&EPA
United States
Environmental Protection
Agency
Office of Atmospheric Programs (6207J) EPA 430-R-10-001
Washington, DC 20460 April 2010
Methane and Nitrous Oxide Emissions From
Natural Sources
Permafrost
-------
How to Obtain Copies
You can electronically download this document from EPA's Web page at
http://www.epa.gov/methane/sources.html.
For Further Information
For further information or questions about this report, please e-mail Jason
Samenow (samenow.iason(S)epa.gov) or call the EPA Climate Change Division
hotline at 202-343-9990.
-------
Methane and Nitrous Oxide Emissions
From Natural Sources
April 2010
United States Environmental Protection Agency
Office of Atmospheric Programs (6207J)
1200 Pennsylvania Ave., NW
Washington, DC 20460
-------
Acknowledgments
This report was prepared under a contract between the U.S. Environmental Protection Agency
and Eastern Research Group, Inc. (ERG).
Co-authors of the report were:
• Brian Anderson, West Virginia University
• Karen Bartlett, University of New Hampshire
• Steven Frolking, University of New Hampshire
• Katharine Hayhoe, ATMOS Research and Consulting
• Jennifer Jenkins, Carbon Dynamics, LLC
• William Salas, Applied Geosolutions
EPA thanks the following individuals who served as peer reviewers for this report:
• Stephen Del Grosso, USDA Agricultural Research Service
• Giuseppe Etiope, Istituto Nazionale di Geofisica e Vulcanologia
• Aslam Khalil, Portland State University
• Joel S. Levine, NASA Langley Research Center
• Paulette Middleton, Panorama Pathways
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY ES-1
ES.l INTRODUCTION ES-1
ES.2 SUMMARY OF NATURAL SOURCE EMISSIONS ES-2
ES.2.1 Wetlands ES-6
ES.2.2 Upland Soils and Riparian Zones ES-7
ES.2.3 Oceans, Estuaries, and Rivers ES-9
ES.2.4 Permafrost ES-11
ES.2.5 Lakes ES-12
ES.2.6 Gas Hydrates ES-13
ES.2.7 Terrestrial and Marine Geologic Sources of Methane ES-14
ES.2.8 Wildfires ES-14
ES.2.9 Vegetation ES-16
ES.2.10 Terrestrial Arthropods and Wild Animals ES-16
ES.3 FUTURE NEEDS ES-17
ES.4 REFERENCES ES-18
CHAPTER 1. INTRODUCTION 1-1
-1
-1
-2
-2
-3
-4
-5
-7
1.1 Importance of Methane and Nitrous Oxide as Greenhouse Gases ....
1.1.1 Methane
1.1.2 Nitrous Oxide.
1.2 Sources of Methane and Nitrous Oxide
1.1.1 Methane
1.1.2 Nitrous Oxide.
1.2 Overview of This Report..
1.3 Summary of Methods
1.4 References
CHAPTER 2. WETLANDS 2-1
2.1 Description of Emission Source 2-1
2.2 Factors That Influence Emissions 2-2
2.2.1 Process-Level Controls 2-2
2.2.2 Ecosystem-Level Controls 2-4
2.2.3 Methane Emission Pathways 2-5
2.3 Current Global Methane and Nitrous Oxide Emissions From Wetlands 2-7
2.3.1 Techniques for Making Global Estimates 2-7
2.3.2 Global Wetland Methane Emissions 2-8
2.3.3 Global Wetland Nitrous Oxide Emissions 2-14
2.4 Future Emission Scenarios 2-15
2.5 Areas for Further Research 2-18
2.6 References 2-19
CHAPTER 3. UPLAND SOILS AND RIPARIAN AREAS 3-1
3.1 Description of Emission Source 3-1
3.1.1 Soils as Nitrous Oxide Sources 3-2
3.1.2 Soils as Methane Sinks 3-3
-------
3.2 Factors That Influence Emissions 3-3
3.2.1 Factors That Influence Emissions of Nitrous Oxide 3-4
3.2.2 Factors That Influence Methane Sink Strength 3-4
3.3 Current Global Emissions 3-5
3.3.1 Current Global Emissions of Nitrous Oxide From Soils 3-5
3.3.2 Current Global Sink Estimates for Methane 3-6
3.4 Future Scenarios of Nitrous Oxide and Methane Fluxes 3-8
3.4.1 Future Emissions of Nitrous Oxide From Soils 3-8
3.4.2 Future Emissions of Methane From Soils 3-9
CHAPTER 4. OCEANS, ESTUARIES, AND RIVERS 4-1
4.1 Description of Emission Source 4-1
4.2 Factors That Influence Emissions 4-2
4.2.1 Nitrous Oxide 4-2
4.2.2 Methane 4-3
4.3 Current Global Emissions 4-4
4.3.1 Current Ocean, Estuarine, and Riverine Nitrous Oxide Fluxes 4-4
4.3.2 Current Ocean, Estuarine, and Riverine Methane Fluxes 4-15
4.4 Future Oceanic, Estuarine, and Riverine Nitrous Oxide and Methane Emission Scenarios ..4-18
4.5 Areas for Further Research 4-19
4.6 References 4-19
CHAPTER 5. PERMAFROST 5-1
5.1 Description of Emission Source 5-2
5.2 Factors That Influence Emissions 5-2
5.3 Current Global Emissions 5-3
5.4 Future Emission Scenarios 5-5
5.5 Areas for Further Research 5-5
5.6 References 5-6
CHAPTER 6. LAKES 6-1
6.1 Description of Emission Source 6-1
6.2 Factors That Influence Emissions 6-2
6.2.1 Methane 6-2
6.2.2 Nitrous Oxide 6-4
6.3 Current Global Emissions 6-4
6.3.1 Methane 6-4
6.3.2 Nitrous Oxide 6-6
6.4 Future Emission Scenarios 6-6
6.4.1 Methane 6-6
6.4.2 Nitrous Oxide 6-7
6.5 Areas for Further Research 6-7
6.6 References 6-7
CHAPTER 7. GAS HYDRATES 7-1
7.1 Description of Emission Source 7-1
7.2 Factors That Influence Emissions 7-3
7.3 Current Global Emissions 7-4
7.4 Future Emission Scenarios 7-5
7.4.1 Continental Hydrates 7-5
7.4.2 Oceanic Hydrates 7-6
IV
-------
7.5 Areas for Further Research 7-7
7.6 References 7-8
CHAPTER 8. TERRESTRIAL AND MARINE GEOLOGIC SOURCES 8-1
8.3.1 Mud Volcanoes in Petroliferous Sedimentary Regions 8-8
8.3.2 Seepage in Petroliferous Sedimentary Regions 8-9
8.3.3 Submarine Seepage 8-10
8.3.4 Volcanoes, Vents, and Other Geothermal Sources 8-11
8.3.5 Global Emissions 8-13
CHAPTER 9. WILDFIRES 9-1
9.1 Description of Emission Source 9-1
9.2 Factors That Influence Emissions 9-2
9.2.1 Type of Vegetation 9-3
9.2.2 Influence of Weather and Climate 9-3
9.2.3 Human Influence 9-5
9.3 Current Global Emissions 9-5
9.3.1 Estimating Emission Factors 9-6
9.3.2 Estimating Amount of Biomass Burned 9-7
9.4 Future Emission Scenarios 9-11
9.5 Areas for Further Research 9-12
9.6 References 9-13
CHAPTER 10. VEGETATION 10-1
10.1 Description of Emission Source 10-1
10.1.1 Identification of High Methane Concentrations Over Tropical Forests 10-2
10.1.2 Measurement of Methane Emissions From Plants Under Aerobic Conditions 10-2
10.1.3 Measurement of Ecosystem Methane Flux 10-3
10.2 Factors That Influence Emissions 10-4
10.3 Current Global Emissions 10-4
10.4 Future Emission Scenarios 10-6
10.5 Areas for Further Research 10-6
10.6 References 10-7
CHAPTER 11. TERRESTRIAL ARTHROPODS AND WILD ANIMALS 11-1
11.1 Description of Emission Source 11-1
11.2 Factors That Influence Emissions 11-2
11.3 Current Global Emissions 11-2
11.4 Future Emission Scenarios 11-3
11.5 Areas for Further Research 11-3
11.6 References 11-4
CHAPTER 12. SUMMARY AND CONCLUSIONS 12-1
12.1 Summary of Methane Emissions 12-1
12.2 Summary of Nitrous Oxide Emissions 12-2
12.3 Future Needs 12-4
12.4 Summary 12-5
12.5 References 12-5
APPENDIX A: GLOSSARY A-l
-------
VI
-------
Executive Summary
In 1993, EPA prepared a Report to Congress entitled Current and Future Methane Emissions from
Natural Sources (U.S. EPA, 1993). That report provided global estimates of current and future emissions
of methane (ClrU), a "greenhouse gas," from natural sources. Much new knowledge has emerged since
1993. For example, the Intergovernmental Panel on Climate Change's recent Fourth Assessment Report
(referred to hereafter as "AR4") (Solomon et al., 2007) reviewed the scientific evidence and reached the
strongest conclusions to date regarding climate change. The AR4 focused largely on anthropogenic (or
human-produced) sources, however, and included only a limited assessment of natural source emissions.
This report serves as an update to EPA's original 1993 report on natural sources. Building on the AR4
and other recent efforts, this report summarizes the latest research and provides global estimates of
current and future emissions of CH4 from natural sources, including emissions from newly identified
sources. It also provides global estimates of current and future emissions for nitrous oxide (N2O), another
important greenhouse gas, from natural sources.
ES.1 Introduction
CH4 and N2O are "greenhouse gases," meaning that they trap infrared radiation (heat) from the earth's
surface and increase the temperature of the earth. Without this natural "greenhouse effect," temperatures
would be about 33°C (60°F) lower than they are now, and life as we know it today would not be possible.
During the past century, humans have substantially added to the amount of greenhouse gases in the
atmosphere through activities such as burning fossil fuels and deforestation. The added gases are
enhancing the natural greenhouse effect, and very likely contributing to an increase in global average
temperature and related climate changes.
CH4 and N2O are emitted from both natural and anthropogenic sources. Natural sources of CFLt include
fires, geologic processes, and bacteria that produce CH4 in a variety of settings (most notably, wetlands).
N2O is also produced by bacteria. Major anthropogenic sources of these gases include fossil fuel
combustion and agriculture. Some sources can be related to both natural and anthropogenic processes. For
example, forest and grassland fires, which produce CFLj, can be either human-initiated (e.g., for land
clearing) or the result of lightning ignition or other natural causes.
While much attention is currently focused on anthropogenic sources of greenhouse gases, there is ample
evidence that emissions of these gases from natural sources have also changed overtime. Human
activities have significant potential to change emissions from these sources, both directly (e.g., decreased
CH4 from wetlands, due to wetland loss from draining and filling) or indirectly through human-induced
climate change (e.g., increased CFLt emissions from wetlands due to rising temperature, or from wildfires
that are more frequent and severe). Therefore, to address greenhouse gas emissions, it is important not
only to quantify the current magnitude of natural sources, but also to understand how human activities
and climate change affect emissions from these sources.
This report presents many different estimates of CFLj and N2O emissions, reflecting the variety of
approaches that scientists use to characterize emissions. These approaches generally fall into two
categories: "bottom-up" and "top-down." "Bottom-up" estimates work from a small scale to a larger
scale, extrapolating actual measurements of flux (that is, the release or uptake of a gas) to larger scales, or
developing a model of the processes controlling fluxes and then applying it to a larger scale. "Top down,"
or inverse, methods use atmospheric concentration measurements, atmospheric transport models, and
statistical methods to estimate emissions from individual sources.
ES-1
------- An error occurred while trying to OCR this image.
-------
Executive Summary
It is important in reading this report to keep in mind that the earth is a system of interacting components,
and change often affects many components as well as their interactions. This report is organized into
chapters covering natural sources by type (e.g., wetlands, lakes, oceans, gas hydrates, etc.). However, the
earth is a mosaic of these different source types, the boundaries between these source types are sometimes
inexact (e.g., between a wetland and the emergent vegetation of a lake margin), and system changes that
affect one source can also affect one or more other sources.
The issue of methane in permafrost regions exemplifies this interconnectedness. The projected thawing of
permafrost with climate warming may contribute to increased natural source methane emissions to the
atmosphere. However, this is a complex system response. There is not a lot of methane frozen into
permafrost (unless it is a gas hydrate formation), so permafrost thaw will not release much methane
directly. The methane it does release has a reasonable probability of being oxidized as it diffuses through
1 to 100 meters of thawed soil before reaching the atmosphere (see Chapter 5, "Permafrost"). Therefore,
the report concludes that permafrost thawing is not likely to be a strong methane source. However, if gas
hydrates are associated with permafrost, then thawing permafrost and destabilization of these hydrates
may co-occur (though they are not exactly the same thing), releasing methane, potentially in large
quantities (see Chapter 7, "Gas Hydrates"). Another consideration is changes in wetland vegetation and
moisture status associated with permafrost thawing; this would be an issue for some, but not all,
permafrost landscapes. If the landscape gets wetter and the vegetation composition becomes more
dominated by sedges, this could lead to increased methane emissions from a wet landscape - at least for
years to decades; however, if the landscape gets drier (or stays relatively dry), then methane emissions
would probably stay low (see Chapter 2, "Wetlands"). Finally, permafrost thaw can be associated with
thermokarst erosion, which can form (or drain) lakes; these lakes can also be methane sources (See
Chapter 6, "Lakes"). Each chapter of this report specifies the natural sources addressed by that chapter.
ES-3
-------
Executive Summary
Permafrost
0-1 TgCIVyr
Terrestrial and marine
geologic sources
Figure ES-1. Estimated annual emissions of CH4 and N2O from natural sources. N2O emissions are presented as Tg of
nitrogen (N). Note that permafrost and any permafrost sources of methane occur mostly at high latitudes, not high elevations.
See Table ES-1 for additional detail.
ES-4
-------
Executive Summary
Table ES-1. Current Methane and Nitrous Oxide Emissions From Natural Sources
Source
Wetlands
- Northern/bogs
- Tropical/swamps
Upland soils and riparian
areas
Oceans, estuaries, and
rivers
Permafrost
Lakes
Gas hydrates
Terrestrial and marine
geologic sources
Wildfires
Vegetation
Terrestrial arthropodsf
Wild animals
All natural sources
All sources to the
atmosphere (anthropogenic
and natural)
Natural sources as a percent
of the total
Methane (Tg CH4/year)
Emissions
Estimate3
170.3
42.7
127.6
-30
9.1
0.5
30
20
8
208
566j
37%
Rangeb
24-72
81-206
Not
available
2.3-15.6
0-1
10-50
2-9e
42-64
2-5
Not a
source
or 20-60
2-22
2-15
See
note9
503-
610k
See
note9
513C (%0)c
-62
-58.9
-58
-53.8
-62.5
-41.8
-25
Not
available
-63
-60.5
-57h
-54.5C
n/a
Nitrous Oxide (Tg N/year)
Emissions
Estimate3
Rangeb
515N (%0)d
Negligible
6.6
5.4
3.3-9.0
1.5-9.1
-38 to +2
-2 to +12
Negligible
0.1
12.1
18. 8'
64%
0.004-
0.04
Not
available
See
note9
8.5-27.7
See
note9
8.81
?m
n/a
a In some cases, a point estimate cannot be provided due to large uncertainty.
b Ranges presented here may reflect a compilation of several different estimates. Published estimates vary due in
part to uncertainty in estimating the global number of point and diffuse sources and the average annual emissions
from each individual source or source area.
c Mean value from Whiticar and Schaefer, 2007, and references therein.
d Range from Rahn and Wahlen, 2000, and references therein.
e The emission estimates for gas hydrates correspond to the flux of methane to the ocean, most of which is likely to
be oxidized in the ocean water column.
f Estimates for terrestrial arthropods include termites. It is estimated that other arthropods could contribute up to 100
Tg ChU/year.
ES-5
-------
Executive Summary
g Because the relative contributions of emissions from each source to the total budget are not independent of each
other (i.e., if one source is at the lower end of its estimated range, another may be at the higher), the ranges
cannot be summed.
h Lasseyetal., 2007.
1 Based on change from pre-industrial to present as estimated by Rockmann et al., 2003; assuming that pre-
industrial emissions are primarily natural.
1 Mean value for anthropogenic emissions from Wuebbles and Hayhoe, 2002; natural emissions from this work.
k Range in total anthropogenic and natural emissions from Denman et al., 2007, and references therein.
1 Estimates of anthropogenic emissions from Denman et al., 2007; natural emissions from this work.
m Observed tropospheric values from Rahn and Wahlen, 2000, and references therein.
ES.2.1 Wetlands
Wetlands are ecosystems in which saturation with water is the dominant factor controlling soil
development and the species of plants and animals that occur. Covering about 5 percent of the Earth's
surface, they are concentrated in the high latitudes, where frozen soils can inhibit water drainage, and in
the tropics, where precipitation rates are high. Bacteria in the moist, anoxic (oxygen-free) wetland soil
produce CFU as they decompose dead plant material, making wetlands an important CFU source. The
opposite is true for N2O. Although bacteria found in wetlands do produce N2O, flooded conditions tend to
favor bacteria that consume N2O and produce nitrogen gas (N2).
Emissions of CF^ and N2O from wetlands to the atmosphere are a small residual of the much larger
amounts produced and consumed in wetland soils. The different types of bacteria in wetlands that produce
and consume these gases are affected differently by environmental factors (e.g., temperature, water level,
and organic matter supply and characteristics). Therefore, a relatively small environmental change can
result in a large change in flux by changing the balance between production and consumption.
Methane
Current Emissions Estimates
The earlier version of this report (U.S. EPA, 1993) estimated total emissions to be 109 Tg CHVyr,
including 38 Tg CHVyr from high latitude wetlands, 5 Tg CHVyr from temperate wetlands, and 66 Tg
CHVyr from tropical wetlands. The 1993 report extrapolated emissions based on measurements of actual
gas flux from wetlands, however, so that only a relatively small number of sites represented these diverse
ecosystems. Measurements have shown that gas fluxes are highly variable over time (e.g., from day to
day, week to week, and year to year). Fluxes also vary from place to place, even within the same wetland.
Although there are clear seasonal changes as well as differences between wetland types, the high
variability makes large-scale estimates difficult. Average emissions therefore have large uncertainty
estimates. High variability occurs because many small-scale factors combine to influence fluxes and small
environmental differences can cause large differences in flux. Variability is also high because of the many
ways emissions can occur. Bubbling, for example, only happens sporadically but can release large
amounts of gas.
In recent years, more sophisticated models have emerged for estimating emissions from wetlands. There
are still substantial uncertainties attached to emissions estimates, but using models to calculate emissions
also allows estimation effluxes under changed environmental or climate conditions. The emissions
estimates in this report (see Table ES-1) were derived by taking a simple average of the many estimates
that have been made since 2004. Flux estimates for high latitude wetlands range from 24 to 72 Tg CHVyr,
ES-6
-------
Executive Summary
with an average of 43 Tg CHVyr. Reported emissions from tropical wetlands range from 81 to 206 Tg
CHVyr, with an average of 128 Tg CHVyr. Overall, these estimates reflect an increase over the figures in
the 1993 report, although uncertainties are still quite high. Although the 1993 report suggested that
emissions from the tropical latitudes made the greatest contribution to global fluxes, the recent numbers
have increased their importance from roughly 61 percent to about 75 percent of the total CH4 emissions
from wetlands worldwide.
Future Emissions Scenarios
Changes in land use and climate will affect QrU emissions from wetlands, with the potential for both
large increases and large decreases. At high latitudes, changes in climate are thought to be the major
factors driving changes in CUt emissions. For example, models using altered temperatures and rainfall
suggest that emissions from northern wetlands could double by the end of the 21st century. In the tropics,
changes in land use, such as draining or filling wetlands for other uses, are believed to be the major driver
of change.
Models suggest that sulfur deposited in wetlands by acid rain may decrease CFLt flux. Changes in wetland
plants (for example, in the species present or in their growth rate) can affect emission pathways
(movement through soil layers, bubbling, or movement through plants), as well as the quality and
quantity of organic material available to be decomposed.
Nitrous Oxide
The available research indicates that wetlands are a negligible source of N2O. There is some evidence that
they may be a small sink (i.e., removing N2O from the atmosphere) but no global estimates have been
made.
If water levels in wetlands drop significantly under altered rainfall patterns, it is possible that these
systems may more closely resemble upland soils and may therefore emit N2O.
ES.2.2 Upland Soils and Riparian Zones
Upland soils are well-aerated, not water-saturated, and generally oxic (that is, containing oxygen), with
dry soil conditions. These conditions favor microbial processes that make these soils a sink for ClrU and a
source of N2O. Natural sources include upland soils associated with forests and grasslands under natural
vegetation, but not agricultural lands.
Riparian zones, located at the interface of terrestrial and aquatic environments, are often permanently wet
and rich in organic matter, with saturated soil conditions and microbially available carbon that contribute
to higher rates of production of N2O than dry upland soils.
Many interrelated factors determine both the magnitude of emissions of N2O from upland and riparian
soils and the sink strength for CFLj. As the carbon and nitrogen cycles in soils are linked, changes in
nitrogen and carbon availability strongly influence the rate of emission or sequestration. Recent studies
have found that soil organic carbon content, vegetation type, soil pH, bulk density, and drainage are the
major factors influencing N2O emissions. The strength of soil as a sink for CFLj depends on oxidation by
methanotrophic (methane-using) microbes in the soil, and therefore is influenced by environmental
factors that control this oxidation rate. The primary factor is soil diffusivity, which controls the amount of
CH4 transferred into the soil and, therefore, its availability to methanotrophs.
Both upland soils and riparian soils have been significantly impacted by human activities. The arable
lands composed of dry upland soils have been cleared for agricultural use, which is responsible for an
ES-7
-------
Executive Summary
estimated 80 percent of anthropogenic emissions of N2O through soil emissions, biomass burning, and
animal production. Riparian zones have been significantly impacted by agricultural activities as well.
Riparian buffer zones serve as sites for nitrate removal from agricultural runoff, and are often loaded with
high levels of nitrogen.
Methane
Current Emissions Estimates
CH4 sink strength of soils under natural vegetation (including upland and riparian soils) is estimated at 30
Tg CHVyr. Process-level (bottom-up) methods of estimating ClrU budgets (i.e., the balance of sources and
sinks) contain significant uncertainties due to the aggregation of local measurements, taken on short time
scales and with large spatial variability. While recent strides have been made in collecting and analyzing
emissions measurements and the source strength of tropical soils has been characterized, there still exist a
lack of field measurements and significant model uncertainties.
Future Emissions Scenarios
Future CFLj oxidation by soils will depend on the changing human activities on these soils, as well as on
climate patterns that are shifting as a result of global climate change. Clearing land for agricultural use
has been shown to lead to a decreased capacity for QrU oxidation, for example. Global climate models
show patterns of temperature and precipitation changes worldwide. As soil moisture is a key determinant
of the microbial processes that consume QrU, these shifting climate patterns will determine the fluxes of
CH4 into the future.
Nitrous Oxide
Current Emissions Estimates
Based on the available data, emissions of N2O from soils under natural vegetation (including upland and
riparian soils) are estimated at 6.6 Tg N/yr.
Many microbiological, chemical, and physical parameters affect N2O emissions, and complex interactions
among these factors make extrapolating global emissions budgets difficult and uncertain. Further, the vast
majority of studies in the past have focused on N2O emissions from agricultural, not natural, soil sources.
Since the publication of the AR4, the number of N2O emissions measurements has been increasing
steadily, allowing for improvements in emission models and budgets. However, while recent strides have
been made in collecting and analyzing emissions measurements and the source strength of tropical soils
has been characterized, there still exist a lack of field measurements and significant model uncertainties.
Future Emissions Scenarios
As with CFi4, future N2O emissions will depend on changing human activities and climate patterns.
Clearing land for agricultural use has been shown to lead to increased N2O emissions, for example.
Because soil moisture is a key determinant of the microbial processes that consume or produce N2O,
shifts in global precipitation and temperature patterns will affect N2O fluxes in the future.
ES-8
-------
Executive Summary
ES.2.3 Oceans, Estuaries, and Rivers
In water bodies, CH4 and N2O are produced by microbial processes that occur both in the water column
and in sediments. Ultimately these gases can enter the atmosphere. This report considers a range of
aquatic sources, including the deep waters of the open ocean, shallower coastal waters (i.e., on the
continental shelves), freshwater rivers, and estuaries where fresh and salt waters mix.
Methane
Sampling conducted in the 1960s and 1970s found that in general, surface aquatic waters are relatively
small sources of CFU to the atmosphere. Methane production is greater in areas under freshwater and in
shallow waters with highly organic sediments. In the open ocean, enhanced CH4 production and flux can
be found in upwelling areas, which are areas where the prevailing winds and currents bring nutrient-rich
deep water to the ocean surface. Upwelling areas tend to have higher rates of primary productivity, which
in turn leads to more organic material falling to depth, depleting oxygen levels and creating favorable
conditions for methane-producing bacteria.
Current Emissions Estimates
CH4 fluxes from water bodies are typically calculated from surface dissolved concentrations and wind
speeds. Because the global flux of CH4 from oceans, estuaries, and rivers is relatively small, these sources
have not been the focus of extensive research, and the available data have geographic and seasonal
limitations that add considerable uncertainty to any overall estimates. For example, no data are available
for tropical latitudes or upwelling zones.
Open ocean
Open ocean emissions are low and dispersed over large areas, and thus they are difficult to resolve with
techniques such as inverse modeling that use changes in atmospheric concentrations to estimate flux. The
most recent estimates for the open ocean fall into two groups, with some estimates of less than 1 Tg
CtVyr and several others around 4 Tg CHVyr. For this report an estimate was derived by calculating a
simple average of 1.8 Tg CHVyr.
Coastal ocean areas
Emissions from the continental shelves are somewhat higher than those from the open ocean, even though
these environments cover a much smaller area. This difference likely reflects greater organic inputs and
an increase in sedimentary contributions. An average of recent estimates gives a total flux of 5.5 Tg
CHVyr.
Estuaries and rivers
Estuaries and rivers cover limited areas, yet they contain much biological activity. They are also sites of
active mixing, enabling CF^ produced in adjacent wetlands and shallow-water environments to be rapidly
released to the atmosphere. Published global emissions estimates from estuaries and rivers are sparse.
Averaging the available estimates for estuaries and adding the only available estimate for rivers results in
a total flux of 1.85 Tg CHVyr.
Overall, natural CFU emissions from oceans, estuaries, and rivers are estimated at 9.1 Tg CHVyr (see
Table ES-2). Uncertainties are due to sparse data, poor habitat coverage, and lack of any tropical or
southern latitude sampling. It is, however, very similar to estimates made since the mid-1970s (10 to 15
Tg CHVyr). Although better flux estimates would improve confidence in emissions, it is unlikely that
they would change global estimates by more than a few Tg, as natural emissions from oceans, estuaries,
and rivers represent only about 2 percent of the total global CH4 emissions to the atmosphere every year.
ES-9
-------
Executive Summary
Table ES-2. Natural Emissions of CH4 From Oceans, Estuaries, and Rivers
Open ocean
Continental shelves
Estuaries and rivers
Total
Annual Emissions, Tg CH4/yr
1.8
5.5
1.85
9.1
Percent of Total
Emissions
20%
60%
20%
Future Emissions Scenarios
Natural emissions of ClrU from oceans, estuaries, and rivers are expected to remain largely unchanged in
the future.
Nitrous Oxide
The oceans are believed to be one of the largest natural sources of N2O emissions. Estuaries and rivers
also contribute N2O to the atmosphere; however, emissions of N2O from these other aquatic environments
are typically classified as anthropogenic because the majority of nitrogen entering these systems is
believed to be associated with human activities such as agriculture.
Current Emissions Estimates
Open ocean
Published estimates of open ocean fluxes generally range from 3 to 6 Tg N/yr. For example, the AR4
estimates natural emissions from oceans at 3.8 Tg N/yr, with a range of 1.8 to 5.8 Tg N/yr. This report
combines several recent estimates to calculate an open ocean N2O flux of 3.2 Tg N/yr, consistent with the
AR4 estimate. The most significant recent development in open ocean emissions estimates has been a
better understanding of the geographic distribution effluxes. Earlier reports calculated that the ocean area
from 30°S to 90°S made the largest contribution to global emissions (about 45 percent of the total). More
recent work has decreased this flux estimate, and inverse modeling results suggests that this region may
contribute as little as 7 percent to the global total.
Coastal ocean areas
Because continental shelves receive drainage from rivers and estuaries, they are impacted by humans.
Several approaches have been used to estimate emissions, and published estimates vary by about an order
of magnitude. By combining these estimates with recent models of natural and anthropogenic nitrogen
export from rivers and estuaries, this report estimates that the flux of natural N2O from the continental
shelves is 1.5 Tg N/yr. Enhanced emissions are expected in upwelling zones; however, exact estimates are
complicated because upwelling intensity can vary overtime and space. Currently, regional/short-term
upwelling estimates do not agree with global/annual flux estimates. A simple average of reported global
fluxes from upwelling zones is 0.4 Tg N/yr. The estimates for both continental shelves and upwelling
areas have substantial uncertainties, but these are unlikely to change estimates by more than 1-2 TgN/yr.
Estuaries and rivers
Estuaries and rivers are highly impacted by human activities and the corresponding changes in continental
nitrogen budgets. The AR4 estimates N2O emissions of 1.7 Tg N/yr from coasts, estuaries, and rivers
ES-10
-------
Executive Summary
(with a range of 0.5 to 2.9 Tg N/yr), and assumes that these emissions are entirely anthropogenic. Based
on published models, we estimate that the flux of natural N2O from estuaries is approximately 0.24 Tg
N/yr and the natural flux from rivers is 0.09 Tg N/yr, for a total of 0.33 Tg N/yr.
Together, the estimated natural N2O fluxes from oceans, estuaries, and rivers total 5.4 Tg N/yr (see Table
ES-3). This figure is at the upper range in uncertainty that the AR4 predicts for oceans, estuaries, and
rivers (5.8 TgN/yr); however, the AR4 assumes that all N2O from coasts, estuaries, and rivers reflects
anthropogenic sources only. If total global N2O emissions are 18.8 Tg N/yr (Table ES-1), then natural
emissions from these environments contribute about 29 percent of all N2O emissions worldwide.
Table ES-3. Natural Emissions of N2O From Oceans, Estuaries, and Rivers
Open ocean
Continental shelves
Upwelling zones
Estuaries
Rivers
Total
Annual Emissions, Tg N/yr
3.2
1.5
0.4
0.2
0.1
5.4
Percent of Total
Emissions
59%
28%
7%
4%
2%
Future Emissions Scenarios
Like CH/i, natural emissions of N2O from oceans, estuaries, and rivers are expected to remain largely
unchanged. Estimates should become more accurate, however, as the accumulating database of
atmospheric N2O measurements makes it possible to use inverse modeling techniques to determine
emissions on smaller time and spatial scales. Greater precision is important because emissions change
seasonally and some sources (e.g., upwelling areas) are highly episodic and occur only over relatively
small areas.
Overall, it is not expected that possible changes in oceanic emissions of N2O will greatly affect climate
policy. While ocean N2O emissions do make a significant contribution to global emissions, most of these
emissions are from the open ocean and are less susceptible to anthropogenic impacts. Based on the
current understanding of emissions, the major controls are fundamental physical oceanic properties (e.g.,
wind parameters and ocean mixing) that are not easily influenced by human activities.
ES.2.4
Permafrost
Permafrost is soil, sediment, or rock that is continuously frozen (temperature < 0°C) for at least two
consecutive years. Permafrost is widespread and nearly continuous in the arctic, but also exists
intermittently in the sub-arctic and boreal regions, and at high elevation. EPA's 1993 report included
permafrost as a natural source of ClrU, because early results were emerging on QrU frozen within
permafrost which could be released from the permafrost as it melts.
Methane
Emissions estimates are based on measured concentrations of ClrU in permafrost and estimates of
contemporary permafrost degradation rates. Current QL releases from permafrost are estimated to be 0 to
1 Tg QrL/yr. The AR4 does not include permafrost as a natural source of CH/j.
There is now strong evidence that permafrost is melting, and that a substantial fraction of permafrost
existing now will be melted within the next 100 years due to global climate change. However, it now also
ES-11
-------
Executive Summary
seems clear that this permafrost melting will be only a small direct source of CH4. CFL, concentrations in
permafrost are not high, and as CH4 is released from melting permafrost, it must pass through the
overlying thawed soil before reaching the atmosphere. During this transport, some or most of the CFL, will
be oxidized to carbon dioxide (CO2) before reaching the atmosphere.
Indirectly, however, thawing permafrost is already impacting CFL, emissions from other natural sources,
particularly lakes and wetlands. Permafrost can contain ice wedges, which are lenses of frozen water that
can be up to several meters thick. As permafrost thaws and ice wedges melt, this water can sometimes
drain away, leading to ground subsidence or collapse, which in turn can alter drainage patterns. In this
process, known as thermokarst erosion, lakes and wetlands can form or can dry up. This process affects
CH4 emission rates from high latitude lakes and wetlands.
Nitrous oxide
N2O concentrations are estimated to be about 1,000 times smaller than CFL, although few studies have
measured them, and global emissions of N2O from permafrost are considered negligible. The AR4 does
not include permafrost as a natural source of N2O. Because N2O concentrations in permafrost are very
low, little N2O is likely to be released on melting.
ES.2.5 Lakes
Lakes and ponds are naturally formed permanent water bodies contained on a body of land. This source
category includes natural freshwater lakes but excludes impoundments and reservoirs (water bodies
formed by dams), as greenhouse gas emissions from impoundments, reservoirs, and other engineering
works are considered to be anthropogenic.
Lakes contribute to both CFLand N2O global emissions, although analysis of this source has been limited
to date. CH4 is produced by the activity of methane-generating bacteria in anoxic sediments, while N2O is
produced by microbial activity in sediments and water as an intermediate product of both an aerobic
process called nitrification and an anaerobic process called denitrification.
Methane
CFL production rates depend on temperature, organic matter availability (food for the bacteria), and
isolation from oxygen; these factors are influenced by climate, lake size and depth, and productivity of
microscopic and macroscopic plants and animals, which create organic matter for CFL production when
they die and sink to the bottom. There are four pathways for CFL emissions from lakes: bubbling,
diffusion, plant-mediated transport, and seasonal overturning. Bubbling has been determined to be the
dominant pathway for CFL flux, accounting for more than 90 percent of CFL emissions from lakes. Wind
speed is an important control on gas exchange between a lake and the atmosphere. Flux rates by all
pathways generally increase with increasing wind speed.
Current Emissions Estimates
Based on recent estimates, lakes emit approximately 30 Tg CFL to the atmosphere per year. One key
uncertainty involves the total surface area of lakes and ponds. The number and total area of large lakes is
well known, but the number and total area of small lakes and ponds is not. A recently published estimate
suggests that there are about 300 million natural lakes and ponds worldwide, 90 percent of which are
smaller than 1 hectare (0.01 km ). Lakes smaller than 1 km constitute about 40 percent of the total global
lake surface area. Because small lakes and ponds generally emit more CFL per unit area than large lakes,
uncertainties about total surface area are a major factor in the overall uncertainty of the estimate.
ES-12
-------
Executive Summary
Future Emissions Scenarios
Climate warming impacts on permafrost and the development of thermokarst lakes could substantially
affect future CIL emissions from lakes. It is estimated that emissions from lakes north of 45°N will
eventually decrease, due to lake area loss and permafrost thaw. Before this long-term decline, though,
would come a period of increased CIL emissions associated with thermokarst lake development in the
zone of continuous permafrost. CIL emission rates from northern lakes could rise as high as 50 to 100 Tg
CIL/yr during this transitional period, which would last hundreds of years.
Nitrous Oxide
Because nitrification and denitrification are highly sensitive to oxygen availability, oxygen concentrations
are an important factor controlling the balance between the two processes. Lake oxygen concentration is
affected by water temperature, water depth, and the rate at which oxygen is consumed by organisms
living in lake water and sediments. The amount of nitrogen available as ammonium (for nitrification) and
nitrate (for denitrification) is also an important control on N2O production.
Current Emissions Estimates
Lakes are generally considered a weak source of N2O to the atmosphere (estimated at 0.004-0.04 Tg
N2O/yr). Accordingly, few data have been collected. Global emissions from lakes are likely to be much
smaller than emissions from other natural N2O sources such as soils, oceans, and estuaries.
Future Emissions Scenarios
No estimates of future N2O emissions from lakes have been published. Increased nitrogen loading and
increased temperatures may cause an increase in N2O fluxes from lakes, but total N2O flux from lakes is
likely to remain a very small fraction of total global N2O emissions from natural sources.
ES.2.6 Gas Hydrates
Gas hydrates are ice-like crystals formed between water and a gas molecule under high pressure and
ambient temperatures. Gas hydrates can store large amounts of the gases that they trap, and are stable
within a specific range of temperature and pressure known as the hydrate stability zone. Large quantities
of CIL are currently trapped in hydrate form, occurring mainly on continental shelves and to a lesser
extent below permafrost regions.
Current Emissions Estimates
Under current conditions, the CIL emissions from gas hydrates are small (estimated at 2 to 9 Tg CtL/yr);
however, the potential for significant CH4 release from gas hydrates warrants close examination of this
source. A significant fraction, if not all, of these emissions are expected to be oxidized in the ocean water
column.
Since 1993, there has been limited discussion of the current flux of CIL from gas hydrate reservoirs.
Oceanic and onshore continental reserves are believed to be stable at present, which means that they are
not currently emitting CIL. Offshore continental shelf reserves are currently unstable, however, and are
believed to emit CIL. Estimates assume that the CIL being liberated from the gas hydrate form is released
into the atmosphere. It is possible, however, that some or all of this gas is not actually emitted to the
atmosphere—that, instead, it is oxidized or absorbed within the sediment or dissolved into the water
column.
ES-13
-------
Executive Summary
Future Emissions Scenarios
Due to their proximity to the Earth's surface (< 2,000 meters) and the sensitivity of the hydrate stability
zone, gas hydrates will likely be affected by climate change. Pressure on hydrates is expected to change
as a result of sea level rise and the melting of polar ice caps; however, temperature changes are likely to
have a far more significant effect. QrU emissions from this source are likely to increase as temperature
rises. Based on recent research, it is estimated that the increase in methane emissions due to an increase in
ocean temperatures may be sufficient to overcome oxidation in the water column and result in significant
atmospheric methane emissions. The magnitude of the methane emissions expected to reach the
atmosphere due to release from methane hydrates upon ocean warming is, however, uncertain.
ES.2.7 Terrestrial and Marine Geologic Sources of Methane
CH4 and other hydrocarbons can seep naturally from geologic sources deep within the Earth's crust. Some
geologic CH4 emissions are produced via what is known as macroseepage, which includes relatively large
localized emissions from identified geologic features and events such as mud volcanoes and localized
vents. Emissions can also result from diffuse soil exhalation or degassing in volcanically active or other
geothermal regions, and from microseepage in petroliferous or hydrocarbon-containing sedimentary
basins. Sources include both marine (underwater) and terrestrial (land-based) faults.
Current Emissions Estimates
Previous estimates of natural sources have either ignored this source or only evaluated "traditional" but
actually minor sources such as high temperature magma-producing volcanoes. More recent estimates
include emissions from mud volcanoes, other macroseepage locations, terrestrial microseepage, and
submarine seeps.
Submarine estimates are extremely uncertain, particularly when estimating the proportion of emissions
that are absorbed by ocean water before reaching the surface. In contrast, estimates of onshore emissions
can be based on direct measurements and standard emission factor concepts applied to point sources (for
individual features such as mud volcanoes) and more diffuse area sources (for microseepage). "Bottom-
up" emissions estimates for both marine and terrestrial sources generally lie in the range of 32 to 74 Tg
CHVyr. This range largely reflects uncertainty in estimating both the global number of sources and the
proportion of emissions that actually reach the atmosphere, rather than being absorbed by ocean water.
Recent isotopic constraints on the budget suggest a narrower range of 42 to 64 Tg CHVyr, based on the
total burden of "fossil" (radiocarbon-free) methane in the atmosphere.
Future Emissions Scenarios
Relatively few climate- or human-related factors are believed to be capable of influencing emissions of
CH4 from geologic sources. Some reports suggest decreased emissions associated with large-scale
extraction of oil and gas, and increased emissions following deglaciation events and the corresponding
increase in seismic activity (i.e., post-glacial rebound). While geologic CH4 emissions have almost
certainly changed in the past and are likely to continue to change in the future, these mechanisms are too
speculative to use as a basis to estimate even the potential direction of future changes in geologic ClrU
emissions.
ES.2.8 Wildfires
Wildfires are fires in forests, grasslands, savannas, and shrublands. They can either be ignited by
lightning strikes or started accidentally by humans, but do not include deliberate controlled burns for
land-clearing activities. As they burn, wildfires release a number of greenhouse gases, particulates, and
ES-14
-------
Executive Summary
other air pollutants. When combustion is complete — for example, in dominantly flaming fires — the
carbon in biomass is generally converted to CC>2 and the nitrogen is converted to oxidized forms such as
N2O. When combustion is incomplete, particularly in smoldering fires, some carbon is released in the
form of carbon monoxide (CO) and
Although emissions from wildfires may be virtually indistinguishable from controlled burns, an important
distinction must be made between natural fires (induced by lightning or accidentally started by humans)
and anthropogenic fires (deliberately human-initiated). Almost 90 percent of all biomass burning is
considered to be deliberate or planned anthropogenic burning. Much of it occurs in the tropics, where
savanna and forest fires are driven by land clearing for agriculture and the need for fuelwood. Prescribed
burning for forest management and agricultural waste burning is also prevalent in temperate-boreal
regions such as the boreal forests of Canada and the temperate forests of the eastern U.S.
Methane
Current Emissions Estimates
Wildfire emissions of QrU are estimated to range from 2 to 5 Tg per year. This range depends on the
frequency and strength of wildfires, which in turn are determined by a number of factors, including type
of vegetation burned, influences from weather (e.g., wind, humidity, temperature) and climate (e.g., large-
scale climatic patterns such as El Nino), and influences from humans (e.g., settlement, transportation, and
recreation patterns).
As noted above, it is extremely difficult to distinguish between the burned areas from natural wildfires
and those from anthropogenic fires, which makes it difficult to quantify wildfire ClrU emissions relative to
the global total from all biomass burning. Estimates of wildfire ClrU emissions are currently based on the
assumption that about 10 percent of global biomass burning is natural, although this proportion is likely to
vary from year to year.
Future Emissions Scenarios
Future climate change is likely to lead to enhanced frequency of weather conditions associated with high
wildfire risks in many regions of the world. Climate change could affect multiple elements of wildfires,
including fire behavior, ignition, fire management, and vegetation fuels. The complex interactions
between each of these factors will determine future spatial and temporal distribution of wildfires and their
emissions.
Currently, no scenarios for future QrU emissions from global wildfires exist in the literature. Efforts are,
however, directed toward developing models that can predict or forecast wildfire events and can therefore
be used to estimate emissions.
Nitrous Oxide
Current Emissions Estimates
N2O emissions depend not only on the amount of biomass burned, but also on the nitrogen content of this
biomass and the type of fire (smoldering versus flaming), which can vary from one ecosystem to another.
Based on the limited emission factors available, EPA estimates global wildfire N2O emissions to be
approximately 0. 1 Tg N/yr, although no reference for this value is available in the peer-reviewed
literature. Given the methodological problems associated with estimating the amount of biomass burned
in wildfires, a valid statistical error analysis of the emission estimates cannot be performed.
ES-15
-------
Executive Summary
Future Emissions Scenarios
No scenarios for future N2O emissions from global wildfires exist in the literature.
ES.2.9 Vegetation
Plants have long been recognized as important conduits for ClrU emissions, transporting QrU from
anaerobic soils and sediments to the atmosphere. However, it is only recently that plants themselves have
been considered a possible source of CH4 production. In 2006, it was reported that plants emit QrU
through an unidentified process under aerobic conditions, and that this previously unrecognized source
could add up to as much as 62 to 236 Tg CWyr, or 10 to 40 percent of global QrU emissions. A
significant plant ClrU source could help explain a number of gaps in the understanding of current and past
global CH4 budgets, including an apparent large unidentified ClrU source in the tropics.
Current Emissions Estimates
The initial estimate of the size of a possible plant ClrU source has since been revised downward in a
number of analyses that have either scaled emission rates measured in the laboratory by estimates of
global plant production ("bottom-up" estimates) or worked backward from global budgets to determine
how large a plant ClrU source could be reconciled with current estimates of other ClrU sources ("top-
down" estimates). The range of greatest agreement among these estimates is about 20 to 60 Tg CHVyr.
However, few studies have attempted to measure direct ClrU emissions from plants, and those reports
include a finding of no significant emissions. Given this finding and the uncertainties regarding the
underlying mechanism, a best estimate of global plant emissions must also include the possibility of zero
emissions — i.e., that plants are in fact not a direct source of QrU.
Future Emissions Scenarios
The recently proposed aerobic plant QrU source has not yet been incorporated into simulations of future
CH4 emissions. However, future plant emissions would likely depend on changes in the distribution of
different vegetation types, as well as changes in environmental factors that might control emission rates.
Current estimates attribute 35 to 50 percent of global plant emissions to tropical forests, with the second
largest source, tropical savanna and grasslands, contributing about 20 percent. These estimates suggest
that future plant emissions will depend largely on changes in climate and land use in the tropics.
ES.2.10 Terrestrial Arthropods and Wild Animals
Termites and other terrestrial arthropods produce QrU as a result of microbial degradation of ingested
organic matter. QrU is also produced by enteric fermentation, a normal digestive process that occurs in
ruminant animals such as bison, deer, elk, mountain goats, and sheep, as well as in some smaller rodent
species. In the 1993 report, EPA cited termites as a contributor to natural QrU emissions, but did not
discuss this source. It also did not consider contributions from any other terrestrial arthropods or from
enteric fermentation in animals. Note that enteric fermentation also occurs in cattle and other
domesticated ruminants; however, emissions from livestock are considered anthropogenic, so this report
estimates emissions from wild animals only.
Current Emissions Estimates
In the years since the publication of the 1993 report, additional investigation of ClrU emissions from
termites has resulted in more refined estimates of emissions from the various termite species, and has
suggested that some QrU may be oxidized in termite mounds prior to atmospheric release. The factors
that determine the magnitude of emissions of CH4 from terrestrial arthropods include the species of
ES-16
-------
Executive Summary
arthropod, including the specific type of termite. The highest rates of QrU are produced by arthropods
with methanogenic (methane-producing) bacteria, which are found in many species of termites. Other
terrestrial arthropods have been studied to assess whether they generate CH4 and should be included in
any estimates of global emissions. Based on this new research, termites and other terrestrial arthropods
continue to be a small but not insignificant contributor to global QrU emissions, producing an estimated 2
to 22 Tg CH4 per year.
Estimates of emissions from wild animals range from 2 to 15 Tg CHVyr.
Future Emissions Scenarios
Emissions from terrestrial arthropods (including termites) and wild animals are not expected to change
substantially in the future. Changes to land use, which alter the type of plants available for wild
ruminants, could affect the diets of these animals and subsequently their rate of enteric fermentation. As
human activity encroaches on wildlife ecosystems, wild animal populations will likely decrease due to
habitat unavailability. The habitats for terrestrial arthropods and wild animals are also linked to climate
effects, resulting in shifting ecosystems (e.g., in more northern environments) or drought, which are likely
to decrease populations. Currently, no scenarios for future QrU emissions from this source exist in the
literature.
ES.3 Future Needs
It continues to be difficult to estimate contributions from natural sources, and uncertainties can be large,
as evidenced by the large ranges associated with the emissions estimates. Additional research focused on
improving our understanding of the processes that result in CH4 and N2O emissions should improve
current flux estimates and help refine future estimates under altered environmental conditions. High
uncertainty in some sources is a result of a lack of basic data - flux measurements may be sparse from
some geographic regions and/or seasons. For a number of sources such as wetlands, uncertainties are high
in part because these are highly dynamic systems that respond to short-term climate and weather
variability with changed emissions. This source of uncertainty will always be present. A number of
sources currently rely on inventory-type data to extrapolate small-scale measurements. While this is
reasonable for some sources (for example, the number of mud volcanoes is unlikely to change quickly or
drastically), this means that they are largely static estimates. Even if modeled, these flux estimates will be
limited by the spatial and temporal resolution of the data used for their extrapolation. Reliance on
inventory or long-term average data also means that it is difficult to fully take advantage of the
accumulating database of atmospheric mixing ratios and isotopic signatures. These data are highly
dynamic and this short-term variability is a crucial part of their utility in inverse modeling approaches.
These techniques have proven that they can both help to constrain "bottom up" estimates and provide a
way to integrate highly variable natural systems.
For wetlands, the major natural source contributing to CFI4 emissions, research in tropical areas remains
sparse and incomplete. Increased work linking emissions to environmental controls, long-term studies to
capture seasonality and inter-annual variability, and work on the importance of episodic emissions will
help resolve difficulties in modeling these systems. In addition, more work should examine the
relationships between CFLj flux and net primary productivity (the rate at which biomass is produced, for
example by photosynthesis), since these relationships appear to be habitat-specific. Because emissions to
the atmosphere are a function of the competing processes of CFLj production and consumption, both
processes and their responses to environmental controls must be understood across the landscape.
Episodic emissions, which may release a sizeable fraction of annual flux, remain difficult to measure and
include in models. Failure to adequately incorporate these fluxes, however, can yield inaccurate and
misleading results.
ES-17
-------
Executive Summary
For upland soils and riparian zones, the major natural source contributing to N2O emissions, more field
measurements and improvements in global emissions models are needed. While field measurements of
N2O have increased steadily in the past several years, coverage of global vegetation zones remains
incomplete. More measurement data are needed, especially for the dry tropical forest, savanna, tundra,
and temperate ecosystems not affected by nitrogen deposition. These measurements should be carried out
over extended periods, to help improve our understanding of the complex factors that impact emissions as
well as to assess natural variability.
There are many additional areas where research would help improve flux estimates. These are discussed
in more detail in the source-specific chapters, but we briefly list a number of them here, where
uncertainties are notably high. They include:
• Data from tropical and southern latitude oceans, estuaries, and rivers, as well as estimates of
upwelling sources.
• Improvements to permafrost models to account for lateral water movement, dynamic vegetation
algorithms, and detailed soil physics.
• Data to quantify lake fluxes, particularly in the Arctic, boreal region, and tropics.
• Better quantification of CFL, reserves stored as gas hydrates, as well as better estimation of the
rate of CFL, absorption into oceans and CH4 oxidation in the water column.
• Rates of CFL, from seeps and mud volcanoes oxidized in sediments, as well as better
quantification of the source locations (e.g., number of mud volcanoes, frequency of eruption).
• Activity data for wildfires, including area and amount of biomass, burned area estimates
associated with natural wildfires, and additional research on emissions related to different "fuels"
(i.e., different types of vegetation).
• Confirmation or rejection of vegetation as a source of CH4.
• Research that better quantifies the oxidation of CFL, through termite mounds, confirmation of CH4
from non-termite terrestrial arthropods, and activity data for arthropods and wild animals.
ES.4 References
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohman, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: Solomon, S.,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). 2007.
Climate Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY:
Cambridge University Press.
Lassey, K.R., D.M. Etheridge, D.C. Lowe, A.M. Smith, and D.F. Ferretti. 2007. Centennial evolution of
the atmospheric methane budget: What do the carbon isotopes tell us? Atmos. Chem. Phys. 7: 2119-
2139.
Rahn, T., and M. Wahlen. 2000. A reassessment of the global isotopic budget of atmospheric nitrous
oxide. GlobalBiogeochem. Cycles 14(2): 537-543.
Rockmann, T., J. Kaiser, and C.A.M. Brenninkmeijer. 2003. The isotopic fingerprint of the pre-industrial
and the anthropogenic N2O source. Atmos. Chem. Phys. 3: 315-323.
ES-18
-------
Executive Summary
Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.)
2007. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and
New York, NY: Cambridge University Press.
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Whiticar, M., and H. Schaefer. 2007. Constraining past global tropospheric methane budgets with carbon
and hydrogen isotope ratios in ice. Phil. Trans. R. Soc. A 365: 1793-1828.
Wuebbles, D.J., and K. Hayhoe. 2002. Atmospheric methane and global change. Earth-Sci. Rev. 57: 177-
210.
ES-19
-------
Executive Summary
ES-20
-------
Chapter 1. Introduction
Methane (CH4) and nitrous oxide (N2O) are important greenhouse gases, present in the atmosphere, that
are produced in part by natural sources. Greenhouse gases prevent heat emitted by the Earth from
escaping to space (Figure 1-1). The natural
greenhouse effect is necessary to life as we know it.
It maintains the Earth's surface temperature at an
average of 15°C, 33°C warmer than it would be
otherwise (NOAA, 2007). Because greenhouse
gases can absorb infrared radiation, changes in their
atmospheric concentrations can alter the energy
balance of the climate system. Increases in
greenhouse gas concentrations in the atmosphere
produce a net increase in the absorption of energy
by the Earth, leading to climate change such as a
warming of the Earth's surface—as has been
observed in recent decades (U.S. EPA, 2007).
The previous version of this report (U.S. EPA,
1993) provided global estimates of current and
future CFLj emissions from natural sources. Since
the release of that report, there has been significant
new research on a number of previously identified
sources (e.g., wetlands), newly identified potential
sources (e.g., vegetation), and the contribution of
N2O from natural sources. This report provides a
summary of the latest research, and presents global
estimates of current emissions of both CFUand N2O
from natural sources, as well as global estimates of
future changes in those emissions where data are
available.
What is climate change?
Climate change, as defined in the IPCC's
Fourth Assessment Report, refers to a change
in the state of the climate that can be identified
(e.g., by using statistical tests) by changes in
the mean and/or the variability of its properties,
and that persists for an extended period,
typically decades or longer (Solomon et al.,
2007b).
The Greenhouse Effect
Some soiar radiation
ts reflected by the
earth and the
atmosphere
Some of the Infrared radiation passes
through tie atmosphere, and some is
absorbed and re-emitted in all
dlnec&ons by greenhouse gas
molecules, the effect of this is to warm
Che earth's surface and the lower
atmosphere.
Most radiation Is absorbed
by the earth's surface
and warms it
Figure 1-1. The Greenhouse Effect
1.1 Importance of Methane and Nitrous Oxide as Greenhouse Gases
Long-lived greenhouse gases such as CF^and N2O are chemically stable and persist in the atmosphere
over time scales of a decade (in the case of CFL,) to centuries or longer (for N2O). For this reason,
emissions of these gases have a long-term influence on climate. These gases become well-mixed
throughout the atmosphere much faster than they are removed, and their global concentrations can be
accurately estimated from data at a few locations (Solomon et al., 2007a).
1.1.1 Methane
The chemical lifetime of CFL, from removal through reactions with the hydroxyl radical (OH) is estimated
at 9.6 years (Folland et al., 2001). Once emitted, however, CFL, actually remains in the atmosphere for
what is known as a "perturbation lifetime" of approximately 12 years before removal and ultimate
conversion to carbon dioxide (CO2), mainly by chemical oxidation in the lower atmosphere, or
troposphere (Solomon et al., 2007b). The longer perturbation lifetime of CFL, is primarily a result of
feedbacks between CFL,, OH, and its byproduct CO which is also removed by reactions with OH. Minor
removal processes include reaction with chlorine in the marine boundary layer, a soil sink, and
1-1
-------
Chapter 1. Introduction
stratospheric reactions. Increasing emissions of CIL, reduce the concentration of OH, a feedback that may
increase the atmospheric lifetime of CIL, (Solomon et al., 2007b). As OH also reacts with other short-
lived pollutants including volatile organic compounds (VOCs) and tropospheric ozone, changing
emissions of VOCs and ozone precursor species are also likely to affect CIL, lifetime (Wuebbles et al.,
2000).
