Kolchugina and Vinson 1994b).
b. Forests
Using Approach I and data from Table III.4, the NAPh of 814 X
106 ha forest ecosystems in the FSU was estimated at 519.5
47
-------
TgC/yr (Table III.9). The NACWD was estimated at 118.1 TgC/yr.
The NASOM was estimated at 40 TgC/yr. Density of NAPh, NACWD,
and NASOM of FSU forests (Table III.10) was obtained by
dividing total NAPh and NACWD (Table III.9) by the area of
specific forest types (Table III.l). The sequestration
potential or NEP of FSU forestsr-if they were left undisturbed-
was estimated at 0.678 PgC/yr and consistent with assessment
made by Bonan (1991a). Russian forests represent 95% of the
total NEP of the FSU. Previously NEP of FSU forests was
estimated at 0.83 PgC/yr (Kolchugina and Vinson 1993d) and 0.88
PgC/yr (Kolchugina and Vinson 1993e). The first assessment was
based on the assumption that forests in the FSU are fully
stocked and did not consider CWD accumulation/decomposition.
The second assessment ignores CWD decomposition in undisturbed
forest ecosystems. Because the present estimate is based on
the results of the 1988 Forest Inventory (the Inventory is
conducted every decade for the area under State management), it
reflects the age, quality, and stocking of FSU forests as well
as losses of forest productivity owing to forest fires or the
influence of pests, diseases, and pollutants. However, the
precise assessment of the average age of FSU forests may
represent a problem. Losses of phytomass, CWD, litter, and
soils burned directly as well as post-fire emissions are not
included in forest statistical accounting.
Using Approach II to estimate NAPh and NACWD for the mature/
overmature forests, it appeared that in 1983 the GS of the
coniferous mature/overmature forests reported for 654 X 106 ha
was 40,856 Mm3, but it was only 37,999 Mm3 in 1988 (Alimov et
al. 1989). Therefore, between 1983 and 1988 the GS of the
coniferous mature/overmature category of FSU forests decreased
by 2,856 Mm3. However, according to the U.S.S.R. State Forestry
48
-------
Table III.9. Net Accumulation of Phytomass (NAPh),
and Soil Organic Matter Accumulation
Coarse Woody Debris
(NASOM), FSU
(NACWD),
Ecosystem/Land-Use Type*
NAPh
(TgC/yr)
NACWD
(TgC/yr)
NASOM
(TgC/yr)
Approach*
I II
I II
Polar desert
0
0
0
Tundra (cold grasslands/small
shrubs)
0
0
0.5*
Coniferous (needleleaf)
forest (including sparse
northern extreme)
372.0* 96.0*
92.2* 240.7*
20.8*
Deciduous hardwood forest
(oak, stony birch, etc.)
23.8* 37.6*
7.3* 10.9*
1.5'
Deciduous softwood forest
(birch, poplar, aspen,
linden)
123.7* 119.0*
18.6* 8.3*
5.7'
Cold meadows (partially
tundra)
0
0
0.6'
Boreal and temperate meadows
(moderately moist)
0
0
3.3s
Dry temperate grasslands
0
0
0.8s
Desert/semidesert
0
0
0.2$
Peatlands
0
0
52. 7#
Arable land
0
0
-94.2*
Total
519.5* 252.6*
118.1* 269.9*
-8.2
*Polar desert, tundra, grasslands, and desert from Ryabchikov 1988.
Coniferous forest, deciduous hardwood forest, deciduous softwood forest, cold
meadows, as well as boreal and temperate meadows from Ryabchikov 1988;
Vorobyov 1985; and U.S.S.R. State Forestry Committee 1990. Peatlands from
Botch et al. 1995. Arable land from Kolchugina and Vinson 1994b and WRI
1992.
tTwo different approaches (described in Section Ill.D.l.b) were used to estimate
NAPh and NACWD of the mature/overmature forest ecosystem.
'Based on Alimov et al. 1989; U.S.S.R. State Forestry Committee 1990; Isayev
et al. 1993 (see methodology, Section Ill.D.l.b)
'Based on Kobak 1988 and Ryabchikov 1988.
'Based on Kolchugina and Vinson 1993d.
*Based on Botch et al. 1995.
*Based on Kolchugina and Vinson 1994b.
49
-------
Table III.10.
Densities of Net Accumulation in Phytomass (NAPh), Coarse Woody
Debris (NACWD), and Soil Organic Matter Accumulation (NASOM), PSU
Bcosystem/Land-use Type*
NAPh
(MgC/ha/yr)
NACWD
(MgC/ha/yr)
NASOM
(MgC/ha/yr)
Approach* I
Polar desert
Tundra (cold grasslands/
small shrubs)
Coniferous (needleleaf)'
forest (including sparse
northern extreme) 0.61*
Deciduous hardwood forest
(oak, stony birch, etc.) 0.68*
Deciduous softwood forest
(birch, poplar, aspen,
linden) 0.73*
Cold meadows (partially
tundra)
Boreal and temperate
meadows (moderately
moist)
Dry temperate grasslands
Desert/semidesert
Peatlands
Arable land
II I II
0 0 0.001*
0 0 0.002*
0.16* 0.15* 0.39* 0.034*
1.0* 0.21* 0.31* 0.042*
0.71* 0.11* 0.11* 0.034*
0 0 0.005*
0 0 0.011*
0 0 0.007*
0 0 0.001*
0 0 0.320*
0 0 -0.4155
*Polar desert, tundra, grasslands, and desert from Ryabchikov 1988.
Coniferous forest, deciduous hardwood forest, deciduous softwood forest, cold
meadows, as well as boreal and temperate meadows from Ryabchikov 1988;
Vorobyov 1985; and U.S.S.R. State Forestry Committee 1990. Peatlands from
Botch et al. 1995. Arable land from Kolchugina and Vinson 1994b and WRI
1992.
*Two different approaches were used to estimate NAPh and NACWD of the
mature/overmature forest ecosystems. Obtained by dividing NAPh, NACWD and
NASOM totals (Table III.9) by the area of ecosystems/land-use types (Table
III.l).
*Based on Kobak 1988 and Ryabchikov 1988.
'Based on Botch et al. 1995.
^Obtained by dividing NASOM total (Table III.9) by the area of arable land
(Table III.l).
50
-------
Committee (Alimov et al. 1989), the area of the coniferous
forests included in the 1988 inventory decreased by 0.7 X 106 ha
compared to the 1983 inventory, which corresponds to a difference
of 95.2 Mm3 (assuming the average density of the GS of the
mature/ overmature forests at 136 m3/ha [Table III.4]). This
volume was assumed to have remained on the forest land base.
Thus, for the equivalent area, the annual loss of the GS of the
coniferous mature/overmature forests was 552.1 Mm3/yr, or 687.2
Mm3/yr estimated for 814 X 106 ha. Between 1983 and 1988, the
average rate of principal cuttings of the coniferous forests was
233.4 Mm3/yr. Therefore, between 1983 and 1988, for the equiva-
lent area, there was a loss of 453.8 Mm3/yr of standing timber
which was independent of principal cuttings (i.e., caused by
natural mortality or other than logging disturbances). This loss
of the GS corresponded to 172.5 TgC/yr (assuming 0.38 conversion
factor for the coniferous stands [Table III.8]). This loss of C
was assumed to result in the formation of a new pool of CWD. The
decomposition of CWD may be assumed at 2%/yr (as for the
coniferous stands). Thus, the NACWD in the mature/overmature
forests was estimated at 169 TgC/yr.
In 1983, the GS of the hardwood deciduous mature/overmature
forests was 977 Mm3, and it was 1,049 Mm3 in 1988 (reported for
the area of 654 X 106 ha) (Alimov et al. 1989). In 1988, the
area included in the inventory was 0.1 X 105 ha lower than in
1983. The decrease in the inventoried area of the hardwood
deciduous forests may account for 12.6 Mm3 decrease in the GS
(assuming 126.4 m3/ha the GS density [Table III.8]). Therefore,
between 1983 and 1988, for the equivalent area, the GS of the
hardwood deciduous mature/overmature forests actually increased
by 84.6 Mm3 (5 years), which corresponded to 21.1 Mm3/yr esti-
mated for 814 X 106 ha. Between 1983 and 1988, the average rate
of principal cuttings of the hardwood deciduous forests was 17.0
assuming Mm3/yr. This value was calculated from the data
assuming total volume of deciduous timber logged (95.2 Mm3/yr)
51
-------
and the relative areas of the hardwood and softwood deciduous
forests (1:4.6) (Alimov et al. 1989). Between 1983 and 1988,
standing timber increased by 38 Mm*/yr (before logging). This
corresponded to NAPh of 18.6 TgC/yr (assuming 0.49 conversion
factor for the hardwood deciduous stands [Table III.4]).
Furthermore, NACWD may be assumed at 27% of HAPh (Table III.4;
middle-age forest category), or 5.0 TgC/yr.
In 1983, the GS of the softwood deciduous mature/overmature
forests was 5,819 Mm3 and 5,900 Mm3 in 1988 (reported for 654 X
106 ha) (Alimov et al. 1989). The area included in the 1983 and
1988 inventories was similar. Therefore, between 1983 and 1988,
the GS of the softwood deciduous mature/overmature forests
actually increased by 80.9 Mm3, which corresponded to 20.2 Mm3/yr
estimated for 814 X 106 ha. Between 1983 and 1988, the average
rate of principal cutting of the softwood deciduous forests was
78.2 Mm3/yr. Between 1983 and 1988, the standing timber in-
creased by 98.4 Mm3/yr (before cuttings), which corresponded to
NAPh of 38.4 TgC/yr (assuming 0.39 conversion factor for the
softwood deciduous stands [Table III.4]). Furthermore, NACWD may
be assumed at 14% of NAPh (Table III.4; middle-age category), or
5.4 TgC/yr.
Using Approach II, the NAPh of the coniferous and deciduous
mature/overmature FSU forests was estimated at -115.5 TgC/yr and
the NACWD of these forests was estimated at 179.0 TgC/yr. The
NAPh of the young, middle-age, and premature FSU forests
(determined under Approach I) was estimated at 369.2 TgC/yr
(estimated for 814 X 106 ha) and the NACWD of these forests was
estimated at 91.0 TgC/yr. Thus, the total NAPh of the FSU
forests was 253.7 TgC/yr and the total NACWD was 270.0 TgC/yr.
Approach II relied on a correct representation of the GS volume
of the mature/overmature forests in 1983 and 1988. The loss of
453.8 Mm3/yr is equivalent to the area of approximately 3.3 X 106
ha. The loss of the GS of the coniferous mature/overmature
-------
forests may be explained by several reasons, e.g., natural
mortality of the even-age old forests, dieback from pests,
diseases, fires, windfalls, etc.
The decrease in the GS of the mature/overmature coniferous
forests which was not attributed to logging could also be
associated with the fact that between 1983 and 1985 the area of
the even-age old forests with mortality (but which still have
growing trees) increased and these forest stands were not
included in the inventory of commercial timber. Furthermore,
forests subjected to forest fires and diseases or attacked by
insects and, as a result, containing large number of dead trees
or trees with burn-marks could also be excluded from the
inventory. In addition, under-reported logging could account for
some decrease in the GS of the mature/overmature forests.
The direct fire emissions from low-intensity understory fires
were estimated at 45 TgC/yr for the total burned area of 9.5 X
106 ha. The replenishment of this phytomass may be quick enough
to make the emission only a short-lived efflux. The direct fire
emissions within 0.6 X 106 ha of intense crown fires were esti-
mated at 57 TgC/yr. Post-fire emissions were estimated at 3
TgC/yr. Thus total fire emissions were estimated at 105 TgC/yr.
The overall release from logging related to wood and slash
burned, plus slash and wood products decomposition was 115
TgC/yr.
In the mid-1980s, FSU forests were a net sink for 0.343-0.458
PgC/yr of atmospheric C (Table III.11). The difference in the
assessment of the net C flux associated with FSU forests (natural
ecosystems, disturbances, and wood products) was substantial (127
TgC/yr) and was associated with the uncertainty in the estimates
of the NEP of the mature/overmature age-class. More data on the
GS, age distribution, and the description of disturbances for the
mature/overmature category of FSU forests are needed for the
53
-------
Table III.11. Summary: Carbon Pools, Fluxes, Net Sequestration, and Emissions,
FSU
Carbon
PgC
PgC/yr
(-) Pgc/yr
Pools
Phytomass
Coarse woody debris (CWD)
Litter
Soil organic matter (SOM)
Subtotal
60.4
19.3
15.2
506.0
601.1
Fluxes
Net primary productivity
8.2
Net Accumulation
(Sequestration)
Forest ecosystems
Approach
I
II
Phytomass
0.520
0.253
CWD
0.118
0.270
SOM (labile, stable)
0.040
0.040
Subtotal
0,678
Pt563
Non-forest ecosystems
Peat
Soil
Subtotal
Total
0.053
0.005
0.058
0.736 0.621
Emissions
Forest ecosystems/disturbances
Fires
Logging
Subtotal
Nonforest ecosystems
Arable land (soil)
Peat utilization
CH4 from peatlands
Subtotal
Industrial
Total
(0.105)
(0.115)
(0.2201
(0.094)
(0.100)
(0.030)
<0-2241
f1.020)
(1.4641
Net flux (sequestration and disturbance
in terrestrial ecosystems, as well as
industrial emissions)
Net flux (sequestration and disturbance
in terrestrial ecosystems)
Net flux forest (sequestration and
disturbance in forest ecosystems)
0.728-0.843
0.292-0.177
0.458-0.343
54
-------
improved assessment of the carbon flux within this age-class.
These estimates are within the range previously reported for FSU
forests (0.4-0.5 PgC/yr) (Sedjo 1992; Kolchugina and Vinson
1993d,e; Kolchugina and Vinson 1994a). Kolchugina and Vinson
(1993d) over-estimated stocking of FSU forests. Kolchugina and
Vinson (1993e) used data on the total mean annual increment
reforested by the U.S.S.R. State Forestry Committee, and they
reported data (1994a) obtained by utilizing Approach I presented
herein. The present estimate was greater than estimates made by
Melillo et al. (1988) and Krankina and Dixon (1993) who did not
consider the age-structure of Russian forests. The net C sink
for FSU forests exceeded the net C sink estimated for Canadian
forests (Kurz and Apps 1993). The area of 20- to 60-yr FSU
forests (maximum rate of C sequestration) is 265 X 105 ha, i.e.
1.7 times the area of the same age-group in Canada. Terrestrial
ecosystems of the FSU were a net sink for 0.177-0.293 PgC/yr
(Table III.11). Methane emissions from industrial processes and
peatlands (Kolchugina and Vinson 1994a) amount to 56 TgC/yr.
Emissions of chlorofluorocarbons are insignificant (WRI 1992;
Subak et al. 1993).
Russian forests accounted for 95% of the net sink of C estimated
for FSU forests. In the 1990s, logging of Russian forests was
only half as intensive as in the 1980s (Korovin, pers. comm.,
1994) . The C emissions associated with logging may thus be
reduced significantly. At present, Russian industrial emissions
are 0.7 to 0.8 PgC/yr (Karaban et al. 1993). Currently, Russian
forests may be offsetting approximately 43% of the national
industrial emissions.
55
-------
E. Forest Management Options for Mitigating
Carbon Emissions
At present, it is possible to provide only expert assessment of
forest management options available in the FSU (Shvidenko and
Rozhkov, pers. comm., 1992).
1. Reforestation
It is possible to afforest 95 X 106 ha of currently unforested
area under the jurisdiction of the federal government (from 145 X
106 ha theoretically available for reforestation). This option
would provide a net sequestration of 0.18 PgC/yr (an average of
1.9 MgC/yr).
2. Increased Fire Control and Prevention
The extent of forest fires could be decreased by 50%. Forest
fires in the FSU are a natural occurrence in forest ecosystems.
Fire control and prevention may cause an accumulation of a
significant amount of fuels available for burning, thus leading
to catastrophic fires. Hence, the feasibility of this management
option should be investigated further.
3. Enhanced Forest Productivity
This goal could be achieved mainly by reconstructing low-stocked
stands and partial replacement of deciduous stands with long-
rotation conifers. It may be possible to reconstruct 140 X 106
ha of low-stocked stands in the FSU and replace 60 X 106 ha of
deciduous stands (low productive, small leaf, deciduous forests)
in the European Part of the FSU. However, the replacement of
deciduous stands by conifers may present a serious problem for
FSU (Russian) forests. Deciduous and mixed stands may be more
56
-------
resistant to pests, diseases, and fires, than coniferous stands,
especially monoculture stands. The issue on the replacement of
old-growth forests by young forest plantations is also question-
able considering the large quantities of coarse woody debris
stored in mature and over-mature forest ecosystems compared to
young forests.
4. Improved Disease and Pest Control
This measure would include the decrease and subsequent elimina-
tion of the pests and diseased areas, removal and selective
cutting of dead and sick trees, etc. This option may help to
prevent fires and decrease C emissions.
5. Reconstruction of the Forestry
Industrial Sector
The reconstruction includes a transition to ecologically
protective methods of wood harvesting that would allow rapid
regeneration of forest stands and an increase in efficiency of
wood utilization from 50 (present assessment) to 70%. This is a
very significant management option. At present, forest phytomass
is utilized (efficiency of logging) at levels between 30 to 70%
of the total. Poor extraction of timber causes overcutting of
significant forest areas in the FSU. It would be a better use of
forest resources designated for logging to reduce clear cut
areas.
6. Increased Rotation Length
Increase in forest logging age in the European part of the FSU is
theoretically realistic and would result in substantial addi-
tional sequestration of C. If the age of main cuttings were
increased by 20 years on 20 percent of the forest areas presently
57
-------
subjected to harvesting, an estimated 0.2 PgC/yr could be
sequestered. However, high demand for wood by domestic and
international users might prevent the implementation of this
option.
58
-------
Box 2: United States
The United States is a predominantly
temperate zone country (Figure
IV.1), occupying an area of 936 x
106 ha (765 x 106 ha in the
conterminous U.S.). The total
population is 263 x 106, and the
population density is 280/1000 ha.
Recent estimates of population
project increases of nearly 1% per
year over the next few decades, a
rate that is low by comparison with
many less developed countries but is
considerably higher than many other
industrialized, temperate-zone
countries (WRI 1992). Per capita
Gross National Product is relatively
high at $21,000 per year and per
capita energy consumption (2,956 GJ)
is nearly the highest in the world
(WRI 1992).
Land cover in the conterminous U.S.
is approximately 1/3 forest, 1/3
rangeland, and 1/3 cropland,
pastureland or other uses (USDA
1989). The proportion of the land
base which is arable is relatively
high (25%), and agricultural exports
are a significant component of the
national economy. The current ratio
of potential forest land to actual
forest land is 1.8 to 1 (Turner et
al. 1993b), reflecting periods of
extensive land-cover change over the
last several centuries. Much of the
remaining forest has been logged
over the last century. After modest
reversions of land from agricultural
to forest use in the middle decades
of this century, the rates of
interconversions among the land-use
categories have slowed (Alig et al.
1990; MacCleery 1992). A slow
decline in the area of forest land
is projected over the coming decades
(4% by 2040) driven primarily by
urbanization and conversion of
forest land to agriculture.
Land ownership in the U.S. is
predominantly private with the
cropland, pastureland, and other
categories almost exclusively found
in private. Federal ownership of
forest land and rangeland amounts to
33% of the total land in those
categories (USDA 1989). Approxi-
mately 6% of the total forest land,
primarily in public ownership, is
reserved for conservation and
recreational uses.
The forest products industry is
important at the national scale and
dominates the economy at the com-
munity level in some cases. Indus-
trial forest lands, however, cover
only 15% of the forest lands in the
50 states (Alig et al. 1990).
Harvest levels have increased over
the last few decades and now account
for about 75% of net growth on
forests land available for harvest
(Powell et al. 1993). Public foreBt
provide 18% of the national harvest
(Powell et al. 1993).
The U.S. imports and exports logs as
well as finished forest products.
In the case of forest products, the
value of imports were nearly double
that of exports in 1985 (Haynes
1990). However by 1990, the value
of imports had grown by 38% while
that of exports grew by 133%, making
exports equal to 84% of imports.
For logs, the opposite was the case
with virtually no imports but
extensive exports. Exports were 8
times the imports in 1985 (by
volume) and further increased to 11
times by 1990. At the national
level, the net effect is an import
of tree C.
59
-------
kilometers
Forest - conifer
Forest - hardwood
Forest - mixed
Mixed woodland
Desert/shrubland
I] Alpine - tundra
I Grassland/pasture
] Cropland
I Barren
__j Water
Figure IV.1.
United States land-cover map derived from the
AVHRR sensor (Loveland et al. 1991).
60
-------
IV. United States Case Study
A. Introduction
The United States is a large, heavily industrialized nation and
is characterized by relatively high per capita emissions of
fossil C02. The country is also extensively forested. Reversion
of land from agriculture, and pastureland to forest in the mid
decades of this century, and the recovery of lands in the eastern
U.S. from earlier heavy cutting (USDA 1992a), has created a
significant C sink (Heath and Birdsey 1993; Turner et al. in
press a). The net C emissions of the U.S. thus reflect both
sources associated with fossil fuel combustion and net sinks
associated with the forest land base and forest products. In
this section, previous studies and recent analyses are used to
develop for the conterminous U.S. a C budget which accounts for
both anthropogenic and biological factors.
The complete C budget includes C pools and principal C sources
and sinks. Pools are subdivided into soil, litter, woody debris,
and phytomass components. The C fluxes include those, such as
fossil fuel emissions and accumulation of C in forest products
and landfills, which are controlled primarily by anthropogenic
factors, as well as those controlled by biological
processes—notably net ecosystem production.
The C flux associated with the forest land base receives the
greatest attention in the U.S. analysis. Recent studies suggest
that the rate of conversion of land from one use to another is
relatively low in the U.S. (Alig et al. 1990), so the C source
from deforestation is minimal. However, much of the U.S.
forestland is managed for timber production, and the relative
rates of growth and harvest determine the direction and magnitude
of the associated C flux. The harvest rates also influence the
61
-------
rate of C sequestration in wood products still in use and in
landfills.
B. Methods
The objective of this study was to estimate C pools and flux for
the complete land base of the conterminous U.S. As an initial
step, the land base was subdivided into the 5 land-use types used
in the USDA Resource Protection Act reports (SCS 1987). The
approach for estimating C pools and flux varied according to
land-use category, and each will be discussed separately.
1. Forestland
In this category, all forested land capable of producing 1.4
m3/ha/yr of industrial wood is treated. Lands available for
harvest as well as reserved lands were included. "Other Forest"
land or woodland (Powell et al. 1993), with relatively low total
ecosystem carbon (TEC) and productivity, are treated below.
The basic approach to quantifying C pools and flux on forestland
was to link a complete forest inventory for 1990 with a set of
stand-level C budgets. The inventory contained information on
the area and stocking level for each combination of region,
forest type, age-class, productivity level, and management
intensity. The stand-level C budgets contained estimates of C in
the tree, understory, litter, woody debris, and soil (organic C)
at each age class. The total C pool was quantified as the
product of the areal extent, the stocking level of each age class
(within each inventory type), and the C pool values in the
associated stand-level C budget. The C flux was estimated by
advancing the inventory one time step and considering harvest-
related transfers.
62
-------
a. The Forest Inventory
A forest inventory for private lands in 1990 was prepared as part
of the USDA Forest Service 1993 Resources Planning Act Assessment
(J. Mills, USDA Forest Service, Portland OR, pers. comm., 1994;
Haynes 1990). The inventory reflects survey data from regional
Forest Inventory and Analysis (FIA) work units (USDA 1992b) and
was structured in the framework of the ATLAS inventory projection
model (Mills and Kincaid 1992). The ATLAS-based inventory
includes information on the area and stocking level within each
age class for 422 combinations of region, forest type, productiv-
ity level, and timber management intensity. Uncertainty
associated with FIA inventories is discussed in Peterson and
Turner (1993).
Data for the age-class distributions on public lands within each
forest type exist for particular areas, but they have not been
compiled on a national scale. An estimate of the age-class
distributions and stocking levels on public lands was developed,
therefore, in part from data in recent state-level studies.
These reports collated age-class data for national forests,
reserved lands and other public lands in the Pacific Northwest
region, where the largest volumes of public timber are found
(MacLean et al. 1991; Sessions 1991; Adams et al. 1992). Other
sources of data included the areas by forest type and the volumes
by forest type on public lands reported in Waddell et al. (1989)
and the age-class distributions on private lands (Turner et al.
1993a).
b. Construction of Stand-Level
Carbon Budgets
The Forest Carbon Model (Turner et al. 1993a; Turner et al.
1994a, is a set of stand-level C budgets (e.g. Figure IV.2),
63
-------
which relate the growth and yield tables from ATLAS to trends in
total ecosystem C over the course of stand development.
Allometric relationships are used for converting merchantable
volume in the growth and yield tables into tree C. A modeling
approach involving harvest residue, rates of tree mortality, and
rates of dead wood decay is employed for estimating the pool of
woody debris in each age class. Literature studies (Vogt et al.
1986) provide the basis for estimating the C in the forest floor
and understory over the course of stand development. To arrive
at representative soil C pools for each forest type, a map of
soil C density (Kern 1994) was overlain with the spatial
distribution of each forest type (Eyre 1980). Soil C was assumed
to be stable over the course of stand development in accord with
the recent literature (Alban and Perala 1992; Johnson 1992).
c. Estimation of Carbon Flux
The net C flux was considered to be the net accumulation or loss
of C from the forestland base over a one-year period. The net
flux thus includes C transfer between the forest and the
atmosphere in either direction (i.e., growth, autotrophic
respiration, and heterotrophic respiration) and C transfer out of
the forest for human use. In an effort to isolate the effects of
harvesting, we distinguished between biologically driven C flux
and harvest-driven C flux.
d. Biologically-Driven Carbon Flux
The biologically driven C flux is an estimate of the net change
in C storage expected over a one-year period assuming there had
been no harvest. The physiological processes involved are
primarily photosynthesis and autotrophic plus heterotrophic
respiration. This flux for individual forest components and the
total flux, which reflects net changes in all the pools (i.e.,
64
-------
s 300-
0
20
40
60
80
100
Stand Age (year)
¦o—Tree • Woody debris —o— Forest floor —a— Understory
Figure IV.2. Pacific Northwest Region sytandard carbon budget
for medium productivity of Douglas-fir (Turner et
al. 1994a).
65
-------
net ecosystem production), were based on the inventories and
stand-level C budgets described earlier. Using the ATLAS model
for private lands and a spreadsheet approach for public lands,
the entire inventory was advanced one time step (5 years in the
southern regions, 10 years elsewhere) and flux was based on
annualized changes in pools over the time increment.
In order to derive an estimate of annual forest NPP, information
on wood growth was combined with literature data on rates of leaf
(fine litterfall) and fine root production. To obtain a ratio of
wood production to foliage production, data on these two
variables was extracted from 68 relevant studies in Cannell
(1982). The studies covered a range of stand ages and included
thirty-one coniferous stands and thirty-seven deciduous stands.
Average aboveground wood production from these studies (6.7
Mg/ha/yr) was multiplied by 1.15 to approximate inclusion of
coarse root production (Koch 1989). The average ratio of wood
production to foliage production was then 2.3 ± 1.4. Average
fine root production was estimated from average foliage produc-
tion using the regression of Nadelhoffer and Raich (1992), which
was developed from 59 studies in the literature wherein above-
ground fine litter production and fine root production were
measured. Thus the average foliage production of 3.5 ± 0.9
Mg/ha/yr was estimated to be associated with a fine root
production of 2.7 Mg/ha/yr. The final ratio of total production
to wood production was 2.0.
e. Harvest-Driven Carbon Flux
The volume of harvest removals or growing stock reduction for the
1990s for private lands was that used in the 1993 Resource
Planning Act (RPA) Assessment, and the distribution of the
harvest among regions and forest types was derived from the ATLAS
66
-------
model. The distribution of the removals among regions and
ownerships is similar to that reported in Powell et al. (1993).
Estimates of growing stock from public lands were projections for
the 1990s from the 1989 RPA Assessment. A harvest of 87 x 106
m3/yr was assumed here, which is higher than the estimate of 81
x 106 m3/yr for 1992 in Powell et al. (1993) but probably close
to what the decade average will be since harvest levels are
projected to increase. In order to estimate the area harvested
on public lands, the harvest volume within a region was distrib-
uted across forest types and age classes by generally assuming an
equivalence in the distribution of harvest across forest types
and age classes between public and private lands within a region.
In the forests of the Pacific Northwest region, data were used
from Sessions (1991) on the distribution of the harvest among
age-classes on public lands.
The tree C on harvested lands which was not accounted for as
growing stock removals or formation of woody debris residue
(calculated as the difference in the stand level C budgets
between post-harvest and pre-harvest woody debris on all
harvested areas) was assumed to be emitted within the year of
harvest. This residual could be considered to include slash
burns and perhaps additional material which might be removed from
the site but returned to the atmosphere relatively quickly (e.g.
fuelwood). The understory and part of the litter layer of
harvested stands were also assumed to be either burned or rapidly
decomposed, and this loss was likewise treated as instantaneous.
The. harvest in ATLAS included some lands which are cut but not
returned to the timberland base. In the 1993 RPA Assessment
projection, the average rate of reduction in the forestland base
in the 1990s was 201 x 103 ha/yr (Alig et al. 1990). All carbon
in woody debris on these lands was assumed to return to the
67
-------
atmosphere. This source of C was termed the "land conversion
flux."
f. Fire Emissions
A separate estimate of forest wildfire emissions is not in
principle necessary for this C budget. Mortality derived from
noncatastrophic wildfire is a component of the woody debris
budgets (Turner et al. 1993a). Catastrophic wildfires are
somewhat analogous to harvests when they are salvage logged which
is often the case. To the degree that this growing stock
contributes to the harvest targets and direct fire emissions
mimic slash emissions, these C fluxes are captured in our budget.
Direct wildfire C emissions appear to be on the order of 20 Tg/yr
(approximately 106 ha) , using a long term average (A. Auclair,
Science and Policy Associates, Washington D.C., pers. comm.,
1994). Approximately one-third of the wildfire emissions are
associated with woodlands, which tend to maintain a long-term C
equilibrium on those lands. Prescribed fire, including slash
burns, may generate an additional 10-20 TgC (Yamate 1974),
however, this flux is also captured in the modeling of the
harvest-driven C flux. Because fire was not treated explicitly,
formation of charcoal has not been considered in this analysis.
2. Woodlands
Woodland is defined as forested land with potential growing stock
accumulation rates of less than 1.4 m3/ha/yr (Waddell et al.
1989) e.g., the juniper (Juniperous occidentalis) woodlands of
eastern Oregon. For woodland phytomass, an average value of 13.0
Mg/ha from the compilation of Peterson (1993) was used. Woodland
NPP was based on an estimated average wood production of 0.7
m3/ha/yr (half the maximum accumulation rate), and the same ratio
68
-------
of total production to wood production was used here as on
forestland. Woodland NEP was assumed to be zero because climatic
factors and wildfire tend to keep TEC accumulation low. There
was no database from which to estimate woodland soil C density.
As a preliminary estimate, a value midway between the average for
forestland and rangeland soil C was assigned to woodland.
3. Rangeland, Pastureland, Cropland,
and Other Land
For non-wooded land, litter and woody debris pools were assumed
to be zero. The C pool for vegetation phytomass was assumed to
be the same as the NPP. The effect of that assumption is to
place an upper bound on the phytomass pool rather than an annual
average. The annual NPP was assumed to be balanced by annual
decomposition (i.e. NEP equal to zero) or fire. The average soil
C pool for nonforested land is based on soil C, by great group,
determined from the SCS National Soil Survey Laboratory Pedon
Database as calculated by Kern (1994), adjusted for soil depth
and rock fragment for the subsoil by data from the SCS Soil-5
database and rock fragment content for the surface horizon based
on the texture modifier from the 1982 National Resources
Inventory (NRI).