CH4 is one of several greenhouse gases responsible for increased radiative forcing on the climate system
(see box on "Radiative forcing"). Over a period of 100 years, each molecule of CH4 has 25 times the
direct global warming potential of a molecule of CO2 (Solomon et al., 2007b). Ice core records indicate
that, over the last 650,000 years, atmospheric CIL, concentrations have varied from lows of about 400
parts per billion (ppb) during glacial periods to highs of about 700 ppb during interglacial periods.
Atmospheric CIL, has increased by about 1,000 ppb
since the beginning of the industrial era in the late
1700s, representing the fastest changes in this gas
over at least the last 80,000 years. In 2005, the
global average abundance of CH4 measured by
NOAA in both hemispheres was 1,774.62 ± 1.22
ppb, more than double its pre-industrial value
(Forster et al., 2007).
1.1.2 Nitrous Oxide
Like CH4, N2O is a long-lived greenhouse gas
responsible for increased radiative forcing on the
climate system. N2O has an atmospheric lifetime of
about 114 years, and over a 100-year period, each
molecule of N2O has a direct global warming
potential 298 times that of a single molecule of CO2
(Solomon et al., 2007b). Ice core data for N2O have
been reported extending back more than 2,000 years
from the present. These data show relatively little
change in mixing ratios over the first 1,800 years of
this record. Since the beginning of the industrial
revolution, however, N2O levels exhibit a relatively
rapid rise. Since 1998, atmospheric N2O levels have
steadily risen, reaching 319 ± 0.12 ppb in 2005 (Forster et al., 2007).
1.2 Sources of Methane and Nitrous Oxide
To understand the role of natural sources of CH4 and N2O, it helps to be familiar with a few key concepts
that describe the movement of gases into and out of the atmosphere. The atmosphere is considered a
reservoir, where each of these gases resides for a specific lifetime. Other reservoirs include oceans and
soils. Material can be transferred from one reservoir into another—a process described as a flux. Fluxes
into a reservoir such as the atmosphere are known as sources, while fluxes out are called sinks. Each
reservoir also has an overall budget, which represents the balance sheet of all sources and sinks.
This report examines the flux of CFL, and N2O into the atmospheric reservoir, with a particular focus on
natural sources. Because sources of CH4and N2O to the atmosphere are essentially processes that release
gases into the air, this report will use the term "emissions" to describe the actual movement of these gases
into the atmosphere.
Radiative forcing is a measure of how the
energy balance of the Earth-atmosphere
system is influenced when factors that affect
climate are altered. Radiative forcing is usually
quantified as the "rate of energy change per
unit area of the globe as measured at the top
of the atmosphere," and is expressed in units
of watts per square meter (W/m2). When
radiative forcing from a factor or group of
factors is evaluated as positive, the energy of
the Earth-atmosphere system will ultimately
increase, leading to a warming of the system.
In contrast, for a negative radiative forcing, the
energy will ultimately decrease, leading to a
cooling of the system. As of 2005, atmospheric
CH4 and N2O are the second- and third-largest
contributors to radiative forcing among
greenhouse gases, after CO2 (IPCC, 2007):
2
C02
CH4
N20
+1.66 W/rrf
+0.48 W/m2
+0.16 W/m2
1-2
-------
Chapter 1. Introduction
1.1.1 Methane
A large portion of CH4 emissions can be linked to the biological process of anaerobic decomposition, in
which bacteria break organic matter down in the absence of oxygen. Methanogens are a specialized group
of microbes that break down certain molecules, such as hydrogen (H2) and CO2, to produce CH4 through
the process of methanogenesis. Some of this QrU can be partly or completely oxidized by another group
of bacteria, called methanotrophs, while the remainder can ultimately enter the atmosphere. Microbial
production of methane occurs in a variety of settings, some deemed natural and others attributed to
anthropogenic activities. For example, methanogenesis can occur both in natural wetlands (a natural
source) and in human-influenced wetlands such as rice paddies (an anthropogenic source).
Methanogenesis also takes place in other water bodies, submerged sediments, landfills and waste
treatment facilities, and the digestive systems of certain animals—some domesticated (an anthropogenic
source) and some wild (a natural source).
Other natural sources of CH4 include wildfires and Methanogens are microorganisms (Archaea)
geologic processes. Other anthropogenic sources
include natural gas handling, biomass burning, and
fossil fuel combustion.
presence of oxygen. There are over 50
Scientists use several methods to determine the described species of methanogens.
relative contributions of different sources to the
atmosphere (see Section 1.3 for more detail on
methods). For example, the relative abundance of
the three principal isotopes of carbon (stable C,
stable 13C, and unstable or radioactive 14C) can
indicate which processes are producing CFi4. This is
because microbes producing CFi4 discriminate
against the heavier isotopic form, 13C, resulting in to form ^aldehyde, which is then
., . . , , . j • iv, T 4. 4. u 4.- incorporated into organic compounds.
that is depleted in C. In contrast, combustion
that produce methane as a metabolic
byproduct in anoxic conditions. Methanogens
are anaerobic; most are rapidly killed by the
Methanotrophs are bacteria that are able to
grow using methane as their only source of
carbon and energy. They can grow aerobically
oranaerobically and require single-carbon
compounds to survive. Under aerobic
conditions, they combine oxygen and methane
processes (wildfires, biomass burning) do not
discriminate. Fossil fuel sources of CFI4 (coal, oil, natural gas hydrates, geological sources) release
ancient carbon that lacks radioactive 14C, while more modern CFI4 contains 14C. Information on the
isotopic signature of the hydrogen in CH4 has also been used to characterize sources. In theory, weighting
individual source isotopic signatures by their fluxes and accounting for fractionation by sink processes
should yield the isotopic signature of atmospheric CFLj, but the signatures of many sources overlap and
signatures can change seasonally. The addition of isotopic data however, can be a powerful constraint on
possible source budgets.
By analyzing ice cores, scientists can compare present-day methane concentrations and sources with
historical (i.e., pre-industrial) records (Chappellaz et al., 2000). This has been helpful in estimating how
natural sources respond to changes in the environment before there was a significant anthropogenic input.
Changes in 14C isotopic ratios in ice cores can also help evaluate natural fossil CFLt sources such as
geological sources.
The total flux of CFLt into the atmosphere from all sources is currently 566 teragrams2 of CH4 per year
(Tg CtVyr), which is more than double the pre-industrial value (Solomon et al., 2007b). Based on
analysis of known CFLt sources, observed isotopic abundances, and budget modeling, the
Intergovernmental Panel on Climate Change (IPCC) estimates that the significant increase in atmospheric
CFLj levels observed over the last two centuries is primarily due to increasing anthropogenic emissions of
CFLt, which are currently approximately 1.5 to 2.5 times the magnitude of natural emissions (Forster et
al., 2007). In this report, we reach a similar estimate by adding together the best estimates for the main
21 teragram (Tg) equals 1 x 1012 grams or 1 million metric tons.
1-3
-------
Chapter 1. Introduction
natural source categories, which suggest that natural sources contribute 208 Tg CHVyrtothe atmosphere,
or 37 percent of total global emissions. The remainder comes from anthropogenic sources, the largest of
which are livestock, landfills, and natural gas systems (U.S. EPA, 2007).
1.1.2 Nitrous Oxide
Global production of N2O is attributed largely to microbial processes. Bacteria produce N2O through
nitrification and denitrification, which are key processes within the natural nitrogen cycle (Figure 1-2).
Nitrification is the main source of N2O under aerobic conditions, while denitrification dominates under
anoxic conditions.
In nitrification, bacteria oxidize nitrogen through a two-step aerobic process. Two groups of nitrifying
bacteria are responsible: those that oxidize ammonium (NFLj) to nitrite (NO2) and those that oxidize NO2
to nitrate (NO3). This multi-step process produces some N2O as a byproduct or as an alternate product of
ammonium oxidation. In denitrification, bacteria reduce oxidized inorganic forms of nitrogen. This
process may form N2O as an intermediate byproduct, or it may consume N2O. Therefore, the process of
denitrification can be either a source or a sink for this gas, depending on environmental conditions such as
oxygen levels, nitrogen levels, pH, and temperature (Sorai et al., 2007; Capone, 1991).
Microbial sources and sinks of N2O can be considered either anthropogenic or natural, depending on the
setting. Anthropogenic sources of N2O largely relate to agricultural soils, especially production of
nitrogen-fixing crops and forages, the use of synthetic and manure fertilizers, and manure deposition by
livestock. Natural sources reflect microbial processes in uncultivated soils, oceans and other aquatic
systems, and possibly wetlands.
Other anthropogenic sources that produce N2O include fossil fuel combustion (especially from mobile
sources), industrial processes, wastewater treatment and waste combustion, and biomass burning (U.S.
EPA, 2007).
Global estimates of N2O emissions from natural sources have utilized top-down approaches that rely on
atmospheric mixing ratios to estimate natural sources and sinks of N2O (Prather et al., 2001; Crutzen et
al., 2008). For example, Prather et al. (2001) estimated that the net pre-industrial flux of N2O was 10.2
TgN/yr).
The total flux of N2O into the atmosphere from all sources is currently estimated at 18.8 Tg per year as
nitrogen (Tg N/yr), which represents an increase since the pre-industrial era (Solomon et al., 2007b). This
increase primarily reflects human activities, particularly agriculture and associated land use change. The
IPCC estimates that about 60 percent of all N2O emissions come from natural sources, but individual
source estimates remain subject to significant uncertainties (Forster et al., 2007). In this report, we
estimate that natural sources contribute 12.1 Tg N/yr to the atmosphere, or 64 percent of the total of all
emissions worldwide.
1-4
-------
Chapter 1. Introduction
Nitrogen in «* -^ >
atmosphere (N j) fr
™IML ~
t.^t,;v> ,
Assimilation
itrifying
'tcria
Nitrogen-fixing
bacteria In
root nodules
at legumes
Decomposers
Ammo n ifi ca b o n
Nitrification
Nitrifying
bacteria
Ammonium
Nitrogen -fixing
soil bacteria
Figure 1-2. The Nitrogen Cycle
Nitrifying
bacteria
1.2 Overview of This Report
This report focuses on identifying and quantifying QrU and N2O emissions from the following natural
sources:
• Wetlands. Chapter 2 of this report discusses wetlands, which are ecosystems where saturation
with water is the dominant factor controlling soil development, as well as the species of plants
and animals that are present. Wetlands are transitional areas, at the interface between upland
environments and aquatic systems, and are thought to cover about 5 percent of the Earth's
surface. Because saturated soils create anoxic conditions, wetlands are an important natural
source of CH4, which is produced by anaerobic microbial processes. Wetlands are believed to be
negligible sources of N2O, though, and may at times act as minor sinks.
• Upland soils and riparian zones. Chapter 3 addresses upland and riparian soils. Upland soils are
well aerated and generally oxic. These dry soil conditions favor microbial processes which make
them a sink for CUt and a large source of N2O. Natural sources include upland soils associated
with forests and grasslands under natural vegetation, but not agricultural lands. Riparian zone
soils are often permanently wet and rich in organic matter, with saturated soil conditions and
microbially available carbon that contribute to higher rates of production of N2O than dry upland
soils.
• Oceans, estuaries, and rivers. Chapter 4 of this report considers a range of aquatic sources,
including the deep waters of the open ocean, shallower coastal waters (i.e., on the continental
shelves), freshwater rivers, and estuaries where fresh and salt waters mix. Microbial processes
that produce CH4 and N2O can occur both in sediments and in the water column. The oceans are
believed to be one of the two largest natural sources of N2O to the atmosphere, as well as a minor
natural source of CH/L. Continental shelf areas, estuaries, and rivers also contribute emissions of
CH4 and N2O. These water bodies typically have higher organic inputs and nutrient levels than the
1-5
-------
Chapter 1. Introduction
open ocean, and because they are relatively shallow, mixing is active and can transport gases
produced in the sediments into near-surface water where it can be released to the atmosphere.
However, there is some controversy as to how much of the N2O emissions associated with rivers
and estuaries should be counted as natural source emissions, because they are largely driven by
anthropogenic contributions of nitrogen to these water bodies.
• Permafrost. Permafrost is soil, sediment, or rock that is continuously frozen (temperature < 0°C)
for at least two consecutive years. As discussed in Chapter 5, Or^and N2O can be frozen within
permafrost, and thus permafrost represents a stock of Or^and N2O than can be released upon
thawing. Gas hydrates, which can co-occur with permafrost but which also are common in non-
permafrost regions, are discussed separately in Chapter 7, "Gas Hydrates."
• Lakes. Chapter 6 addresses lakes and ponds, which are naturally formed permanent water bodies
contained on a body of land. This report excludes impoundments and reservoirs, since gas
emissions from water bodies formed by dams and other engineering works are considered to be
anthropogenic. Lakes contribute to natural emissions of QL,, but appear to be a minor source of
N2O.
• Gas hydrates. As described in Chapter 7, gas hydrates are ice-like crystals formed between water
and gas molecules such as ClrU under high pressure and ambient temperatures. Large quantities of
CH4 are currently trapped in hydrate form, occurring mainly on continental shelves and to a lesser
extent below permafrost regions. Under current conditions, global CH4 emissions from gas
hydrates are small; however, the potential for significant Qr^ release from gas hydrates warrants
close examination.
• Terrestrial and marine geologic sources. Chapter 8 discusses natural seeps of CIL, and other
hydrocarbons from geologic sources deep within the earth's crust. Geologic CH4 emissions can be
produced via what is known as "macroseepage," which includes relatively large localized
emissions from identified geologic features and events such as mud volcanoes and localized
vents. Emissions also can result from "microseepage" in volcanically active or other geothermal
regions. Sources include both marine (underwater) and terrestrial (land-based) faults.
• Wildfires. Wildfires are fires in forests, grasslands, savannas, and shrublands. This report
includes fires ignited by lightning strikes or started by humans, but not controlled burns. As
Chapter 9 explains, wildfires release a number of greenhouse gases, particulates, and other air
pollutants as they burn. Incomplete combustion or smoldering of biomass, consisting of both
living and dead organic matter, is the primary source of emissions of CH^ from wildfires.
Wildfires also produce N2O, with the amount produced depending on the nitrogen content of the
biomass burned.
• Vegetation. Chapter 10 addresses direct QL, emissions from vegetation, including upland
tropical forests and other unflooded ecosystems. It has not been previously considered a potential
source of ClrL,, because drier soils act as an oxidative sink for CIL,. However, recent findings
suggest that such ecosystems may be a significant unrecognized source of CH4 although the
existence and magnitude of this source remains to be confirmed.
• Terrestrial arthropods and wild animals. Chapter 11 of this report discusses the production and
emission of Cl^by termites and other terrestrial arthropods as a result of microbial degradation of
ingested organic matter, as well as CH4 emissions caused by enteric fermentation in wild ruminant
animals such as bison, deer, elk, mountain goats, and sheep, and also in some smaller rodent
species.
It is important to note that there can be both ambiguity in source definitions (e.g., are small ponds in
natural wetlands a wetland source or a lake source?), and a mixture of sources (e.g., co-location of
1-6
-------
Chapter 1. Introduction
permafrost, gas hydrates, and geologic sources). Field studies may include flux measurements that have
contributions from more than one of the sources listed here. This is discussed further in the chapters on
each source.
It is also important to keep in mind that the earth is a system of interacting components, and change often
affects many components as well as their interactions. This report is organized into chapters covering
natural sources by type (e.g., wetlands, lakes, oceans, gas hydrates, etc.). However, the earth is a mosaic
of these different source types, the boundaries between these source types are sometimes inexact (e.g.,
between a wetland and the emergent vegetation of a lake margin), and system changes that affect one
source can also affect one or more other sources.
The issue of methane in permafrost regions exemplifies this interconnectedness. The projected thawing of
permafrost with climate warming may contribute to increased natural source methane emissions to the
atmosphere. However, this is a complex system response. There is not a lot of methane frozen into
permafrost (unless it is a gas hydrate formation), so permafrost thaw will not release much methane
directly. The methane it does release has a reasonable probability of being oxidized as it diffuses through
1 to 100 meters of thawed soil before reaching the atmosphere (see Chapter 5, "Permafrost"). Therefore,
the report concludes that permafrost thawing is not likely to be a strong methane source. However, if gas
hydrates are associated with permafrost, then thawing permafrost and destabilization of these hydrates
may co-occur (though they are not exactly the same thing), releasing methane, potentially in large
quantities (see Chapter 7, "Gas Hydrates"). Another consideration is changes in wetland vegetation and
moisture status associated with permafrost thawing; this would be an issue for some, but not all,
permafrost landscapes. If the landscape gets wetter and the vegetation composition becomes more
dominated by sedges, this could lead to increased methane emissions from a wet landscape - at least for
years to decades; however, if the landscape gets drier (or stays relatively dry), then methane emissions
would probably stay low (see Chapter 2, "Wetlands"). Finally, permafrost thaw can be associated with
thermokarst erosion, which can form (or drain) lakes; these lakes can also be methane sources (See
Chapter 6, "Lakes"). Each chapter of this report specifies the natural sources addressed by that chapter.
1.3 Summary of Methods
This report builds on information provided in EPA's 1993 report (U.S. EPA, 1993) as well as the
Intergovernmental Panel on Climate Change's Fourth Assessment Report, or AR4 (Solomon et al.,
2007b). The 1993 EPA report focused primarily on natural wetlands and other fossil sources (including
gas hydrates and permafrost) as contributors to the global CH4 budget, but briefly acknowledged other
sources such as termites, ocean and freshwater systems, and non-wetland soils. The 1993 report did not
include any estimates of these other CFL sources, however, nor did it address natural sources of N2O.
Although several of the major greenhouse gases occur naturally, the AR4 attributes increases in their
atmospheric concentrations over the last 250 years largely to human activities. Therefore, the majority of
research discussed in the AR4 focused on anthropogenic sources, and the AR4 includes only a limited
assessment of natural source emissions.
Scientists use a variety of approaches to characterize emissions. These approaches generally fall into two
categories: "bottom-up" calculations and "top-down" or inverse modeling. Bottom-up estimates are based
on an estimate of activity (e.g., population contributing to the source) combined with an emission factor
reflecting the amount of emission per unit of activity. The estimate also can include direct emission
measurements from individual sources or other variables that contribute to the calculation of an emission
factor (e.g., temperature, geographic location) (Olivier, 2002). Although conceptually simple, bottom-up
methods contain numerous uncertainties. First, point measurements taken at a limited number of locations
and times must be assumed to be representative of global emissions. These point measurements must be
extrapolated to larger scales which can introduce significant error and most extrapolations are performed
1-7
-------
Chapter 1. Introduction
using inventory-type data (e.g., population size, burned area, or vegetation type) which tends to be a
"snap-shot" of conditions. As data coverage improves, however, researchers are focusing more attention
on the need to understand and predict temporal and spatial variations in those emissions. Such variations
are particularly important for biogenic sources such as wetlands, whose emissions can vary by several
orders of magnitude depending on the location, time of day, season, or year. Although these uncertainties
continue to produce a range of values in the emission estimates for individual sources, bottom-up ranges
also can be constrained by top-down analyses.
Inverse modeling methods use atmospheric concentration measurements, atmospheric models, and
statistical tools to estimate emissions from individual sources. This method requires initial (apriori)
emissions information distributed overtime and space (Olivier, 2002). Such analyses use observed spatial
and temporal changes in atmospheric mixing ratios and isotopic abundances as input to mass-balance or
more sophisticated three-dimensional chemical transport models of the atmosphere. In the case of ClrU,
data on abundances and isotopic ratios are usually taken from ground-based observing stations and ice
core samples, although satellite observations are also being incorporated into more recent analyses (e.g.,
see Frankenburg et al., 2006). The isotopic composition of atmospheric CFLt provides particularly
valuable information, as it reflects the relative strength of bacterial versus nonbacterial and modern versus
fossil sources of CFLj, as well as differentiating among various formation pathways (Cicerone and
Oremland, 1988; Conny and Currie, 1996; Whiticar, 2000).
Mass-balance modeling approaches use spatial and regional variations in isotopic composition as well as
hemispheric and global averages as input to box models. Atmospheric observations are compared with the
flux-weighted composition of total emissions plus the fractionation effects of sinks, taking into account
the effects of atmospheric transport. In this way, the magnitudes of individual sources as well as entire
budgets have been estimated (e.g., Khalil and Rasmussen, 1983; Stevens and Engelkemeir, 1988;
Cicerone and Oremland, 1988).
Three-dimensional chemical transport models in combination with knowledge of sources and sinks
perform calculations that match modeled estimates of concentrations and isotopic composition with
observed abundances. Because the model should ultimately agree with observations, the initial conditions
can then be evaluated and changes made if needed. Uncertainties in sources and sinks can be reduced
through this process and the relationships between emissions, atmospheric chemistry, and the climate
system better quantified. Numerous studies (e.g., Fung et al., 1991; Hein et al., 1997) have used this
approach, while others have developed new inversion methods (Brown, 1995; Kandlikar and McRae,
1995; Kandlikar, 1997; Houweling et al., 1999). These analyses, combined with multi-box and source-
specific studies, produce comprehensive budgets of sources and sinks.
Bottom-up and top-down methods each have advantages and disadvantages. For example, bottom-up
methods require accurate activity data and emission factors, while top-down methods require the
development of models which may have many unknowns. Top-down methods can provide
comprehensive coverage, but cannot easily attribute emissions to specific activities in specific regions.
Inventories of anthropogenic greenhouse gas emissions typically use bottom-up methods.
1.4 References
Brown, M. 1995. The singular value decomposition method applied to the deduction of the emissions and
the isotopic composition of atmospheric methane. J. Geophys. Res. 100(11): 425-446.
Capone, D.G. 1991. Aspects of the marine nitrogen cycle with relevance to the dynamics of nitrous and
nitric oxide. In: J.E. Roders and W.E. Whitman. (eds.).Microbial Production and Consumption of
Greenhouse Gases. Am. Soc. Microbiol. Washington, DC. pp. 255-275.
1-8
-------
Chapter 1. Introduction
Chappellaz, J., D. Raynaud, T. Blunier, and B. Stauffer. 2000. The ice core record of atmospheric
methane. In: M.A.K. Khalil (ed.). Atmospheric Methane: Its Role in the Global Environment. Berlin:
Springer-Verlag. pp. 9-24.
Cicerone, R.J., and R.S. Oremland. 1988. Biogeochemical aspects of atmospheric methane. Global
Biogeochemical Cycles 2(4): 299-327.
Conny, J.M., and L.A. Currie. 1996. The isotopic characterization of methane, non-methane hydrocarbons
and formaldehyde in the troposphere. Atmospheric Environment 30(4): 621-638.
Crutzen, P.J., A.R. Mosier, K.A. Smith, and W. Winiwater. 2008. N2O release from agro-biofueled
production negates global warming reduction by replacing fossil fuels. Atmospheric Chemistry and
Physics 8: 389-395.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohman, S. Ramachandran, P.L. da Silva Bias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report
of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press.
Folland, C.K., T.R Karl, J.R. Christy, RA. Clarke, G.V. Gruza, J. Jouzel, M.E. Mann, J. Oerlemans, M.J.
Salinger and S.-W. Wang, 2001: Observed Climate Variability and Change. In: J.T. Houghton, Y.
Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A. Johnson (eds.)
Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY:
Cambridge University Press.
Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Berts, D.W. Fahey, J. Haywood, J. Lean, D.C.
Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz, and R. Van Dorland. 2007. Changes in
Atmospheric Constituents and in Radiative Forcing. In: S. Solomon, D. Qin, M. Manning, Z. Chen, M.
Marquis, K.B. Averyt, M.Tignor, and H.L. Miller (eds.). Climate Change 2007: The Physical Science
Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge University Press.
Frankenberg, C., et al. 2006. Satellite chartography of atmospheric methane from SCIAMACHY on board
EMVISAT: Analysis of the years 2003 and 2004. J. Geophys. Res. Ill, doi:10.1029/2005JD006235.
Fung, I., J. John, J. Lerner, E. Matthews, M. Prather, L.P. Steele, and P.J. Fraser, 1991: Three-
dimensional model synthesis of the global methane cycle. J. Geophys. Res. 96: 13033-13065.
Hein, R., P. Crutzen, and M. Heimann, 1997. An inverse modeling approach to investigate the global
atmospheric methane cycle. Glob. Biogeochem.Cycles 11: 43-76.
Houweling, S. 1999. Global Modeling of Atmospheric Methane Sources and Sinks. Netherlands:
Universal Press.
Houweling, S., T. Kaminski, F. Dentener, J. Lelieveld, M. Heimann. 1999. Inverse modeling of methane
sources and sinks using the adjoint of a global transport model. Journal of Geophysical Research—
Atmospheres 104(D21): 26137-26160.
Kaandlikar, M. 1997. Bayesian inversion for reconciling uncertainties in global mass balances. Tellus B
49(2): 123-135.
Kandlikar, M., and G.J. McRae. 1995. Inversion of the global methane cycle using chance constrained
programming: Methodology and results. Chemosphere 30(6): 1151-1170.
1-9
-------
Chapter 1. Introduction
Khalil, M.A.K. (ed.). 2000. Atmospheric Methane: Its Role in the Global Environment. Springer.
Khalil, M.A.K., and R.A. Rasmussen. 1983. Sources, sinks, and seasonal cycles of atmospheric methane.
Journal of Geophysical Research 88: 5131-5144.
NOAA (National Oceanic and Atmospheric Administration). 2007. Frequently asked questions.
http://www.esrl.noaa.gov/gmd/infodata/faqcat-3.html. Date accessed: December 2007.
Olivier, J.G.J. 2002. On the Quality of Global Emission Inventories: Approaches, Methodologies, Input
Data, and Uncertainties. Amersfoort, Netherlands: Wilco BV.
Prather, M., D. Ehhalt, F. Dentener, R.G. Derwent, E. Dlugokencky, E. Holland, I.S.A. Isaksen, J.
Katima, V. Kirchhoff, P. Matson, P.M. Midgley, and M. Wang. Chapter 4: Atmospheric chemistry and
greenhouse gases. In: J.T. Houghton et al. (eds.). 2001. Climate Change 2001: The Scientific Basis.
Cambridge University Press, pp. 239-287.
Schaefer, S., M.J. Whiticar, E.J. Brook, V.V. Petrenko, D.F. Ferretti, and J.P. Severinghaus. 2006. Ice
record of 513C for atmospheric CIL, across the younger Dryas-Preboreal Transition. Science 313: 1109-
1112.
Solomon, S., D. Qin, M. Manning, R.B. Alley, T. Berntsen, N.L. Bindoff, Z. Chen, A. Chidthaisong, J.M.
Gregory, G.C. Hegerl, M. Hermann, B. Hewitson, B.J. Hoskins, F. Joos, J. Jouzel, V. Kattsov, U.
Lohmann, T. Matsuno, M. Molina, N. Nicholls, J. Overpeck, G. Raga, V. Ramaswamy, J. Ren, M.
Rusticucci, R Somerville, T.F. Stacker, P. Whetton, RA. Wood, and D. Wratt. 2007a. Technical
summary. In: S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and
H.L. Miller (eds.). Climate Change 2007: The Physical Science Basis. Contribution of Working Group
I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK,
and New York, NY: Cambridge University Press.
Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.).
2007b. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and
New York, NY: Cambridge University Press.
Sorai, M., N. Yoshida, and M. Ishikawa. 2007. Biogeochemical simulation of nitrous oxide cycle based
on the major nitrogen processes. J. Geophys. Res. 112: G01006, 10.029/2005JG000109.
U.S. EPA (United Status Environmental Protection Agency). 2007. Inventory of U.S. Greenhouse Gas
Emissions and Sinks: 1990-2005. EPA-430-R-07-002.
U.S. EPA. 1993. Current and Future Methane Emissions From Natural Sources. EPA-430-R-93-011.
Whiticar, M.J., 2000. Can stable isotopes and global budgets be used to constrain atmospheric methane
budgets? In: M.A.K. Khalil (ed.). Atmospheric Methane: Its Role in the Global Environment. Berlin:
Springer-Verlag. pp. 63-85.
Whiticar, M., and H. Schaefer. 2007. Constraining past global tropospheric methane budgets with carbon
and hydrogen isotope ratios in ice. Phil. Trans. R. Soc. A 365: 1793-1828. doi:10.1098/rsta.2007.2048.
Wuebbles, D., K. Hayhoe, and R. Kotamarthi. 2000. Methane in the Global Environment. In: M. Khalil,
Ed, Atmospheric Methane: Its Role in the Global Environment. Springer-Verlag, Berlin, 351 pp.
1-10
-------
Chapter 2. Wetlands
Wetlands are ecosystems where saturation with water is the dominant factor controlling soil development
and the species of plants and animals that are present (Cowardin et al., 1979). Because water saturation is
a defining characteristic of wetlands, these areas are an important natural source of CH4, which is
produced by bacteria requiring oxygen-free conditions. Wetlands are believed to be a negligible source of
N2O and may at times act as a minor sink.
Researchers have known for some time that wetlands produce QrU (Koyama, 1963; Swain, 1973).
However, interest in measuring emissions from wetlands accelerated in the mid- to late 1970s as the
atmospheric importance of these releases became clear (Ehhalt, 1974; Baker-Blocker et al., 1977; Harriss
and Sebacher, 1981; Harriss et al., 1982). A large database has accumulated over the last 30 years. These
data were summarized for the period before 1993 in the earlier version of this report (U.S. EPA, 1993).
The wetlands flux review portion of that publication was published by Bartlett and Harriss (1993).
The measurement of wetland ClrU flux remains an area of very active research and the flux database from
diverse wetland types continues to grow. The focus of work, however, has moved beyond the initial phase
of assessing source magnitude as understanding has increased. Most of the current work examines
environmental controls and processes. Although the most recent IPCC report (AR4) does not calculate
wetland QrU fluxes, work cited in the report estimates emissions ranging from 100 to 23 1 Tg CHVyr. If
the total CH4 flux from all sources is roughly 566 Tg CHVyr (see Table ES-1), then wetlands may
contribute 18 to 41 percent.
Measurements of the flux of N2O from wetlands are much sparser. Initial work demonstrated that
emissions are relatively low; efforts, therefore, have focused on ecosystems that are more important
globally. Estimates of the wetlands flux of N2O are not included in the AR4.
2.1 Description of Emission Source
Wetlands are transitional areas at the interface between upland, terrestrial environments and aquatic
systems. They are distinctly different from both terrestrial and aquatic environments, but depend on them
both. The global area of wetlands has been estimated at 5.2 to 5.86 x 106 km2 and they are thought to
cover about 5 percent of the Earth's surface (Matthews and Fung, 1987; Prigent et al., 2007).
Although topography (the lay of the land and its elevation) is not the defining characteristic of wetlands, it
is an important factor contributing to their presence, since it can control where water goes and how long it
remains. The water itself may come from precipitation, subsurface flow from ground water, or surface
flow from a surrounding watershed or water body such as an ocean, river, or lake. Characteristics of the
water from these different types of sources can determine wetland type (for example, saline or fresh, low
nutrient or high, active water flow or still and stagnant).
The presence of water in wetlands creates significant physiological problems for both plants and animals,
and adaptations to it have resulted in distinctive communities. The line between wet and dry environments
is often a gradual one, and water saturation may occur on a constant, seasonal, daily, or even sporadic
basis and still be the dominant factor defining a landscape.
Wetlands are diverse and can be classified in many different ways. The U.S. National Wetlands Inventory
uses a combination of water depth (deepwater vs. shallow water), landscape location (marine, estuarine,
riverine, and lacustrine or bordering lakes), and vegetation to classify habitats. The colloquial terms
"marsh," "swamp," "fen," "bog," "muskeg," and "pocosin" identify environments largely on the basis of
vegetation. Wetlands are found from the tundra to the tropics and on every continent except Antarctica. In
the tropics, where rainfall is strongly seasonal, many wetlands (such as varzea or flooded forests) also
2-1
-------
Chapter 2. Wetlands
occur seasonally. At high latitudes, although the amount of precipitation is often relatively low, frozen
soils inhibit drainage over large regions and create a landscape of abundant wetlands and shallow
standing water.
Vegetation characteristics and water dynamics, used to classify wetlands, are also useful for
characterizing ClrU and N2O emissions, because they integrate many environmental factors that affect the
production of the gases. Major controls on soil conditions include organic inputs (the quantity and quality
of vegetative material) and the balance between aerobic and anaerobic soil environments, which is largely
controlled by the presence or absence of water.
Although some wetlands in the United States have been constructed for water processing, environmental,
or management reasons, they will not be discussed here due to their human origin and ecological
differences from natural systems. Riparian zones, usually defined as the interface between uplands and
flowing water, occur in many forms (grassland, forest, unvegetated) and may or may not be wetlands. If
they are vegetated by plant communities characteristic of water-saturated soils, they are included in this
report's definition of wetlands and are addressed in this chapter. Other types of riparian zones are
addressed in Chapter 3, "Upland Soils and Riparian Areas."
2.2 Factors That Influence Emissions
Both CH4 and N2O are produced by bacteria in wetland soils and are therefore affected by a suite of
environmental variables. Emission of both gases is a function of the balance between their production and
consumption, which are carried out by different functional groups of bacteria. Environmental controls
may affect these groups differently, resulting in non-linear responses to small changes. For example,
bacteria producing CFLt (methanogens) have been found to be more sensitive to temperature than those
oxidizing CUt (methylotrophs and methanotrophs). This means that a change in temperature may change
the balance between production and consumption, and thus change emissions. Flux to the atmosphere is
commonly a small residual of the larger amounts of the gases that are produced and consumed in the soil.
This means that the potential is present for large changes in flux in response to what may appear to be
minor environmental change.
Whalen (2005) recently summarized wetland biogeochemical controls on CFI4 emissions and grouped
these into process-level and ecosystem-level factors.
2.2.1 Process-Level Controls
Process-level controls include organic material quality and quantity, temperature, and pH. All of these
affect how rapidly and well bacteria grow. Bacterial responses to controls are frequently non-linear, as
demonstrated by the classic logarithmic Qi0 temperature response model (the reaction rate change for a
10°C temperature change).
The presence of water inhibits the diffusion of O2 (here water is considered an ecosystem-level control;
see Section 2.2.2). Because methanogens require oxygen-free conditions to grow, soils saturated with
water provide this critical condition for methanogenesis to occur. Once anaerobic conditions have been
established, however, organic matter supply and temperature have been shown to be primary controls
(Valentine et al., 1994; Coles and Yavitt, 2002).
Since N2O is an intermediate product of nitrification and denitrification (see Chapter 1), controls on its
atmospheric release are also complex. As for CFLb interactions between different functional groups mean
that changes in environmental parameters have multiple effects. Rates of nitrogen cycling overall are
impacted by soil fertility and texture, available N (NO3, NO2, or NFL^, and oxygen, needed for respiration
(Groffman, 1991). Davidson (1991) has proposed a conceptual model of the controls on N2O release from
soils that has been widely used. Known as the "hole-in-the-pipe," this model uses the analog of water
2-2
-------
Chapter 2. Wetlands
flowing through a leaky pipe for the flow of N in soils during organic matter decomposition. The sizes of
the holes in the "pipe" are analogous to the relative rates of nitrification and denitrification. In fertile
soils, flow through the pipe is large, as are the "leaks." The converse is true in infertile soils, and neither
gas is produced in large amounts. In dry soils, where O2 is present, the nitrification "leak" is greater and
NO, which is more oxidized than N2O and N2, is the dominant gas. In wetter soils, with less soil O2,
denitrification is dominant and more N2O is produced. In very wet soils, denitrification also dominates,
but proceeds all the way to the most reduced end product, N2. Testing this model against observations
from a wide range of sites suggests that it can help explain both the amount of N gases released as well as
their relative proportions (Davidson and Verchot, 2000).
Reported Qi0 temperature coefficients for methanogenesis (see Whalen, 2005, and cited references) are
relatively high, indicating a strong response to temperature change by methanogens. Bacteria oxidizing
CH4 appear to have a somewhat weaker response. Although laboratory studies have suggested that
optimal pHs for many methanogens and methanotrophs lie in the neutral range, many wetland soils are
acidic and both functional groups appear to tolerate what may be sub-optimal conditions. Moore and
Roulet (1995) suggest that pH is a secondary control on production and oxidation o
CH4 oxidation by bacteria occurs in a wide variety of soils and environments, including those of wetlands.
Bacteria can consume CF^ from both the soil and from the air. The uptake of atmospheric CF^ is largely
determined by its rate of diffusion into the soil and, less critically, temperature. Controls on diffusion
rates in soils are physical — primarily moisture content and soil texture. Rates of CH4 uptake vary over a
much more limited range than do CF^ emissions, which is consistent with the greater importance of
physical controls. The uptake of atmospheric CH4 largely takes place in the surface of we 11 -drained soils
such as forests rather than wetlands. The role of well-drained soils as a sink for atmospheric CFU is
discussed in more detail in Chapter 3. Subsurface CFU oxidation, however, is an important process in
wetlands, where it consumes a significant portion of the CF^ produced. Understanding controls on
emissions therefore requires assessing both CH4 production and CH4 oxidation.
Recent work has suggested that nutrient inputs (sulfur and nitrogen) may affect CH4 emissions. These
studies have focused on the atmospheric deposition of sulfate (SO4) by acid rain and the wet and dry
deposition of N. Because SO4 may also be used by bacteria to decompose organic material, its addition
can alter carbon flow pathways, reducing the energy flow through methanogenesis and therefore reducing
emissions. For example, this competitive interaction and its effect on emissions has been observed in salt
marshes, where natural SO4 inputs from seawater decrease fluxes (Bartlett et al., 1987). Field experiments
suggest that reductions due to atmospheric deposition may be substantial, up to 40 percent of controls
(Dise and Verry, 2001; Gauci et al., 2002). Examining a gradient in SO4 deposition, Vile et al. (2003)
found that CFU production decreased as deposition rates increased. Gauci et al. (2005) followed the
response of emissions after the addition of a pulse of SO4. They report that emissions were suppressed for
at least 5 years and estimate that as many as 10 years may be required for recovery to previous levels.
Although it might be expected that chronic N additions to wetlands would also decrease CH4 emissions
since NO3 is also used during bacterial organic decomposition, field simulations suggest that effects are
relatively small (Dise and Verry, 2001). Zhang et al. (2007), however, report significant increases in CFU
flux over a range of N additions. They suggest that this increase is most likely an indirect effect of the N-
induced increase in plant biomass. Effects on N2O emissions are unknown, but would not be expected to
be large since water saturation (and the low O2 levels as well as slowed diffusion it creates) appears to be
the major control on emissions rather than N inputs. Sensitivity to N inputs is likely to depend upon a
wetland 's nutrient status.
The complex and non-linear response of bacteria to process-level controls means that at larger scales
where these variables can be highly dynamic, fluxes can be highly variable both in time and space. This
has created significant difficulties in attempting to derive large-scale flux estimates.
2-3
-------
Chapter 2. Wetlands
2.2.2 Ecosystem-Level Controls
At the ecosystem level, the most important control on emissions of CFL and N2O is the presence of water
and the position of the water table. For N2O, this single variable appears to be the dominant control on
whether a site is a small source or a sink, since wetland systems appear to be poised at near-equilibrium
with respect to the atmosphere. The process of denitrification is thought to dominate in wet soils (more
than 60 percent water-filled pore spaces), producing a greater fraction of N2O than NO (Davidson, 1991).
However, water-saturated soils (more than 80 percent water-filled pore spaces) slow diffusion, which
enhances N2O consumption, resulting in little release to the atmosphere. For saturated wetland soils,
therefore, N2 becomes the dominant gas released (Davidson, 1991). Under moderately wet conditions,
N2O may be produced by denitrifiers and some fraction may diffuse to the atmosphere before
consumption occurs. It may also remain in the soil and be released if soils dry (Davidson et al., 1993).
This may happen seasonally, as in seasonal wetlands and lakes in the tropics, or as a consequence of
drought. Wetting and drying cycles have been shown to enhance N2O emissions and bacterial responses
to wetting appear to be rapid (Davidson et al., 1993; Venterink et al., 2002). Nitrogen dynamics in
seasonal wetlands in the tropics undergo annual transitions between nitrification-domination and
denitrification-domination (Koschorreck and Darwich, 2003). Koschorreck (2005) reports high rates of
N2O flux during the transition between flooded, anaerobic soils and dry, aerobic conditions.
For CFI4, the presence of water serves as the primary and required condition in wetlands. There is
essentially no flux and there may even be uptake from the atmosphere in its absence. In the Great Dismal
Swamp, for example, a lack of rainfall transformed much of the system from a strong source to a sink as
the water level fell (Harriss et al., 1982). If the system is wet, then other variables such as organic quality
and quantity come into play. Summarizing work over a 12-year period in northern wetlands with high
water tables, Christensen et al. (2003) found that temperature and substrate availability combined to
explain virtually all of the variation in annual emissions. Wetlands located in riparian zones receive water,
C, and nutrients from both adjacent water bodies and from upland downslope surface or subsurface flows
(Itoh et al., 2007). Under high flow conditions, the rapid resupply of oxygenated water may suppress
methanogenesis even through soils are water-saturated.
The climatic setting as well as the plant species present in a wetland exerts important controls on
emissions at larger scales. Plants have been shown to affect CFL emissions both directly and indirectly.
Species differ in the amount of biomass they produce as well as how easily that biomass may be
decomposed. In addition, some wetland plant species can move gases produced in sediments such as CFL
through stem and leaf gas spaces and release them to the atmosphere (Sebacher et al., 1985). There is
evidence that N2O may also be transported and released to the atmosphere by some plants (Mosier et al.,
1990; Rusch and Rennenberg, 1998; Chang et al., 1998; Yan et al., 2000; Kreutzweiser et al., 2003).
Net ecosystem CC>2 exchange (NEE, a measure of plant primary productivity) and CFL exchange have
been found to be correlated in some wetlands (Whiting and Chanton, 1993; Christensen et al., 2000) and
this relationship has been used as a basis for modeling fluxes at large scales (Walter and Heimann, 2000).
Labeling studies, which mark organic material with low levels of radiation, indicate that much of the CFL
released to the atmosphere has come from recently produced organic material, rather than from older soil
organic material (King and Reeburgh, 2002; Megonigal et al., 1999). In fact, King et al. (2002) estimate
that more than 75 percent of average CFL flux from tundra wetlands is derived from recently fixed
carbon.
Recently, von Fischer and Hedin (2007) used a radioactive tracer technique to explore hypotheses about
controls on CFL. fluxes at the landscape scale. Their study evaluated the relative importance of three
mechanisms in controlling emissions: (1) fluxes were controlled by bacterial CFL oxidation; (2) fluxes
were controlled by substrate availability, the amount of C mineralization; and (3) fluxes were determined
by the relative C flow through methanogenic vs. non-methanogenic decomposition pathways
(nitrification/denitrification, aerobic decomposition, SC>4 reduction). In contrast to the work indicating a
2-4
-------
Chapter 2. Wetlands
Redox (short for reduction/oxidation)
describes the relative oxidation status of a
soil. Soils with a low redox status have little
available oxygen, which limits the types of
reactions that can take place.
strong link between emissions and C supply (Whiting and Chanton, 1993), they found that rates of C
mineralization did not explain variations in fluxes at these large scales. Rather, they found that
surprisingly small changes in C flow between mineralization pathways resulted in large differences in
CFLj production and subsequent flux. On average, if more than 0.04 percent of total C mineralization was
through methanogenic pathways, soils were net sources of CFU. This result indicates that CFU production
is highly sensitive to soil redox status and the development of anaerobic microsites where energy flow can
be diverted to methanogenesis. Such microenvironments could develop due to differences in C supply,
nutrients, or moisture, von Fischer and Hedin suggest
that while ecosystem C supply may constrain CFU
production from a mass-balance perspective, it is the
fine-scale biophysical factors that create the high
variability that has been observed in emissions within
and across environments and which must be included
in models.
2.2.3 Methane Emission Pathways
Emissions to the atmosphere from wetlands occur through a variety of mechanisms (Figure 2-1): by
diffusion across the soil or water interface, by ebullition (bubbling) when concentrations are high and
exceed saturation levels, and by plant-mediated transport, as noted above. The relative importance of
these pathways can vary between habitats as well as over the course of a season, as production and soil
concentrations change seasonally and plants mature and die. In a number of wetland types, ebullition has
been found to release a significant fraction of total emissions (Bartlett et al., 1988; Marani and Avala,
2007; Happell and Chanton, 1993). Assessing the importance of ebullition can be difficult, however,
because it is episodic and may release a considerable volume of gas in a very brief period. If sampling
was not occurring at this time, emissions may be significantly underestimated.
Wetland plants can also serve as important transport pathways to the atmosphere. By moving rapidly
through the air spaces in plant stems and leaves, CFU can bypass the near-surface unsaturated layer where
most CFLj oxidation occurs and thus a greater fraction of production may be released to the atmosphere.
Transport through plants can be by diffusion, following the air pressure gradient produced by the plant
during respiratory consumption (Garnet et al., 2005), or by bulk flow, pressurized ventilation. Although
not all plant species transport CFLj (Sebacher et al., 1985; Chanton and Dacey, 1991), transport through
plants may make a major contribution to total emissions. For example, Whiting and Chanton (1992)
estimated that 90 percent of the CFU flux from a sub-arctic fen was plant-mediated. Oquist and Svensson
(2002) found that plants functioned as important controls on emissions, but species acted in different
ways in the wetlands they investigated.
2-5
-------
Chapter 2. Wetlands
t
170.3Tgyr
02 CH4
CH4
Methanogenesis
Figure 2-1. Conceptual model of ChU cycling in wetland and upland environments, showing plant-mediated transport,
bubbling, and diffusion. Adapted from Whalen, 2005.
The importance of plant-mediated N2O fluxes is difficult to evaluate, since only a few wetland species
have been investigated. Rice plants, particularly under flooded conditions, appear to be an important
pathway for N2O release from paddy soils (Yan et al., 2000). Emissions of N2O have been reported from
both the prop roots and pneumatophores (specialized root structures used for air transfer) of several
mangrove species (Kreutzwieser et al., 2003; Krithika et al., 2008). Black alder, a common European tree
in wetlands, develops internal air spaces in response to flooding, as do a number of other trees found in
wetlands. These air spaces were found to enhance the transport of both CH4 and N2O from soils (Rusch
and Rennenberg, 1998) and the authors suggest that this may be a common phenomenon in wetland tree
species.
The variety of transport mechanisms to the atmosphere, their high spatial and temporal variability, and
their dependence upon small-scale environmental factors complicate efforts to derive larger-scale flux
models. They also add significant uncertainty to flux estimates and can be a large source of error if flux
measurement techniques create artifacts.
2-6
-------
Chapter 2. Wetlands
2.3 Current Global Methane and
Nitrous Oxide Emissions From
Wetlands
This section discusses the techniques used to make
global flux estimates from wetlands and then review
current estimates. The majority of this discussion
focuses on QrU fluxes because of their global
importance and because few large-scale estimates of
N2O from wetlands have been made. However, the
techniques used for global flux estimation are fully
applicable to N2O.
An empirical model just uses observations to
make estimates. For example, summer
emissions might be found to contribute 90
percent of annual emissions. It implies no
deeper understanding.
A process model uses an understanding of
the underlying processes that control flux to
make estimates. For example, fluxes might
be modeled using soil temperature, as a
result of the control of bacterial growth by
temperature.
2.3.1 Techniques for Making Global Estimates
This section discusses three techniques for making global estimates of CH4 and N2O emissions from
wetlands: surface flux extrapolation, process modeling, and inverse modeling. Each approach has
strengths and weaknesses.
2.3.1.1 Surface Flux Extrapolation
Surface flux extrapolation uses actual emission measurements to calculate a global estimate. Although the
scale of possible measurement approaches ranges from less than 1 m2 (using chambers) to perhaps 100
km2 (using eddy correlation towers), all estimates of surface flux must be extended in time and space
beyond actual observations. The implicit assumption is that the measurements are representative of the
true flux and that measurement error is small relative to true variability. For wetlands, however,
extrapolating flux is complex because of non-linearities in the processes affecting emissions and high
surface heterogeneity on a range of scales. Wetland vegetation can serve as an indicator or integrating
variable for environmental parameters that are more difficult to measure over large scales (i.e., nutrient
status, frequency, depth, and length of inundation, salinity, organic content). Therefore, it has often been
used as a way to organize, characterize, and extrapolate emissions. Vegetation is also a parameter that can
be remotely sensed.
Scaling up wetland flux measurements by using vegetation and moisture or water level relies on inventory
databases of these variables. These databases may be sources of significant error to the resulting flux
estimate (Frey and Smith, 2007). They are also commonly static, lacking the dynamics required for
estimating biological processes. The increasing sophistication of satellite sensors of land cover and
inundation offers significant improvement in temporal and spatial resolution of these variables. Melack et
al. (2004) used seasonal synthetic aperture radar measurements to improve regional flux estimates from
the Amazon Basin. Prigent et al. (2007) have recently developed a multi-satellite method combining
passive microwave land surface emissions, active microwave backscatter measurements, and visible and
near-infrared reflectances to derive the first monthly estimates of inundated area for the years 1993
through 2000 at global scales. The inclusion of seasonal variability in fluxes, however, usually involves
either an empirical model or the development of a process-based model.
2.3.1.2 Process Modeling
Although there are still knowledge gaps, it is likely that major improvements in estimating flux will
require approaches in addition to simply adding to the flux database. The extrapolation of flux
measurements has provided the basis for initial global estimates, but it is clear that measurements cannot
be made in every environment under all conditions. Enough is understood about the dynamics of CFi4 in
2-7
-------
Chapter 2. Wetlands
natural systems to permit the development and testing of process-based models, and over approximately
the last 10 years these models have increased in number and complexity. Refining and improving models,
in combination with increasing the spatial and temporal resolution of the data used for their extrapolation,
should yield improved flux estimates.
Process-based models use an understanding of emissions and their environmental controls to calculate
fluxes. Because estimates are not based on actual measurements, there are additional sources of
uncertainty. A critical assumption is that the model actually captures the variables driving fluxes and that
its formulation realistically reflects the all of the important variables and flux responses. Currently many
process models are dependent upon the use of the NEE/CIrU flux relationship described by Whiting and
Chanton (1993), which does not appear to be applicable to all wetlands and may be species-specific
(Strom and Christensen, 2007).
As for extrapolating flux measurements, process models depend on the databases of environmental
drivers such as temperature, inundation, and primary productivity (for those using the NEE/CH4
relationship). An additional difficulty is the inclusion of the high spatial and temporal variability
characteristic of natural systems. This variability contains much of the information of interest for dynamic
flux estimations and for predictions under altered climate conditions. Currently, high variability creates
correspondingly high uncertainty in emissions. Although it is unlikely that all of this variability can be
resolved, improved statistical methods to include small-scale heterogeneity in models should reduce the
uncertainty. A technique known as nested sampling has been used with success in several recent
measurement programs.
2.3.1.3 Inverse Modeling
The concentration of CH4 in the atmosphere at any one time and place is a function of its sources and
sinks, modified by mixing and transport. Knowledge of how concentrations vary at multiple scales thus
can provide data on the factors that control its distribution. The addition of atmospheric concentrations as
a constraint on flux is a powerful tool. Inverse modeling attempts to optimize flux estimates given
knowledge of these distributions. Inverse ("top down") models incorporate atmospheric observations, a
model of atmospheric transport, and prior estimates of source distributions and magnitudes. Sources are
then evaluated to determine if they can "account" for observations and are adjusted, if required, to be
more consistent. Difficulties can arise in a variety of ways. Discriminating between co-located sources is
a challenge, and inadequate observations can bias results. Complex meteorology and topography are
difficult to simulate. If transport models have errors, these may be misinterpreted as source/observation
mismatches. Errors in initial assumptions about sources (spatial distribution, magnitude, variability, or
even types) can make model interpretation difficult. The spatial and temporal integration of various data
sets can also introduce errors. Inverse techniques, however, have increased in sophistication and power
over the past 10 years and offer a way to integrate highly variable emissions over large scales and to bring
other data sources to bear on the problem of estimating emissions. An increasing database on isotopic
ratios of atmospheric CUt is a valuable addition to inverse modeling, since these data can be used to help
discriminate between source types. Over the last several years, measurements of QrU mixing ratios from
satellites have been added to the observations available, significantly enhancing data coverage.
2.3.2 Global Wetland Methane Emissions
This subsection summarizes global wetland CH4 emissions developed using each of the three techniques
described in Section 2.3.1. The first global estimates from wetlands were made by extrapolating flux
measurements, which gave the initial indication of their significance to the global budget. As the
databases of flux, atmospheric measurements, and the variables needed for flux extrapolation has grown,
our ability to make more sophisticated estimates has increased.
2-8
-------
Chapter 2. Wetlands
2.3.2.1 Estimates Based on Flux Measurements
Table 2-1 summarizes existing QrU flux measurements and their extrapolation to a global scale from the
first version of this report. These estimates are in agreement with other bottom-up estimates based on the
existing flux database, shown in Table 2-2. These estimates reduced the initial Matthews and Fung (1987)
calculation of the relative importance of northern wetlands as more data became available. Note that
although uncertainties are not explicitly estimated, they are known to be large. All estimates are static,
since the databases on temperature and ecosystem area used for extrapolation are based either on long-
term means or a combination of compiled literature sources. Small-scale flux variability, differences
within ecosystem types, as well as possible year-to-year differences are included in error bars on mean
fluxes.
Table 2-1. Wetlands Methane Emissions (Tg CH4/Year),
From Bartlett and Harriss (1993) and U.S. EPA (1993)
Ecosystem
Tropical
Temperate
All northern
Boreal
Arctic
Well-drained tundra
Total
Non-
Forested
Swamps
31
1.5
1.4
0.2
34.1
Forested
Swamps
27
1.6
0.5
0.1
29.2
Non-
Forested
Bogs
0.5
0
4.9
4.9
10.3
Forested
Bogs
2.4
2.1
12.6
8.9
26
Alluvial
Formations
5.0
0.3
0
0
5.3
Annual
Emissions
66
5
38
20
14
4
109
Table 2-2. Global Wetland Methane Flux Estimates Made by Extrapolating
Flux Measurements (Tg CH4/Year)
Global
Estimate
Matthews and
Fung, 1987
Aselmann and
Crutzen, 1989
Bartlett etal.,
1990
Fung et al.,
1991
Bartlett and
Harriss, 1993
Climate Zone
High Latitude
(Nof50°N;
S of 50°S)
65
25
39
35
38
Temperate
(50-30°N;
30-50°S)
14
12
17
5
Tropical
(30°N-30°S)
32
43
55
80
66
Total
111
80
111
115
109
Comment
First global estimate,
based on relatively few
measurements
Uses a different wetland
area database than
Matthews and Fung
(1 987)
Significant increase in
number of tropical
emissions
Same formulation as
Matthews and Fung
(1 987) but more
measurements and
compares fluxes to
atmospheric
concentrations
Earlier version of this
report
2-9
-------
Chapter 2. Wetlands
Estimates of global wetland emissions based on flux extrapolation range from 80 to 115 Tg CFL/yr but
are not independent of one another, since several use the same area and inundation databases. Bartlett and
Harriss (1993) calculated emissions using two different wetlands area estimates and demonstrated that
although global emissions were similar, there were significant smaller-scale differences in geographic
distribution and between wetland types. Excluding the early Matthews and Fung (1987) calculations,
northern wetlands were extrapolated to release 30 to 35 Tg CFL/yr, an average of 33 percent of the global
total; temperate systems emitted 5 to 17 Tg CFL/yr (10 percent); and tropical systems emitted 50 to 70 Tg
CFL/yr (59 percent). After about 1993, few global flux estimates were made by extrapolating
measurements, since the database of atmospheric observations had grown, providing an additional
constraint on estimates (Fung et al., 1991), and knowledge of the processes controlling fluxes had
permitted the development of large-scale process-models (Cao et al., 1996).
2.3.2.2 Estimates Based on Process Models
Much of the more recent process and inverse modeling simulations is based on the work of Fung et al.
(1991). Later work uses the geographic and seasonal distributions of CFL, sources reported there and
updates emissions from other sources. For this reason, wetland emissions are commonly broken down
into the bog (forested and non-forested, largely located in the high latitudes) and swamp (forested and
non-forested, dominating tropical wetland areas) categories used by Fung and co-workers.
Global wetland emissions estimated from process models vary more widely than do those based on flux
extrapolation. This is in part due to the different approaches taken and to the use of different ways to
estimate the environmental factors used for model extrapolation. Global estimates range from 92 to 260
Tg CFL/yr, but the majority fall into the range of 140 to 160 Tg CFL/yr—somewhat greater than those
calculated from measurements, although uncertainties are quite large (Table 2-3). Note that the Gedney et
al. (2004) estimate includes rice paddies and so is not directly comparable to strictly natural wetlands.
Northern ecosystems in particular have been a focus of modeling efforts, so a greater number of estimates
have been made from these systems. As shown in Table 2-3, modeling calculations range from 20 to 72
Tg CFL/yr, for high latitudes. Note, however, that some authors calculate fluxes on the basis of wetland
types (i.e., bogs vs. swamp), which correspond to broad latitude classes, but are not an exact match.
Tropical (or swamp systems) are calculated to have greater emissions than northern wetlands, and range
from 41 to 195 Tg CFLVyr. For those estimates that include both northern and tropical wetlands, the global
contribution from northern systems ranged from 15 to 49 percent and averaged 31 percent. Tropical or
swamp systems were calculated to contribute 44 to 85 percent and averaged 64 percent. These relative
ecosystem contributions are similar to those calculated from flux measurement extrapolations.
Estimates published by Walter et al. (2001) are relatively high (65 Tg CFL/yr for bogs and 195 Tg CFL/yr
for swamps). This is believed to be due to the flux database upon which the model was based, rather than
a problem with the model structure, and the model has been used in other work yielding estimates more
similar to others reported in the literature (Shindell et al., 2004). Emissions calculated by Harder et al.
(2007) for high latitudes are also relatively high (72 Tg CFL/yr, making up 49 percent of the global total),
but it is difficult to determine how they were estimated other than that a variety of literature sources were
used. The only estimate of temperate wetland emissions from process modeling is that of Cao et al.
(1996). At 25 Tg CFL/yr, it is high in comparison to those derived from measurement extrapolation.
2-10
-------
Chapter 2. Wetlands
Table 2-3. Global Wetland Methane Flux Estimates Made by Extrapolating
Process Models (Tg CH4/Year)
Global
Estimate
Cao et al.,
1996
Christensen et
al., 1996
Lelieveld et
al., 1998
Walter etal.,
2001
Houweling et
al., 2000
Kaplan, 2002
Shindell et al.,
2004
Gedney et al.,
2004
Zhuang et al.,
2004
Zhuang et al.,
2006
Harder et al.,
2007
Bergamaschi
etal. ,2007
Climate Zone
High
Latitude
(Nof50°N;
S of 50°S)
25
20
(±13)
54
(bogs and
tundra)
65
24
57
36
72
62
Temperate
(50-30°N;
30-50°S)
25
Tropical
(30°N-30°S)
41
91
(swamps)
195
133
74
113
Total
92
145
(115-
175)
260
147
140
156
297
146
Comment
Modeled using NPP
and flux as a fraction
of heterotrophic
respiration
From a pre-industrial
budget using ice core
data and a CTM
Christensen et al.
(1996) structure,
BIOME4 model; much
larger estimated
wetland area
Based on Walter
model;
northern = 32°N-
90°N;
tropical = 32°N-32°S
Includes rice paddies
Uses Terrestrial
Ecosystem Model
(TEM); Nof45°N
Uses TEM
Based on Fung et al.
(1991) but revised by
combining many
sources
Uses Kaplan wetland
areas, Christensen
model, and a global
vegetation model — a
priori model input; total
flux specified = 174.5
2-11
-------
Chapter 2. Wetlands
Most process models employ a relationship between QrU flux and a measure of plant productivity as the
core structure that relates flux to environmental controls (Walter et al., 2001). Christensen et al. (1996)
use BIOME2 modeled net primary productivity (NPP) estimates, as moderated through calculations of
heterotrophic soil respiration, to estimate flux. Several models attempt to include all three transport
pathways of CH4 to the atmosphere. This requires inputs and/or calculations of vegetation type, density,
root distribution, and soil QrU distributions. Zhang et al. (2002) have modified a well-known model of
soil carbon and nitrogen dynamics (DNDC) to include functions unique to wetlands. Their Wetlands-
DNDC model uses the Walter methanogenesis process model and includes water table changes, the
growth of mosses as well as vascular plants, and anaerobic soil processes.
In addition to the models referenced in Table 2-3 that have been used to derive large-scale flux estimates,
a number of models have been developed that focus on regional or ecosystem-specific estimates. These
include those of Frolking and Crill (1994), Potter (1997), Grant and Roulet (2002), Frolking et al. (2002),
and Kettunen (2003). Potter et al. (2006) use the NEPiQrU relationship in combination with a model to
estimate NEP, and a land cover database to calculate U.S. wetland emissions at 5.5 Tg CHVyr.
2.3.2.3 Estimates Based on Inverse Modeling
After the initial work of Hein et al. (1997) and Houweling et al. (1999), improvements in inverse
modeling techniques have led to a number of recent global CH4 budgets, as shown in Table 2-4. Early
estimates from wetlands are believed to be high due to an overestimate of the magnitude of the OH sink
(Wang et al., 2004). As Table 2-4 illustrates, variability in emission estimates is still high, although values
are broadly similar to those generated from process models. Estimated flux from bogs or high-latitude
wetlands is relatively consistent and ranges from 21 to 47 Tg CHVyr, with an average of 35 Tg CHVyr.
Estimated flux from swamps or tropical wetlands is higher and more variable, ranging from 81 (an
estimate believed to include a contribution from co-located rice areas) to 206 Tg CHVyr. Swamp
estimates average 144 Tg CHVyr. Wetlands classified as bogs are calculated to contribute 9 to 29 percent
of the global total, averaging 20 percent, while swamp systems contribute 56 to 89 percent, averaging 78
percent of total emissions.
The increasing database of atmospheric observations, sampled both at the ground surface and from
satellites, has permitted the calculation of interannual (year-to-year) variability in flux. This is an
important capability since emissions have a strong response to variation in environmental drivers.
Variability in wetlands emissions (driven by decreases in temperature and other variables), for example,
has been suggested as an important factor in slowing the growth rate of atmospheric CFLj in the early
1990s (Hogan and Harriss, 1994; Wang et al., 2004). The large increase in atmospheric QrU observed in
1998 has also been attributed to variability in wetland emissions (Chen and Prinn, 2006). These authors
estimate that from year to year, wetland fluxes (and perhaps co-located rice agriculture) may increase as
much as 19 Tg CHVyr or decrease 15 Tg CHVyr, a significant fraction of total global emissions. Bousquet
et al. (2006) suggest that between 1991 and 1993, wetland emissions decreased by 24 ± 6 Tg QrU due to a
decrease in solar radiation after the eruption of Mt. Pinatubo. They find that emissions from wetlands are
the dominant source contributing to the interannual variability in emissions, explaining 70 percent of the
global changes in growth rate over the last 20 years.
2-12
-------
Chapter 2. Wetlands
Table 2-4. Global Wetland CH4 Flux Estimates Made by Inverse
Modeling (Tg CH4/Year)
Global
Estimate
Hein et al.,
1997
Houweling et
al., 1999
Wang et al.,
2004
Mikaloff
Fletcher et al.,
2004
Chen and
Prinn, 2006
Bousquet et
al., 2006
Bergamaschi
etal.,2007
Climate Zone
High Latitude
(Nof50°N;
S of 50°S)
44 ±7
(bogs)
31 .4 (90-45°N)
0 (45-90°S)
27 ±3
(bogs and
tundra)
21 ±14
(bogs)
4±4
(tundra)
34
43 ±8
(bogs and
tundra)
47
Temperate
(50-30°N;
30-50°S)
Tropical
(30°N-30°S)
192 ±19
(swamps)
48.4 (45-0°N)
65.1 (0-45°S)
149 ±10
(swamps)
206 ± 44
(swamps)
81
104 ±12
(swamps)
161
Total
237 (±20)
145 (±41)
176
231
143-148
147 (±15)
208
Comment
As cited in Dentener et
al. (2003)
1998-1999
Includes a contribution
from Southeast Asia
rice; interannual wetland
variability = +19 and -15
Tg CH4/yr
Long-term mean for
1984-2003
Model constrained by
satellite and surface
observations
2.3.2.4 Summary of Methane Emissions From Wetlands
The range in calculated ClrU flux from wetlands is summarized in Table 2-5 according to the techniques
used to make the calculations. Uncertainties are not included here, but often range from roughly 10 to 75
percent. This chapter's analysis derives a "best guess" wetlands emissions estimate by taking an average
of process and inverse model estimates published since 2004. Calculations made by the extrapolation of
flux measurements are not included, since this technique has been largely superseded. Uncertainties in
these averages are based only on the range in estimates. Table 2-5 suggests that calculated emissions from
northern wetlands have fallen as the modeling techniques have improved and that the overall magnitude
of emissions from wetlands, illustrated by summed averages and ranges, has increased.
Here we estimate that wetlands release on the order of 170 Tg CHVyr. The majority (75 percent) of
emissions are from swamp systems located largely in the tropics. Estimates of these emissions range
widely, however. If global emissions of ClrU from all sources are roughly 566 Tg CHVyr (see Table ES-
1), then wetlands contribute about 30 percent. Interannual variability, driven largely by climate
variability, is high. Using the estimates of interannual variability derived by Chen and Prinn (2006) and
Bousquet et al. (2006) of 15 to 19 Tg, the wetlands contribution may vary by ±3 percent from year to
year.
2-13
-------
Chapter 2. Wetlands
Table 2-5. Summary of Estimated Wetland CH4 Fluxes by Technique (Tg CH4/Year)
Approach
Flux extrapolation
Process modeling
Inverse modeling
Current best guess
(process and inverse
modeling since 2004)
Northern/Bogs
31-48a
avg = 38 (37%)
20-72°
avg = 44 (31%)
21-47
avg = 36 (20%)
24-72
avg = 42.7 (25%)
std. dev. = 16.6; n = 10
Tropical/Swamps
49-80
avg = 65 (63%)
41-133
avg = 90 (64%)
81-206
avg = 144(78%)
81-206
avg = 127.6(75%)
std. dev. = 44.0; n = 8
Total
80-115
sum of avgs = 103
n = 4
92-156
sum of avgs = 134
n = 8 (bogs); 5
(swamps)
145-237
sum of avgs = 180
n = 6
170.3
range = 105-278 by
summing minima and
maxima
For flux extrapolation, temperate emissions are split equally between bogs and swamps. Values in parentheses
indicate percentage contribution to wetland total emissions.
Walter et al. (2001) estimates excluded.
2.3.3 Global Wetland Nitrous Oxide Emissions
There are very few global estimates of the N2O released from wetlands, and wetlands are not included in
N2O budgets as significant sources. Bouwman et al. (1993) modeled global N2O based on the input of
organic matter, soil fertility, moisture status, temperature, and oxygen level and concluded that wetlands
appeared to make only a minor contribution. A number of measurements have been made of emissions
from coastal mangrove swamps, some of which received outside N inputs from sewage or bird droppings.
Corredor et al. (1999) calculate global emissions from mangroves ranging from 0.004 to 0.17 Tg N/yr.
Similar magnitude fluxes were measured by Allen et al. (2007) over nearly an annual period. Barnes et al.
(2006) calculated that coastal mangrove systems release 0.076 Tg N/yr and find that these systems make
only a minor contribution to global fluxes. Emissions of N2O were measured from intertidal mud flats in
the Dutch Scheldt estuary by Middelburg et al. (1995). They calculated a small global estuarine intertidal
flux of 0.0013 Tg N/yr. 18.8 Tg N/yr (see Table ES-1), If the total N2O to the atmosphere is 18.8 Tg N/yr
(see Table ES-1), then these wetland systems would contribute 0 to 1 percent. Working in the Hudson
Bay Lowlands, Schiller and Hastie (1994) also report low emissions. Overall, the authors found that these
wetlands acted as a small source to the atmosphere and calculate that N2O emissions from the Hudson
Bay Lowlands would contribute on the order of 0.0005 to 0.005 percent to global N2O emissions. Open
fens released the majority of the regional N2O emissions, while treed fens and bogs had the lowest flux.
Although the Third IPCC Assessment Report (Ehhalt et al., 2001) includes a source called "tropical wet
forest" which is a major source to global emissions, this category would be more accurately titled
"tropical humid forest" and is not actually wetland. Wetlands are not included as a source of N2O in the
most recent IPCC report (Denman et al., 2007).
The possibility that wet soils may at times act as a sink for atmospheric N2O was suggested a number of
years ago based on laboratory measurements (Blackmer and Bremner, 1976). Scattered measurements
have found that this occurs in a variety of wetlands (Keller et al., 1986; Regina et al., 1996; Schiller and
Hastie, 1994). Schiller and Hastie, for example, found that about 34 percent of their measurements in the
Hudson Bay Lowlands showed atmospheric uptake. Chapuis-Lardy et al. (2007) summarize
measurements of N2O uptake. They suggest that low NOs availability and conditions in soils that slow
diffusion, such as the water-saturated soils of wetlands, promote N2O consumption. If a small soil sink for
N2O is a widespread occurrence in wet soils, understanding its magnitude and dynamics would improve
global budget estimates. In summary, wetlands appear to contribute negligibly to atmospheric N2O; their
2-14
-------
Chapter 2. Wetlands
role as a sink can not be currently estimated, but is likely to be small in comparison to the stratospheric
sink.
2.4 Future Emission Scenarios
This section focuses on possible future emissions of CIL,. Because wetlands do not appear to be a
significant global source of N2O, no work has been undertaken to estimate how these emissions may
change. In general, changes in soil moisture would be expected to have the greatest impact on N2O fluxes
by changing the relative proportions of NO, N2O, and N2. If wetlands act as a small, diffuse sink for
atmospheric N2O, it is possible that climate changes may alter the magnitude and timing of this part of the
budget.
Before the Industrial Revolution, natural wetlands were the dominant source of CIL, to the atmosphere
(Brook et al., 2000; Etheridge et al., 1998). The increase in human population has decreased the relative
importance of wetland sources by both increasing anthropogenic sources of CH4 and decreasing the
wetlands source through drainage and land use change. Both direct and indirect changes in wetland fluxes
will continue because many of the controls on flux—such as temperature, rainfall, and vegetation type—
are among those expected to change under projected altered climate regimes. The process model
developed by Walter et al. (2001) suggests, for example, that a ± 1°C change in temperature may change
CH4 emissions by ± 20 percent and that a ± 20 percent change in precipitation may change emissions by ±
8 percent. The response of emissions to climatic change will vary with latitude. The complex link
between climate and wetland CIL, emissions may also be a positive feedback mechanism accelerating
changes in climate (Woodwell et al., 1998; Monson and Holland, 2001). High-latitude wetlands store a
considerable fraction of the global soil C pool as a result of the relatively small long-term imbalance
between C production and decay. Understanding the sensitivity of this pool to changes in environmental
conditions is crucial to understanding the effects of global climate change.
A number of studies have attempted to model the response of wetlands to climate change. These
simulations commonly use a process model to estimate current emissions, re-run the model under altered
environmental conditions, and then compare subsequent results. Climate models have indicated that there
may be large changes in temperature and moisture patterns in the high latitudes over the next 100 years.
Because temperature is also the major driver of northern seasonal cycles, much of this work concentrates
on boreal wetland responses. The response of wetland CIL, emissions to changes in climate has also been
investigated by examining the link between climate and atmospheric CH4 in the past through ice core
records.
Working on a regional scale, Zhuang et al. (2007) have attempted to model Alaskan CIL, fluxes and
simulate changes in response to the region's expected changes in climate over the 21st century. They
calculate that emissions from wet soils will be enhanced more than will oxidation in dry tundra and forest
soils. As a result, projected CH4 emissions from northern wetlands nearly double by the end of the
century. In the context of the overall carbon budget (CO2 and CIL,), it is estimated that in Alaska, net CIL,
emissions will be greater than C sequestration and that there will be a positive feedback between radiative
forcing due to changes in wetland C cycling and climate change.
Work on estimated changes in wetland CH4 emissions is summarized in Table 2-6. Studies to date
indicate that emissions will be strongly affected by projected climate scenarios. Altered climate
conditions impact emissions in both direct ways (e.g., temperature and precipitation change) and indirect
ways (e.g., CO2 fertilization effects on plant productivity, vegetation change, permafrost thaw, effects on
thermokarst lakes, changes in biomass burning regimes). A more detailed discussion on some of these
indirect effects on CH4 emissions can be found in Chapter 5 ("Permafrost") and Chapter 6 ("Lakes").
2-15
-------
Chapter 2. Wetlands
Table 2-6. Projected Changes in Wetland CH4 Emissions
(Tg CH4/Year)
Ecosystem
Global
Global
Northern, <
50°N
Northern, <
40°N
Global
Time
Frame
2100
2100
2030
2080
2030
2080
Environmental
Change
2XCO2
Model simulation of
IPCC IS92a
emissions scenario
Climate change from
Integrated Global
Systems Model
(IGSM)
GHGsa
GHGs + SO4
deposition13
GHGs + SO4dep +
SO4 aerosols
GHGs
GHGs + SO4
deposition
GHGs + SO4dep +
SO4 aerosols
GHGs
GHGs + SO4
deposition
GHGs + SO4dep +
SO4 aerosols
GHGs
GHGs+ S04
deposition
GHGs+SO4 dep+SO4
aerosols
Flux Change
Increase of 78% from 156
to 277
Doubling, an increase from
-300 to 500-600;
flux increase similar to
projected anthropogenic
flux increase
More than double from
current 41 to 58
Increase -13% from 1960
estimate (35 to 40)
Increase 13% from 1960
estimate (31 to 35)
Unchanged at 35
Increase 30% from 1960
estimate (35 to 45.5)
Increase 39% from 1960
estimate (31 to 43)
Increase 23% from 1960
estimate (35 to 43)
Increase -10% from 1960
estimate (165 to 181)
Increase -3% from 1960
estimate (155 to 160)
Decrease -1 % from 1 960
estimate (1 57 to 1 56)
Increase -17% from 1960
estimate (165 to 193)
Increase -19% from 1960
estimate (155 to 185)
Increase -13% from 1960
estimate (1 57 to 1 78)
Reference
Shindelletal., 2004;
majority of increase from
tropics
Gedney etal., 2004;
change due largely to
temperature response
Zhuang etal., 2006;
permafrost thaw and CO2
fertilization increase fluxes
Gaucietal., 2004; uses
Walter process model and
tests a combination of
environmental changes; S
pollution decreases 1960
flux and reduces GHG
effect
Gaucietal., 2004; S
suppression becomes less
important, fluxes similar to
GHG alone
Gaucietal., 2004; flux
reduced 21-25Tg by S
pollution.
Gaucietal., 2004;
increases under S
suppression scenario are
smaller; S pollution
reduced by anticipated
cleaner technologies
Greenhouse gases.
Simulates acid rain.
A rising sea level is also likely to affect emissions. Coastal inundation will flood terrestrial systems and
may transform some of them to CH/t sources. In addition, some current freshwater wetlands may
transition to more saline systems as they are flooded, which is likely to decrease emissions. Coastal
inundation and/or an increase in storm frequency is also likely to increase shoreline erosion, resulting in
wetland loss (Nicholls, 2004). Shoreline retreat, the natural response to sea level rise, is made difficult
2-16
-------
Chapter 2. Wetlands
due to human land use in coastal zones. The rate of sea level rise is an important variable in determining
the fate of coastal fresh and saline wetlands. If it is approximately equal to the rate of soil or peat
accretion in these systems, they are likely to remain relatively stable (Moorhead and Brinson, 1995;
McFadden et al., 2007). Carbon stored in the peats of coastal wetlands after inundation has a largely
unknown future, but has the potential to be a significant C source to the atmosphere (Henman and Poulter,
2008).
Anthropogenic activities in addition to those that release greenhouse gases also impact emissions from
wetlands. Gauci et al. (2004) model the effects of the sulfur in acid deposition on wetlands. They
calculate that sulfur deposited by acid rain has already (year 2000) suppressed fluxes over their pre-
industrial level by roughly 15 Tg CFL/yr. The projected time course of the effects of S deposition are
complex, in part due to regional changes in economic growth and anticipated cleaner technologies, but
sulfur pollution appears to be a potentially important factor in understanding future emissions (Table 2-6).
In addition to modeling extrapolation, there have
been manipulative field experiments as well as
work comparing naturally existing
environmental gradients that can be used as
analogs for the anticipated changes under altered
climate scenarios. As for modeling climate
scenarios, the majority of this work has focused
on northern systems. A major unknown is the effect of the loss of permafrost on tundra vegetation,
surface water distribution, soil respiration, and organic matter accumulation (Jorgenson et al., 2006;
Walter et al., 2007; Riodan et al., 2006). Turetsky et al. (2002) suggest that, in the short term at least,
permafrost melting is associated with a 30-fold increase in CH4 flux. Strom and Christensen (2007) found
that the change in plant species caused by permafrost degradation and hydrologic changes created
changes in plant productivity, photosynthetic rate, and CH4 flux. These changes may ultimately mean that
CH4from melting permafrost will act as an increasing source of radiative forcing in the future. Wickland
et al. (2006) found that the formation of
thermokarst wetlands by permafrost melting
caused a 13-fold increase in a site's annual
CH4 flux, an increase similar to that reported
by other studies (Turetsky et al., 2002;
Bubier et al., 1995; Christensen et al., 2004;
Johansson et al., 2006). Changes in
emissions due to the formation of
Manipulative field experiments are observations
made in natural systems in which something
(usually a controlling variable such as water level,
nutrient inputs, or temperature) has been altered in
order to assess ecosystem response.
Thermokarst wetlands are wetlands formed in
depressions by meltwater from thawing permafrost.
These depressions are often produced by ground
subsidence associated with permafrost melt and may
also form small lakes (see Chapter 3 for more detail on
flux from thermokarst lakes). Continued melting of the
permafrost can lead to the drainage and eventual
., , .,, ,. j • /~>u * /: disappearance of thermokarst wetlands and lakes.
thermokarst lakes are discussed in Chapter 6 I MM
("Lakes").
Changes in precipitation are clearly crucial, since small changes in water saturation may result in large
flux changes (von Fischer and Hedin, 2007). Strack and Waddington (2007) demonstrate, however, that
the response of a wetland to water table drawdown is complex and will vary with the existing
microtopography. In their work, relatively high areas showed an increase in net global warming potential
(CO2 and CFL,) over controls while the global warming potential of topographic lows decreased. Roulet et
al. (1992) compare the effects of temperature and precipitation changes associated with a 2 x CO2 climate
change scenario (if CO2 concentrations were to double) on emissions from a northern fen. Their model
found that increased peat temperatures could raise flux by 5 to 40 percent, but that the projected drop in
water table decreased flux by 74 to 81 percent, suggesting that northern peatlands may be more sensitive
to changes in moisture than temperature. A decrease in wetland water level would impact other
radiatively important gases such as CO2 and N2O in addition to CFL,. Laine et al. (1996) have attempted to
integrate the response of all three gases to a water table drop by comparing drained and undrained
peatlands of varying productivity. They suggest that a general drying may decrease the impact of northern
2-17
-------
Chapter 2. Wetlands
peatlands by about 0. 1 W/m due to a decrease in CFU flux, small changes in peat C storage, and an
increase in tree-stand biomass.
An important, but difficult factor to include in future emission scenarios is the difference between short-
term and long-term change. Part of the difficulty is that changing systems never reach a steady state, but
in addition, biological systems respond to change with both short-term and long-term adaptations.
Experimental manipulations seldom observe long-term responses since they typically do not last for more
than a year or two. For this reason, some studies have used natural spatial gradients, for example in
nutrient deposition, water table drainage, or disturbance, as analogs to changes over time since these can
potentially include long-term environmental responses. Summarizing the apparently contradictory
literature on peatland responses to lowered water levels, Laiho (2006) suggests that differences may be
due to observation time scales. Short-term changes in an environmental control such as water level
represent a disturbance to a system that is in relative equilibrium with previous conditions, while long-
term changes result in adaptations to a new regime. He proposes that although disturbed systems will
always lose C in the short term, the long-term response can be highly variable. These adaptations to a new
hydrologic regime will depend on variables such as the peatland type, climate, and the amount of change
from previous conditions. Frolking et al. (2006) also emphasize the difference between short- and long-
term responses to an environmental perturbation and the necessity of taking a holistic approach when
evaluating the impacts of climate change. They suggest that the net effect of peatland gas fluxes is a
balance between the rapid, strong warming due to CF^flux and the slow cooling due to CO2 uptake. In
the short term, the CH4 response will dominate radiative impacts, but after several decades, changes in
CO2 emissions will take over. Short-term and long-term effects are thus quite different.
emissions from wetlands have been impacted by humans in the past and will continue to be impacted
into the future as a consequence of climate change and land use change. Current knowledge suggests that
emissions in the Northern Hemisphere will likely be significantly altered by anticipated changes in
temperature and water regime, two of the key variables that control flux. In the tropics, it is likely that
land use changes will have the greatest impact on emissions as these regions undergo developmental and
population growth. The potential exists for large emission changes from both regions, both positive and
negative.
2.5 Areas for Further Research
Continued improvements in large-scale process models should yield an increased understanding of the
smaller-scale variability in flux. One way to approach this problem is to test models against long-term
measurements to assess natural variability. These long-term datasets are few in number, which limits
model development and verification. Existing long-term measurement programs should therefore be
continued and new measurement sites added, particularly in the tropics.
Global emissions estimates have consistently pointed to tropical wetlands as major CFU sources, but work
at these latitudes remains sparse and incomplete. Although satellite measurements have begun to fill gaps
in the atmospheric observational network, increased surface sampling will help resolve current inverse
modeling difficulties with low latitude sources.
Since seasonal variability in these systems is driven largely by changes in water level, improvements in
our ability to model or accurately measure hydrologic change are crucial to being able to apply process
models. The inclusion of ephemeral wetlands at all
latitudes remains difficult.
The empirically based NPPiCFUflux relationship, which
has served as a fundamental core of many process
models, has proven useful, but does not appear to be
universally applicable. Testing of the relationship should
Ephemeral means "short-lived." An
ephemeral wetland, pond, or spring
exists for only a brief period, usually
following precipitation orsnowmelt. An
ephemeral wetland is different from a
seasonal wetland, which may exist for a
longer period (still less than a year).
2-18
-------
Chapter 2. Wetlands
continue and more habitat-specific relationships should be developed.
Incorporating small-scale heterogeneity into large-scale modeling remains difficult. Improvements in the
ability to assign uncertainties will aid in distinguishing short-term variability from true long-term trends
and will make assessment of the response of emissions to climate changes more powerful.
Combining a spatially explicit flux process model with an atmospheric transport model would produce
seasonal and interannual simulations of atmospheric distributions. Comparing these to satellite-derived
distributions and/or surface sampling network distributions should prove informative.
Because emission to the atmosphere is a function of the competing processes of CH4 production and
consumption, both processes and their responses to environmental controls must be understood across the
landscape. Episodic emissions, which may release a sizeable fraction of annual flux, remain difficult to
both measure and include in models. Failure to adequately incorporate these fluxes, however, can yield
inaccurate and misleading results.
Fluxes of N2O from wetlands remain largely unknown, particularly from tropical regions. Although
measurements indicate they are relatively low, at least in the wetlands sampled to date, the seasonal
changes in water level that characterize many tropical wetlands suggest that there may be brief periods of
emissions. Experimental work examining emissions over wetting and drying cycles suggests thatN2O
fluxes may be enhanced by these moisture changes. Understanding of the global extent and importance of
the uptake of atmospheric N2O in wetland soils should be improved.
2.6 References
Allen, D.E., R.C. Dalai, H. Rennenberg, R.L. Meyer, S. Reeves, and S. Schmidt. 2007. Spatial and
temporal variation of nitrous oxide and methane flux between subtropical mangrove sediments and the
atmosphere. SoilBiol. Biochem. 39: 622-631.
Aselmann, I., and P.J. Crutzen. 1989. Global distribution of natural freshwater wetlands and rice paddies,
their net primary productivity, seasonality, and possible methane emissions. J. Atmos. Chem. 8: 307-
358.
Baker-Blocker, A., T.M. Donahue, and K.H. Mancy. 1977. Methane flux from wetland areas. Tellus 29:
245-250.
Barnes, J., R. Ramesh, R. Purvaja, A.N. Raijkumar, B.S. Kumar, K. Krithika, K. Ravichandran, G. Uher,
and R. Upstill-Goddard. 2006. Tidal dynamics and rainfall control N2O and CFLt emissions from a
pristine mangrove creek. Geophys. Res. Lett. 33: L15405, 10.0129/2006GL026829.
Bartlett, K.B., D.S. Bartlett, R.C. Harriss, and D.I. Sebacher. 1987. Methane emissions along a salt marsh
salinity gradient. Biogeochem. 4: 183-202.
Bartlett, K.B., P.M. Crill, D.I. Sebacher, R.C. Harriss, J.O. Wilson, and J.M. Melack. 1988. Methane flux
from the central Amazonian floodplain. J. Geophys. Res. 93: 1571-1582.
Bartlett, K.B., P.M. Crill, J.A. Bonassi, J.E. Richey, and RC. Harriss. 1990. Methane flux from the
Amazon River floodplain: Emissions during rising water. J. Geophys. Res. 95: 16773-16788.
Bartlett, K.B., and R.C. Harriss. 1993. Review and assessment of methane emissions from wetlands.
Chemosphere, 26(1-4): 261-320.
Bergamaschi, P., C. Frankenberg, J.F. Meirink, M. Krol, F. Dentener, T. Wagner, U. Platt, J.O. Kaplan,
S. Korner, M. Hermann, E.J. Dlugokencky, and A. Goede. 2007. Satellite chartography of atmospheric
methane from SCHIAMACHY on board ENVISAT: 2. Evaluation based on inverse model simulation.
J. Geophys. Res. 112: D02304, 10.1029/2006JD007268.
2-19
-------
Chapter 2. Wetlands
Blackmer, A.M., and J.M. Bremner. 1976. Potential of soil as a sink for atmospheric nitrous oxide.
Geophys. Res. Lett. 3: 739-742.
Bousquet, P., P. Ciais, J.B. Miller, E.J. Dlugokencky, D.A. Hauglustaine, C. Prigent, G.R. Van der Werf,
P. Peylin, E.-G. Brunke, C. Carouge, R.L. Langenfelds, J. Lathiere, F. Papa, M. Ramonet, M. Schmidt,
L.P. Steele, S.C. Tyler, and J. White. 2006. Contribution of anthropogenic and natural sources to
atmospheric methane variability. Nature 443(7110): 439-443.
Bouwman, A.F., I. Fung, E. Matthews, and J. John. 1993. Global analysis of the potential for N2O
production in natural soils. Global Biogeochem. Cycles 7(3): 557-597.
Brook, E.J., S. Harder, J. Severinghaus, E.J. Steig, and C.M. Sucher. 2000. On the origin and timing of
rapid changes in atmospheric methane during the last glacial period. Global Biogeochem. Cycles. 14:
559-572.
Bubier, J.L., T.R. Moore, L. Bellisario, N.T. Comer, and P.M. Crill. 1995. Ecological controls on
methane emissions from a northern peatland complex in the zone of discontinuous permafrost,
Manitoba, Canada. Global Biogeochem. Cycles. 9(4): 455-470.
Cao, M., S. Marshall, and K. Gregson. 1996. Global carbon exchange and methane emissions from
natural wetlands: Application of a process-based model. J. Geophys. Res. 101:14399-14414.
Chang, C., H.H. Janzen, C.M. Cho, and E.M. Nakonechny. 1998. Nitrous oxide emission through plants.
Soil Sci. Soc. Am. J. 62: 35-38.
Chanton, J.P., and J.W.H. Dacey. 1991. Effects of vegetation on methane flux, reservoirs, and carbon
isotope composition. In: T.D. Sharkey, E.A. Holland, and H.A. Mooney (eds.). Trace Gas Emissions by
Plants. New York: Academic Press, pp. 65-92.
Chapuis-Lardy, L., N. Wrage, A. Metay, J.L. Chotte, and M. Bernoux. 2007. Soils, a sink for N2O? A
review. Global Change Biol. 13(1): 1-17.
Chen, Y.-H., and R.G. Prinn. 2006. Estimation of atmospheric methane emissions between 1996 and
2001 using a three-dimensional global chemical transport model. J. Geophys. Res. Ill: D103 07,
10.1029/2005JD006058.
Christensen, T.R., I.C. Prentice, J. Kaplan, A. Haxeltine, and S. Sitch. 1996. Methane flux from northern
wetlands and tundra. Tellus 48B: 652-661.
Christensen, T.R., T. Friborg, M. Sommerkorn, J. Kaplan, L. Illeris, H. Soegaard, C. Nordstroem, and S.
Jonsson. 2000. Trace gas exchange in a high arctic valley, 1. Variations in CO2 and CH4 flux between
tundra and vegetation types. Global Biogeochem. Cycles. 14: 701-713.
Christensen, T.R., A. Ekberg, L. Strom, M. Mastepanov, N. Panikov, M. Oquist, B.H. Svensson, H.
Nykanen, P.J. Martikainen, and H. Oskarsson. 2003. Factors controlling large scale variations in
methane emissions from wetlands. Geophys. Res. Lett. 30(7): 1414, 10.1029/2002GL0116848.
Christensen, T.R., T. Johansson, H.J. Akerman, M. Mastepanov, N. Malmer, T. Friborg, P. Crill, and
B.H. Svensson. 2004. Thawing sub-arctic permafrost: Effects on vegetation and methane emissions.
Geophys. Res. Lett. 31: L04501, 10.1029/GL18680.
Coles, J.R.P., and J.B. Yavitt. 2002. Control of methane metabolism in a forested northern wetland, New
York. Geomicrobiol. J. 19: 293-315.
Corredor, J.E., J.M. Morell, and J. Bauza. 1999. Atmospheric nitrous oxide fluxes from mangrove
sediments. Mar. Poll. Bull. 38: 473-478.
Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of Wetland and Deepwater
Habitats of the United States. U.S. Department of the Interior, Fish and Wildlife Service.
2-20
-------
Chapter 2. Wetlands
Davidson, E.A.. 1991. Fluxes of nitrous oxide and nitric oxide from terrestrial ecosystems, In: J.E.
Rodgers and W.B. Whitman (eds.). Microbial Production and Consumption of Greenhouse Gases:
Methane, Nitrogen Oxides, andHalomethanes. Washington: American Society for Microbiology, pp.
219-235.
Davidson, E.A., P.A. Matson, P.M. Vitousek, R. Riley, K. Dunkin, G. Garcia Mendez, and J.M. Maass.
1993. Processes regulating soil emissions of NO and N2O in a seasonally dry tropical forest. Ecol. 74:
130-139.
Davidson, E.A., and L.V. Verchot. 2000. Testing the hole-in-the-pipe model of nitric and nitrous oxide
emissions from soils using the TRAGNET database. Global Biogeochem. Cycles 14: 1035-1043.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press, pp. 499-587.
Dentener, F., M. van Weele, M. Krol, S. Houweling, and P. van Velthoven. 2003. Trends and inter-
annual variability of methane emissions derived from 1979-1993 global CTM simulations. Atmos.
Chem. Phys. 3: 73-88.
Dise, N.B., and E.S. Verry. 2001. Suppression of peatland methane emission by cumulative sulfate
deposition in simulated acid rain. Biogeochem. 53:143-160.
Ehhalt, D.H. 1974. The atmospheric cycle of methane. Tellus 26: 58-70.
Ehhalt, D., M. Prather, F. Dentener, R. Derwent, E. Dlugokencky, E. Holland, I. Isaksen, J. Katima, V.
Kirchhoff, P. Matson, P. Midgley, and M. Wang. 2001. Atmospheric Chemistry and Greenhouse Gases.
In: Houghton, J.T., Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A.
Johnson (eds.). Climate Change 2001: The Scientific Basis, Contribution of Working Group I to the
Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and
New York, NY: Cambridge University Press.
Etheridge, D.M., L.P. Steele, R.J. Francey, and R.L. Langenfelds. 1998. Atmospheric methane between
1000 A.D. and present: Evidence of anthropogenic emissions and climatic variability. J. Geophys. Res.,
103: 15979-15993.
Frey, K.E., and L.C. Smith. 2007. How well do we know northern land cover? Comparison of four global
vegetation and wetland products with a new ground-truth database for West Siberia. Global
Biogeochem. Cycles. 21: GB1016, 10.1029/2006GB002706.
Frolking, S., and P.M. Crill. 1994. Climate controls on temporal variability of methane flux from a poor
fen in southeastern New Hampshire. Global Biogeochem. Cycles. 8: 385-397.
Frolking, S., N.T. Roulet, T.R Moore, P.M. LaFleur, J.L. Bubier, and P.M. Crill. 2002. Modeling
seasonal to annual carbon balance of Mer Bleu Bog, Ontario, Canada. Global Biogeochem. Cycles. 16,
1030, 10.1029/2001GB001457.
Frolking, S., N. Roulet, and J. Fuglesvedt. 2006. How northern peatlands influence the Earth's radiative
budget: Sustained methane emission versus sustained carbon sequestration. J. Geophys. Res. Ill:
G01008, 10.1029/2005JG000091.
Fung, I., J. John, J. Lerner, E. Matthews, M. Prather, L.P. Steele, and P.J. Fraser. 1991. Three-
dimensional model synthesis of the global methane cycle. J. Geophys. Res. 96: 13033-13065.
2-21
-------
Chapter 2. Wetlands
Garnet, K.N., J.P. Megonigal, C. Litchfield, and G.E. Taylor Jr. 2005. Physiological control of leaf
methane emission from wetland plants. Aquatic Bot. 81: 141-155.
Gauci, V, N. Dise, and D. Fowler. 2002. Controls on suppression of methane flux from a peat bog
subjected to simulated acid rain sulfate deposition. Global Biogeochem. Cycles 16(1): 1004,
10.1029/GB0001370.
Gauci, V., N. Dise, and S. Blake. 2005. Long-term suppression of wetland methane flux following a pulse
of simulated acid rain. Geophys. Res. Lett., 32(12): L12804.
Gauci, V., E. Matthews, N. Dise, B. Walter, D. Koch, G. Granberg, and M. Vile. 2004. Sulfur pollution
suppression of the wetland methane source in the 20th and 21st centuries. PNAS 101(34): 12583-12587.
Gedney, N., P.M. Cox, and C. Huntingford. 2004. Climate feedback from wetland methane emissions.
Geophys. Res. Lett. 31: L20503, 10.1029/2004GL020919.
Grant, R.F., and N.T. Roulet. 2002. Methane efflux from boreal wetlands: Theory and testing of the
ecosystem model Ecosys with chamber and tower flux measurements. Global Biogeochem. Cycles 16:
1054, 10.1029/2001GB001752.
Groffman, P.M. 1991. Ecology of nitrification and denitrification in soil evaluated at scales relevant to
atmospheric chemistry. In: J.E. Rodgers and W.B. Whitman (eds.).MicrobialProduction and
Consumption of Greenhouse Gases: Methane, Nitrogen Oxides, andHalomethanes. Washington:
American Society for Microbiology, pp. 201-217.
Happell, J.D., and J.P. Chanton. 1993. Carbon remineralization in a north Florida swamp forest: Effects
of water level on the pathways and rates of soil organic matter decomposition. Global Biogeochem.
Cycles 7: 475-490.
Harder, S.L., D.T. Shindell, G.A. Schmidt, and E.J. Brook. 2007. A global climate model study of CH4
emissions during the Holocene and glacial-interglacial transitions constrained by ice core data. Global
Biogeochem. Cycles 21: GB1011, 10.1029/2005GB002680.
Harriss, R.C., and D.I. Sebacher. 1981. Methane flux in forested freshwater swamps of the southeastern
United States. Geophys. Res. Lett. 8: 1002-1004.
Harriss, R.C., D.I. Sebacher, and P.P. Day Jr. 1982. Methane flux in the Great Dismal Swamp. Nature.
297: 673-674.
Hein, R., P.J. Crutzen, and M. Heimann. 1997. An inverse modeling approach to investigate the global
atmospheric methane cycle. Global Biogeochem. Cycles 11: 43-76.
Henman, J., and B. Poulter. 2008. Inundation of freshwater peatlands by sea level rise: Uncertainty and
potential carbon cycle feedbacks. J. Geophys. Res. 113: G01011, 10.1029/2006JG000395.
Hogan, K.B., and R.C. Harriss. 1994. Comment on "A dramatic decrease in the growth rate of
atmospheric methane in the Northern Hemisphere during 1992" by E.J. Dlugokencky et al. Geophys.
Res. Lett. 21(22): 2445-2446.
Houweling, S., T. Kaminski, F. Dentener, J. Lelieveld, and M. Heimann. 1999. Inverse modeling of
methane sources and sinks using the adjoint of a global transport model. J. Geophys. Res. 104: 26137-
26160.
Houweling, S., F. Dentener, and J. Lelieveld. 2000. Simulation of preindustrial atmospheric methane to
constrain the global source strength of natural wetlands. J. Geophys. Res. 105: 17243-17255.
Itoh, M., N. Ohte, K. Koba, M. Katsuyama, K. Hayamizu, and M. Tani. 2007. Hydrologic effects on
methane dynamics in riparian wetlands in a temperate forest catchment. J. Geophys. Res. 112: G01029,
10.1029/2006JG000240.
2-22
-------
Chapter 2. Wetlands
Johansson, T., N. Malmer, P.M. Crill, T. Friborg, J.H. Akerman, M. Mastepanov, and T.R. Christensen.
2006. Decadal vegetation changes in a northern peatland, greenhouse gas fluxes, and net radiative
forcing. Global Change Biol 12(12): 2352-2369.
Jorgenson, M.T., Y.L. Shur, and E.R. Pullman. 2006. Abrupt increase in permafrost degradation in Arctic
Alaska. Geophys. Res. Lett. 33: L02503, 10.1029/2005GL024960.
Kaplan, J.O. 2002. Wetlands at the Last Glacial Maximum: Distribution and methane emissions.
Geophys. Res. Lett. 29: 1079, 10.1029/2001GLO13366.
Keller, M., W.A. Kaplan, and S.C. Wofsy. 1986. Emissions of N2O, CFL,, and CO2 from tropical forest
soils. J. Geophys. Res. 91: 11791-11802.
Kettunen, A. 2003. Connecting methane fluxes to vegetation cover and water table fluctuations at
microsite level: A modeling study. Global Biogeochem. Cycles 17: 1051, 10.1029/2002GB001958.
King, J.Y., and W.S. Reeburgh. 2002. A pulse-labeling experiment to determine the contribution of
recent plant photosynthates to net methane emission in arctic wet sedge tundra. Soil Biol. Biochem. 34:
173-180.
King, J.Y., W.S. Reeburgh, K.K. Thieler, G.W. Kling, W.M. Loya, L.C. Johnson, and K.J. Naderhoffer.
2002. Pulse-labeling studies of carbon cycling in arctic tundra ecosystems: The contribution of
photosynthates to methane emissions. Global Biogeochem. Cycles 16:1062-1075.
Koschorreck, M. 2005. Nitrogen turnover in drying sediments of an Amazon floodplain lake. Microb.
Ecol. 49(4): 567-577.
Koschorreck, M., and A. Darwich. 2003. Nitrogen dynamics in seasonally flooded soils in the Amazon
floodplain. Weil. Ecol. Manag. 11: 317-330.
Koyama, T. 1963. Gaseous metabolism in lake sediments and paddy soils and the production of
atmospheric methane and hydrogen. J. Geophys. Res. 68: 3971-3973.
Kreutzwieser, J., J. Buchholz, and H. Rennenberg. 2003. Emission of methane and nitrous oxide by
Australian mangrove ecosystems. Plant Biol. 5: 423-431.
Krithika, K., R. Purvaja, and R. Ramesh. 2008. Fluxes of methane and nitrous oxide from an Indian
mangrove. Curr. Sci. 94: 218-224.
Laiho, R. 2006. Decomposition in peatlands: Reconciling seemingly contrasting results on the impacts of
lowered water levels. Soil Biol. Biochem. 38(8): 2011-2024.
Laine, J., J. Silvola, K. Tolonen, J. Aim, H. Nykanen, H. Vasander, T. Sallantaus, I. Savolainen, J.
Sinisalo, and P.J. Martikainen. 1996. Effect of water-level drawdown on global climatic warming:
Northern peatlands. Ambio. 25(3): 179-184.
Lelieveld, J., P.J. Crutzen, and F.J. Dentener. 1998. Changing concentration, lifetime, and climate forcing
of atmospheric methane. Tellus 50B: 128-150.
Marani, L., and P.C. Avala. 2007. Methane emissions from lakes and floodplains in Pantanal, Brazil.
Atmos. Environ. 41: 1627-1633.
Matthews, E., and I. Fung. 1987. Methane emission from natural wetlands: Global distribution, area, and
environmental characteristics of sources. Global Biogeochem. Cycles 1: 61-86.
McFadden, L., T. Spenser, and R.J. Nicholls. 2007. Broad-scale modeling of coastal wetlands: What is
required? Hydrobiologia 577: 5-15.
2-23
-------
Chapter 2. Wetlands
Megonigal, J.P., S.C. Whalen, D.T. Tissue, B.D. Bovard, D.B. Albert, and A.S. Allen. 1999. A plant-soil-
atmosphere microcosm for tracing radiocarbon from photosynthesis through methanogenesis. Soil Sci.
Soc. Am.J., 63: 665-671.
Melack, J.M., L.L. Hess, M. Gastil, B.R. Forsberg, S.K. Hamilton, I.B.T. Lima, and E.M.L.M. Novo.
2004. Regionalization of methane emissions in the Amazon Basin with microwave remote sensing.
Global Change Biol. 10:530-544.
Middelburg, J.J., G. Klaver, J. Nieuwenhuize, R.M. Markusse, T. Vlug, and F.J.W.A. van der Nat. 1995.
Nitrous oxide emissions from estuarine intertidal sediments. Hydrobiol. 311: 43-55.
Mikaloff Fletcher, S.E., P.P. Tans, L.M. Bruhwiler, J.B. Miller, and M. Heimann. 2004. CIL, sources
estimated from atmospheric observations of CH4 and its 13C/12C isotopic ratios: 1. Inverse modeling of
source processes. Global Biogeochem. Cycles 18: GB4004, 10.1029/2004GB002223.
Monson, R.K., and E.A. Holland. 2001. Biospheric trace gas fluxes and their control over tropospheric
chemistry. Ann. Rev. Ecol. Syst. 32: 547-576.
Moore, T.R., and N. Roulet. 1995. Methane emissions in Canadian peatlands. In: R. Lai et al. (eds.). Soils
and Global Change. Boca Raton, FL: Lewis Publishers, pp. 153-164.
Moorhead, K.K., and M.M. Brinson. 1995. Response of wetlands to rising sea level in the lower coastal
plain of North Carolina. Ecol. Appl. 5: 261-271.
Mosier, A.R., S.K. Mohanty, and A. Bhadrachalam. 1990. Evolution of dinitrogen and nitrous oxide from
the soil to the atmosphere through rice plants. Biol. Fertil of Soils 9: 61-67.
Nicholls, R.J. 2004. Coastal flooding and wetland loss in the 21st century: changes under the SRES
climate and socio-economic scenarios. Global Environ. Change. 14: 69-86.
Oquist, M.G., and B.H. Svensson. 2002. Vascular plants as regulators of methane emissions from a
subarctic mire ecosystem. J. Geophys. Res. 107(21): 4580, 10.1029/2001JDOO1030.
Potter, C., S. Klooster, S. Hiatt, M. Fladeland, V. Genovese, and P. Gross. 2006. Methane emissions from
natural wetlands in the United States: Satellite-derived estimation based on ecosystem carbon cycling,
Earth Interactions 10: paper 22.
Potter, C.S. 1997. An ecosystem simulation model for methane production and emission from wetlands,
Global Biogeochem. Cycles 11(4): 495-506.
Prigent, C., F. Papa, F. Aires, W.B. Rossow, and E. Matthews. 2007. Global inundation dynamics
inferred from multiple satellite observations, 1993-2000. J. Geophys. Res. 112: D12107,
10.1029/2006JD007847.
Regina, K., H. Nykanen, J. Silvola, and P.J. Martikainen. 1996. Fluxes of nitrous oxide from boreal
peatlands as affected by peatland type, water table level, and nitrification capacity. Biogeochem. 35(3):
401-418.
Riodan, B., D. Verbyla, and A.D. McGuire. 2006. Shrinking ponds in subarctic Alaska based on 1950-
2002 remotely sensed images. J. Geophys. Res. Ill: G04002, 10.1029/2005JG0000150.
Roulet, N., T. Moore, J. Bubier, and P. LaFleur. 1992. Northern fens—methane flux and climate change.
Tellus 44B (2): 100-105.
Rusch, H., and H. Rennenberg. 1998. Black alder (Alnus glutinosa (L.) Gaertn.) trees mediate methane
and nitrous oxide emission from the soil to the atmosphere. Plant and Soil 201: 1-7.
Schiller, C.L., and D.R. Hastie. 1994. Exchange of nitrous oxide within the Hudson Bay lowland. J.
Geophys. Res. 99: 1573-1588.
2-24
-------
Chapter 2. Wetlands
Sebacher, D.I., R.C. Harriss, and K.B. Bartlett. 1985. Methane emissions to the atmosphere through
aquatic plants. J. Env. Qual. 14(1): 40-46.
Shindell, D.T., B.P. Walter, and G. Faluvegi. 2004. Impacts of climate change on methane emissions
from wetlands. Geophys. Res. Lett. 31: L21202, 10.1029/2004GL021009.
Strack, M., and J.M. Waddington. 2007. Response of peatland carbon dioxide and methane fluxes to a
water table drawdown experiment. Global Biogeochem. Cycles 21: GB1007, 10.1029/GB002715.
Strom, L., and T.R. Christensen. 2007. Below ground carbon turnover and greenhouse gas exchanges in a
sub-arctic wetland. Soil Biol. Biochem. 39(7): 1689-1698.
Swain, P.M. 1973. Marsh gas from the Atlantic coastal plain, United States. Adv. Org. Geochem. 1: 673-
687.
Turetsky, M.R., R.K. Wieder, and D.H. Vitt. 2002. Boreal peatland C fluxes under varying permafrost
regimes. Soil Biol. Biochem. 34: 907-912.
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Valentine, D.W., E.A. Holland, and D.S. Schimel. 1994. Ecosystem and physiological controls over
methane production in a northern wetland. J. Geophys. Res, 99: 1563-1571.
Venterink, H.O., T.E. Davidsson, K. Kiehl, and L. Leonardson. 2002. Impact of drying and re-wetting on
N, P, and K dynamics in a wetland soil. Plant and Soil 243: 119-130.
Vile, M.A., S.D. Brigham, R.K. Weider, and M. Novak. 2003. Atmospheric sulfur deposition alters
pathways of gaseous carbon production in peatlands. Global Biogeochem. Cycles. 17(2): 1058,
10.1029/2002GB001966.
von Fischer, J.C., and L.O. Hedin. 2007. Controls on soil methane fluxes: Tests of biophysical
mechanisms using stable isotope tracers. Global Biogeochem. Cycles 21: GB2007,
doi: 10.029/2006GB002687.
Walter, B.P., and M. Heimann. 2000. A process-based, climate-sensitive model to derive methane
emissions from natural wetlands: Application to five wetland sites, sensitivity to model parameters, and
climate. Global Biogeochem. Cycles 14: 745-765.
Walter, B.P., M. Heimann, and E. Matthews. 2001. Modeling modern methane emissions from natural
wetlands 1. Model description and results. J. Geophys. Res. 106: 34189-34206.
Walter, K.M., M.E. Edwards, G. Grosse, S.A. Zimov, and F.S. Chapin. 2007. Thermokarst lakes as a
source of atmospheric CH4 during the last deglaciation. Science 318: 633-636.
Wang, J.S., J.A. Logan, M.B. McElroy, B.N. Duncan, LA. Megretskaia, and RM. Yantosca. 2004. A 3-D
model analysis of the slowdown and interannual variability in the methane growth rate from 1988 to
1997. Global Biogeochem. Cycles 18: GB3011, 10.1029/2003GB002180.
Whalen, S.C. 2005. Biogeochemistry of methane exchange between natural wetlands and the atmosphere.
Environ. Eng. Sci. 22(1): 73-94.
Whiting, G.J., and J.P. Chanton. 1992. Plant-dependent CLL emission in a subarctic Canadian fen. Global
Biogeochem. Cycles 6(3): 225-231.
Whiting, G.J., and J.P. Chanton. 1993. Primary production control of methane emission from wetlands.
Nature 364: 794-795.
2-25
-------
Chapter 2. Wetlands
Wickland, K.P., R.G. Striegl, J.C. Neff, and T. Sachs. 2006. Effects of permafrost melting on CO2 and
CH4 exchange of a poorly drained black spruce lowland. J. Geophys. Res. Ill: G02011,
10.1029/2005JG000099.
Woodwell, G.M., F.T. Mackenzie, R.A. Houghton, M. Apps, E. Gorham, and E. Davidson. 1998. Biotic
feedbacks in the warming of the earth. Climate Change 40: 496-518.
Yan, X., S. Shi, L. Du, and G. Xing. 2000. Pathways of N2O emission from rice paddy soil. Soil Biol.
Biochem. 32: 437-440.
Zhang, L.H., C.C. Song, D.X. Wang, Y.Y. Wang, and X.F. Xu. 2007. The variation of methane emission
from freshwater marshes and response to the exogenous N in Sanjiang Plain Northeast China. Atmos.
Environ. 41(19): 4063-4072.
Zhang, Y., C. Li, C.C. Trettin, H. Li, and G. Sun. 2002. An integrated model of soil, hydrology, and
vegetation for carbon dynamics in wetland ecosystems. Global Biogeochem. Cycles 16(4): 1061,
10.1029/GB001838.
Zhuang, Q., J.M. Melillo, D.W. Kicklighter, R.G. Prinn, A.D. McGuire, P.A. Steudler, B.S. Felzer, and S.
Hu. 2004. Methane fluxes between terrestrial ecosystems and the atmosphere at northern high latitudes
during the pas century: A retrospective analysis with a process-based biogeochemistry model. Global
Biogeochem. Cycles 18: GB3010, 10.1029/2004GB002239.
Zhuang, Q., J.M. Melillo, M.C. Sarofirm, D.W. Kicklighter, A.D. McGuire, B.S. Felzer, A. Sokolov,
R.G. Prinn, P.A. Steudler, and S. Hu. 2006. CO2 and CFL, exchanges between land ecosystems and the
atmosphere in northern high latitudes over the 21st century. Geophys. Res. Lett. 33: L17403,
10.1029/2006GL026972.
Zhuang, Q., J.M. Melillo, A.D. McGuire, D.W. Kicklighter, RG. Prinn, P.A. Steudler, B.S. Felzer, and S.
Hu. 2007. Net emissions of CFL, and CO2 in Alaska: Implications for the region's greenhouse gas
budget. Ecol. Appl. 17(1): 203-212.
2-26
-------
Chapter 3. Upland Soils and Riparian Areas
This chapter focuses on upland soils and riparian areas and their contribution to global greenhouse gas
emissions. The original 1993 report (U.S. EPA, 1993) focused solely on the trace gas QrU, and only
considered the potential source strength of the
wetland soils category in contributing to global
CH4 emissions. The report did not discuss upland
and riparian soils as significant sources or sinks of
CH4 and other trace gases such as N2O. which ,suPP°rt f°rests< grasslands, and
& agricultural lands.
In the years since the publication of this 1993 Rjparjan aregs gre vegetated ecosystems
Upland soils are well-aerated soils with lower
moisture content than wetland soils. They are
found throughout the world, and include soils
along a waterbody through which energy,
materials, and water pass. As they form the
interface between terrestrial and aquatic soils,
they often contain characteristics of both dry
and water-saturated soils.
report, the role of upland soils as a sink for global
atmospheric ClrU has become better understood.