Net primary productivity for non-forested lands was estimated
using land-cover data from the 1982 NRI database (SCS 1987) and
several different NPP databases (King 1993b). Estimates of
rangeland aboveground net primary production (ANPP) were obtained
from the Soils-5 interpretation records (SCS 1983). Rangeland
ANPP was converted to* NPP by reference to ANPP/NPP ratios (-3:1)
derived from field studies in the literature. Pastureland NPP
for each state was estimated by assuming a NPP equivalent to the
average hayland total NPP for the same state. Cropland NPP was
obtained by converting crop yield data (SCS 1987) to estimates of
69
-------
total plant yield. "Other Land" NPP was assumed equal to the
average NPP of the other three general cover types.
C. Results
The total land of the conterminous U.S. is 765.6 X 106 ha, of
which 200.7 X 106 ha is timberland (Table IV.1). A check of
forest area derived from this inventory (243 X 106 ha) with
forest cover derived from a recent land-cover map based on
satellite remote sensing (252 X 106 ha) indicated reasonable
agreement (within 4%) at the national level (Turner et al.
1993b). At the state and regional levels, greater differences
were apparent. The area of rangeland is comparable to the sum of
forestland and woodland. Cropland and pastureland cover 154.7
and 54 X 106 ha respectively.
The average nonsoil (phytomass + CWD + litter) C density in
forestland (93.3 Mg/ha) is an order of magnitude greater than
that in the other land-cover types (2.3—13.0 Mg/ha, Table IV.1).
Forestland dominates the total in the phytomass C pool (Table
IV.2). Considering all C pools (87 Pg), the soil C component is
largest at 75% with phytomass C accounting for much of the
remainder.
Average NPP density is highest in cropland (Table IV.3),
reflecting subsidies of nutrients and protection from pests and
pathogens. Cropland also has the highest total NPP among the
vegetation types. Only forestland is considered to sequester C
on an annual basis in this analysis, and forest sequestration
(including phytomass,- CWD, and litter) amounted to 95 Tg/yr
(Table IV.4).
Detailed results by land-cover type are given below.
70
-------
Table IV.l. Area and Carbon Density by Land-Use Category,
United States
Area
C Density
( McrC/ha 1
Ecosystem/
Land Use
(Mha)
Phytomass Litter
Woody
debris
Soil
Forestland
200.7
63.1
11.6
18.6
91.1
Woodland*
41.2
13.0
NA
NA
78.6
Pastureland*
54.0
4.3
NA
NA
87.6
Range1and*
246.8
2.3
NA
NA
66.2
Cropland1
154.7
5.3
NA
NA
105.2
Other*
68.2
2.5
NA
NA
86.3
Total
765.6
"Peterson 1993.
fKing 1993a.
NA Not Applicable.
71
-------
Table IV.2. Total Carbon Pools for the Conterminous United
States
C
Pool (TaC^
Ecosystem/
Land Use
Phytomass
Litter
Woody
debris
Soil
Total
Forestland
12,663
2,324
3,730
18,280
36,997
Woodland*
1,550
NA
NA
3,238
4,788
Pastureland*
230
NA
NA
4,730
4,960
Rangeland1
578
NA
NA
16,3 38
16,916
Cropland*
815
NA
NA
16,274
17,089
Other*
171
NA
NA
5,886
6, 057
Total
16,007
2,324
3,730
64,746
86,807
*Peterson 1993.
fKing 1993b.
NA Not Applicable.
72
-------
Table IV.3. Total Net Primary Production for the
United States
Ecosystem/Land Use
Average
(MgC/ha/yr)
Total
(Tg/yr)
Forestland
3.46
695
Woodland
1.05
43
Pastureland
4.26
230
Rangeland
2.34
578
Cropland
5.29
815
Other
2.51
171
Total
NA
2,532
NA Not applicable.
73
-------
Table IV.4. Carbon Balance on the Forestland Base, United
States
Factors
Carbon (TgC/yr)
Tree
Woody
Debris
Litter Understory
Net
Biological*
359
-66
32 7
332
Harvest*
-275*
73
I
CO
CM
1
-237
Combined
84
7
4 0
95
'Biological and harvest refer to the net change in pool size as a
function of biologically-related or harvest-related factors.
includes growing stock removals (-125), woody debris formation
(-73) and harvest emissions (-77).
74
-------
1. Forestland
a. Carbon Pools
Half of the total forestland C is in the mineral soil (Table
IV.2). Phytomass C, which includes coarse roots and understory,
is the next largest component at 33%, followed by woody debris
(10%), litter (6%), and understory, (1%). The total C storage in
living trees on forestland in the U.S. is estimated at 12.2 Pg
(Table IV.2). The average storage for total phytomass C was 63.1
Mg/ha (Table IV.1). While direct comparisons are difficult, the
tree C pools appear to be consistent with earlier analyses based
on USDA FIA statistics (Cost et al. 1990, Birdsey 1992).
b. Biologically-Driven Flux
The estimated NPP for forestland was 695 TgC/yr (Table IV.3)
based on wood production derived from the inventory advancement
procedure and the ratios of fine root and le;af production to wood
production. The net flux of C into the forest from tree growth
(after accounting for mortality and turnover of foliage and fine
roots) was 359 TgC/yr (Table IV.4). Carbon also accumulated in
the understory (7 TgC/yr) and forest floor (32 TgC/yr) pools.
The only pool to show a net C emission (66 TgC/yr) was the woody
debris pool. This efflux was driven by the decay of woody debris
in the early to middle stages of stand development. During these
periods, the quantity of C 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
carbon inputs to the compartment originating in tree mortality
(Harmon 1993). The overall net flux of C driven by biological
processes was 332 TgC/yr moving from the atmosphere into forest
stands.
75
-------
Limited opportunities exist to validate these biological flux
estimates. Powell et al. (1993) estimated timber volume growth on
U.S. available timberland at 598 X 106 m3 for 1991 based on
repeated sampling of permanent plots. The comparable value in
the present analysis is 590 X 106 m3. The tree growth estimate
here (359 TgC/yr) compares closely with the estimate in Birdsey
(1992) .
c. Harvest-Driven Carbon Flux
The effect of timber harvest is manifested in several ways (Table
IV.4). Most significant is the transfer of a large quantity of C
out of the tree phytomass pool (275 Tg). In the U.S., this
reduction in tree C is about 80% of the annual C accumulation
from tree growth.
Of the 275 Tg total reduction in the phytomass C pool, 125 Tg was
associated with harvest and removal of the growing stock. The
remainder of the tree C on the harvested land (-150 Tg) was
partitioned between the harvest emissions (-77 Tg) and a transfer
to the woody debris pool (+73 Tg). Partitioning of the harvest
residue between formation of woody debris (49%) and short term
emissions (51%) is consistent with the observation that roots and
stumps generally remain after harvest.
The rapid change in forest floor and understory C associated with
or immediately following harvesting amounted to a 35 Tg/yr loss
of C, which brought the harvest emissions total to 112 TgC/yr
when combined with the 77 Tg of tree carbon (Table IV.4).
d. Complete Forestland Flux Analysis
The net effect of forest growth and decomposition was a transfer
of 332 TgC/yr from the atmosphere into organic matter. This rate
76
-------
of potential accumulation is possible only because the removal of
C by harvesting has tended to maintain a large area in the early
to mid-stages of stand development which favor C accumulation.
Under unmanaged conditions, the rate of C accumulation on the
same land base would be lower or even negative because fire and
increased rates of woody debris decomposition would balance
uptake. The potential accumulation of 332 TgC/yr was partially
offset by harvest-related removals from the forestland base
totaling 125 TgC/yr (Table IV.4) and harvest emissions of 112
TgC/yr (Table IV.4). The difference between the biological flux
and the losses associated with harvest is a gain of 95 TgC/yr on
the forestland base. A loss of 4 TgC/yr of woody debris C was
estimated for lands which were harvested but not returned to the
forest land base.
The live tree C pool is increasing because tree growth is greater
than the C in mortality and trees harvested. The simplicity of
the assumptions about woody debris dynamics lends a high degree
of uncertainty to the net flux estimate for woody debris (7
TgC/yr). Over the last several decades utilization of harvest
residues has tended to increase, and the inventory of older
forests carrying large quantities of woody debris has decreased.
At the same time, the average age-class of stands in the U.S. has
probably fallen because of increasing harvesting and management.
Younger stands carry more woody debris in the form of stumps and
harvest residues. These counteracting factors make it difficult
to assess the overall change in the storage of woody debris.
More recently, forest management regulations have begun to
require that some woody debris be left on-site for wildlife
habitat (Swanson and Franklin 1992), and that trend will also
favor woody debris accumulation.
The results here for net accumulation of tree C may be compared
with other recent national analyses. Heath and Birdsey (1993)
used current inventory data and inventory projections to estimate
77
-------
accumulation of live and dead C for timberland in the contermi-
nous U.S. The live C accumulation for the period 1987-2010 was
17 TgC/yr. That value does not include reserved timberland (11.7
X 106 ha), which is estimated here to sequester about 10 Tg/yr,
and does reflect the general trend of a decreasing tree C sink on
the forest land base over the coming decades (Turner et al.
1994b). Our estimate of tree C gain for the 1990s (84 Tg/yr) is
roughly consistent with the growing stock accumulation for 1991
reported in Powell et al. (1993). They indicated an annual
accumulation (net growth - removals) on available timberland in
the conterminous U.S. of about 140 x 106 m3 of growing stock
which represents about 34 Tg of growing stock or about 70 Tg of
whole tree C. Subak et al. (1993) similarly estimated the 1990 C
accumulation on the commercial forestland base at 60 TgC/yr based
on inventory statistics from the Food and Agriculture Organiza-
tion of the United Nations.
2. Woodland
Woodland represents 5.4% of the total area of the conterminous
U.S. but woodland NPP is only 1.7% of total NPP (Table IV.3).
Wood production within woodlands is about one-third of NPP, so it
amounts to 15 Tg. Our earlier estimate of emissions from
woodland wildfire was approximately 7 TgC/yr, thus the remainder
(8 TgC/yr) is assumed to be balanced by decomposition such that
NEP for woodland as a whole is zero.
3. Grassland, Rangeland, Pastureland,
Cropland and Other Land
Non-forested land constitutes about 70% of the land surface of
the conterminous United States. Rangeland is the dominant non-
forest cover type, accounting for about half of the non-forested
land. Cropland is the most productive cover type, with a mean
78
-------
NPP of 5.3 MgC/ha/yr, while rangeland is the least productive,
with a mean NPP of 2.3 MgC/ha/yr (Table IV.3). Overall, NPP for
non-forested land in the U.S. is estimated to vary by about 20%
between favorable and unfavorable years, or about 0.7 PgC (King
1993b).
Normal tillage practices in croplands of temperate zones tend to
decrease soil C until a stable equilibrium is reached. Much of
the agricultural land in the U.S. has reached this equilibrium
but a small C source (1.5-2.6 Tg/yr) is still indicated (Kern and
Johnson 1991).
D. Discussion
NPP is by far the largest C flux among the processes evaluated in
this analysis. At 2,532 Tg/yr (Table IV.3), C uptake via NPP is
nearly twice C release via fossil fuel emissions (Table IV.5).
However, uncertainty about the magnitude of NPP is large and the
great majority of NPP is returned to the atmosphere relatively
rapidly. Only 95 Tg/yr of NPP accumulates on the forestland
¦\
base. Another 36 Tg/yr accumulates in forest products still in
use (Birdsey et al 1993; Turner et al 1993a), and 10 Tg/yr in
landfills (Subak et al. 1993). C emissions from land conversion
and agricultural tillage (Table IV.5) are relatively small and
are to some degree counteracted by factors untreated here such as
urban forestation (Rowntree and Nowak 1991). Thus, in terms of
V
net flux, fossil fuel emissions are an order of magnitude larger
than the biologically based net C sinks. The accumulation of C
on the forestland base of the U.S. is similar to trends in many
northern temperate zone countries (Clawson 1979; Armentano and
Ralston 1980; Sedjo 1992; Kolchugina and Vinson 1993d; Kauppi et
al. 1992; Birdsey et al. 1993). Recovery from earlier periods of
extensive forest harvest and limited management is now resulting
in C accumulation. Specific factors in the U.S. include
79
-------
Table IV.5.
Net Carbon Flux for the Conterminous
United States*
Source
Energy
Cement Manufacture
Forest Sector
Forest Land Base
Forest Products
Conversion Emissions
Agricultural Sector
Agricultural Land Base
Landfills
Carbon (TgC/yr)
-1,296*
-10f
95
36*
-4
-2s
10*
Net
-1,171
'Emission of CO, methane, and other hydrocarbons are
excluded. Negative values are carbon transfers to
the atmosphere.
~Subak et al. 1993.
*Birdsey et al. 1993.
JKern and Johnson 1991.
80
-------
reversion of agricultural land to forest (Williams 1988) and fire
suppression (USDA 1992a). The combined effects of these factors
in northern temperate zone forests is believed to be an accumula-
tion of 700 Tg/yr (Sedjo 1992). A terrestrial sink of this
magnitude is needed to balance the global C budget (Post et al.
1990; Siegenthaler and Sarmiento 1993).
Several studies have examined.the possibility of significantly
increasing the sink on the forestland base. Suggestions have
included increased paper recycling, reduction in the harvest on
public land, and afforestation (Turner et al. 1993a, Turner et
al. 1994b; Winnett et al. 1993). An increase in paper recycling
to 48% by the year 2000 was estimated to increase sequestration
on the forestland base by about 5 Tg/yr over a 50 year scenario
(Turner et al. 1993a). The harvest reduction scenario in Turner
et al. (1993a) called for a 20% reduction in the harvest on
national forestland (3.4% of the total U.S. harvest). Isolating
the public land base, the effect of the harvest reduction was an
additional C sink on the order of several Tg/yr. The harvest on
private.land could potentially rise in response and negate the
sink; however, the forest economic model used in these scenarios
did not indicate such a response in this case (Figure IV.2).
Likelihood confounding factors include potential changes in
imports and exports of logs and forest products.
The potential for C sequestration from afforestation of marginal
pasturelands and croplands is much larger (Moulton and Richards
1990; Sampson and Hair 1992; Barker et al. [submitted]). As much
as 47 X 106 ha of such land could be converted to forest (Parks
et al. 1992), and the. C sink associated with a land conversion of
this magnitude would sequester a significant fraction of U.S.
fossil fuel emissions. The costs of establishing the forests and
the long-term land rents are substantial, on the order of
billions of dollars per year. A scenario in which $220 million
per year for 10 years was invested in afforestation (5 X 106 ha)
81
-------
according to the scheme of Moulton and Richards (1990) resulted
in an average C sink of 15 Tg/yr over a 50 year scenario (Turner
et al. 1993a). Land rent costs beyond the initial 10 years were
not accounted for, and these might raise costs significantly.
82
-------
Box 3i Mexico
Mexico straddles the tropic of
Cancer, extending from 33*N to 14*N
latitude. This nation of approxi-
mately 196 X 106 ha comprises 31
states and the Federal District of
Mexico, the capital. The 1990
population of Mexico was 88.6 mil-
lion people (the largest Spanish
speaking country in the world),
corresponding to a population den-
sity of 452 people/1000 ha (WRI
1993). However, greater than 70% of
this population lives in urban
areas, with Mexico City considered
to be the largest city in the world
(1990 population approximately 20
million people). The population
growth rate of Mexico, 2.2%/yr
between 1985 and 1990 (WRI. 1993), is
similar to that of Brazil.
Highly organized indigenous civili-
zations in Mexico today date back at
least to 1200 B.C. with the rise of
the Olmec civilization in eastern
Mexico!, Spain began the European
conquest of Mexico in 1519, and
today the Spanish language is the
predominant language in most regions
of the country. Indigenous popula-
tions exist throughout the country,
though the Mayan populations in
southern Mexico are perhaps the
largest. The Mayans are only par-
tially assimilated into the dominant
Latin culture.
Mexico is a rapidly industrializing
country. In Latin America, Mexico
is second only to Brazil in popu-
lation and GNP, yet it exceeds
Brazil in commercial energy produc-
tion and estimated C emissions from
industrial sources (WRI 1993). At
the same time, FAO (1993) estimates
that Mexico had the second highest
amount of deforestation in Latin
America during the 1980's, though
for the humid tropical forest type
it is less than one fifth of that
deforested in Brazil during the same
time period.
A broad diversity of tropical and
temperate ecosystems exists in
Mexico (Figure V.l) due to its
position at the junction of the
neotropical and holarctic zones, and
its extensive vertical relief.
Drylands and temperate forests of
coniferouB and broadleaf trees in
the northern highlands grade into
tropical deciduous and evergreen
ecosystems in the southern lowland
areas. Mangrove forests and wet-
lands exist along both the Atlantic
and Pacific coasts.
Once extensively forested, many
Mexican forests, especially the
tropical forests, have been con-
verted to pastoral and agricultural
land uses. Rzedowski (1978) esti-
mated that the current extent of
tropical evergreen forests repre-
sents approximately 10% of the
extent of these forests prior to
human-induced land conversion. In
some areas, deforestation occurs
within the context of various tradi-
tional systems of shifting agricul-
ture involving forest-fallow (Ita-
Martinez et al., Universidad Nacion-
al Autonoma de Mexico, unpublishedi
manuscript; Bellon 1991). In other
areas, government land tenure or
agricultural subsidy policies have
influenced forestland conversion
(Ramos and del Amo 1992; Williams-
Linera 1983). Recent land tenure
reforms were enacted to slow forest
conversion and to promote sustain-
able use and restoration of Mexican
forest resources (Masera et al.
1995.
The wood products industry of Mexico
is small relative to the national
economy, and Mexico is not self-
sufficient in terms of meeting
national wood products demand (CNIF
1992). Although roundwood produc-
tion has historically exceeded
demand, production fell short of
consumption by 10% in 1991. Mexican
production of wood fiber products
meets less than 50% of its national
demand (CNIF 1992).
83
-------
CO
fpo^£STASERVic^
u4s<
2 ^ Si
Temperate forest
Tall and medium statured
tropical evergreen forest
Low statured tropical
deciduous forest
Non-wooded
Water
Clouds
INEGI
i' itn i" if n fin i • l s I HII I \ F. tp I I s Fy > (I F / l
»• - ti o fi i r » i ' ii r» i n r> f If/
Figure v.l. Forest map of Mexico produced by classification of composite 1990 and 1991
AVHRR images.
-------
V. Mexico Case Study
A. Introduction
A C budget for Mexico, as for other nations, requires information
on four broad elements: (1) an estimate of land cover; (2)
transition rates between land-cover classes (e.g., forest
conversion to pastureland); (3) carbon density estimates of each
land-cover component; and (4) estimates of the effects of
land-use changes on C pools. Published C budgets for Mexico,
focusing on forest ecosystems, have brought together much of the
published and unpublished data collected in Mexican forest
ecosystems (Masera et al. 1992). Though a slightly different
approach is used, this section draws upon that study and other
sources of data to model forest C. This section extends C budget
analysis to all land-cover types and incorporates recently
available land-use and vegetation information.
B. Methods
The types of land use in Mexico can be divided into six broad
categories (Table V.l). Approximately 36% of Mexico is classi-
fied as forestland. This category is composed of both fully-
stocked forests (greater than 5 m in height, crown cover not
specified) and what Secretaria de Agricultura y Recursos
Hidr&ulicos labeled perturbed forests (SARH 1992), though the
interpretation of this category is uncertain. The fully-stocked
forest category includes both temperate and tropical forests.
Temperate forests include coniferous (9% of total area of Mexico)
and broadleaf forests (4%). Tropical forests include humid
tropical (5%), dry tropical (7%) and mangrove forests (< 1%).
Perturbed forests (11% of total area) are not distinguished by
vegetation type (SARH 1992; Table V.2).
85
-------
Table V.l. Areas of Major Land-Cover
Types in Mexico*
Land-Cover Type
Area
X 10e
' ha
%
Forest*
71.
.28
36
Woodland/scrub*
66.
.42
34
Pastureland'
38.
,37
20
Cropland*
13.
,14
7
Wetland1
1.
,20
1
Mangrove forest*
0.
,53
<1
Other*
4.
,88
2
Totalf
195.
,82
100
Based on SARH (1992) which is the most
recent and comprehensive survey of
land use in Mexico to date.
* Developed from multidate composites of
NOAA-11 satellite AVHRR data and ancil-
lary maps (Evans et al. 1992). Uncer-
tainty levels for 11 of estimates area
are unknown.
* Masera et al. (1992) provide estimates of
the proportion of non-forest lands in
cropland and pastureland based in part
on data from SARH (1992). These propor-
tions were used in this section to
estimate the area of cropland (13.14 X
106 ha) and pastureland (38.37 X 106 ha)
for the purpose of modeling standing
stock of C.
86
-------
After forests, the next largest land area is composed of
woodland/scrub (34% of total area of Mexico) (Table V.l). This
class includes matorral or scrub (28%), chaparral (4%), mesquite
(1%), and miscellaneous shrubland covers. Other land-cover
classes include agricultural/pastoral (27%), wetlands (1%),
desert (1%), water (1%), and urban (< 1%). The
agricultural/pastoral land use class was not distinguished
according to previous land cover (SARH 1992). In this section,
the agricultural/pastoral land-cover class was divided into
cropland and pastoral classes according to percentage estimates
used by Masera et al. (1992). Thus, cropland is estimated to
represent 7% of the total land cover of Mexico and pastureland
20% (Table V.l).
Estimates of the area, TEC, and rate of conversion of each land-
use type were entered into a spreadsheet model. Carbon fluxes
from past land-use changes (i.e., regrowth and continuing
decomposition of slash and soil organic matter on sites converted
from forest within the past 10 years) are incorporated as
"inherited emissions" (i.e., C released to the atmosphere
predominantly due to phytomass burning and, to a small degree,
from decomposition of wood products or soil C). These were added
to base year, "prompt emissions" to calculate total C emissions.
This follows the convention of Makundi et al. (1992). The
standing C stocks and fluxes were estimated for a nominal base
year of 1990.
1. Land Cover
Land-cover change is extremely dynamic in some regions of Mexico
(e.g., 7.7% annual deforestation rate from 1974-1986 in Chiapas;
Cuar6n 1991). The first Mexican National Forest Inventory (SARH
1986), completed in 1985 after a 24 year study, was largely out
of date by the time it became available. Masera et al. (1992)
87
-------
Table V.2. Estimates of Land Cover and Annual Deforestation Rates in Mexico
Total
(X 106
Area
ha)
Annual
Deforestation
(X 106 ha)
Land-Cover Type
SARH
(1992)*
Masera
et al.
(1992) *
Masera et al.
(1992)t
Temperate conifer forest
17.0
16.9
0.108
Temperate broadleaf forest
8.6
8.8
0.059
Humid tropical forest
10.5
9.7
0.195
Dry tropical forest
13.7
16.1
0.306
Total closed forest
49.8
51.5
0.668
Open forest
Woodland/scrub'
Perturbed forest
ND
66.4
21. 611
31.1
ND
ND
NAS
ND
ND
Other forest (mangrove)
0. 5
ND
ND
Total other forest
88.5
31.1
NA
Pastureland
38.4#
79.9*
ND
Cropland
13.1*
27.4*
ND
Wetland
1.2
ND
ND
Total non-forest
52. 7*
107.3*
ND
Other
4.9
6.7
ND
Total Land Area
195.8
196.6
0.668*
88
-------
Table V.2 (Continued)
*Land cover estimates were derived from analysis of AVHRR imagery and
existing national forest inventory information.
tLand cover estimates were derived from various data sources reported
by Masera et al. (1992) including earlier SARH estimates.
'This class includes matorral (55.5 X 106 ha), mesquite (2.8 X 106 ha)
chaparral (7.7 X 106 ha), and miscellaneous shrublands (0.4 X
106 ha).
'Not available. Masera et al. (1992) declined to include open forests
in their C analyses due to a lack of data on deforestation for
this class and the low C density associated with its vegetation.
*SARH (1992) includes moderately perturbed (18.1 X 10* ha) and
severely perturbed forests (3.5 X 106 ha) in this category.
*Masera et al. (1992) provide estimates of the proportion of non-
forest lands in cropland and pastureland based in part on data
from SARH (1992). These proportions were used in this section
to estimate the area of cropland (-13 X 106 ha) and pastureland
(-38 X 10s ha) (Table V.l) for the purpose of modeling standing
stocks of C.
*Area-weighted mean of deforestation rates for all forest zones of
closed forests as defined by Masera et al. (1992).
NA No data available.
ND No datum.
89
-------
drew upon these and other data to estimate 51.5 X 106 ha of
closed forest and 31.1 X 106 ha of open forest (Table V.2).
A national forest inventory was recently completed (Sorani et al.
1993). The preliminary estimates of land cover used to stratify
the current Periodic National Forest Inventory include land area
under non-forest land uses (SARH 1992; Evans et al. 1992). This
analysis is based on interpretation of 1-km AVHRR imagery from
1990-1991 combined with existing forest inventory data.
Estimates of closed forest cover by SARH (1992) differ only
slightly from those used by Masera et al. (1992) (Table V.2).
The land-cover estimates of Masera et al. (1992) and SARH (1992)
diverge, however, in estimating the proportion of land dedicated
to open forest, pastureland, and cropland. Masera et al. (1992)
use an estimate of pastureland and cropland area (107.3 X 106 ha)
which is double that used by SARH (1992) (51.5 X 106 ha) with a
consequently smaller area labelled as open forest (Table V.2).
The difference is due in part, to differences in the definition
of forest cover. Many areas used for pastureland in Mexico hold
significant phytomass in woody vegetation and trees either as
isolated individuals or as live fencing (Guevera et al. 1992).
Some of these areas may be included in the SARH (1992) estimate
of woodland/scrub forest or of perturbed forest. This section
uses the area estimates of SARH (1992) since that data is based
on the most recent and most comprehensive survey of land use in
Mexico to date. The differences in area estimates of the
woodland/scrub and forest categories, however, will have an
effect on estimates of C pools and on C flux. A significant
portion of the perturbed forest area estimated by SARH (1992) may
have potential for sequestering C through forest growth even if
these lands are not highly productive.
Masera et al. (1992) further provide estimates of the proportions
of deforested areas which are converted to pastureland or
90
-------
cropland (i.e., 79.9 X 106 and 27.4. X 106 ha; Table V.2). These
same proportions (75% and 25%) are used in this section to
estimate the proportion of pastureland or cropland converted from
the humid and dry tropical forest types during the last 10 years
(i.e., 0.195 and 0.306 X 106 ha/yr, respectively, from Table
V.2). Therefore, this section estimates that 3.9 X 106 ha of
forestland has been converted to pastureland and 1.1 X 106 ha has
been converted to cropland during the past 10 years.
The main elements contributing to the net flux of C in the
woodland/scrub category are vegetative growth which fixes C, and
phytomass removals by pastoral and woodcutting activities which
cause C losses. Since definitive data are not available for
these processes, this section considers the woodland/scrub
category to be in approximate C equilibrium, with vegetation
growth balanced by pastoral and woodcutting phytomass removals.
For perturbed forests, Eggen-Mclntosh (USDA Forest Service, pers.
comm., 1994), interprets this SASH (1992) category as being
fragmented forest roughly equivalent to the 10-40% forest cover
class of FAO (1993). Much of this land is continually perturbed
through repeated cultivation. Nasera et al. (1995) also
considered the perturbed forest category to include many areas
that have long sustained shifting cultivation. They state that
forest productivity on these lands is low and that forest
plantations are generally not commercially viable.
For this section, there was insufficient information to make a
definitive estimate of the C dynamics of perturbed forest
systems; therefore, two possible scenarios were evaluated. In
the first scenario, only the area equivalent to those lands that
had been recently (during the last 10 years) deforested through
timber harvesting or wildfires (about 13.5%; SARH 1992) are
modeled as having net C sequestration. In the alternative
scenario, fully 50% of the moderately disturbed category (SARH
91
-------
1992) is modeled with a low net C sequestration; the other 50% is
presumed to be in rough C equilibrium as continuing human
perturbations balance C sequestration. These scenarios are
extremely tenuous; from the viewpoint of C sequestration they are
likely conservative since fragmented forests are widespread
throughout this land-cover type. The assumptions underlying
these scenarios, however, dramatically affect the final emissions
estimate for Mexico.
2. Carbon Density Estimates
There are few published TEC data for Mexican ecosystems. Masera
et al. (1992) used phytomass estimates generated from five case
studies, one of which involved the direct measurement of forest
phytomass. Direct measurements of forest TEC pools have been
published for only a few sites in Mexico (Williams-Linera 1983;
Wadsworth et al. 1988; Castellanos et al. 1991; Martinez-Yrizar
et al. 1992).
For this section, estimates of Olson et al. (1983, 1985) were
used for vegetation C content (Table V.3) because these are some
of the only published estimates commonly in use for all Mexican
vegetation types present in Mexico. It is recognized that these
estimates are based on a global database and may be high for
Mexico (J.S. Olson, Global Patterns Company, Lenoir City, TN,
pers. comm., 1993). To account for this, the low estimates
published in Olson et al. (1983) or the revised values in Olson
et al. (1985) were used in this section (Table V.3). Olson
phytomass values were converted to above- and belowground
phytomass pools according to published conversion factors using
root/shoot ratios (Table V.4).
92
-------
Table V.3. Phytomass Carbon Density (above- and belowground) in Mexico by
Land-Cover Type
Phytomass (MgC/ha)
Land-Cover Type
Olson et al.*
(1983, 1985)
Masera et al.
(1992)
Others
Tropical humid
forest
150
144
139-265*
Tropical dry
forest
50
67.58
60*
52*
Temperate coniferous
forest
60
56
ND
Temperate broadleaf
forest
50
39
ND
Wetland
1.5
ND
ND
Mangrove forest
30
ND
ND
Grassland/shrub
20
ND
ND
Desert
2
ND
ND
Cropland
6
ND
ND
Pastureland
5
ND
ND
Perturbed forest
40
ND
ND
'This Section uses data of Olson (1983, 1985) for C density estimates.
+Mean value of four plots calculated from data provided by M. Ricker
and M. Quinlan (Yale, 1993, personal communication) using the phytomass
equations of Brown et al. (1989) and a root/shoot ratio of 0.2.
*Martinez-Yrizar et al. (1992).
'This value is based on Martinez-Yrizar et al. (1992), but Masera et al. (1992) use
a root/shoot value of 0.59. The root/shoot value of 0.4 measured by
Castellanos et al. (1991) within the same forest type and natural area is used
here.
'Castellanos et al. (1991).
ND Ho datum.
93
-------
Table V.4. Total Area, Flux Parameters, and Estimates Used to Compute Net Carbon Balance for Mexico
Land-Cover
Type Parameter Estimate Source/Basis
Forest
Root/shoot ratios
humid tropical
dry tropical
coniferous
broadleaf
mangrove forest
Proportion of litter & coarse
woody debris C vs. phytomass C:
humid tropical
dry tropical
coniferous
broadleaf
mangrove forest
C dynamics during fires:
% aboveground C to charcoal
% aboveground C released
humid tropical
dry tropical
coniferous
' broadleaf
% litter released
% soil C released
0.20
0.40
0.20
0.20
0.73
Saldarriaga (1985); Masera et al.