This sink strength has been quantified and modeled
under various scenarios, and upland soils have
been shown to provide a significant sink for
atmospheric ClrU.
In addition, there have been profound advances in the understanding of the role of N2O as a greenhouse
gas and of the biogeochemical processes that regulate N2O emissions from both upland and riparian soils.
These breakthroughs have led to refined top-down and bottom-up estimates of global N2O budgets, and
increased understanding of the effects of anthropogenic influences, such as agricultural practices, on
emissions from upland and riparian soils.
In the AR4, the IPCC reported that dry upland soils serve as one of the primary global ClrU sinks by
biologically oxidizing ClrU They estimate that this source accounts for 30 Tg ClrU removed per year from
the atmosphere. The same upland soils have also been shown to release 6.6 Tg N/yr as N2O through the
microbiological processes of nitrification and denitrification (Denman et al., 2007). Although the IPCC
report did not consider riparian zone soils separately, several other recent studies have described their
potential as ClrU sinks and N2O sources.
This chapter reviews the current scientific understanding of upland and riparian soils as sources and sinks
for N2O and QrU. It describes recent findings on biogeochemical processes that regulate emissions of
these greenhouse gases in soils. It also reviews the factors that influence gas fluxes in these soils and their
ultimate role as sources and sinks of greenhouse gases, with particular attention to spatial variations.
Estimates of current global emissions from upland and riparian soils are provided and compared, and
projections of future emissions scenarios and sensitivity to climate disruptions are summarized.
3.1 Description of Emission Source
Upland soils can be distinguished from wetland soils based upon their lower moisture content, which
significantly alters the biogeochemical processes that take place in these soils. Upland soils are well-
aerated, not water-saturated, and generally oxic (Conrad, 1996). The dry conditions favor microbial
processes that make dry upland soils a sink for ClrU and a source of N2O (these processes are described in
Sections 3.1.1 and 3.1.2). Upland soils include forests and grasslands under natural vegetation and arable
lands. Chapter 10 of this report covers the contribution of vegetation growing on these soils to global QrU
budgets; this chapter focuses only on the contribution of the soils themselves.
Riparian zone soils are often permanently wet and rich in organic matter. Riparian soil moisture content is
distributed spatially, with gradients from the hill slope down toward the stream (Hefting et al., 2006).
These saturated conditions and microbially available C contribute to higher rates of production of N2O
3-1
-------
Chapter 3. Upland Soils and Riparian Areas
than in dry upland soils. The degree of soil moisture also determines whether riparian zone soils will
serve as net sources or net sinks of QrU (Hope et al., 2004).
3.1.1 Soils as Nitrous Oxide Sources
Soil N2O production in upland and riparian soils with natural vegetation cover is influenced by various
microbiological, chemical, and physical properties
and processes in the soil. N2O emissions from soils
are produced predominantly by the microbial
processes of nitrification and denitrification.
Nitrification is the main source of N2O under
Globally, about 65 percent of all N2O
emissions arise from nitrification and
denitrification processes in soil (Smith and
Conen, 2004).
aerobic conditions, while denitrification dominates
under anoxic conditions. Both processes can occur simultaneously in soils, and the production of N2O
depends on the balance between these two microbial processes (Butterbach-Bahl et al., 2004; Conrad,
1996).
Nitrification
NO
NO i
' x^"\ !
[HNO],^ x
-------
Chapter 3. Upland Soils and Riparian Areas
3.1.2 Soils as Methane Sinks
Soils can act as either net sources or net sinks
depending primarily upon soil moisture content. ClrU is
produced in waterlogged, wetland soils, but in well-aerated
upland soils, ClrU is oxidized (and therefore consumed).
Upland soils may contain anoxic microsites (small volumes
of soil) where CH4 is produced, but dry upland soils are
overall net sinks of QrU (Castaldi et al., 2006; Del Grosso et
al., 2000; Hein et al., 1997; Jang et al., 2006; Wuebbles and
Hayhoe, 2002). Dry soil oxidation of QrU makes up about 5
percent of the global QrU sink (Wuebbles and Hayhoe, 2002).
One of the most important global biological sinks for ClrU is
forest soils, where methanotrophic bacteria oxidize
atmospheric QrU to CO2 in the presence of O2. In most soils,
atmospheric ClrU consumption seems to be located in the
subsoil, usually between the A and B soil horizons (Conrad,
1996). The concentration of CH4 in soil microsites drives this
CH4 oxidation, along with soil water content and other
physical properties (Del Grosso et al., 2000).
The CH4 consumption potential of upland and riparian soils has been characterized in a variety of
ecosystems and land use categories (Groffman et al., 2006; Hope et al., 2004; Kaye et al., 2005; Werner et
al., 2007b). Most ClrU consumption has been found to occur in the well-drained soils of temperate and
tropical areas (Ridgwell et al., 1999).
Figure 3-2. Typical soil profile. (Source:
U.S. Department of Agriculture.)
2
e
at
18
16
14
12 ••
Tola!
- - - Total Anthropogenic
— - Nature)
3.2 Factors That Influence Emissions
The factors that determine both the magnitude of emissions of N2O from upland and riparian soils and the
sink strength for CH4 are numerous and interrelated. As the carbon and nitrogen cycles in soils are linked,
changes in N and C availability strongly influence
the rate of emission or sequestration.
Both upland and riparian soils have been
significantly impacted by human activity. Many
of the arable lands where dry upland soils occur
have been cleared for agricultural use, which is
responsible for an estimated 80 percent of
anthropogenic emission of N2O through soil
emission, biomass burning, and animal
production (Kroeze, 1999). During the 20th
century, an expansion of agricultural land coupled
with intensification of use of N fertilizer inputs
caused a net increase of global N2O emissions
from 11 Tg N/yr in 1850 to 18 Tg N/yr in 1994
(Kroeze, 1999). Land use change, as forests and
grasslands are converted to agriculture, has also
decreased the global soil ClrU sink (King, 1997).
Arable land has a much smaller QrU uptake rate
than untreated soils, particularly when treated
ill
6
1700 1725 1750 1775 1SDO 1625 1850 1875 1900 1925 1950 1975
Time (year)
Figure 3-3. Global emissions of N2O (1700-1994) used in
the base scenario calculations. (Source: Kroeze, 1999.)
3-3
-------
Chapter 3. Upland Soils and Riparian Areas
with certain fertilizers (Wuebbles and Hayhoe, 2002).
Riparian zones, located at the interface of terrestrial and aquatic environments, have been significantly
impacted by agricultural activities as well. Riparian buffer zones serve as sites for nitrate removal from
agricultural runoff, and are often loaded with high levels of N. Recent studies have suggested that
prolonged exposure to N from agricultural runoff leads to higher N2O emissions compared to emission
rates from non-exposed forest soils (Dhondt et al., 2004; Hefting et al., 2003, 2006; Ullah and Zinati,
2006). The increased N levels in riparian soils may also suppress QrU oxidation processes in these soils.
Clearly, the human-induced changes in N2O emissions and ClrU sink potential of upland and riparian soils
are significant. As this chapter focuses on the natural processes occurring in these soils, however,
anthropogenic effects on these trace gases are not considered.
3.2.1 Factors That Influence Emissions of Nitrous Oxide
One method of identifying factors that influence N2O emissions has been the development of process-
based models such as DNDC (Li et al., 1992a,b, 1994) and DAYCENT (Parton et al., 1996), which
simulate trace gas fluxes from soils. In addition, summaries of emission measurement data from field
studies, using statistical techniques, have led to the development of emission factors such as those used by
the IPCC (Bouwman, 1996).
Field measurements of N2O emissions from soils under natural vegetation cover have been undertaken
worldwide, in a variety of ecosystems and land uses. In a recent paper that reviewed 207 studies of
emissions measurements from soils under natural vegetation, it was found that soil organic C content,
vegetation type, soil pH, bulk density, and drainage were the major factors influencing N2O emissions
(Stehfest and Bouwman, 2006).
As soil organic content increases, emissions of N2O increase as well due to increased availability of C for
denitrifying bacteria (Kanerva et al., 2007; Stehfest and Bouwman, 2006). Vegetation type also influences
emissions, as N2O emissions decrease with increasing plant species diversity, and increase in the presence
of legumes (Niklaus et al., 2006).
Chemical and physical characteristics of soil also influence emissions of N2O. As soil pH increases, N2O
emissions decrease (Stehfest and Bouwman, 2006). As soil bulk density decreases, so do N2O emissions
(Stehfest and Bouwman, 2006). Soil moisture also plays a role in N2O emissions: as Khalil and Baggs
(2005) reported, N2O emissions were the highest from the wettest soils (at 75 percent water-filled pore
spaces). In the same study, it was found that 90 percent of the N2O was produced through
denitrification—proof that these water-filled soil microsites were primarily anaerobic.
The various microbiological, chemical, and physical properties of soil that influence N2O emissions are
distributed throughout upland and riparian soils worldwide. However, some general trends for particular
biomes (ecological communities in particular climates) appear. Specifically, emissions of N2O from
rainforests are significantly higher than from grasslands, savannah, and tropical dry forest and.emissions
from grasslands are significantly lower than those from deciduous forests and rainforests (Stehfest and
Bouwman, 2006). High nitrogen availability, coupled with high moisture content, makes tropical soils
especially likely to emit N2O (Bouwman et al., 2002; Hirsch et al., 2006; Keller et al., 2005).
3.2.2 Factors That Influence Methane Sink Strength
Soil CH4 sink strength depends on oxidation by methanotrophic microbes in the soil, and therefore is
influenced by environmental factors that control this oxidation rate. The primary factor is soil diffusivity,
which controls the amount of QrU transferred into the soil and, therefore, its availability to
methanotrophs. Soil diffusivity is influenced primarily by soil moisture content (King, 1997). Soil
moisture strongly controls the uptake of atmospheric ClrU by limiting the diffusion of QrU into the soil,
3-4
-------
Chapter 3. Upland Soils and Riparian Areas
resulting in a negative correlation between soil moisture and CH4 uptake rates under most non-drought
conditions (Borken et al., 2006). Seasonal changes in soil moisture have also been shown to affect the
exchange of ClrU and, therefore, the soil sink/source strength (McLain, 2006).
CH4 oxidation factors reflect the amount of ClrU converted by the microbes. In a study by Jang et al.
(2006) that reviewed 28 studies of QrU oxidation factors, the main variables shown to control CH4
oxidation rate were soil water content and inorganic N presence or absence. The inhibitory effects of
nitrate on CH4 oxidation in forest soils were reported to range from 10 to 86 percent (Jang et al., 2006).
The influence of temperature on oxidation rate is unclear. One global model suggests that the colder
winter temperatures in the Northern Hemisphere inhibit microbial activity, therefore slowing CH4 uptake
and affecting global budgets (Potter et al., 1996). Site-based measurements suggest that temperature does
affect QrU oxidation rates, with the highest rates shown in warm, dry soils (Price et al., 2003). However,
several models indicate that QrU oxidation in soil is insensitive to temperature increase (Jang et al., 2006;
Ridgwell et al., 1999; Zhuang et al., 2004).
CH4 oxidation also varies spatially, with CH/j oxidation rates shown to be higher in upland soils located
adjacent to CH4 sources such as subalpine wetlands (Wickland et al., 1999). Soil:atmosphere fluxes of
CH4 are also strongly influenced by exposure to an urban land use matrix and atmosphere (Groffman et
al., 2006). Also, CH4 oxidation rates tend to be lower in coniferous forests than in deciduous forests (Jang
et al., 2006).
3.3 Current Global Emissions
Estimating global emissions budgets of trace gases such as N2O and ClrU has thus far relied on three
principal techniques: (1) extrapolation from field measurements, (2) process-based modeling (bottom-up
approach), and (3) inverse modeling (top-down approach). Each of these methods contains uncertainties,
as they depend upon the complex set of interrelated factors detailed above. However, the increasing
numbers of emissions measurements in field studies coupled with refinements of existing models have led
to updated global emissions estimates for both N2O and CH4.
3.3.1 Current Global Emissions of Nitrous Oxide From Soils
The various microbiological, chemical, and physical parameters that determine N2O emissions create
complex interactions that make extrapolating global emissions budgets difficult and uncertain. Also, the
vast majority of studies to date have focused on N2O emissions from agricultural, not natural, soil
sources. However, some global budgets of N2O emissions from natural soils have been established, based
on both top-down and bottom-up estimates (Bouwman et al., 2002; Butterbach-Bahl et al., 2004; Del
Grosso et al., 2005; Galloway et al., 2004; Hirsch et al., 2006; Kroeze, 1999; Stehfest and Bouwman,
2006; Werner et al., 2007a). The AR4 provides a global emissions total for soils under natural
vegetation (including upland and riparian soils) of 6.6 Tg N/yr (with an uncertainty range of 3.3 to
9.0 Tg N/yr) (Denman et al., 2007). This global budget was based on data for the 1990s, provided by key
studies from Bouwman et al. (2002).
Since the publication of the AR4, the number of N2O emissions measurements has been increasing
steadily, allowing for improvements in emission models and budgets. In particular, Stehfest and
Bouwman (2006) revised previous global N2O emissions estimates based on 207 field measurements of
soils under natural vegetation conditions. This study provides the first comprehensive statistical analysis
of published measurement data from N2O emissions measurements of soils under natural conditions.
Table 3-1 summarizes the emissions estimates by ecosystem provided by this statistical approach.
However, this statistical model relies upon incomplete coverage of global vegetation zones and a high
uncertainty in the developed statistical models. Models such as this are useful for site-specific estimates,
but cannot be used to create a global N2O budget (Stehfest and Bouwman, 2006).
3-5
-------
Chapter 3. Upland Soils and Riparian Areas
Table 3-1. N2O Emissions From Soils Under Natural Vegetation
Vegetation Classes
Temperate forest
Open tropical forest
Closed tropical forest
Grassland/steppe
Area
(Mha)
230
1,598
854
2,765
Emissions Estimates
(Tg N20 - N/Yr)
0.147
0.333
1.170
0.403
Source: Stehfest and Bouwman, 2006.
Werner et al. (2007a) recently used the biogeochemical model ForestDNDC-tropicato estimate the global
source strength of tropical rainforest soils between 1991 and 2000, at about 1.34 Tg N/yr (0.88 to 2.37 Tg
N/yr). According to this study, detailed biogeochemical models provide useful methods for global N2O
emissions estimates; however, there are insufficient field measurements and soil and vegetation data for
this model to be applied to other ecosystems (Werner et al., 2007a). Table 3-2 shows the tropical
rainforest emissions data from Werner et al. (2007a), separated by the contribution of different continents.
Kesik et al. (2005) also used the Forest-DNDC model to estimate N2O emissions from European forest
soils from 1990 to 2000 to be between 77.6 and 86.8 kt N/yr (0.07 to 0.086 Tg N/yr).
Table 3-2. Global N2O Emissions From Tropical Rainforest Soils
Continents
South America
Africa
Asia
Central America
Oceania/Australia
Total
Area
(10s km2)
6.026
3.055
1.432
0.310
0.104
10.926
N2O Source Strength
(Tg N/Yr)
0.671 ±0.154
0.344 ± 0.084
0.258 ±0.063
0.051 ±0.011
0.011 ±0.003
1.335 ±0.315
Source: Werner et al., 2007a.
While recent advances have been made in collecting and analyzing emissions measurements (Stehfest and
Bouwman, 2006) and the source strength of tropical soils has been characterized (Werner et al., 2007a),
there still exist a lack of field measurements and significant model uncertainties. Therefore, the estimate
provided by Bouwman et al. (and used by the IPCC) of 6.6 Tg N/yr (3.3 to 9.0 Tg N/yr) stands as the
most comprehensive estimate of global N2O emissions from soils under natural vegetation cover. Tropical
rainforest soils, with emissions of 1.34 Tg N/yr, clearly play a significant role in the global N2O emissions
scenarios. Figure 3-4 below illustrates the global distribution of N2O emissions from natural sources, and
highlights the significance of tropical rainforests to the global N2O budget.
3.3.2 Current Global Sink Estimates for Methane
CFi4 budget estimates developed using process-level (bottom-up) measurement techniques contain
significant uncertainties due to the aggregation of local measurements, taken on short time scales and at
large spatial variability (Mikaloff Fletcher et al., 2004). Inverse modeling is a top-down approach that
uses observations of atmospheric mixing ratios, a model of atmospheric transport, and the spatial
distributions of sources or sinks to estimate magnitudes and match observations in bottom-up estimates
(Mikaloff Fletcher et al., 2004). While inverse modeling also contains uncertainties, the AR4 used atop-
down method to estimate the global sink strength of CFI4, with an uncertainty of + 5 percent constrained
mainly by uncertainty in sink estimates and the choice of CFI4 lifetime used in the mass balance
calculation (Solomon et al., 2007).
3-6
-------
Chapter 3. Upland Soils and Riparian Areas
Figure 3-4. Simulated annual N2O emission rates for natural ecosystems for 1998 land cover. Agricultural
area, regrowth forest, arid climate, and polar climate are excluded. (Source: Stehfest and Bouwman, 2006.)
The IPCC reported an overall CH4 sink strength of soils of 30 Tg CH4/yr (Solomon et al., 2007), a
value that closely reflects estimates made in other top-down studies (Table 3-3).
Table 3-3. Estimates of Global Methane Sink
Study
Potter etal., 1996
Hein etal., 1997
Wang etal., 2004
Mikaloff Fletcher et al.,
2004
Solomon etal., 2007
Data Collection
Period
1982-1994
1983-1989
1988-1997
1998-2000
2000-2004
Global Sink Estimate
(Tg CH/Yr;
17-23
26
34
30
30
One process-level estimate by Ridgwell et al. (1999) calculated the global QrU sink to be 37.8 Tg CHVyr.
In this study, annual ClrU sink rates were calculated for aggregated Holdridge life zones, a set of
characteristic life zone classes as predicted by climate (Table 3-4). While this process-level estimate did
not take into account seasonal moisture fluxes that influence oxidation, the estimate did reveal the
importance of dry tropical forest ecosystems as sites of QrU uptake, representing 28 percent of the global
soil QrU sink.
3-7
-------
Chapter 3. Upland Soils and Riparian Areas
Table 3-4. Methane Consumption by Soils (From Ridgwell et al., 1999)
Aggregated Holdridge Life Zone
(From Leemans, 1992)
Tundra
Cold parklands
Forest tundra
Boreal forest
Cool desert
Steppe
Temperate forest
Hot desert
Chapparal
Warm temperate forest
Tropical semiarid
Tropical dry forest
Tropical seasonal forest
Tropical rain forest
Area
(106km2)
10.5
2.8
8.9
15.2
4.0
7.4
10.0
20.9
5.6
3.2
9.5
14.9
15.1
8.5
Annual Sink
(Ta CH/Yr;
1.05
0.59
1.45
2.99
1.11
2.36
2.98
3.83
2.42
1.24
4.42
7.5
6.64
2.92
In light of the similar estimates achieved by various inverse modeling scenarios, the global budget
provided by the IPCC of 30 Tg CHVyr represents the best estimate of the CH4 sink strength provided by
soils. As indicated by Werner et al., the dry, relatively undisturbed soils of dry and seasonal tropical forest
regions provide a significant sink for
3.4 Future Scenarios of Nitrous Oxide and Methane Fluxes
Future emissions of N2O and CH4 oxidation by soils will depend on the changing human activities on
these soils, as well as on climate patterns that are shifting as a result of global climate change. The
clearing of land for agricultural use has been shown to lead to increased N2O emissions and a decreased
capacity for CFL oxidation. Predictive models of global climate show changed patterns of temperature
and precipitation worldwide. As soil moisture is a key determinant of the microbial processes that
consume or produce N2O and CFL, these shifting climate patterns will likely determine the fluxes of these
greenhouse gases into the future.
3.4.1 Future Emissions of Nitrous Oxide From Soils
The AR4, while not specifically predicting future N2O emissions scenarios, highlights a few studies that
have speculated about future emissions from upland and riparian soils. These studies do not make global
predictions of N2O emissions, but rather use site-specific parameters that underscore the importance of N
supply, temperature, and soil moisture as regulators of N2O emissions (Solomon et al., 2007).
Agriculture remains the single biggest source of anthropogenic N2O (Bouwman et al., 2002). In a model
analysis of major U.S. cropping systems, Del Grosso et al. (2005) found that modern agricultural N2O
emission was more than 2 times that of pre-1940 management and about 6 times that of native vegetation.
In the future, intensification of modern agricultural techniques that use N fertilizers or N-fixing crops
could lead to a further increase in N2O flux (Neill et al., 2005).
Land use changes that involve clearing of forests for agricultural use can also increase N2O flux from
soils. Logging in the Amazon has been shown to increase N2O and NO emissions by 30 to 350 percent
(Keller et al., 2005). If similar perturbations in land management in tropical rainforests continue, then
regional emissions of N2O (as well as CFL) from the Brazilian Amazon could be increased by 5 to 10
percent (Keller et al., 2005). As tropical soils are already the largest natural source of N2O to the
3-8
-------
Chapter 3. Upland Soils and Riparian Areas
atmosphere, any land use changes such as logging or modern agriculture would further increase their
contribution to global N2O emissions.
Future climate changes that affect soil moisture will also affect N2O emissions: as Khalil and Baggs
(2005) pointed out, N2O emissions are highest from the wettest soils. Riparian soils, with their higher soil
moisture content and potentially high N deposition from adjacent agricultural areas, appear to have the
potential to create hotspots for N2O production (Neill et al., 2005). Kesik et al. (2006) used the Forest-
DNDC process model to simulate changes in N2O emissions from EU forests for a climate change
scenario for 2030 through 2039. They predicted a 6 percent decrease in N2O emissions relative to the
1990-2000 period, primarily due to an increase in denitrification with an increase in NO production
relative to N2O.
3.4.2 Future Emissions of Methane From Soils
The AR4 did not predict the future global CFi4 sink strength of upland and riparian soils; however, it did
state that future changes to Earth's climate will influence future CFi4 oxidation in these soils.
Temperature and precipitation shifts that accompany global climate change could substantially affect
global CFi4 stocks, while a doubling of atmospheric CO2 would likely change the sink strength only
marginally (in the range of-1 to +3 Tg CWyr) (Ridgwell et al., 1999). As in the case of N2O,
anthropogenic land use changes will also significantly affect the degree to which the CFi4 sink potential of
upland soils is affected.
Several model studies indicate that CFi4 oxidation in dry, upland soils is relatively insensitive to
temperature increase (Ridgwell et al., 1999; Zhuang et al., 2004). However, any temperature changes that
alter the amount and pattern of precipitation may significantly affect the CFi4 oxidation capacity of soils
(Solomon et al., 2007). As CFi4 oxidation is a function of soil moisture content, if rising temperatures
create drier conditions, CFi4 oxidation rates may increase and provide some negative feedback on the
accumulation of CFi4 in the atmosphere (Del Grosso et al., 2000). In one model based on temperate soils
in New England, results suggest that the extension of snow periods may decrease the annual rate of CH4
oxidation while summer droughts may increase soil CFI4 oxidation rates of temperate forest soils (Borken
et al., 2006).
In addition to the temperature and precipitation effects of climate change, human-induced disturbances to
the CH4 oxidation capacity of soils may also significantly affect the global CHt sink. As forests and
grasslands are converted to agriculture, the soil CFLj sink decreases, and if land use changes continue, the
decrease is likely to continue into the future. Agricultural land has a much smaller CFLt uptake rate than
untreated soils, especially when treated with N fertilizers (Jang et al., 2006; Suwanwaree and Robertson,
2005). Any intensification of agriculture in the dry tropical forest regions will have especially important
effects, due to those regions' significant share of the global CFLt sink (Ridgwell et al., 1999). Even when
converted back to its previous state, agricultural land has a lower oxidation rate than before clearing. This
points to an apparent irreversibility of human impacts on these soils, and has implications for future land
management strategies as a growing population exerts pressure to use more upland and riparian soils for
agriculture (Wuebbles and Hayhoe, 2002).
3.5 Areas for Further Research
The complex interactions of soil microbial, chemical, and physical properties that regulate N2O and CFLj
fluxes are still being detailed. This complexity is compounded by the rapid climate changes that
accompany global warming. The variability among estimates makes it difficult to monitor and model
trace gas emissions. In order to improve emissions scenarios of the trace gases N2O and CFLj, researchers
have called for both increasing numbers of field measurements and refined global emissions models.
3-9
-------
Chapter 3. Upland Soils and Riparian Areas
While field measurements of N2O have increased steadily in the past several years, incomplete coverage
of global vegetation zones remains. Stehfest and Bouwman (2006) point out that far more measurement
data are needed, especially for the dry tropical forest, savanna, tundra, and temperate ecosystems not
affected by N deposition. These measurements should be carried out over prolonged periods, to improve
understanding of the complexity of interactions (Kanerva et al., 2007; Stehfest and Bouwman, 2006). The
lack of detailed field measurements was also cited by Werner et al. (2007a) as a crucial step in narrowing
the uncertainty range of the biogeochemical models used to generate global inventories of N2O emissions.
While advances have been made in constraining estimates of CFL sources and sinks, estimates of global
emissions are still constrained by uncertainties (Solomon et al., 2007). Aggregation of spatially distinct
source estimates to achieve global estimates introduces a source of error (Mikaloff Fletcher et al., 2004),
as does reliance on incomplete and short-term field measurements (Wang et al., 2004). One improvement
recommended by Zhuang et al. (2004) involves combining terrestrial CFL flux data with atmospheric CFL
transport models to more accurately simulate seasonal and interannual fluxes. Also, as seasonally dry
ecosystems have emerged as a significant sink of CFL, more study in these ecosystems is needed (Potter
etal., 1996).
3.6 References
Borken, W., E.A. Davidson, K. Savage, E.T. Sundquist, and P. Steudler. 2006. Effect of summer
throughfall exclusion, summer drought, and winter snow cover on methane fluxes in a temperate forest
soil. SoilBiol andBiochem. 38: 1388-1395.
Bouwman, A.F. 1996. Direct emission of nitrous oxide from agricultural soils. Nutr. Cyc. inAgro. 46: 53-
70.
Bouwman, A.F., L.J.M. Boumans, and N.H. Batjes. 2002. Modeling global annual N2O and NO emissions
from fertilized fields. Global Biogeochem. Cycles 16(4): 1080.
Bremner, J.M. 1997. Sources of nitrous oxide in soils. Nutr. Cyc. inAgro. 49: 7-16.
Butterbach-Bahl, K., M. Kesik, P. Miehle, H. Papen, and C. Li. 2004. Quantifying the regional source
strength of N-trace gases across agricultural and forest ecosystems with process based models. Plant
and Soil 260(1-2): 311-329.
Castaldi, S., A. Ermice, and S. Strumia. 2006. Fluxes of N2O and CFL from soils of savannas and
seasonally-dry ecosystems. J. ofBiogeography 33(3): 401-415.
Chatskikh, D., J. Olesen, J. Berntsen, K. Regina, and S. Yamulki. 2005. Simulation of effects of soils,
climate and management on N2O emission from grasslands. Biogeochem. 76(3): 395-419.
Conrad, R. 1996. Soil microorganisms as controllers of atmospheric trace gases (H2, CO, CFL, OCS, N2O,
andNO). Microbiological Review (Dec.): 609-640.
Del Grosso, S.J., A.R. Mosier, W.J. Parton, and D.S. Ojima. 2005. DAYCENT model analysis of past and
contemporary soil N2O and net greenhouse gas flux for major crops in the USA. Soil & Tillage Res.
83(1): 9-24.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, D.S. Ojima, C.S. Potter, W. Borken, R. Brumme, K.
Butterbach-Bahl, K. D.P.M. Crill, and K.A. Smith. 2000. General CFL oxidation model and
comparisons of CFL oxidation in natural and managed systems. Global Biogeochem. Cycles 14(4): 999-
1019.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
3-10
-------
Chapter 3. Upland Soils and Riparian Areas
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor, and H.L. Miller (eds.) Climate
Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY:
Cambridge University Press.
Dhondt, K., P. Boeckx, G. Hofman, and O. Van Cleemput. 2004. Temporal and spatial patterns of
denitrification enzyme activity and nitrous oxide fluxes in three adjacent vegetated riparian buffer
zones. Biol. Fertil. Soils 40(4): 243-251.
Firestone, M.K., and E.A. Davidson. 1989. Microbiological basis of NO and N2O production and
consumption in soil. In: M.O. Andreae and D.S. Schimel (eds.) Exchange of Trace Gases Between
Terrestrial Ecosystems and the Atmosphere. New York, NY: John Wiley & Sons. pp. 7-21.
Galloway, J.N., F.J. Dentener, D.G. Capone, E.W. Boyer, R.W. Howarth, S.P. Seitzinger, G.P. Asner,
C.C. Cleveland, P.A. Green, E.A. Holland, D.M. Karl, A.F. Michaels, J.H. Porter, A.R. Townsend, and
C.J. Vorosmarty. 2004. Nitrogen cycles: Past, present, and future. Biogeochem. 70(2): 153-226.
Groffman, P.M., R.V. Pouyat, M.L. Cadenasso, W.C. Zipperer, K. Szlavecz, I.D. Yesilonis, L.E. Band,
and G.S. Brush. 2006. Land use context and natural soil controls on plant community composition and
soil nitrogen and carbon dynamics in urban and rural forests. Forest Ecol. &Mgmt. 236: 177-192.
Hefting, M.M., R. Bobbink, and H. deCaluwe. 2003. Nitrous oxide emission and denitrification in
chronically nitrate -loaded riparian buffer zones. J. ofEnv. Qual. 32(4): 1194-1203.
Hefting, M.M., R. Bobbink, and M.P. Janssens. 2006. Spatial variation in denitrification and N2O
emission in relation to nitrate removal efficiency in a n-stressed riparian buffer zone. Ecosystems 9(4):
550-563.
Hein, R., P.J. Crutzen, and M. Heimann. 1997. An inverse modeling approach to investigate the global
atmospheric methane cycle. Global Biogeochem. Cycles 11(1): 43-76.
Hirsch, A.I., A.M. Michalak, L.M. Bruhwiler, W. Peters, E.J. Dlugokencky, and P.P. Tans. 2006. Inverse
modeling estimates of the global nitrous oxide surface flux from 1998-2001. Global Biogeochem.
Cycles 20(1).
Hope, D., S.M. Palmer, M.F. Billet, and J.J.C. Dawson. 2004. Variations in dissolved CO2 and CFL. in a
first-order stream and catchment: An investigation of soil-stream linkages. Hydrol. Process. 18: 3255-
3275.
Jang, I., S. Lee, and J. Hong. 2006. Methane oxidation rates in forest soils and their controlling variables:
A review and a case study in Korea. Ecol. Res. 21(6): 849-854.
Kanerva, T., K. Regina, K. Ramo, K. Ojanpera, and S. Manninen. 2007. Fluxes of N2O, CFL. and CO2 in a
meadow ecosystem exposed to elevated ozone and carbon dioxide for three years. Env. Pollution
145(3): 818-828.
Kaye, J.P., R.L. McCulley, and I.C. Burke. 2005. Carbon fluxes, nitrogen cycling, and soil microbial
comunities in adjacent urban, native and agricultural ecosystems. Global Change Biol. 11: 575-587.
Keller, M., R. Varner, J. Dias, H. Silva, P. Crill, R. C. deOliviera, and G. Asner. 2005. Soil-atmosphere
exchange of nitrous oxide, nitric oxide, methane, and carbon dioxide in logged and undisturbed forest
in the tapajos national forest, brazil. Earth Interactions 9(23): 1.
Kesik, M., N. Bruggemann, R. Forkel, R. Kiese, R. Knoche, C. Li, G. Seufert, D. Simpson, and K.
Butterback-Bahl. 2006. Future scenarios of N2O and NO emissions from European forest soils. J.
Geophy. Res. Ill: G02018, doi:10.1029/2005JG000115.
3-11
-------
Chapter 3. Upland Soils and Riparian Areas
Kesik, M., P. Ambus, R. Baritz, N. Bruggemann, K. Butterbach-Bahl, M. Damm, J. Duyzer, L. Horvath,
R. Kiese, B. Kitzler, A. Leip, C. Li, M. Pihlatie, K. Pilegaard, D. Seufert, D. Simpson, U. Skina, G.
Smiatek, T. Vesala, and S. Zechmeister-Boltenstern. 2005. Inventories of N2O and NO emissions from
European forest soils. Biogeosciences 2: 353-375.
Khahil, M.I., and E.M. Baggs. 2005. CFL oxidation and N2O emissions at varied soil water-filled pore
spaces and headspace CFL concentrations. SoilBiol. & Biochem. 37(10): 1785-1794.
King, G.M. 1997. Responses of atmospheric methane consumption by soils to global climate change.
Global Change Biol 3: 351-362.
Kroeze, C. 1999. Closing the global N2O budget: A retrospective analysis 1500-1994. Global
Biogeochem. Cycles 13(1): 1-8.
Leemans, R. 1992. Global Holdridge life zone classifications. Digital Raster Data on a 0.5-degree
Cartesian orthonormal geodetic (lat/long) 360x720 grid. In: Global Ecosystems Database Version 2.0.
Boulder, CO: NOAA National Geophysical Data Center.
Li, C., S. Frolking, and R.C. Harriss. 1994. Modeling carbon biogeochemistry in agricultural soils. Global
Biogeochem. Cycles. 8: 237-254.
Li C., S. Frolking, and T.A. Frolking. 1992a. A model of nitrous oxide evolution from soil driven by
rainfall events: 1. Model structure and sensitivity. J. Geophy. Res. 97: 9759-9776.
Li C., S. Frolking, and T.A. Frolking. 1992b. A model of nitrous oxide evolution from soil driven by
rainfall events: 2. Model applications. Journal of Geophysical Research 97: 9777-9783.
McLain, J.E.T. 2006. Moisture controls on trace gas fluxes in semiarid riparian soils. Soil Sci. Soc. of Am.
Journal 70(2): 367-377.
Mikaloff Fletcher, S.E., P.P. Tans, L.M. Bruhwiler, J.B. Miller, and M. Heimann. 2004. CFL sources
estimated from atmospheric observations of CFL and its 13c//12c isotopic ratios: 1. Inverse modeling of
source processes. Global Biogeochem. Cycles 18: 1-17.
Neill, C., P.A. Steudler, D.C. Garcia-Montiel, J.M. Melillo, B.J. Feigl, M.C. Piccolo, and C.C. Cerri.
2005. Rates and controls of nitrous oxide and nitric oxide emissions following conversion of forest to
pasture in rondonia. Nutr. Cyc. inAgro. 71(1): 1-15.
Niklaus, P.A., D.A. Wardle, and K.R. Tate. 2006. Effects of plant species diversity and composition on
nitrogen cycling and the trace gas balance of soils. Plant and Soil 282(1-2): 83-98.
Parton, W.J., M. Hartman, D. Ojima, and D. Schimel. 1996. DAYCENT and its land surface submodel:
Description and testing. Global and Planetary Change 19(1-4): 35-48.
Potter, C.S., E.A. Davidson, and L.V. Verchot. 1996. Estimation of global biogeochemical controls and
seasonality in soil methane consumption. Chemosphere 32(11): 2219-2246.
Price, S.J., R.R. Sherlock, P.M. Kelliher, T.M. McSeveny, K.R. Tate, and L.M. Condron. 2003. Pristine
New Zealand forest soil is a strong methane sink. Global Change Biol. 10: 16-26.
Ridgwell, A.J., S.J. Marshall, and K. Gregson. 1999. Consumption of atmospheric methane by soils: A
process-based model. Global Biogeochem. Cycles 13(1): 59-70.
Smith, K.A., and F. Conen. 2004. Impacts of land management on fluxes of trace greenhouse gases. Soil
Use andMgmt. 20(255-263).
Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.).
2007. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and
New York, NY: Cambridge University Press.
3-12
-------
Chapter 3. Upland Soils and Riparian Areas
Stehfest, E., and L. Bouwman. 2006. N2O and NO emission from agricultural fields and soils under
natural vegetation: Summarizing available measurement data and modeling of global annual emissions.
Nutr. Cyc. inAgro. (74): 207-288.
Suwanwaree, P., and G. P. Robertson. 2005. Methane oxidation in forest, successional, and no-till
agricultural ecosystems: Effects of nitrogen and soil disturbance. Soil Sci. Soc. of Am. Journal 69(6):
1722-1729.
Ullah, S., and G.M. Zinati. 2006. Denitrification and nitrous oxide emissions from riparian forests soils
exposed to prolonged nitrogen runoff. Biogeochem. 81(3): 253-267.
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Wang, J.S., J.A. Logan, M.B. McElroy, B.N. Duncan, LA. Megretskaia, and R.M. Yantosca. 2004. A 3-D
model analysis of the slowdown and interannual variability in the methane growth rate from 1988 to
1997. Global Biogeochem. Cycles 18(3).
Werner, C., K. Butterbach-Bahl, E. Haas, T. Hickler, and R. Kiese. 2007a. A global inventory of N2O
emissions from tropical rainforest soils using a detailed biogeochemical model. Global Biogeochem.
Cycles 21(3): 18.
Werner, C., R. Kiese, and K. Butterbach-Bahl. 2007b. Soil-atmosphere exchange of N2O, CLL, and CO2
and controlling environmental factors for tropical rain forest sites in western Kenya. J. Geophys. Res.
112: 1-15.
Wickland, K.P., R.G. Striegel, S.K. Schmidt, and M.A. Mast. 1999. Methane flux in subalpine wetland
and unsaturated soils in the southern rocky mountains. Global Biogeochem. Cycles 13: 101-113.
Wuebbles, D.J., and K. Hayhoe. 2002. Atmospheric methane and global change. Earth Sci. Rev. 57(3-4):
177-210.
Zhuang, Q., J.M. Melillo, D.W. Kicklighter, R.G. Prinn, A.D. McGuire, P.A. Steudler, B.S. Felzer, and S.
Hu. 2004. Methane fluxes between terrestrial ecosystems and the atmosphere at northern high latitudes
during the past century: A retrospective analysis with a process-based biogeochemistry model. Global
Biogeochem. Cycles 18(3).
3-13
-------
Chapter 3. Upland Soils and Riparian Areas
3-14
-------
Chapter 4. Oceans, Estuaries, and Rivers
Oceans cover roughly 71 percent of the Earth's surface and have a surface area of approximately 361
million km2. They are critical to controlling the planet's temperature and serve as both a source and a sink
for a number of atmospheric trace gases. This chapter covers emissions from flowing waters. This
includes open and coastal oceans, estuaries, and rivers. It does not include emissions from lakes, which
are discussed in Chapter 6. Lakes and ponds differ from oceanic, estuarine, and riverine systems in many
ways, but a key difference is the effect of active and continuous water movement on biological, chemical,
and physical characteristics.
Earth's interconnected waters form a gradient from freshwater rivers, to estuaries where fresh and salt
waters mix, through the relatively shallow coastal ocean on the continental shelves, to the deepwater,
open ocean. These environments increase in water depth and salinity as distance from shore increases.
Nearshore waters are generally shallow and can be well mixed throughout the water column. Sediments
and shallow waters are often linked, and these nearshore waters are often heavily impacted by the
adjacent land. In deeper water, the water column is often stratified by sharp changes in temperature called
thermoclines. These serve to separate shallow and deep waters, which may be only loosely linked. Direct
land impacts decrease with distance from the coast. In this chapter, emissions from all of these
environments will be assessed, although the primary focus will be on the full-salinity coastal and open
ocean. The oceans are believed to be one of the major natural sources of N2O to the atmosphere. In the
most recent IPCC assessment (the AR4: Denman et al., 2007), natural emissions from oceans were
estimated at 3.8 (1.8 to 5.8) Tg N/yr. Oceans, estuaries, and rivers are thought to be a relatively minor
natural source of CFL and are not explicitly estimated in the 2007 IPCC report or in the earlier version of
this report (U.S. EPA, 1993).
4.1 Description of Emission Source
N2O in aquatic environments is produced by microbial communities using the processes of nitrification
and denitrification, through a series of complex and interacting pathways. N2O is produced in both the
water column and in sediments. CFL is also produced in both sediments and the water column, although
the relative importance of the sediments is much greater. The environments of open ocean, coastal ocean
or continental shelf, estuaries, and rivers differ in the quantity and source of their organic inputs, nutrient
inputs, water depth (and therefore the amount of interaction between sediments and the water column),
mixing dynamics, and salinity. Because of these differences, their emissions of trace gases can be very
different.
Fluxes from the deep open ocean come from the water column, not the sediments. Deep ocean sediments,
because of their great depth and low organic matter inputs, are thought to have little impact on N2O and
CFL. budgets. The continental shelves occupy a much smaller area than the open ocean, but their
emissions per area are greater. Estuaries and rivers typically have higher organic inputs and nutrient levels
than the oceans. Because they are relatively shallow, mixing is active and commonly extends throughout
the entire water column. This mixing can transport gases produced in the sediments into near-surface
water where they can be released to the atmosphere.
The most recent IPCC assessment of greenhouse gas sources estimates that oceans, coastal zones,
estuaries, and rivers release roughly 5.5 Tg N/yr. Oceans (3.8 Tg N/yr) are categorized as a natural
source, while the other aquatic environments (releasing 1.7 Tg N/yr) are classified as anthropogenically
controlled (Denman et al., 2007). The oceans and upland soils under natural vegetation are the major
natural sources of N2O to the atmosphere. At 3.8 Tg N/yr, oceans are thought to contribute roughly 21
percent of global emissions. The AR4 does not explicitly estimate the flux of CFL from oceanic,
estuarine, and riverine environments. However, it cites literature estimates that range from 4 to 15 Tg
4-1
-------
Chapter 4. Oceans, Estuaries, and Rivers
CWyr and does not treat the various environments separately. If the aquatic source of ClrU is on the order
of 4 to 15 Tg CHVyr, it would contribute roughly 1 to 3 percent of the global source of CH4.
4.2 Factors That Influence Emissions
This chapter discusses aquatic N2O and CH4 emissions separately, because of differences in the
importance of their relative contributions as well as in the factors that control fluxes. It also treats the
environments of open ocean, coastal ocean (continental shelf), estuaries, and rivers separately.
4.2.1 Nitrous Oxide
Nitrification and denitrification are often closely coupled in aquatic systems, where denitrification may be
limited by the rate of production of nitrate (NOs) through nitrification (Capone, 1991; Suntharalingam
and Sarmiento, 2000). N2O production and consumption at low oxygen levels is complex. Concentrations
may be low in anoxic waters due to denitrifiers using the gas for respiration, but may be higher in suboxic
waters or on the periphery of anoxic areas, suggesting
production in these environments. In a summary analysis
integrating 136 published reports of denitrification rates,
Pina-Ochoa and Alvarez-Cobelas (2006) found that annual
rates of denitrification were highest in lakes, followed by
rivers, coastal ecosystems, and estuaries. Rates were
highly correlated with NOs levels, which explained 70 percent of the observed variability, and were
inversely correlated with O2 levels.
Oxic means "containing oxygen."
Anoxic means "without oxygen."
Suboxic means "oxygen deficient" and
may describe the transition zone
between the two extremes.
Reactive N refers to forms of N that can be used by
bacteria. It includes inorganic reduced forms (ammonia
{NH3] and ammonium [NH4]), inorganic oxidized forms
(nitrite [NO2], nitrate [NO3], nitric acid [HNO3], nitric oxide
[NO], and nitrous oxide [N2O]), and organic forms (urea,
amines, and proteins). Nitrogen gas (N2) is the most
common form of N but is chemically inert and cannot be
used by bacteria.
Nitrogen is a critical nutrient for plant
growth, and productivity in many
environments is controlled by its
availability. Adding usable N increases
crop yields, and man's use of fertilizers
has resulted in changes in the N cycle on
a global scale. It is estimated that
anthropogenic sources of reactive N
increased by an order of magnitude from
the late 19m century to the early 1990s, and that human sources now make up 40 percent of the global
total (Galloway et al., 2004). Much of the N added to land surfaces and cycled through human food and
energy production is carried away from its original site of introduction. This may occur, for example,
through volatilization and runoff. Green et al. (2004) calculate that the total global N flux from river
basins has doubled since the pre-industrial period as a result of losses from the land. Understanding the
fate of N added to the land has been an active area of research for many years, for economic as well as
environmental reasons. The
anthropogenic changes in the N cycle
have implications for aquatic emissions
of N2O, since rates of nitrification and
denitrification depend on reactive N. The
anthropogenic influence on emissions
decreases with distance offshore and
impacts the approaches used to estimate
fluxes.
Gases dissolved in water bodies will come into balance
with their concentration in the air above. The balance (or
equilibrium) concentration depends on temperature,
salinity, and the how soluble the gas is in water. If a
dissolved concentration is higher than the equilibrium
concentration, it is termed "supersaturated." Gas will be
lost to the air, and thus the water will be a source of the
gas. If it is lower than the equilibrium concentration, it is
"undersaturated." It will absorb gas from the air, and thus
will act as a sink to the atmosphere.
Early studies of oceanic N2O found
widespread supersaturations of the gas,
which varied seasonally. This suggested that oceans could be an important atmospheric source (Nevison
4-2
-------
Chapter 4. Oceans, Estuaries, and Rivers
et al., 1995). N2O production in the surface layer is thought to be small because oxygen inhibits
denitrifiers and light inhibits nitrifiers, so subsurface production and then transport to the surface is
believed to occur. In the open ocean, high apparent oxygen utilization (AOU, a measure of aerobic
decomposition) in the 100 meters below the near-
AOL) (apparent oxygen utilization) is the
difference between a measured dissolved O7 correlated with both N2O production and
concentration and that expected when at
atmospheric equilibrium saturation. It is therefore a . . . ... XT „ , . , . .,,
measure of the sum of the biological activity that consistent with N2O production by nitrifiers.
has occurred since last surface exposure.
surface mixed layer was found to be highly
elevated surface concentrations. This is
An oceanic gyre is a relatively stationary region of
the open ocean with a circular current created by
the Coriolis effect. Gyres are permanent large-
scale water circulation features whose circulation
tends to isolate them from the rest of the ocean.
Sampling suggests that high dissolved levels are
seen in biologically productive upwelling regions
along the eastern margins of ocean basins where low oxygen areas are also seen. Concentrations closer to
atmospheric equilibrium or even small undersaturation are seen in the oceanic gyres (Nevison et al., 2003;
Suntharalingam and Sarmiento, 2000). Making up less than 1 percent of the ocean's volume, low-oxygen
areas are located primarily in the northeast tropical Pacific, the northern Arabian Sea, and off the coast of
Peru. Based on model calculations, Suntharalingam et al. (2000) suggest that these regions could make up
roughly 25 percent of the total open ocean source of N2O.
Concentrations of N2O are also high relative to atmospheric equilibrium in coastal and estuarine waters
and in general have been found to be inversely correlated with salinity. This suggests a land impact and
then losses with mixing (Bange et al., 1996).
Measurements indicate that N exported by rivers
to estuarine systems is much less than the N put
into rivers, meaning that it is transformed within
the river system. Most of this N is lost by
denitrification or burial. A range of
denitrification rates in rivers have been
measured. Lowest rates are generally in unpolluted systems and highest rates are where pollution inputs
are significant (Seitzinger et al., 2006). The fraction of N removed by denitrification may be affected by
river geometry (length and depth), flow rate and water residence time, oxygen content, sediment organic
content, and season/temperature (Seitzinger and Kroeze, 1998). In oxygenated water, ammonium is
quickly used by nitrifying bacteria. Nitrification rates have been found to be influenced by ammonium
concentrations, temperature, oxygen, suspended particulate material, and light (Seitzinger and Kroeze,
1998).
4.2.2 Methane
Sampling conducted in the 1960s and 1970s found that surface aquatic waters in general were commonly
supersaturated in QrU with respect to atmospheric equilibrium, indicating that they were likely small
sources to the atmosphere. Since surface waters are highly aerobic, it was thought unlikely that these
levels were generated in place. Possible sources of oceanic ClrU include horizontal transport, diffusion
from sediments, in situ production, and, in the coastal ocean, submarine groundwater discharge. Recently,
Karl et al. (2008) have also proposed a possible
aerobic metabolic pathway that produces ClrU from
an organic, phosphorous-containing compound. In
the open ocean, depth profiles frequently reveal a
subsurface maximum, from which the gas is mixed
to the surface and released (Holmes et al., 2000).
A pycnocline is a water layer with a large
usually called a thermocline. Mixing is
impeded across such a layer.
Oremland (1979) hypothesized that this QrU was produced in the anaerobic guts of zooplankton and fish.
Current thought is that the subsurface maxima occur at the pycnocline, and that CH4 is produced from the
suspended organic particles that accumulate there (Holmes et al., 2000). Budget calculations suggest that
flux across the air-water interface is also from production within anaerobic micro-environments within
4-3
-------
Chapter 4. Oceans, Estuaries, and Rivers
particles in the near-surface. This production was found to be independent of diurnal cycles, indicating
that it is not related to photosynthesis, grazing, and the vertical migrations of zooplankton (Holmes et al.,
2000). Mixing and loss across the air-water interface is calculated to be the major sink of QrU in
seawater. A model that allowed particles to settle and decompose according to simple mixing/diffusion
equations was able to reproduce oceanic QrU profiles (Nihous and Masutani, 2006). This suggests that
these processes did a good job describing the system and that these particles were the primary QrU
source.
Upwelling areas, with higher rates of organic material falling to depth and lower O2 levels, are sites of
enhanced ClrU production and flux (Sansone et al., 2001; Rehder et al., 2002). Rapid transport to the air-
water interface also decreases time in the oxygenated water column and therefore reduces oxidation.
In shallower environments closer to continents, , ,, , , , ,
... ., . . ., ,. . . ., . Upwelling refers to a pattern of coastal and
sedimentary methanogenesis is thought to contribute
_ . , , ,. open water oceanic circulation. It is created
to water column concentrations. Continental shelf by persistent winds blowing across the
waters may also receive inputs from hydrocarbon
seeps and from hydrate reservoirs. These sources are
discussed in Chapters 7 and 8. The relative
importance of submarine groundwater discharge
varies significantly, but in some areas it has been
found to account for 83 to 99 percent of the total
input to the water column (Bugna et al., 1996).
Although its global significance is largely unknown
,.,,..,,. , j + j occurs along some coasts as well as along
and the driving factors poorly understood, equatoryRegions of coasta| upwe||ingy
ocean surface. As winds move surface
waters, they are replaced by deeper waters
that are richer in nutrients and can support
increased phytoplankton growth.
Phytoplankton in turn support higher
populations offish and other consumers,
making these areas some of the most
productive fisheries in the world. Upwelling
include coastal Peru and Chile, the Arabian
Sea, western South Africa, eastern New
Zealand, and the coast of California.
groundwater discharge into coastal waters may rival
riverine inputs in many systems and serves as a
transport mechanism delivering nutrients and other
dissolved compounds such as ClrU offshore (Kim and
Hwang, 2002; Slomp and Van Cappellen, 2004). Bange et al. (1994) and Rehder and Suess (2001)
suggest that continental shelves may contribute the majority of oceanic CH4 sources because they receive
high-CH4 waters from the continents and are areas of active mixing.
The distribution of CH4 in some estuaries appears to be controlled by inputs from rivers and simple
mixing between river waters and lower-CH4 seawater (de Angelis and Lilley, 1987). Mixing behavior is
more complex in other estuaries, with production, oxidation, and de-gassing occurring (de Angelis and
Scranton, 1993; Abril and Iverson, 2002; Sansone et al., 1999). Middelburg et al. (2002) found that CH4
was only partially correlated with salinity in well-mixed, long-residence-time European estuaries. In these
systems, CH4 initially decreased as salinity rose, then increased to a maximum at medium to high
salinities before decreasing again offshore. In river-dominated systems with relatively short residence
times, there was little correlation between ClrU and salinity.
4.3 Current Global Emissions
This section discusses N2O, then CH4 emissions. For N2O, emissions from rivers and estuaries, the
continental shelves, upwelling regions, and the open oceans will be estimated separately. Emissions from
all of the environments are summarized in Section 4.3.1.4.
4.3.1 Current Ocean, Estuarine, and Riverine Nitrous Oxide Fluxes
Global estimates of N2O fluxes have generally been made through two approaches: (1) gas transfer
calculations that combine measurements of near-surface concentrations and wind speed through gas
transfer coefficients and (2) calculations based on organic matter decomposition, using the "yield" of N2O
as a fraction of nitrification and denitrification or nitrate. Anthropogenic sources of N dominate natural
4-4
-------
Chapter 4. Oceans, Estuaries, and Rivers
sources in nearshore environments, so efforts to estimate emissions from these areas have used the well-
developed literature on anthropogenic N export to rivers to calculate cycling through nitrification and
denitrification and therefore N2O production. Both of these approaches are considered "bottom up" and
use a combination of inventories, measurements, and emission coefficients to calculate fluxes. Recently,
the expanding database on atmospheric N2O mixing ratios and the development of inverse modeling
techniques has permitted the use of this "top-down" approach to constrain the distribution and magnitude
effluxes. None of the estimates of emissions from aquatic environments are based on actual flux
measurements using chambers or eddy correlation techniques.
4.3.1.1 River and Estuarine Nitrous Oxide Fluxes
Although rivers and estuaries are not large areas globally, they are active sites for aquatic productivity
and biogeochemical cycling. Flux estimates are relatively high in comparison to those from the open
ocean, which led to the suggestion that oceanic fluxes were underestimates because they did not include
coastal, estuarine, or riverine emissions (Bange et al., 1996; Bange, 2006a). Although shallow water
emissions are relatively high (Table 4-1) as a result of the impact of humans on the N cycle, only a
modest fraction of this N2O arises from natural N sources (Seitzinger et al., 2000). For this reason,
although these river and estuarine fluxes were newly added to global sources in the AR4 (Denman et al.,
2007), they were classified as anthropogenic, rather than natural sources.
emissions, Tg N yr1
S 1
Oceanic N2O f
•
-V • .
-r- ^»^
• • A •
9
•A
(A)
Year of publication
Figure 4-1. Global N2O ocean emission estimates. Red points include coastal fluxes; black triangles are IPCC
estimates. Horizontal lines mark the 2001 IPCC mean and standard deviation. From Bange, 2006b.
The first estimates of these river and estuarine fluxes were based on surface dissolved concentrations and
gas-transfer calculations (Bange et al., 1996). More recent estimates are based on estimating N inputs and
calculating the fraction released as N2O (Seitzinger and Kroeze, 1998; Kroeze and Seitzinger, 1998;
Kroeze et al., 2005). The basic assumption for these calculations is that N2O production is related to rates
of nitrification and denitrification, which are in turn related to inputs of N. Published estimates of aquatic
N inputs have been based both on inventories (Seitzinger and Kroeze, 1998; Kroeze and Seitzinger, 1998;
Seitzinger et al., 2000) as well as models (Boyer et al., 2006; Dumont et al., 2005; Schaefer and Alber,
2007). However, the actual production of N2O from nitrification and denitrification has been an assumed
constant fraction of the N2 produced, 0.3 percent N2O:N2 for denitrification in rivers and estuaries with
low N loadings (0 to 10 kg N/ha/yr) and 3 percent for those with high N inputs (> 10 kg N/ha/yr)
(Seitzinger and Kroeze, 1998). The same constants were used for N2O yield from nitrification, largely
because there were few data available. Seitzinger et al. (2000) and Seitzinger (1988) indicate that
laboratory and field studies have estimated the ratio of N2O to N2 generally between 0.1 and 0.5 percent,
with values ranging up to 6 percent in highly polluted sediments, although the database is quite small.
4-5
-------
Chapter 4. Oceans, Estuaries, and Rivers
Clearly this is a critical but poorly known value in calculating emissions and depends upon a suite of
biogeochemical variables (Schlesinger et al., 2006). The importance of this ratio may be demonstrated by
comparing global estimates made by Capone (1991), who assumed different yield values—0.3 percent for
nitrification under aerobic conditions, 1 percent for nitrification linked to denitrification in low-oxygen
waters and sediments, and 5 percent for denitrification. Although estimates of N aquatic reservoirs are
also derived differently, Capone's nearshore/estuarine N2O flux of 0.74 Tg N/yr is appreciably higher
than that made by the Seitzinger/Kroeze research group (0.22 Tg N/yr, see Table 4-1).