(1992); Castellanos et al. (1991)
Cannell (1982)
Cannell (1982)
Cannell (1982)
0.14
0.04
0.04
0.04
0
Salati et al. (1991); Uhl et al.
(1988); Uhl and Kauffman (1990)
Assumed
Assumed
Assumed
3.6
20
40*
40
40
95
5
Fearnside (1992); Houghton et al.
(1991a)
Masera et al. (1992)
Masera et al. (1992)
Masera et al. (1992)
Masera et al. (1992)
Assumed
Cerri et al. (1991); Sanchez et al.
(1983)
94
continued
-------
Table V.4 (continued)
Land-Cover
Type
Parameter
Perturbed
forest
Root/shoot ratio
Proportion of litter & coarse
woody debris C vs. phytomass C
Net vegetation C sequestration
rate (MgC/ha/yr)
Estimate
Source/Basis
0.2
0.33
Saldarriaga (1985)
Saldarriaga (1985); Brown and Lugo
(1990)
1.0
Assumed*
Grassland/ Root/shoot ratio 1.6 Singh and Joshi (1979)
shrub Litter C mass (MgC/ha) 2.0 Kauffmann et al. (1994b)
Pastureland
Root/shoot ratio
Litter & coarse woody debris C
mass (MgC/ha)
1.6
2.0
Same as grassland/shrub
Kauffman et al. (1994b)
C decomposition post-conversiont
decomposition period
(MgC/ha)
% released/yr
% soil C released (total)
10
10
20
Feamside (1992); Houghton et al.
(1991a)
Assumed
Houghton et al. (1991a)
95
continued
-------
Table V.4 (continued)
Land-Cover
Type
Parameter
Estimate
Source/Basis
0.18
0
10
10
30
Sanchez et al. (1989)
Assumed
Fearnside (1992)
Assumed
Houghton et al. (1991a); Davidson
and Ackernan (1993)
0.73
0
Assumed similar to mangrove forest
Assumed
0.23
0
Cannell (1982)
Assumed
Soil and belowground phytomass were
not modeled. These ecosystems were
assumed to be in carbon equilibrium
for lack of data.
Cropland
Root/shoot ratio
Litter & coarse woody debris C
C decomposition post-conversion:
decomposition period - yr
% released/yr
% soil C released (total)
Wetland
Root/shoot ratio
Proportion of litter & coarse
woody debris C vs. phytomass C
Desert
Root/shoot ratio
Litter & coarse woody debris C
Urban,
water
96
continued
-------
Table V.4 (continued)
'Gonzalez Floras (1992) reports consumption rates of approximately 60% for two fires in tropical dry forest
in Jalisco, Mexico.
*As discussed in the text, the severely perturbed forest area is assumed to be in C equilibrium. Within
the moderately perturbed forest area, two scenarios are used. Scenario 1: most of this category is in
approximate C equilibrium with only a portion equivalent to the land area of forests recently (last 10
years) deforested through timber harvest or wildfires supporting a net C sequestration. Scenario 2: 50% of
the perturbed forest area is presumed to be in carbon equilibrium with continued human disturbance balancing
vegetation regrowth. The sequestration rate of 1 MgC/ha/yr applies to the remaining 50% of perturbed forest
area.
97
-------
For estimates of soil organic matter C in forest systems, this
section relied on global estimates of soil C from Zinke et al.
(1984) (Table V.5). Zinke et al. (1984), similarly to Olson et
al. (1983), used the globally available estimates of soil carbon
and estimated soil C for each continent and the globe in general.
The global averages of soil C estimated by Zinke et al. (1984)
are similar to the estimates of Kern (1994) and lower than
Sombroek et al. (1993). The estimates used by Masera et al.
(1992) appear to be low by approximately a factor of two relative
to other global estimates.
3. Transition Rates
Obtaining good estimates of the rates of land-cover conversion,
especially deforestation, is perhaps the greatest difficulty in
estimating C pools and flux in Mexico. The rate of forest
conversion in recent decades has most likely not been constant as
many factors, including government resettlement and land tenure
policies, have served to impede or promote forest clearing
(Masera et al. 1995).
For the main analyses of this section, the closed forest
deforestation estimate by Masera et al. (1992) is used (0.668 X
106 ha/year) (Table V.2). The FAO (1993) estimate of 0.678 X 106
ha/year is similar. A direct estimate of the deforestation rate
is not available for the perturbed or open forest category
(Masera et al. 1992; SARH 1992).
4. Carbon Flux Estimates
To estimate C flux in response to land-use change, parameters
were taken from the literature. Where information was not
available, assumptions were made based upon the best available
knowledge. In some cases, these assumptions are distinctly
98
-------
Table V.5. Carbon in Mexican Soils by Land-Cover Type*
Carbon in Soil (Depth to 1 m) (MgC/ha)
Zinke et al. Masera et al.
Land-Cover Type (1984) (1992) Other
Tropical humid 109*
forests 104 66 145*
Tropical dry forest 112 29.5 105'
Temperate coniferous
forest 132 109.1 102*
Temperate broadleaf
forest 161 29.5 102*
Wetland 234 ND ND
Mangrove forest 234 ND ND
Shrub/grassland 77 ND ND
Desert 25 ND ND
Cropland 75 ND ND
Pastureland 75 ND ND
Perturbed Forest 94 ND ND
*This section uses data of Zinke et al. (1984) for soil C data.
'Generated by overlaying the forest cover map of Evans et al. (1992) and
the map by Kern (1994) then calculating an area-weighted mean carbon
density value.
'Sombroek et al. (1993).
ND No datum
99
-------
different than those used by Masera et al. (1992) and account for
some of the difference in net C emissions between this analysis
and Masera et al (1992). Estimates of net terrestrial C flux for
Mexico are especially sensitive to certain assumptions made about
(1) C uptake by phytomass growth and accumulation in soils and
(2) C loss through decomposition and fire (Table V.4). Further,
these assumptions are based upon assumptions about the character
of vegetation and human perturbation in the land-cover types
used. Most of the land area of Mexico is assumed to be in
approximate C equilibrium with losses due to fire, harvesting,
and decomposition balanced by vegetative growth. The exceptions
are recently deforested areas (defined as forests converted to
pastureland or cropland within the last 10 years [Table V.2]) and
perturbed forests. In both scenarios for the extent of perturbed
lands (see Section V.B.I. Land Cover), a net C sequestration rate
of l.o Mg/ha/yr (Table V.4) was used, with the assumption that C
sequestration will be low on most perturbed forestlands because
of continuing human and natural disturbance and low site
productivity (Masera et al. 1995). Humid forest regrowth
following disturbance in the region of Uxpanapa, Veracruz, has
been measured at approximately 3.75 MgC/ha/yr (Williams-Linera
1983) during the first seven years following abandonment. This
is most likely an upper limit to C sequestration in perturbed
forests (Masera et al. 1995). The perturbed forest category of
SARH (1992) is not broken out into forest types, so it is
difficult to determine the proportion represented by productive
humid forest lands. In addition, much of the pastureland and
cropland in Uxpanapa has been abandoned (Williams-Linera 1983),
and human disturbance pressures in these forests are probably
lower than in perturbed forests elsewhere.
Masera et al. (1992) assume that 2 0-40% of aboveground phytomass
C is carbonized during land-clearing fires, with the remainder
either passing to the atmosphere immediately through combustion
or slowly through decomposition. An alternative estimate of 4%
100
-------
or less has been suggested (Table V.4) (Houghton et al. 1991a;
Fearnslde 1992; Kauffman et al. 1994a). Although values higher
than 4% nay be reasonable for areas that are repeatedly burned,
values as high as 40% are probably extreme. This section uses an
estimate of 8% to reflect the lower estimate of carbonization
through repeated fires.
Data to estimate long-term C storage in wood products and
landfills are not available for Mexico. Due to this lack of
information, this section makes no estimate of C storage in
landfills, and C storage in wood products is estimated only for
current year harvest levels. Clearly, ignoring C storage in
landfills and wood products will lead to an underestimate of
total C storage in Mexico. For current year withdrawals,
estimates of harvest volume for Mexico were taken from three
sources (UNAM 1991; Masera et al. 1995; WRI 1993). The propor-
tion of wood products from each of four forest types (conifer,
broadleaf, humid tropical, and dry tropical) are given by UNAM
(1991). The volume of wood from each forest type was converted
to mass of C using wood density estimates of 0.5 for conifer
species and 0.6 for all others (Chudnoff 1984; Masera et al.
1992).
C. Results
An estimated 54% of the total TEC of Mexico is in forests (Table
V.6), which cover 36% of the total land area (Table V.l).
Temperate coniferous forests are the largest reservoir at -13% of
the country's TEC, followed by humid tropical forests with -11%,
dry tropical forest with -9%, and broadleaf temperate forests
with -7%. Although perturbed forests have been assigned
relatively low C density values, they are estimated to contain
-13% of the TEC of Mexico due to their large area.
101
-------
Table V.6. Carbon Pools of Mexico by Land-Cover Type, 1990*
Carbon Pool (PgC)
Phytomass
Land-Cover
Type
Total
Above-
ground*
Below-
ground
Litter
/ CWD
Soil
Total
Forest
4.56*
3.72
0.84
0.50
8.44
13.50
Woodland/
scrub
1.33
0.51
0.82
0.13
5.11
6.57
Pasture-
land
0.19
0.07
0.12
0.14
3.00
3.33
Cropland
0.08
0.068
0.012
0.015
1.02
1.115
Wetland
0.002
0.001
0.001
0
0.29
0.292
Mangrove
forest
0.015
0.009
0.006
0
0.12
0.135
Other
0.008
0.007
0.001
0
0.097
0.105
Total
6.185
4.385
1.800
0.785
18.077
25.047
* Values developed for this Mexico case study based upon Tables V.l through
V. 5.
t Forest C pool was computed by multiplying 71.28 X 106 ha (Table V.l) times
the area-weighted mean C density, 64 MgC/ha, based upon values in
Tables V.2 and V.3 and appearing in Table V.7.
* Aboveground phytomass values were determined by partitioning the total
phytomass values with the root/shoot ratios in Table V.3, e.g. Mangrove
forest: (0.015 PgC) + (0.73 roots + 1.0 shoots [Table V.3]) = 0.009 PgC
in aboveground phytomass.
102
-------
Woodland/scrub ecosystems are estimated to hold only ~26% of the
national C pool (Table V.6) although this ecosystem is approxi-
mately equivalent in land area to that of forests (Table V.l).
Pastureland is estimated to contain -13% of TEC, primarily as
soil organic matter, due largely to the extensive area of this
land-cover type. As previously discussed, much of the perturbed
forest and woodland/scrub categories are presumed to include
areas used for pastureland but which carry significant woody
phytomass. Croplands, which are assumed to lose more soil
organic matter than pasturelands after conversion from forest
(Brown and Lugo 1990; Houghton et al. 1991a; Davidson and
Ackerman 1993) and which represent -7% of the area of Mexico,
contain an estimated 4% of the national TEC.
The area-weighted mean C density for phytomass, litter/CWD and
soil in each major land-cover type (Table V.l) was calculated
(Table V.7). Although forest C densities for phytomass are
estimated to range as high as 150 Mg/ha in tropical humid forest
(Olson et al. 1983), the much larger areas of forests in Mexico
with relatively low phytomass C density (conifer, broadleaf, and
dry tropical) and perturbed forests dominate the forest category
(i.e. resulting in an area-weighted mean of 64 Mg/ha, Table V.7).
If perturbed forests are not included in this calculation, the
area-weighted mean phytomass C density of forests is 74 Mg/ha.
Most of the land surface of Mexico is estimated to be in C
equilibrium, though this is a tenuous assumption (Lugo and Brown
1992; Brown et al. 1992). Fluxes of C were assigned (Table V.8)
to (1) land-cover conversion (e.g., fires associated with
deforestation); (2) long-term losses due to decomposition after
land conversion (e.g., slash or soil organic matter decomposi-
tion); and (3) C sequestration through regrowth of a portion of
the perturbed forest and pastureland classes. The total annual C
flux to the atmosphere from deforestation was estimated at 75.8
103
-------
Table V.7. Area-Weighted Mean Carbon Densities by Major Land-Cover Type in
Mexico*
Carbon Density (MgC/ha)
Land-Cover Type
Forest
Woodland/scrub
Pastureland
Cropland
Wetland
Mangrove forest
Other
Total Area-
Weighted Mean
Phytomass
64
20
5
6
2
30
14
32
Litter/
CWD
7
2
4
1
0
0
0
Soil
118
77
78
78
234
234
20
92
Total
189
99
87
85
236
266
34
128
Values developed from Tables V.l through V.5.
104
-------
Table V.8. Carbon Flux for Mexico*
Reported C Flux (TgC/yr)
(base year]
Component
This section Makundi Subak et WRI
et al. al. (1993) (1993)
(1992)
[1990] [1985] [1985-1988] [1989]
Land-base flux
Phytomass
Litter and CWD
Soil
Total land-base flux
Land-use change
Removals
Emissions:
Prompt
Inherited
Total land-use change
+4.0 to +12.0
+0.2
+0.2
+4.4 to +12.4
NC
-30.8
-45.0
-75.8
+3.5*
ND
ND
ND
-8.6*
-27.9
-26.1
-62.6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Net flux of C between
terrestrial ecosystems
and the atmosphere9
Fossil fuels
Fossil fuel as C02-C
Cement manufacture C
-63.4 to
-71.4
NC
NC
-59.1
-83.1?
-3.2s
-55.9
-74.4
-3.1
-60.9
-93.7
-3.6
Total industrial
emissions
NC
-86.3
-77.5
-97.3
Total nation
-149.7 to -157.7*
-145.4
-133.4
-158.2
'Positive values are C uptake and negative values are C emission to the
atmosphere.
*For forests only (Makundi et al. 1992).
Conversion of Makundi et al. (1992) wood removal values (lumber, paper,
posts, and fuelwood) to equivalent phytomass C quantities.
'From Masera et al. (1992). '
*The "total nation" values for "this section" utilizes the Makundi et al.
(1992) values for total industrial emissions of -86.3 TgC/yr; values from
Subak et al. (1993) and WRI (1993) could also be used resulting in a wider
range of national totals.
NC Not calculated for this section because of insufficient data; however, the
value is not assumed to be zero.
ND No datum.
105
-------
Tg/yr (Table V.8) of which 45.0 Tg/yr are estimated to be
inherited emissions. The remaining 30.8 Tg/yr are prompt C
emissions arising directly from base-year land conversions,
predominantly through fires.
D. Discussion
For Mexico, the estimate of 30.8 Tg/yr for prompt C emissions and
75.8 Tg/yr combined (prompt + inherited for 1990) emissions are
higher than those estimated by Makundi et al. (1992) (based on
work by Masera et al. 1992) of 27.9 and 62.6 TgC/yr, respec-
tively, for 1985 (Table V.8). The difference in base years
between the two studies has a negligible effect on the estimates
since most of the deforestation data is averaged for the 1980s.
Differences in land area between the two dates (1985 and 1990)
are well within the uncertainty associated with any single-date
estimate. Also, the fact that this section includes all Mexican
lands, rather than only closed forests, has little effect on
emissions estimates since most non-forest lands are assumed to be
in approximate C equilibrium.
The emission differences between this section and Makundi et al.
(1992) are primarily attributable to two factors. First, this
section uses much lower estimates of percent phytomass C
converted to charcoal during fires (8% vs. 20-40%). Second, this
section generally used higher phytomass C estimates for forests
and much higher soil C estimates. These two factors account for
approximately 70% of the difference between the estimates of this
section and those of Makundi et al. (1992) and Masera et al.
(1992) .
A few comparisons are possible between the Olson et al. (1983)
estimates for C density and the limited data published for
Mexican forest sites (Table V.3). Castellanos et al. (1991)
106
-------
estimated total phytomass (above and belowground) in a dry
tropical forest in Jalisco state at 104.5 Mg/ha. If half of this
amount of phytomass is C (as commonly assumed), then about 52
MgC/ha is comparable to Olson's estimate of 50 MgC/ha.
Martinez-Yrizar et al. (1992), working in the same area,
estimated aboveground phytomass at 85 Mg/ha or by the same
assumption, 42.5 MgC/ha. At a root/shoot ratio of 0.42
(Castellanos et al. 1991), total phytomass C would be 60 MgC/ha.
For mature Mexican humid tropical forest, no comprehensive
studies of phytomass are published, yet the Olson value of 150
MgC/ha is at the low end of C estimates based upon stand-table
data from Veracruz (M. Ricker, Yale University, pers. comm.,
1993). The Olson et al. (1983) estimates appear to concur
reasonably well with the few published C density values reported
for Mexico.
The soil C estimates used by Masera et al. (1992) are lower by a
factor of approximately two relative to estimates by Zinke et al.
(1984). The Zinke estimates use globally available soil C data
for calculating values for each continent and the globe. The
global averages of Zinke et al. (1984) are similar to the
estimates of Kern (1994) and lower than Sombroek et al. (1993)
(Table V.5).
Studies are underway in Mexico to improve the estimates of
phytomass and soil C densities (Cairns 1994) through satellite-
based mapping and field TEC measurements. Until they are
available, it is best to consider that the estimates of Makundi
et al. (1992) and this section define a range within which the
actual level of C emissions lie.
Emissions of C may be partially offset in Mexico through
sequestration in growing vegetation. Although there is little
information available on the status of second growth forests or
regrowth of vegetation in perturbed forests, this section defines
107
-------
an upper and lower limit of C sequestration. This rough
approximation suggests that C sequestration may range from 4 to
12 TgC/yr (Table V.8), equivalent to 5 to 16% of emissions due to
land-use change. Makundi et al. (1992) estimated that 3.5 TgC/yr
were sequestered annually on recently deforested lands within the
four closed forest categories of Mexico.
Opportunities exist in Mexico for promoting conservation and
sequestration of C through ecosystem management that would
contribute to global greenhouse gas mitigation. Recent changes
in national policies (Masera et al. 1995) may remove government
support for forest conversion to pastureland. Further, changes
in land-tenure laws may also encourage long-term forest manage-
ment .
Masera et al. (1995) suggest that investments in improved
management of temperate and moist tropical forests in Mexico are
economically feasible and would, at the same time, provide
substantial C sequestration and conservation. Masera et al.
(1995), also considered enhanced natural area protection,
restoration planting, pulpwood plantations, and improved
wood-burning stoves as viable mechanisms for sequestering or
conserving C. These authors estimate an upper limit of 60 TgC/yr
could be sequestered over the next 20 years at a cost of
approximately $5.1 billion if all examined options were imple-
mented. This represents a substantial fraction of the annual
national energy emissions of C02-C (Table V.8), although the
energy sector emissions may double or even triple in the next 40
to 100 years (Mendoza et al. 1991; Masera et al. 1995).
An analysis of the potential of reforestation practices in 34
forested nations of the world by Dixon et al. (1993) show that
reforestation practices in Mexico could sequester an average of
about 100 MgC/ha at $4.50/MgC based on initial costs of planta-
tion establishment. Among the 34 forested nations analyzed,
108
-------
Mexico's reforestation cost per MgC was one of 12 with costs
below $5/MgC.
Carbon may also be sequestered in wood products. This section
estimates that approximately 8.6 TgC/yr of C were removed from
Mexican forests in 1990 as wood products. The proportion of this
C which is quickly converted into C02 through burning or decompo-
sition is not known, though fuelwood comprises approximately 70%
of all wood removals. Due to these uncertainties, the 8.6 TgC/yr
removed in wood products was not included in the C budget for
Mexico. O. Masera (Universidad Nacional Autonoma de Mexico,
pers. comm., 1994) suggests that solid wood products may last at
least an average of 5 years in Mexico. If even a small propor-
tion of the 8.65 TgC/yr removed as wood products each year is
used as durable products, this represents a significant C pool,
leading to an underestimate of C storage in Mexico.
Estimates of net C emissions from Mexico's terrestrial ecosystems
by all studies range from 56 to 71 TgC/yr as compared to
industrial emissions of 77 to 97 TgC/yr (Table V.8). These
emissions due to land-use change represent approximately 0.25% of
the entire C pool in all ecosystems of Mexico (Table V.6) being
transferred to the atmosphere each year. The estimates of land-
use change emissions represent approximately 3% of estimated
global emissions of C from deforestation (WRI 1993).
Of the total C emissions from land-use change in Mexico,
disturbance of the humid tropical forests are the major source of
C emissions. Deforestation in Mexican tropical forests account
for approximately 75% of net deforestation emissions. Most
emissions come from land conversion of forests to pastureland,
cropland or other non-forest activities rather than from
harvesting. Little more than 2% of total emissions are derived
from forests that have been logged versus a 16% contribution from
forest fires.
109
-------
Recent estimates of deforestation rates in Mexico have largely
been site- or region-specific (Dirzo and Garcia 1992; Masera et
al. 1992). The Masera et al. (1992) estimate of total area
deforested (. 668 X 106 ha/year, Table V.2) is about the same as
FAO's (1993) estimate. Other deforestation estimates in Mexico
range from 365 X 103 to 1500 X 103 ha/yr (Cairns et al. 1995).
FAO (1993) stratified forest types in a manner not readily
comparable to other available sources of information for Mexico.
For example, an exact equivalence between the forest zones of FAO
(1993) and the forest types of Masera et al. (1992) is not
possible because the zones and types do not match in area or
definition. Deforestation rates in the wet and moist zones of
the FAO (1993) classification, however, are lower than estimated
by Masera et al. (1992) for the humid tropical forests (evergreen
and semi-evergreen tropical forests). Conversely, deforestation
in the hill and montane zones is higher than the conifer and oak
types of Masera et al. (1992).
E. Conclusions
This section reports that estimated net emissions of C from
Mexican terrestrial ecosystems for 1990 range from 63.4 to 71.4
TgC or about 44% to 49% of the total national C emissions as
estimated by Makundi et al. (1992). The net C emissions
estimated in this section are higher than in studies by Makundi
et al. (1992), Subak et al. (1993) and WRI (1993). Using any of
these studies, however, it is apparent that land-use change is
driving a significant loss of the C capital of the nation of
Mexico and is making a contribution to the net C flux to the
atmosphere on a global basis.
In Mexico, as with many other tropical nations, C emissions from
land-use change are substantial when compared to fossil fuel C
emissions. Mexico cannot long sustain the second highest areal
110
-------
extent of deforestation in Latin America (FAO 1993). However,
conservation of remaining forests may be difficult to achieve
based on trends in Mexico's agricultural expansion.
The relatively close agreement between results from this section
and other sources is not necessarily an indication of accuracy or
precision. There is considerable uncertainty in the magnitude of
both source and sink components of the Mexican terrestrial C
cycle, and the results of this study must be considered extremely
tentative. Several assumptions employed in this section
contribute to the uncertainties in this analysis.
First, the land cover area estimates used in this study are
preliminary. Both this section and Masera et al. (1992) use data
from SARH (SARH 1986 and SARH 1992 respectively), which may lead
to common errors.
Second, as discussed previously in this section and at greater
length earlier in Section II.B.2.C, this section assumes that
most undisturbed lands are in C equilibrium with the atmosphere.
This assumption has been challenged globally by Lugo and Brown
(1992). Carbon equilibrium in mature forests is also implicit in
the assumption that forests disturbed in the past ten years are
considered to have C sequestration potential while older
perturbed forests are not modeled as c sinks. Specifically, this
section employs one scenario that assumes that only forests
recently and moderately perturbed (i.e., including perturbation
by wildfire and timber harvest, but not by conversion to
agricultural use) are experiencing net C sequestration at the
rate of 1 Mg/ha/yr (see part 1. Land Cover). Another scenario
assumes that 50% of the recently perturbed forest is sequestering
C at 1 Mg/ha/yr. Both the assumption that older perturbed
forests are in equilibrium and the estimate of C sequestration
rate may be in error.
Ill
-------
Third, this section uses the relative proportions of pastureland
and croplands estimated by Masera et al. (1992) to estimate
forestlands recently deforested. This was done even though the
total area of non-forestland estimated by Masera et al. (1992)
(107.3 X 106) is different from the total estimated by SARH
(1992) (51.5 X 106 ha) used here. The assumption that the
proportions used by Masera et al. (1992) can be applied to the
SARH data may not be valid.
Fourth, this study relies on phytomass C densities from Olson et
al. (1983, 1985). Olson C densities are based on limited numbers
of field measurements extrapolated to continental scales.
Although the low Olson C density estimate are in reasonable
agreement with the few available published values for Mexico,
their accuracy and relevance to Mexico are uncertain.
Fifth, the assumptions used here regarding loss of soil C after
conversion of forest to pastureland or cropland are strongly
debated in the scientific literature. There is neither consensus
as to the magnitude or, in some cases, the sign of C flux after
forest conversion (Davidson and Ackerman 1993, Houghton et al.
1991b): A number of factors are critical in controlling ecosystem
response to forest conversion. Clearly, however, the type of
conversion coupled with the severity and duration of disturbance
(e.g., repeated fires, mechanical clearing, and cultivation) will
lead different C dynamics post-disturbance (Brown and Lugo 1990) .
Finally, the equation of land-use and land-cover change with
deforestation is a recognition of our lack of data on the
ultimate fate of disturbed land. It is known that indigenous
Mexican Mayans have managed southern forests for several
centuries, creating a patchwork of secondary forests. If, due to
lower population densities today than in the past, the net effect
is an increase in growing stock, then a large amount of C may be
actively sequestered in tropical Mexican forests. Not all land
112
-------
cover change has C dynamics impacts similar to deforestation,
e.g. selective harvesting.
Although these may be the major uncertainties involved in this
study, they are likely not the only ones. This section has
attempted to incorporate improvements in estimates of C dynamics
which have become available since the studies of Masera et al.
(1992) and Makundi et al. (1992) were completed and to expand the
scope of the C budgets. At present, it is prudent to consider
the results of this section and those of the previous studies to
define a range within which the actual level of C pools and flux
lie. Because data are lacking to confirm several assumptions
presently employed, research is required in each of these areas.
113
-------
Box 4: Brazil
Brazil lies between latitudes 5°N
and 32°S and longitudes 35°W and
75°W, occupies an area of 845 X 10s
ha, and dominates South America. It
is the fifth largest nation in the
world in land area behind the former
Soviet Union (FSU), Canada, China,
and the United States. The nation
is divided into 27 states and the
Federal District of Brasilia, the
capital (Cottle et al. 1990) (Figure
VI.1).
In 1990, the population of Brazil
was about 150 million people cor-
responding to a population density
of 18/km2. By comparison in 1990,
the population density of FSU was 13
people/km2, China, 122, Canada, 29,
and the U.S., 27 (WRI 1992). In
Brazil, however, 58% of the people
live in the five most southern and
southeastern states with 25% living
in five metropolitan areas located
in these states (NYT 1988). In
total, Brazil has 14 cities with
over 1 million people each. The
annual growth rate of the population
is 2.2% which projects to a
population of 165 million by 1995
(WRI 1992).
Brazil was explored and settled by
the Portuguese starting in about
1500 AD. Today the Portuguese
influence remains strong within
Brazil, and Portuguese is the
official language. However, many
remnants of the indigenous Indian
populations are scattered through
the nation with higher concen-
trations of these people in the
Amazon region (Eden 1990).
Currently, Brazil has a democrati-
cally elected government, though
prior to 1985, Brazil had a long
period of military rule. New
elections are scheduled for 1994.
Economically, Brazil is considered a
developing nation with a 1990 GNP of
$U.S. 2,550/capita compared to $U.S.
9,211 for the FSU, $US 360 for
China, $U.S. 19,020 for Canada, and
$U.S. 21,100 for the U.S. (WRI
1992). In the early 1990s, Brazil's
economy has been in a recession with
inflation rates in 1993 near 40% per
month (EIU 1994).
Brazil has the largest area of
forest of any of the nations in the
tropical latitudes, over 370 X 106
ha of closed forests plus 70 X 106
ha of secondary forests (Stone et
al. 1994) (Tables VI.1 and VII.1).
Thus, Brazil's forest area is
several times that of Zaire at 113 X
106 ha; Indonesia at 110 X 106 ha;
and Peru at 68 X 106 ha, the next
most forested nations in the tropics
(FAO 1993).
The climate of Brazil has four broad
geographic regions:
1) Tropical humid in the rolling
lowlands (<300 m of elevation)
of the Amazonia area to the
north (Figure VI.1). The
annual precipitation ranges
from 1500 to 3000 mm with the
primary wet season occurring
from November through April
(Eden 1990). Mean monthly
temperatures vary between 25
to 28 °C year around. The
natural vegetation type is
predominantly closed moist
forest (Stone et al. 1994).
About 10% of closed moist
tropical forest has been
deforestated during the last
several decades (Fearnside
1993). These areas are now in
secondary forests, pasture,
shifting subsistence
agriculture, dam reservoirs,
and open-pit mines (Fearnside
1992).
114
-------
(2) Cooler tropical humid in the
Atlantic coast (lat5°S to
15°S) at elevations near 500
m, though with coastal
mountains up to 2000 m (Figure
VI.1). The temperatures vary
from 21 to 24 °C and annual
precipitation rangeB from 1000
mm to 4000 mm, mostly in
December through March (though
lat5°S to 15°S, there is a
rainy season from April
through July). Droughts,
however, are frequent
throughout the region.
Vegetation types are a mix of
cerrado (i.e., savannah) and
woodlands. Though in many
areas much land is in a
degraded condition (Stone et
al. 1994), subsistence farming
and sugar cane production
widely occur.
(3) Tropical semi-humid in the
central cerrado region at
elevations near about 1000 m
(Figure VZ.l). For the
cerrado plateau, the annual
precipitation varies between
1000 to 2000 mm (primarily in
December through March) with
annual temperatures averaging
20 to 28 °C. Vegetation types
are dominated by cerrado (a
mix of campo limpo or
grassland to woody cerrado)
with widespread clearing for
commercial farm crops such as
soybeans and corn (Stone et
al. 1994).
(4) Cooler subtropical to warm
temperate humid conditions at
elevations below 800 m in the
south though interspersed with
higher elevations in mountain
ranges (NYT 1988; Cottle et al.
1990). The southern subtropical
area has an annual precipitation
rate between 1000 and 2000 mm
distributed year around and with
mean monthly temperatures generally
between 10 °C in the winter up to 27
°C in the summer. Vegetation types
of the lower elevations are grass-
lands widely used for farming.
Mountain areas have degraded closed
forests extensively used for farming
with some degraded lands that have
reverted to secondary forests (Stone
et al. 1994).
The tropical region of Brazil is
dispersed within the basin of the
world's largest river system, the
Amazon. In total, the basin covers
705 X 106 ha (NYT 1988), and 66% is
within Brazil (Fearnside 1992). The
river has an annual discharge of
5.5 X 1012 m3 (Sioli 1975). Compar-
able figures for the Mississippi
River in the U.S. are 325 X 10 ha
in area drained, an annual discharge
of 0.5 X 1012 m3 of water, and about
0.8% of the world's total fresh
water (Goulding 1980).
Brazil's biodiversity is one of the
richest in the world with an
estimated total of 30,000 species of
higher plants (470-*- tree species)
and over 1 million animal species
including insects, birds, fish,
rodents, and others (Eden 1990).
Conservation of these unique
ecological features along with
sustainable development has become
one of the greatest contemporary
environmental concerns for both
natural resource managers and
scientists in Brazil as well as
around the world.
115
-------
VI. Brazil Case Study
A. Introduction
Brazil plays an important role in the tropical and global C
cycle. This importance emanates from Brazil's (1) large
geographic area (Figure VI.1), placing it fifth in size among all
nations at about 850 X 106 ha (WRI 1992) ; (2) over 350 X 106 ha
of C-dense, closed moist forest (Stone et al. 1994 and Table
VI.1) , the largest in the tropics (FAO 1993); and (3) intense
level of land-use change associated with the nation's attempt to
improve its developing economy (Box 4).
The most renowned land-use change in Brazil is deforestation of
the tropical forests to meet land needs for more agriculture,
pastureland, logging, mining, dams, roads, and settlements as the
nation strives to develop economically. The level of deforesta-
tion in Brazil has been one of the world's highest during the
past decade ranging from 1 to 2 X 106 ha/yr (INPE 1992) . The
annual deforestation rate for the total tropical regions of the
world during the period 1981-1990 was 15.4 X 106 ha/yr (FAO
1993). Fragmentation of the remaining forest habitat is also a
concern (Skole and Tucker 1993).
These characteristics make Brazil a most useful case study to
assess the contribution of land use and changes in land use to a
national C flux. The main focus of the Brazil study is on (1)
the C pools of terrestrial ecosystems in Brazil; (2) the net flux
of C02 through ecosystem phytomass and soils; and (3) the effect
on C pools and flux associated with human activities, predomi-
nantly land-use change. All data on pools and fluxes are for the
nominal base year of 1990. Information and data for the study
come from a review of the technical literature and contacts with
scientists engaged in ecological research in Brazil.
116
-------
I Closed tropical moist forest
CD Recently degraded
tropical moist forest
¦ ^1 !¦» jir . ~i I .-h jiiji fj £ jfc m * hj a
Utner ctoseo torest
O Degraded dosed forest
D Cerrado woodlands
H Degraded woodlands
D Grasslands
O Degraded passiands
¦ Water
¦ Wetlands
Equator
Brazil
0 1000
1 I I
Kilometers
Figure VI.1 Vegetation map of Brazil developed by Stone et al
(1994) .
117
-------
Table VI.1.
Vegetation and Land-Use Types, Brazil, 1994*
Vegetation and
Land-Use Type
Closed moist forest (TMF)
Recently degraded TMF
Secondary forest
Degraded forest/mixed
agriculture
Pastureland
Other closed forest
Degraded closed forest
Agriculture
Secondary Forest
Cerrado woodland
Degraded woodland
Savannah/grassland
Degraded savannah
(converted to agriculture)
Wet1and/mangrove
Total
Area Area
(X 106 ha) (%)
352.2 42
17.4 2
19.4 1
15.1 2
16.4 2
99.7 12
17.6 2
155.6 19
33.0 4
74.0 9
17.9 2
12.4 1
830.7 100
'Stone et al. (1994).
118
-------
B. Methods
A national biotic C flux study is essentially an accounting of
the annual net exchange of C between the atmosphere and the
terrestrial ecosystems of that nation. For Brazil, this study
considered both processes that take up C and those that release
C, and it encompasses the entire geographic region of the nation.
Many previous studies for tropical regions consider either only
the uptake of C in forests (e.g., Brown et al. 1992; Lugo and
Brown 1992) or the release of C, particularly through human
activities such as deforestation (e.g., Detwiler and Hall 1988;
Houghton et al. 1991a and b). For Brazil, most analyses of
phytomass C dynamics have only considered the nation's Amazon
region (e.g., Fearnside 1992).
1. Conceptual Model
The approach for the Brazil case study is based on a simple
conceptual model of ecosystem C storage and flux (Figure VI.2).
The primary C pools in the model are the atmosphere, phytomass,
litter and coarse woody debris (CWD), and soil. Flux is
represented as a transfer of C between these pools that occur as
a result of land-use change, natural disturbance such as by
storms (Nelson et al. 1994), or recovery from past changes or
disturbances by plant growth and accumulation of organic matter.
The key flux is that between components of the land and the
atmosphere, i.e., through growth (carbon removed from the
atmosphere), decomposition and deforestation/burning (carbon
emitted to the atmosphere) (Figure VI.2).
2. Vegetation and Land-Use Types
The conceptual C model was applied to each of the vegetation
types for a recently completed vegetation map of Brazil (Figure
119
-------
Growth
Deforestation/
Burning
Decomposition
Litterfall
Mortality
Humification
Vegetation
Litter
Coarse Woody Debris
Soil
Atmosphere
Figure VI.2. Conceptual model of biotic net carbon balance.
120
-------
VI.1) (Stone et al. 1994). Other vegetation maps of Brazil were
considered and found less useful for this analysis (see Section
VI.D.2.).
The map by Stone et al. (1994) was based primarily on Advanced
Very High Resolution Radiometer (AVHRR) satellite data from the
late 1980s and early 1990s with a resolution of approximately 1.1
km. These data were supplemented by higher resolution satellite
imagery available for certain sections of South America, by
photographs, by earlier potential vegetation maps, and by field
observations. Because cloud-free, 1.1 km data were not available
for the whole nation, a three-year weekly data set of 15 km
Global Vegetation Index (GVI) data was also utilized for areas
with missing data (Stone et al. 1994). This explains the
different pixel sizes in the vegetation map (Figure VI.1).
The two degraded forest types (Table VI.1) were subdivided in
order to more readily estimate their C densities. Stone et al.
(1994) use the term degraded in connection with any of the types
to include "areas which have been altered or converted from
natural or primary vegetation by man, grazing or for cultiva-
tion." The degraded forest vegetation classes are very heteroge-
neous and can include land uses with quite different C densities.
First, the degraded tropical moist forest (TMF) type was split
into three subdivisions: secondary forest, degraded forest with
mixed agriculture, and pastureland. Of the total area (51.9 X
106 ha; Table VI. 1) in this type, 9.4 X 106 ha are located
outside the Amazon basin are dealt with below. Another 12.3 X
106 ha are in the states of Para and Maranhao in areas that have
been settled for a long time. The resulting landscape has been
heavily influenced by human occupation and consists of a mix of
small scale agriculture, fallow fields, palms, and degraded
secondary vegetation. In this study, this diverse landscape is
called "degraded forest/mixed agriculture" (Table VI.1).
121
-------
An additional area of degraded TMF of about 30.2 X 106 is located
elsewhere in the Amazon basin. This area includes secondary
forest and, although it could not be clearly distinguished from
the AVHRR data, is also believed to include areas of pasture (Tom
Stone, Woods Hole Research Center, pers. comm., 1994). A number
of references from the literature indicate what might be a
plausible weighing between the two. A Landsat Thematic Mapper
(TM) analysis in Para by Lefebvre and Stone (1994) showed that
about 45% of the area of cleared or altered forest was in
secondary forest. Fearnside (1992) estimated that 63% of
deforested areas Amazon-wide would develop into secondary forest
and degraded pasture. Degraded in this context usually implies
invasive secondary vegetation including woody shrubs and trees.
In Paragominas County in Para, Mattos and Uhl (1994) estimated
that up to 50% of previously cleared pastures had been abandoned.
Hecht (1993) estimated 50% of pastures throughout the Amazon had
been abandoned, presumably to secondary vegetation. Based on
these studies, the area of degraded TMF was partitioned into 50%
each (i.e., 15.1 X 106 ha) for secondary forest and pastureland
(though in Table VI.1 the area for secondary forest, is 17.4 X 106
ha because 2.3 X 106 ha were added for plantations as noted
below).
Two other studies, however, illustrate the uncertainty in this
partitioning. A Landsat TM study in Rondonia estimated that only
33% of deforested areas were in secondary forest (Skole et al.
1994), but another Landsat TM study in western Para found that
70-80% of cleared areas were occupied by secondary forest (Moran
et al. 1994). For this analysis of the c pools of Brazil, exact
partitioning of this 30.2 X 106 ha vegetation type has some
influence on the results, but not a critical one. The impact on
C flux, however, is potentially more significant as shown in the
sensitivity analysis for the importance of secondary forests (see
Section VI.D.3).
122
-------
Finally, 9.4 X 105 ha of degraded TMF are located outside of the
Amazon basin along the Atlantic coast in eastern and southeastern
Brazil (Table VI.1 and Figure VI.1). For this area, 2.3 X 106 ha
were allocated to tree plantations and secondary forest (Camara
1991; H.T. do Couto, Universidade de Sao Paulo, pers. comm.,
1992) and the remainder to degraded forest/mixed agriculture.
Tree plantations were thus included in the secondary forest type
bringing the total in Table VI.1 to 17.4 X 106 ha (15.1 from
above plus 2.3 here).
Second, the definition of the degraded closed forest vegetation
type includes both secondary forest and agricultural areas.
Because the C densities of these land uses are quite different,
the type was subdivided into these two land uses using
proportions reported by Fonseca (1985). He estimated that the
Atlantic forest had been converted to approximately 85%
agriculture and 15% secondary forest so that the resulting areas
were 99.7 and 17.6 X 106 ha, respectively (Table VI.1).
The combination of information sources just described produced
the following set of major vegetation types or land-use types
used for this study (Table VI.1):
a. Closed tropical moist e. Cerrado woodland
forest (TMF)
f. Degraded woodland
b. Recently degraded TMF
Secondary forest g. Savannah/grassland
Degraded forest/
mixed agriculture h. Degraded savannah
Pastureland (converted to agriculture)
c. Other closed forest i. Wetland/mangrove
Degraded closed forest
Subdivided into:
Agriculture and
Secondary forest
123
-------
3. Brazil's Carbon Pools
Estimates of the terrestrial C pools of Brazil depend upon two
primary types of information: (1) the area of major vegetation
and land-use types and (2) reported data on the C density in
MgC/ha within each area. Areas of vegetation came from the map
by Stone et al. (1994).
The biotic C content in each vegetation or land-use type included
estimates of three major components: phytomass, woody debris
(including litter and course woody debris [CWD]), and soil. An
important source for C density data was a report compiled by
Olson et al. (1983) characterizing C in the world's phytomass.
This global scale study developed a 0.5s latitude X 0.5°
longitude vegetation map based on data from the late 1970s.
Plant mass and C were estimated for 47 vegetation types based on
a very wide variety and extensive number of ecosystem studies
(Olson et al. 1983). Some C density values were revised in a
subsequent report (Olson et al. 1985), and these data were also
used where appropriate. Finally, C density values were assigned
to the Stone vegetation types from the most similar types in the
Olson classification system (Table VI.2).
Though the Olson (1983, 1985) C density data for the terrestrial
ecosystems of the world are widely used, more accurate site-
specific data for Brazilian ecosystems is a definite research
need. To illustrate, some of the vegetation types in the map by
Stone et al. (1994) had no counterpart in the Olson map so that C
density estimates came from other sources. One was degraded
woodlands, and here the derivation of the C density estimate was
from several references. Tropical woodlands are most often
degraded by excessive burning which lowers the phytomass density
and opens the woody vegetation (Eiten 1972; Coutinho 1982; Frost
and Robertson 1985). It was assumed that if intact woodland
124
-------
Table VI.2. Carbon Density in Phytomass by Vegetation and
Land-Use Type, Brazil
Phytomass C (MgC/ha)
Vegetation and
Land-Use Type*
Closed tropical moist
forest (TMF)
Recently degraded TMF
Secondary forest
Degraded forest/
mixed agriculture
Pastureland
Brown & Lugo5
Fearnside' Kauffman*
Olson1
200
40
2 511
10*
Other closed forest 60
Degraded closed forest
Agriculture 8
Secondary Forest 4 0
Cerrado woodland 3 0
Degraded woodland 2 0#
Savannah/grassland 9
Degraded savannah
(converted to agriculture) 8
Wetland/mangrove 3 0
191
ND
ND
ND
172
ND
ND
22
ND
ND
ND
190
137
ND
ND
ND
ND
ND
ND
ND
ND
3.4*
ND
ND
*Stone et al. (1994).
^Olson et al. (1983 and 1985).
'Fearnside (1992).
'Brown and Lugo (1992). Estimated from their published value for
above-ground phytomass by using a root/shoot ratio of 0.2
and C concentration of 0.5.
^Author's approximation, see text for explanation.
'Deduced from several sources, see text for explanation.
*Kauffman et al. (1994b).
ND No datum.
125
-------
contained 30 MgC/ha and if agricultural areas contain 8 MgC/ha
(both values from Olson et al. 1983 and 1985) then degraded
woodlands should be intermediate, i.e., approximately 20 MgC/ha
(Table VI.2).
The degraded forest/mixed agriculture type also had no
counterpart in the Olson classification. This type actually
represents an intermixed variety of different types of vegetation
and land uses. No data were available on the C density for this
type. The best approach seemed to try to approximate an average
C density that could be used for the whole type. Similar to
degraded woodlands, it seemed that an appropriate average C
density should be somewhere between that for secondary forest (40
MgC/ha) and that for pastureland (10 MgC/ha). A mid-point value
of 25 MgC/ha was used. The C density estimate for pastureland
was taken from Fearnside (1992) .
These values (the Olson set plus the estimates for degraded
woodlands and Fearnside#s value for pastureland) furnished one
series of C densities (Table VI.2). Carbon values from other
sources were also used in order to provide a range of C pool
estimates for some vegetation and land-use types. This approach
reflects the uncertainty in the available data that results from
the complexity, size, and diversity of the nation and its
vegetation. By using a variety of data sources, the objective
was to produce a range of results that would bound this
uncertainty. Alternative C densities for four types were taken
from Fearnside (1992), i.e., closed moist forest, other closed
forest, cerrado woodland, and wetland/mangrove (Table VI.2). In
addition, two alternatives were taken from Brown and Lugo (1992)
and Kauffman et al. (1994b), i.e., closed moist forest and
savannah/grassland (Table VI.2). When no other published
datawere available, however, the Olson C estimates were used as
default values (see Section VI.D for further considerations).
126
-------
Published estimates of root/shoot ratios were used to partition
phytomass C into aboveground and belowground components (Table
VI.3). Further, published results were used to estimate C in
litter and CWD. A companion report to Olson et al- (1983)
provides soil organic C estimates for all of the Olson vegetation
types (Zinke et al. 1984). These values were also applied to the
Stone map (Table VI.4).
By these methods, estimates of C density for terrestrial phyto-
mass were determined for each of the vegetation or land-use types
(Table VI.4). Nation-wide pools were then calculated by
aggregating C pools across the major vegetation or land-use types
(Table VI.6).
4. Brazil's Biotic Carbon Flux
Published data were used to calculate flux estimates (Table
VI.5). It was assumed that mature vegetation was in an
equilibrium or steady state where uptake and release via
biological processes were balanced for all pools on a per unit
area basis with no net gain or loss of C. There is growing
uncertainty, at least for closed moist forest, that this
assumption is appropriate. Brown and Lugo (1992) believe that
most forests in the Brazilian Amazon, for example, are not in C
steady state because of past widespread exploitation. The
implications of altering the steady state assumption are
addressed in more depth in Section VI.D.6.
Besides the background flux associated with each vegetation and
land-use type, changes were accounted for in the area of any of
these vegetation types (i.e., annual change from one vegetation
type to another land use). Each such land-use change created an
important C flux.
127
-------
Table VI.3. Root/Shoot Ratios and Factors Used in the Brazilian Study
Vegetation and
Land-Use Type*
Root/Shoot
Ratio
Litter/CWD (in
relation to above-
ground phytomass C)
Closed tropical moist
forest (TMF)
0.2*
0.135*
Recently degraded TMF
Secondary forest
Degraded forest/mixed
agriculture
Pastureland
0.2*
0.2**
1.6*
0.135"
0.135**
o.ott
Other closed forest
0.23'
0.04"
Degraded closed forest
Agriculture
Secondary Forest
0.18*
0.23*
o.ott
0.04"
Cerrado woodland
1.6*
2 MgC/ha**
Degraded woodland
1.61
0.0OT
Savannah/grassland
1.6*
0.0"
Degraded savannah
(converted to agriculture)
0.18*
o.ott
Wetland/mangrove
0.75a
o.ott
*Stone et al. (1994).
tpearnside (1992); Jordan and Escalante (I960).
'Saldarriaga (1985).
'Cannell (1982) (India).
*As for shrub/grassland.
'Sanchez et al. (1989).
*Same as other closed foreBt.
*Singh and Joshi (1979); Fiala et al. (1991) (Cuba).
xSame as woodland.
*Cannell (1982) (mangroves).
*Salati et al. (1991); Uhl et al. (1988); Uhl and Kauffman (1990).
aSame as closed moist foreBt.
"Kauffman et al. (1993), Schacht et al. (1988) for caatinga forest.
^Assumption.
Same as other closed forest.
"Kauffman et al. (1994b).
"Assumes frequent burning and no litter accumulation.
**Same as secondary forests.
128
-------
Table VI.4. Carbon Density in Phytomass Components and Soil by
Vegetation and Land-Use Type, Brazil*
Vegetation and
C Density
(MgC/ha)
Land-Use Type
Total
Phytomass'
Above-
ground
Roots
Litter/
cwd'
Soil5
Closed tropical moist
forest (TMF)
137-200"
114-167
22-33
15-22
104
Recently degraded TMF
Secondary forest
Degraded forest/
mixed agriculture
Pastureland
40
25
10
33
21
4
7
4
6
4
3
35-52*
94
94*
95
Other closed forest
60-172
49-140
11-32
2-6
100
Degraded closed forest
Agriculture
Secondary Forest
8
40
7
32
1
8
0
1
75
90*
Woodland
22-30
8-12
14-18
2
60
Degraded woodland
20
8
12
0
60
Savannah/grassland
3-9
1-3
2-6
0
60
Degraded grassland
(converted to
agriculture)
8
7
1
0
54*
Wetland/mangrove
30-190
17-108
13-81
0
234
*Developed from Tables VI.2 and VI.3
'The sum of the C densities for "Aboveground" and "Roots" may not equate
exactly with that for "Total Phytomass" because of rounding.
'Values are the product of proportion factors in Table VI.3 and the above-
ground C densities of Table VI.4.
'values from Zinke et al. (1984).
'See Section VI.D. Discussion for consideration of some reports of higher C
density values by dos Santos (1987) and Anderson and Spencer (1991).
^Values are area-weighted means for above-ground residual C from remaining
snags and logs; in addition, the area-weighted means for remaining dead
roots are 10 to 15 MgC/ha; see Methods in text for full explanation.
*Assumes 10% C loss due to disturbance.
129
-------
Table VI.5. Total Area, Area Cleared, Flux Parameters per Hectare, and Estimates Used to
Compute Net Carbon Balance, Brazil
Vegetation and
Source/Basis
Land-Use Type
Flux Parameter
Estimate
Closed tropical moist
forest (TMF)
Total Area (X 106 ha)
352.2
Stone et al. (1994)
Area Cleared (X 106 ha)
1.38
Fearnside (1992)
Deforestation/burning factorst
% aboveground C to charcoal/yr
3.6
Fearnside (1992)
% aboveground C released/yr
27.5
Fearnside (1992)
Crutzen & Andrae
(1990)
% litter released/yr
95.0
Assumption
% soil C released/yr
S.O
Cerri et al. (1991)
Sanchez et al. (1983)
Net flux (MgC/ha/yr) -!
50.0 to -70.3
Recently degraded TMF
Stone et al. (1994)
Secondary forest
Total Area (X 10* ha)
17.4
Tree growth (MgC/ha/yr)
3.0
Saldarriaga (1985)
Brown & Lugo (1990)
Litter accumulation (MgC/ha/yr)
1.0
Walschburger & von
Hildebrand (1991)
Soil C accumulation (MgC/ha/yr)
1.0
Lugo et al. (1986)
Net flux (MgC/ha/yr)
+5.0
Deforestation/burning factors:
Area cleared (X 106 ha)
0.35
Fearnside (1992)
% aboveground C to charcoal/yr
3.6
Same as closed moist
forest
% aboveground C released/yr
60.0
Ewel et al. (1981)
% litter released/yr
95.0
Same as closed moist
forest
% soil C released/yr
5.0
Same as closed moist
forest
Net flux (MgC/ha/yr)
-28.3
Degraded forest/
mixed agriculture
Total Area (X 106 ha)
19.4
Pastureland
Total Area (X 106 ha)
15.1
Stone et al. (1994)
Vegetation growth (MgC/ha/yr)
1.0
Fearnside (1992)
Decomposition/reburning
(MgC/ha/yr)
4.2 to 6.1
See text
% soil C released/yr
0.5
Fearnside (1992)
Detwiler (1986)
Houghton et al.
(1991a)
Davidson et al.
(1992)
Net flux (MgC/ha/yr)
-3.2 to -5.1
Other closed forest
Area (X 106 ha)
16.4
Stone et al. (1994)
130
continued
-------
Table VI.5.
(Continued)
Vegetation and
land-use type
Flux parameter
Estimate
Source/Basis
Degraded closed forest
Agriculture Area (X 106 ha)
% soil C released/yr
Net flux (MgC/ha/yr)
99.7
1.0
-0.75
Stone et al. (1994)
Assumed to be larger
than releases from
pasture
Secondary forest Area (X 106 ha)
17.6
Stone et al. (1994)
Cerrado woodland
Total Area (X 106 ha)
Deforestation/burning
factors:
Area cleared (X 106 ha)
155.6
1.0
Stone et al. (1994)
Fearnside (1992)
% aboveground C to charcoal
% aboveground C released
% root C released
0.0
100.0
100.0
Based on assumptions
described in text
% soil C released
5.0
As for forest
Net flux (MgC/ha/yr)
-35.0
Degraded woodland
Area (X 106 ha)
33.0
Stone et al. (1994)
Savannah/grassland
Area (X 106 ha)
74.0
Stone et al. (1994)
Degraded savannah
(converted to
agriculture)
Area (X 106 ha)
% soil C release/yr
Net flux (MgC/ha/yr)
17.9
1.0
-0.56
Stone et al. (1994)
Assume to be larger
than releases from
pasture.
Wetland/mangrove
Area (X 106 ha)
12.4
Stone et al. (1994)
131
-------
Table vi.6.
Total Carbon Pools In Brazil by Vegetation and Land-Use Types Calculated with a Variety of Carbon Density Values (Tables
VI.2 and VI.3)*
reams Ida 1992
Brown 6 Lugo 1992]
Kauffman et al. 1994b
Zinke
et al.
1984
Vegetation and
Land-use Type
Phytomass
fPaCl
%
Below-
ground
fPOCl
Above-
ground
-------
It was assumed that most of the degraded vegetation types were in
a disequilibrium state (i.e., either accumulating or releasing
C), so data were sought to quantify the C flux. Information for
degraded woodlands, however, was insufficient to form a basis for
any kind of flux estimates. This vegetation type, however, only
occupies 4% of the area of the nation (Table VI.1).
There was also insufficient information available to quantify
flux for the areas of degraded closed forest listed as secondary
forest, 17.6 X 106 ha (Table VI.1). This area is located
primarily along the Atlantic coast of Brazil where human
population densities are high. For this reason, it is likely
that these secondary forests experience fairly intensive use to
supply wood for fuel, housing construction, and other purposes.
Therefore, it is unlikely that these secondary forests are
experiencing recovery and regrowth as is the case in the less-
populated Amazon basin. In summary, the secondary forests of the
former Atlantic forest region could be net sources of C to the
atmosphere, but data were not found to support or refute this
idea.
Much the same can be said for the area that was designated as
degraded forest with mixed agriculture in the TMF zone of eastern
Amazonia. This area, in the states of Para and Maranhao, has
been settled for a long time. No information was available to
estimate C fluxes for this vegetation class which forms a very
heterogeneous landscape of different land uses.
Major fluxes in Brazil's C budget are released to the atmosphere
resulting from land-use change. Three types of land-use change
were included in the analysis: conversion of closed moist forest
to pasture or low intensity shifting cultivation, conversion of
secondary forest to pasture or low intensity shifting cultiva-
tion, and conversion of cerrado woodlands to intensive agricul-
ture (e.g. for soybean). In 1990, 1.38 X 106 ha of closed forest
133
-------
were cleared in Brazil (INPE 1992) with an additional 1.0 X 106
ha of cerrado woodland (Fearnside 1992).
No nationwide estimates of clearing of secondary forest were
available, although it seems reasonable that some clearing of
this forest type must occur. Only regional studies have been
published. Skole et al. (1994) reported that in an area of
Rondonia, clearing of secondary vegetation was an important
source of new agricultural land. From 1988 to 1989, the area of
secondary forest cleared was about 70% as large as the area of
mature forest cleared. However, Moran et al. (1994) cite data
from Para that indicate abandonment of pastures and agricultural
lands was 1.5 to 2.5 times as large as the rate of new clearing.
The rapid expansion of secondary forest in their study area
implies that relatively little was being cleared. Hecht (1993)
and Fearnside (1989) also seem to imply that most pastures are
created from mature forest and that abandoned areas are usually
left to secondary vegetation. These differing reports make it
difficult to decide on an appropriate factor for secondary forest
clearing. Therefore, it was assumed, more or less arbitrarily,
that the clearing rate for secondary forests was 25% that of
mature forests, or about 0.35 X 106 for 1990 (Table VI.5).
Releases of C from deforestation were calculated by multiplying C
density by estimates of combustion efficiency (deforestation/
burning factors, Table VI.5). The combustion efficiency, the
proportion of aboveground phytomass C released to the atmosphere
at the time of deforestation, for closed moist forest (27.5% from
Fearnside 1992) and secondary forest (60% derived from Ewel et
al. 1981) reflect the. assumption that these areas are cleared for
pastureland or low intensity shifting cultivation. The factor
for cerrado woodlands (100%) is based on the assumptions that (1)
these areas are cleared for types of intensive agriculture (e.g.,
soybean cultivation) requiring they be completely cleared and (2)
the vegetation at the time of burning is very dry because of the
134
-------
arid environment. A factor was also included for release of soil
C due to the disturbance caused by forest clearing and burning
(5% of soil C in the upper 1 m; Table VI.5).
The calculation of C release from maintaining pasturelands, via
decomposition and reburning of remaining residual forest
phytomass, is more complicated. The initial burning of the
forest releases only .a little less than 30% of aboveground C and
no belowground root C (given the assumptions reflected in Table
VI.5 that burning does not consume root phytomass). This leaves
a significant amount of residual C that must be accounted for in
the overall flux estimates.
As a working base, the pastureland management scenario described
by Fearnside (1992) was adopted. This is a typical management
scenario that includes three reburnings over a period of 10
years. Bacterial decomposition and termite activity also release
C under this scenario which results in the consumption of all of
the residual C by the end of the assumed 10 year period. A mean
was calculated for the residual aboveground C density in all
pasturelands by (1) using the charcoal and combustion factors in
Table VI.5; (2) dividing the total pastureland area into 10 equal
sized age classes; and (3) assuming linear rate of release over
10 years. Further, a mean residual root C was calculated in a
similar manner. The means for residual aboveground C density
(i.e., CWD) ranged from 35 to 52 MgC/ha with a range of 10 to 15
MgC/ha in residual root C (Table VI.4).
Working with proportions of C released in initial and reburns
presented by Fearnside (1992), it was calculated that about 76%
of the residual aboveground C would be released by decomposition;
the remainder (24%) would be released by reburning. It was
assumed that all of the residual belowground phytomass would be
decomposed. With a linear decomposition rate over 10 years, this
calculation gave an annual decomposition release of 3-5 MgC/ha/yr
135
-------
aboveground and about 1 MgC/ha/yr belowground. The C releases
from reburning were estimated on an area-vide basis because it
does not occur on every ha annually. The total annual area-wide
release from reburning was calculated as a proportion of the
total area-wide decomposition release, i.e., about 30%
(0.24/0.76), of the total area-wide decomposition release.
Pasturelands also gradually accumulate C both in roots and in the
aboveground biomass of invading woody vegetation (Table VI.5).
C. Results
Key results for the Brazil case study are in two parts, C pools
and C flux. The nominal base year for all results is 1990.
1. Carbon Pools
The total phytomass C pool for all ecosystems was 81 PgC, 80 PgC,
and 58 PgC with the Olson, Fearnside, and Brown and Lugo/Kauffman
et al. C estimates, respectively (Table VI.6). In all three
cases, closed moist forest predominated with 86%, 83% and 81% of
the total phytomass C, respectively. By contrast, the other
geographically important ecosystems together accounted for only
14-19% of phytomass C. Differences in total pool size were due
primarily to differences in estimates of C in closed moist forest
(Table VI.6). Nationwide, the closed moist forest phytomass C
pool was 70 PgC, 67 PgC, and 48 PgC based on the Olson,
Fearnside, and Brown and Lugo/Kauffman et al. C density esti-
mates, respectively.
The litter and CWD C pool was an order of magnitude smaller than
the phytomass pool. The range was from less than 6 PgC to about
9 PgC (Table VI.6). Most of this pool was also contained in the
closed moist forest (5-8 PgC). The total estimated pool size for
soil C was 72 PgC (Zinke et al. 1984) (Table VI.6). Combined,
136
-------
the phytomass, litter plus CWD, and soil pools ranged from 136 to
162 PgC, 55-62% of which was belowground. Compared to the total
C pool for the nation, the total closed moist forest pool
represented 66-71% (Table VI.6).
2. Carbon Flux
Because the magnitude of C emissions was largely dependent on C
density and pool sizes, there was also a range of results for the
net C flux of the nation as a whole (Table VI.7). This resulted
from both biological activity and land-use changes. The range
was from a net emission of 174 TgC/yr to 233 TgC/yr. The
difference was caused mostly by differing estimates in the C
density of closed moist forest which resulted in different flux
rates from deforestation and burning. Combustion efficiency
values, however, also have a significant influence (see Section
VI.D.4).
As expected, direct C emissions from clearing and burning closed
moist forest were large, 69-97 TgC/yr (Table VI.7 and further
consideration in Section VI.D.l). However, emissions from
decomposition and reburning of existing pasturelands were nearly
as large, 62-93 TgC/yr. Two other classes of C releases resulted
from (1) clearing and burning cerrado woodlands, 27-35 TgC/yr and
(2) releasing soil C by intensive agriculture: 75 and 10 TgC/yr
from converted closed forest and converted grassland, respec-
tively (Table VI.7).
Net changes in the total nationwide pool sizes of the C budget
model varied. For all three of the C density matrixes, the
litter/CWD pool increased by 70-90 TgC/yr (Figures VI.3-5). This
was largely the result of large transfers of 140-205 TgC/yr from
the vegetation C pool due to deforestation. The vegetation C
137
-------
Table VI.7.
Net Carbon Flux by vegetation and Land-Use Types,
Brazil
Net Flux*
Vegetation and Land-Use Type (TgC/yr)
Closed tropical moist forest (TMF) -97, -93, -69
Recently degraded TMF
Secondary forest +77
Degraded forest/mixed agriculture ND
Pastureland -93, -89, -62
Other closed forest ND
Degraded closed forest
Agriculture -75
Secondary forest ND
Cerrado woodland -35, -27, -35
Degraded woodland ND
Savannah/grassland ND
Degraded savannah -10
(converted to agriculture)
Wetland/mangrove ND
Total Net Flux -233, -217, -174
*Flux is uptake minus release. A plus sign indicates net c
uptake by vegetation or human activity. A minus sign indicates
net C release. Multiple values result from the three mixes of C
density values used. From left to right, the results in the table
were derived from the Olson (1983, 1985), Fearnside (1992), and
Brown and Lugo (1992j/Kauffman et al. (1994b) C density mixes.
Where no net flux is given, the type is assumed to be in C steady
state or available information was insufficient to estimate flux.
A row with only one value implies that the value applies to
calculations for all three sources.
ND No datum.
138
-------
pool, therefore, decreased 240 and 230 TgC/yr for the Olson and
Fearnside C density values, respectively, and 160 TgC/yr for the
Brown and Lugo/Kauffman et al. C densities. The lesser estimate
of C in closed moist forest of the latter produced a smaller C
transfer from deforestation. The soil pool lost 80 TgC/yr
because of soil disturbance associated with deforestation and
ongoing soil C loss from intensive cultivation.