Although natural aquatic N inputs were known to vary geographically, Seitzinger and Kroeze (1998)
calculate natural N contributions by assuming a globally constant natural N input rate. They use a model
to calculate anthropogenic sources (human sewage, fertilizer, and atmospheric deposition). The model
does not explicitly include either agricultural or non-agricultural soil N fixation inputs to watersheds,
although these were known to be large. They calculate that natural river and estuarine DIN (dissolved
inorganic nitrogen, often the most abundant and available N form) makes up about 25 percent of total
inputs. Recent calculations by Boyer et al. (2006) suggest a significantly different breakdown. This more
sophisticated model also finds important regional differences in natural N sources. Boyer et al. (2006)
estimate natural N exports to rivers based on calculations of biological N fixation in forests and other
natural vegetation with minor inputs from N fixation by lightning. They find that natural sources
dominate riverine N inputs in Africa, Latin America, and Oceania. Anthropogenic sources dominate in
Europe and the former Soviet Union, North America, and Asia. These calculations indicate roughly equal
N export from natural and anthropogenic sources rather than the 25:75 percent breakdown of Seitzinger
and Kroeze. Overall, anthropogenic sources are concentrated in the Northern Hemisphere, but as land use
changes and food and energy production patterns continue to change, this geographic distribution will be
altered. Asian sources, with high inputs and a growing population and economic base, were found to drive
the overall global N budget.
Dumont et al. (2005) also describe a more sophisticated and spatially explicit model based on the work of
Seitzinger and Kroeze (1998). Although modeling DIN rather than total N, they also find that natural
biological N fixation is the dominant N source over large areas. Globally, the model calculates that
natural N sources contribute 36 percent of total exports to rivers; the authors hypothesize that this fraction
may be somewhat lower than that for total N because natural systems export N primarily in forms other
than DIN.
These more sophisticated models suggest that the IPCC AR4 classification of all riverine and estuarine
N2O as wholly anthropogenic may be an oversimplification. Given a conservative assumption that 41
percent of the N exported has natural origins (an average of the Boyer and Dumont models) rather than 25
percent, by ratio riverine N2O from natural sources would be 8 percent of the total rather than 5 percent
and estuarine N2O would be 15 rather than 9 percent. Natural N2O fluxes are then calculated to be on the
order of 0.09 (rivers) to 0.24 (estuaries) Tg N/yr, for a total of 0.33 Tg N/yr (Table 4-1). This estimate is
about 19 percent of the AR4 estimate from rivers, estuaries, and coastal environments, assumed to have
wholly anthropogenic sources (1.7 Tg N/yr, Table 4-1). It is on the same order, however, as N2O sources
such as fossil fuel combustion and industrial processes, biomass burning, human excreta, and
anthropogenic atmospheric deposition (Denman et al., 2007).
4-6
-------
Chapter 4. Oceans, Estuaries, and Rivers
Table 4-1. River and Estuarine Nitrous Oxide Emissions
(Tg N/Year)
Habitat
Rivers
Rivers
Rivers
Rivers
Nearshore/estuaries
Nearshore/estuaries
Estuaries
Estuaries
Estuaries
Estuaries
Rivers, estuaries,
and coast
Rivers and
estuaries
Flux (Range)
1990s
1.05
(0.19-1.87)
0.05
1.26
0.09
(0.08-0.10)
0.74
3.0
(2.34-3.63)
0.22
(0.07-0.69)
0.02
0.25
0.24
(0.03-0.45)
1.7
(0.5-2.9)
0.33
(0.11-0.55)
Source
Seitzinger and
Kroeze, 1998;
Seitzinger etal., 2000
Galloway etal., 2004
Kroeze etal., 2005
Capone, 1991
Bange etal., 1996
Seitzinger and
Kroeze, 1998;
Seitzinger etal., 2000
Galloway etal., 2004
Kroeze etal., 2005
Denman etal., 2007
(AR4)
Comment
Only 5% (0.05 Tg N/yr) considered
natural; range based on use of low
and high emission factors
Natural only; cites Seitzinger and
Kroeze, 1998; Kroeze and
Seitzinger, 1998; Seitzinger et al.,
2000, as basis
Cited as basis for AR4; based on N
export model; mix of natural and
anthropogenic
8% of range of total estimates,
midpoint (Boyer et al., 2006;
Dumont et al., 2005)
Estimated from N2O released during
nitrification and denitrification
Midpoint of range; mix of natural
and anthropogenic
Only 9% (0.02 Tg N/yr) considered
natural; range based on use of low
and high emission factors
Natural only; cites Seitzinger and
Kroeze, 1998; Kroeze and
Seitzinger, 1998; Seitzinger et al.,
2000, as basis
Cited as basis for AR4; based on N
export model; mix of natural and
anthropogenic
15% of range of total estimates,
midpoint (Boyer et al., 2006;
Dumont et al., 2005)
Considered all anthropogenic; new
source added since the IPCC Third
Assessment Report; cites Kroeze et
al., 2005; Nevison et al., 2004
(upwelling)
4.3.1.2 Coastal Ocean Nitrous Oxide Fluxes
The coastal ocean bordering the continents can be broken into two systems of importance for N2O fluxes:
the continental shelves and upwelling regions. Making up roughly 7 percent of the ocean's area, the
continental shelves are commonly defined as the relatively flat regions, less than 150 to 200 meters deep,
adjacent to continents. All are sites of greater primary productivity, biological activity, and sedimentation
than the deeper open ocean. Upwelling areas, where nutrient-rich deep water is brought to the ocean
surface, are usually located along the eastern margins of ocean basins. These regions support enhanced
primary productivity, which in turn supports greater populations offish and other consumers. Regions of
4-7
-------
Chapter 4. Oceans, Estuaries, and Rivers
upwelling include coastal Peru, Chile, the Arabian Sea, western South Africa, eastern New Zealand, and
the California coast.
Continental Shelves
The continental shelves receive riverine and estuarine drainage and are therefore impacted by
anthropogenic activities. Most of the N input to rivers and estuaries, however, is not exported to coastal
waters. Although the amount exported from rivers to estuaries varies, it is commonly estimated at about
25 percent globally and can vary from roughly 8 to 40 percent (Boyer et al., 2006; Schaefer and Alber,
2007). Export from estuaries is dependent on a number of variables such as estuarine water residence
time, extent of intertidal area, and sediment O2 consumption. Seitzinger and Kroeze (1998) estimate that
perhaps 50 percent of external DIN inputs to estuaries are removed by denitrification. These
approximations would suggest that perhaps 10 to 50 percent of the N input to rivers and estuaries may
reach coastal waters. The continental shelves also receive N via atmospheric deposition and from the
adjacent continental slope waters.
Because terrestrial N exports are not the major driver of N2O production on the continental shelves,
emissions estimates for these areas have been made using different methodologies than those for rivers
and estuaries. Seitzinger and Giblin (1996) report a model correlating rates of shelf denitrification with
phytoplankton primary productivity through rates of sediment O2 uptake. Predicted rates varied with
latitude, with higher rates at lower latitudes. Grouping latitudes into three classes, Seitzinger and Kroeze
(1998) use this relationship to estimate average global denitrification rates and then apply the constant
N2O:N2 fraction for river and estuarine lower N loading to calculate emissions. Rates of nitrification are
estimated differently, by assuming an average shelf depth and constant nitrification rates from 25 to 50
meters deep and below 50 meters. The emission of N2O, therefore, is calculated as a simple function of
the area of sediment and water surface multiplied through constant rates. Natural N sources to the
continental shelf are estimated by assuming that rivers and estuaries supply 40 to 50 percent of total N
inputs (50 percent of which are anthropogenic in origin). The balance is assumed to be from the open
ocean and considered natural in origin (Seitzinger and Kroeze, 1998). Natural sources would therefore
make up 75 percent of the total. Note that Seitzinger and Kroeze conservatively assume that only 50
percent of river and estuarine N inputs to continental shelves were of anthropogenic origin, rather than 75
percent as their model had suggested (see discussion above). This conservative figure would be in better
agreement with later modeling results. Other estimates of continental shelf N2O flux are based on surface
dissolved concentrations and gas-transfer calculations (Bange et al., 1996; Rhee et al., 2009) and
nitrification/denitrification yields and oceanic reservoir estimates (Capone, 1991). Although some of
these flux estimates tend to be higher than those of Seitzinger and Kroeze (Table 4-2), all are highly
uncertain. The Capone (1991) estimate includes both coastal and upwelling systems. The most recent
estimate is that of Rhee et al. (2009), based on samples from the Atlantic basin. They find appreciably
lower levels of N2O in coastal surface waters than Bange and co-workers, with calculated fluxes similar
to those of Seitzinger and Kroeze. If the fraction of natural N sources is assumed to be 75 percent as in
Seitzinger and Kroeze (1998) and in agreement with the more recent Boyer and Dumont model results,
using all four estimates for the continental shelf results in a range of estimates of 0.37 to 3.52 Tg N/yr,
differing by about an order of magnitude. The midpoint of this range is a flux of 1.5 Tg N/yr (Table 4-2).
Upwelling Regions
Sampling of oceanic dissolved N2O concentrations has found unusually high levels of supersaturation in
the vicinity of upwelling areas. Since N2O is largely produced in subsurface waters due to light inhibition,
upwelling provides a rapid way to the surface, where it degasses. Areas of upwelling are also regions of
enhanced primary productivity, which results in higher fluxes of organic material sinking into mid-waters.
In turn, this organic input depletes O2 levels and creates conditions favorable for denitrifiers as well as
nitrifiers.
4-8
-------
Chapter 4. Oceans, Estuaries, and Rivers
Table 4-2. Coastal Ocean Nitrous Oxide Fluxes
(Tg N/Year)
Habitat
Coast/upwelling
Coastal ocean and
marginal seas
Continental shelves
Continental shelves
Coastal ocean
Continental
shelves
Upwelling
Upwelling
Upwelling
Upwelling
Upwelling
Coastal ocean
(continental shelf
and upwelling)
Flux
(Range)
4.7
2.22
(1.74-2.71)
0.64
(0.64-6.43)
0.4
0.37
(0.23-0.51)
1.5
(0.37-3.52)
0.26
(0.2-0.32)
1.0
0.2
(0.06-0.34)
0.003
(0.002-0.003)
0.37
(0.0003-1.0)
1.87
(0.37-4.52)
Source
Capone, 1991
Bangeetal., 1996
Seitzinger and Kroeze,
1998; Seitzinger etal.,
2000
Galloway et al., 2004
Rheeetal., 2009
Bangeetal., 1996
Suntharalingam et al.,
2000
Nevison etal., 2004
Rheeetal., 2009
Comment
Estimated from N2O released
during nitrification and
denitrification
Midpoint of range; mix of
natural and anthropogenic
Only 75% (0.48 Tg N/yr)
considered natural; range
based on use of low and high
emission factors
Natural only; cites Seitzinger
and Kroeze, 1998; Kroeze and
Seitzinger, 1998; Seitzinger et
al., 2000, as basis
Natural (75% of original figures
of 0. 3 1-0. 68 Tg N/yr)
75% of range of total
estimates (Seitzinger et al.,
2000; Boyer etal., 2006;
Dumont et al., 2005)
Midpoint of range
Model for low-O2 zones; given
as model best fit (25% of fixed
ocean flux of 3.6 Tg N/yr)
Range given as V 70%;
classified as anthropogenic in
AR4
Uses smaller area for
extrapolation
Average of reported fluxes
Although the areas where strong upwelling occurs are relatively well known, the intensity of the process
depends upon wind speed, velocity, and duration and can therefore be quite episodic. This also
complicates both sampling and modeling efforts, since air-sea flux calculations commonly employ long-
term means of wind speeds.
Regional estimates suggest that areas of upwelling contribute significantly to the global oceanic source.
Law and Owens (1990) calculate that upwelling in the northwest Indian Ocean, an area covering only
0.43 percent of the global ocean, could contribute 5 to 18 percent of the total oceanic flux. Measurements
in the Arabian Sea over several years suggest that conservative estimates of annual flux range from 0.35
to 0.48 Tg N/yr, on the order of 13 to 17 percent of oceanic flux (Lai and Patra, 1998). Sampling over
only a 2-month period in the Somali Basin, De Wilde and Helder (1997) calculate that even over this brief
time, emissions were 0.4 to 0.8 percent of the oceanic total from an area < 0.011 percent of the world
4-9
-------
Chapter 4. Oceans, Estuaries, and Rivers
ocean surface. Upstill-Goddard et al. (1999) calculate a 6-month N2O flux of 0.26 to 0.48 Tg N during the
southwest monsoon and inter-monsoon transition in the northwestern Indian Ocean based on in situ wind
speeds and surface N2O levels. Their measurements suggest that seasonal changes in wind speed are the
dominant control on air-sea exchange. Bange et al. (2001) summarize N2O measurements from the
Arabian Sea from 1977 to 1997; emissions ranged from 0.21 to 0.44 Tg N/yr and were dominated by
fluxes during monsoonal periods. In addition, atmospheric N2O sampling at the coast of California has
captured enhanced N2O levels for periods lasting for several days that can be traced to upwelling (Lueker
et al., 2003). This suggests that emissions resulting from upwelling have a significant regional impact. In
the past, denitrification in upwelling regions may have been a major driver of atmospheric N2O variability
(Agnihotri et al., 2006).
Global estimates of the upwelling N2O flux have been made using several approaches. Bange et al. (1996)
calculate a mean surface saturation level from tabulated literature reports and derive a flux based on
average wind speeds and gas-transfer coefficients. They estimate that upwelling areas, covering 0.2
percent of the world ocean, contribute 3 percent of the global aquatic N2O total flux (Table 4-2).
Suntharalingam et al. (2000) use a modeling approach that treats N2O as a tracer in an ocean general
circulation model. Model results are then compared to observed profiles of water column dissolved N2O
distributions. The authors find that although the modeled low-O2 regions occur primarily in known
upwelling areas (to the north and south of the equator in the eastern tropical Pacific and in the northwest
Arabian Sea), these environmental conditions were not limited to narrow coastal bands and were found to
include deep-water O2 minima zones existing over hundreds of km2. Model results suggest that low-O2
regions could supply a significant fraction of the global oceanic source. Model simulations in which 25
and 50 percent of the oceanic flux was attributed to these regions had the best fit to observed
distributions. If the global oceanic source is on the order of 4 Tg N/yr (Nevison et al., 1995; Table 4-3),
this suggests that the flux from low-O2 regions would be 1 to 2 Tg N/yr.
Nevison et al. (2004) extend an upwelling atmospheric flux model developed by Lueker et al. (2003) to
estimate the fluxes required to account for enhanced atmospheric N2O levels. The model is driven by
wind and sea surface temperature data (which therefore predict the occurrence of upwelling) and by
relationships between temperature and subsurface N2O levels. Verifying the model is difficult due to
spatial and temporal mismatches between modeled surface N2O levels and observations of annual means,
but the two appear to be consistent. However, in regions of large subsurface O2 deficits (Peru, Mexico,
the Arabian Sea), the model is thought to significantly underestimate the amount of N2O brought to the
surface—by a factor of 3 or more. Because of this, Nevison et al. (1995) suggest that traditional gas-
transfer flux calculations are likely to underestimate coastal upwelling sources by a factor of 3 to 8 from
poorly sampled regions such as the western coasts of South America and Africa.
Comparison between regional and global upwelling fluxes reveals an apparent disagreement, although
uncertainties are large. Much of the disagreement may be due to the relatively sparse global database on
N2O from these small areas and to mismatches in spatial and temporal resolutions, since the upwelling
phenomenon is episodic. Global upwelling estimates are on the order of 0.5 Tg N/yr (Table 4-2).
However, regional/site-specific estimates from small areas and for periods of less than 1 year, as
discussed above, are similar in size and range from 0.21 to 0.48 Tg N. If we simply sum regional and
short-term estimates and conservatively assume that emissions are zero when there are no measurements,
upwelling from the Arabian Sea, Somali Basin, and northwestern Indian Ocean alone would release 0.5 to
l.OTgN/yr.
4.3.1.3 Open Ocean Nitrous Oxide Fluxes
Since the vast proportion of anthropogenic N input to aquatic systems is consumed and recycled in rivers
and estuaries, Galloway et al. (2004) suggest that the terrestrial and open ocean N budgets are essentially
disconnected. Although there is some anthropogenic impact on the amounts of atmospheric N deposition,
4-10
-------
Chapter 4. Oceans, Estuaries, and Rivers
contributions decrease away from continental regions and all oceanic N2O emissions are considered
natural in origin. Estimates of open ocean N2O fluxes have been made using several different techniques,
which are discussed in more detail below. Initial calculations used measurements of dissolved
concentrations in surface waters and estimates of gas-transfer based on wind speeds. Estimates have also
been made using observations of the correlation between dissolved N2O and other gases. Recently,
inverse modeling techniques have applied to making relatively coarse scale emissions estimates such as
Northern vs. Southern Hemisphere and land vs. ocean fluxes.
The majority of aquatic N2O flux estimates have been made for the open ocean. Surface sampling during
the 1970s and 1980s demonstrated that much of the surface of the world ocean was supersaturated with
respect to atmospheric levels of N2O, indicating that it was a likely net global source. Nevison et al.
(1995) tabulate more than 60,000 N2O measurements and use gas-transfer calculations to estimate a
global flux of 4 Tg N/yr, with a range of 1.2 to 6.8 Tg N/yr. This large range was thought to be largely
due to uncertainties in gas-transfer coefficients and to seasonally biased data collection. Incomplete
geographic and temporal sampling coverage contributes additional uncertainty. Other early global
estimates, made by Capone (1991) and Najjar et al. (1994) using estimates of N2O yield from nitrification
and denitrification, are in rough agreement (4 to 5.8 Tg N/yr; Table 4-4).
Based on observations of water column profiles, N2O is «,-,..- „ ,. • ,. „,
., ,. . , , , . . . .•,.,.•,. Ohqotrophic means nutrient-poor and
thought to be produced a intermediate depths in the therefore usually having low productivity.
ocean. Nitrification is believed to be the dominant
Euphotic means "having light"; aphotic
means "having no light."
source process because N2O and O2 are commonly
inversely correlated and N2O and NO3 are positively
correlated. Based on isotopic measurements of N2O in the oligotrophic subtropical North Pacific, Dore et
al. (1998) find that nitrification in the lower euphotic and upper aphotic zones (100 to 300 meters) could
supply 70 to 90 percent of the oceanic N2O source.
Nevison et al. (1995) and Global Emissions Inventory Activity (GEIA) inventories based on this work
(Bouwman et al., 1995; see Table 4-3) estimated that most oceanic emissions (45 percent) are located in
the Southern Ocean (30 to 90°S), where high winds
combine with high levels of supersaturation. Some
,. • , , • ., AT w-u A+I +• part of the International Geosphere-Biosphere
areas in subtropical gyres and in he North Atlantic ^ gp intema Jna|
may change seasonally from weak sinks (winter) to
weak sources (summer).
The strong correlation between the degree of N2O
saturation relative to atmospheric equilibrium and sin9|e 'ar9est Contributor. GEIA develops gas and
A~TT ., ,,... , . M ^. . -KT r. aerosol emissions inventories of both natural and
AOU provides an additional way to estimate N2O anthropogenic sources and contributes to global
GEIA (the Global Emissions Inventory Activity) is
organization that studies global change and earth
system science. The IGBP is funded by roughly 40
countries, among which the United States is the
assessments such as the IPCC.
flux by estimating N2O production relative to O2
consumption. Suntharalingam and Sarmiento (2000)
use this relationship in combination with a global ocean biogeochemical model. They derive an open
ocean source (excluding N2O from low-O2 regions such as upwelling) that is similar to those based on
gas-transfer estimates (Table 4-4). However, there are difficulties in using this relationship to estimate
fluxes. There is variability in the oceanic N2O saturation/AOU relationship that may be the result of
differences in the N2O/NO3 yield due to the sensitivity of nitrifiers to O2 (Nevison et al., 2003). These
authors find that the correlation of N2O supersaturation and AOU is a poor predictor of the N2O
yield/mole O2 consumed because the relationship is strongly affected by mixing. By calculating N2O
production relative to O2 consumption as a function of O2 and depth, Nevison et al. (2003) estimate the
oceanic N2O production at 5.8 Tg N/yr. They suggest this figure could be decreased by perhaps 1 to 3 Tg
N/yr by denitrifier consumption in low-O2 environments. Approximately 70 percent of production was
calculated to occur in the tropics, between 30°N and 30°S, in contrast to earlier estimates of flux
dominance by southern latitudes (Table 4-3).
4-11
-------
Chapter 4. Oceans, Estuaries, and Rivers
The apparent disagreement in the relative importance of southern latitude oceans was examined in more
detail by Nevison et al. (2005). They used the 10-year atmospheric N2O record at Cape Grim, Tasmania,
to estimate the seasonal atmospheric cycle and calculate the transfer of N2O from the subsurface to the
atmosphere. The calculations make use of observations from the coast of California during periods of
upwelling (Leuker et al., 2003). At these times, the subsurface correlation between N2O saturation and
AOU (N2O/AOU) is mirrored by an inverse relationship between atmospheric N2O and O2/N2.
Comparison between atmospheric N2O observations and those predicted by an ocean biogeochemical
model coupled with an atmospheric transport model suggests that the Southern Ocean N2O flux is roughly
0.9 Tg N/yr, although the large corrections applied to the atmospheric N2O data create significant
uncertainty (Table 4-3).
Nevison et al. (2005) find that oceanic N2O saturation should not be treated as an annual constant, as was
done earlier in Nevison et al. (1995), since it exhibits strong seasonality. This seasonality is a result of
thermal effects during the summer and a larger mixing effect during the winter as the depth of the surface
mixed layer deepens and deeper N2O-enriched waters are mixed in. Since the majority of surface N2O
measurements have been made during the summer, this implies that annual mean fluxes may be
overestimates. It may also help explain some of the apparent discrepancy in Southern Ocean emissions.
Table 4-3. Open Ocean Nitrous Oxide Fluxes (Post-1990)
Broken Down by Latitude Band (Tg N/Year)
Latitude
Northern
Tropical
Southern
Northern
Tropical
Southern
Southern
Northern
Tropical
Southern
Flux
(V Std. Dev.
or Range)
1990s
1.2V1.0
0.9V 3.3
1.7 V 2.0
0.9
4.1
0.9
0.9
1.8(1.7-2.1)
3.5 (2.6-4.1)
0.4 (0-0.8)
Source
Bouwman et al.,
1995
Nevison etal., 2003
Nevison etal., 2005
Hirsch et al., 2006
Comment
Based on Nevison (1994), subtracting 0.2 Tg
N/yr for scaling conversion; used forGEIA
inventory
N2O production rate; may be 1-3 Tg N/yr less
due to N2O consumption by denitrifiers;
based on )N2O/AOU relationship
Modeled using atmospheric N2O seasonal
cycle and production as a function of O2
consumption (Jin and Gruber, 2003)
Inverse model; mean and range of model
scenarios; uses GEIA estimate of 3.8 Tg N/yr
as a priori; likely includes upwelling and
continental shelves; latitude grouping broken
at 15° rather than 30°
Northern = 90°N-30°N; tropical = 30°N-30°S; southern = 30°S-90°S.
Recently, Hirsch et al. (2006) have applied inverse modeling techniques to atmospheric N2O
measurements in an attempt to infer fluxes and estimate their uncertainties between 1998 and 2001. Due
to its relatively long atmospheric lifetime, applying large-scale inverse techniques to N2O presents some
challenges because its atmospheric concentration is nearly constant. Seasonal cycles, for example, are on
the order of 0.1 percent of the mean concentration (Jiang et al., 2007) and until recently they have been
difficult to isolate from necessary sink corrections. Using the GEIA inventory (Bouwman et al., 1995) as
initial fluxes, Hirsch et al. (2006) test whether these emissions are consistent with atmospheric
concentration distributions. Results suggest that flux from the Southern Ocean is lower than the initial
estimate of 1.7 Tg N/yr; this is consistent with other recent reports. They also find that tropical emissions
4-12
-------
Chapter 4. Oceans, Estuaries, and Rivers
are substantially higher than GEIA estimates. However, because sampling is sparse in the tropical region
(15°N to 15°S), it is difficult to separate and determine land and ocean fluxes. The ratio of Northern
Hemisphere to Southern Hemisphere modeled fluxes ranged from 1.9 to 5.2. These values are
significantly higher than the initial GEIA inventory ratio of 1.5 and highlight the importance of emissions
from the Northern Hemisphere. Overall, Hirsch et al. (2006) calculate that oceans contribute 26 to 36
percent of total flux. This is consistent with GEIA estimates (29 percent), but generally higher. Hirsch et
al. (2006) also compare land and ocean emissions. Globally, the mean ratio of land to ocean N2O
emissions was calculated at 2.04, less than the initial GEIA inventory ratio of 2.41 and implying a greater
relative contribution from oceans. However, because tropical land and ocean fluxes are difficult to
distinguish, this figure is somewhat uncertain. Ratios of land to ocean fluxes increased from northern
latitudes (an average of 1.6) through tropical (1.9) to southern (3.7), reflecting the large estimated
decrease in Southern Ocean fluxes. Globally, oceans were estimated to release an average of 5.7 Tg N/yr
with estimates ranging from 4.5 to 6.5 Tg N/yr.
An important consideration, however, is what is included in the Hirsch et al. classification of "ocean."
Since the authors are working at large geographic scales, discriminating only between "land" and
"ocean," the ocean class includes all oceanic environments (open ocean, continental shelves, upwelling
areas). Because rivers and estuaries are relatively small-scale features, they are likely included into the
land category. If we remove the flux estimates for continental shelf and upwelling regions derived in
Table 4-2 from the Hirsch et al. ocean flux estimate, it is reduced to a figure similar to earlier estimates
(Table 4-4).
Seasonal and interannual variability in atmospheric N2O also yields insight into sources and sinks. Only
recently has the accumulation of high-quality data made it feasible to examine variability on these shorter
time scales. Nevison et al. (2007) compare observations of N2O and a suite of chlorofluorocarbons (used
as tracers) with atmospheric transport model simulations. They generate monthly mean oceanic N2O
fluxes from a model described by Jin and Gruber (2003), which combines a number of 3-D coupled
biogeochemical models with an N2O component based on Suntharalingam et al. (2000). These
calculations estimate an oceanic flux of 3 Tg N/yr but find that uncertainties in stratospheric mixing tend
to overwhelm source variability on shorter time scales.
Jiang et al. (2007) also examine the seasonal cycle of atmospheric N2O and derive information on sources
and sinks from its latitudinal distribution. They find that the peak to trough amplitude of the seasonal
cycle increases systematically from 0.29 ppb at the South Pole to 1.15 ppb at Alert, Nunavut, Canada. The
month of the seasonal minimum concentration also changes systematically, from April at the South Pole
to September at Alert. The Northern Hemisphere seasonal cycle appears to be driven largely by
stratospheric influences, which control the loss of N2O. In the Southern Hemisphere, surface sources such
as the oceans appear to have greater impact. Over the 3-year period from 2000 to 2002, the mean N2O
concentration in the Northern Hemisphere was 0.73 ppb greater than that in the Southern Hemisphere.
This difference requires that sources in the Northern Hemisphere be 4.7 Tg N/yr higher, significantly
greater than the inter-hemispheric difference of 2.657 Tg N/yr derived from the GEIA inventory. This
value, however, is lower than the inter-hemispheric source difference estimated by Hirsch et al. (2006),
which ranged from 5.5 to 11.9 Tg N/yr and averaged 8.8 Tg N/yr. A greater inter-hemispheric difference
may in part be due to the increase in Northern Hemisphere anthropogenic emissions since 1995 as well as
the downward revision in Southern Ocean fluxes.
4-13
-------
Chapter 4. Oceans, Estuaries, and Rivers
Table 4-4. Open Ocean Nitrous Oxide Fluxes (Post-1990)
(Tg N/Year)
Flux
(Range)
1.3
5.8
2
(1 .4-2.6)
4
3.8
(2.8-5.7)
4
(1 .2-6.8)
3
(1-5)
3.6
1.8
(1.34-2.18)
3.5
3.85
(2.7-8.0)
3.0/3.6
(0.9-5.7)
5.8
(3.8-7.8)
3.5
3.7
5.7
(4.5-6.5)
3
3.8
(1 .8-5.8)
1.2
(0.9-1 .7)
3.2
(1-4)
Source
Prinn etal., 1990
Capone, 1991
Houghton etal., 1992
Najjar, 1994
Nevison, 1994
Nevison et al., 1995
Houghton etal., 1995
Bouwman et al., 1995
Bange etal., 1996
Seitzingeretal., 2000
Suntharalingam and
Sarmiento, 2000
Ehhalt etal., 2001
Nevison etal., 2003
Galloway etal., 2004
Nevison et al., 2005
Hirsch etal., 2006
Nevison et al., 2007
Denman etal., 2007
Rhee etal., 2009
Open ocean only,
best guess
Comment
Early inverse modeling; adopts 4 Tg N/yr a priori by
subtraction (the few previous estimates varied widely);
atmospheric lifetime used too long
Estimated from N2O released during nitrification and
denitrification
The first IPCC estimate
Estimated from N2O released during nitrification
Dissertation; based on compiling cruise data
Based on surface measurements, upwelling not treated
separately
The IPCC Second Assessment Report estimate
Based on Nevison (1994), subtracts 0.2 Tg N/yr for scaling
conversion; used forGEIA inventory
Midpoint of range
Based on Nevison (1994) with continental shelf subgrids
subtracted; 100% considered natural
Model of organic matter remineralization and )N2O/AOU
relationship
The IPCC Third Assessment Report estimate
N2O production rate; may be 1-3 Tg N/yr less due to N2O
consumption by denitrifiers; based on )N2O/AOU relationship
Continental shelf estimate subtracted from Nevison et al.
(1995)
Modeled using atmospheric N2O seasonal cycle and
production as a function of O2 consumption (Jin and Gruber,
2003)
Inverse model; mean and range of model scenarios; uses
GEIA estimate of 3.8 Tg N/yr as initial fluxes; likely includes
upwelling and continental shelves — with these subtracted (1 .9
Tg N/yr; Table 4-2), would = 3.8 Tg N/yr, range = 3-5 Tg N/yr
Based on Jin and Gruber (2003) model
Cites Nevison et al. (2003; 2004 — upwelling) as sources,
combining uncertainties in production and exchange; this is
the estimate from AR4
Average of Seitzingeret al. (2000), Suntharalingam and
Sarmiento (2000), Nevison et al. (2005), Rhee et al. (2009),
and Hirsch et al. (2006) with continental shelf (1 .5 Tg N/yr;
Table 4-2) and upwelling (0.37; Table 4-2) removed
4-14
-------
Chapter 4. Oceans, Estuaries, and Rivers
In summary, mean open ocean N2O flux estimates have remained in the range of 3 to 6 Tg N/yr despite
the variety of approaches used to calculate their magnitude (Table 4-4). A 3 Tg N/yr range, however, is
significant, making up about 16 percent of the global total of 18.8 Tg N/yr (see Table ES-1). Inclusion of
the uncertainties for individual estimates increases this range further. The most significant revision of
oceanic flux estimates in recent years has been the reduction of the relative importance of the Southern
Ocean (Table 4-3). Once thought to make up nearly half of the oceanic source, it is now thought to
contribute 7 to 15 percent. Assuming open ocean flux is roughly 3.2 Tg N/yr (Table 4-4), open oceans
contribute on the order of 17 percent of the total.
4.3.1.4 Summary: Rivers, Estuaries, Continental Shelves, Upwelling, and Open Ocean
An important improvement in our capability to estimate oceanic and other emissions is the accumulating
database of high-quality atmospheric N2O measurements. This database makes it possible to use inverse
modeling techniques (Hirsch et al., 2006) and examine sources at higher spatial and temporal resolutions,
a strategy that has yielded important insights for other trace gases. Because many important N2O sources,
including those from aquatic environments, are biogenic in nature, seasonality is a critical consideration.
As the spatial scales of estimates are reduced, discrimination between different biogeochemical
environments (open ocean, regions of upwelling, continental shelves) is enhanced and the inclusion of
episodic emissions is made possible. These improvements should allow significant progress in refining
emissions, which in turn will aid in predicting changes in emissions as the effects of climate change are
felt.
Mean fluxes from Tables 4-1 through 4-4 are summarized below. The total flux of 5.4 Tg N/yr is higher
than that calculated in the AR4 report (an average of 3.8 Tg N/yr with a range of 1.8 to 5.8 Tg N/yr;
Denman et al., 2007), but includes natural emissions from rivers and estuaries that are classified as
entirely anthropogenic there. The relatively shallow oceanic environments of the continental shelves and
upwelling regions are treated explicitly here. Upwelling may be included in the AR4 anthropogenic class,
although it is natural in origin. Continental shelves, here calculated as releasing 1.5 Tg N/yr and
contributing 28 percent of flux, are highly uncertain. The few reported emissions vary by roughly an order
of magnitude (Table 4-2). If the low estimate of 0.46 Tg N/yr based on the work of Seitzinger and Kroeze
(1998) and Seitzinger et al. (2000) is used, rather than the mean of reported fluxes, total emissions are
reduced to 4.88 Tg N/yr. If the global atmospheric source of N2O is 18.8 Tg N/yr (see Table ES-1), then
ocean, estuarine, and riverine sources would contribute on the order of 29 percent.
Table 4-5. Summary of Natural Ocean, Estuarine, and Riverine
Nitrous Oxide Flux Estimates
Environment
Rivers
Estuaries
Continental shelves
Upwelling areas
Open ocean
Total
Avg. Flux
(Tg N/Yr)
0.09
0.24
1.5
0.37
3.2
5.4
Range
0.08-0.10
0.03-0.45
0.37-3.52
0.0003-1.0
1-4
1.5-9.1
Comment
Average taken as midpoint of range; 1% of total
Average taken as midpoint of range; 4% of total
28% of total
7% of total
59% of total
Range from summing minima and maxima
4.3.2 Current Ocean, Estuarine, and Riverine Methane Fluxes
The flux of CH4 from oceans, estuaries, and rivers has been calculated to make a relatively small
contribution to total emissions. Initial estimates were made by Ehhalt (1974), based on only a few
4-15
-------
Chapter 4. Oceans, Estuaries, and Rivers
dissolved concentrations and gas-transfer calculations. Ehhalt and Schmidt (1978) suggest that the oceans
release 1.3 to 16.6 Tg CFL/yr, based on limited sampling and gas-transfer calculations. Although they
indicate that upwelling regions may transfer deep water enriched in CFL to the surface and therefore be
areas of enhanced emissions, there were no data available to make estimates. Because the oceans have
been considered to be a minor source, they have not been the focus of extensive research and these early
estimates have been carried forward (Cicerone and Oremland, 1988; Fung et al., 1991). Oceanic fluxes
are frequently not included in inverse modeling studies because emissions are relatively low and are
distributed over large areas. Atmospheric samples, upon which inverse techniques are based, tend not to
be sensitive to them therefore (Chen and Prinn, 2006). Mikaloff Fletcher et al. (2004), for example,
simply define oceanic CFi4 flux at 5 Tg CFL/yr.
A small database reported by Lambert and Schmidt (1993) yielded estimates similar to those used earlier,
with coastal (continental shelf) emissions roughly twice that of the open ocean despite making up about 7
percent of ocean surface area (Table 4-6). Bange et al. (1994) report CFL saturation levels in the Baltic
and North Seas and summarize literature reports to derive the first ecosystem-specific flux estimates.
They derive a somewhat higher annual oceanic flux of 11 to 18 Tg CFL/yr, but acknowledge that there are
significant uncertainties in the estimate. The majority of measurements upon which it is based are from
summer months, which is likely to bias the estimate high. In addition, northern water bodies are ice-
covered for much of the year, which impedes gas exchange. All early estimates neglect possible
seasonality and assume emissions are constant year-round. Differences in assumed wind speeds, gas
transfer model, and coefficients also add considerable uncertainty to estimates. Despite caveats, Bange et
al. (1994) calculate that the majority (75 percent) of emissions occur in the shallow water habitats of
estuaries and continental shelves, even though they have relatively small surface areas. Although the
range of continental shelf emissions is relatively large (see Table 4-6), it is similar to that of Lambert and
Schmidt (1993).
Bates et al. (1996) report a multi-year database of open ocean CFL concentrations from the Pacific Ocean.
Surface waters in the equatorial tropical Pacific were supersaturated with respect to atmospheric
equilibrium, while those outside of the tropics were often under-saturated during fall and winter. Their
measurements provide the first estimates of seasonality and the large-scale distribution of emissions.
Open water CFL fluxes based on this dataset are roughly an order of magnitude lower than those
calculated by Bange et al. (1994) and Lambert and Schmidt (1993). Although both the Kelley and Jeffrey
(2002) and Rhee et al. (2009) estimates are calculated based on a single cruise, they also report low open
ocean emissions, consistent with those of Bates et al. (1996). Seasonal sampling in the tropical South
Atlantic gyre suggests intermediate levels and demonstrates the importance of seasonal variability, with
higher emissions in the fall than the spring (Robinson et al., 2006). In contrast, sampling near Hawaii over
the course of a year found little variation in fluxes (Holmes et al., 2000).
Regional open ocean sampling has yielded CFL fluxes that range from 1.2 to 3.2 (imol/m2/d (e.g.,
Scranton and Brewer, 1977; Ward, 1992; Tilbrook and Karl, 1995; Holmes et al., 2000; Oudot et al.,
2002). If we assume emissions are relatively constant year-round and that the open ocean surface area is
roughly 340 x 106 km2 (Bates et al., 1996), these regional estimates would extrapolate to global emissions
of2to6TgCH4/yr.
There are very few global estimates for CIL, from upwelling regions, although Bange et al. (1998) and
Upstill-Goddard et al. (1999) have noted their importance in the Arabian Sea and Sansone et al. (2001)
have done the same for the eastern tropical North Pacific. Working in the Atlantic, Rhee et al. (2009)
calculate very low fluxes, 0.001 Tg CFL/yr. Owens et al. (1991) found that high CFLt production in the
Arabian Sea was associated the increased phytoplankton biomass supported by monsoon-driven
upwelling. They calculate that fluxes were up to 5 times that of the average ocean flux and that the
Arabian Sea (making up < 1 percent of the global ocean surface area) could account for 1.3 to 133 percent
of the estimated open ocean flux.
4-16
-------
Chapter 4. Oceans, Estuaries, and Rivers
Although there are a limited number of estimates of CH^ flux from estuaries, they fall within the same,
relatively narrow range of roughly 1 to 3 Tg CHVyr (Table 4-6). It is important to note, however, that
these estimates are derived from limited sampling, much of which is from the North and Baltic Seas.
There are no samples from tropical systems. While QrU distributions in a number of rivers have been
studied and budgets have been attempted, few estimates of the global contributions from these systems
have been calculated, aside from that of Upstill-Goddard et al. (2000). Working in the North Sea basin,
they report that Clr^ was highly, but non-linearly correlated with salinity. They calculate that over 90
percent of the riverine, low-salinity QrU input to estuaries was lost there, the vast majority through air-
water exchange. Recently, Saarnio et al. (2009) have attempted to calculate the flux of QrU from
European lakes and rivers. Their river estimate of 0.14 Tg CHVyr is relatively high compared to the
Upstill-Goddard et al. global estimate of 0.25 Tg CHVyr. This may be due to the inclusion of emissions
from small streams draining peatlands and wetlands in addition to those from rivers. We calculate the flux
of CH4 from oceans, estuaries, and rivers by averaging recent habitat-specific estimates and summing
(Table 4-6). Emissions total 9.1 Tg CHVyr, with the majority released from the continental shelves (5.5
Tg CHVyr; 60 percent of total). Fluxes from the open ocean and from estuaries and rivers are similar at
1.8 and 1.85 Tg CHVyr, respectively. Emissions compiled in Table 4-6 suggest that additional research
since the early 1980s has served to confirm initial estimates of oceanic, estuarine, and riverine CH4 fluxes
of 10 to 15 Tg CHVyr. This report's estimate of roughly 9 Tg CHVyr falls within this range, given the
uncertainties involved in these estimations. If the global annual flux of CH4 to the atmosphere is roughly
566 Tg CHVyr (see Table ES-1), oceans, estuaries, and rivers would contribute 2 percent.
There are, however, significant uncertainties in these estimates. Sampling is poor in tropical and southern
latitudes and is seldom performed over an annual or multi-year scale. Virtually all estimates are calculated
from sporadic samples assumed to be representative of the entire year. This is a simplification, since CH4
saturation levels and wind speeds vary seasonally. Calculated open ocean emissions vary by an order of
magnitude, due perhaps to methodological problems and/or unrepresentative sampling. Current estimates
do not include emissions from upwelling areas. It would not be expected that such emissions would
greatly change total fluxes; they would be unlikely to increase fluxes by more than a few Tg. Perhaps
unsurprisingly, emissions appear to be dominated by coastal and shelf systems, where organic sources are
greater and contact with anaerobic sediments is greater.
Table 4-6. Ocean, Estuarine, and Riverine Methane Fluxes
Flux
(Tg CH/Yr)
4-6.7
1.3-16.6
10
(5-20)
10
3.6
3.6
(2.8-4.4)
10
0.4
15 (±10)
10-15
Environment
Open ocean
Open ocean
Open ocean
Oceans
Open ocean
Open ocean
Open ocean
Open ocean
Open ocean
Open ocean
Reference
Ehhalt, 1974
Ehhalt and Schmidt, 1978
Cicerone and Oremland,
1988
Fung etal., 1991
Lambert and Schmidt,
1993
Bange etal., 1994
Houghton etal., 1995
Bates etal., 1996
Houweling etal., 2000
Ehhalt etal. ,2001
Comment
Cites Ehhalt (1974)
Cites Ehhalt (1974)
Total of 93 measurements
Flux calculated as midpoint of range
The IPCC Second Assessment
Report; used by Lelieveld et al.
(1998)
Cites Lambert and Schmidt (1993)
The IPCC Third Assessment Report;
cites Lelieveld et al. (1998); Fung et
4-17
-------
Chapter 4. Oceans, Estuaries, and Rivers
Flux
(Ta CH/Yr;
0.8 ±0.6
4
(0.2-2)
0.44
(0.41-0.46)
0.7-14
6.1
9.7
(7.3-12.1)
0.58
(0.41-0.77)
0.0009
(0.0009-0.001)
1.06
(0.8-1 .32)
2.1
(1.1-3)
0.1-0.4
Environment
Open ocean
Open ocean
Open ocean
Shelf
Coastal ocean
Shelf
Coastal ocean
Upwelling
Estuaries
Estuaries
Rivers
Reference
Kelley and Jeffrey, 2002
Wuebbles and Hayhoe,
2002
Rhee etal., 2009
Ehhalt, 1974
Lambert and Schmidt,
1993
Bangeetal., 1994
Rhee et al., 2009
Rhee etal., 2009
Bange etal., 1994
Middelburg etal., 2002
Upstill-Goddard et al.,
2000
Comment
al. (1991); Houweling etal. (2000)
Single cruise, 4 1°S-27°N
Cites Bates et al. (1996); Holmes et
al. (2000)
Single Atlantic cruise, 50°S-50°N
Flux calculated as midpoint of range
Flux calculated as midpoint of range
Flux calculated as midpoint of range
Summary
1.8
(0.4-3.6)
5.5
(0.58-9.7)
1.85
(1.06-2.1
estuaries)
(0.25 rivers)
9.1
(2.3-15.6)
Open ocean
Shelf
Estuaries and
rivers
Total
Avg. of Rhee et al. (2009),
Kelley and Jeffrey (2002),
Bates et al. (1996), Bange
et al. (1994), and Lambert
and Schmidt (1993)
Average of Rhee et al.
(2009), Bange et al.
(1994), and Lambert and
Schmidt (1993)
Average of Bange et al.
(1994) and Middelburg et
al. (2002) + range
midpoint of Upstill-
Goddard et al. (2000)
20% of total
60% of total
20% of total
Range calculated from summing
minima and maxima
Future Oceanic, Estuarine, and Riverine Nitrous Oxide and
Methane Emission Scenarios
4.4
In general, natural oceanic, estuarine, and river emissions of N2O and QrU are expected to remain largely
unchanged. Even though atmospheric levels of both gases have increased sharply in the last several
hundred years, this is primarily due to increases in anthropogenic sources. The impact of human activities
is centered on the continents, so aquatic ecosystems closer to continents are affected more than the open
ocean. Galloway et al. (2004), for example, suggest that open ocean N cycles are largely unconnected to
those in terrestrial systems. In extrapolating N2O emissions for 2050, Kroeze and Seitzinger (1998) have
estimated that natural sources will remain constant and that their relative contribution will decrease as
anthropogenic sources continue to rise.
4-18
-------
Chapter 4. Oceans, Estuaries, and Rivers
The sensitivity of denitrification and nitrification to O2 levels suggests that in some systems, coastal
eutrophication may enhance denitrification rates, leading to flux increases (Naqvi et al., 2000). Increases
in hypoxia and the extent of anaerobic sediments may also increase CH4 emissions, although there are no
estimates available. These increases, however, would be considered largely anthropogenic in origin. Since
submarine groundwater discharge appears to be an important source of nutrients and other dissolved
materials to some coastal systems, this may provide a mechanism by which anthropogenic impacts are
conveyed offshore.
Because N2O and ClrU production in upwelling areas is a function of the enhanced primary production
there, it is possible that changes in production due to increases in CO2 may also change their flux (Altabel
et al., 2002). For example, ocean iron fertilization, proposed to enhance productivity and mitigate
increasing atmospheric CO2, may also increase N2O emissions. This feedback would therefore offset
possible radiative benefits (Jin and Gruber, 2003).
Overall, it is not expected that possible changes in oceanic emissions of N2O and QrU will greatly affect
climate policy. Emissions of CFLj are a relatively small fraction of total emissions; significant changes
would require large environmental alterations. While ocean N2O emissions do make a significant
contribution to global emissions, the majority of emissions are from the open ocean and are less
susceptible to anthropogenic impacts. Based on the current understanding of emissions, major controls are
fundamental physical oceanic properties (wind parameters, mixing) that would not be easily changed.
4.5 Areas for Further Research
Because there are almost no estimates of the global upwelling source of CH4, focused sampling and the
compilation of available literature on CH4 saturation ratios to derive a first-order estimate would likely be
a useful way to decrease this uncertainty. Sampling and modeling the episodic nature of these emissions
is challenging.
Current estimates of N2O fluxes from the continental shelves vary by an order of magnitude, so refining
emissions from this region should be a priority.
Because the ratio of N2O/N2 is a critical but poorly constrained variable in estimating the production of
N2O from N exported to rivers and estuaries, improving our understanding of its magnitude and controls
would help refine both natural and anthropogenic emissions from these environments.
Data in general from tropical and southern latitude environments are sparse for both gases. Since tropical
systems in particular are suggested as globally important (Hirsch et al., 2006), this region should receive
increased attention and research. Most current flux estimates do not include seasonal or episodic inputs,
which also greatly increases uncertainties. Because the enlarging atmospheric database has higher
temporal resolution (hourly to weekly), the full exploitation of this valuable resource requires a better
understanding of the variability in fluxes. Currently, there is considerable temporal mismatch between
emissions and atmospheric samples for inverse modeling.
4.6 References
Abril, G., and N. Iversen. 2002. Methane dynamics in a shallow non-tidal estuary (Randers Fjord,
Denmark). Mar. Ecol Progr. Ser. 230: 171-181.
Agnihotri, R. M.A. Altabet, and T.D. Herbert. 2006. Influence of marine denitrification on atmospheric
N2O variability during the Holocene. Geophys. Res. Lett. 33: L13704, 10.1029/2006G1025864.
Altabel, M.A., M.J. Higginson, and D.W. Murray. 2002. The effect of millennial-scale changes in
Arabian Sea denitrification on atmospheric CO2. Nature 415: 159-162.
4-19
-------
Chapter 4. Oceans, Estuaries, and Rivers
Bange, H.M. 2006a. New directions: The importance of oceanic nitrous oxide emissions. Atmos. Environ.
40: 198-199.
Bange, H.M. 2006b. Nitrous oxide and methane in European coastal waters. Est. Coastal Shelf Sci. 70:
361-374.
Bange, H.W., M.O. Andreae, S. Lai, C.S. Law, S.W.A. Naqvi, P.K. Patra, T. Rixen, and R.C. Upstill-
Goddard. 2001. Nitrous oxide emissions from the Arabian Sea: A synthesis. Atmos. Chem. Phys. 1:61-
71.
Bange, H.W., R. Ramesh, S. Rapsomanikis, M.O. Andreae. 1998. Methane in the surface waters of the
Arabian Sea. Geophys. Res. Lett 25: 3547-3550.
Bange, H.W., S. Rapsomanikis, and M.O. Andreae. 1996. Nitrous oxide in coastal waters. Global
Biogeochem. Cycles 10: 197-207.
Bange, H.W., U.K. Bartell, S. Rapsomanikis, and M.O. Andreae. 1994. Methane in the Baltic and North
Seas and a reassessment of the marine emissions of methane. Global Biogeochem. Cycles 8(4): 465-
480.
Bates, T.S., K.C. Kelly, J.E. Johnson, and R.H. Gammon. 1996. A reevaluation of the open ocean source
of methane to the atmosphere. J. Geophys. Res. 101: 6953-6961.
Bouwman, A.F., K.W. Van der Hoek, and J.G.J. Olivier. 1995. Uncertainties in the global source
distribution of nitrous oxide. J. Geophys. Res. 100(D2): 2785-2800.
Boyer, E.W., RW. Howarth, J.N. Galloway, F.J. Dentener, P.A. Green, and C.J. Vorosmarty. 2006.
Riverine nitrogen export from the continents to the coasts. Global Biogeochem. Cycles 20: GB1S91,
10.1029/2005GB002537.
Bugna, G.C., J.P. Chanton, J.E. Cable, W.C. Burnett, and P.H. Cable. 1996. The importance of
groundwater discharge to the methane budgets of nearshore and continental shelf waters of the
northeastern Gulf of Mexico. Geochim. Cosmochim. Acta. 60(23): 4735-4746.
Capone, D.G. 1991. Aspects of the marine nitrogen cycle with relevance to the dynamics of nitrous and
nitric oxide. In: J.E. Roders and W.E. Whitman, (eds.). Microbial Production and Consumption of
Greenhouse Gases. Am. Soc. Microbiol. Washington, DS. pp. 255-275.
Chen, Y.-H., and R.G. Prinn. 2006. Estimation of atmospheric methane emissions between 1996 and
2001 using a three-dimensional global chemical transport model. J. Geophys. Res. Ill: D103 07,
10.1029/2005JD006058.
Cicerone, R.J., and R.S. Oremland. 1988. Biogeochemical aspects of atmospheric methane. Global
Biogeochem. Cycles 2: 299-327.
de Angelis, M.A., and M.D. Lilley. 1987. Methane in surface waters of Oregon estuaries and rivers.
Limnol. Oceanogr. 32(2): 716-722.
de Angelis, M.A., and M.I. Scranton. 1993. Fate of methane in the Hudson River and estuary. Global
Biogeochem. Cycles 7(3): 509-523.
De Wilde, H.P.J., and W. Helder. 1997. Nitrous oxide in the Somali Basin: The role of upwelling. Deep
Sea Res. II. 44: 1319-1340.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohman, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment
4-20
-------
Chapter 4. Oceans, Estuaries, and Rivers
Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY:
Cambridge University Press.
Dore, J.E., B.N. Popp, D.M. Karl, and F.J. Sansone. 1998. A large source of atmospheric nitrous oxide
from subtropical North Pacific surface waters. Nature. 396: 63-66.
Dumont, E., J.A. Harrison, C. Kroeze, E.J. Bakker, and S.P. Seitzinger. 2005. Global distribution and
sources of dissolved inorganic nitrogen export to the coastal zone: Results from a spatially explicit,
global model. Global Biogeochem. Cycles 19: GB4s02, 10.1029/2005GB002488.
Ehhalt, D., M. Prather, F. Dentener, R. Derwent, E. Dlugokencky, E. Holland, I. Isaksen, J. Katima, V.
Kirchhoff, P. Matson, P. Midgley, and M. Wang. 2001. Atmospheric chemistry and greenhouse gases.
In: J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A.
Johnson (eds.). Climate Change 2001: The Scientific Basis, Contribution of Working Group I to the
Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and
New York, NY: Cambridge University Press.
Ehhalt, D.H. 1974. The atmospheric cycle of methane. Tellus 26: 58-70.
Ehhalt, D.H., and U. Schmidt. 1978. Sources and sinks of atmospheric methane. Pageoph. 116: 452-464.
Fung, I., J. John, J. Lerner, E. Matthews, M. Prather, L.P. Steele, and P.J. Fraser. 1991. Three-
dimensional model synthesis of the global methane cycle. J. Geophys. Res. 96: 13033-13065.
Galloway, J.N., F.J. Dentener, D.G. Capone, E.W. Boyer, R.W. Howarth, S.P. Seitzinger, G.P. Asner,
C.C. Cleveland, P.A. Green, E.A. Holland, D.M. Karl, A.F. Michaels, J.H. Porter, A.R. Townsend, and
C.J. Vorosmarty. 2004. Nitrogen cycles: Past, present, and future. Biogeochem. 70(2): 153-226.
Green, P.A., C.J. Vorosmarty, M. Meybeck, J.N. Galloway, B.J. Peterson, and E.W. Boyer. 2004. Pre-
industrial and contemporary fluxes of nitrogen through rivers: A global assessment based on typology.
Biogeochem. 68(1): 71-105.
Hirsch, A.I., A.M. Michalak, L.M. Bruhwiler, W. Peters, E.J. Dlugokencky, and P.P. Tans. 2006. Inverse
modeling estimates of the global nitrous oxide surface flux from 1998-2001. Global Biogeochem.
Cycles 20: GB1008, 10.029/2004GB002443.
Holmes, M.E., F.J. Sansone, T.M. Rust, and B.N. Popp. 2000. Methane production, consumption, and air-
sea exchange in the open ocean: An evaluation based on carbon isotopic ratios. Global Biogeochem.
Cycles 14(1): 1-10.
Houghton, J.T., B.A. Callender, and S.K. Varney. 1992. Climate Change 1992. The Supplementary
Report to the IPCC Scientific Assessment. Cambridge, UK: Cambridge University Press.
Houghton, J.T., L.G. Meira Filho, B.A. Callender, N. Harris, A. Kattenberg, and K. Maskell. 1995.
Climate Change 1995: The Science of Climate Change. Contribution of Working Group I to the Second
Assessment of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge
University Press.
Houweling, S., F. Dentener, and J. Lelieveld. 2000. Simulation of preindustrial atmospheric methane to
constrain the global source strength of natural wetlands. J. Geophys. Res. 105(D13): 17243-17255.
Jiang, X., W.L. Ku, R-L. Shia, Q. Li, J.W. Elkins, RG. Prinn, and Y.L. Yung. 2007. Seasonal cycle of
N2O: Analysis of data. Global Biogeochem. Cycles 21: GB1006, 10.029/GB002691.
Jin, X., and N. Gruber. 2003. Offsetting the radiative benefit of ocean iron fertilization by enhancing N2O
emissions. Geophys. Res. Lett. 30(24): 2249, doi:10.029/2003GL018458.
Karl, D.M., L. Beversdorf, K.M. Bjorkman, M.J. Church, A. Martinez, and E.F. DeLong. 2008. Aerobic
production of methane in the sea. Nat. Geosci. 1:473-478.
4-21
-------
Chapter 4. Oceans, Estuaries, and Rivers
Kelley, C.A. and W.H. Jeffrey. 2002. Dissolved methane concentration profiles and air-sea fluxes from
41°Sto27°N. Global Biogeochem. Cycles 16(3): 1040, 10.1029/2001GB001809.