Total gross C input to the system was 100 TgC/yr of which 60
TgC/yr was vegetation growth and 4 0 TgC/yr was added to lit-
ter-CWD and soil pools in secondary forests (Figures VI.3-5).
Total C releases were 120-150 TgC/yr from deforestation/burning
and 150-180 TgC/yr from decomposition (Figures VI.3-5). Most C
released from burning derived from aboveground vegetation, but a
significant amount also came from litter and CWD, including
reburning of residual logs on pasturelands.
D. Discussion
Brazil's large geographic size and predominantly tropical climate
combine to make a vast vegetative C pool. The pool ranges from
58-81 PgC in live aboveground vegetation and roots (Table VI.6).
The size of the vegetation C pool estimated here for Brazil
represents approximately 10% of the total global vegetation pool
of 560-830 PgC (Schneider 1989), or 16-22% of global forest
vegetation pool (Dixon et al. 1994). Brazil's soil C pool (71.7
TgC, Table VI.6) is 4-7% of the global soil C pool of 1100-1700
PgC (Dixon and Turner 1991).
By far the largest portion of the vegetation pool in Brazil is
contained in the closed moist forest; the estimates were 81-86%
of the total vegetation pool (Table VI.6). Future attempts to
refine this type of analysis, therefore, should focus on the
closed moist forest. There is undoubtedly variation in the
139
-------
Atmosphere
" +0.23*
T
0.15
Deforestation/^ 0.105
Burning
~
0.05*
Litter
Coarse Woody Debris
9
+0.09
Humification
0.02;
Soil
72
-0.08
Growth
0.06^0.04'
I
Vegetation
81
-0.24
0.18
Decomposition
~
Litterfall/
Mortality
0.20T + 0.04T
t
0.08 0.10
I
Figure VI.3 Annual not carbon balance for Brazil using the Olson (1983, 1985) carbon density
values. Boxes are pools (Table VI.6), arrows are fluxes, and all units are FgC.
The upper.number in each box is the pool size. The lower number is the net annual
change in the pool size.
' The 0.23 PgC represents a net gain of C to the atmosphere (shown as an emission of 233 TgC in
Table VI.7) as a result of summing the emissions and uptake in PgC from closed moist
forest, -0.097; secondary forest, +0.077; pastureland, -0.093; degraded closed forest, -
0.075; cerrado woodland, -0.035; and degraded savannah, -C.010.
" The uptake of 0.06 PgC is 0.C5 PgC from secondary forest plus 0.C1 PgC from pastures
: Litterfall accumulation is 0.04 PgC primarily in secondary forests which then contributes 0.02
PgC to the litter C pooi and ultimately 0.02 PgC to the soil C pool through humification.
5 The emission of 0.10 PgC is the sum of 0.06 PgC from deforestation and burning (including
removals of harvested logs) in the closed moist forest; 0.01 PgC from burning in secondary
forests; and 0.03 PgC from burning in the cerrado woodlands in preparation for agriculture
crops.
' The 0.20 PgC consists of logs and other CUD left aboveground after deforestation.
* Decomposition emissions of 0.08 PgC from litter and CWD consists primarily from decay of legs
and roots in patsures and roots in cleared cerrado woodlands.
* Decomposition emissions of 0.10 PgC from soils is the result of human disturbances in six
vegetation typest closed moist forests, 0.01; secondary forests < 0.01; pastureland, 0.01;
degraded closed forest, 0.07; cerrado woodlands, < 0.01; and degraded savannah, 0.01.
' Burning of litter and CWD results in emissions of 0.05 PgC which consists of closed moist
forest, 0.03 secondary forests, < 0.01; pastureland, 0.02; and cerrado woodlands, < C.01.
140
-------
Growth
0.06 + 0.04
0.14
0.18
Deforestation/
Burning
Decomposition
Litterfall/
Mortality
0.20 + 0.04
0.08
0.10
0.05
Humification
0.02
-0.08
Soil
Vegetation
-0.23
Litter
Coarse Woody Debris
+0.09
Atmosphere
+0.22'
Figure VI.4. Annual net C balance for Brazil during 1990 using the
Fearnside (1992) carbon density values. Boxes are pools,
arrows are net fluxes, and all units are PgC. The upper
number in each box is the pool size (Table VI. 6), and the
lower number is the net annual change in the pool size.
Following from the footnote above, the net gain of C to the atmosphere (shown
as an emission of 217 Tg in Table VI.7) is 0.01 PgC less than the value in
Figure VI.3.
s The C density value of Fearnside (1992) for closed moist forest of 191 MgC/ha
(Table VI.2) is lower than the Olson (1983, 1985) value of 200 MgC/ha.
Similarly, for the cerrado woodlands, Fearnside's C density value is 22'
Mg/ha compared to Olson's value at 30 MgC/ha. These differences cause the
emission of C from deforestation and burning in both vegetation types to
be 0.01 PgC less than the value in Figure VI.3.
Other footnotes for the net flux arrows in Figure VI.3 also apply in this figure.
141
-------
Growth
0.06 + 0.04
0.12
0.15
Vegetation
Deforestation/^ 0.085 ;
Burning *
Decomposition
-0.16
Litterfall/
Mortality
0.14 + 0.04
0.05
0.10
0.04
Humification
0.02
-0.08
Soil
Litter
Coarse Woody Debris
+0.07
Atmosphere
+0.17*
Figure VI.5. Annual net C balance for Brazil during 1990 using the Brown
and Lugo (1992) carbon density values. Boxes are pools,
arrows are net fluxes, and all units are PgC. The upper
number in each box is the pool size (Table VI.6), and the
lower number is the net annual change in the pool size.
Following from the footnote above, the net change of C to the atmosphere (shown
as a release of 174 TgC in Table VI.7) is a 0.06 PgC difference than the
value in Figure VI.3 for the uptake of C by terrestrial ecosystems in
Brazil.
8 The C density value of Brown and Lugo (1992) for closed moist forest of 137
MgC/ha (Table VI.2) is lower than the Olson (1983, 1985) value of 200
MgC/ha. This causes the emission of C from deforestation and burning to
be 0.02 PgC less than the value in Figure VI.3. Also the same lower C
density value used in this figure results in less emissions from burning
(0.04 vs 0.05 PgC) and decomposition (0.05 vs 0.08 PgC) of litter/CWD than
in Figure VI.3.
Other footnotes for the net flux arrows in Figure VI.3 also apply in this figure.
142
-------
phytomass and C content of these forests that is not adequately
reflected in the analysis and which is an important source of
uncertainty. For example, dos Santos (1987) reported phytomass
densities for eastern Amazonia ranging from 266 to 915 ton/ha
(i.e., about 133-458 MgC/ha). Similarly, Anderson and Spencer
(1991) reported a range of 185-513 ton/ha (i.e., about 93-257
MgC/ha). The lower end of these ranges are comparable to the
values used in this analysis (Table VI.4). The values apply to
southeastern and southern portions of the Brazilian closed moist
forest where the majority of the recent deforestation has
occurred (Fearnside 1993). The high-end values would not affect
the estimates of C flux for the vegetation type. However, the
higher values, if shown to be representative of a large portion
of the Brazilian closed moist forest, would increase the C pool
estimate. This is an important topic for refinement in future
analyses of the C budget for Brazilian forests.
Based upon the literature, the C flux estimates of this study for
all major vegetation and land-use types in Brazil were the first
that reached beyond the Amazon region. Thus the estimated 1990
net C flux of the vegetation and land-use types for the nation as
a whole ranged between -174 and -233 TgC/yr (Table VI.7). The
study demonstrated the dominance of the Amazonian region in the C
dynamics of the nation. It contains the largest C pool (closed
moist forest), the two largest release flux rates (deforestation
at 69-97 TgC/yr and decomposition of residual wood on pasture-
lands at 62-93 TgC/yr), and the largest uptake flux (growth of
secondary forests at 77 TgC/yr [Table VI.7 and possibly by the
mature TMF at 176 TgC/yr as suggested in Section VI.D.6]).
Characterization of Brazil's C flux could be largely accomplished
by focusing on the Amazonia region. Improved information on C
accumulation by both mature and secondary forests and the fate of
C during and after deforestation would be useful in future
studies.
143
-------
The largest biotic C emissions from Brazil are driven by human
factors related to land-use change. These include a developing
economy, an expanding population, prevalent poverty, land
speculation, and government tax-breaks or subsidies to large land
owners in conjunction with the existence of a large resource-rich
region of undeveloped land in th$ Amazon basin (Hecht 1992;
Schmink and Wood 1992). Gross C emissions to the atmosphere
resulting from current and past land-use change may be as high as
300 TgC/yr (Table VI.7). In addition to this possible contribu-
tion to greenhouse gas emissions, the environmental degradation
caused by deforestation and the threats to biological diversity
that it entails are well known (Randall 1991). Throughout the
early 1990s, Brazilian deforestation rates have been declining,
but this may be mostly due to the nation's current economic
crisis. Once Brazil's economy recovers, deforestation rates are
expected to rise again (Fearnside 1992).
1. Comparisons to Other Studies
The most intensive analysis of Brazilian C dynamics, heavily
cited here, was performed by Fearnside (1992) for the Amazon
basin. That study determined that clearing 1.38 X 106 ha/yr of
mature forest in the period 1989-1990.was responsible for the
release of approximately 270 TgC02-C/yr. That study, however,
was focused on C emissions and did not include C uptake by
secondary forests. It, therefore, estimated C release rather
than net c balance. There is also a difference in accounting for
the fate of C released from cleared forest. Fearnside's approach
was to account in the base year for all C that would ultimately
be released due to deforestation. This approach, sometimes
called ,the committed C approach to estimating emissions,
attributes all C release to the year' in which the deforestation
occurs. In the nationwide analysis for Brazil that is presented
here, the method was to develop an estimate of the annual C flux
144
-------
that focused on the actual C released from deforestation in 1990
and also accounted for release of residual C from decomposition
and reburning of pastures that resulted from previous deforesta-
tion .
Another recent study (Subak et al. 1993) estimated that net
biotic C02-C releases from Brazil were 213 X 106 MgC/yr, which is
within the range of our estimates in this analysis. Although
this study included C uptake by vegetation, it only estimated it
for plantations. By not including non-commercial secondary
forests, Subak et al. (1993) may have underestimated C uptake.
2. Base Map Used for Study
The map used here by Stone et al. (1994) is very current and was
derived from a continental-scale data set. AVHRR-derived
vegetation maps, however, are subject to their own disadvantages
(Townshend et al. 1991). Even a resolution of 1.1 km can be
rather coarse for vegetation mapping, particularly in areas of
heterogeneous vegetation cover. Mixed pixels can produce signals
that do not accurately depict the actual vegetation present.
Another problem inherent in all broad-scale maps is that it is
impossible to fully verify the accuracy of the various types of
vegetation and land use and of the map itself. This is due
simply to the size of the area being mapped; it is much too large
to be directly ground-truthed. The Stone group addressed this
problem by conducting a series of accuracy assessments and ground
study comparisons. They examined the variation in information
sources used to develop classifications within the map, and they
compared other estimates of the areal extent of land-cover types
to their findings. In addition, they compared land-cover type
assignments in a widely distributed set of sites to external
land-cover descriptions from field studies. They concluded that
the vegetation types covering 85% of the whole map of South
145
-------
America were correctly assigned 90% of the time. They also
determined that the vegetation types for Brazil were among the
most reliable for the entire continent (Stone et al. 1994).
Although the resolution of the Stone map may have some disadvan-
tages in a strictly cartographic sense, it seemed very suitable
for estimating major C pools over a very large area. More
detailed vegetation maps of Brazil with much finer stratification
of vegetation types are available, for example, the UNESCO (1980)
and IBGE/IBDF (1993) maps. The usefulness of these maps for this
Brazil study, however, was limited by the availability of C
density data for the more numerous vegetation types they contain.
Lack of C density data for many vegetation types would require
some amount of aggregation which would result in a loss of
resolution. A wider sampling of C densities for terrestrial
ecosystems in Brazil is an important research need (see Section
VIII.D).
In another recent study using satellite technology, Skole and
Tucker (1993) used a combination of Landsat TM and MSS imagery to
estimate the amount of deforestation in the Amazon Basin of
Brazil between 1978 and 1988. With this methodology, the authors
indicate that deforestation is about 50% less than estimates made
with coarser resolution AVHRR satellite data for the southern
Amazon Basin of Brazil. Across the whole Brazilian Amazon Basin,
the authors compared the area deforested in 1978 with that in
1988. A breakdown of the increase in deforested area into esti-
mates of current land uses such as secondary forests or pasture-
land was not indicated. While the improved precision of
deforestation estimates by Skole and Tucker (1993) is important
to note, the Brazil analysis described in the current report
required estimates for major vegetation and land use types within
the deforested areas throughout the nation for 1990. Thus it was
felt that the recent 1.1 km resolution map by Stone et al. (1994)
was appropriate for the sub-continental and nationwide analysis
146
-------
as required here for Brazil. Further, a comparative advantage of
the AVHRR-based map is that the use of training sets enabled the
Stone group to identify the spectral signature of secondary
forests. In addition, this approach recognizes that deforesta-
tion occurred in decades prior to 1978 and continued after 1988.
3. The Role of Secondary Forests:
A Sensitivity Analysis
At least two recently published articles have commented on the
widespread occurrence of secondary forests in Amazonia (Skole et
al. 1994; Moran et al. 1994). However, there is no reliable
estimate for the area of these forests in Brazil. This analysis
used an area of 15.1 X 106 ha of secondary forest in Amazonia as
the major part of the total of 17.4 X 106 ha (see Section VI.B.2)
by partitioning the very heterogeneous degraded TMF vegetation
class defined by Stone et al. (1994). An important factor in
this partitioning was weighting the areas of secondary forest and
pastureland. Based on several references from the literature,
these two vegetation classes were weighted at 50% each, but there
were also higher and lower published values (Schroeder and Winjum
1995a, 1995b).
To evaluate the importance of this uncertainty on total C
emissions from Brazil, a sensitivity analysis was conducted by
altering the areas of secondary forest and pastureland. The
secondary forest/pasture ratio by was changed ±20% to correspond
to a lower estimate reported by Skole et al. (1994) and a higher
estimate derived from Moran et al. (1994). This resulted in
large changes in C releases (Table VI.8). Partitioning up to 70%
of the area in question to secondary forest resulted in a
decrease in total C emissions of about 30%. When 30% of the area
was partitioned to secondary forest, total C emissions rose by
about 30%. This is equivalent to a change of ±56-69 MgC/yr.
147
-------
Secondary forests play an important role in Brazil's C budget
because they are very active in fixing and storing atmospheric C.
However, there is much uncertainty about specifically how large
that role is. A nationwide inventory of the area and age class
distribution of secondary forests would help to resolve this
uncertainty.
The results also suggest that tropical secondary forests may be
more important in the global C cycle than previously realized.
Current estimates of the global C budget show an imbalance of 1.1
PgC/yr; i.e., known C sources exceed known and modeled sources
(Dixon et al. 1994). There has been much discussion and debate
about the fate of this "missing" C. Perhaps it is possible that
tropical secondary forests are a larger C sink than they are
accounted for in current global C budgets.
4. Changes in Combustion Efficiency
An important variable in this analysis was the combustion
efficiency of deforestation burning. The combustion efficiency
of 27.5% used in this analysis (Table VI.5) has been observed in
actual fires in the Amazon (Fearnside 1992), but there likely is
wide variation that depends on an array of factors (Molion
submitted). To test the importance of variation in combustion
efficiency, a flux calculation was made with a substituted value
of 50%. The result was that total net C releases to the
atmosphere were actually reduced by 5-6 TgC/yr. A higher
combustion efficiency releases more C during the initial bum,
and results in a smaller pool of residual C on pasturelands.
Therefore, pasturelands release far less C from reburning and
decomposition than when the initial combustion efficiency is
lower.
148
-------
Table VI.8.
Sensitivity Analysis
Secondary Forest on
of the Effect
Brazil's Net C
of the Area of
Balance*
Secondary
forest area
(X 106 ha)
C Density Matrix (TgC/yr)
Olson
(1983 & 1985)
Fearnside
(1992)
Brown &
Lugo (1992)
11.4
-302
-283
-228
17.4
-233
-217
-174
23.5
-166
-151
-118
*Minus signs indicate releases to the atmosphere. Areas of
secondary forests represent alternative partitioning of 30.1
X 106 ha of the degraded TMF vegetation class in Amazonia
between secondary forests and pastureland (i.e., proportions
are .7, .5, and .3 as explained in Section VI.D.3).
149
-------
In this example, decreased releases of C from pasturelands were
slightly greater than increased releases during the initial burn.
Although this can have some impact on Brazil's annual C balance,
the long-term effect on total emissions to the atmosphere is
unchanged. The same amount of C will be released as a result of
deforestation, but the passage of time may vary.
5. Hood Products and Fossil Fuels
An important C flux not represented in the conceptual model
(Figure VI.2) is release resulting from commercial timber
harvesting. Some timber harvesting in Brazil is part of the
deforestation process (Westoby 1989). Road construction and
selective logging establish access to the forest which can then
be subject to agricultural colonization (Cottle et al. 1990;
Fearnside 1989; Mather 1990). This analysis, therefore,
distinguished between timber harvest that may be associated with
deforestation processes and timber harvest in managed planta-
tions. It was assumed that C fluxes associated with the former
are accounted for in the net flux for deforestation in the closed
TMF, i.e., -50.0 to -70.3 MgC/ha/yr (Table VI.5). Harvesting
plantations is accounted for separately using the following
assumptions. Brazil provides 60% of its domestic industrial wood
needs from 7 X 106 ha of tree plantations (Mather 1990) . Timber
harvesting from these managed plantations represents a process
that is separate from and in addition to what is normally thought
of as deforestation. In 1990 about 72.0 X 106 m3 of the total
harvest in Brazil was in industrial roundwood used for fuel,
sawnwood, panels, and paper (WRI 1992). Brazil's timber exports
are insignificant (< 1% of total harvest) (WRI 1992). Thus 60%
of the total harvest of 72 X 106 m3 or 43 X 10® m3 comes from
managed plantations. This volume of wood harvested can be
converted to C released by making two additional assumptions:
(1) 1 m3 of wood is equivalent to 0.26 MgC/m3 (i.e., an average
150
-------
wood density of 0.52 Mg/m3 X 0.5 MgC/Mg wood) (Marland 1988;
Sedjo 1989) and (2) a unit of stemwood phytomass is associated
with a total 1.59 units of whole tree phytomass which includes
not only bole but also branches, leaves, and roots (Marland 1988;
Sedjo 1989). Multiplying all of these factors together produces
an estimate of an additional 18 TgC/yr released from timber
harvesting of plantations (Table VI.9). This is an estimate of
gross release; C uptake by plantations is included with secondary
forests. Truly sustained yield plantation management could fully
balance C release from harvest with C uptake by regrowth.
Release of 18 TgC/yr is comparable to about 8-10% of gross C
releases from land-use change and intensive agriculture (Table
VI.9 and Figures VI.3-5).
A small amount of C is stored annually in the harvested wood
manufactured into durable-wood products. WRI (1992) reported
that for 1990 about 21 X 106 m3 of Brazil's wood harvest was used
in this manner. Converting this volume to its approximate weight
in C (i.e., 1 m3 = 0.26 tC from above) indicates that about 5.4
TgC was newly stored in durable-wood products in 1990. However,
for such products produced in previous years, there is some level
of decomposition which releases C. It is assumed here that the
annual decomposition emissions of C is about half that of newly
produced so the net gain in 1990 was approximately 3 TgC (Table
VI. 9).
In addition to C released to the atmosphere from biotic sources
and timber harvest, there are also releases from burning fossil
fuels. Brazil released approximately 50 TgC from fossil fuel
usage in 1987 (Graca and Ketoff 1991; Subak et al. 1993). In
addition, 3 TgC was emitted annually in the late 1980's from the
manufacture of cement (Subak et al. 1993). Combined, these
emissions contribute about 18-22% of the total net C flux from
Brazil (Table VI.9).
151
-------
Table VI.9. Total Net Annual Carbon Flux from
Brazil, All Sources
Sources
Net Annual Flux4
(TgC/yr)
Biotic sources (Table VI.7)
-174 to -233
Timber harvestt
-18
Fossil fuels*
-50
Cement*
-3
Durable-wood products11
+3
Total Net Flux
-242 to -301
Potential flux mature TMF*
+176
Net potential flux
-66 to -125
*Flux is uptake minus release. A plus sign indicates
net terrestrial C uptake. A minus sign indicates
net C release.
^Emissions from plantation harvests as developed in
Discussion (section VI.D.5) from Mather (1990) and
WRI (1992).
'Graca and Ketoff (1991); Subak et al. (1993).
*Subak et al. (1993).
'Accounts for that portion of Brazil's wood harvest
manufactured into durable-wood products (see
Section VI.D.5).
'Potential value for mature TMF if uptake by vegeta-
tion is 1 MgC/ha/yr for half the area (Lugo and
Brown 1992), i.e., 176.1 X 10f ha (see Table VI. 1
and detailed consideration [Section VI.D.6]).
152
-------
Overall, net annual C emissions from all known sources in 1990
were estimated to range from 242 to 301 TgC/yr (Table VI.9). Net
annual global emissions from all sources are about 7 PgC/yr
(Sundquist 1993). This means that Brazil's contribution may
represent 3-4% of the total. With more definitive data on the
potential C uptake by mature tropical moist forests, the nation's
contribution might be closer to 1-2% (Table VI.9 and next
discussion, Section VI.D.6).
6. Unquantified Flux Processes
Although the major components of the biotic C budget of Brazil
are C02 uptake in terrestrial ecosystems and emissions by land-
use changes, other processes are ongoing which may affect net C
balance. At least two are possible: C uptake by mature forests
and river transport of C to the ocean. These other processes are
potentially important, but detailed data are unavailable or
limited in both cases.
Undisturbed tropical mature forests not undergoing land-use
change have often been assumed to be in C equilibrium or steady
state with no net C gain or loss. However, there may be reason
to question this steady state assumption. Lugo and Brown (1992)
make the case that the TEC and TEC accumulation rates of tropical
forests change over time as a result of past disturbance by
people and natural catastrophe. They point out that the Earth's
total biotic C pool (560 PgC) is currently considerably smaller
than the total pool size possible (900 PgC) for interglacial
optimal climatic conditions estimated by Faure (1990). Lugo and
Brown (1992) suggest that this means the current biota could
store additional C and still be within the estimated optimal
storage capacity. They further cite plot-level field data from
Venezuela and Puerto Rico that show mature forests accumulating C
at rates of 1-2 MgC/ha/yr. In addition, an eddy correlation
153
-------
study by Fan et al. (1990) estimated that a nature closed forest
near Manaus in Brazil had a net C uptake of 2.2 Mg/ha/yr.
Brazil has such a huge area of mature forest in the Amazon region
(i.e., > 350 X 10e ha, Table VI.1) that the slightest increase in
phytomass per ha would result in very significant amounts of C
accumulation. Data are limited that quantify such increases
region-wide (Molion submitted), but as noted earlier, Lugo and
Brown (1992) suggest that possibly half of the mature tropical
forests of Brazil could average 1.0 MgC/ha/yr in net uptake. The
range may be from as little as 0.25 Mg/ha/yr to as high as 1.75
Mg/ha/yr. Thus a simple matrix of these values indicate the
potential C uptake of the mature tropical moist forests of
Brazil:
The uncertainly related to this topic clearly indicates that
additional research on this potentially significant flux is very
important. Not only would more definitive data aid in balancing
the global C budget, but they would greatly improve the
calculations for the C dynamics of Brazil as well. To illus-
trate, using an uptake value of 176 TgC/yr for half of Brazil's
mature TMF would indicate that the net biotic flux of Brazil is
nearly in balance (Table VI.7).
Globally, most C is cycled among three major reservoirs: the
atmosphere, the terrestrial biosphere, and the oceans (Holmen
1992). Out analysis on Brazil largely focused on the C exchanged
between the first two, although some C is exchanged between the
latter two as well. Carbon in water draining from the terres-
trial biosphere is transported by rivers to the oceans.
Globally, this flux is estimated to be 540 TgC annually, and for
C
Uptake
(Mg/ha/yr)
Mature TMF IX 106 ha)
Half Total (Table VI.1)
0.25
1.00
1.75
176
44
176
308
352
99
352
616
154
-------
the rivers of the humid tropics the estimate is 250 TgC/yr
(Meybeck 1993).
The vast size of the Amazon River makes river transport of C a
potentially important flux in Brazil's C budget. The Amazon
transports 68 TgC/yr to the ocean in various forms (Kempe 1988).
The fate of this carbon, however, is not completely understood.
Some is oxidized and returns to the atmosphere either directly
from rivers or estuaries or after entering the ocean. By either
route, this C is accounted for in the Brazil C budget by the net
change approach used for computing release and uptake. The C not
oxidized goes into long-term storage in the sediments of lakes,
river beds, and the ocean bottom. Over time in river basins such
as the Amazon, some riverine meanders develop into cutoff loops
that become large C pools. The total C pool for cutoff loops in
the Brazilian Amazon basin is unquantified, but if such data
becomes available it should be considered in later analyses
(L.C.B. Molion, Universidade Federal de Alagoas, Brazil, pers.
comm., 1994). For river transport of C worldwide, about 10% is
deposited into lakes (Meybeck 1993).
It is the ocean sink that remains to be accounted for in Brazil's
C budget. Currently only 6% of the total C transported by the
Amazon is known to be deposited on the continental shelf near the
river mouth (Showers and Angle 1986). It is estimated that this
known flux is about 3 TgC/yr (68 TgC/yr X 6% X [66%, the portion
of the Amazon basin in Brazil]). The fate of the remaining
Amazon-transported C is unknown.
7. Forest Management for C02 Mitigation
Brazil appears to have significant potential to expand the
management of its forest resources for C storage and other
benefits. By these means, the nation's forest C sink would be
155
-------
increased, aiding global C02 mitigation. Prospects are particu-
larly attractive in the tropical moist forest which occupies
slightly over half of Brazil. The same environment that favors
vigorous and rich growth in mature forests will also (in theory)
produce rapid growth rates in managed forests on deforested
lands. In terms of aboveground C uptake, rates reported for
forest areas in or similar to the Brazilian Amazon region show
the higher potential for managed forests and plantations:
Land use MaC/ha/vr References
Pastureland 1 Fearnside 1992 and Table VI.5
Secondary forest
Unmanaged 2-5 Saldarriaga 1985; Brown & Lugo 1990
Managed 4-20 Brown & Lugo 1990
Plantations 4-25 Brown et al. 1986; Postel & Heise
1988
As of 1990, the Brazilian Amazon had considerable land area that
was under producing. Estimates are that in 1990 there was a
total of 51.9 X 106 ha of recently degraded tropical moist forest
divided into secondary forests at 17.4 X 106 ha, degraded
forest/mixed agriculture at 19.4 X 10® ha, and pasturelands at
15.1 X 106 ha (Table VI.1). Productivity on these lands could be
greatly increased if significant portions were placed under
management to grow new secondary forests and plantations or to
keep losses in existing stands to a minimum.
The amount of managed secondary forests reported for Brazil is
only a few thousand hectares (WRI 1992). Most of the 51.9 X 106
ha was originally cleared for pastureland or slash-and-burn
agriculture, became degraded, and was left to become unmanaged
secondary forest (Fearnside 1992). Management practices
suggested for secondary tropical forests include tree thinning,
understory control, and opening crown cover to promote seed
germination and natural regeneration_(Brown and Lugo 1990). By
these practices, the production of many useful forest goods can
156
-------
be enhanced in secondary forests. Examples are logs, fuelwood,
fruit, nuts, oils, medicinal plants, and local construction
materials (Brown and Lugo 1990). To the extent that deforesta-
tion of mature forest is for the same goods, forest management
practices that, as an alternative, provide them on deforested
lands reduces the pressure to clear more mature forest, thereby
conserving existing phytomass C.
The 1990 total plantation area in Brazil was 7 X 106 ha or only
0.8% of the nation's land area (FAO 1993). Key hardwood species
are eucalyptus (i.e., Eucalyptus deglupta and E. ur.ophylla) and
melina (Gmelina arborea), and softwoods are predominantly pine
such as Pinus caribaea var. hondurensis (McNabb et al. 1994).
The proportion of hardwood plantations to softwood plantations is
about two to one (Cottle et al. 1990). Fast growing plantations
of these species are grown in 10- to 20-year rotations. High
wood yields can be sustained and increased through economically
attractive silviculture practices such as intensive site
preparation, genetic improvement, fertilization, and particularly
weed control throughout plantation rotation periods (McNabb et
al. 1994). Plantation establishment continues in Brazil at
roughly 450 X 103 ha/yr (WRI 1992). This annual planting rate
should be sustained or perhaps expanded because in the future it
is anticipated that more plantation wood will be needed in Brazil
to meet increased demands in the paper and cellulose industries
(Serrao and Homma 1993).
Thus with high growth rates and large areas of suitable land, the
potential exists to sequester and conserve more C on forestlands
in Brazil's Amazon region. Many approaches are possible such as
(1) protect young secondary forests so that they continue to
store and accumulate C, (2) convert more degraded pastureland to
either secondary forests or to plantations, and (3) manage more
efficiently mature forests such as is done by the extractive
reserve projects (Serrao and Homma 1993).
157
-------
The nation of Brazil must decide whether C mitigation through
expansion of forest management is important to their interests.
Vegetatively, the nation as a whole seems close to having a zero
net C flux (Table VI.7), and it has only modest emissions through
timber harvest and fossil fuels burning (Table VI.9). Growing
more phytomass through expanding forest management, however,
would (1) increase the nation's managed forest resources and
reduce deforestation of mature forests, (2) provide a significant
step toward sustainable development called for in 1992 at UNCED
for all forested nations (Keating 1993), and (3) place Brazil in
a position to participate in joint implementation forest projects
with developed nations. These outcomes could offer Brazil added
economic benefits and simultaneously sequester C to help reduce
global atmospheric C02.
158
-------
VII. Discussion: Comparison of the C
Budgets for Four Nations
Natural terrestrial ecosystems and land-use types of the FSU,
U.S. (48 conterminous states), Mexico, and Brazil total 3.92 X
109 ha (Table VII.1). The total land area of the world, exclud-
ing Antarctica, is 13.13 X 109 ha (WRI 1992). Thus, out of
approximately 200 nations or territories in the world, the four
case-study nations contain about 30% of the total land area.
They also play a dominant role in the amount of land in three
major vegetation and land-use types.
Closed forests: The four nations of this study contain approxi-
mately 1.26 X 109 ha or 35% of the world total which is 3.59 X
109 ha of closed forests (WRI 1992) (Table VII.1). Forest C
densities (TEC) range from 184 to 271 MgC/ha (Table VII.2).
Woodlands; The four nations have 0.45 X 109 ha or 26% of the
world total which is 1.70 X 109 ha of woodlands (WRI 1992) (Table
VII.1). Woodland C densities (TEC) range from 88 to 173 MgC/ha
(Table VII.2).
Croplands: The four case-study nations contain 0.57 X 109 ha of
croplands or 38% of the world total of 1.48 X 109 ha (WRI 1992)
(Table VII.1). Cropland C densities (TEC) range from 85 to 156
MgC/ha based mostly on soil C (Table VII.2).
Clearly, understanding the extent of the C dynamics of the four
nations in this report contributes significantly to an under-
standing of the terrestrial C dynamics of the world.
159
-------
Table VII.1. Land Area of Major Vegetation Types by Nation
Land Area (X 10® ha)
Vegetation
Type
FSU*
USA*
Mexico*
Brazil*
Totals
Forestland
Coniferous
Decid. Hardwd
Decid. Softwd
Total
461
24
138
623
Forested
201
Forest
71
Closed TMF
Other closed TMF
Total
352
16
368
1263
Woodland
Coniferous
Decid. Hardwd
Decid. Softwd
Total
150
11
30
191
Woodland
41
Woodland/
Scrub
66
Cerrado
156
454
Cropland
Arable
277
Cropland
155
Cropland
13
Agriculture in
degraded lands»
TMF
Closed For.
Savannah
Total
10
100
18
128
573
Sub-totals
1091
397
150
652
2290
Other types
1041
368
46
179
1634
Totals
2132
765
196
831
3924
Former Soviet Union from Table III.l. The sum of the area for foreBts and woodlands is 814 X 10 ha and is
equal to the sum of the three forest types in Table III.l which are subdivided here into their forest and
woodland components based on Kolchugina and Vinson (1993d) and Ryabchikov (1988).
'Conterminous United States from Table IV.1.
^Mexico from Table V.1.
'Brazil from Table VI.1. The total area for forest here is only for mature forests: i.e., does not include
secondary forests which are estimated to have covered in 1990 17.4 X 10 ha in the degraded TMF type plus
17.6 X It) ha in the degraded close forest type.
160
-------
Table VII.2.
Comparison of Carbon Densities Among Nations
C Density (MgC/ha)
Vegetation Type/
Component
Forest
Phytomass
Litter & CWD
Soil
Total
Woodlands
Phytomass
Litter & CWD
Soil
Total
FSU*
USAt
70
27
122
219
34
16
123
173
63
30
91
184
13
ND
79
92
Mexico*
64
7
118
189
20
2
77
99
Brazil6
152
17
102
271
26
2
60
88
Croplands
Phytomass
Litter & CWD
Soil
Total
3
0
153
156
5
0
105
110
6
1
78
85
8
0
72
80
'Values for forest and woodlands of the FSU are area-weighted means based on
areas in Table VII.7 and C densities in Tables III.2, III.3, and III.5.
Values for croplands are those for arable land in Tables III.l, III.2,
III.3 and III.5.
Values are area-weighted means for the conterminous United States as
presented in Table IV.1.
Values for Mexico are means as presented in Table V.7.
Values for Brazil on forest phytomass (i.e., of mature closed forests) and
agriculture soil are area-weighted means based on Tables VI.2 and VI.4;
the value for woodlands phytomass is a mid-point in the range 22-30
MgC/ha appearing in Table VI.4.
ND No datum
161
-------
A. Carbon Pools
The terrestrial ecosystems of the world have been estimated to
contain 560 PgC aboveground in phytomass, litter, and CWD;
belowground estimates are 1100 to 1400 PgC in roots and soil C
(Smith et al. 1993). The total range, therefore, is 1660 to 1960
PgC. The total aboveground and belowground C for the FSU,
conterminous U.S., Mexico, and Brazil is estimated in this
analysis to be 866 PgC (Table VII.3) or about 44-52% of the world
total. Thus, on 30% of the land area of the world represented by
the four nations, there is from 45% to 53% of the world's
phytomass and soil C pool. Because of its vast size and large
areas of C-rich forests, peatlands, and chernozem soils, the FSU
alone contains about a third of the world's total terrestrial C
pool with 601.1 PgC (Table VII.3).
For forests and woodlands, phytomass and soils (to a depth of 1
m) worldwide are estimated to contain 359 PgC and 787 PgC,
respectively, for a total of 1146 PgC (Dixon et al. 1994). The
estimated 354 PgC for the forests and woodlands of the four
nations of this analysis is about 30% of the worldwide forest-
woodland total (Table VII.3). The FSU and Brazil have the
largest forest C pools with 138 and 109 PgC, respectively. The
forest C pools of the U.S. and Mexico are significantly smaller
at 37 and 14 PgC, respectively.
Trends in C densities seem to follow expected changes from boreal
to tropical climates. For all nations, forests have more
phytomass than woodlands or croplands by more than three fold
(Table VII.2). The Brazilian moist tropical forest ecosystems
have the highest phytomass density of 152 MgC/ha, which is more
than double that for the more northerly forests. The latter
forests are remarkably similar in phytomass density to each other
162
-------
Table VII.3. Estimates of Carbon Pools for the Vegetation Types Common to
All Four Nations, 1990
C Pool (PgC)
Vegetation Type/
Component
FSU
USA*
Mexico*
Brazil1
Total
Forest
Phytomass
Litter & CWD
Soil
Total
43.3
19.2
75.7
138.2
12.6
6.0
18.3
36.9
4.6
0.5
8.4
13.5
64.0
6.6
38.3
108.9*
124.5
32.3
140.7
297.5
Woodlands
Phytomass
Litter & CWD
Soil
Total
6.4
3.0
23.5
32.9
0.5
ND
3.2
3.7
1.3
0.1
5.1
6.5
4.1
0.3
9.4
13.8
12.3
3.4
41.2
56.9
Croplands
Phytomass
Litter & CWD
Soil
Total
0.8
0
42.5
43.3
0.8
0
16.3
17.1
0.1
<0.1
1.0
1.2
1.0
0
9.2
10.2
2.7
<0.1
69.0
71.8
Total - all vegetation types
Phytomass 60.4
Litter & CWD 34.5
Soil 506.2
Total 601.1
16.0
6.1
64.7
86.8
73.1
8.2
71.7
153.0
155.7
49.6
660.7
866.0
*Values for forest, woodlands, and croplands of the FSU are the product of
areas in Tables VII.1 and C densities in Table VII.2 (some differences
due to rounding). Values for "all vegetation types" are from Tables
III.5, III.6 and III.7.
^Values for forest, woodlands, and croplands for the conterminous United
States are the product of areas in Table VII.1 and C densities in Table
VII.2 (some differences due to rounding). Values for all vegetation
types are from Table IV.2.
'Values for Mexico are meanB as presented in Table V.6.
'Values for forest (mature closed), woodlands, and croplands for Brazil are
the product of areas in Tables VII.1 and C densities in Table VII.2.
The national totals are means of three estimates appearing in Table
VI. 5.
'secondary forests are estimated to have an additional 4.7 PgC based upon
phytomass at 40 TgC/ha, litter/CWD at 2 TgC/ha, and soil at 92 TgC/ha
(Table VI.4) and a land area of 35.0 X 10* ha (Table VI.1).
ND No datum
163
-------
ranging (in MgC/ha)'from Mexico at 64 to U.S. at 63 to the FSU at
70. For litter and CWD, the forest ecosystems of the U.S. and
FSU are highest with estimates similar to each other at 30 and 25
MgC/ha, respectively. For soil C, the FSU is highest at 123
MgC/ha; however, estimates for the other nations are not too
different, ranging from 91 to 118 MgC/ha. The total C density,
however, remains highest in the Brazilian moist tropical forest,
i.e., 271 MgC/ha (Table VII.2).
For woodlands and croplands, the FSU ecosystems have the highest
C densities at 173 and 156 MgC/ha, respectively (Table VII.2).
For cropland phytomass alone, the FSU is lowest at 3 MgC/ha.
However, for soil C, these vegetation types in the FSU have
C-rich chernozem soils associated with the temperate forest
steppe region. These soils are very high in organic matter per
ha (123 MgC and 153 MgC) when compared to that of the other three
nations whose soil c ranges from 72 to 105 MgC/ha.
The C densities and pools appear to reflect the influence of
climate on terrestrial ecosystems (Table VII.2). That is, the
colder climates of the boreal and temperate latitudes have
ecosystems with more C stored belowground than above, while the
tropical ecosystems of Brazil under warm climates have the
opposite trend (Barbour et al. 1987). Mexico does not fit the
tropical pattern perhaps because of the strong human influence on
landscapes within this country. These relationships among the C
densities and pools of the world's ecosystems have historically
influenced their management by practices designed to provide
goods and services for human needs. As the world moves toward
managing more ecosystems for sustainable development, these
relationships will take on even greater significance.
164
-------
B. Carbon Flux
In this section, an effort is made to bring the flux estimates
from the case studies into a common framework. As has been
noted, there is considerable variation in the quantity of
information available for the different countries as well as
relative importance of different carbon flux processes.
Nevertheless, it will be important in advancing the implementa-
tion of the Framework Convention for Climate Change for cross-
country comparisons to be made. The focus here is on net flux.
Where ranges were reported in the case studies, a mid range value
was used here. In some cases, data from the case studies had to
be disaggregated; however, the net carbon flux values used here
are in agreement with the values in Tables III.11, IV.4, V.8,
VI.7, and VI.8.
The components of the land base flux (Table VII.4) have been
separated from the set of all flux terms (Table VII.5). The land
base flux is divided into a net biological term and a land-use
term. The net biological flux refers to the expected change in C
storage on the land base if no further anthropogenic impacts were
imposed. The three components of the net biological flux are the
net changes in the pools of phytomass, CWD (woody debris and
litter), and soil organic matter.
With the exception of soil and woody debris losses associated
with previous disturbances (inherited emissions), nonforest lands
are considered to be in C equilibrium with regard to the net
biological flux. The mature forest of Brazil and Mexico are also
considered to be in C equilibrium. Caveats regarding this
assumption have been discussed in the sections covering Brazil
and Mexico. The effect of these assumptions is that a consider-
able proportion of the land-base in each country is considered to
165
-------
Table VIZ.4.
Land Base Carbon Flux for the Four Case Studies*
Land Base
FSU
Net Biological1
Phytomass*
CWD® & Litter
Soil
Total
453*
1031
_ZiA
631
C Flux (Tg/yr)
USA
366**
-34»
0
332
Mexico
8.0
-44.8*
0.2
-36.6
ii
Brazil*
54™
-75m
-16
Land Use
Harvest Removals1
Harvest Emissions'
Conversion Emissions*
Agricultural Emissions
Total
-101°
-47"
0
-194*»
-342
-125®
-112"
-4"
-2"
-243
-2.6
-1.6**
-30.8*"
0
-35.0
-11
-132"'
-85"*
-235
Net
289n
89*
-72
-2 51*"
Unless otherwise noted, the estimates are taken directly from the case
study chapters. The approximate time period associated with the
estimates varies somewhat by country. For the FSU the period is
the late 1980s; for the USA it is the early 1990s; for Mexico it
is the 1980s; and for Brazil it is 1990. This net land base flux
also appears on Table VII.5.
f Net biological is the change in the pool, assuming no harvest or land
use change.
* Tree growth less mortality.
8 Coarse woody debris. Includes decomposition on formerly disturbed lands.
11 Harvest removals refer only to growing stock removed from the forest.
* Harvest emissions refer only to burning of tree C at the time of
harvest.
* Conversion emissions refer to burning of tree, CWD, or soil C at the
time of conversion.
* 520 (phytomass, Table III.11) - 67 (fires, Table III.11 and discussion
in Section III.D.l.c, p. 44-46) = 453.
1 118 (CWD, Table III. 11) - 15 (fires, Table III.11 and discussion in
Section III.D.l.c) = 103, litter flux is not included.
A 53 (peat, Table III. 11) + 5 (non-forest, soil, Table III. 11) + 40
(forests, soil Table III.11) - 23 (fires, Table III.11 and
discussion in Section III.D.l.c) = 75.
A 389 Mm1 (Section III.D.l.c) of merchantable wood is harvested. This
converts to 101 Tg using 260 TgC/m3 (Section III.D.l.c).
" 389 Mm* of merchantable wood converts to 148 Tg of tree carbon (Section
III.D.l.c). 148 (tree C) - 101 (merchantable C) = 47 (slash C).
100 (peat combustion, Table III.11) + 94 (agricultural land emission,
Table III.11) = 194.
' 292 (Table III.11) - [-30 CCH, emissions, Table III.11] - 33 (increase
in products pool accounted for in Table III.11) = 289.
: Tree + understory, Table IV.4.
1 Table IV.4.
1 -28 (litter, Table IV.4) + [-7 (understory, Table IV.4)] + [-77 (tree C,
Table IV.4)] = (-112).
166
(continued)
-------
Table VII.4 Continued
«•
**
~~
2:
A1
ttt
ttt
ss
Section IV.C.l.d.
Table IV.5.
Table IV.5, 95 (forest land base) - 4 (conversion emissions) - 2
(agricultural land base) = 89.
Values from Table V.8.
-45 (inherited emissions) + 0.2 (CWD + litter accumulation) = (-44.8).
-2.6 (merchantable) x 1.6 (assumed ratio of tree C to merchantable C)
-4.2 (tree C harvested. Then (-4.2) - [-2.6, merchantable] = -1.6
(slash).
Prompt emissions.
Based on Tables VI.7 (Olson biomass densities) and VI.9.
44 (secondary forests) + 10 (pastureland) = 54.
18 (secondary forests) + (-93, pastureland) = -75.
15 (secondary forests) + (-10, tropical moist forest) = 5.
-97 (tropical moist forests + [(-35), cerrado] = -132.
-75 (degraded closed forest) + [(-10), degraded savannah] = -85.
-233 (Table VI.7) - [(-18), harvest removals and harvest emissions] =
(-251).
167
-------
Table VII.5.
Net Carbon Flux for the Four Case Studies
C Flux (Tg/yr)
Component
Land base*
Forest
products
Energy*
Cement*
Net
FSU*
USA*
289
33
-1,020
-19
-717
89
36
-1/296
-10
-1,181
Mexico*
-72
1*
-74
-3
-148
Brazil1
-251
3
-50
-3
-301
Total
55
73
-2,440
-35
-2,347
'Subak et al. 1993
Values from Table III.11.
^Values from Table IV.5.
^Values from Table v.8. Net flux of -148 here compares to -148.9
reported by Makundi et al. (1992) in Table V.8.
'Values from Table VI.8.
'One half of the input to the forest products pool is assumed to be
accumulating.
"Values from Table VII.4 in this row only.
168
-------
be in C equilibrium (43% for FSU, 54% for the U.S., 86% for
Mexico, and 86% for Brazil).
Considering only the net biological flux, the change in the
phytomass pool—a balance of woody phytomass growth and mortal-
ity—is positive and relatively high in the FSU, the U.S., and
Brazil because large areas of young forests promote C accumula-
tion. The origin of these fluxes differs greatly: in the U.S.,
extensive logging? in FSU, extensive wildfire and anthropogenic
disturbances? and in Brazil, extensive deforestation followed by
reversion after abandonment. The annual change in CWD storage—a
balance of natural tree mortality and decomposition—is negative
in three of the nations. In the U.S. decomposition of woody
residues in young secondary forests determines the sign of the
flux. Woody debris is also a relatively large source of C in
Brazil and Mexico because of decomposition of woody residue on
pasturelands created within the last 10 years. The large C sink
in the CWD pool of the FSU would be approximately 25 TgC/yr lower
if the decomposition of logging residue was modeled as it was in
the U.S. rather than as a component of the harvest emissions.
The remaining CWD sink may be in part a function of a generally
cooler climate which tends to slow the rate of decomposition.
The soil organic matter pool is a source in areas of active
deforestation because of the imbalance of inputs from litter fall
and fine root turnover and outputs from decomposition (i.e.
inherited emissions). In Brazil, these soil emissions are
counteracted by a large accumulation assumed to be occurring in
the young secondary forests. The FSU budget also indicates a
gain in soil C associated with secondary forests as well as a
large sink from peat formation.
Besides the net biological flux, the other main land base flux is
that associated with anthropogenic land uses. The components are
(1) removal of wood during commercial forestry, (2) direct
169
-------
emissions from phytomass burning (i.e. deforestation in the
tropics and slash burns after timber harvests in the temperate
and boreal countries), and (3) soil C emissions from tillage of
agricultural soils (Table VII.4). Utilization of peatlands for
fuel was included in the agricultural emissions.
Planned harvest removals were highest in the U.S. and Russia. In
both cases over 100 TgC/yr of wood was removed from the forest
land base. Smaller harvest removals in Brazil and Mexico reflect
less industrial development, lack of infrastructure, few species
identified to be commercially valuable, and less use of timber in
construction. The direct emissions associated with commercial
harvest are likewise largest in the U.S. and Russia because of
the level of logging. Fuelwood removals are greater than
commercial logging removals in the tropics but we have not made
an attempt to isolate this flux; it tends to be a dispersed
activity and is subsumed to some degree by other components of
the C budgets.
The conversion emissions are greatest in Brazil (-132 Tg/yr)
because of its high deforestation rate. Mexico also has a large
source associated with deforestation (-31 Tg/yr). The land bases
in the U.S. and FSU are relatively stable; thus, the land
conversion emissions are quite low.
Agricultural lands in Russia are a 94 Tg/yr source and emissions
from peatland utilization adds another 100 Tg/yr source.
Agricultural lands in Brazil are extensive and represent a 85
Tg/yr source. Losses of soil C from tillage of agricultural
lands in Mexico were not estimated separately here, but some
losses may be captured in the estimate of inherited emissions.
Much of the agricultural land in the U.S. has been under
cultivation for many decades and, except for erosion, is probably
approaching a steady state or even increasing due to improved
soil conservation practices.
170
-------
The sum of the four components of the land-use flux is largest in
FSU (-342 Tg/yr), followed by the U.S. (-243), then Brazil (-235
Tg/yr), and Mexico (-35 Tg/yr). The net effect of the biological
and anthropogenic factors on the land base is a C sink in the FSU
and the U.S. and a C source in the tropical countries (Table
VII.4) .
In the complete country-level budgets, four additional flux
components are included (Table VII.5). The accumulation of C in
forest products still in use is greatest in the U.S. and FSU,
which follows from the more extensive commercial logging in those
countries. Accumulation in landfills is greatest in the U.S.,
but this flux is a relatively small term in these budgets.
Emissions from energy use and production dominate the calculation
of the U.S. and FSU budgets.
In all cases, C sources exceed C sinks, i.e. all countries are
net emitters of C. The U.S. is the largest emitter because of
its high per capita industrial emissions and large population.
FSU has a similar distribution of sources and sinks, and like the
U.S., is a large net emitter. Brazil is notable in that the net
land-use source is larger than the industrial source. In Mexico,
industrial emissions are of about the same magnitude as the net
land base source. The summed net emissions from these four
countries (230 Tg/yr) represents about 30% of the estimated
global anthropogenic C02-C emission.
C. Policy Perspective and Prospects for
Stabilization of Emissions
All four countries considered in this report make a significant
contribution to global greenhouse gas emissions (Table VII.6) and
have signed the Framework Convention on Climate Change (FCCC).
Thus, they are committed to the policy stated in the FCCC, namely
171
-------
Table VII.6 Greenhouse Harming Potentials
Country
TgC as C02
Equivalents (100-yr time horizon)*
co2
CH*
N20
CFCs
Total
FSU
827
100
19
132
1,078
USA
1,241
98
56
357
1,752
Mexico
133
11
3
10
157
Brazil
269
38
10
13
330
Total
2,470
247
88
512
3,317
World
6,410
791
278
1,303
8,782
*Subak et al. 1993.
172
-------
to stabilize atmospheric concentrations of greenhouse gases at
levels that pose no threat to the planet. The quantitative
meaning of this commitment and the policies that may be required
for its implementation remains a subject for negotiation through
the International Negotiating Committee.
Given the uncertainty regarding the meaning and implementation of
the FCCC policy, no country has committed to a specific, long-
term target for emissions. The U.S. has committed to an
intermediate goal of reducing net emissions to the 1990 level by
the year 2000. This goal, in conjunction with estimates of
"business as usual" increases in emissions of greenhouse gases,
commits the U.S. to off-setting 106 TgC/yr through a combination
of decreased emissions and increased sinks (Clinton and Gore
1993). The lack of quantitative emission goals, and the lack of
estimates of "business as usual" increases in emissions of
greenhouse gases (a topic not considered in this report),
restricts discussion of the potential for stabilizing emissions
in the other countries to very general terms.
The four countries studied in this report differ greatly in the
relative extent of C02 flux associated with the land base. In
Mexico and Brazil the land base is a carbon source and that
source appears to more than double emissions from other sources
(see Table VII.4, VII.5). In these countries, land-use policies
could potentially contribute substantially to implementing the
FCCC policy.
In contrast, the land base plays a relatively small role in the
net emissions from the U.S. Sequestration of carbon in terres-
trial ecosystems and in forest products offsets less than 10% of
the greenhouse gases emitted from other sources. Thus, the
opportunities for conserving and sequestering C in the U.S.
through changes in land use are relatively small. Nevertheless,
land use may have a substantial role (i.e. 10-20%) in meeting the
173
-------
intermediate national goal of implementing 106 TgC/yr of reduced-
C-emission/increase-C-sequestration by the year 2000 (Clinton and
Gore 1993) and in sustaining this achievement beyond 2000. In
the absence of strong policy shifts over the next few decades, it
appears that land use will play a continually decreasing role in
offsetting industrial C emissions in the U.S.
In FSU, forest growth is estimated to offset up to 30% of
emissions from other sources. Given the large forest land base,
the opportunities to increase carbon sequestration rates appear
to be significant. ' A possible counteracting factor will be
potentially large increases in the harvest levels which may
produce a carbon source in the near term.
174
-------
VIII. Uncertainties and Research Needs
A. Global
In assessing C flux due to land-cover change, data are required
for (1) classifying land cover and its relative change over time
and (2) phytomass, or TEC, mass per unit area of the various
land-cover classes—C density. Until recently, the most uncertain
factor has been the accurate characterization of areal extent of
land-cover classes through time and its resultant rate of land-
cover conversion. Due to recent advances in satellite remote
sensing and GIS technology, increasingly more accurate estimates
of terrestrial land-cover change are becoming widely available.
The C density and its change over time is now the limiting factor
in accurate estimates of C flux due to changes in terrestrial
biota. Because pasturelands and croplands typically possess C
densities that are from one to two orders of magnitude less than
the forest ecosystems they replace, attention is justifiably
focused on the forest systems.
As interest in global C dynamics has grown in recent years,
estimates of C pools and flux are required at larger spatial
scales such as regions, nations, subcontinents, and even
continents. Investigative research to provide the estimates have
commonly relied on aggregating small-scale values for key
components through multiplication and summation operations. The
case studies in this analysis serve as examples. In simple
terms, the total terrestrial C pool of each country was estimated
by multiplying the mass of C/ha in phytomass and soil (usually
from limited field data) for each vegetation type times number of
hectares in that type (from vegetation maps), and then summing
all products for a national estimate.
175
-------
A concern is the degree of uncertainty in the estimates of C
pools and flux at these large spatial scales. Robinson (1989)
states "...error enters estimation through uncertainties and
biases in the data, and compounds as data are combined to draw
(often) lengthy chains of inferences." The subject is also
referred to as error propagation.
Theoretically, if the variance around the value of each component
in the chain calculation were known, statistical methods exist to
determine the level of uncertainty for the estimate. The smaller
the variances the less the overall uncertainty. Small variances,
however, depend upon large representative samples, preferably
taken randomly and with uniform methodology, in each homogeneous
unit of every component in the chain. At the large spatial
scales in national or global analyses, obtaining data in such a
manner is currently not possible.
In terrestrial ecology, the issue is further complicated by a
shortage of information on the belowground C density. These
components have typically been estimated by applying factors
developed from an even more limited amount of soil sampling
around the world to the aboveground phytomass.
In time preliminary calculations of uncertainty in large-scale
estimates of C pools and flux can be undertaken. Ongoing
research is providing better data. For instance, more studies
are focusing on the C density of phytomass and soil in more areas
of the world. Better vegetation maps are being produced as a
result of improved use of remote sensing from satellites together
with increased ground-truth validation. Standard methodology is
developing. One early improvement would be at least to have
ranges around the values for chain components. Ranges along with
clues about frequency distributions would enable first approxima-
tions in computations of uncertainty.
176
-------
At present, most of the values in this document that are used for
estimating national C pools and flux are based upon very limited
data along with a few assumptions as discussed in each section.
In general, the uncertainty issue underscores the research needs
outlined above to make significant advances in the science of
global environmental change in the next few decades. The
modeling approaches, assumptions and data bases used for these
analyses are supported by scientific literature as referenced.
Projections of C dynamics compared favorably with those of other
researchers.
Carbon density is most adequately understood in standing live
aboveground phytomass because of its association with commer-
cially important wood volumes. However, many forest ecosystems
hold significant amounts of C in CWD aboveground and belowground
in soils and roots. Globally, these C budget compartments offer
areas for future research.
Processes important to understanding C dynamics and reducing C
budget uncertainties include (1) phytomass accretion in regrowth
forests following abandonment of pastureland or other agricul-
tural uses, (2) carbon dynamics in small scale (1-2 ha) land uses
causing forest degradation and subsequent recovery (e.g., milpa
agriculture in Latin America), and (3) growth in mature forests
which have been assumed to be in C equilibrium with the atmo-
sphere .
Reducing uncertainties associated with the various forest
ecosystem compartments and with the internal forest processes
will only be achieved through intensive and extensive field
research to validate and calibrate the models currently used to
assess national and global C flux.
177
-------
B. Boreal Zone
Accumulation of C in CWD, soil organic natter and peat is quite
important in boreal countries because cool wet climates tend to
slow decomposition rates. The frequency, areal extent, and
intensity of forest fires is also critical because of the
propensity for catastrophic wildfire. Because these forests may
not have been of commercial interest, due to slow growth rates or
inaccessibility, comprehensive information relevant to forest C
budgets may not have been generated or be available. Data needs
include information on forest inventories, CWD pools, allometric
relationships, and rates of logging. Postharvest dynamics of
soil C may be of particularly interest in the boreal zone because
of the increased potential for soil waterlogging, another factor
that may slow decomposition.
C. Temperate Zone
Large uncertainties are inevitably associated with working at the
spatial and temporal scale of C budget studies at the national
scale, and a number of specific issues relating to forest C
dynamics need increased attention. In the temperate zone, the
importance of an age-class based inventory has been discussed.
Because of the economic importance of the forest sector, many
countries appear to have reasonably complete inventories.
However, approaches to inventory development and sampling
intensity vary widely. Intercomparisons among approaches would
be useful. For the U.S. in particular, an improved forest
inventory on public lands is needed since historically the USDA
Forest Service's Forest Inventory and Analysis units have
surveyed only private timberland in the western U.S. where mostly
the federal timberland is located (USDA 1992b). More generally,
the inventory on forest lands reserved for habitat conservation
178
-------
and recreation may not be tracked well because of the relatively
low economic importance of these lands.
A better understanding of soil C and woody debris dynamics
following harvest is also needed. Assumptions of significant
losses in soil C after harvest, and gains later in stand
development, which have been employed in earlier C budget models
(Houghton et al. 1983? Heath and Birdsey 1993) may not be
warranted in temperate zone forests. The dynamics of woody
debris are also highly uncertain and will benefit from additional
observations during forest surveys as well as more complex
modeling approaches that account for stand history.
D. Tropical Zone
To date, C budgets for tropical nations or regions are based upon
limited amounts of published data. Research is required to
substantially increase the availability of data vital to more
definitive estimates of C dynamics in tropical terrestrial
ecosystems.
The C budgets of Mexico and Brazil presented here point to
several key research needs. First, C budgets would be improved
with more precise data on the areal extent of vegetation types
and various land uses. The size of such areas is not clearly
known in many tropical regions. In Mexico for instance, it is
unclear how much agricultural land is used for pastureland versus
the amount used for croplands.
Second, C density is another important general area of needed
research. Within the cover and land-use types of the tropics,
there is a general lack of data on C density as well as on the
temporal changes caused by growth and disturbance, both natural
and human-caused. Two vegetation cover types are of particular
179
-------
interest, particularly in Brazil and Mexico: the mature and the
secondary forests, which are both quite important to the overall
accounting of C because of both their C density and their size.
For example, as noted in the Brazil section, despite high
deforestation rates, there remains over 350 X 106 ha of mature
forests. It has been suggested that many mature tropical forests
are not in equilibrium, but actually experience net C accretion
in response to natural and human-caused disturbances (Lugo and
Brown 1992). Because of the large land areas involved, there is
an important need to clearly determine the rate of change in C
density of mature forests even if the annual rate of C sequestra-
tion per ha is small.
Both Brazil and Mexico have large areas of secondary forests
resulting from past deforestation and other human disturbance
(-70 and -20 X 106 ha, respectively). Yet little is known about
the exact extent of secondary forests in these and other tropical
nations. Further, the C density or annual C uptake of these
forests is not well known. Quantifying these secondary forest
processes in tropical nations such as Brazil and Mexico is a
vital missing segment of the world's terrestrial C budget (Skole
et al. 1994).
Several other topics on which the availability of additional data
would improve tropical C budgets warrant research attention.
These include rates of land conversion such as deforestation,
soil C contents and their rates of change following ecosystem
disturbance, long-term C storage in wood products and landfills,
the role of C transport by rivers (particularly for the Amazon
basin), and assumptions about C equilibriums for the vegetation
and land-use types as discussed in the Brazil and Mexico
sections.
180
-------
E. Carbon Dynamics Through Modeled Projections
The global C dynamics will likely continue to be impacted for at
least several decades by human activities that cause high rates
of C02 emissions and changes in vegetation and soil C pools. The
dynamics of C at various scales need to be adequately projected
to better understand the long-term consequences of these
continued anthropogenic influences. Various vegetation and
land-use change models are useful tools for projecting the
results of changing C pools and flux, especially in forests.
Examples follow.
1. Individual Tree Growth Models
Individual tree growth models have been developed that are based
on physiological processes (Dixon et al. 1990). TREGRO
(Weinstein et al. 1991) is an example of this type of model.
TREGRO simulates flux of C, water and nutrients through a
tree-soil continuum. Photosynthesis is calculated on an hourly
basis, determined by the availability of resources such as light,
water, and nutrients. Carbon is partitioned to various compo-
nents based on phenology and the relative requirements of fixed C
by the individual components. Tree growth can be simulated over
a time frame of years.
Tree growth models do not simulate forest stand dynamics because
they only consider growth of individual trees. They also do not
consider ecological processes such as nutrient cycling or
disturbance. However, individual tree growth models can be used
to parameterize other models that do evaluate the vegetation
dynamics of forest stands.
181
-------
2. Forest Gap Models
Forest gap models were developed to simulate plant succession in
a forest canopy opening (ca. 0.1 ha) caused by mature tree death
or fall (Shugart et al. 1992). Tree growth is simulated by
reference to empirical relationships among the growing tree,
growing-degree days, and solar radiation. Recent developments
include the addition of soil nutrients (e.g., FORTNITE, Aber and
Melillo 1982; FORTNUT, Weinstein et al. 1982; LINKAGES, Pastor
and Post 1985, 1986), soil water (FORENA, Solomon and Shugart
1984, Solomon 1986), seed dispersal (ZELIG, Urban et al. 1991),
and high latitude environment (LOKI, Bonan 1991b).
The advanced forest gap models offer many advantages in assessing
C pools and flux (King 1993a). These models couple ecological
processes such as disturbance, nutrient cycling, water balance,
and tree competition. Thus, they can simulate the response of
forest stands to a changing environment. Disadvantages are that
they require species-specific parameterization which can be quite
detailed, tree migration is not simulated, and modeling the
direct effects of atmospheric C02 is still being refined.
3. Biogeographic Models
Biogeographic models have been developed to simulate changes in
natural vegetation at regional to global scales (Smith et al.
1993). The Holdridge Life Zone Classification System is an
example of the simplest models, and it has been used to simulate
the effects of climate change on the global distribution of
vegetation (Prentice 1990, Smith et al. 1992). More sophisti-
cated models are now available such as BIOME (Prentice et al.
1992), CCVM (Lenihan 1992), and MAPPS (Neilson et al. 1992).