Kim, G., and D.W. Hwang. 2002. Tidal pumping of groundwater into the coastal ocean revealed from
submarine Rn-222 and CH4 monitoring. Geophys. Res. Lett. 29(14): 1678.
Kroeze, C., and S.P. Seitzinger. 1998. Nitrogen inputs to rivers, estuaries and continental shelves and
related nitrous oxide emissions in 1990 and 2050: A global model. Nutr. Cycling inAgrosyst. 52: 195-
212.
Kroeze, C., E. Dumont, and S.P. Seitzinger. 2005. New estimates of global emissions of N2O from rivers
and estuaries. Environ. Sci.s. 2(2 & 3): 159-165.
Lai, S., and P.K. Patra. 1998. Variabilities in the fluxes and annual emissions of nitrous oxide from the
Arabian Sea. Global Biogeochem. Cycles 12(2): 321-327.
Lambert, G., and S. Schmidt. 1993. Reevaluation of the oceanic flux of methane: Uncertainties and long
term variations. Chemosphere. 26(1-4): 579-589.
Law, C.S., and N.J.P. Owens. 1990. Significant flux of atmospheric nitrous oxide from the northwest
Indian Ocean. Nature. 346:826-828.
Lelieveld, J., P.J. Crutzen, and F.J. Dentener. 1998. Changing concentration, lifetime and climate forcing
of atmospheric methane. Tellus 50B: 128-150.
Lueker, T.J., S.J. Walker, M.K. Vollmer, R.F. Keeling, C.D. Nevison, and R.F. Weiss. 2003. Coastal
upwelling air-sea fluxes revealed in atmospheric observations of O2/N2, CO2, and N2O. Geophys. Res.
Lett. 30: 1292, 10.1029/2002GL016615.
Middelburg, J.J., J. Nieuwenhuize, N. Iversen, N. H0gh, H. De Wilde, W. Helder, R. Seifert, and O.
Christof 2002. Methane distribution in European tidal estuaries. Biogeochem. 59: 95-119.
Mikaloff Fletcher, S.E., P.P. Tans, L.M. Bruhwiler, J.B. Miller, and M. Hermann, CFL, sources estimated
from atmospheric observations of QL, and its 13C/12C isotopic ratios: 1. Inverse modeling of source
processes. Global Biogeochem. Cycles 18: GB4004,doi:10.029/2004GB002223.
Najjar, R.G., J.L. Sarmiento, and J.R. Toggweiler. 1994. Downward transport and fate of organic matter
in the oceans: Simulations with a general circulation model. Global Biogeochem. Cycles 6: 45-76.
Naqvi, S.W.A., D.A. Jayakumar, P.V. Navekar, H. Nalk, V.V.S.S. Sarma, W. D'Souza, S. Joseph, and
M.D. George. 2000. Increased marine production of N2O due to intensifying anoxia on the Indian
continental shelf. Nature 408: 346-349.
Nevison, C. 1994. A model analysis of the spatial distribution and temporal trends of nitrous oxide
sources and sinks. Ph. D. dissertation. Stanford, CA: Stanford University.
Nevison, C., J.H. Butler, and J.W. Elkins. 2003. Global distribution of N2O and the AN2O-AOU yield in
the subsurface ocean. Global Biogeochem. Cycles 17(4): 1119, 10.1029/2003GB002068.
Nevison, C.D., N.M. Mahowald, R.F. Weiss, and R.G. Prinn. 2007. Interannual and seasonal variability
in atmospheric N2O. Global Biogeochem. Cycles 21: GB3017,doi:10.029/2006GB002755.
Nevison, C.D., R.F. Keeling, RF. Weiss, B.N. Popp, Z. Jin, P.J. Fraser, L.W. Porter, and P.O. Hess.
2005. Southern ocean ventilation inferred from seasonal cycles of atmospheric N2O and O2/N2 at Cape
Grim, Tasmania. Tellus 57B: 218-229.
Nevison, C.D., R.F. Weiss, and D.J. Erickson III. 1995. Global oceanic emissions of nitrous oxide. J.
Geophys. Res. 100 (C8): 15809-15820.
4-22
-------
Chapter 4. Oceans, Estuaries, and Rivers
Nevison, C.D., T.J. Lueker, and R.F. Weiss. 2004. Quantifying the nitrous oxide source from coastal
upwelling. Global Biogeochem. Cycles 18: GB1018, doi:10.029/2003GB002110.
Nihous, G.C., and S.M. Masutani. 2006. A model of methane concentration profiles in the open ocean. J.
Mar. Res. 64: 629-650.
Oremland, R.S. 1979. Methanogenic activity in plankton samples and fish intestines—Mechanism for in
situ methanogenesis in oceanic surface waters. Limnol. Oceanogr. 24(6): 1136-1141.
Oudot, C., P. Jean-Baptiste, E. Fourre, C. Mormiche, M. Guevel, J.F. Ternon, and P. LeCorre. 2002.
TransAtlantic equatorial distribution of nitrous oxide and methane. Deep-Sea Res. I. 49: 1175-1193.
Owens, N.J.P., C.S. Law, R.F.C. Mantoura, P.H. Burkill, and C.A. Llewellyn. Methane flux to the
atmosphere from the Arabian Sea. Nature. 354: 293-296.
Pina-Ochoa, E., and M. Alvarez-Cobelas. 2006. Denitrification in aquatic environments: A cross-system
analysis. Biogeochem. 81: 111-130.
Prinn, R., D. Cunnold, R. Rasmussen, P. Simmons, F. Alyea, A. Crawford, P. Fraser, and R. Rosen. 1990.
Atmospheric emissions and trends of nitrous oxide deduced from 10 years of ALE-GAGE data. J.
Geophys. Res. 95: 18369-18385.
Rehder, G., and E. Suess. 2001. Methane and/?CO2 in the Kuroshio and the South China Sea during
maximum summer surface temperatures. Mar. Chem. 75: 89-108.
Rehder, G., RW. Collier, K. Heeschen, P.M. Kosro, J. Earth, and E. Suess. 2002. Enhanced marine CH4
emissions to the atmosphere off Oregon caused by coastal upwelling. Global Biogeochem. Cycles 16:
1081, 10.1029/2000GB001391.
Rhee, T. S., A. J. Kettle, and M. O. Andreae. 2009. Methane and nitrous oxide emissions from the ocean:
A reassessment using basin-wide observations in the Atlantic. J. Geophys. Res. 114: D12304,
doi:10.1029/2008JD011662.
Robinson, C., A.J. Poulton, P.M. Holligan, A.R Baker, G. Foster, N. Gist, T.D. Jickells, G. Malin, R.
Upstill-Goddard, R.G. Williams, E.M.S. Woodward, and M.V. Zubkov. 2006. The Atlantic Meridional
Transect (AMT) Programme: A contextual view 1995-2005. Deep-Sea Res. II. 53: 1485-1515.
Saarnio, S., W. Winiwarter, and J. Leitao. 2009. Methane release from wetlands and watercourses in
Europe. Atmos. Environ. 43: 1421-1429.
Sansone, F.J., B.N. Popp, A. Gase, A.W. Graham, and T.M. Rust. 2001. Highly elevated methane in the
eastern tropical North Pacific and associated isotopically enriched fluxes to the atmosphere. Geophys.
Res. Lett. 28: 4567-4570.
Sansone, F.J., E. Holmes, and B.N. Popp. 1999. Methane stable isotopic ratios and concentrations as
indicators of methane dynamics in estuaries. Global Biogeochem. Cycles 13(2): 463-474.
Schaefer, S.C., and M. Alber. 2007. Temperature controls a latitudinal gradient in the proportion of
watershed nitrogen exported to coastal ecosystems. Biogeochem. 85: 333-346.
Schlesinger, W.S., K.H. Reckhow, and E.S. Bernhardt. 2006. Global change: The nitrogen cycle and
rivers. Water Res. Res. 42: W03S06, doi:10.1029/2005WR004300.
Scanton, M.I., and P.G. Brewer. 1977. Occurrence of methane in the near-surface waters of the western
subtropical North Atlantic. Deep-Sea Res. 24: 127-138.
Seitzinger, S., J.A. Harrison, J.K. Bohlke, et al. 2006. Denitrification across landscapes and waterscapes:
A synthesis. Ecol. Appl. 16: 2064-2090.
4-23
-------
Chapter 4. Oceans, Estuaries, and Rivers
Seitzinger, S.P. 1988. Denitrification in freshwater and coastal marine ecosystems: Ecological and
geochemical significance. Limnol. Oceanogr. 33: 702-724.
Seitzinger, S.P., and A.E. Giblin. 1996. Estimating denitrification in North Atlantic continental shelf
sediments. Biogeochem. 35: 235-260.
Seitzinger, S.P., and C. Kroeze. 1998. Global distribution of nitrous oxide production and N inputs in
freshwater and coastal marine ecosystems. Global Biogeochem. Cycles 12: 93-113.
Seitzinger, S.P., C. Kroeze, and R.V. Styles. 2000. Global distribution of N2O emissions from aquatic
systems: Natural emissions and anthropogenic effects. Global Change Sci. 2: 267-279.
Slomp, C.P., and P. Van Cappellen. 2004. Nutrient inputs to the coastal ocean through submarine
groundwater discharge: Controls and potential impact. J. Hydrol. 295: 64-86.
Suntharalingam, P., and J.L. Sarmiento. 2000. Factors governing the oceanic nitrous oxide distribution:
Simulations with an ocean general circulation model. Global Biogeochem. Cycles 14(1): 429-454.
Suntharalingam, P., J.L. Sarmiento, and J.R. Toggweiler. 2000 Global significance of nitrous-oxide
production and transport from oceanic low-oxygen zones: A modeling study. Global Biogeochem.
Cycles 14(4): 1353-1370.
Tilbrook, B.D., and D.M. Karl. 1995. Dissolved methane distributions, sources, and sinks from California
coastal waters to the oligotrophic Northern Pacific gyre. Mar. Chem. 49: 51-64.
U.S. EPA (United Status Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Upstill-Goddard, R.C., J. Barnes, and N.J.P. Owens. 1999. Nitrous oxide and methane during the 1994
SW monsoon in the Arabian Sea/northwestern Indian Ocean. J. Geophys. Res. 104(C12): 30067-30084.
Upstill-Goddard, R.C., J. Barnes, T. Frost, S. Punshon, and N.J.P. Owens. 2000. Methane in the southern
North Sea: Low-salinity inputs, estuarine removal, and atmospheric flux. Global Biogeochem. Cycles
14(4): 1205-1217.
Ward, B.B. 1992. The sub-surface methane maximum in the Southern California Bight. Cont. Shelf Res.
12: 735-752.
Wuebbles, D.J., and K. Hayhoe. 2002. Atmospheric methane and global change. Earth-Sci. Rev. 57: 177-
210.
4-24
-------
Chapter 5. Permafrost
Permafrost is soil, sediment, or rock that is continuously frozen (temperature below 0°C) for at least two
consecutive years. It forms when ground is sufficiently cold during winter to remain frozen during the
following summer. It occurs both on land and in sediments of the shelf regions of the Arctic Ocean. On
land, there is a thin surface layer—called the active layer—that thaws each summer, typically reaching a
maximum thickness of about 0.5 to 1.0 meters in late summer. Ground below the active layer can be
perennially frozen to depths of less than 10 meters to more than several hundred meters. Permafrost
occurs primarily at high latitudes, but also at high elevations. The area of permafrost, and its potential as a
source of CH4 and N2O emissions to the atmosphere, are much greater at high latitudes than at high
altitudes (except the Tibetan Plateau), so most field measurements have occurred at high latitudes.
Permafrost regions are classified, based on the prevalence of frozen ground, as continuous (over 90
percent of exposed ground underlain by permafrost), discontinuous (50 to 90 percent), sporadic (10 to 50
percent), or isolated (below 10 percent). Permafrost currently underlies about 23 million km2 of exposed
land surface (i.e., land not under ice sheets) in the Northern Hemisphere (Zhang et al., 1999). Of this,
about 11 million km2 is classified as continuous, and about 4 million km2 each as discontinuous, sporadic,
and isolated (Zhang et al., 1999). Note that, for example, 4 million km2 of land with isolated permafrost
has less than 10 percent of its area, or less than 400,000 km ,of actual permafrost.
,**
Figure 5-1. Distribution of permafrost and ground ice in the Northern Hemisphere, based on the EASE-Grid version
of the International Permafrost Association (IPA) map. "High," "Med," and "Low" refer to ice content, and "T" and "t"
refer to thick and thin overburden, respectively, image courtesy of the IPA, supplied by Tingjun Zhang, National Snow
and Ice Data Center, University of Colorado, Boulder.
5-1
-------
Chapter 5. Permafrost
5.1 Description of Emission Source
CH4 or N2O can be frozen within permafrost, and thus permafrost represents a stock or reservoir
and N2O that can be released upon thawing. This chapter focuses on that potential source, but also briefly
addresses the role of permafrost in other natural source categories covered in other chapters (Chapter 6,
"Lakes"; Chapter 2, "Wetlands"; Chapter 4, "Oceans"). Gas hydrates or clathrates, which can co-occur
with permafrost but which also are common in non-permafrost regions, are discussed as a QrU and N2O
source in Chapter 7, "Gas Hydrates," and the presence of thermogenic (non-microbial) CH4 in deeper
permafrost strata has been noted (Collett and Dallimore, 1999; Yakushev and Chuvilin, 2000); geologic
sources of ClrU are discussed in Chapter 8.
There have been no direct field measurements of QrU or N2O emissions from permafrost, because gas
losses are negligible while the permafrost remains frozen (Rivkina et al., 2004, 2007). However, when
permafrost melts, gas bound within the frozen soil is released, and can then diffuse to the surface and be
emitted to the atmosphere. Since permafrost, when thawing from above, is overlain by a seasonally
thawed active layer, field measurements of surface emissions can include QrU and N2O generated in the
active layer and/or QrU and N2O released from the thawing permafrost; little effort has been made to
discriminate between these two possibilities. Permafrost can also thaw at its base, due to geothermal heat
flux from deeper in the earth (temperatures increase with depth into the earth), but gas released by this
thawing is likely to remain trapped under the overlying frozen ground. Studies of permafrost as a direct
greenhouse gas source have included (1) measurements of greenhouse gas concentrations in frozen
samples of permafrost collected from below the active layer; (2) incubations of these samples at a range
of temperatures, above and below freezing, to measure potential greenhouse gas production; and (3)
modeling studies of permafrost thaw (over centuries) to estimate the amount of gas that might be released
during future climate change.
Permafrost was included among the natural sources of QrU reviewed in the earlier version of this report
(U.S. EPA, 1993) when some early measurements of QrU concentration in permafrost were just being
reported. The U.S. EPA (1993) reported the magnitude of emissions to be highly uncertain and probably
relatively small. However, permafrost has not been listed as a separate, specific source of ClrU or N2O in
the IPCC assessments of 1990, 2001, or 2007 (Watson et al., 1990; Prather et al., 2001; Lemke et al.,
2007), but CH4 emissions are discussed in the U.S. Climate Change Science Program Synthesis and
Assessment Product 3.4 on abrupt climate change (Brook et al., 2008). As discussed elsewhere (e.g.,
Section 6.1), there can be overlap between the source categories this report is organized around, and
operationally these overlapping or co-occurring sources may be difficult to distinguish in the field (e.g., if
permafrost and gas hydrates are co-located).
5.2 Factors That Influence Emissions
Gas diffusion rates in frozen permafrost are very slow (Rivkina et al., 1998), so the release of greenhouse
gases directly from permafrost occurs when permafrost melts. This can happen as a result of climatic
warming, surface disturbance that changes the land surface energy balance (e.g., fire), exposure of
permafrost due to thermokarst or coastline erosion, and new construction (e.g., roads and buildings) that
changes the surface energy balance, although most construction now is designed to preserve the
underlying permafrost in order to increase the future integrity of the structures. Thermokarst erosion
occurs when ice-rich permafrost melts, and the land surface slumps or collapses. The rate of melting (or
formation) of permafrost can be highly variable, as it is a function of several factors: the land surface
energy balance, the depth of the permafrost in the ground, and the thermal properties (i.e., heat
conductivity and heat capacity) of the permafrost and overlying ground.
The quantity of CIL, or N2O released from permafrost also depends directly on several factors. First, it
depends on the total extent of permafrost, which is a primarily a function of climate (Zhang et al., 1999).
5-2
-------
Chapter 5. Permafrost
Second, it depends on the concentration of CFU or N2O in the permafrost. Finally, it depends on the rate at
which the permafrost is melting, which (as noted above) itself can depend on several factors.
An important consideration that has received little study is the transport of greenhouse gases from the
melting permafrost, through the overlying melted soil or sediment, to the atmosphere, and potential
transformations of these gases before reaching the atmosphere (e.g., oxidation of CH4 to CO2 by
methanotrophic (or "methane-eating") bacteria, or reduction of N2O to N2 by denitrifying bacteria in the
unfrozen soil column above the permafrost). The amount of gas reaching the atmosphere (net emission)
could be significantly less than the amount released from the thawing permafrost (Kvenvolden et al.,
1993). For example, in non-permafrost systems that emit CFi4, such as wetlands and rice paddies, a
substantial fraction of the CFi4 produced in the flooded soil is oxidized in overlying oxic soils and fresh
water (i.e., soils and fresh water containing enough dissolved oxygen to support aerobic organisms)
before reaching the atmosphere (Reeburgh, 2004).
The high-latitude permafrost region has been warming over the past several decades, but this warming has
not been spatially or temporally uniform (ACIA, 2004). Permafrost has been observed to be degrading
(thawing) over the past several decades in several regions (e.g., Christensen et al., 2004; Camill et al.,
2001), but at this time there is no large-scale quantification of the rate of melt. These changes in
permafrost distribution and prevalence, and in active layer depth, will impact regional hydrology,
vegetation distributions, ecosystem productivity, and soil organic matter decomposition rates. These
landscape effects of melting permafrost in turn will impact net greenhouse gas emission rates from
wetlands, soils, and lakes in permafrost landscapes. These indirect emissions are not considered in this
chapter.
5.3 Current Global Emissions
In permafrost regions, the ground surface thaws each summer (this is called the active layer), so the actual
permafrost is below the ground surface. Direct CFI4 or N2O emissions from in situ permafrost are not
measured, and estimates of emissions are based on measured concentrations of CFt^ in permafrost, and
estimates of contemporary permafrost degradation rates.
5.3.1 Methane
Similar ranges in permafrost CFI4 concentrations have been measured near Prudhoe, Alaska (0.4-8.7 mg
CFLAg: Rasmussen et al., 1993), near Fairbanks, Alaska (<0.001-22 mg CFIVkg: Kvenvolden et al.,
1993), and in northeastern Siberia (0-6 mg CFLAg: Rivkina et al., 2004). Brouchkov and Fukuda (2002)
also measured permafrost CFLj in northeastern Siberia, but their reported values (2-6,000 ppmv in bubbles
in permafrost) cannot be converted into mg per kg. Moraes and Khalil (1993) and Kvenvolden et al.
(1993) used a model of heat transfer into permafrost soils to estimate climate change impacts on CFU
emission from permafrost (discussed in more detail in Section 5.4, "Future Emission Scenarios"). These
studies did not report or discuss contemporary emissions, but the initial emission rates of their future
scenarios provide an estimate of current emissions. Both studies estimated that contemporary CFLt release
from permafrost is about 0 to 1 Tg CFL/yr. These simulations did not include CH4 oxidation during
transport from the melting permafrost to the surface, and so these are likely to be an overestimate or an
estimate of the maximum net emission to the atmosphere.
5.3.2 Nitrous Oxide
Global emissions of N2O from permafrost are negligible. Very few studies have measured N2O
concentrations in permafrost. Rasmussen et al. (1993) collected shallow-core permafrost samples from
Alaska's North Slope, near Prudhoe Bay, and measured N2O concentrations of ug/kg, or parts per billion
by weight, about 1,000 times smaller than for CFU (mg/kg, or parts per million by weight). Using this
factor of 1,000 and the CFLj emission rate estimated from the same study (Moraes and Khalil, 1993), the
5-3
-------
Chapter 5. Permafrost
current N2O flux would be about 1 Gg N2O/yr, or about 0.005 percent of the total annual global N2O flux
from all sources. The very low concentrations reported by Rasmussen et al. (1993) are consistent with
biogeochemical measurements made on permafrost samples from northern Siberia by Rivkina et al.
(1998). They observed that viable microbes (including denitrifiers, bacteria that generate N2O) were
present in permafrost samples collected from 1 meter to about 35 meters below the surface, and that
sample Eh (or oxidation-reduction potential) was +40 to -250 mV. Eh values in this range indicate highly
reduced conditions, in which most nitrogen has been reduced by denitrifiers to N2, and typical N2O
concentrations are very low. Low N2O concentrations in permafrost are also consistent with negligible
N2O production measured in incubations of soil samples collected from the active layer and shallow
permafrost in northern Siberia by Rodionow et al. (2006).
5.3.3 Indirect Impacts of Permafrost on Methane Emissions
Permafrost has a number of indirect effects on CIL, emissions. These are not included in this chapter's
quantification of direct emissions from permafrost.
Some scientists consider permafrost to be the nominal source of greenhouse gases emitted when organic
matter frozen into permafrost decomposes after thawing (e.g., Zimov et al., 1997). A number of studies
have established that the old soil organic matter frozen into permafrost readily decomposes if thawed, and
that microbial populations that can decompose the organic matter are present and viable in the permafrost
(Rivkina et al., 1998, 2004, 2007; Zimov et al., 1997, 2006). It has been demonstrated in laboratory
incubations that microbial metabolism and CIL, production can occur, albeit at very low rates, with soil
samples held at temperatures as low as -16°C (Brouchkov and Fukuda, 2002; Rivkina et al., 2004). It is
likely, however, that most of the CtL, contained in permafrost was produced before the soils froze (e.g., in
the Pleistocene), that the CH4 froze into the permafrost as the permafrost developed, and that the CtL, has
remained stably bound in the permafrost for thousands of years (Rivkina et al., 2007).
Permafrost in the ground acts as an impervious boundary to water infiltration, and thus can generate
wetter soils and the conditions necessary for denitrification (which can generate N2O) and CtL,
production. Melting permafrost leads to changes in land surface hydrology, which, together with the
warming that caused the melting, affects emission rates (e.g., Christensen et al., 2004; Van Huissteden et
al., 2005; Tarnocai et al., 2007). The effect of hydrology on N2O and CtL, emissions is discussed in more
general terms in Chapter 2, "Wetlands," and Chapter 3, "Soils"), which also include quantitative
summaries of emissions.
Ice-rich permafrost has large ice wedges (tens of centimeters to several meters in scale) within the frozen
soil and rock. As these ice wedges melt, the water can drain away, leaving voids; these can collapse,
causing ground surface subsidence and the formation of small lakes and ponds called thermokarst lakes.
Measurements in northeastern Siberia have shown that these thermokarst lakes can be significant sources
of CH4 to the atmosphere (Zimov et al., 1997; Walter et al., 2006, 2007); this source is discussed and
quantified in Chapter 6, "Lakes." Coastal erosion of permafrost can also lead to the release of CH4 into
coastal waters; Shakhova et al. (2005) measured very high CUt concentrations (supersaturation) in Arctic
Ocean coastal waters (East Siberian and Laptev Seas on the Siberian coast). Shakhova et al. (2005)
suggest four possible mechanisms for these high CH4 concentrations: (1) coastal erosion and the
subsequent release of CUt trapped in permafrost ice; (2) release of CUt trapped in sub-sea permafrost; (3)
release of CtL, trapped in shallow gas hydrates; and (4) biogenic CtL, generated from decomposition of
eroded carbon in sub-sea sediments. They did not measure CtL, fluxes.
Melting of permafrost will likely be accompanied by hydrate destabilization in regions where permafrost
and hydrates co-occur (e.g., Harvey and Huang, 1995). Emissions to the atmosphere from this situation
are likely to be dominated by hydrate sources, which make up a substantially larger pool; this source is
discussed and quantified in Chapter 7, "Gas Hydrates."
5-4
-------
Chapter 5. Permafrost
5.4 Future Emission Scenarios
Like the contemporary emission estimates discussed above, estimates of future emissions of CFL, or N2O
from permafrost are based on estimates of climate-change-induced melting of permafrost, and the
concentrations of CFL, or N2O in that permafrost.
5.4.1 Methane
Moraes and Khalil (1993) and Kvenvolden et al. (1993) estimated future CFL, emissions from permafrost
melting for the next several hundred years. Their estimates were based on several factors: (1) air
temperature increases of 5° and 8.5°C in 2100, (2) permafrost CFL, concentrations measured in Alaska
(Rasmussen et al., 1993; Kvenvolden et al., 1993), and (3) one-dimensional modeling of soil heat transfer
(Moraes and Khalil, 1993). Both studies concluded that emissions would rise slowly from current low
values (about 0 to 1 Tg CHVyr) to rates of 5 to 30 Tg CHVyr over the next 30 to 60 years, and then
decline back to low values within another 50 to 100 years. Neither of these analyses included any CIL,
oxidation during the transport of CIL, from the melting permafrost through the overlying thawed soil to
the atmosphere. As permafrost melts, the active layer eventually thickens to at least tens of meters in most
regions, and the potential for CIL, oxidation before reaching the atmosphere is high. In addition, neither
model included any changes in soil thermal dynamics due to changing vegetation cover (e.g., see
Christensen et al., 2004) and/or hydrological regimes (e.g., see Smith et al., 2005); these changes can
either enhance or diminish the rate of thawing.
More recently, estimates of global permafrost melting were made with a land-surface model coupled to a
general circulation model (CLM3 in the CCSM3; Lawrence and Slater, 2005). In these simulations,
permafrost area declined by 60 to 90 percent by 2100, depending on the greenhouse gas emission scenario
used to drive the model. Under the assumption that CtL, emissions are proportional to rate of permafrost
decline, the permafrost decline simulated by Lawrence and Slater (2005) would generate similar results to
the earlier studies of Moraes and Khalil (1993) and Kvenvolden et al. (1993), again noting that changes in
vegetation and hydrology were not taken into account in the simulations reported by Lawrence and Slater.
Permafrost melting may also indirectly have significant impacts on landscape-scale CtL, emissions to the
atmosphere, through related changes in vegetation and hydrology (e.g., see Christensen et al., 2004).
Future emissions from permafrost melting will be directly related to the rate and magnitude of climatic
warming at high latitudes. Direct net CFL emissions from permafrost melting are not likely to be very
high, though they would be a small positive feedback on the climate system. Given other, more important
issues related to high-latitude warming (including potential indirect effects of permafrost melting), direct
net CFL, emissions from permafrost are not a key issue in climate policy development (Brook et al., 2008).
5.4.2 Nitrous Oxide
Given the small stock of N2O frozen into permafrost in northern Alaska (Rasmussen et al., 1993) and the
negligible N2O production measured in incubations of permafrost and active layer soils in northern
Siberia (Rodionow et al., 2006), it is unlikely that future emissions of N2O from permafrost will be
significant over at least the next several decades to hundred years. It is important to note, however, that
this conclusion is based on a very limited number of field studies.
5.5 Areas for Further Research
Recent modeling of permafrost melting (e.g., Lawrence and Slater, 2005) has a more realistic treatment of
the soil physical processes involved, and much more detail about their variability across permafrost
landscapes, than the heat transfer modeling of Moraes and Khalil (1993) that drove the initial estimates of
CFL, release. However, even the recent models lack detailed treatment of vegetation and landscape
5-5
-------
Chapter 5. Permafrost
hydrology changes that will occur with high-latitude warming, and that can have both significant direct
and indirect impact on CIL, emissions. A second significant uncertainty is the estimation of the fraction of
CH4 released from melting permafrost that will be oxidized (to CO2) before reaching the atmosphere. The
only global estimates of direct CIL, emissions from melting permafrost did not account for any oxidation.
In some cases, however, particularly for deeper permafrost, most of the CIL, is likely to be oxidized
before reaching the atmosphere.
Fully addressing these issues will require sophisticated models that include several features not common
in current models: (1) lateral movement of surface water, soil water, and shallow groundwater (as
opposed to just vertical percolation); (2) dynamic vegetation algorithms to simulate changes in plant
community composition; and (3) detailed soil physics, including gas transport, to simulate CH^ oxidation.
This type of model will also be better able to estimate indirect impacts of permafrost melting on CFLj
emissions. At this time, it is not known whether direct or indirect effects will have a stronger impact on
net CFLt emissions.
Several field studies indicate that N2O emissions from melting permafrost, both directly and indirectly,
are likely to be very low. However, very few field studies have been conducted, and so it is difficult to
even estimate the uncertainty of this conclusion.
5.6 References
ACIA (Arctic Climate Impact Assessment). 2004. Impacts of a Warming Arctic: Arctic Climate Impact
Assessment. Cambridge University Press.
Brook, E., D. Archer, E. Dlugokencky, S. Frolking, D. Lawrence. 2008. Chapter 5. Potential for abrupt
changes in atmospheric methane. In: U.S. Climate Change Science Program Synthesis and Assessment
Product 3.4: Abrupt Climate Change. [http://www.climatescience.gov/Library/sap/sap3-4/final-report/]
Brouchkov, A., M. Fukuda. 2002. Preliminary measurements on methane content in permafrost, Central
Yakutia, and some experimental data. Permafrost andPeriglacialProcesses 13: 187-197.
Camill, P., J.A. Lynch, J.S. Clark, J.B. Adams, and B. Jordan. 2001. Changes in biomass, aboveground
net primary production, and peat accumulation following permafrost thaw in the boreal peatlands of
Manitoba, Canada. Ecosystems 4: 461-478.
Christensen, T.R., T. Johansson, H.J. Akerman, M. Mastepanov, T. Firborg, P. Crill, B.H. Svensson.
2004. Thawing sub-arctic permafrost: Effects on vegetation and methane emissions. Geophysical
Research Letters 31: L04501, doi:10.1029/2003GL018680.
Collett, T.S., and S.R. Dallimore. 1999. Hydrocarbon gases associated with permafrost in the Mackenzie
Delta, Northwest Territories, Canada. Applied Geochemistry 14: 607-620.
Harvey, L.D.D., and Z. Huang. 1995. Evaluation of the potential impact of methane clathrate
destabilization on future global warming. Journal of Geophysical Research-Atmospheres 100(D2):
2905-2926.
Kvenvolden, K.A., and T.D. Lorenson. 1993. Methane in permafrost—Preliminary results from coring in
Fairbanks, Alaska. Chemosphere 26(1-4): 609-616.
Lawrence, D.M., and A.G. Slater. 2005. A project of severe near-surface permafrost degradation during
the 21st century. Geophysical Research Letters 32, L24401, doi.10.1029/2005GL025080.
Lemke, P., J. Ren, R.B. Alley, I. Allison, J. Carrasco, G. Flato, Y. Fujii, G. Kaser, P. Mote, R.H. Thomas
and T. Zhang. 2007. Observations: changes in snow, ice and frozen ground. In: S. Solomon, D. Qin, M.
Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate Change 2007:
The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the
5-6
-------
Chapter 5. Permafrost
Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press.
Moraes, F., and M.A.K. Khalil. 1993. Permafrost methane content: 1. Modeling theory and results.
Chemosphere 26(1-4): 595-607.
Prather, M., Ehhalt D., Dentener F., Derwent R., Dlugokencky E., Holland E., Isaksen I., Katima J.,
Kirchhoff V., Matson P., Midgley P., Wang M. 2001. Atmospheric chemistry and greenhouse gases. In:
J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A.
Johnson. Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge
University Press, pp. 239-287.
Rasmussen, R.A., M.A.K. Khalil, and F. Moraes. 1993. Permafrost methane content: 1. Experimental data
from sites in northern Alaska. Chemosphere 26(1-4): 591-594.
Reeburgh, W.S. 2004. Global methane biogeochemistry. In: R.F. Keeling (ed.) Treatise on Geochemistry,
Volume 4: The Atmosphere. Amsterdam: Elsevier. pp. 65-89.
Rivkina, E., D. Gilichinsky, S. Wagener, J. Tiedje, J. McGrath. 1998. Biogeochemical activity of
anaerobic microorganisms from buried permafrost sediments. Geomicrobiology 15: 187-193.
Rivkina, E., K. Laurinavichius, J. McGrath, J. Tiedje, V. Shcherbakova, D. Gilichinsky. 2004. Microbial
life in permafrost. Advances in Space Research 33: 1215-1221.
Rivkina, E., V. Shcherbakova, K. Laurinavichius, L. Petrovskaya, K. Krivushin, G. Kraev, S.
Pecheritsina, and D. Gilichinsky. 2007. Biogeochemistry of methane and methanogenic archaea in
permafrost. FEMSMicrobiology Ecology 61(1): 1-15.
Rodionow, A., H. Flessa, O. Kazansky, and G. Guggenberger. 2006. Organic matter composition and
potential trace gas production of permafrost soils in the forest tundra of northern Siberia. Geoderma
135: 49-62.
Shakhova, N. I. Semiletov, and G. Panteleev. 2005. The distribution of methane on the Siberian Arctic
shelves: Implications for the marine methane cycle. Geophysical Research Letters 32: L09061,
doi.10.1029/2005GL022751.
Smith, L.C., Y. Sheng, G.M. MacDonald, and L.D. Hinzman. 2005. Disappearing Arctic lakes. Science
308: 429.
Tarnocai, C., C.-L. Ping, and J. Kimble. 2007. Carbon cycles in the permafrost region of North America.
In: A.W. King, L. Dilling, G.P. Zimmerman, D.M. Fairman, R.A. Houghton, G. Marland, A.Z. Rose,
and T.J. Wilbanks (eds.). The First State of the Carbon Cycle Report (SOCCR): The North American
Carbon Budget and Implications for the Global Carbon Cycle. A report by the U.S. Climate Change
Science Program and the Subcommittee on Global Change Research, Washington, DC. {Draft Four:
Subsequent From Government Review}
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Van Huissteden, J., T.C. Maximov, and A.J. Dolman. 2005. High methane flux from an Arctic floodplain
(Indigirka lowlands, eastern Siberia). J. Geophysical Research 110: G02002,
doi.10.1029/2005JG000010.
Walter, K.M., L.C. Smith, and F.S. Chapin III. 2007. Methane bubbling from northern lakes: Present and
future contributions to the global methane budget. Philosophical Transactions of the Royal Society A
365: 1657-1676.
5-7
-------
Chapter 5. Permafrost
Walter, K.M., S.A. Zimov, J.P. Chanton, D. Verbyla, and F.S. Chapin III. 2006. Methane bubbling from
Siberian thaw lakes as a positive feedback to climate warming. Nature 443: 71-75.
Watson, R.T., H. Rodhe, H. Oeschger, and U. Siegenthaler. 1990. Greenhouse gases and aerosols. In: J.T.
Houghton, G.J. Jenkins, and J.J. Ephraums (eds.). Climate Change: The IPCC Scientific Assessment.
Cambridge, UK: Cambridge University Press, pp. 1-40.
Yakushev V.S., and E.M. Chuvilin. 2000. Natural gas and gas hydrate accumulations within permafrost in
Russia. Cold Regions Science and Technology 31: 189-197.
Zhang T., R.G. Barry, K. Knowles, J.A. Heginbottom, and J. Brown. 1999. Statistics and characteristics
of permafrost and ground ice distribution in the Northern Hemisphere. Polar Geography 23(2): 132-
154.
Zimov, S.A., S.P. Davydov, G.M. Zimova, A.I. Davydova. E.A.G. Schuur, K. Durra, and F.S. Chapin III.
2006. Permafrost carbon: Stock and decomposability of a globally significant carbon pool. Geophysical
Research Letters 33: L20502, doi.10.1029/2006GL027484.
Zimov, S.A., Y.V. Voropaev, I.P. Semiletov, S.P. Davydov, S.F. Prosiannikov, F.S. Chapin III, M.C.
Chapin, S. Trumbore, and S. Tyler. 1997. North Siberian lakes: A methane source fueled by
Pleistocene carbon. Science 277: 800-802.
5-8
-------
Chapter 6. Lakes
Lakes and ponds are naturally formed permanent water bodies contained on a body of land. (This report
only addresses freshwater lakes.) Recently, Downing et al. (2006) estimated that there are about 300
million natural lakes and ponds globally, with a total area of about 4.2 million km , or about 3 percent of
the global land surface area. Ninety percent of these lakes are smaller than 1 hectare (0.01 km2), and 43
percent of the total lake area is from lakes smaller than 1 km2 (Downing et al., 2006). Downing et al.
(2006) estimated that impoundments with large, engineered dams cover about 0.25 million km and that
small farm impoundments occupy an additional 0.1 million km2, but some earlier estimates of
impoundment/reservoir area are substantially higher (e.g., St. Louis et al., 2000).
In the 1990 IPCC assessment, tabulated sources of CFL include both natural wetlands and fresh waters
(Watson et al., 1990), with freshwater source strength estimated at 5 Tg CHVyr (estimated range of 1-25
Tg CFL/yr). This freshwater source could include natural lakes, impoundments/reservoirs, and rivers, but
no specifics were discussed. Cicerone and Oremland (1988) noted that this range came from a very early
estimate by Ehhalt (1974) and was very uncertain. In the 2001 and 2007 IPCC assessments, the only
comparable CFL source is natural wetlands (Prather et al., 2001; Denman et al., 2007). Lakes are not
tabulated as a source of N2O in the IPCC assessments of 1990, 2001, or 2007, although the 2007
assessment includes rivers/estuaries/coastal zones (Watson et al., 1990; Prather et al., 2001; Denman et
al., 2007).
6.1 Description of Emission Source
This quantification of freshwater lakes as a source of CH4 and N2O to the atmosphere includes both lakes
and ponds, but excludes impoundments/reservoirs. Gas emissions from impoundments/reservoirs, water
bodies formed by dams and other engineering works, are considered to be anthropogenic. (Although not
discussed further here, reservoirs can be a significant source of greenhouse gases; see, for example, St.
Louis et al., 2000.)
There is the potential for overlap between natural lakes and wetlands as methane and N2O sources for
several reasons. Some (but not all) wetland area inventories include shallow lakes as wetlands (Matthews,
2000), but lake area estimates (e.g., Downing et al., 2006) also include shallow lakes. The littoral (or
shoreline) zone of many lakes has emergent vegetation, and could be considered wetland in some
inventories. Juutinen et al. (2003) define the littoral zone as the area between the highest shoreline and the
outer limit of floating-leaved vegetation; this can include temporarily flooded grass and sedge-dominated
zones, continuously inundated reed-dominated zones, and slightly deeper zones dominated by floating
vegetation. Many wetlands contain numerous small ponds (e.g., Hamilton et al., 1994; Repo et al., 2007),
and in scaling up wetland area these ponds are likely to be included, but they may also be included in lake
and pond inventories such as that of Downing et al. (2006). Beaver ponds are small, temporary, natural
impoundments that can be sources of CFL (e.g., Naiman et al., 1991; Bubier et al., 1993; Roulet et al.,
1997) and that can be interspersed among wetland or rivers, and difficult to map separately from them.
Finally, in a tropical regional assessment of CFL emissions for the floodplains of the central Amazon
Basin, Melack et al. (2004) used remote sensing to quantify the area of different landscape classes. They
noted that there is much seasonal inundation of the major and minor river floodplains. Their "wetland
habitat" classes include open water (lakes and channels), flooded forests, and aquatic macrophytes
(floating emergent herbaceous plants during inundated periods). Some fraction of this could be classified
as lake, some as river, and some as wetland.
6-1
-------
Chapter 6. Lakes
6.2 Factors That Influence Emissions
6.2.1 Methane
In lakes, QrU is produced by methanogenic or methane-producing bacteria in anoxic (oxygen-free)
sediments (Cicerone and Oremland, 1988; Conrad, 1989; Kiene, 1991). CH4 production rates depend on
temperature, organic matter availability, and isolation from oxygen; these factors are influenced by
climate, lake size and depth, and productivity of microscopic and macroscopic plants and animals. The
death of these plants and organisms generates organic matter that serves as a substrate (or 'food') for
methane production in the sediments. Roulet et al. (1997) made high-frequency, near-continuous ClrU flux
measurements from a beaver pond in Central Canada. They found that daytime QrU flux was highly
correlated with sediment temperature and inversely correlated with dissolved oxygen concentration, and
that nighttime flux was most correlated with wind speed. Oxidation in overlying oxic sediments and/or
lake water can consume a large fraction of the ClrU generated in the sediments, and reduce the net flux to
the atmosphere (e.g., Reeburgh, 2003; Kiene, 1991;
Galchenko et al., 1989).
There are four pathways for CH4 emissions from lakes
(e.g., see Bastviken et al., 2004; Whalen, 2005):
bubbling, diffusion, plant-mediated transport, and
seasonal overturning (see text box). In their synthesis
study, Bastviken et al. (2004) refer to the seasonal
overturning flux as a storage flux (i.e., ClrU is stored in
deeper lake water during a season, then emitted
episodically upon overturning). Seasonal lake
overturning can occur in both spring and autumn. In
spring, overturning occurs following ice-out (e.g., see
Rudd and Hamilton, 1978; Striegl and
Michmerhuizen, 1998). Autumn overturn occurs in
northern lakes when thermal stratification established
during the summer breaks down. During summer
months when incoming solar radiation is high, the
surface water is warmed. This makes it less dense than
the colder deeper waters, and it floats above the denser
deep water (this is thermal stratification). As the
surface waters cool in the autumn, the thermal
stratification breaks down, and the deeper water,
which can have relatively high dissolved QrU
concentrations, mixes to the surface. This ClrU is then
emitted to the atmosphere (e.g., see Kankaala et al.,
2007). Michmerhuizen et al. (1996) estimated that up
to 40 percent of the annual QrU flux from small lakes
could occur during spring turnover.
Wind speed is an important control on gas exchange between a lake and the atmosphere (Sebacher et al.,
1983). Flux rates by all pathways generally increase with increasing wind speed.
Four pathways for CH4 emissions from lakes
•Bubbling (orebullition), in which bubbles
contain substantial methane concentrations
and transport this methane from a lake's
sediments (bubble formation) to the
atmosphere (bubble bursting at the water
surface).
• Plant-mediated transport. Aquatic plants
often contain arenchymous tissue (spongy
tissue with large pores) in their stems and roots
that allows air to move quickly between the leaf
surface and the roots. These gas-conducting
voids provide a conduit for diffusive flux from
the sediment to the atmosphere that bypasses
diffusion through the lake water itself.
• Diffusive emissions from surface water
whose methane concentration is greater than
the concentration that would be in equilibrium
with the overlying atmosphere.
•Seasonal lake-overturning, in which methane
builds up in deeper lake water during a season
in which the lake water is stratified and not
well-mixed, and then is emitted episodically
when there is lake water overturning.
Bubbling has been determined to be the dominant pathway for CFLt flux (more than 90 percent) in a
variety of field studies in both the tropics (Keller and Stallard, 1994) and the permafrost zone (Walter et
al., 2006). In their review of temperate and boreal lake emissions, Bastviken et al. (2004) estimate bubble
fluxes to account for roughly half of the ClrU flux for lakes with areas ranging from 10" to 10 km . Water
depth, and therefore water pressure, are important factors controlling ebullition (Keller and Stallard,
1994).
6-2
-------
Chapter 6. Lakes
Walter et al. (2006) measured extremely high CH4 fluxes from thermokarst lakes in northeastern Siberia,
following up on earlier work by Zimov et al. (1997). Thermokarst (or thaw) lakes form in permafrost
regions when massive ground-ice wedges melt, causing the ground surface to subside and lakes to form.
Further ground melting can eventually lead to drainage and disappearance of thermokarst lakes (Smith et
al., 2005). In Siberia, there are huge quantities of soil organic carbon (SOC) frozen into permafrost in
which these thermokarst lakes form (about 450 Pg C; Zimov et al., 2006), and this SOC has been shown
to be readily decomposable and contributing to the observed CFLj flux (Walter et al., 2006).
University of Alaska Fairbanks researcher Katey Walter
lights a pocket of methane on a thermokarst lake in
Siberia in March of 2007. Igniting the gas is a way to
demonstrate, in the field, that it contains methane.
Credit: Photo by Sergey Zimov
Many studies have shown that QrU fluxes are generally higher from regions with rooted or floating
vegetation (e.g., Juutinen et al., 2003; Melack et al., 2004). Kankaala et al. (2005) found that seasonal
variation in QrU flux from a boreal lake was more correlated with plant growth than variation in sediment
temperature for sites with relatively low emission rates, while sites with dense vegetation and high flux
rates showed a seasonal flux rate variation more correlated with sediment temperature. Since vegetation is
more common in shallower water, the ratio of shoreline length to lake area is an important factor
influencing mean lake emissions per unit area (Bergstrom et al., 2007), and this ratio generally increases
as the total lake area decreases.
Stadmark and Leonardson (2005), measuring ClrU emissions from shallow ponds constructed in southern
Sweden for nitrogen retention/removal, found that QrU fluxes were strongly dependent on bottom water
temperature.
Lake Baikal has been shown to emit methane derived from gas seeps (Schmid et al. 2007). Accordingly,
an unknown but likely small fraction of lake fluxes may belong to the "geologic source" category
(Chapter 8); attribution of bubble fluxes can be difficult without CFL. isotopic analyses.
6-3
-------
Chapter 6. Lakes
6.2.2 Nitrous Oxide
N2O is produced by microbial activity in soils, sediments, and water. Under aerobic conditions, N2O is an
intermediate product of nitrification (the oxidation of ammonium to nitrate); under mildly anaerobic
conditions N2O is an intermediate product of denitrification (the reduction of nitrate to dinitrogen, or N2).
In general, high rates of N2O production are more commonly associated with denitrification than
nitrification (Firestone and Davidson, 1989). In lakes, denitrification occurs under several conditions: in
aquatic sediments with overlying oxic water, in which nitrate is generated in a thin layer of oxic
sediments and diffuses down to a thin band of sediments in which it can be denitrified; in seasonally
stratified aquatic systems, in which nitrate produced in situ or from external inputs is available for
denitrification when the water stratifies and the bottom layer becomes anoxic; and in permanently
stratified (or rarely mixed) aquatic systems with anoxic bottom water, where nitrate diffuses from
overlying oxic water (Seitzinger et al., 2006). However, in a study of 15 Swiss lakes of varying size and
trophic or nutrient status, Mengis et al. (1997) found that N2O was produced in three zones: (1) near the
surface and attributed to actively growing algae and/or co-existing denitrifiers, (2) in the oxic
hypolimnion (or water below the thermocline or thermal boundary in thermally stratified lakes) and
attributed to nitrification, and (3) at oxic/anoxic interfaces in the hypolimnion and attributed to both
denitrification and nitrification. They also found that denitrification consumes N2O (i.e., denitrifiers
further reduce N2O to N2) in the anoxic hypolimnia/sediment.
Huttunen et al. (2003) found that most N2O flux from a boreal lake came from the littoral (or shoreline
vegetated) zone. Fluxes were highest from the eulittoral (or temporarily flooded) zone, and lower from
the infralittoral (or permanently flooded) zone. They refer to the temporarily flooded sites as meadow and
marsh, and these may be classified as wetlands, while the permanently flooded and vegetated sites are
reed and water lily, and are more likely to be considered as lake sites.
Because nitrification and denitrification are highly sensitive to oxygen availability, oxygen concentration
is an important factor in the balance between the two processes. Nitrogen levels in available forms such as
ammonium and nitrate are also an important control. Pina-Ochoa and Alvarez-Cobelas (2006)
summarized literature reports on denitrification rates in aquatic ecosystems. Annual rates of
denitrification were found to be higher in lakes than in rivers, coastal ecosystems, and estuaries. Across
ecosystems, rates were correlated with nitrate levels and inversely correlated with O2 levels. Stadmark
and Leonardson (2005), measuring N2O emissions from shallow (<1.5m) ponds constructed in southern
Sweden for nitrogen retention/removal, found that N2O fluxes from the ponds were below their detection
limit, but that N2O production in sediment and water incubations increased with increasing nitrate
concentration. Liikanen et al. (2002) found a similar effect in incubations of shallow (4m) and deep (8m)
sediments from a eutrophic or nutrient-rich lake in Finland.
6.3 Current Global Emissions
6.3.1 Methane
As noted above, early estimates of 5 Tg CHVyr emitted from lakes/freshwater (range 1 to 25 Tg CHVyr)
were based on very limited data (Cicerone and Oremland, 1988). Khalil and Shearer (2000) included
lakes as one of a group of 10 "minor sources that each emit between 10-50 Tg CHVyr, but collectively are
a significant fraction of the global budget," but provided no additional assessment.
Bastviken et al. (2004) estimated global CFLj emissions from lakes at 8 to 48 Tg CHVyr, with 6 to 25 Tg
CHVyr from open water and 2 to 23 Tg CHVyr of plant-mediated flux. Their estimate is based on CFLt
flux relationship to lake area, based on fluxes from about 20 field studies, and a global lake area database
from Kalff (2002). There are two reasons to re-evaluate this estimate. First, the global lake area reported
by Kalff (2002) is less than half the lake area of the more recent estimate of Downing et al. (2006), and
much of this discrepancy may be due to representation of small lakes and ponds. Other studies have also
6-4
-------
Chapter 6. Lakes
noted that very small lakes cover substantial areas and must not be overlooked in regional and global lake
assessments (Hanson et al., 2007; Walter et al., 2007); small lakes are more likely to have more
vegetation, and thus have larger fluxes per unit lake area. Second, the data used to develop an empirical
relationship between lake area and ClrU flux in the analysis of Bastviken et al. (2004) came from
temperate and boreal lakes, with no data from arctic or tropical lakes. Bastviken et al. (2004) also note
that, since flux for most lakes is dominated by bubbles and since it is difficult to design a measurement
program that will capture all of these events, some very large events may be missed, introducing a
potential low bias in measured fluxes. Based on the regression equations in Bastiviken et al. (2004), mean
lake CH4 fluxes for bubbles plus diffusion ranged from about 5 mg CH4/m2/day for lakes larger than
1,000 km2 to 40 mg CH4/m2/day for lakes smaller than 1 hectare. Episodic storage fluxes (spring and/or
fall lake overturning) were estimated at one-third of the total annual flux for small lakes and only a few
percent for large lakes, while bubble fluxes were 50 to 60 percent of the total flux across the range of lake
sizes.
Data from arctic sites, boreal and sub-boreal beaver ponds, and tropical lakes (Table 6-1) show that flux
rates in these lakes are generally higher than the empirical fits to boreal and temperate lake data in
Bastviken et al. (2004). Bartlett et al. (1992) measured fluxes from small arctic lakes and ponds, and
observed lower fluxes for large lakes and higher fluxes for small lakes, and that their range (Table 6-1)
was comparable to other measurements that had been made in Alaska. Walter et al. (2006) reported that
CH4 fluxes from thermokarst (thaw) lakes were highly variable, and predominantly from bubbles, with
small hotspots accounting for most of the flux. Walter et al. (2007) extrapolated this and other
measurements to an annual flux of 24.2 ± 10.5 Tg CHVyr from all lakes north of 45°N.
CH4 fluxes from beaver ponds in boreal and sub-boreal North America (Table 6-1) were generally higher
than the fluxes reported in Bastviken et al. (2004).
In the tropics, most measurements of lake CH4 fluxes have been made in the neotropics (South and
Central America), with very little sampling in tropical Africa or Asia. Again, flux rates (Table 6-1) are
generally higher than boreal and temperate zone rates used in the analysis of Bastviken et al. (2004).
Bubble fluxes generally accounted for a majority of the total flux (e.g., Keller and Stallard, 1994). Fluxes
were typically higher in shallower water (Engle and Melack, 2000; Keller and Stallard, 1994), and if
vegetation was present (e.g., Melack et al., 2004) were higher in shallower water, and increased with (or
were triggered by) wind speed.
Based on higher flux ranges from thermokarst lakes, other small arctic lakes, beaver ponds, and tropical
lakes than the range of values used by Bastviken et al. (2004), the likely underestimation of small lake
area in the analysis of Bastviken et al. (2004), and the general pattern across many studies of higher fluxes
from smaller lakes, the actual CH4 emissions from lakes are likely to be at the higher end of the range
presented by Bastviken et al. (2004), perhaps 30 ± 20 Tg CHVyr.
Table 6-1. Observed Methane Fluxes from Tropical and Arctic Lakes and Boreal
and Sub-Boreal Beaver Ponds
Location
Alaska
Siberia
Ontario
Manitoba
Minnesota
Panama
Panama
CH4 Flux
(mg CH/m/day)
4-77
70 ±6
290
100
50-70
10-200
300-2,000
Reference
Bartlett etal, 1992
Walter etal, 2006
Bubieretal., 1993
Roulet etal., 1997
Naimanetal., 1991
Keller and Stallard, 1994
Keller and Stallard, 1994
Site Type
Arctic lakes
Arctic thermokarst lakesb
Boreal beaver pond
Boreal beaver pond
Sub-boreal beaver pond
Tropical lake, deeper sections (> 7 m)
Tropical lake, shallower sections (< 2 m)
6-5
-------
Chapter 6. Lakes
Location
Brazil
Brazil
Brazil
Brazil
Brazil
Brazil
CH4 Flux
(mg CHj/m/day)
53-330
2-25
50 ±8
120 ±40
320 ± 70
140 ±310
Reference
Engle and Melack, 2000
Engle and Melack, 2000
Melack etal, 2004
Melack etal., 2004
Melack etal. ,2004
Marani and Alvala, 2007
Site Type
Tropical lake, low and falling water
Tropical lake, high and rising water0
Tropical lake, open water
Tropical lake, with aquatic vegetation, shallow water
Tropical lake, aquatic vegetation, deeper water
Tropical lake and floodplain ranged
a Higher fluxes for small lakes, lower fluxes for large lakes.
b Mean value for thermokarst lakes, with most flux from hotspots occupying a small fraction of the lake surface.
c Bubble flux not measured, assumed to be 20 percent of total.
d Lake fluxes somewhat lower than floodplain fluxes.
6.3.2 Nitrous Oxide
There has been no quantification of freshwater lakes as a global source of N2O to the atmosphere. In their
analysis, Mengis et al. (1997) measured fluxes of 0.01 to 0.84 micromoles/m2/hour, which, integrated
over a 250-day open-water year, is 0.02 to 1.6 kg N2O-N/ha/yr. Huttenen et al. (2003) report N2O
emissions from boreal lakes of 0.003 to 0.015 kg N2O-N/ha during spring ice-out, 0.006 to 0.025 kg N/ha
during the open water season, and negligible during winter. Applying a mean annual flux of 0.01 to 0.1
kg N2O-N/ha across 4.2 million km of lakes and ponds (Downing et al., 2006) would generate an annual
flux of 0.004 to 0.04 Tg N2O-N/yr. This represents much less than 1 percent of the global annual N2O
flux from natural sources.
6.4 Future Emission Scenarios
6.4.1 Methane
Climate warming impacts on permafrost and the development of thermokarst (thaw) lakes could
significantly affect future Clr^ emissions from high-latitude lakes. Walter et al. (2007) used a space-for-
time substitution based on the current and projected lake distributions in permafrost-dominated and
permafrost-free terrains north of 45 °N, and representative flux rates for these lakes, to generate two
estimates of future QrU emissions from lakes north of 45°N. One estimate is based on the disappearance
of all permafrost in the Northern Hemisphere, with a consequent decrease in lake area of about 60 percent
in permafrost regions. The other estimate is based on what they consider to be a more probable transition
to a reduced extent of permafrost, with a 10 percent increase in lake area in continuous permafrost regions
due to warming and partial melting and a 60 percent decrease in lake area in non-continuous permafrost
regions due to melting. Based on mean lake emission rates from their earlier work, Walter et al. (2007)
estimated that northern lake emissions (lakes north of 45 °N) will eventually decrease by about 12 percent
to 22 Tg CHVyr in a probable transitional permafrost scenario, and by approximately 53 percent to 12 Tg
CHVyr in a "permafrost-free" Northern Hemisphere. Before this long-term decline in QrL; emissions from
lakes, due to lake area loss and permafrost thaw, there would be a period of increased QrL; emissions
associated with thermokarst lake development in the zone of continuous permafrost. Walter et al. (2007)
estimate that QrL; emission rates from northern lakes could rise as high as 50 to 100 Tg CHVyr during this
transitional period lasting hundreds of years, due to thermokarst lake development in current regions of
continuous permafrost with abundant soil organic matter. At this time, however, there are few published
studies of thermokarst lake Clr^ emissions, and much uncertainty in future projections related to rates of
permafrost decay and thermokarsting and the potential role of oxidation to reduce fluxes from warmer
lakes.