182
-------
BIOME, CCVM, and MAPPS have certain advantages: they are based on
plant physiological principles and their simulations of vegeta-
tion distribution, based on changing climates, are likely to be
more robust than those based only on correlation (King 1993a).
The mechanistic treatment of water balance allows an evaluation
of C02 effects on plant distribution. Also, since these are
broad-scale models, they can be used as tools for assessing
forest management effects on C dynamics. The main disadvantages
of these models are that they were developed for equilibrium
approaches and fail to simulate transient vegetation response to
climate change.
4. Ecosystem Process Models
Several vegetation productivity models are being used to examine
the potential effects of global change on NPP and biogeochemical
cycles. These models attempt to simulate the processes of plant
energy dynamics and C balance at the canopy level. The consider-
ation of ecosystem C dynamics allows the simulation of changes in
net C flux for a given location under changing climatic condi-
tions (Smith et al. 1993). Such information could be valuable
for large-scale C budget analysis if coupled with spreadsheet
models such as COPATH (Makundi et al. 1991). Ecosystem models
such as CENTURY (Parton et al. 1988), FOREST-BGC (Running and
Coughlan 1988), and GEM (Rastetter et al. 1991) are useful for C
dynamics evaluations because they are based on plant physiologi-
cal and ecosystem processes.
The main advantage of ecosystem models is that they couple
vegetation productivity with water balance and nutrient avail-
ability. In GEM, C allocation and dynamics in response to
changing climate can be simulated from year to year (Rastetter et
al. 1992). Ecosystem models require an enormous number of
parameters, many which are lacking data or poorly understood.
183
-------
5. Modeling Terrestrial Carbon Dynamics
Recent modeling efforts have investigated the potential transient
response of vegetation to climate change and the impact on
terrestrial C pools and flux. Regional and global C models are
used to simulate C dynamics based on various climate scenarios.
Examples include PULSE (Neilson 1990; King and Neilson 1992),
IMAGE 2 (Vloedbeld and Leemans 1993), and the Holdridge bio-
climatic model (Smith and Shugart 1993).
The PULSE model tracks forest tree dieback and regrowth to
estimate the possible magnitude of C flux (Neilson 1990; King and
Neilson 1992). Carbon loss to the atmosphere is simulated
through an exponential decay process, while C sequestration by
forest trees is simulated by a logistic function constrained by
maximum C density. PULSE was developed for non-tropical forests.
The IMAGE 2 model projects changes in land cover as well as plant
physiological processes and industrial production to simulate C
flux between the terrestrial biosphere and atmosphere (Vloedbeld
and Leemans 1993). IMAGE 2 explores C flux based on atmospheric
C02 concentrations, water-use efficiency, plant photosynthesis
and respiration, and changes in vegetation cover and agricul-
tural-use patterns with the subsequent effects on land-use
change.
In another global modeling effort, Smith and Shugart (1993) used
the Holdridge Life Zone Classification to simulate the potential
impacts of global climate change on patterns of potential
terrestrial C storage and flux with the atmosphere. The
Holdridge bioclimatic model relates current vegetation distribu-
tion to global climate patterns and potential impacts of climate
change. Climate change scenarios were based on general circula-
tion model predictions for an atmosphere with doubled C02
184
-------
concentration. Current estimates of C pools in vegetation and
soil are then used to calculate changes in potential terrestrial
C reserves given the climate change scenario.
6. Land-Use Change Models
The models presented above are used to assess C dynamics at
various temporal and spatial scales. The function of these
models is based mainly on climate, vegetation distribution,
and/or plant physiological processes. Only the IMAGE 2 model
includes land-use change as a parameter in calculating changes in
C pools and flux (Vloedbeld and Leemans 1993). IMAGE 2 is unique
in its ability to simulate changes in land cover and land use by
considering socioeconomic factors such as regional population and
gross national product (Vloedbeld and Leemans 1993). These
drivers set the demand for forestland conversion to agricultural
land. The transitions from original land cover to agricultural
land is used to determine the immediate C flux into the atmo-
sphere with the emission rate dependent on environmental factors
and the C content of the various pools.
7. Spreadsheet Modeling of Carbon Dynamics
The impact of terrestrial ecosystem processes on climatic change
and C cycling is significant. However, several efforts to
quantify and balance the global C budget have not been success-
ful. These attempts have been constrained due to unreliable data
on the role of vegetation and soil as a net C source or sink and
the rate of land-use change. One way to balance the global C
budget is the effort to understand the C dynamics at the regional
and country scale. One approach is the COPATH spreadsheet model
(Makundi et al. 1991). The spreadsheet format allows entry of
inventory data such as forest type, land area, and various C
pools.
185
-------
COPATH is a unified approach to estimate C pools, flux and the
rate of deforestation in the major forested countries of the
world. The aggregation of the country-level C assessments can be
used to assess climate and land-use implications at the global
scale. A disadvantage of COPATH is that process-based simula-
tions such as vegetation response to increasing atmospheric C02
or temperature changes cannot be assessed. However, COPATH could
be linked with processed-based ecosystem or biogeographical
models to overcome this limitation.
8. Modeling of Carbon Dynamics! In Conclusion
The models suggested above are only examples, and they are not
all inclusive regarding available models. Detailed discussions
of modeling C dynamics as it relates to global change and forests
are presented by King (1993a) and Smith et al. (1993). Over
time, projections of forest C pools and flux through modeling and
with improved data will further advance the understanding of
global C dynamics as well as how it is influenced by human
activities.
186
-------
IX. Literature Cited
Aber, J.D., and J.M. Melillo. 1982. FORTNITE: A computer model of
organic matter and nitrogen dynamics in forest ecosystems.
Research Bulletin R3130. University of Wisconsin, Madison,
WI.
Adams, D.M., R.J. Alig, and D.K. Anderson. 1992. Future prospects
for western Washington's timber supply. University of
Washington, College of Forest Resources, Seattle, WA.
Agafonov, A.F. 1989. Drying of pure pine stands on fire areas.
Lesnoe Khozyaistvo 12: 37-38.
Alban, D.H., and D.A. Perala. 1992. Carbon storage in Lake States
aspen ecosystems. Canadian Journal of Forest Research 22:
1107-1110.
Aleksandrova, V.D. 1977. Geobotanical regionalization of Arctic
and Antarctic. Nauka Press, Leningrad.
Alig, R.J., W.G. Hohenstein, 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. Forest
Service General Technical Report SE-64. U.S. Department of
Agriculture, Forest Service, Southeastern Forest Experiment
Station, Asheville, NC.
Alimov, Y.P., I.V. Golovikhin, L.B. Zdanevich, and I.V. Yunov
(eds.). 1989. Dynamics of forests under forest management
organization regarding the main forest forming species in
1966-1988. U.S.S.R. State Forestry Committee, Moscow.
Allan, T., and J.P. Lanly. 1991. Overview of status and trends of
world forests. Paper No. 3.1 in Proceedings of a Technical
Workshop to Explore Options for Global Forest Management,24-
30 April 1991. Bangkok, Thailand.
Anderson, J.M., and T. Spencer. 1991. Carbon, nutrient, and water
balances of tropical rain forest ecosystems subject to
disturbance: Management implications and research proposals.
MAB Digest 7. UNESCO, Paris, France.
Armentano, T.V., and C.W. Ralston. 1980. The role of temperate
zone forests in the world carbon cycle. Canadian Journal of
Forest Research 10: 53-60.
Auclair, A.N.D. 1985. Postfire regeneration of plant and soil
organic pools in Picea mariana - Cladonia stellaria ecosys-
tems. Canadian Journal of Forest Research 15: 291-297.
187
-------
Barbour, M.G., J.H. Burk, and W.D. Pitts. 1987. Terrestrial plant
ecology. Benjamin/Cummings Publishing Company, Menlo Park,
CA.
Barker, J.R., G.A. Baumgardner, D.P. Turner, and J.J. Lee. 1994.
Carbon sequestration potential of the Conservation and
Wetland Reserve Programs. Submitted to Environmental
Management.
Bazilevich, N.I. 1986. Biological productivity of soil-vegetation
formations in the U.S.S.R. Bulletin of Academy of Sciences
of the U.S.S.R., Geographical Series 2: 49-66.
Bazilevich, N.I., O.S. Grebenshikov, and A.A. Tyshkov. 1986.
Natural association of geographical patterns with the
structure and functioning of ecosystems. Nauka Press,
Moscow.
Bellon, M.R. 1991. The ethnoecology of maize variety management:
A case study from Mexico. Human Ecology 19: 389-418.
Birdsey, R.A. 1992. Carbon storage and accumulation in United
States forest ecosystems. USDA Forest Service General
Technical Report wo-57. U.S. Department of Agriculture,
Forest Service.
Birdsey, R.A., A.J. Plantinga, and L.S. Heath. 1993. Past and
prospective carbon storage in United States forests. Forest
Ecology and Management 58: 33-40.
Blake, D.R., and F.S. Rowland. 1988. Continuing worldwide
increase in tropospheric methane, 1978-1987. Science 239:
1129-1131.
Bohn, H.L. 1982. Estimates of organic carbon in world soils: II.
Journal of Soil Science Society of Arizona 4: 1118-1119.
Bonan, G.B. 199la. Boreal forests, the carbon cycle and global
change: A challenge for ecologists. Pages 139-153 in T.
Kolchugina and T. Vinson (eds.) Proceedings of the workshop
on carbon cycling in boreal forests and subarctic ecosys-
tems, 9-14 September. Corvallis, OR.
Bonan, G.B. 1991b. Seasonal and annual C fluxes in a boreal
forest landscape. Journal of Geophysical Research 96:
7032-7312. "
Bormann, F.H., and G.E. Likens. 1979. Pattern and processes in a
forested ecosystem. Springer-Verlag, New York, NY.
188
-------
Botch, M.S., K.I. Kobak, T.S. Vinson, and T.P. Kolchugina. 1995.
Carbon pools and accumulation in peatlands of the former
Soviet Union. Global Biogeochemical Cycles 9(1): 37-46.
Brown, S., and A.E. Lugo. 1982. The storage and production of
organic matter in tropical forests and their role in global
carbon cycle. Biotropica 14: 161-187.
Brown, S., and A.E. Lugo. 1990. Tropical secondary forests.
Journal of Tropical Ecology 6: 1-32.
Brown, S., and A.E. Lugo. 1992. Aboveground biomass estimates for
tropical moist forests of the Brazilian Amazon. Interciencia
17: 8-18.
Brown, S., A.E. Lugo, and J. Chapman. 1986. Biomass of tropical
tree plantations and its implications for the global carbon
budget. Canadian Journal of Forestry Research 16: 390-394.
Brown, S., A.J.R. Gillespie, and A.E. Lugo. 1989. Biomass
estimation methods for tropical forests with application to
forest inventory data. Forest Science 35: 881-902.
Brown, S., A.E. Lugo, and L.R. Iverson. 1992. Processes and lands
for sequestering carbon in the tropical forest landscape.
Water, Air and Soil Pollution 64: 139-155.
Cairns, M.A. 1994. Land cover characterization for carbon
dynamics assessment in southeast Mexico. Presented at
IGBP-DIS Land Cover Working Group workshop on strategies for
a fast-track 1-km global land cover characterization, 23-24
February. Las Vegas, NV.
Cairns, M.A., R. Dirzo, and F. Zadroga. 1995. Forests of Mexico:
A diminishing resource. Submitted to Journal of Forestry.
Camara, I. de G. 1991. Piano de Acao Para a Mata Atlantica.
Fundacao SOS Mata Atlantica, Sao Paulo.
Cannell, M.G.R. 1982. World forest biomass and primary production
data. Academic Press, London.
Castellanos, J., M. Maass, and J. Kummerow. 1991. Root biomass of
dry deciduous tropical forest in Mexico. Plant and Soil 131:
225-228.
Cerri, C.C., B. Volkoff, and F. Andreaux. 1991. Nature and
behavior of organic matter in soils under natural forest and
after deforestation burning and cultivation, near Manaus.
Forest Ecology and Management 38: 247-257.
189
-------
Cherdantsev, G.N. 1961., Map: Arable land in the U.S.S.R. in 1954.
Scale 1:50,000,000. Page 85 in J.P. Cole (ed.) A geography
of the U.S.S.R.- Background to a planned economy. F.C.
German, Butterworths, London, UK.
Chudnoff, M. 1984. Tropical timbers of the world. Agriculture
Handbook Number 607. U.S. Department of Agriculture, Forest
Service, Washington, DC.
Clawson, M. 1979. Forests in the long sweep of American history.
Science 204: 1168-1174.
Clinton, W.J., and A. Gore, Jr. 1993. The climate action plan.
The White House, Washington, DC. October.
CNIF (Camara Nacional de la Industria Forestal). 1992. Memoria
Economica. Production data. Mexico, D.F., Mexico.
Cost, N.D., J.O. 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. U.S. Department of Agriculture, Forest
Service.
Cottle, L.E., G.F. Schreuder, and A.A.A. de Barros. 1990. Brazil:
A country profile of the forests and forest industries.
Working Paper 27. Center for International Trade in Forest
Products, University of Washington, Seattle, WA.
Coutinho, L.M. 1982. Ecological effects of fire in Brazilian
cerrado. Pages 273-291 in B.J. Huntley and B.H. Walker
(eds.) Ecology of tropical savannas. Springer-Verlag, New
York, NY.
Crutzen, P.J., and M.O. Andreae. 1990. Biomass burning in the
tropics: Impact on atmospheric chemistry and biogeochemical
cycles. Science 250: 1669-1678.
Cuardn, A.D. 1991. Conservacibn de los primates y sus habitats en
el sur de Mexico. Masters thesis. Sistema de Estudios de
Posgrado, Universidad Nacional, Heredia, Costa Rica.
Davidson, E. A., and I. L. Ackerman. 1993. Changes in soil carbon
inventories following cultivation of previously untilled
soils. Biogeochemistry 20: 161-193.
Davidson, E.A., D. Nepstad, and S.E. Trumbore. 1992. Completing
belowground carbon budgets for pastures, recovering forests,
and mature forests in Amazonia. Research proposal submitted
to the Terrestrial Ecology Program, Ecosystems Dynamics and
Biogeochemical Cycles Branch, National Aeronautics and Space
Administration, Washington, DC.
190
-------
Detwiler, R.P. 1986. Land-use change and the global carbon cycle:
The role of tropical soils. Biogeochemistry 2: 67-93.
Detwiler, R.P., and C.A.S. Hall. 1988. Tropical forests and the
global carbon cycle. Science 239: 42-47.
Dirzo, R., and M.C. Garcia. 1992. Rates of deforestation in Los
Tuxtlas, a neotropical area in southeast Mexico. Conserva-
tion Biology 6: 84-90.
Dixon, R.K, and D.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., J.K. Winjum, and P.E. Schroeder. 1993. Conservation
and sequestration of carbon: The potential of forest and
agroforest management practices. Global Environmental Change
3: 159-173.
Dixon, R.K., P.E. Schroeder, and J.K. Winjum. 1991. Assessment of
promising forest management practices and technologies for
enhancing the conservation and sequestration of atmospheric
carbon and their costs at the site level. EPA/600/3-91/067.
U.S. Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR.
Dixon, R.K., R.S. Meldahl, G.A. Ruarkk, and W.G. Warren, (eds.).
1990. Process modeling of forest growth responses to
environmental stress. Timber Press, Portland, OR.
Dixon, R.K., S. Brown, R.A. Houghton, A.M. Solomon, M.C. Trexler,
and J. Wisniewski. 1994. Carbon pools and flux of global
forest ecosystems. Science 263: 185-190.
dos Santos, J.M. 1987. Climate, natural vegetation, and soils in
Amazonia: An overview. Pages 25-36 in P.E. Dickson (ed.)
Geophysiology of Amazonia. John Wiley & Sons, New York, NY.
Dylis, N.V., and L.M. Nosova, 1977. Phytomass of forest bioceno-
sis. Nauka Press, Moscow.
Eden, M.J. 1990. Ecology and land management in Amazonia.
Belhaven Press, London, UK.
Eiten, G. 1972. The cerrado vegetation of Brazil. Botanical
Review 38: 201-341.
EIU (Economist Intelligence Unit). 1994. Brazil. EIU country
report 1st quarter 1994. The Economist Intelligence Unit,
London, UK.
191
-------
Evans, D.L., Z. Zhu, S. Eggen-Mclntosh, P.G. Mayoral, and J.L.
Ornelas de Anda. 1992. Mapping Mexico's forest lands with
advanced very high resolution radiometer. Research Note SO-
367. U.S. Department of Agriculture, Forest Service,
Southern Forest Experiment Station, New Orleans, LA.
Ewel, J., C. Berish, B. Brown, N. Price, and J. Raich. 1981.
Slash and burn impacts on a Costa Rican wet forest site.
Ecology 62: 816-829.
Eyre, F.H. (ed.). 1980. Forest cover types of the United States
and Canada. Society of American Foresters, Washington, DC.
Fan, S.M., S.C. Wofsy, P.S. Bakwin, and D.J. Jacob. 1990.
Atmosphere-biosphere exchange of C02 and 03 in the Central
Amazon Forest. Journal of Geophysical Research 95(D10):
16,851-16,864.
FAO (Food and Agriculture Organization of the United Nations).
1993. Forest resources assessment 1990, tropical countries.
FAO Forestry Paper 112. FAO, Rome, Italy.
Faure, H. 1990. Changes in the global continental reservoir of
carbon. Global Planet Change 21: 247-252.
Fearnside, P.M. 1989. Deforestation and agricultural development
in Brazilian Amazonia. Interciencia 14: 291-297.
Fearnside, P.M. 1992. Greenhouse gas emissions from deforestation
in the Brazilian Amazon. In W. Makundi, J. Sathaye, and 0.
Masera (eds.) Carbon emissions and sequestration in forests:
Case studies from seven developing countries, Volume 1:
Summary. LBL-32119. U.S. Environmental Protection Agency,
Climate Change Division, Washington, DC and the Lawrence
Berkeley Laboratory, Energy and Environment Division,
Berkeley, CA.
Fearnside, P.M. 1993. Deforestation in Brazilian Amazonia: The
effect of population and land tenure. Ambio 22(8): 537-545.
Fiala, K., L. Hernandez, E. Perez, and A. Cardenas. 1991. Biomass
partitioning in Paspalum notatum stands on slope relief in
the anthropic savanna of Cuba. Preslia 63: 269-289.
Flint, E.P,, and J.Fr Richards. 1994. Trends in carbon content of
vegetation in south and southeast Asia associated with
changes in land use. Pages 201-299 in V.H. Dale (ed.)
Effects of land use change on atmospheric COz concentra-
tions: Southeast Asia as a case study. Springer-Verlag, New
York, NY.
192
-------
Fonseca, G.A.B da. 1985. The vanishing Brazilian Atlantic forest.
Biological Conservation 34: 17-34.
Franklin, J.F., and M.A. Hemstrom. 1981. Aspects of succession in
the coniferous forests of the Pacific Northwest. Pages 212-
292 in D.C. West, H.H. Shugart, and D.B. Botkin (eds.)
Forest succession: Concepts and application. Springer-
Verlag, New York, NY.
Frost, P.G.H., and F. Robertson. 1985. The ecological effects of
fire in savannas. Pages 93-140 in B.H. Walker (ed.) Determi-
nants of tropical savannas. International Union of Biologi-
cal Sciences Monograph Series, No. 3. IRL Press, Oxford, UK.
Furyaev, V.V. 1989. Assessment and mapping of stands according to
their fire resistance. Lesnoe Khozyaistvo, 4: 47-48.
Gaston, G.G., and T.P. Kolchugina. 1994. Use of global vegetation
index images in a framework to estimate phytomass and net
primary productivity in terrestrial ecosystems of the former
Soviet Union. Presented at the 5th Global Warming Interna-
tional Conference and EXPO, 4-7 April 1993. San Francisco,
CA.
Gaston, G.G., T.P. Kolchugina, and T.S. Vinson. 1993. Potential
effect of no-till management on carbon in the agricultural
soils of the former Soviet Union. Agriculture Ecosystems and
Environment 45: 295-309.
Gonzalez Flores, P.C. 1992. El manejo del fuego en el sistema de
roza, tumba y quema en la selva baja caducifolia de Chamela,
Jalisco. Dissertation, Universidad Nacional Autonoma de
Mexico City, Mexico.
Goulding, M. 1980. The fishes and the forest: Explorations in
Amazonian natural history. University of California Press,
Berkeley, CA.
GPEP (Global Processes and Effects Program). 1993. Effects of C02
and climate change on forest trees. EPA/600/R-93/146. U.S.
Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR.
Graca, M.G., and A.N.. Ketoff. 1991. C02 savings in Brazil. Energy
Policy 19: 1003-1010.
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.
193
-------
Guevara, S., J. Meave, P.Moreno-Casasola, and J. Laborde. 1992.
Floristic composition and structure of vegetation under
isolated trees in neotropical pastures. Journal of Vegeta-
tion Science 3: 655-664.
Hansen, J., I. Fung, A. Lacis, D. Rind, S. Lebedeff, R. Ruedy, G.
Russell, and P. Stone. 1988. Global climate changes as
forecast by Goddard Institute for Space Studies three-
dimensional model. Journal of Geophysical Research 93:
9341-9364.
Harmon, M. 1993. Woody debris budgets for selected forest types
in the U.S. Pages 151-178 in D.P. Turner, J.J. Lee, G.J.
Koerper, and J.R. Barker (eds.) The forest sector carbon
budget of the United States: Carbon pools and flux under
alternative policy options. EPA/600/3-93/093. U.S. Environ-
mental Protection Agency, Environmental Research Laboratory,
Corvallis, OR.
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 States: 1989-2040. USDA Forest Service General
Technical Report RM-199. U.S. Department of Agriculture,
Forest Service, Fort Collins, CO.
Heath, L.S., and R.A. Birdsey. 1993. Carbon trends of productive
temperate forests on the coterminous United States. Water,
Air and Soil Pollution 70: 279-293.
Hecht, S.B. 1992. Chapter 2, Logics of livestock and deforesta-
tion: The case of Amazonia. Pages 7-26 in T.E. Downing, S.B
Hecht, H.A. Pearson, and C. Garcia-Downing (eds.) Develop-
ment and destruction: The conversion of tropical forest to
pasture in Latin America. Westview Press, Boulder, CO.
Hecht, S. 1993. The logic of livestock and deforestation in
Amazonia. Bioscience 43(10): 687-695.
Holmen, K. 1992. The global carbon cycle. Pages 239-262 in R.S.
Butcher, R.J. Charlson, G.H. Orians, and G.V. Wolfe (eds.)
Global biogeochemical cycles. Academic Press, London, UK.
Houghton, J.T., B.A. Callender, and S.K. Varney (eds.). 1992.
Climate change 1992: The supplementary report to the IPCC
scientific assessment. Cambridge University Press, New York,
NY.
194
-------
Houghton, J.T., G.J. Jenkins, and J.J. Ephraums (eds.). 1990.
Climate change: The IPCC assessment. Cambridge University
Press, New York, NY.
Houghton, R.A. , and D.L. Skole. 1990. Carbon. Pages 393-408 in
B.L. Turner, W.C. Clark, R.W. Kates, J.F. Richards, J.T.
Mathews, and W.B. Meyer (eds.) The Earth as transformed by
human action. Cambridge University Press, New York, NY.
Houghton, R.A., D.S. Lefkowitz, and D.L. Skole. 1991a. Changes in
the landscape of Latin America between 1850 and 1985: I.
Progressive loss of forests. Forest Ecology and Management
38: 143-172.
Houghton, R.A., D.L. Skole, and D.S. Lefkowitz. 1991b. Changes in
the landscape of Latin America between 1850 and 1985: II.
Net release of C02 to the atmosphere. Forest Ecology and
Management 38: 173-199.
Houghton, R.A., J.D. Unruh, and P.A. Lefebvre. 1993. Current land
cover in the tropics and its potential for sequestering
carbon. Global Biogeochemical Cycles 7: 3 05-320.
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: A net release of C02 to the atmosphere.
Ecological Monographs 53: 235-262.
Houghton, R.A., R.D. Boone, J.R. Fruci, J.E. Hobbie, J.M.
Melillo, C.A. Palm, B.J. Peterson, G.R. Shaver, G.M.
Woodwell, B. Moore, D.L. Skole, and N. Myers. 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.
IBGE/IBDF. 1993. Mapa de vegetacao do Brasil. Map edition 1993,
Scale 1:5,000,000. Ministerio da Agricultura. Brasilia,
Brazil.
INPE. 1992. Deforestation in Brazilian Amazonia. Instituto
Nacional de Pesquisas Espaciais, Sao Jose dos Campos,
Brazil.
Isachenko, A.G. (ed.). 1988. The Landscape Map of the U.S.S.R.
Scale 1:4,000,000. Institute of Geography, Leningrad Stte
University, GUGK of the U.S.S.R., Moscow.
Isayev, A.S., G.N. Korovin, A.I. Utkin, A.A. Pryazhnikov, and
D.G. Zamolotchikov. 1993. Deposition in assessment of stores
and annual sequestration of carbon in phytomass of forest
ecosystems of Russia. Lesovedenie 5: 3-10.
195
-------
Ita-Martinez, C., J. Sarukhan, and J.M. Maass. (Unpublished
manuscript). Land use patterns on a tropical deciduous
forest ecosystem on the Pacific coast of Jalisco, Mexico.
Universidad Nacional Autonoma de Mexico.
Izrael, Y.A., and F.Ya. Rovinskyi (eds.). 1990. Overview of state
of natural environment in the U.S.S.R. U.S.S.R. State
Committee on Hydrometeorology, Hydrometeoizdat Press,
Moscow.
Johnson, D.W. 1992. Effects of forest management on.soil carbon
storage. Water, Air and Soil Pollution 64: 83-120.
Jordan, C.F., and 6. Escalante. 1980. Root productivity in an
Amazonian rain forest. Ecology 61(1): 14-18.
Karaban, R.T., A.O. Kokorin, I.M. Nazarov, and A.Z. Shvidenko.
1993. Sequestration of C02 in Russian forests. Metereologia
I Gidrologia 1: 5-14.
Kauffman, J.B., D.L. Cummings, R.F. Hughes, and E.A. Castro.
1994a. Biomass and nutrient losses associated with anthropo-
genic fires. International Meeting of the Society for
Conservation Biology and the Association for Tropical
Biology, 7-11 June. Universidad de Guadalajara, Guadalajara,
Jalisco, Mexico
Kauffman, J.B., D.L. Cummings, and D.E. Ward. 1994b. Relationship
of fire, biomass, and nutrient dynamics along a vegetation
gradient in the Brazilian cerrado. Journal of Ecology.
Forthcoming.
Kauffman, J.B., R.L. Sanford, D.L. Cummings, I.H. Salcedo, and
E.V.S.B. Sampaio. 1993. Biomass and ^nutrient dynamics
associated with slash fires in neotropical dry forests.
Ecology 74: 140-151.
Kauppi, P.E., K.K.K. Mielikainen, and K. Kuusela. 1992. Biomass
and carbon budget of European forests, 1971 to 1990. Science
256: 70-74.
Kazantseva, T.I. 1980. Productivity and dynamics of above-ground
phytomass of deserts. Problems of Desert Utilization 2: 76-
83.
Kazimirov, N.I., and R.M. Morosova. 1973. Biological turnover of
substances in spruce forests of Karelia. Nauka Press,
Leningrad.
Keating, M. 1993. The earth summit's agenda for change. Centre
for Our Common Future, Geneva, Switzerland.
196
-------
Kempe, S. 1988. Freshwater carbon and the weathering cycle. Pages
197-223 in A. Lerman and M. Meybeck (eds.) Physical and
chemical weathering in geochemical cycles. Kluwer Academic
Publishers, The Netherlands.
Kern, J. 1994. Spatial patterns of soil organic carbon in the
contiguous United States. Soil Science Society of America
Journal 58: 439-455.
Kern, J.S., and M.G. Johnson. 1991. The impact of conservation
tillage use on soil and atmospheric carbon in the contiguous
United States. EPA/600/3-91/056. U.S. Environmental Protec-
tion Agency, Environmental Research Laboratory, Corvallis,
OR.
King, G.A. 1993a. Conceptual approaches for incorporating
climatic change into the development of forest management
options for sequestering carbon. Climate Research 3: 61-78.
King, G.A. 1993b. Net primary productivity of non-forested lands.
Pages 179-194 in D.P. Turner, J.J. Lee, G.J. Koerper, and
J.R. Barker (eds.) The forest sector carbon budget of the
United States: Carbon pools and flux under alternative
policy options. EPA/600/3-93/093. U.S. Environmental
Protection Agency, Environmental Research Laboratory,
Corvallis, OR.
King, G.A., and R.P. Neilson. 1992. The transient response of
vegetation to climate change: A potential source of C02 to
the atmosphere. Water, Air and Soil Pollution 64: 365-384.
Kobak, K.I. 1988. Biotical components of carbon cycle.
Hydrometeoizdat, Leningrad.
Koch, P. 1989. Estimates by species group and region in the USA
of below-ground root weight as a percentage of ovendry
complete tree weight and carbon content of tree portions.
Consulting Report to the USDA Forest Service.
Kolchugina T.P., and T.S. Vinson. 1991. Framework to quantify the
natural terrestrial carbon cycle of the former Soviet Union.
Pages 257-273 in T. Vinson and T. Kolchugina (eds.) Proceed-
ings of the workshop on carbon cycling in boreal forests and
subarctic ecosystems, 9-14 September. Corvallis, OR.
Kolchugina, T.P., and T.S. Vinson. 1993a. Equilibrium analysis of
carbon pools and fluxes of forest biomes in the former
Soviet Union. Canadian Journal of Forest Research 23: 81-88.
197
-------
Kolchugina, T.P., and T.S. Vinson. 1993b. Carbon balance of the
continuous permafrost zone of Russia. Climate Research 3:
13-21
Kolchugina, T.P., and T.S. Vinson. 1993c. Climate warming and the
carbon cycle in the permafrost zone of the former Soviet
Union. Permafrost and Periglacial Processes 4: 149-163.
Kolchugina, T.P., and T.S. Vinson. 1993d. Carbon sources and
sinks in the forest biomes of the former Soviet Union.
Global Biogeochemical Cycles 7: 291-304.
Kolchugina, T.P., and T.S. Vinson. 1993e. Comparative analysis of
carbon budget components for forest biomes in the former
Soviet Union. Water, Air and Soil Pollution 70: 207.
Kolchugina, T.P., and T.S. Vinson. 1994a. Production of green-
house gases in the former Soviet Union. World Resource
Review 6, N3: 291-303.
Kolchugina, T.P., and T.S. Vinson. 1994b. Mitigation options in
the agricultural sector of the former Soviet Union. Pre-
sented at IPCC WG II Subgroup D mitigation options in
agriculture, May 1994. Colorado State University, Fort
Collins, CO.
Kolchugina, T.P., and T.S. Vinson. 1995. Role of Russian forests
in the global carbon balances. Ambio. Forthcoming.
Kolchugina, T.P., T.S. Vinson, G.G. Gaston, V.A. Rozhkov, and
A.Z. Shvidenko. 1995. Carbon pools, fluxes, and sequestra-
tion potential in soils of the former Soviet Union. Advances
in Soil Science. Forthcoming.
Korovin, G.N. 1993. The structure of the Russian forest wildfire.
Presented at the international conference Fire in Ecosystems
of Northern Eurasia, 29 June-2 July 1993. Krasnoyarsk,
Russia.
Kozlovski, V.B., and V.M. Pavlov. 1967. Growth rates of the main
forest forming species in the U.S.S.R. Moscow, Lesnaya
Promyshlennost.