6-6
-------
Chapter 6. Lakes
6.4.2 Nitrous Oxide
There have been no published estimates of future N2O emissions from lakes. Increased nitrogen loading
(e.g., Galloway and Cowling, 2002) and increased temperatures may cause an increase in N2O fluxes
from lakes, but total N2O flux from lakes is likely to remain a very small fraction of total global N2O
emissions from natural sources.
6.5 Areas for Further Research
Lake fluxes of CFL and N2O are still not well-quantified across the globe. For example, in a recent CFL
budget assessment for a large catchment in Sweden, Christensen et al. (2007) excluded lakes and rivers
(about 15 percent of the total area in their study catchment) due to a lack of data on CFL fluxes. All large-
scale extrapolations are based on empirical relationships developed from relatively limited field sampling.
(This is especially true for N2O fluxes, of which only a few have been reported, probably because the low
flux rates reported to date have discouraged additional field studies.) For CFL., there are now dozens of
studies reporting lake fluxes, but since bubbling is a significant and highly episodic source, there is still
substantial range and uncertainty in the seasonal to annual flux rate. In addition, field sampling of lake
CH4 fluxes has not occurred in many parts of the arctic, boreal region, and tropics. Isotopic analysis of
methane fluxes may identify deep geological sources (see Chapter 8) that are bubbling through lakes.
There has been no process-based modeling of regional or global lake fluxes of CFL or N2O.
6.6 References
Bartlett, K.B., P.M. Crill, R.L. Sass, R.C. Harriss, and N.B. Dise. 1992. Methane emissions from tundra
environments in the Yukon-Kuskokwim Delta, Alaska. Journal of Geophysical Research 97: 16645-
16660.
Bastviken, D., J. Cole, M. Pace, and L. Tranvik. 2004. Methane emissions from lakes: Dependence of
lake characteristics, two regional assessments, and a global estimate. Global Biogeochem. Cycles 18:
GB4009, doi:10.1029/2004GB002238.
Bergstrom, I., S. Makela, P. Kankaala, and P. Kortelainen. 2007. Methane efflux from littoral vegetation
stands of southern boreal lakes: An upscaled regional estimate. Atmospheric Environment 41: 339-351.
Bubier, J.L., T.R. Moore, and N.T. Roulet. 1993. Methane emissions from wetlands in the midboreal
region of northern Ontario, Canada. Ecology 74: 2240-2254.
Christensen, T.R., T. Johansson, M. Olsrud, L. Strom, A. Lindroth, M. Mastepanov, N. Malmer, T.
Friborg, P. Crill, and T.V. Callaghan. 2007. A catchment-scale carbon and greenhouse gas budget of a
subarctic landscape. Philosophical Transactions of the Royal Society A 365: 1643-1656.
Cicerone, R.J., and R.S. Oremland. 1988. Biogeochemical aspects of atmospheric methane. Global
Biogeochemical Cycles 2: 299-327.
Conrad, R. 1989. Control of methane production in terrestrial ecosystems. In: M.O. Andreae and D.S.
Schimel (eds.). Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere. New
York, NY: John Wiley & Sons. pp. 39-58.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor, and H.L. Miller (eds.) Climate
Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY:
Cambridge University Press.
6-7
-------
Chapter 6. Lakes
Downing, J.A., Y.T. Prairie, J.J. Cole, C.M. Duarte, L.J. Tranvik, R.G. Striegl, W.H. McDowell, P.
Korelainen, N.F. Caraco, J.M. Melack, and J.J. Middelburg. 2006. The global abundance and size
distribution of lakes, ponds, and impoundments. Limnology and Oceanography 51: 2388-2397.
Ehhalt, D.H. 1974. The atmospheric cycle of methane. Tellus 26: 58-70.
Engle, D., and J.M. Melack. 2000. Methane emissions from an Amazon floodplain lake: enhanced release
during episodic mixing and during falling water. Biogeochemistry 51: 71-90.
Firestone, M., E. Davidson. 1989. Microbiological basis of NO and N2O production and consumption in
soil. In: M.O. Andreae and D.S. Schimel (eds.). Exchange of Trace Gases Between Terrestrial
Ecosystems and the Atmosphere. New York, NY: John Wiley & Sons. pp. 7-21.
Galchenko, V.F., A. Lein, and M. Ivanov. 1989. Biological sinks of methane. In: M.O. Andreae and D.S.
Schimel (eds.). Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere. New
York, NY: John Wiley & Sons. pp. 59-71.
Galloway J.M., and E.B. Cowling. 2002. Reactive nitrogen and the world: 200 years of change. Ambio
32: 64-71.
Hamilton, J.D., C.A. Kelly, J.W.M. Rudd, RH. Hesslein, and N.T. Roulet. 1994. Flux to the atmosphere
of CFL, and CO2 from wetland ponds on the Hudson Bay Lowlands (HBLs). Journal of Geophysical
Research 99: 1495-1510.
Hanson, P.C., S.R. Carpenter, J.A. Cardille, M.T. Coe, and L.A. Winslow. 2007. Small lakes dominate a
random sample of regional lake characteristics. Freshwater Biology. 52: 814-822.
Huttenen, J.T., S. Juutinen, J. Aim, T. Larmola, T. Hammar, J. Silvola, and P.J. Mrtikainen. 2003. Nitrous
oxide flux to the atmosphere from the littoral zone of a boreal lake. Journal of Geophysical Research
108(D14): 4421, doi:10.1029/2002JD002989.
Juutinen, S., J. Aim, T. Larmola, J.T. Huttunen, M. Morero, P.J. Martikainen, and J. Silvola. 2003. Major
implication of the littoral zone for methane release from boreal lakes. Global Biogeochemical Cycles
17(4): 1117, doi:10.1019/2003GB002105.
Kalff, J. 2002. Limnology. Old Tappan, NJ: Prentice Hall.
Kankaala, P., S. Taipale, H. Nykanen, and R.I. Jones. 2007. Oxidation, efflux, and isotopic fractionation
of methane during autumnal turnover in a polyhumic boreal lake. Journal of Geophysical Research
112: G02003, doi:10.1029/2006JG000336.
Kankaala, P., K. Kaki, S. Makela, A. Ojala, H. Pajunen, and L. Arvola. 2005. Methane efflux in relation
to plant biomass and sediment characteristics in stands of three common emergent macrophytes in
boreal mesoeutrophic lakes. Global Change Biology 11: 145-153.
Keller, M., and R.F. Stallard. 1994. Methane emission by bubbling from Gatun Lake, Panama. Journal of
Geophysical Research 99(8): 307-8319.
Khalil, M.A.K., and M.J. Shearer. 2000. Sources of methane: An overview. In: M.A.K. Khalil (ed.).
Atmospheric Methane. Berlin: Springer-Verlag. pp. 88-111.
Kiene, R.P. 1991. Production and consumption of methane in aquatic systems, pp. 111-146 in J. Rogers
and W. Whitman (eds.)Microbial Production and Consumption of Greenhouse Gases: Methane,
Nitrogen Oxides, and Halomethanes. American Society of Microbiology, Washington DC, 1991.
Liikanen, A., L. Flojt, and P. Martikainen. 2002. Gas dynamics in eutrophic lake sediments affected by
oxygen, nitrate, and sulfate. Journal of Environmental Quality 31: 338-349.
Marani, L., and P.C. Avala. 2007. Methane emissions from lakes and floodplains in Pantanal, Brazil.
Atmospheric Environment. 41: 1627-1633.
6-8
-------
Chapter 6. Lakes
Matthews, E. 2000. Wetlands. In: M.A.K. Khalil (ed.) Atmospheric Methane. Berlin: Springer-Verlag. pp.
202-233.
Melack, J.M., L.L. Hess, M. Gastil, B.R. Forsberg, S.K. Hamilton, I.B.T. Lima, and E.M.L.M. Novo.
2004. Regionalization of methane emissions in the Amazon Basin with microwave remote sensing.
Global Change Biology 10: 530-544.
Mengis, M., R. Gacher, and B. Wehrli. 1997. Sources and sinks of nitrous oxide (N2O) in deep lakes.
Biogeochemistry 38: 281-301.
Michmerhuizen, C.M., R.G. Striegl, and M.E. McDonald. 1996. Potential methane emission from north-
temperate lakes following ice melt. Limnology and Oceanography 41: 985-991.
Naiman, R.J., T. Manning, C.A. Johnston. 1991. Beaver population fluctuations and tropospheric
methane emissions in boreal wetlands. Biogeochemistry 12: 1-15.
Pina-Ochoa, E., and M. Alvarez-Cobelas. 2006. Denitrification in aquatic environments: A cross-system
analysis. Biogeochemistry 81: 111-130.
Prather, M., D. Ehhalt, F. Dentener, R. Derwent, E. Dlugokencky, E. Holland, I. Isaksen, J. Katima, V.
Kirchhoff, P. Matson, P. Midgley, and M. Wang. 2001. Atmospheric chemistry and greenhouse gases.
In: J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A.
Johnson (eds.). Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the
Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK:
Cambridge University Press, pp. 239-287.
Reeburgh, W.S. 2004. Global methane biogeochemistry. In: R.F. Keeling (ed.). Treatise on
Geochemistry, Volume 4: the Atmosphere. Amsterdam: Elsevier. pp. 65-89.
Repo, M.E., J.T. Huttunen, A.V. Naumov, A.V. Chichulin, E.D. Lapshina, W. Bleuten, P.J. Martikainen.
2007. Release of CO2 and CIL, from small wetlands lakes in western Siberia. Tellus 59B: 788-796.
Roulet, N. T., P.M. Crill, N.T. Comer, A. Dove, R.A. Boubonniere, R. A. 1997. CO2 and CH4 flux
between a boreal beaver pond and the atmosphere. Journal of Geophysical Research 102: 29313-
29319.
Rudd, J.W.M., and R.D. Hamilton. 1978. Methane cycling in a eutrophic shield lake and its effects on
whole lake metabolism. Limnology and Oceanography 23: 337-348.
Schmid, M., M. de Batist, N.G. Granin, V.A. Kapitanov, D.F. McGinnis, I.E. Mizandrontsev, A.I.
Obzhirov, and A. Wuestl. 2007. Sources and sinks of methane in Lake Baikal: A synthesis of
measurements and modelling. Limnology and Oceanography 52: 1824-1837.
Sebacher, D.I., R.C. Harriss, and K.B. Bartlett, 1983, Methane flux across the air-water interface: Air
velocity effects. Tellus 35B: 103-109.
Seitzinger, S., J.A. Harrison, J.K. Bohlke, A.F. Bouwman, R. Lowrance, B. Peterson, C. Tobias, and G.
Van Drecht. 2006. Denitrification across landscapes and waterscapes: A synthesis. Ecological
Applications 16: 2064-2090.
Smith, L.C., Y. Zheng, G.M. MacDonald, and L.D. Hinzman. 2005. Disappearing arctic lakes. Science
308: 1429.
Stadmark, J., and L. Leonardson. 2005. Emissions of greenhouse gases from ponds constructed for
nitrogen removal. Environmental Engineering 25: 542-551.
St. Louis, V.L., C.A. Kelly, E. Duchemin, J.W.M. Rudd, and D.M. Rosenberg. 2000. Reservoir surfaces
as sources of greenhouse gases to the atmosphere: a global estimate. BioScience 50: 766-775.
6-9
-------
Chapter 6. Lakes
Striegl, R.G., and C.M. Michmerhuizen. 1998. Hydrological influence on methane and carbon dioxide
dynamics at two north-central Minnesota lakes. Limnology and Oceanography 43: 1519-1529
Walter, K.M., L.C. Smith, and F.S. Chapin III. 2007. Methane bubbling from northern lakes: present and
future contributions to the global methane budget. Philosophical Transactions of the Royal Society A
365: 1657-1676.
Walter, K.M., S.A. Zimov, J.P. Chanton, D. Verbyla, and F.S. Chapin III. 2006. Methane bubbling from
Siberian thaw lakes as a positive feedback to climate warming. Nature 443: 71-75.
Watson, R.T., H. Rodhe, H. Oeschger, and U. Siegenthaler. 1990. Greenhouse gases and aerosols. In: J.T.
Houghton, G.J. Jenkins, and J.J. Ephraums (eds.). Climate Change: The IPCC Scientific Assessment.
Cambridge UK: Cambridge University Press, pp 1-40.
Whalen, S.C. 2005. Biogeochemistry of methane exchange between natural wetlands and the atmosphere.
Environmental Engineering Science 22: 73-94
Zimov, S.A., S.P. Davydov, G.M. Zimova, A.I. Davydova. E.A.G. Schuur, K. Durra, and F.S. Chapin III.
2006. Permafrost carbon: Stock and decomposability of a globally significant carbon pool. Geophysical
Research Letters 33: L20502, doi.10.1029/2006GL027484.
Zimov, S.A., Y.V. Voropaev, I.P. Semiletov, S.P. Davydov, S.F. Prosiannikov, F.S. Chapin III, M.C.
Chapin, S. Trumbore, and S. Tyler. 1997. North Siberian lakes: A methane source fueled by
Pleistocene carbon. Science 277: 800-802.
6-10
-------
Chapter 7. Gas Hydrates
Currently, natural gas hydrates (also called clathrate-hydrates) play two very distinct roles in the global
carbon cycle. They act as a dynamic storage capacitor for CF^ (i.e., a large sink of CFU, which fluctuates
in size as atmospheric concentrations of CH4 change) (Dickens, 2003), but they also feed CFU directly
into the ocean water column. Under current conditions, CFU emissions from gas hydrates are small, but
there is the potential for significant CFU release from gas hydrates.
U.S. EPA (1993) examined and summarized CFU emissions from gas hydrates as a natural source of CH4,
as did the AR4 (Denman et al., 2007). The AR4 discusses potential CFU emissions from hydrates but does
not examine them quantitatively. The estimates for future emissions from hydrates reported in U.S. EPA
(1993) have been updated significantly and more advanced models have been developed. The emission
source discussed in this chapter includes CFU currently stored in the form of gas hydrate; CFU emitted
from gas hydrates is introduced into the ocean water column, where it is dissolved, oxidized, or emitted to
the atmosphere. This source differs from the oceanic CFU source discussed in Chapter 4: what this report
describes as oceanic CFU is generated in the water column, but CH4 from gas hydrates is emitted into the
water column.
-T A r\ ... r_ . . _ Figure 7-1. Small cage of Structure I
7.1 Description of Emission Source gas hydrate cage.
Gas hydrates are an ice-like compound formed between water
and a gas molecule such as CFU, under high pressure and at
temperatures near the freezing point of water. Pressure and
temperature conditions conducive to CH4 gas hydrate
formation are found at ocean depths typically greater than 500
meters and are characterized by the hydrate stability zone
(HSZ) shown in Figure 7-2. With increasing temperature, a
higher pressure is necessary to stabilize gas hydrates in the
solid form. The hydrate stability curve shown in blue in Figure
7-2 is determined from the pressure-temperature data for
methane hydrate equilibrium, while the geothermal
temperature gradient is an example of a typical temperature-
depth curve for an ocean depth of approximately 1.7 km. The
water temperature decreases with depth throughout the water
column and increases below the seafloor due to the heat flux
through the earth's crust. Due to the low solubility of methane
in water and thus the low concentration of methane in the
ocean, hydrate deposits are mostly contained in ocean
sediments.
Three major types of crystal structures are formed by gas hydrates: structure I (si), structure II (sll), and
structure H (sH). Structure I is the preferred form for pure CH4, ethane, and carbon dioxide hydrates.
Hydrocarbons including propane and iso-butane and small gases such as argon, krypton, nitrogen, and
oxygen form structure II hydrate, as well as certain gas mixtures of CFU and ethane.
Gas hydrates have an incredible storage capacity for the gases that they trap. For instance, one cubic
meter (1 m3) of pure methane hydrate, which contains approximately one CFU molecule for every six
water molecules, can store 170 to 180 standard cubic meters (SCM) of gas (Kvenvolden, 1991; Sloan,
2003). As a result of this storage capacity and their widespread occurrence, gas hydrates have recently
gained interest as a potential energy resource. This section, however, examines the worldwide inventory
and gas hydrates' potential as a CFU emission source.
Water molecules are illustrated as
red (oxygen) and gray (hydrogen)
atoms, with hydrogen bonds
indicated with yellow dotted lines.
7-1
-------
Chapter 7. Gas Hydrates
!/)
1/1
10
15
20
25
30
35
40
Geothermal
Temperature
Gradient
0.0
I 0.5
- 1.0
- 1.5
- 2.0
- 2.5
- 3.0
I 3.5
o.
o»
Q
4.0
10
15
20
25
30
Temperature (°C)
Figure 7-2. Illustration of the water temperature and hydrate stability zone through the water column, with an ocean
depth of 1.7 km and a geothermal gradient (i.e., temperature increase with depth below the ocean floor) of 15
degrees Celsius per km.
Methane gas hydrates form if CH4 is generated within or enters the HSZ (see Figure 7-2). The source of
CFLj generation can be either microbial or thermal in origin. Microbial QrU is created by the anaerobic
digestion of organic matter by microorganisms in shallow, oceanic, and continental sediments.
Thermogenic QrU may contain a mixture of other higher hydrocarbons such as ethane and propane, but
microbially produced CFLj is generally pure ClrU. The introduction of the higher hydrocarbon gases will
lower the pressures necessary for hydrate formation, thus enlarging the HSZ.
Thermogenesis of QrU takes place under high pressure (>20 megapascal [MPa]) and temperature (>80 to
120°C) (MacDonald, 1990). Typically, such conditions exist well below the zone of hydrate stability.
When CH4 is created in this manner, it enters hydrate form only when a migration pathway exists from
the lower sediment to the upper sediment layer (Kvenvolden, 1988). Tests of the isotopic composition of
the CH4 extracted from hydrates at several depths and locations worldwide reveal that most of it is
microbially generated, not thermally produced (MacDonald, 1990).
Although gas hydrates currently have low ClrU emission rates, the potential for ClrU release is great. It is
therefore important to detail the amount of CFLj currently stored as gas hydrate globally. Since the
7-2
-------
Chapter 7. Gas Hydrates
previous EPA report on natural sources of QrU (U.S. EPA, 1993), the global methane hydrate inventory
estimates have changed substantially. As shown in Table 7-1, the estimates for the amount of QrU stored
worldwide have varied significantly. Early estimates of the amount of QrU in oceanic hydrates ranged
from 2.3xl06 Tg CH4 (Mclver, 1981) to 5.5xl09 Tg CH4(Dobrynin et al., 1981), and estimates of the
amount of QrU in continental hydrates ranged from 1x10 Tg QrU (Meyer, 198 1) to 2.4x 10 Tg
(Dobrynin et al., 198 1). Later, the inventories were thought to have converged to about 1 x 107 Tg
in oceanic reserves and 5x 105 Tg for continental reserves (MacDonald, 1990; Kvenvolden, 1991).
However, in recent years the estimates have, for the most part, decreased. Since the amount of continental
hydrate is now believed to be only a small fraction of the total hydrate volume, most current estimates
focus on the marine hydrate inventory. In the mid-1990s, estimates for oceanic CH4 in hydrates ranged
from 1 . 8 x 1 07 Tg (Gornitz and Fung, 1 994) to 3 .2 x 1 07 Tg (Harvey and Huang, 1 995) while in the current
decade estimates include 1 x 107 Tg (Kvenvolden and Lorenson, 2001), 6.7x 106 Tg in both hydrates and
CH4 bubbles (Buffett and Archer, 2004), 0.7 to 3.4x 106 Tg (Milkov, 2004), and 7.4x 107 Tg (Klauda and
Sandier, 2005). Harvey and Huang estimated that the ClrU in continental hydrates totals 1 x 106 Tg,
approximately 3 percent of the total ClrU found in hydrate deposits (Harvey and Huang, 1995). Milkov
(2000) estimated that there are 7-700 Tg CH4 in hydrates associated with mud volcanoes.
Table 7-1. Summary of Methane Inventories in Marine Gas Hydrates
Study
(Mclver, 1981)
(Dobrynin etal., 1981)
(MacDonald, 1990)
(Gornitz and Fung, 1994)
(Harvey and Huang, 1995)
(Kvenvolden and Lorenson, 2001)
(Buffett and Archer, 2004)
(Milkov, 2004)
(Klauda and Sandier, 2005)
CH4 Mass/100
Tg
2.3
5,500
10
18
32
10
6.7
2a
74
a Midpoint value.
7.2 Factors That Influence Emissions
Gas hydrates, when found within the HSZ, emit ClrU only though the dissolution of QrU into
undersaturated water (i.e., water that contains less than the maximum amount of QrU at any given
temperature). Dissolution rates for QrU and carbon dioxide hydrates (i.e., the rate at which they dissolve
in ocean water) have been measured experimentally on the seafloor (Rehder et al., 2004). It was
concluded that the rate of dissolution of both QrU and €62 from hydrates into the seawater was limited by
the diffusion of the gas away from the hydrate as opposed to any characteristics of the hydrate itself.
Dissolution rates of several millimeters per day were found for a sample exposed to the open ocean
currents (Rehder et al., 2004). Dissolution rates are influenced by the ocean currents, the exposed hydrate
area, and the level of QrU saturation in the water at the exposed hydrate interface.
The minimum depth of hydrate stability is proportional to pressure and inversely proportional to
temperature (i.e., hydrates are stable if the pressure is sufficiently high and the temperature is sufficiently
low). Hydrates do not exist at the ocean's surface because, even under the coldest conditions, the pressure
is not sufficient. Conversely, hydrates do not exist at great depths because temperature increases with
depth in sediment (according to a geothermal gradient of 0.016 to 0.053°C/meter), and by about 2
kilometers below the sediment surface, temperature is almost always too high for the hydrate structure to
be viable. The HSZ can be anywhere from 0 to 2 kilometers thick, and can expand or contract in response
to changing temperature and pressure conditions. If it expands (as a result of increased pressure and/or
7-3
-------
Chapter 7. Gas Hydrates
decreased temperature), it can incorporate more CFU from the surrounding sediment. If it contracts (as a
result of decreased pressure and or increased temperature), large amounts of CFU can be liberated from
hydrates into the sediment and can migrate into the atmosphere. The global average HSZ is expected to
respond to atmospheric temperature changes over time scales on the order of several thousand years.
However, initial decreases could occur in as little as 200 years in shallow high-latitude seafloor areas that
underlie regions of sea ice loss (Fyke and Weaver, 2006). These areas that would experience sea ice loss
would be subject to lowered albedo and increased absorption of incoming radiation at the sea surface. The
combination of initially shallow HSZ depths and rapid, large seafloor temperature increases in these
regions makes them the first to experience rapid hydrate destabilization (Fyke and Weaver, 2006).
Currently, the major source of CFU flux due to methane hydrates is offshore continental hydrates
(Kvenvolden, 1991). Offshore continental hydrates are found on the nearshore continental shelf, where
melting subsea permafrost has continued since times of lower sea level. Since the last ice age, 18,000
years ago, sea level has risen about 100 to 125 meters, and the temperature of the present shelf has risen
about 15°C (Hill et al, 1985; Kvenvolden et al., 1991). Additional increases in shelf temperatures would
result in additional areas of hydrates becoming destabilized.
7.3 Current Global Emissions
Since 1993, there has been limited discussion of the current flux of CFU from gas hydrate reservoirs.
Oceanic and onshore continental reserves are believed to be stable at present, which means that they are
not currently emitting CFU. However, offshore continental shelf reserves are currently unstable, and may
emit 2 to 5 Tg of CFU annually to the seafloor (Kvenvolden, 1988, 1991; Dickens, 2003), of which a large
fraction would likely be oxidized in the ocean water column(MacDonald et al., 2002; Niemann et al.,
2006). The Kvenvolden estimates are based on emissions resulting from climate changes within the last
18,000 years, the time since the last glaciation. Since that time, sea level has risen about 100 to 125
meters, inundating large areas of permafrost which contain methane hydrates. Inundation has increased
the pressure in this region by about 9 atmospheres or 132 psi, which would be expected to increase the
stability of the hydrates. However, over the same period, the temperature at the sediment surface has
increased by 15°C, which is more than enough to offset the increase in pressure and destabilize the
hydrates (Kvenvolden, 1991). Due to the slow rates of downward thermal diffusion in sediments, this
temperature change is still in the process of penetrating downward, melting the permafrost and associated
gas hydrates. The emission estimate of 3 to 5 Tg CFU/yr was determined by assuming that the difference
between the calculated and the known amount of Cfk in sub-sea permafrost hydrates (160,000 - 43,000 =
1 17,000 Tg) has been uniformly released over the last 18,000 years (Kvenvolden, 1991). These estimates
are very sensitive to the estimates of the total amount of CFU stored in sub-sea permafrost hydrates. As
discussed in Section 7.1, these estimates can vary significantly.
In 2003, Dickens published models of two plausible gas hydrate capacitors that consisted of 13*106 Tg of
in hydrate, 0.67x 106 Tg office CFU beneath the hydrate reservoirs, and 1.3* 106 Tg of dissolved
. The two scenarios were a high and low flux of CFU into the gas hydrate capacitor from
methanogenesis, FueOi = 2.2 and 9.3 Tg/yr. The resulting fluxes of CFU to the seafloor are 2.2 and 9.3
Tg/yr from the low and high flux cases, respectively (Dickens, 2003). These likely represent the most
accurate estimate of CFU flux from hydrate deposits globally and are of the same magnitude of the
estimates reported in the previous report on natural sources (U.S. EPA, 1993), as well as the AR4
(Denman et al., 2007); however, Dickens' estimates correspond to emissions to the seafloor, while the
EPA report and the AR4 refer to atmospheric emissions.
These estimates assume that the CH4 being liberated from the gas hydrate form is released into the
atmosphere. It is possible, however, that most or all of this gas is not actually emitted to the atmosphere.
Instead it is oxidized or absorbed within the sediment or dissolved into the water column. Recent work by
Yamamoto et al. (2009) showed while the saturation of methane in the water column does not have to
7-4
-------
Chapter 7. Gas Hydrates
reach 100 percent of its capacity in order for methane to be released to the atmosphere, the current rate of
methane release from ocean sediments is insufficient to result in methane being released to the
atmosphere.
7.4 Future Emission Scenarios
Due to their proximity to the earth's surface (< 2,000 meters), gas hydrates will likely eventually be
affected by climate change. ClrU emissions from this source are likely to increase if temperature
significantly rises. While pressure on hydrates is also expected to change as a result of sea level rise and
the melting of polar ice caps, temperature changes are likely to be far more significant than changes in
pressure in determining changes in future emissions due to anthropogenic climate change and on the HSZ.
The following two sections summarize existing estimates of potential future CH4 emissions from
hydrates.
7.4.1 Continental Hydrates
Kvenvolden (1991) estimates that the most likely additional future source of QrU emissions from hydrates
is from subsea permafrost. This is a region of QrU hydrates that has been submerged since the last glacial
maximum, and is currently unstable. The accelerated oceanic temperature rise that would result from an
expected atmospheric temperature increase due to climate change over the coming century could
penetrate to the subsea permafrost and increase the current rate of emissions from these hydrates by an
order of magnitude (from 4-5 Tg CHVyr to 40-50 Tg CHVyr). This increase would be expected to take
place sometime after the 21st century (Kvenvolden, 1991). Similarly, this scenario predicts that onshore
gas hydrates will eventually be destabilized by climate change and the permeation of rising temperatures
through the permafrost. The rate of emissions from destabilization of onshore hydrates is predicted to be
twice that of subsea permafrost hydrates, or about 100 Tg CWyr. However, the time lag before this
emission rate is achieved is likely to be greater than the lag for subsea permafrost emissions, with
estimates of the time lag for continental hydrates ranging from hundreds to thousands of years.
MacDonald (1990) assumes, based on the analysis of temperature changes in the Arctic by Lachenbruch
and Marshall (1986), that Arctic surface temperature will increase by 2°C by 2080 (an extremely
conservative estimate). MacDonald predicts that about 0.01 percent of the continental hydrate zone will
degas per year, resulting in annual emissions of 50 Tg CHVyr.
Based on the assumption that air temperatures in the high latitudes will increase by 10°C as a step change,
Bell (1982) predicts that the destabilization of hydrates below the Arctic permafrost region will begin
within a few hundred years of the initial temperature rise. Consequently, the permafrost between the -5°C
and -15°C isotherms of annual mean air temperature becomes unstable. Bell estimates that half of the
continental reserves (2.7* 106 Tg QrU) of hydrates will dissociate uniformly over 4,000 years resulting in
a flux of about 300 Tg CHVyr.
Based on a 19°C air temperature rise within a century in the Arctic Islands resulting from a quadrupling
of atmospheric CCh concentration, Nisbet (1989) predicts that a large fraction of the hydrates in the
Arctic Islands will destabilize. Nisbet assumes that hydrates in this region are stable below 50 to 100
meters of sediment due to the extreme cold temperatures. Because of the large temperature rise and the
proximity of hydrates relative to the surface, Nisbet predicts that this region of hydrates will dissociate
within approximately the 100 years following the temperature increase, emitting 100 Tg/yr of ClrU from
hydrates. Deeper hydrates would begin degassing in 500 to 1,000 years (Nisbet, 2002). The shallow
nature of hydrate deposits in the Arctic Islands and the potential for large temperature increases make
these hydrates very vulnerable to warming of the surface (Nisbet, 2002).
Harvey and Huang (1995) built a one-dimensional model with vertical columns on a 1 degree by 1 degree
global grid to predict hydrate distribution and potential impact of thermal perturbations on hydrate
7-5
-------
Chapter 7. Gas Hydrates
destabilization. Their base-case clathrate distribution estimates about 24,000 Gt C (3.2* 107
stored as methane clathrate in marine sediments and about 800 Gt C in terrestrial sediments. They
predicted that only a small fraction of current stores could be destabilized by future climate change. For
the continental hydrates, they examined three global surface warming scenarios (5, 10, and 15°C) and the
cumulative release of CFU at times ranging from 100 to 5,000 years after the warming. They predict that,
after 100 years, no CFU will have been released from terrestrial hydrates; after 500 years, 1.1, 7.3, and
14.4 Gt CH4 will have been released (Harvey and Huang, 1995) for 5, 10, and 15°C warming. If the
release is assumed to be uniform over years 100 to 500, the rates of CFU emissions are 2.7, 18.3, and 36.0
Tg CWyr. Maximum CH4 rates of 14.0, 74.5, and 132.1 Tg CWyr are reached between 1,000 and 2,000
years.
A current Department of Energy project (DE-NT0005665) led by the University of Alaska Fairbanks and
the U.S. Geological Survey is currently assessing the quantity of methane seepage from methane hydrate
deposits under thermokarst lakes on the Alaskan North Slope. Terrestrial permafrost gas hydrates may be
a significant source of methane seepage, but remain unqualified on a global scale. These gas hydrate
deposits have a much thinner water column for oxidation processes and may be more susceptible to
climate forcing than oceanic hydrates. Emission of methane from these terrestrial methane hydrate
deposits may present a more direct and potentially more effective method of transfer of methane to the
atmosphere.
7.4.2 Oceanic Hydrates
Kvenvolden (1991) predicts that oceanic hydrates will begin to destabilize after thousands of years
following a rise in global temperatures, estimating that CFU emissions from this reservoir could
eventually exceed the emissions from continental hydrates (150 Tg CFU/yr).
Bell (1982) assumes that the surface water temperature of the Norwegian Sea, which feeds the Arctic
Ocean, will rise about 3.5°C over an unspecified timeframe, resulting in a 3.5°C temperature rise along
half the length of the 300 meter depth contour in the Arctic Ocean. Given this temperature change, the
hydrates extending from the ocean sediment interface to a depth of 40 meters below the sea floor would
be destabilized where ocean depths are between 280 and 370 meters. Bell estimates that the CFU in this
40 meter zone, which represents about 1 percent of the oceanic methane hydrate reserves, would be
entirely and uniformly released over a 100-year span. A global estimate of 160 Tg CHVyr results from the
assumption that the global oceanic CFU reserve of 1.3><106 Tg is uniformly distributed in the top 250
meters of ocean sediments, at depths between 200 and 1,000 meters.
Revelle (1983) estimates hydrate emissions under a scenario in which both mean annual air temperature
and mean ocean surface temperature rise 3°C globally. The ocean bottom temperature would therefore
increase by 1 to 4°C. After approximately 100 years, the top 100 meters of all oceanic hydrates around
the world would be destabilized and begin releasing CFU into the ocean at a rate of 800 Tg CFU/yr, with
20 percent absorbed by the water. The net CFU released to the atmosphere would be 640 Tg CHVyr.
Fyke and Weaver (2006) performed a series of climate sensitivity and potential future climate change
experiments using the University of Victoria Earth System Climate Model. They found that the global
HSZ responds significantly to elevated atmospheric CO2 overtime scales of thousands of years, with
initial decreases of the HSZ occurring after 200 years in shallow high-latitude seafloor areas. The
majority of the global HSZ adjustment to warmer seafloor temperatures occurs within the first 5,000 years
after the atmospheric CO2 increase. They estimate that, for average seafloor temperature increases of 1.0,
2.0, and 3.9°C (corresponding to atmospheric CO2 levels that are allowed to increase exponentially based
on the observed 1850-1990 increase, capped at years 2000, 2050, and 2100, and then held constant) that
7, 14, and 27 percent of the global hydrate reservoir is dissociated for their median thermal diffusivity
(i.e., rate of temperature propagation) value. About half of this loss occurs in the first 5,000 years
resulting in average CFU emissions of 67, 138, and 300 Tg CFU/yr.
7-6
-------
Chapter 7. Gas Hydrates
Gornitz and Fung (1994) provide estimates of the magnitudes and spatial distribution of potential oceanic
methane hydrate reservoirs and examine implications for eventual atmospheric CFLj release due to climate
change. Two models were used: in situ bacterial production and pore fluid expulsion. The potential
sensitivity to projected climate change was explored by estimating ClrL volumes contained within the
uppermost 1 meter of the HSZ that lie within 2°C and 2 meters of the equilibrium curve. Uniform release
of this CH4 (according to the bacterial model) would occur over the 100 years following a 2°C increase in
ocean bottom temperatures at rates of 6.7 to 45.3 Tg CWyr.
In addition to the continental CFLj release estimates by Harvey and Huang (1995) discussed in Section
7.4.1, release rates from oceanic hydrates were calculated at ocean bottom temperature increases of 1, 2,
3, 4, 5, and 6°C over a timeframe of 2,000 years. The global hydrate reservoir is assumed to contain
24,000 Gt C (3.2x 107 Tg OL) (Harvey and Huang, 1995). The peak CH4 fluxes predicted range from 133
to 4,360 Tg CHVyr (for 1 to 6°C ocean bottom warming), with this peak occurring at the initial
temperature spike.
Recent work by Reagan and Moridis (2008) at the Lawrence Berkeley National Laboratory studied the
potential response of oceanic gas hydrate deposits to 1, 2, and 5°C increases in seafloor temperatures
using TOUGH+HYDRATE, a multiphase flow and transport model developed for methane hydrates.
They studied two cases, one similar to hydrates deposits below the Gulf of Mexico at 570 m depth and
one at a depth of 320 m, similar to oceanic methane hydrate deposits on the Arctic continental shelf. Their
estimates indicate that fluxes from hydrate deposits exposed to significant temperature increases may
exceed the ability of the seafloor environment to consume the released methane via oxidation pathways,
similar to observations by Leifer et al. (2006).Their model indicates that a 1 to 5°C rise in ocean
temperature could result in 120 to 200 g CHVyr/m . Assuming that the global average methane hydrate
saturation in oceanic sediments and a worldwide methane hydrate reservoir containing 107 Tg QrU, this
would correspond to a methane flux of about 1,200-2,000 Tg QrL/yr to the seafloor.
Table 7-2. Ocean Hydrate Scenarios
Ocean
temperature
rise (°C)
CH4 reserves
(Tg)
Time until
destabilization
begins (years)
Time to fully
destabilize
(years)
Avg emissions
factor3
(10~6yr~1 °C~1)
Annual
emissions'
(Tg/yr)
Kvenvolden
1x1Q7
1 ,000s
1 ,000s
5.0
>150
Revelle
1-4
1.8x107
100s
150
800
Bell
-0.5-3
1.3x107
100s
100s
3.4
160
Fyke
1-4
8.7x106
200
5,000
8.9
67-300
Gornitz
1-2
8.7x106
100
N/A
flux
3.1
6.7-45
Harvey
1-6
1.8x107
100
1,000s
11.0
93-2,133
Composite
3
1.0x107
6
186
3 Emissions factor at a hTocean of 3°C, either calculated or interpolated.
b ChU emitted to the ocean.
7.5 Areas for Further Research
Areas for future research in the area of gas hydrates as they relate to atmospheric ClrL start first and
foremost with the quantification of the reserves of QrU stored in the form of gas hydrates. Since most of
7-7
-------
Chapter 7. Gas Hydrates
the future emission scenarios depend on the total amount of methane stored as gas hydrates globally, it is
very important to have the most accurate estimates possible of the total methane stored as hydrate.
Secondly, accurate estimates of the rates of QrU absorption into the oceans and the rates of CFU oxidation
within the water column are absolutely essential. Estimates of future emissions from methane hydrates
presented in Table 7-2 only consist of how much CH4 would reach the oceans from hydrate deposits. In
order to close the loop on CFU emissions to the atmosphere, absorption and oxidation rates must be
known. It is also not clear that the current flux of CFU from hydrate deposits has been accurately assessed.
Not only are the continental shelf hydrates persisting in an unstable state (Buffett and Zatsepina, 1999),
but any hydrates exposed to seawater undersaturated in CFU are vulnerable to dissolution (Rehder et al.,
2004).
7.6 References
Bell, P. R. 1982. Methane hydrates: an estimate of their contribution to carbon dioxide-induced
atmospheric warming, Oak Ridge Associated Univ. Inc., Oak Ridge, TN, USA.: 30 pp.
Buffett, B. and D. Archer. 2004. Global inventory of methane clathrate: sensitivity to changes in the deep
ocean. Earth and Planetary Science Letters 227(3-4): 185-199.
Buffett, B. A. and O. Y. Zatsepina. 1999. Metastability of gas hydrate. Geophysical Research Letters
26(19): 2981-2984.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohman, S. Ramachandran, P.L. da Silva Bias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: IPCC, 2007.
Climate Change 2007: The Physical Basis. Contribution of Working Group I to the 4th Assessment
Report of the IPCC. Cambridge, UK and New York, NY: Cambridge Univ. Press.
Dickens, G. R. 2003. Rethinking the global carbon cycle with a large, dynamic and microbially mediated
gas hydrate capacitor. Earth and Planetary Science Letters 213(3-4): 169-183.
Dobrynin, V. M., Y. P. Korotajev and D. V. Plyuschev. 1981. Gas Hydrates: A possible energy resource.
Long-Term Energy Resources. R. F. Meyer and J. C. Olson. Boston, MA., Pitman. 1: 727-729.
Fyke, J. G. and A. J. Weaver. 2006. The effect of potential future climate change on the marine methane
hydrate stability zone. Journal Of Climate 19(22): 5903-5917.
Gornitz, V. and I. Fung. 1994. Potential Distribution Of Methane Hydrates In The Worlds Oceans. Global
Biogeochemical Cycles 8(3): 335-347.
Harvey, L. D. D. and Z. Huang. 1995. Evaluation Of The Potential Impact Of Methane Clathrate
Destabilization On Future Global Warming. Journal Of Geophysical Research-Atmospheres 100(D2):
2905-2926.
Hill, P. R., P. J. Mudie, K. Moran and S. M. Blasco. 1985. A sea-level curve for the Canadian Beaufort
Shelf. Can. J. Earth Sd^22(W): 1383-1393.
Klauda, J. B. and S. I. Sandier. 2005. Global distribution of methane hydrate in ocean sediment. Energy &
Fuels 19(2): 459-470.
Kvenvolden, K. A. 1988. Methane hydrates and global climate. Global Biogeochemical Cycles 2(3): 221-
9.
Kvenvolden, K. A. 1991. A Review of Arctic Gas Hydrates as a Source of Methane in Global Change.
International Conference on the Role of the Polar Regions in Global Change, Geophysical Institute,
University of Alaska Fairbanks.
7-8
-------
Chapter 7. Gas Hydrates
Kvenvolden, K. A. and T. D. Lorenson. 2001. The Global Occurrence of Natural Gas Hydrates. Natural
Gas Hydrates: Occurrence, Distribution, and Detection. C. K. Paull and W. P. Dillon. Washington DC,
American Geophysical Union. 124: 3-18.
Kvenvolden, K. A., T. D. Lorenson and T. C. Collett. 1991. Arctic Shelf Gas Hydrates as a Possible
Source of Methane. Abstracts with Programs, Geological Society of America 23(5): A238.
Lachenbruch, A. H. and B. V. Marshall. 1986. Changing Climate - Geothermal Evidence From
Permafrost In The Alaskan Arctic. Science 234(4777): 689-696.
Leifer, I., B. P. Luyendyk, J. Boles and J. F. Clark. 2006. Natural marine seepage blowout: Contribution
to atmospheric methane. Global Biogeochemical Cycles 20(3).
MacDonald, G. J. 1990. Role Of Methane Clathrates In Past And Future Climates. Climatic Change
16(3): 247-281.
MacDonald, I. R., I. Leifer, R. Sassen, P. Stine, R. Mitchell and N. Guinasso. 2002. Transfer of
hydrocarbons from natural seeps to the water column and atmosphere. Geofluids 2(2): 95-107.
Mclver, R. D. 1981. Gas Hydrates. Long-Term Energy Resources. R. F. Meyer and J. C. Olson. Boston,
MA., Pitman. 1:713-726.
Meyer, R. F. 1981. Speculations on Oil and Gas Resources in Small Fields and Unconventional Deposits.
Long-Term Energy Resources. R. F. Meyer and J. C. Olson. Boston, MA., Pitman. 1: 49-72.
Milkov, A. V. 2000. Worldwide distribution of submarine mud volcanoes and associated gas hydrates.
Marine Geology 167(1-2): 29-42.
Niemann, H., M. Elvert, M. Hovland, B. Orcutt, A. Judd, I. Suck, J. Gutt, S. Joye, E. Damm, K. Finster
and A. Boetius. 2005. Methane emission and consumption at a North Sea gas seep (Tommeliten area).
Biogeosciences 2(4): 335-351.
Nisbet, E. G. 1989. Some Northern Sources Of Atmospheric Methane - Production, History, And Future
Implications. Canadian Journal Of Earth Sciences 26(8): 1603-1611.
Nisbet, E. G. 2002. Have sudden large releases of methane from geological reservoirs occurred since the
Last Glacial Maximum, and could such releases occur again? Philosophical Transactions of the Royal
Society of London Series a-Mathematical Physical and Engineering Sciences 360(1793): 581-607.
Reagan, M. T. and G. J. Moridis. 2008. Modeling of Oceanic Gas Hydrate Instability and Methane
Release in Response to Climate Change. Proceeding of the 6th International Conference on Gas
Hydrates, Vancouver, British Columbia, Canada.
Rehder, G., S. H. Kirby, W. B. Durham, L. A. Stern, E. T. Peltzer, J. Pinkston and P. G. Brewer. 2004.
Dissolution rates of pure methane hydrate and carbon-dioxide hydrate in undersaturated seawater at
1000-m depth. Geochimica Et Cosmochimica Acta 68(2): 285-292.
Revelle, R. R. 1983. Methane Hydrate in Continental Slope Sediments and Increasing Atmospheric
Carbon Dioxide. Changing Climates. Washington D.C., National Academy Press: 252-261.
Sloan, E. D. 2003. Fundamental principles and applications of natural gas hydrates. Nature 426(6964):
353-359.
U.S. EPA (United States Environmental Protection Agency) 1993. Current and Future Methane
Emissions from Natural Sources, EPA 430-R-93-011, Office of Air and Radiation, EPA, Washington
DC
Yamamoto, A., Y. Yamanaka and E. Tajika. 2009. "Modeling of methane bubbles released from large
sea-floor area: Condition required for methane emission to the atmosphere." Earth and Planetary
Science Letters 284(3-4): 590.
7-9
-------
Chapter 7. Gas Hydrates
7-10
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Natural seeps of CFU and other hydrocarbons from geologic sources deep within the Earth's crust have
been known to human civilization for millenia. Perpetual fires, crude oil and gas seeps, and the trance-
inducing properties of natural gas emissions from rock fissures have all been objects of fear, wonder, and
even worship throughout the centuries (e.g., Kvenvolden and Rogers, 2005; Etiope et al., 2006).
Until recently, however, the contribution of geologic emissions to the global CFU budget—arising from a
wide range of types of fissures and fractures in the Earth's crust—has been largely ignored. (Geologic
seeps are not considered to be a source of nitrous oxide emissions.) Most top-down analyses seeking to
resolve the global CFU budget through inverse modeling do not include emissions from this source in
their a priori estimates (Hein et al., 1997; Milakoff et al., 2004; Frankenberg et al., 2005; Bosquet et al.,
2006; Chen and Prinn, 2006; Bergamaschi et al., 2007). Bottom-up CFU budget estimates summarized in
the first, second, and third IPCC reports do not include a terrestrial geologic source (e.g., Ehhalt et al.,
2001). The few inverse modeling studies that do include a geologic source (Houweling et al., 1999;
Ferretti et al., 2005; Harder et al., 2007) tend to significantly underestimate the magnitude of these
emissions relative to bottom-up estimates (Etiope and Klusman 2002; Etiope, 2004, 2005; Etiope et al.,
2008b; Kvenvolden and Rogers, 2005) and/or misattribute emissions to single sources such as hydrates or
magma-emitting volcanoes.
Over the last decade, systematic measurements of CH4 emission rates from volcanic geothermal regions
and hydrocarbon sedimentary basins around the world—from the mud volcanoes (MVs) and seeps of
Italy to those of Azerbaijan—have firmly established both the reality of these emissions and their
significance on the global scale (Etiope, 1999; Etiope et al., 1999, 2002, 2004a,b, 2007a, 2008a;
Huseynov and Guliyev, 2004; Martinelli and Judd, 2004; Yang et al., 2004). Analyses of the atmospheric
abundance ratios of CFI4 also find a higher fossil component than can be accounted for through fossil fuel
emissions alone (Lassey et al., 2007a,b; Etiope et al., 2008b), verifying the likelihood of a relatively large
geologic source.
Specifically, new interpretation of atmospheric 14CFl4 measurements (Etiope et al., 2008b) suggests that
emissions from all fossil sources make up 30±5 percent of global CFI4 emissions. For a global source
estimated at 582±87 Tg CFIVyear, this implies a natural plus anthropogenic fossil source of 175±39 Tg
CHVyear. Anthropogenic emissions associated with fossil fuel extraction and consumption are estimated
at 90 to 100 Tg CFU/year. This range therefore leaves ample room for both additional unidentified
anthropogenic fossil sources, as well as geologic CFI4 emissions estimated to be on the order of 30 to 70
Tg CFU/year (Table 8-2). The best available estimate of present-day emissions of CFI4 from both
terrestrial and submarine geologic sources lies in the range of 42 to 64 Tg CFU/year, suggesting that, after
wetlands, geological sources may represent the second largest natural source of methane.
These findings have led to geologic sources being explicitly cited in the CFI4 budget section of the most
recent IPCC report (Denman et al., 2007) and assigned a specific emission category by the European
Environment Agency (2009). However, much work still remains to be done in resolving the contribution
of individual source types to the geologic methane budget.
8.1 Description of Emission Source
Geologic CFLj is emitted through fissures and fractures in the Earth's crust. Emissions arise from two
geologically distinct regions: (1) geothermal regions characterized by emissions from geothermal vents,
soil degassing, and magma-producing volcanoes and (2) sedimentary petroliferous or hydrocarbon-
containing basins characterized by emissions from both seepage and MVs. Emissions can also be
characterized by size: macroseepage consists of relatively large, visibly detectable, localized emissions
from identified geologic features and events such as MV, magma-producing volcanic eruptions, mid-
8-1
-------
Chapter 8. Terrestrial and Marine Geologic Sources
ocean ridges, and localized vents; diffuse soil exhalation occurs over broad areas in geothermal regions,
and microseepage overbroad areas in sedimentary basins (Figure 8-1). Finally, sources can be either
submarine (underwater) or terrestrial (land-based), although there is no essential difference in the CFi4
being emitted from these various sources: it is merely a matter of where the fault is geographically
located. However, before reaching the atmosphere, the gas is often modified and sometimes even
absorbed by the environment through which it passes—in the case of submarine sources, by the overlying
sediment layers and ocean water, and in the case of terrestrial sources, by the surface soil layers that can
contain both methanogenic or methanotrophic bacteria. We therefore differentiate CFLt emissions by
geological region of origin, by size, and by submarine vs. terrestrial sources (Figure 8-1).
GEOTHERMAL REGIONS
mag ma-producing
volcanoes
submarine
volcanoes
and
geothermal
vents soil
degassing
PETROLIFEROUS OR
HYDROCARBON-SEDIMENTARY REGIONS
mud
volcanoes
j k other
macroseeps submarine
microseeps macroseeps
_
Figure 8-1. Geologic methane sources can be categorized by region (geothermal or petroliferous), source size
(macro or micro), and location (terrestrial or marine).
Although primarily associated with areas of geothermal activity or hydrocarbon occurrence, known
geologic sources are widespread. Active MV regions with associated vents and microseepage have been
identified along the coastlines and continental shelves of nearly every continent (Dimitrov, 2003). CFi4
seepage has been reported from every sea and ocean, and in a broad range of oceanographic settings and
geological environments (Judd, 2003).
8.1.1 Mud Volcanoes in Petroliferous Sedimentary Regions
MVs are geologic structures formed as a result of the emission of gas, water, and sediments from the
Earth's crust. MVs generally occur in sedimentary areas and are often associated with natural gas and oil
deposits. For that reason, much of the gas emitted from MVs is CFi4, with other hydrocarbons and carbon
dioxide making up the remainder (Kvenvolden and Rogers, 2005). A global dataset surveying more than
140 terrestrial MVs from 12 countries (Etiope et al., 2009) showed that, on average, CFi4 makes up 90
percent of the gases emitted by MVs.
Unlike high-temperature magmatic or "traditional" volcanoes, MV emissions occur at comparatively low
temperatures. Instead of liquid magma, a semi-liquid muddy sediment is formed deep within the Earth's
crust. This mud mixture is then forced up through long, narrow openings or fissures to form a MV cone
or—in some cases—a "mud pie" (Dimitrov, 2003; Kopf, 2002). Individual MVs can also be very
different from one another, with some being less than a meter in diameter while others can cover up to
100 square kilometers. Etiope (2003 and elsewhere) cautions that the term "mud volcano" can actually
refer to a single edifice, a group of vents, or an entire cluster.
8-2
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Figure 8-2. Global distribution of mud volcanoes, including single mud volcanoes, separated mud volcano areas, and
mud volcano belts, as indicated by symbol 1. Symbols 2, 3, and 4 indicate areas of thinner (a) and thicker (b)
sediment, active compressional areas, and subductive zones, respectively.
MV formation can be triggered by a number of pressure-inducing events, including rapid sedimentation,
in situ gas generation, and structural or tectonic compression (Dimitrov, 2003). Similar to magmatic
volcanoes, MVs can also experience both quiescent periods and periods of eruption, with average
emission estimates changing accordingly. Significant emissions from MVs can still occur during
quiescent periods, during which the volcanoes can grow through gradual outflows of semi-liquid mud.
These can often be accompanied by what are known as "eternal flames" when the gases being emitted are
ignited. Eruptions occur periodically, with some volcanoes capable of launching mud and ash several
kilometers into the troposphere (Dimitrov, 2003).
Approximately 1,100 MVs have been documented onshore and in shallow water on continental shelves
(Dimitrov, 2002b). Anywhere from 1,000 to 100,000 MVs may exist below the ocean surface, on
continental slopes and abyssal plains (Milkov, 2000). Volcanoes are generally clustered together in belts
associated with active plate boundary areas (Figure 8-2; Dimitrov, 2003). Specifically, more than half of
the world's terrestrial MVs are located in the Alpine Himalayas Active Belt, which extends from Italy in
the west to Southeast Asia and Indonesia in the east. The largest concentration of terrestrial MVs occurs
in Azerbaijan, with over 700 documented examples. Large terrestrial and marine MV belts also occur
along the eastern and western sides of the Pacific Basin, and along the Caribbean coasts of Central and
South America (Dimitrov, 2003).
8.1.2 Seepage in Petroliferous Sedimentary Regions
Macroseepage in petroliferous or hydrocarbon sedimentary areas refers to emissions of gases from large,
visible features other than MVs. Macroseepage can occur via water-seeps or dry-seeps (Etiope et al.,
2009). In water-seeps, gaseous emissions are accompanied by bubbling springs, groundwater, or even
hydrocarbon wells. The water typically originates deep within the earth's crust and may have interacted
with the gas during its ascent to the surface. In contrast, dry-seeps consist of gaseous emissions only,
8-3
-------
Chapter 8. Terrestrial and Marine Geologic Sources
typically emitted from vents in outcropping rocks, through the soil, or via the beds of rivers or lakes.
Gases bubbling from wells or other shallow water bodies filled by groundwater should be considered dry-
seeps, as the gas only needs to cross surface water before reaching the atmosphere.
Gaseous emissions from dry seeps can be ignited, and many burn naturally. With continuous emissions,
"everlasting" flames from some dry-seeps have been continuously reported for centuries. One example is
the Chimaera seep in Turkey, considered by many to be the site of the first Olympic flame. Globally, the
number of terrestrial macro-seeps is estimated to exceed 10,000 (Clarke and Cleverly, 1991).
Micro-seeps are fissures in the Earth's crust that produce a slow, continuous flux of visibly undetectable
hydrocarbon gases, including CH4. In contrast to macroseepage, fluxes from these sources are not
identified by a visible emitting feature but rather are usually detected using closed-chamber systems, a
technique commonly applied to study the exchange of carbon-bearing gases at the soil-atmosphere
interface (e.g., Norman et al., 1997). This technique is currently used to detect methane fluxes migrating
along faultlines and upward from deep hydrocarbon reservoirs (Etiope, 1999; Klusman et al., 2000).
Figure 8-3. Sedimentary basins where terrestrial microseepage may occur (Kvenvolden and Rogers, 2005, after
Etiope and Klusman, 2002).
Terrestrial microseepage (the only kind that has been broadly documented so far, since locations of
submarine microseepage are very difficult to identify) is basically a diffuse emission of ClrU from soil,
where the CH4 is originating from underground natural gas reservoirs from depths of about 2 to 5
kilometers (Etiope and Klusman, 2002; Etiope, 2004). Microseepage is generally thought to be driven by
the natural buoyancy of gas relative to soil materials, migrating upward in bubble form along faults and
fractured rocks (Etiope and Martinelli, 2002). For that reason, microseepage is much more common in
faulted regions, as the faults give the gas a means to travel from its underground reservoir to the surface
(Etiope, 2005).
Although microseepage emissions often appear to be coming from surface soil itself, in fact the soil lies
over fault lines and other openings in the Earth's crust and—as verified by isotopic analysis—is merely
allowing gases trapped far beneath the surface to escape to the atmosphere (e.g., see Etiope and Klusman,
2002; Etiope, 2004; Kvenvolden and Rogers, 2005). In dry lands, soil is generally a net CFLj sink due to
methanotrophic bacteria in the soil. In areas where soil is a net producer of CFLt, this is an indicator that
8-4
-------
Chapter 8. Terrestrial and Marine Geologic Sources
microseepage is producing more ClrU than can be consumed by the soil (Etiope, 2005). For this reason,
microseepage generally occurs over sedimentary basins in a dry climate, where thermal decomposition of
ancient organic material has created petroleum and gas reservoirs deep below the Earth's surface (Figure
8-3; Kvenvolden and Rogers, 2005; Etiope and Klusman, 2002; Klusman et al., 1998).