Krankina, O.N., and R.K. Dixon. 1993. Forest management options
to conserve and sequester terrestrial carbon in the Russian
Federation. World Resource Review No. 6: 88-101.
Kurbatskyi, N.P., and A.G. Tsykalov. 1988. Forest fires in
central Evenkiya. Lesnoe Khozyaistvo 6: 38-40.
198
-------
Kurz, W.A., and M.J. Apps. 1993. Contribution of northern forests
to the global C cycle: Canada as a case study. Water, Air &
Soil Pollution 70: 163-176.
Kurz, W.A. , M.J. Apps, T.M. Webb, and P.J. McNamee. 1992. The
carbon budget of the Canadian forest sector: Phase 1.
Information Report NOR-X-326, Forestry Canada, Northwest
Region, Northern Forestry Centre.
Lefebvre, P.A., and T.A. Stone. 1994. Monitoring selective
logging in eastern Brazilian Amazonia using multi-temporal
Landsat Thematic Mapper imagery. Pages 288-291 in: Proceed-
ings, International Society for Photogrammetry and Remote
Sensing, Vol. 30 Part 71. 26-30 September, Rio de Janeiro.
Instituto Nacional de Pesquisas Espaciais, San Jose dos
Campos, Brazil.
Lenihan, J.M. 1992. Canadian vegetation response to climate and
projected climatic change. Ph.D. dissertation. Oregon State
University, Corvallis, OR.
Levine, J.S. (ed.). 1991. Global biomass burning: Atmospheric,
climatic, and biospheric implications. The MIT Press,
Cambridge, MA.
Loveland, T.R., J.W. Merchant, D.O. Ohlen, and J.F. Brown. 1991.
Development of a land-cover characteristics database for the
conterminous U.S. Photogrammetric Engineering and Remote
Sensing 57: 1453-1463.
Lugo, A.E., and S. Brown. 1992. Tropical forests as sinks of
atmospheric carbon. Forest Ecology and Management 54:
239-255.
Lugo, A.E., M.J. Sanchez, and S. Brown. 1986. Land use and
organic carbon content of some subtropical soils. Plant and
Soil 96: 185-196.
MacCleery, D.W. 1992. American forests: A history of resiliency
and recovery. USDA Forest Service General Technical Report
FS-540. U.S. Department of Agriculture, Forest Service.
MacLean, C.D., J.L. Ohmann, and P.M. Bassett. 1991. Preliminary
timber resource statistics for southwest Washington.
Resource Bulletin PNW-RB-177. U.S. Department of Agricul-
ture, Forest Service, Pacific Northwest Research Station,
Portland, OR.
Makarov, A.A., and I. Bashmakov. 1990. The Soviet Union: Carbon
emission control strategies. Pages 35-53 in William U.
Chandler (ed.) Case studies in international cooperation.
199
-------
World Wildlife Fund and Conservation Foundation, Washington
D.C.
Makundi, W., J. Sathaye, and A. Ketoff. 1991. COPATH: A spread-
sheet model for estimating carbon flows associated with
tropical forest use. Report No. LBL-35255. Lawrence Berkeley
Laboratory. Berkeley, CA.
Makundi, W., J. Sathaye, and 0. Masera. 1992. Carbon emission and
sequestration in forests: Case studies from seven developing
countries, Volume 1: Summary. Report No. LBL-32119. Lawrence
Berkeley Laboratory, University of California, Berkeley, CA.
Manabe, S., R.J. Stouffer, M.J. Spelman, and K. Bryan. 1991.
Transient responses of a coupled ocean-atmosphere model to
gradual changes of atmospheric co2. Part I: Annual mean
response. Journal of Climate 4: 785-818.
Mann, L.K. 1986. Changes in soil carbon storage after cultiva-
tion. Soil Science 142, N5:279-288.
Marland, 6. 1988. The prospect for solving the C02 problem
through reforestation. Report DOE/NBB-0082. U.S. Department
of Energy, Office of Energy Research, Washington, DC.
Martinez-Yrizar, A., J. Sarukhan, A. Perez-Jimenez, E. Rincon,
J.M. Maass, A. Solis-Magallanes, and L. Cervantes. 1992.
Above-ground phytomass of a tropical deciduous forest on the
coast of Jalisco, Mexico. Journal of Tropical Ecology 8: 87-
96.
Masera, 0., M.R. Bellon, and G. Segura. 1995. Forest management
option for sequestering carbon in Mexico. Biomass and
Bioenergy. Forthcoming.
Masera, 0., M. de Jesus 0rd6nez, and R. Dirzo. 1992. Carbon
emissions from deforestation in Mexico: Current situation
and long-term scenarios. Pages 1-49 in W. Makundi and J.
Sathaye (eds.) Carbon emission and sequestration in forests:
Case studies from seven developing countries, Summary.
Report No. LBL-32665. Lawrence Berkeley Laboratory, Univer-
sity of California, Berkeley, CA.
Mather, A.S. 1990. Global forest resources. Timber Press,
Portland, OR.
Matthews, E. 1983. Global vegetation and land use: New high-
resolution data bases for climate studies. Journal of
Climate and Applied Meteorology 22: 474-487.
200
-------
Mattos, M.M., and C. Uhl. 1994. Economic and ecological perspec-
tives on ranching in the eastern Amazon. World Development.
22(2): 145-158.
McNabb, K., J. Borges, and J. Welker. 1994. Jari at 25: An
investment in the Amazon. Journal of Forestry 92: 21-26.
Melillo, J.M., J.R. Furry, R.A. Houghton, B. Moore, III, and D.L.
Scole. 1988. Land-use change in the Soviet Union between
1850-1980: Causes of a net release of C02 to the atmosphere.
Tellus 40B: 116-128.
Mendoza, Y., 0. Masera, and P. Macias. 1991. Long-term energy
scenarios for Mexico: Policy options for carbon savings and
main barriers. Energy Policy 10: 962-969.
Meybeck, M. 1993. Riverine transport of atmospheric carbon:
Sources, global typology, and budget. Water, Air and Soil
Pollution. 70: 443-463.
Mills, J.R., 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. U.S. Department of Agriculture, Forest Service,
Portland, OR.
Mitchell, J.F.B. 1989. The "greenhouse effect" and climate
change. Reviews of Geophysics 27: 115-139.
Molion, L.C.B. Submitted. Global climate impacts of Amazonia
deforestation. Journal of Climatology.
Moran, E.F., E. Brondizio, P. Mausel, and Y. Wu. 1994. Integrat-
ing Amazonian vegetation, land-use, and satellite data.
Bioscience 44(5): 329-338.
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 General Technical Report
WO-58. U.S. Department of Agriculture, Forest Service.
Nadelhoffer, K.J., and J.W. Raich. 1992. Fine root production
estimates and belowground carbon allocation in forest
system. Ecology 73: 1139-1147.
Neilson, R.P. 1990. Biosphere feedback during climate change.
Pages 57-76 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. EPA/600/3-90/080. U.S. Environmental
Protection Agency, Environmental Research Laboratory,
Corvallis, OR.
201
-------
Neilson, R.P., G.A. King, and J. Lenihan. 1992. Toward a rule-
based biome model. Landscape Ecology' 7: 21-A3.
Nelson, B.W., V. Kapos, J.B. Adams, W.J. Oliveira, O.P.G. Braun,
and and I.L. do Amaral. 1994. Forest disturbance by large
blowdowns in the Brazilian Amazon. Ecology 75: 853-858.
Nikodimov, I.D. 1988. Fire hazardous prophylactic - The basis for
successful forest protection. Lesnoe Khozyaistvo 4: 50-53.
NYT (New York Times). 1988. Atlas of the world. Times Books, New
York, NY.
Odum, E.P. 1953. Ecology. W.B. Saunders Co., Philadelphia and
London.
Olson, J.S., J.A. Watts, and L.J. Allison. 1983. Carbon in live
vegetation of major world ecosystems. Report ORNL-5862. Oak
Ridge National Laboratory, Oak Ridge, TN.
Olson, J.S., J.A. Watts, and L.J. Allison. 1985. Major world
ecosystem complexes ranked by carbon in live vegetation: A
database. Report DOE.NDP-017. Oak Ridge National Laboratory,
Environmental Sciences Division, Oak Ridge, TN.
Parks, P.J., S.R. Brame, and J.E. Mitchell. 1992. Opportunities
to increase forest area and timber growth on marginal crop
and pasture land. Pages 97-121 in R.N. Sampson and D. Hair
(eds.) Forests and global change. American Forestry Associa-
tion, Washington, DC.
Parson, E.A., P.M. Haas, and M.A. Levy. 1992. A summary of the
major documents signed at the Earth Summit and the Global
Forum. Environment 34: 12-15, 34-36.
Parton, W.J., J.W. Stewart, and C.V. Cole. 1988. Dynamics of C,
N, P, and S in grassland soils: A model. Biogeochemistry 5:
109-131.
Pastor, J., and W.M. Post. 1985. Development of linked forest
productivity-soil process model. ORNL-5862. Oak Ridge
National Laboratory. Oak Ridge, TN.
Pastor, J., and W.M. Post. 1986. Influence of climate, soil
moisture, and succession of forest carbon and nitrogen
cycles. Biogeochemistry 2:3-27.
Peterson, C.E. 1993. Biomass and carbon for woodlands in the
Western United States. Pages 143-149 in D.P. Turner, J.J.
Lee, G.J. Koerper, arid J.R. Barker (eds.) The forest sector
carbon budget of the United States: Carbon pools and flux
under alternative policy options. EPA/600/3-93/093. U.S.
202
-------
Environmental Protection Agency, Environmental Research
Laboratory, Corvallis, OR.
Peterson, C.E., and D.P. Turner. 1993. Uncertainty in country-
wide forest biomass estimates. In: The carbon balance of the
world's forest ecosystems: Towards a global assessment, 11-
15 May 1992. Jensuu, Finland.
Phalaleyev, Y.N. 1985. Siberian forests. Krasnoyarsk University
Press, Krasnoyarsk, Russia.
Plantinga, A.J., and R.A. Birdsey. 1993. Carbon fluxes resulting
from U.S. private timberland management. Climatic Change 23:
37-53.
Post, W.M., T-H Peng, W.R. Emanuel, A.W. King, V.H. Dale, and
D.L. DeAngelis. 1990. The global carbon cycle. American
Scientist 78: 310-326.
Postel, S., and L. Heise. 1988. Reforesting the Earth. WorldWatch
Paper 83. WorldWatch Institute, Washington, DC.
Powell, D.S., J.L. Faulkner, D.R. Darr, Z. Zhu, and D.W.
MacCleery. 1993. Forest resources of the United States,
1992. USDA Forest Service General Technical Report, RM-234.
U.S. Department of Agriculture, Forest Service, Fort
Collins, CO.
Prentice, K.C. 1990. Bioclimatic distribution of vegetation for
GCM studies. Journal of Geophysical Research 95: 811-830.
Prentice, I.e., W. Crammer, S.P. Harrison, R.A. Monserud, and
A.M. Solomon. 1992. A global biome model based on plant
physiology and dominance, soil properties and climate.
Journal of Biogeography 19: 117-134.
Raich, J.W., and W.H. Schlesinger. 1992. The global carbon
dioxide flux in soil respiration and its relationship to
vegetation and climate. Tellus 44B: 81-99.
Raison, R.J., P.K. Khanna, and P.V. Woods. 1985. Mechanisms of
element transfer to the atmosphere during vegetation fires.
Canadian Journal of Forest Research 15: 132-140.
Ramos, J.M., and S. del Amo. 1992. Enrichment planting in a
tropical secondary forest in Veracruz, Mexico. Forest
Ecology and Management 54: 289-304.
Randall, A. 1991. The value of biodiversity. Ambio 30: 64-68.
Rastetter. E.B., M.G. Ryan, G.R. Shaver, J.M. Melillo, K.J.
Nadelhoffer, J.E. Hobbie, and J.D. Abner. 1991. A general
203
-------
model describing the responses of the C and N cycles in
terrestrial ecosystems to changes in C02, climate, and N
deposition. Tree Physiology 9: 101-126.
Rastetter, E.B. R.B. McKane, and G.R. Shaver, J.M. Melillo. 1992.
Changes in C storage by terrestrial ecosystems: how ON
interactions restrict responses to C02 and temperature.
Water, Air and Soil Pollution 64: 327-344.
Robinson, J.M. 1989. On uncertainty in the computation of global
emissions from biomass burning. Climate Change 14: 243-262.
Row, C., and R.B. Phelps. 1990. Tracing the flow of carbon
through U.S. forest products sector. Proceedings of the 19th
World Congress of the International Union of Forest Research
Organizations. 5-11 August 1990. Montreal, Canada.
Rowntree, R.A., and D.J. Nowak. 1991. Quantifying the role of
urban forests in removing atmospheric carbon dioxide.
Arboriculture 17: 269-275.
Runkle, J.R. 1985. Disturbance regimes in temperate forests.
Pages 17-33 in S.T.A. Pickett and P.S. White (eds.) The
ecology of natural disturbance and patch dynamics. Academic
Press, Inc., New York, NY.
Running, S.W., and J.C. Coughlan. 1988. A general model of forest
ecosystem processes for regional applications: I. Hydrologic
balance, canopy gas exchange and primary production pro-
cesses. Ecological Modeling 42: 125-154.
Ryabchikov, A.M. (ed.). 1988. Map: Geographical Belts and Zonal
Types of Landscapes of the World. Scale 1:15,000,000. School
of Geography, Moscow State University, Moscow, Russia.
Ryan, M.G., and R.H. Waring. 1992. Maintenance respiration and
stand development in a subalpine lodgepole pine forest.
Ecology 73: 2100-2108.
Rzedowski, J. 1978. Vegetacion de Mexico. Editorial Limusa,
Mexico.
Salati, E., R.L. Victoria, L.A. Martinelli, and J.E. Richey.
1991. Forests: Their role in global change, with special
reference to the Brazilian Amazon. Pages 391-395 in J. Jager
and H.L. Ferguson (eds.) Climate change: Science, impacts
and policy. Proceedings of the second world climate confer-
ence. Cambridge University Press, New York, NY.
Saldarriaga, J.G. 1985. Forest succession in the upper Rio Negro
of Colombia and Venezuela. Ph.D. dissertation. University of
Tennessee, Knoxville, TN.
204
-------
Sampson, R.N. 1992. Forestry opportunities in the United States
to mitigate the effects of global warming. Water, Air and
Soil Pollution 64: 83-120.
Sampson, R.N., and D. Hair (eds.). 1992. Forests and global
change. Volume One: Opportunities for increasing forest
cover. American Forestry Association, Washington, DC.
Sanchez, P.A., C.A. Palm, L.T. Szott, E. Cuevas, and R. Lai.
1989. Organic input management in tropical agroecosystems.
Pages 125-152 in D.C. Coleman, J.M. Oades, and G. Uehara
(eds.) Dynamics of soil organic matter in tropical ecosys-
tems. NifTAL Project. University of Hawaii, College of
Tropical Agriculture and Human Resources, Honolulu, HI.
Sanchez, P.A., J.H. Villachica, and D.E. Bandy. 1983. Soil
fertility dynamics after clearing a tropical rainforest in
Peru. Soil Science Society of America Journal 47: 1171-1178.
SARH (Secretaria de Agricultura y Recursos Hidraulicos). 1986.
Informacion forestal de la Republica Mexicana. SARH, Mexico
D.F., Mexico.
SARH (Secretaria de Agricultura y Recursos Hidraulicos). 1992.
Inventario nacional forestal de gran vision: Reporte
principal. SARH, Mexico D.F., Mexico.
Sarmiento, J.L., and E.T. Sundquist. 1992. Revised budget for the
oceanic uptake of anthropogenic carbon dioxide. Nature 356:
589- 593.
Schacht, W.H., J.N. Long, and J.C. Malechek. 1988. Aboveground
production in cleared and thinned stands of semiarid
tropical woodland, Brazil. Forest Ecology and Management 23:
201-214.
Schindler, D.W., and S.E. Bayley. 1993. The biosphere as an
increasing sink for atmospheric carbon: Estimates from
increased nitrogen deposition. Global Biogeochemical Cycles
7: 717-733.
Schlesinger, W.H. 1984. Soil organic matter: A source of atmo-
spheric C02. Pages 11-127 in G.M. Woodwell (ed.) Scope
23—The role of terrestrial vegetation in the global carbon
cycle: Measurement by remote sensing. Cambridge University
Press, Cambridge, UK.
Schlessinger, M.E., and J.F.B. Mitchell. 1985. Model projections
of the equilibrium climatic response to increased carbon
dioxide. Pages 81-147 in M.D. MacCracken and F.M. Luther
(eds.) Projecting the climatic effects of increasing carbon
205
-------
dioxide. DOE/ER-0237. U.S. Department of Energy, Washington,
DC.
Schmink, M., and C.H. Wood. 1992. Contested Frontiers in
Amazonia. Columbia University Press, New York, NY.
Schneider, S.H. 1989. Global warming: Are we entering the
greenhouse century? Sierra Club, San Francisco, CA.
Schroeder, P.E., and J.K. Winjum. 1995a. Assessing Brazil's
carbon budget: I. Biotic carbon pools. Forest Ecology and
Management. Forthcoming.
Schroeder, P.E., and J.K. Winjum. 1995b. Assessing Brazil's
carbon budget: II. Net carbon balance. Forest Ecology and
Management. Forthcoming.
SCS (Soil Conservation Service, U.S. Department of Agriculture).
1983. National Soils and Agricultural Handbook Number 430.
U.S. Government Printing Office, Washington, DC.
SCS (Soil Conservation Service, U.S. Department of Agriculture).
1987. Basic statistics, 1982 National Resources Inventory.
Statistical Bulletin Number 756. Iowa State University,
Statistical Laboratory, Ames, IA.
Sedjo, R. 1989. Forests to offset the greenhouse effect. Journal
of Forestry 87: 12-15.
Sedjo, R.A. 1992. Temperate forest ecosystems in the global
carbon cycle. Ambio 21: 274-277.
Serrao, E.A.S., and A.K.O. Homma. 1993. Brazil. Pages 265-351 in
Sustainable agriculture and the environment in the humid
tropics. National Research Council, National Academy Press,
Washington, DC.
Sessions, J. (coordinator). 1991. Timber for Oregon's tomorrow:
The 1989 update. Oregon State University, College of
Forestry, Forest Research Laboratory, Corvallis, OR.
Showers, W.J., and D.G. Angle. 1986. Stable isotopic character-
ization of organic carbon accumulation on the Amazon
continental shelf. Continental Shelf Research 6: 227-244.
Shugart, H.H., T.M. Smith, and W.M. Post. 1992. The potential for
application of individual-based simulation models for
assessing the effects of global change. Annual Review of
Ecology and Systematlcs 23: 15-38.
Siegenthaler, U., and J.L. Sarmiento. 1993. Atmospheric carbon
dioxide and the ocean. Nature 365: 119-125.
206
-------
Simpson, L.G., and D.B. Botkin. 1992. Vegetation, the global
carbon cycle, and global measures. Pages 413-425 in D.A.
Dunnette and R.J. O'Brien (eds.) The science of global
change, the impact of human activities on the environment.
ACS Symposium Series 483. American Chemical Society.
Washington, DC.
Singh, J.S., and M.C. Joshi. 1979. Primary production. Pages
197-218 in R.T. Coupland (ed.) Grassland ecosystems of the
world: Analysis of grasslands and their uses. Cambridge
University Press, New York, NY.
Sioli, H. 1975. Amazon tributaries and drainage basins. Pages
199-213 in A.D. Hasler (ed.) Coupling of land and water
systems. Springer-Verlag, Berlin.
Sitarz, D. (ed.). 1993. Agenda 21: The earth summit strategy to
save our planet. EarthPress, Boulder, CO.
Skole, D., and C. Tucker. 1993. Tropical deforestation and
habitat fragmentation in the Amazon: Satellite data from
1978 to 1988. science 260: 1905-1910.
Skole, D.L., W.H. Chomentoski, W.A. Sala, and A.D. Nobre. 1994.
Physical and human dimensions of deforestation in Amazonia.
Bioscience 44: 314-338.
Smith, T.M., and H.H. Shugart. 1993. The transient response of
terrestrial C storage to a perturbed climate. Nature 361:
523-526.
Smith, T.M., H.H. Shugart, G.B. Bonan, and J.B. Smith. 1992.
Modeling the potential response vegetation to global climate
change. Advances in Ecological Research 22: 93-116.
Smith, T.M., W.P. Cramer, R.K. Dixon, R. Leemans, R.P. Neilson,
and A.M. Solomon. 1993. The global terrestrial carbon cycle.
Water, Air and Soil Pollution 70: 19-37.
Solomon, A.M. 1986. Transient response of forests to C02-induced
climate change: Simulation modeling experiments in eastern
North America. Oecologia 68: 567-579.
Solomon, A.M., and H.H. Shugart. 1984. Integrating forest stand
simulations with paleoecological records to examine long-
term forest dynamics. Pages 333-356 in G.I. Agren (ed.)
State and change of forest ecosystems. Swedish University of
Agricultural Science. Uppsala, Sweden.
Solomon, A.M., I.e. Prentice, R. Leemans, and W.P. Cramer. 1993.
The interaction of climate and land use in future terrestr-
207
-------
ial carbon storage and release. Water, Air and Soil Pollu-
tion 70: 595-614.
Sombroek, W.G., F.O. Nachtergaele, and A. Hebel. 1993. Amounts,
dynamics and sequestering of carbon in tropical and subtrop-
ical soils. Ambio 22: 417-426.
Sorani, V, R. Alvarez, J.C. Baca, and S. Varela. 1993. The forest
inventory of Mexico, Phase I: Classification with TM
imagery. 25th International Symposium, Remote Sensing and
Global Environmental Change, 4-10 April 1993. Graz, Austria.
i
Sprugel, D.G. 1985. Natural disturbances and ecosystem energet-
ics. Pages 335-352 in S.T.A. Pickett and P.S. White (eds.)
The ecology of natural disturbances and patch dynamics.
Academic Press, New York, NY.
Stocks, B.J. 1991. The extent and impact of forest fires in
northern circumpolar countries. Pages 197-202 in J.S. Levine
(ed.) Global biomass burning: Atmospheric, climatic and
biogenic implications. MIT Press, Cambridge, MA.
Stone, T.A., P. Schlesinger, R.A. Houghton, G.M. Woodwell, and A.
Merrill. 1994. A map of the vegetation of South America
based on satellite imagery. Photogrammetric Engineering and
Remote Sensing 60: 541-551.
Subak, S., P. Raskin, and D. von Hippel. 1993. National green-
house gas accounts: Current anthropogenic sources and sinks.
Climatic Change 25: 15-58.
Sundquist, E.T. 1993. The global carbon dioxide budget. Science
259: 934-941.
Swanson, F.J., and J.F. Franklin. 1992. New forestry principles
from ecosystem analysis of Pacific Northwest forests.
Ecological Applications 2: 262-274.
Tans, P.P., I.Y. Fung, and T. Takahasi. 1990. Observational
constraints on the global atmospheric C02 budget. Science
247: 1431-1438.
Taylor, J.A., and J. Lloyd. 1992. Sources and sinks of atmo-
spheric C02. Australian Journal of Botany 40: 407-418.
Townshend, J., C. Justice, W. Li, C. Gurney, and J. McManus.
1991. Global land cover classification by remote sensing:
Present capabilities and future possibilities. Remote
Sensing and Environment 35: 243-255.
Turner, D.P., J.J. Lee, G. Koerper, and J.R. Barker. 1993a. The
forest sector carbon budget of the United States: Carbon
208
-------
pools and flux under alternative policy options. EPA/600/3-
93/093. U.S. Environmental Protection Agency, Environmental
Research Laboratory, Corvallis, OR.
Turner, D.P., G.J. Koerper, H. Gucinski, C.E. Peterson, and R.K.
Dixon. 1993b. Monitoring global change: Comparison of forest
cover estimates using remote sensing and inventory
approaches. Environmental Monitoring and Assessment 26: 295-
305.
Turner, D.P., G.J. Koerper, M. Harmon, and J.J. Lee. 1994a. A
carbon budget for forests of the conterminous United States.
Ecological Applications. Forthcoming.
Turner, D.P., G.J. Koerper, M. Harmon, and J.J. Lee. 1994b Carbon
sequestration by forests of the United States: Current
status and projections to the Year 2040. Tellus. Forthcom-
ing.
Tytlyanova A.A. 1977. Biological cycle of carbon in herbaceous
formations. Nauka Press, Novosibirsk, U.S.S.R.
Tyuremnov, S.N. 1976. Peat fields, 3rd ed. Nedra Press, Moscow,
U.S.S.R.
Uhl, C., and J.B. Kauffman. 1990. Deforestation, fire suscepti-
bility, and potential responses to fire in the eastern
Amazon. Ecology 71: 437-449.
Uhl, C., R. Buschbacher, and E.A.S. Serrao. 1988. Abandoned
pastures in eastern Amazonia, II. Patterns of plant succes-
sion. Journal of Ecology. 76: 663-681.
UNAM (Universidad Nacional Autonoma de Mexico). 1991. Recurso y
produccion forestal (VI.4.1). UNAM, Instituto de Geografia,
Mexico D.F., Mexico.
UNCED (U.N. Conference on Environment and Development). 1992.
Framework convention on climate change, Rio de Janeiro,
Brazil.
UNESCO (United Nations Educational, Scientific, and Cultural
Organization). 1980. Vegetation map of South America. Map
1980 edition, Scale 1:5,000,000. UNESCO, Paris, France.
Urban, D.L., G.B. Bonan, T.M. Smith, and H.H. Shugart. 1991.
Spatial applications of gap models. Forest Ecology and
Management 42: 95-110.
USDA (U.S. Department of Agriculture). 1989. An Analysis of the
land base situation in the United States: 1989-2040. Forest
209
-------
Service General Technical Report RM-181. U.S. Department of
Agriculture, Forest Service, Fort Collins, CO.
USDA (U.S. Department of Agriculture). 1992a. American forests: A
history of resiliency and recovery. Forest Service General
Technical Report FS-540. U.S. Department of Agriculture,
Forest Service, Durham, NC.
USDA (U.S. Department of Agriculture). 1992b. Forest Service
resource inventories: An overview. U.S. Department of
Agriculturef Forest Service, Forest Inventory, Economics,
and Recreation Research, Washington, DC.
U.S. EPA (Environmental Protection Agency) 1993. Simulating
climate-biosphere interactions under changing climate
conditions: A research plan. EPA/600/R-93/246. U.S. Environ-
mental Protection Agency, Environmental Research Laboratory,
Corvallis, OR.
U.S.S.R. Committee for the Protection of Nature. 1989. Report on
the state of the environment in the U.S.S.R. U.S.S.R.
Committee for the Protection of Nature, Moscow.
U.S.S.R. State Forestry Committee. 1990. Forest Fund of the
U.S.S.R., Moscow: 1. U.S.S.R. State Forestry Committee,
Moscow.
Vatkovskyi, O.S. 1976. The analysis of primary production of
forest formations. Nauka Press, Moscow.
Vinson, T.S., and T.P. Kolchugina. 1993. Pools and fluxes of
biogenic carbon in the former Soviet Union. Water, Air and
Soil Pollution 70: 223-237.
Vloedbeld, M., and R. Leemans. 1993. Quantifying feedback
processes in the response of the terrestrial carbon cycle to
global change: The modeling approach of IMAGE-2. Water, Air
and Soil Pollution 70: 615-628.
Vogt, K.A., C.C. Grier, and D.J. Vogt. 1986. Production, turn-
over, and nutrient dynamics of above- and below-ground
detritus of world forests. Advances in Ecological Research
15: 303-377.
Vorobyov, G.I. (ed.).- 1985. Forest encyclopedia, Vol. 1.
Sovetskaya Encyclopedia Press, Moscow, U.S.S.R.
Vorobyov, G.I. (ed.). 1986. Forest encyclopedia, Vol. 2.
Sovetskaya Encyclopedia Press, Moscow, U.S.S.R.
Waddell, K.L., D.D. Oswald, and D.S. Powell. 1989. Forest
statistics of the United States. USDA Resource Bulletin PNW-
210
-------
RB-168. U.S. Department of Agriculture, Forest Service,
Pacific Northwest Research Station, Portland, OR.
Wadsworth, G., R.J. Southard, and M.J. Singer. 1988. Effects of
fallow length on organic carbon and soil fabric of some
tropical Udults. Soil Scientific Society of America Journal
52: 1424-1430.
Walschburger, T., and P. von Hildebrand. 1991. The first 26 years
of forest regeneration in natural and man-made gaps in the
Colombian Amazon. Pages 257-263 in A. Gomez-Pompa, T.C.
Whitmore, and M. Hadley (eds.) Rain forest regeneration and
management. Man and the Biosphere Series, Volume 6. The
Parthenon Publishing Group, Pearl River, NY.
Waring, R.H., and W.H. Schlesinger. 1985. Forest ecosystems,
concepts, and management. Academic Press, Orlando, FL.
Weinstein, D.A., R.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
Physiology 9: 137-14 6.
Weinstein, D.A., H.H. Shugart, and D.C. West. 1982. The long-term
nutrient retention properties of forest ecosystems: A
simulation investigation. Report ORNL/TM-8472. Oak Ridge
National Laboratory. Oak Ridge, TN.
Westoby, J.C. 1989. Introduction to world forestry. B. Blackwell,
Ltd., Oxford, UK.
Williams, M. 1988. The death and rebirth of the American forest:
Cearing and reversion in the United States, 1900-1980. Pages
192-206 in J.F. Richards and R.P. Tucker (eds.) World
deforestation in the twentieth century. Duke University
Press, Durham, NC.
Williams-Linera, G. 1983. Biomass and nutrient content in two
successional stages of tropical wet forest in Uxpanapa,
Mexico. Biotropica 15: 275-284.
Winnett, S.M., R.W. Haynes, and W.G. Hohenstein. 1993. Economic
impacts of individual climate change mitigation options in
the U.S. forest sector. Climate Research 3: 121-128.
Woodwell, G.M. 1992. How does the world work? Great issues of
life and government hinge on the answer. Pages 31-37 in R.L.
Peters and T.E. Lovejoy (eds.) Global warming and biological
diversity. Yale University, New Haven, CT.
WRI (World Resources Institute). 1992. World resources 1992-93: A
report by the World Resources Institute in collaboration
211
-------
with the United Nations Environment Program and United
Nations Development Program. Oxford University Press, New
York, NY.
WRI (World Resources Institute). 1993. World resources 1992-93: A
guide to the global environment. Oxford University Press,
Oxford, U.K.
Yamate, G. 1974. Emissions inventory from forest wildfires,
forest managed burns, and agricultural burns. EPA-450/3-74-
062. U.S. Environmental Protection Agency, Office of Air and
Waste Management, Office of Air Quality Planning and
Standards, Research Triangle Park, NC.
Yermolenko, P.M., and L.G. Yermolenko. 1982. Phytomass of
secondary forest phytocenosis of chernevoi subbelt of West
Sayn. Pages 60-71 in A.I. Buzykin (ed.) Formation and
productivity of forest phytocenosis. Institute V.N. Sukache
va for Forest and Wood, Siberian Branch U.S.S.R. Academy of
Sciences, Krasnoyarsk.
Zinke, P.J., A.G. Stangenberger, W.M. Post, W.R. Emanuel, and
J.S. Olson. 1984. Worldwide organic soil carbon and nitrogen
data. Report ORNL/TM-8857, Environmental Sciences Division
Publication No. 2212. Oak Ridgfe National Laboratory, Oak
Ridge, TN.
212
------- |