Some differentiate between microseepage in sedimentary areas and emissions from geothermal regions
(e.g., Kvenvolden and Rogers, 2005). For this reason, microseepage in geothermal zones is generally
referred to as "diffuse soil exhalation" or "diffuse degassing" (see Section 8.1.4 below). However, both
sources consist of CFi4 escaping through a network of small, often indiscernible cracks in the Earth's
crust.
8.1.3 Submarine Seepage
Methane seeping into the marine environment must pass through sea floor sediments and ocean water
before entering the atmosphere. This represents a much more significant barrier to CFi4 production than
exists for terrestrial geologic sources. Methane passing through seafloor sediments is normally oxidized at
the sulfate-methane transition zone; only if CFi4 emissions exceed anaerobic consumption are CFi4
bubbles able to escape into the water column.
Once in the water, CFi4 can still be partially or completely dissolved and oxidized before reaching the
surface. The degree of dissolution in seawater depends mainly on the depth of water, water temperature,
and the size of the bubbles rising towards the surface. The fraction of bubbles reaching the sea surface is
the result of a balance between the rate of bubble dissolution, inflow of air from sea water, and growth
due to decreasing hydrostatic pressure as the bubble rises (Patro et al., 2001). The amount of gas entering
the atmosphere can be estimated as a function of seafloor depth, bubble size, concentration of dissolved
gas around the bubble plume, water temperature, and bulk fluid motions. For seeps at depths shallower
than 20 meters, almost all CFi4 emitted reaches the atmosphere. For deep vents on the order of 50 meters,
at least 50 percent of CFi4 bubbles with a radius greater than 5 mm survive. Below 100 to 300 meters,
CFi4 emissions from submarine seeps are not likely to have a significant impact on the atmosphere (e.g.,
Schmale et al., 2005).
Submarine features such as pockmarks, gas seeps, and gas-charged sediments are well-documented (see,
for example, discussion and references in Etiope and Klusman, 2002; Judd et al., 2002a). Pockmarks are
cone-shaped depressions, produced from the "blow-out" of gas and water, that occur within clays, silts,
and sands at depths down to thousands of meters. Typical pockmarks range in size from less than 1 meter
to 0.5 kilometers in diameter, and depths of more than 20 to 30 meters below the seafloor. Giant
pockmarks with diameters of 100 to 200 meters have been reported in Belfast Bay, Maine, and the
Barents Sea, Norway. Other areas of pockmarks and seeps have been found on the eastern Canadian
continental shelf, the Black Sea, the Adriatic Sea, and even the Arctic Ocean.
8.1.4 Volcanoes, Vents, and Other Geothermal Sources
Regions of geothermal activity can also produce CFi4 during eruptions of high-temperature, magma-
producing volcanoes, or through diffuse soil exhalation or degassing from the vents surrounding major
volcanoes. In contrast to the processes that produce CH4 in petroliferous, sedimentary basins, production
of CFI4 in geothermal regions is relatively low compared to production of other gases. Typical
concentrations of CFI4 in the volcanic gases from fumaroles and crater exhalations is on the order of 0.001
percent or less of total gaseous emissions. Measurements from 27 volcanoes around the world indicate an
average CFI4 concentration of about 0.005 percent or 50 parts per million and a median value of 0.0006
percent or 6 parts per million (Etiope et al., 2007b).
8-5
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Direct estimates of QrU emissions from magmatic volcanoes are relatively rare. Significant CELi
emissions have been measured near volcanoes in the Canary Islands (Hernandez et al., 1998), but other
volcanoes, such as Mauna Loa, do not appear to be emitting QrU on a regular basis (Ryan et al., 2006).
Hence, although magmatic volcanoes are also classified as macro-seeps due to their large physical size,
actual emissions from these types of volcanoes appear to be relatively low, contributing little to net global
emissions from geologic sources.
The distinction between volcanic and other geothermal emissions is based primarily on the form in which
the gases are produced below the surface. Volcanic emissions occur when the gases are released from the
craters or flanks of active, or historically active, volcanoes. Volcanic gases, released directly by magma
without being dissolved into water before reaching the surface, can be identified by their typically high
water vapor content and CO2 to ClrU ratio. In contrast, other geothermal emissions are generally the result
of boiling or degassing from an aqueous hydrothermal solution below ground. These include gases from
extinct volcanoes, paleo-volcanic zones, and CC^-rich "cold" vents in active tectonic zones, where the gas
originates gas from deep thermometamorphic processes and faults.
Macro-seeps through local vents, in terrestrial regions of active geothermal or volcanic activity and
undersea regions such as mid-ocean ridges, are widespread globally in both marine and terrestrial
environments (Figure 8-4; Wilson et al., 1973, 1974). Compared to volcanoes, the concentration of QrU
can be higher in gases emitted by degassing or soil exhalation. Even still, it typically reaches little more
than a few percent of the total volume of gas produced; thus, chemical analyses are generally required to
identify the composition of the gases being emitted (Kvenvolden and Rogers, 2005).
Even when macroseepage locations are relatively close to each other, the mixture of gases being emitted
can still vary widely. An example of this is provided by Kvenvolden and Rogers (2005), who compare
emissions from Yellowstone National Park (where emissions are dominated by carbon dioxide rather than
ClrU) and Grand Teton National Park (where ClrU is the primary constituent of gas seeps), both located in
Wyoming. Higher CH4 concentrations are generally found over faults that intersect the steam cap or the
more extensive liquid-dominated portion of a geothermal reservoir (Etiope et al., 2007b).
Figure 8-4. Documented onshore and offshore oil seeps, many of which also emit ChU (Kvenvolden and Rogers,
2005).
8-6
-------
Chapter 8. Terrestrial and Marine Geologic Sources
8.1.5 Isotopic Signature
As geologic CH4 emissions generally originate from underground reservoirs generated by the decay of
ancient organic matter, the majority of geologic emissions are estimated to be fossil in origin — i.e.,
radiocarbon- or 14C-depleted relative to more "modern" sources from wetlands or ruminants that have a
relatively high radiocarbon content (Etiope et al., 2007b; Judd et al., 2002a). At the same time, however,
CH4 produced in late Pleistocene and Holocene sediments in estuaries, deltas, and bays or trapped
beneath permafrost could also be formally considered geologic even though it does not necessarily
display a "fossil" signature. Contributions of this "recent" gas to geologic CH4 emissions is discussed by
Judd (2004) and Judd and Hovland (2007). Others suggest that the composition of geologic QrU
emissions may be altered by chemical processes during the emission process, including CJrU generation in
shallow sedimentary or soil layers through which the gases pass on their way to the atmosphere
(Dimitrov, 2003; Etiope et al., 2007a).
Typically, QrU produced by MVs and seeps in petroliferous, sedimentary basins has a carbon isotopic
composition (135C) ranging from -25 %oto -65, with an average of about -46 %o and -43 %o, respectively
(Etiope et al., 2009). Microseepage and marine seeps have basically the same signature and range,
although a larger microbial component occurs in marine seeps related to shallow sediments. In contrast,
geothermal and volcanic methane has a distinctive heavier isotopic ratio, ranging from -10%oto -20%o. In
all cases, the isotopic signature at the surface can be modified by isotopic fractionations due to bacterial
oxidation and diffusion. However, higher the gas flux, lower the chance of isotopic fractionation. For this
reason, the largest and most intense sources generally maintain an isotopic signature closer to that of the
original underground reservoir (Etiope et al., 2009).
8.2 Factors That Influence Emissions
Unlike many other natural CFLj sources, geologic CFLj emissions are not directly affected by changes in
climate, or most other factors that change over relatively short time scales from decades to centuries.
Over the last century, the only factor that has been proposed to have affected geologic CH4 emissions is
enhanced oil and gas extraction. It has been hypothesized that extraction of oil and gas from subsurface
reservoirs can decrease CFLj emissions from geologic vents due to a reduction in reservoir pressure that
had previously been forcing these gases to the surface (Quigley et al., 1999). In support of this hypothesis,
Hornafius et al. (1999) compared sonar measurements of bubble plumes near an oil platform in the Santa
Barbara Channel from 1973 and 1995 and estimated an 80 percent reduction in seepage over that time.
Based on isotopic analyses, Etiope et al. (2008a) estimate that pre-industrial emissions from geologic
sources were likely greater than modern emissions because petroleum exploitation has reduced emissions
from associated seeps.
Over longer (geologic-scale) time periods, on the order of millennia, the magnitude of CFLj emissions
from geologic sources is likely to have been affected by changes in surface faulting and seepage.
Specifically, Etiope et al. hypothesize that geologic emissions are likely to have increased during times of
increased seismic activity (Etiope et al., 2008a). These may happen locally over the short term for various
reasons, as one region becomes temporarily more active while another becomes less so. Over the longer
term and at the global scale, however, these shifts are hypothesized to be connected to changes from
glacial to inter-glacial periods as the Earth's crust rebounds, triggering increased seepage rates (Etiope et
al., 2008a).
Geologic CH4 emissions can certainly be altered, however, by the medium through which they pass
before being released into the atmosphere. As discussed previously, methanotropic or methanogenic
bacteria in soils have been observed to alter both the magnitude and the isotopic signature of CFLj
emissions from terrestrial microseepage and even, in some cases, macroseepage features (Dimitrov, 2003;
Etiope, 2005).
8-7
-------
Chapter 8. Terrestrial and Marine Geologic Sources
In the ocean, QrU uptake through the water column affects net emissions to the atmosphere. Under most
conditions, most of the QrU that seeps into the ocean — particularly from microseepage and smaller
macroseepage sources — is consumed by microbial oxidation in sediments and the water column (Judd,
2003). Seep size is a critical factor in the amount of QrU emitted to the atmosphere, with a greater
fraction of seep QrU being emitted to the atmosphere from large seeps (Clark et al., 2003). Using C-CIrU
source partitioning, a method that examines the radiocarbon content of ClrU dissolved in the water column
and emitted from seeps, Kessler et al. (2006) found that only 1.4 to 4.9 percent of QrU from seeps and
clathrates to the water column in the Black Sea reached the atmosphere. The oxidation and dissolution
rates of QrU in ocean water are dependent on the water temperature, the level of QrU saturation in the
water column, seep rate, and bubble size. As these factors may change under different climate conditions,
the fraction of the ClrU released into the water column that reaches the atmosphere is likely to be affected.
Another factor that affects QrU emissions from marine geologic seeps is sea level, which determines the
amount of water through which the ClrU emitted on the sea floor must pass before reaching the
atmosphere. For this reason, a decrease in sea level is likely to lead to an increase in QrU emissions from
marine seepage, while an increase in sea level would lead to a decrease in CFLj emissions (Boles et al.,
2001). More importantly, changes in sea level can cover or reveal near-shore sources. Converting a
previously marine source into a terrestrial source would significantly increase emissions as the overlying
ocean water was removed, while conversion of a terrestrial source into a marine one would decrease
emissions. Hence, Luyendyk et al. (2005) also proposed that greater exposure of continental shelves due
to lower sea level during glacial periods could have increased the amount of geologic CFLj emitted
directly to the atmosphere from continental sources relative to today's emission rates.
8.3 Current Global Emissions
Initially, global emissions of ClrU from geologic terrestrial sources (primarily magma-producing
volcanoes) were estimated relative to measured sulfur emissions from these sources, as these were much
better known (e.g., Houweling et al., 1999; Simpson et al., 1999). However, as recent observations have
shown, magmatic volcanoes likely account for only a small fraction of total geologic QrU emissions each
year.
Instead, systematic measurements in recent years combined with more sophisticated upscaling procedures
have clearly shown that, on a global scale, ClrU emissions from petroliferous, sedimentary areas are far
more important than emissions from geothermal regions. Individual features, including both MVs and
larger areas with active degassing fault lines, can produce on the order of tens to hundreds of tonnes of
CH4 per year (Castaldi and Tedesco, 2005; Etiope et al., 2004a,b, 2007a; Yang et al., 2004).
Global estimates of QrU emissions from the individual source types listed in Section 8.1.1 are provided
below and summarized in Table 8-1. Table 8-2 summarizes global-scale bottom-up geologic
emissions estimates.
8.3.1 Mud Volcanoes in Petroliferous Sedimentary Regions
Currently, estimates of global and annual emissions of QrU from MVs require extrapolation of a limited
number of site -specific emission measurements and eruptive characteristics to MV regions around the
world. This introduces a significant amount of uncertainty into global estimates (Table 8-1).
Annual emissions from individual MVs cover a wide range. Single vents or craters of small MVs (from 1
to 5 meters in height) can produce on the order of tens of tons each year. A single large MV, consisting of
tens or even hundreds of vents, can emit hundreds of tons per year. Eruptions of MVs can release
thousands of tons of ClrU in just a few hours.
8-8
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Methane emissions from MVs are not produced solely by visible craters and vents; significant amounts of
gas are also released as diffuse degassing from the soil. The amount of gas released into the atmosphere
by this type of microseepage, calculated for the whole MV area, is often comparable to or even larger
than the output from the vents themselves (Etiope and Milkov, 2004; Etiope, 2005; Hong and Yang,
2007). Annual average CH4 flux for MV areas, including both microseepage and vents but excluding
episodic eruptions, ranges from 100 to 1,000 tons per square kilometer.
Due at least in part to the relatively large assumptions that must be made when estimating global values
from spatially and temporally variable sources, particularly from submarine MVs, global estimates of
emissions of CFL from MVs have been the source of much debate over the past few years (Kopf, 2003,
2005; Milkov and Etiope, 2005). Kopf (2002) first estimated total CFL emissions to the atmosphere from
terrestrial and marine MVs to lie between 0.08 and 1.29 Tg/year. The estimates, which depended on
volcano size (small, medium, large) and minimum and maximum QrU flux rate estimates averaged from
the literature, ranged from 0.197 to 123 Tg CHVyear, most of which is attributed to marine MVs (Kopf,
2003). According to these estimates, emissions from terrestrial MVs would be negligible when compared
to marine emissions. Kopf (2003) does note that much of the QrU emitted from marine MVs may not
reach the atmosphere (instead being precipitated at the sea floor, oxidized in the water column, or
dissolved into the water column), but does not provide an estimate of the atmospheric emissions. Kopf s
findings were challenged by Milkov and Etiope (2005), who argued, based on their survey of terrestrial
MV regions and associated emissions, that Kopf had used a smaller, not a larger, sample size than
previously, and had committed a number of mathematical errors that led to a best estimate of QrU
emissions from all terrestrial MVs together that was lower than the emissions actually measured at
individual locations in Sicily, eastern Azerbaijan, and eastern Romania (Etiope et al., 2002, 2004a,b).
Alternate estimates for QrU released from onshore and shallow offshore MVs range from 10.2 to 12.6 Tg
CFL/year (Dimitrov, 2002b), neglecting any dissolution or oxidation that may occur in the water column,
and 33 Tg CFL/year, composed of 15.9 Tg CFL/year during quiescent periods and 17.1 Tg CFL/year
during eruptions (Milkov et al., 2003). Of the 33 Tg CFL/year total emissions, 6 Tg CFL/year are
estimated to directly enter the atmosphere from onshore and shallow offshore MVs and the remaining 27
Tg CFL/year are emitted to the water column from deep-water MVs. The only global estimate based on
experimental measurements and emission factors is that proposed by Etiope and Milkov (2004), of 6 to 9
Tg CFL/year.
As increasing numbers of MV locations are identified both on and offshore, and measurements of CFL
emissions made during quiescent and eruptive periods are accumulated, global estimates of MV emissions
will continue to be refined and the uncertainty reduced.
8.3.2 Seepage in Petroliferous Sedimentary Regions
In contrast to emissions from MVs and other macroseepage features, terrestrial microseepage rates from
sedimentary basins characterized by underground petroleum and gas reservoirs have been directly
measured for only a limited number of individual locations. These include regions in Italy (Etiope et al.,
2007b), Greece (Etiope et al., 2006), Romania and Azerbaijan (Etiope et al., 2004a,b), the United States
(Klusman et al., 2000; Klusman, 2003), and the former Soviet Union (Voitov, 1975; Balakin et al., 1981).
Fluxes directly measured or visually estimated from 50 gas seeps in 1 1 countries show annual emissions
between 5 and 100 tons of CFL per year for gas seeps with a diameter greater than 1 meter (Etiope et al.,
2008b). Up to 2,000 tons per year can be emitted from large seeps with diameters exceeding 1 to 2
meters. In a seep site, however, gas is not released only through macroseepage from the vents. Large
amounts are also produced by microseepage through the surrounding soil over broader areas on the order
of 103 to 104 square meters. The large areas are due to the fact that a macro-seep is generally the primary
expression of a larger gas-bearing fracture system.
8-9
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Numerous gas flux surveys show that the amount of gas released from the surrounding soil can be up to 3
times greater than that from the vent alone (Etiope et al., 2004a,b, 2006, 2007a; Hong and Yang, 2007).
Using this methodology, Etiope et al. (2008b) estimate emissions for about 12500 macro- and micro-
seeps worldwide of about 3-4 TgCFL per year (Etiope et al., 2008b).
Global microseepage rates are extrapolated based on the area of known continental areas overlying
sedimentary basins, multiplied by areal averaged emission rates derived from observations such as those
listed above. Preliminary models suggested that the hydrocarbon-prone sedimentary basins in a dry
climate produce a mean microseepage flux of 4.4 mg CH4 per square meter per day (Klusman et al., 1998;
2000). Assuming 90 percent consumption at this microseepage rate in dry soil gives a global annual
emission estimate of at least 7 TgCFL per year.
Global "potential" microseepage of 14 to 28 Tg CFL/year has been estimated by Etiope (2005) on the
basis of the global area covered by Total Petroleum Systems (the system commonly used in petroleum
geology, which includes all elements of gas production, accumulation, and seepage, the area of which can
be estimated from specific maps) and a limited flux data-set. More recently, Etiope and Klusman (2009)
have refined global estimates of annual emissions from terrestrial seepage to lie between 11 and 25 Tg
CHVyear based on 563 measurements from a range of hydrocarbon-prone basins in the United States and
Europe.
8.3.3 Submarine Seepage
CH4 emissions from marine seeps have also been estimated for a limited number of locations. These
include Coal Oil Point, Santa Barbara, California (Hornafius et al., 1999; Clark et al., 2000; Washburn et
al., 2005); the Cascadia Continental Margin (Collier and Lilley, 2005); the Black Sea (Dimitrov, 2002a;
Schmale et al., 2005; Kessler et al., 2006; Wallmann et al., 2006); the continental shelf of the U.K.,
including Tony Bay, Firth of Forth, Scotland (Judd et al., 1997, 2002b); and the Costa Rican coast (Mau
et al., 2006). In Europe, the most intense and largest gas bubble plumes, visible even from the sea surface,
occur in the coastal areas, from inter-tidal zones to 200-300 meters of depth along the coastlines of
Bulgaria, Romania, Ukraine and Georgia. Emission estimates from these areas, although rough and
incomplete, cannot be ignored (Etiope, 2008).
Kessler et al. (2006) estimated basin-wide flux of CFI4 from seeps and clathrates to the water column in
the Black Sea using 14C-CFl4 source partitioning. The approach was based on calculation of gas diffusion
in water, but was not representative of the entire Black Sea area as it only partially accounted for direct
emission from bubbles in coastal zones. Emissions from seeps and clathrates were estimated to be 3.6 to
4.28 Tg CFL/year to the water column, and 0.05 to 0.21 Tg CFL/yearto the atmosphere (Kessler et al.,
2006). This would correspond to a CFI4 transmission efficiency (i.e., the fraction of CFL emitted to the
water column that reaches the atmosphere) of 1.4 to 4.9 percent.
In comparing this estimate to those of Kvenvolden (33 to 60 percent; Kvenvolden et al., 2001;
Kvenvolden and Rogers, 2005), the nature of the CFL seepage must be considered. A greater fraction of
CFLj will reach the atmosphere during more rapid upwelling flow and when larger bubbles are produced
during eruptive episodes such as blowouts (Leifer et al., 2004, 2006). Both factors reduce CH4 loss to the
water column during bubble transport. Therefore, transient "eruptive" emissions are more likely to
transport CFL to the atmosphere than background emissions because bubbles are larger meaning that they
dissolve more slowly, and a greater fraction of CFL remains in bubbles at surface; upwelling flows are
produced, accelerating bubble transport to the surface (Leifer and Clark, 2002); and plume water becomes
saturated, reducing net transport of CFL out of bubbles (Leifer et al., 2000, 2004; Leifer and Judd, 2002).
Leifer and Clark (2002) describe characteristics of the bubble plume from a marine seep field located
offshore from Coal Oil Point, Santa Barbara, California, and found that, based on three seeps of different
sizes, bubbles were about 90 percent CFL at the base of the plume, but only 60 percent CFL at the surface
8-10
-------
Chapter 8. Terrestrial and Marine Geologic Sources
due to dissolution of QrU into the water column. Using a bubble model, MacDonald et al. (2002) found
that only bubbles of a certain intermediate size (about 4 to 5 millimeters) could reach the surface; smaller
bubbles dissolved rapidly and larger bubbles broke into smaller bubbles, which would dissolve.
Several approaches have been used to produce global emissions estimates for submarine seepage into the
ocean and atmosphere. The estimates of Hovland et al. (1993) are based on a review of case studies
representing different geographic and oceanographic environments, while Judd (2004) estimates
emissions to the atmosphere from marine seepage based on previously published estimates rather than
new calculations. Hornafius et al. (1999) extrapolated their Coal Oil Point emission estimates to the high-
seepage-potential areas of the world's continental shelves. Kvenvolden et al. (2001) took two approaches
to estimating QrU seepage to ocean and atmosphere, based on findings of a workshop held to assess the
magnitude of marine ClrU seepage: (1) a review and synthesis of published estimates and (2) an estimate
based on total availability of ClrU for seepage from all global geologic sources and assumptions regarding
the half-life of the geologic reservoir.
From these estimates, extrapolations have been made that approximate global emissions from marine
seepage. These estimates range from 8 to 65 Tg CHVyear emitted to the ocean, with an estimated 10 to 30
Tg CHVyear reaching the atmosphere from marine seeps. The average, 20 Tg CHVyear, is the current
consensus value for emissions from submarine sources reaching the atmosphere (Kvenvolden et al., 2001;
Judd, 2004). Table 8-1 summarizes these emission estimates.
8.3.4 Volcanoes, Vents, and Other Geothermal Sources
Emissions of QrU from magmatic volcanoes were originally estimated through relating QrU to measured
sulfur emissions (e.g., Lacroix, 1993). In this way, the amount of QrU produced by magma-producing
volcanoes on land was estimated to lie between 0.8 and 6.2 Tg CHVyear, averaging about 4 Tg CHVyear.
Based on this work, some inverse modeling studies have mistakenly grouped all geologic emissions into a
"volcanic" category, estimating its magnitude at 3.5 ± 3 Tg CHVyear (e.g., Houweling et al., 1999), or 7
Tg CHVyear for a "misc. ground" category that includes volcanoes and hydrothermal vents (Harder et al.,
2007). However, as noted previously, these estimates are likely both too high for an individual volcanic
source and too low for a total geologic source.
A simple calculation based on the average ratio of CO2 to ClrU emissions from volcanoes, assuming a
global volcanic CO2 flux of 300 Tg per year, suggests a global CH4 source of less than 1 Tg per year. This
suggests that volcanoes are not an important QrU source (Ryan et al., 2006; Etiope et al., 2007b). Overall,
Etiope et al. (2007b) concludes that magmatic volcanoes are not a major source of QrU emissions on an
annual basis, although individual eruptions may sporadically produce globally significant amounts.
Geothermal systems, often independent of active volcanoes, are much more important, as recently shown
by bottom-up estimates in Europe (Etiope, 2008). Etiope et al. (2007a) estimate regional geothermal
emissions for Europe alone (primarily due to emissions from geothermal regions in Italy, Greece, and
Iceland) on the order of 0. 1 Tg CHVyear.
Other minor geological sources include natural exhalation from coal-bearing rocks (influenced by mining
activities), and degassing from crystalline basement and mantle; but no global estimates of atmospheric
emissions from these sources have been proposed as yet.
8-11
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Table 8-1. Estimates of Methane Emissions from Individual Geologic Sources
Annual
Emissions
(Tg CH^/Year)
Sources Included
Reference
Mud Volcanoes
10
5
2-10
5
6-9
6
Terrestrial and shallow water
Terrestrial and shallow water
Terrestrial
Terrestrial
Terrestrial and shallow marine
Terrestrial and marine
Dimitrov, 2002b
Dimitrov, 2003
Etiope and Klusman, 2002
Kvenvolden and Rogers, 2005
Etiope and Milkov, 2004
Milkovetal., 2003
Other Macroseepage
3-4
Terrestrial seeps excluding MVs
Etiope et al., 2008b
Microseepage
>7
14-28
10-25
Terrestrial diffuse
Terrestrial diffuse
Terrestrial diffuse
Etiope and Klusman, 2002
Etiope et al., 2005
Etiope and Klusman, 2009
Submarine Seepage
18-48
8-65
20
30-50
20 (1 0-30)
Reaching the atmosphere
Emitted to oceans
Emitted to oceans
Emitted to oceans
Reaching the atmosphere
Hornafiuset al., 1999
Hovland etal., 1993
Judd, 2004
Kvenvolden et al., 2001; Kvenvolden and
Rogers, 2005
Volcanoes, Vents, and Other Geothermal Sources
1.7-9.4
2.5-6.3
<1
Geothermal and volcanic
Geothermal
Volcanic
Lacroix, 1993
Etiope and Klusman, 2002; Etiope et al.,
2008b
8-12
-------
Chapter 8. Terrestrial and Marine Geologic Sources
8.3.5 Global Emissions
Despite the relatively large uncertainty in emissions estimates from individual geologic sources, more
recent bottom-up estimates of global emissions are relatively consistent (Table 8-2). The most complete
estimate of global ClrU emissions from geologic sources has been made by Etiope et al. (2008b), who
estimates total global geologic emissions of 42 to 64 Tg CHVyear, broken down into emissions of 6 to 9
from MVs (Etiope and Milkov, 2004), 10 to 25 from microseepage (Etiope, 2005), 2.5 to 6.3 from
geothermal flux (Etiope and Klusman, 2002), about 20 from submarine seeps (Kvenvolden et al., 2001),
and less than 1 Tg CHVyear from magma-producing volcanoes. When these additional, documented
sources are included, the geologic CH4 budget increases to 42 to 64 Tg/year.
Table 8-2. Global Estimate of Methane Emissions from Geologic Sources
Annual Emissions
(Mt CH/yr)
Sources Included
Reference
Bottom-Up Estimates
(based on extrapolation of observed emission rates and number of geologic features worldwide)
30-70
35-45
40-60
16-40
7-20
40
53 (42-64)
30-80
Terrestrial and marine mud volcanoes and
vents, microseepage, and geothermal regions
Terrestrial and marine mud volcanoes and
vents, low estimate of microseepage, and
geothermal regions
Terrestrial and marine mud volcanoes and
vents, microseepage, and geothermal regions
Terrestrial (not including microseepage) and
submarine to the ocean
Terrestrial (not including microseepage) and
submarine to the atmosphere
Macroseepage, microseepage, mud volcanoes,
and miscellaneous
Best guess
Extended range (all geologic sources)
Etiope and Klusman,
2002
Etiope and Milkov, 2004
Etiope, 2004
Judd et al., 2002a
Kvenvolden and
Rogers, 2005
Etiope et al., 2008a
Top-Down Estimates
(based on inverse modeling studies using isotopic ratios to constrain the global methane budget)
20a
55-70b
85
All natural "fossil" sources (terrestrial, marine)
Includes natural geologic sources and
abandoned coal mines
May include unresolved anthropogenic fossil
emissions > current estimate of 90 Tg/yr
Ferrettietal., 2005
Lassey etal., 2007a
Etiope etal., 2008b
An a priori estimate that was held constant to estimate the contribution of biomass burning emissions to the
global ChU budget over the last two millennia. Authors acknowledge actual magnitude could be much larger.
Value inferred from Lassey et al. (2007a) estimate of a 30 percent contribution of fossil emissions to the global
budget and a previous fossil source estimate of 90 to 120 Tg CH4/yearfrom Ehhalt et al. (2001) based on a 20
percent contribution from fossil methane, for a total budget of 560 Tg ChU/yr.
8-13
-------
Chapter 8. Terrestrial and Marine Geologic Sources
The largest uncertainty lies in estimating submarine emissions. In the Etiope (2004) budget, marine
geologic emissions are estimated at 20 Tg CFL/year, based on the estimated range of 10 to 30 Tg
CFL/year (Judd, 2004) resulting from theoretical assumptions described in Kvenvolden et al. (2001). In
contrast, estimates of onshore emissions are based on direct measurements and upscaling procedures
based on standard concepts of emission factors as applied to both homogeneous areas (for diffuse
emissions such as microseepage) and point sources (for individual macroseepage features such as MVs).
In terms of other estimates of global CFL emissions from geologic sources, Judd et al. (2002a) estimate
total geologic CFL emissions of 16 to 40 Tg CFL/year, with 6.6 to 19.5 Tg CFL/year reaching the
atmosphere, while Kvenvolden and Rogers (2005) estimate total geologic CFL emissions to the
atmosphere of 45 Tg CFL/year. (Kvenvolden and Rogers' figure is broken down as 25 Tg CFL/year from
macroseeps, 7 Tg CFL/year from microseeps, 5 Tg CFL/year from MVs, and 8 Tg CFL/year from
miscellaneous sources including magmatic volcanoes.) Global geologic CFL emissions estimated using
these methods are summarized in Table 8-2.
New analyses by Lassey et al. (2007a) that combine isotopic records with a mass-balance modeling
approach (as described in Lassey et al., 2007b) find that fossil CFL likely contributed 30 percent of the
global budget for 1986 through 2000. Reconciling this estimate with current bottom-up CFL budgets
requires several significant adjustments to conventional wisdom regarding the relative magnitude of
individual CFL sources. These adjustments include requiring a substantially larger natural source of fossil
CH4, of similar magnitude to the emission estimates provided in Table 8-2. Thanks to this work, global
budgets are now beginning to be reconciled with bottom-up estimates of CFL emissions from geologic
sources as isotopic analyses confirm the likely existence of a geologic source of fossil CH4 significantly
larger than previously thought. Estimates of natural fossil emissions either used (Ferretti et al., 2005) or
estimated (Lassey et al., 2007a; Etiope et al., 2008b; Schaefer and Whiticar, 2008) by inverse modeling
and isotopic analyses are summarized in Table 8-2.
Based on the bottom-up estimates presented in Table 8-2, combined with the isotopic and modeling
analysis of Lassey et al. (2007a,b), it is therefore likely that average annual geologic CFL emissions from
both terrestrial and marine sources could lie within the most recent range of 42 to 64 Tg CFL/year
estimated by Etiope et al. (2008b). Incorporating the additional uncertainty in sub-oceanic emissions
expands this range to 32 to 74 Tg CFL/year, which we present here as the best-available estimate of
present-day emissions of CFL from both terrestrial and submarine geologic sources.
8.4 Future Emission Scenarios
As noted previously, relatively few climate- or human-related factors are hypothesized to be capable of
influencing emissions of CFL from geologic sources. Long-term changes in geologic CFL emissions and
potential causes for these changes inferred from ice core records may have implications for future
emissions, however.
Etiope et al. (2008a) hypothesize that the large-scale extraction of natural gas and oil from underground
reservoirs, over the last century or two, may have decreased geologic CH4 emissions by reducing
underground pressure that had previously been forcing gases up to the surface. This suggestion is
supported by Hornafius et al. (1999), who compared sonar measurements of bubble plumes near an oil
platform in Santa Barbara Channel from 1973 and 1995. They estimated an 80 percent reduction in
seepage over that time, likely due to oil and gas extraction from nearby locations. These past changes
suggest that continued fossil fuel extraction, particularly in areas with significant geologic CFL emissions,
could reduce surface emissions of CFL from that region's soils, fissures, and vents.
Over longer time periods, it has been suggested that geologic CFL emissions may be higher following
deglaciation events, as seismic activity increases (Etiope et al., 2008a). Given that the deglaciation of
Greenland and Antarctica may already be underway, this link suggests a potential positive feedback in the
8-14
-------
Chapter 8. Terrestrial and Marine Geologic Sources
future whereby crustal rebound in these regions could trigger increased seismic activity, increasing
regional CFU emissions from the Earth's crust //these regions lie above petroleum and gas deposits where
significant amounts of CFU gas would be stored.
Judd et al. (2002a) and Luyendyk et al. (2005) suggest a mechanism by which submarine emissions could
have been higher during ice ages. Due to lower sea levels, a greater number of marine sources—including
fissures, vents, and hydrates—were directly exposed to the atmosphere. Greater exposure of continental
shelves due to lower sea level during glacial periods would increase the amount of geologic CFU emitted
directly to the atmosphere from continental sources relative to modern emissions. Given that global sea
level has already increased almost 20 centimeters since pre-industrial times, and is projected to increase
by at least that much—if not more than five times that—by the end of the century (Rahmstorf, 2006), this
would suggest a possible decrease in geologic CFU emissions that reach the atmosphere if sources
currently located in coastal regions were inundated. Higher sea levels would also be likely to decrease
seabed CFU seepage due to an increase in the hydrostatic pressure. This would discourage upward gas
migration. Furthermore, the proportion of seabed seepage gas surviving passage through the water
column decreases as the water depth increases (Judd et al., 2002a).
Hill et al. (2006) estimate a three- to fourfold increase in CFU emissions (0.3 to 0.4 Tg CHVyear) from the
Santa Barbara Channel during the deglaciation relative to modern emissions (>0.10 Tg CHVyear). The
authors estimate global marine seepage source of 90 Tg CFU/year during the deglaciation, although they
note that this would require the unlikely synchronous response of CH4 seeps around the world. Etiope et
al. (2008a) expand on this to hypothesize that global emissions from geologic sources during the
Quaternary could contribute to increased CFU concentrations during the late Quaternary. This assumes
that greater exposure of continental shelves due to lower sea level during glacial periods would increase
the amount of geologic CFU emitted directly to the atmosphere from continental sources relative to
modern emissions, and that pre-industrial emissions from geologic sources were greater than modern
emissions because petroleum exploitation has reduced emissions from associated seeps.
While geologic CFU emissions have very likely changed in the past and are likely to continue to change in
the future, these mechanisms are too speculative to use as a basis to estimate even the potential sign of
future changes in geologic CFU emissions. However, they do serve as an important reminder that even
emissions from seemingly stable sources are not necessarily impervious to the consequences of human
interference with the climate system.
8.5 Areas for Further Research
Although emissions of CFU from a wide variety of geologic sources are now well-documented, much
work remains to be done in extrapolating site-specific and time-limited observations to refine global and
annual estimate of CFU emissions from these sources. Issues range from in situ observational techniques
(including units of measurement) to the statistical methods and geographical data used to extrapolate
local-scale observations to estimate the global average contribution of these sources to the global total.
For example, much of the CFU seepage through submarine seeps and MVs is consumed within the
sediment by anaerobic oxidation (Niemann et al., 2005, 2006; Wallmann et al., 2006). What is not well
known is the overall or average oxidation rate of CFU for all sources. For submarine emissions, the
oxidation rate is highly dependent on the residence time of the CH4 in the water column, the size of the
CH4 bubbles, the rate of CFU release, and the CFU saturation in the surrounding waters. For terrestrial
emissions, the oxidation rate is dependent on the bacterial populations (either methanogenic,
methanotrophic, or both) present in the layers of soil through which the gases must pass before being
released into the atmosphere. In order to obtain an accurate account of the CFU flux into the atmosphere
from these geologic sources, a better understanding of in situ CFU dissolution, oxidation, and production
rates is needed.
8-15
-------
Chapter 8. Terrestrial and Marine Geologic Sources
The primary area where additional research is required, however, is to better quantify macro-seep
emission factors, spatial and temporal distributions of macro-seeps (including, for example, the number of
active seeps and the frequency of MV eruptions), and the extent of micro-seeps over dry land areas. As
microseepage is one of the largest contributors to geologic ClrU emissions, it is particularly important to
develop refined estimates of its emission factors and active areas.
Specific uncertainties relate to identifying:
• The number of MVs around the world.
• The area of faulted land producing a net QrU flux into the atmosphere.
• The frequency of eruption events characterized by significantly higher emission rates.
• Spatial and temporal variations in emission rates.
• CH4 oxidation rates in both ocean water and soils, which determines the transfer efficiency
through the water column for marine sources and sedimentary layers for continental sources.
• Terrestrial microseepage emission rates.
The uncertainties described above all relate to improving the ability of bottom -up budget analyses to
accurately assess global ClrU emissions from geologic sources. In addition to increasing the number of
measurement studies and the identification rates of these geologic features and regions with the aim of
improving bottom-up estimates, however, top-down techniques combining observations with mass-
balance analyses and isotopic modeling also show significant potential to constrain the magnitude of
global CH4 emissions from geologic sources. Recent analyses (Lassey et al., 2007a,b; Etiope et al.,
2008b) find that fossil sources contribute approximately 30 percent of the global budget. This implies that
a fossil geologic source of the magnitude suggested initially by Etiope and Klusman (2002), refined by
Etiope (2004) and Etiope et al. (2008b), is plausible.
8.6 References
Balakin, V.A., G.A. Gabrielants, I.S. Guliyev, F.G. Dadashev, V.M. Kolobashkin, A.I. Popov, and A.A.
Feyzullayev. 1981. Test of experimental study of hydrocarbon degassing of lithosphere of South
Caspian Basin and adjacent mountains systems, using laser gas-analyser "Iskatel-2." Dokl. Akad. Nauk
SSSR 260(1): 154-156 (in Russian).
Bergamaschi, P., C. Frankenberg, J.F. Meirink, M. Krol, F. Dentener, T. Wagner, U. Platt, J.O. Kaplan, S.
Korner, M. Hermann, E.J. Dlugokencky, and A. Goege. 2007. Satellite chartography of atmospheric
methane from SCIAMACHY onboard ENVISAT: 2. Evaluation based on inverse model simulations.
Journal of Geophysical Research-Atmospheres 112(D2): D02304.
Boles, J. R., J. F. Clark, I. Leifer, and L. Washburn. 2001. Temporal variation in natural methane seep
rate due to tides, Coal Oil Point area, California. Journal of Geophysical Research-Oceans 106(C1 1):
27077-27086.
Bousquet, P., P. Ciais, J.B. Miller, E.J. Dlugokencky, D.A., Hauglustaine, C. Prigent, G.R. Van der Werf,
P. Perlin, E.G. Brunke, C. Carouge, R.L. Langenfelds, J. Lathiere, F. Papa, M. Ramonet, M. Schmidt,
L.P. Steele, S.C. Tyler, and J. White. 2006. Contribution of anthropogenic and natural source to
atmospheric methane variability. Nature 443(71 10): 439-443.
Castaldi, S., and D. Tedesco. 2005. Methane production and consumption in an active volcanic
environment of southern Italy. Chemosphere 58(2): 131-139.
8-16
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Chen, Y.H., and R.G. Prinn. 2006. Estimation of atmospheric methane emissions between 1996 and 2001
using a three-dimensional global chemical transport model. Journal of Geophysical Research-
Atmospheres lll(DlO): D10307.
Clark, J.F., L. Washburn, J.S. Hornafius, and B. P. Luyendyk. 2000. Dissolved hydrocarbon flux from
natural marine seeps to the southern California Bight. Journal of Geophysical Research-Oceans
105(C5): 11509-11522.
Clark, J.F., I. Leifer, L. Washburn, and B.P. Luyendyk. 2003. Compositional changes in natural gas
bubble plumes: observations from the Coal Oil Point marine hydrocarbon seep field. Geo-Marine
Letters 23(3-4): 187-193.
Clarke, R., and R. Cleverly. 1991. Leakage and post-accumulation migration. In: W. England, and A.
Fleet (eds.). Petroleum Migration. Geological Society Special Publication No. 59. pp. 265-271.
Collier, R.W., and M.D. Lilley. 2005. Composition of shelf methane seeps on the Cascadia Continental
Margin. Geophysical Research Letters 32(6): L06609.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press, pp. 499-587.
Dimitrov, L. 2002a. Contribution to atmospheric methane by natural seepages on the Bulgarian
continental shelf. Continental Shelf Research 22(16): 2429-2442.
Dimitrov, L.I. 2002b. Mud volcanoes—The most important pathway for degassing deeply buried
sediments. Earth-Science Reviews 59(1-4): 49-76.
Dimitrov, L.I. 2003. Mud volcanoes—A significant source of atmospheric methane. Geo-Marine Letters
23(3-4): 155-161.
Ehhalt, D., M. Prather, F. Dentener, R. Derwent, E. Dlugokencky, E. Holland, I. Isaksen, J. Katima, V.
Kirchhoff, P. Matson, P. Midgley, and M. Wang. 2001. Atmospheric chemistry and greenhouse gases.
In: J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A.
Johnson. Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge
University Press, pp. 239-287.
Etiope, G. 1999. Subsoil CO2 and CH4 and their advective transfer from faulted grassland to the
atmosphere. Journal of Geophysical Research 104(D14): 16889-16894.
Etiope, G. 2004. New directions: GEM—Geologic emissions of methane, the missing source in the
atmospheric methane budget. Atmospheric Environment 38(19): 3099-3100.
Etiope, G. 2005. Mud volcanoes and microseepage: The forgotten geophysical components of
atmospheric methane budget. Annals of Geophysics 48(1): 1-7.
Etiope, G. 2006. Evaluation of geological emissions of methane in Europe. Report for the NATAIR
Project. E.U. Contract 513699. 25 pp.
8-17
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Etiope, G. 2008. Natural emissions of methane from geological seepage in Europe. Atmospheric
Environment 43(7): 1430-1443.
Etiope, G., P. Favali. 2004. Geologic emissions of methane from lands and seafloor: Mud volcanoes and
observing systems. Environmental Geology 46(8): 987-987.
Etiope, G., and R.W. Klusman. 2002. Geologic emissions of methane to the atmosphere. Chemosphere
49(8): 777-789.
Etiope G., and R. Klusman. 2009. Microseepage in drylands: Flux and implications in the global
atmospheric source/sink budget of methane. Global Planetary Change, in press.
Etiope, G., and G. Martinelli. 2002. Migration of carrier and trace gases in the geosphere: An overview.
Physics of the Earth and Planetary Interiors 129(3-4): 185-204.
Etiope, G., A.V. Milkov. 2004. A new estimate of global methane flux from onshore and shallow
submarine mud volcanoes to the atmosphere. Environmental Geology 46(8): 997-1002.
Etiope, G., P. Beneduce, M. Calcara, P. Favali, F. Frugoni, M. Schiattarella, and G. Smriglio. 1999.
Structural pattern and CO2-CH4 degassing of Ustica Island, Southern Tyrrhenian Basin. Journal of
Volcanology and Geothermal Research 88(4): 291-304.
Etiope, G., A. Caracausi, R. Favara, F. Italiano, and C. Baciu. 2002. Methane emission from the mud
volcanoes of Sicily (Italy). Geophysical Research Letters 29(8): Art. No. 1215.
Etiope, G., A. Caracausi, R. Favara, F. Italiano, and C. Baciu. 2003. Reply to comment by A. Kopf on
"Methane Emissions From the Mud Volcanoes of Sicily (Italy)" and notice on CFLt flux data from
European mud volcanoes. Geophysical Research Letters 30(2): Art. No. 1094.
Etiope, G., C. Baciu, A. Caracausi, F. Italiano, and C. Cosma. 2004a. Gas flux to the atmosphere from
mud volcanoes in Eastern Romania. Terra Nova 16(4): 179-184.
Etiope, G., A. Feyzullayev, C.L. Baciu, and A.V. Milkov. 2004b. Methane emission from mud volcanoes
in Eastern Azerbaijan. Geology 32(6): 465-468.
Etiope, G., G. Papatheodorou, D. Christodoulou, P. Favali, and G. Ferentinos. 2005. Gas hazard induced
by methane and hydrogen sulfide seepage in the NW Peloponnesus Petroliferous Basin (Greece).
Terrestrial Atmospheric and Oceanic Sciences 16(4): 897-908.
Etiope, G., G. Papatheodorou, D. Christodoulou, M. Geraga, and P. Favali. 2006. The geological links of
the ancient Delphic Oracle (Greece): A reappraisal of natural gas occurrence and origin. Geology
34(10): 821-824.
Etiope, G., G. Martinelli, A. Caracausi, and F. Italiano. 2007a. Methane seeps and mud volcanoes in Italy:
Gas origin, fractionation and emission to the atmosphere. Geophysical Research Letters 34(14): Art.
No. L14303.
Etiope, G., T. Fridriksson, F. Italiano, W. Winwarter, J. Theloke. 2007b. Natural emissions of methane
from geothermal and volcanic sources in Europe. Journal of Volcanology and Geothermal Research
165(1-2): 76-86.
Etiope, G., A. Milkov, and E. Derbyshire. 2008a. Did geologic emissions of methane play any role in
Quaternary climate change? Global and Planetary Change 22(1-2): 79-88.
8-18
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Etiope, G., K.R. Lassey, R. Klusman, and E. Boschi. 2008b. Re-appraisal of the fossil methane budget
and related emission from geologic sources. Geophysical Research Letters, in press.
Etiope, G., A. Feyzullayev, and C. Baciu. 2009. Terrestrial methane seeps and mud volcanoes: A global
perspective of gas origin. Marine Petroleum Geology 26(3): 333-344.
European Environment Agency. 2009. EMEP/EEA air pollutant emission inventory guidebook—2009.
EEA Technical Report/2009. Available online at: http://www.eea.europa.eu/publications/emep-eea-
emission-inventory-guidebook-2009.
Ferretti, D.F., J.B. Miller, J.W.C. White, D.M. Etheridge, K.R. Lassey, D.C. Lowe, C.M.M. Meure, M.F.
Dreier, C.M. Trudinger, T.D. van Ommen, and R.L. Langenfelds. 2005. Unexpected changes to the
global methane budget over the past 2000 years. Science 309(5741): 1714-1717.
Frankenberg, C., J.F. Meirink, M. van Weele, U. Platt, and T. Wagner. 2005. Assessing methane
emissions from global space-borne observations. Science 308(5724): 1010-1014.
Harder, S.L., D.T. Shindell, G.A. Schmidt, and E.J. Brook. 2007. A global climate model study of CH4
emissions during the Holocene and Glacial-Interglacial Transitions constrained by ice core data. Global
Biogeochemical Cycles 21(1): Art. No. GB1011.
Hein, R., P.J. Crutzen, and M. Heimann. 1997. An inverse modeling approach to investigate the global
atmospheric methane cycle. Global Biogeochemical Cycles 11(1): 43-76.
Hernandez, P.A., N.M. Perez, J.M. Salazar, S. Nakai, K. Notsu, and H. Wakita. 1998. Diffuse emission of
carbon dioxide, methane, and helium-3 from Teide Volcano, Tenerife, Canary Islands. Geophysical
Research Letters 25(17): 3311-3314.
Hill, T.M., J.P. Kennett, D.L. Valentine, Z. Yang, C.M. Reddy, R.K. Nelson, RJ. Behl, C. Robert, and L.
Beaufort. 2006. Climatically driven emissions of hydrocarbons from marine sediments during
deglaciation. Proceedings of the National Academy of Sciences of the United States of America
103(37): 13570-13574.
Hong, W.L., and T. Yang. 2007. Methane flux from accretionary prism through mud volcano area in
Taiwan—from present to the past. Proceedings of the 9th International Conference on Gas
Geochemistry, October 1-8, 2007. National Taiwan University, pp. 80-81.
Hornafius, J.S., D. Quigley, and B.P. Luyendyk. 1999. The world's most spectacular marine hydrocarbon
seeps (Coal Oil Point, Santa Barbara Channel, California): Quantification of emissions. Journal of
Geophysical Research-Oceans 104(C9): 20703-20711.
Houweling, S., T. Kaminski, F. Dentener, J. Lelieveld, and M. Heimann. 1999. Inverse modeling of
methane sources and sinks using the adjoint of a global transport model. Journal of Geophysical
Research-Atmospheres 104(D21): 26137-26160.
Hovland, M., A.G. Judd, and R.A. Burke. 1993. The global flux of methane from shallow submarine
sediments. Chemosphere 26(1-4): 559-578.
Huseynov, D.A., and I.S. Guliyev. 2004. Mud volcanic natural phenomena in the South Caspian Basin:
Geology, fluid dynamics and environmental impact. Environmental Geology 46: 1012-1023.
Judd, A.G. 2003. The global importance and context of methane escape from the seabed. Geo-Marine
Letters 23(3-4): 147-154.
8-19
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Judd, A.G. 2004. Natural seabed gas seeps as sources of atmospheric methane. Environmental Geology
46(8): 988-996.
Judd, A.G., and M. Hovland. 2007. Seabed Fluid Flow: Impact on Geology, Biology and the Marine
Environment. Cambridge, UK: Cambridge University Press.
Judd, A., G. Davies, J. Wilson, R. Holmes, G. Baron, and I. Bryden. 1997. Contributions to atmospheric
methane by natural seepages on the UK continental shelf (vol 137, pg 165, 1997). Marine Geology
140(3-4): 427-455.
Judd, A.G., M. Hovland, L.I. Dimitrov, S.G. Gil, and V. Jukes. 2002a. The geological methane budget at
continental margins and its influence on climate change. Geofluids 2(2): 109-126.
Judd, A.G., R. Sim, P. Kingston, and J. McNally. 2002b. Gas seepage on an intertidal site: Tony Bay,
Firth of Forth, Scotland. Continental Shelf Research 22(16): 2317-2331.
Kessler, J.D., W.S. Reeburgh, J. Southon, R. Seifert, W. Michaelis, and S.C. Tyler. 2006. Basin-wide
estimates of the input of methane from seeps and clathrates to the Black Sea. Earth and Planetary
Science Letters 243(3-4): 366-375.
Klusman, R., M. Jakel, and M. LeRoy. 1998. Does microseepage of methane and light hydrocarbons
contribute to the atmospheric budget of methane and to global climate change? Assoc. Petrol. Geochem.
Explor.Bull. 11: 1-56.
Klusman, R., M. Leopold, and M. LeRoy. 2000. Seasonal variation in methane fluxes from sedimentary
basins to the atmosphere: Results from chamber measurements and modeling of transport from deep
sources. Journal of Geophysical Research 105D: 24661-24670.
Klusman, R. 2003. Rate measurements and detection of gas microseepage to the atmosphere from an
enhanced oil recovery/sequestration project, Rangely, Colorado, USA. Applied Geochemistry 18(12):
1825-1838.
Kopf, A.J. 2002. Significance of mud volcanism. Reviews of Geophysics 40(2): Art. No. 1005.
Kopf, A.J. 2003. Global methane emission through mud volcanoes and its past and present impact on the
Earth's climate. InternationalJournal of Earth Sciences 92(5): 806-816.
Kopf, A.J. 2005. Global methane emission through mud volcanoes and its past and present impact on the
Earth's climate. InternationalJournal of Earth Sciences 94(3): 493-494.
Kvenvolden, K.A., and B.W. Rogers. 2005. Gaia's breath—Global methane exhalations. Marine and
Petroleum Geology 22(4): 579-590.
Kvenvolden, K., T.D. Loreneson, and W.S. Reeburgh. 2001. Attention turns to naturally occurring
methane seepage. Eos Trans. AGU 82(40): 457-458.
Lacroix, A.V. 1993. Unaccounted-for sources of fossil and isotopically-enriched methane and their
contribution to the emissions inventory: A review and synthesis. Chemosphere 26: 507-557.
Lassey, K.R., B.C. Lowe, and A.M. Smith. 2007a. The atmospheric cycling of radiomethane and the
"fossil fraction" of the methane source. Atmospheric Chemistry and Physics 7(8): 2141-2149.
8-20
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Lassey, K.R., D.M. Etheridge, B.C. Lowe, A.M. Smith, and D.F. Ferretti. 2007b. Centennial evolution of
the atmospheric methane budget: What do the carbon isotopes tell us? Atmospheric Chemistry and
Physics 7(8): 2119-2139.
Leifer, I., and J. Clark. 2002. Modeling trace gases in hydrocarbon seep bubbles. Application to marine
hydrocarbon seeps in the Santa Barbara Channel. Geologiya i Geofizika 43(7): 613-621.
Leifer, I., and A.G. Judd. 2002. Oceanic methane layers: The hydrocarbon seep bubble deposition
hypothesis. Terra Nova 14(6): 417-424.
Leifer, I., J.F. Clark, and R.F. Chen. 2000. Modifications of the local environment by natural marine
hydrocarbon seeps. Geophysical Research Letters 27(22): 3711-3714.
Leifer, I., J.R. Boles, B.P. Luyendyk, and J.F. Clark. 2004. Transient discharges from marine hydrocarbon
seeps: spatial and temporal variability. Environmental Geology 46(8): 1038-1052.
Leifer, I., B.P. Luyendyk, J. Boles, and J.F. Clark. 2006. Natural marine seepage blowout: Contribution to
atmospheric methane. Global Biogeochemical Cycles 20(3).
Luyendyk, B., J. Kennett, and J.F. Clark. 2005. Hypothesis for increased atmospheric methane input from
hydrocarbon seeps on exposed continental shelves during glacial low sea level. Marine and Petroleum
Geology 22(4): 591-596.
MacDonald, I.R., I. Leifer, R. Sassen, P. Stine, R. Mitchell, and N. Guinasso. 2002. Transfer of
hydrocarbons from natural seeps to the water column and atmosphere. Geofluids 2(2): 95-107.
Martinelli, G., and A. Judd. 2004. Mud volcanoes of Italy. GeologicalJournal 39(1): 49-61.
Mau, S., H. Sahling, G. Rehder, E. Suess, P. Linke, and E. Seeding. 2006. Estimates of methane output
from mud extrusions at the erosive convergent margin off Costa Rica. Marine Geology 225(1-4): 129-
144.
Mikaloff Fletcher, S.E., P.P. Tans, L.M. Bruhwiler, J.B. Miller, and M. Heimann. 2004. CH4 sources
estimated from atmospheric observations of CH4 and its C-13/C-12 isotopic ratios: 1. Inverse modeling
of source processes. Global Biogeochemical Cycles 18(4): Art. No. GB4004.
Milkov, A.V. 2000. Worldwide distribution of submarine mud volcanoes and associated gas hydrates.
Marine Geology 167(1-2): 29-42.
Milkov, A.V., G. Etiope. 2005. Global methane emission through mud volcanoes and its past and present
impact on the Earth's climate—A comment. InternationalJournal Of Earth Sciences 94(3): 490-492.
Milkov, A.V., R. Sassen, T.V. Apanasovich, and F.G. Dadashev. 2003. Global gas flux from mud
volcanoes: A significant source of fossil methane in the atmosphere and the ocean. Geophysical
Research Letters 30(2): Art. No. 1037.
Morner, N.A., and G. Etiope. 2002. Carbon degassing from the lithosphere. Global and Planetary
Change 33(1-2): 185-203.
Niemann, H., M. Elvert, M. Hovland, B. Orcutt, A. Judd, I. Suck, J. Gutt, S. Joye, E. Damm, K. Finster,
and A. Boetius. 2005. Methane emission and consumption at a North Sea gas seep (Tommeliten area).
Biogeosciences 2(4): 335-351.
8-21
-------
Chapter 8. Terrestrial and Marine Geologic Sources
Niemann, H., J. Duarte, C. Hensen, E. Omoregie, V. H. Magalhaes, M. Elvert, L. M. Pinheiro, A. Kopf,
and A. Boetius. 2006. Microbial methane turnover at mud volcanoes of the Gulf of Cadiz. Geochimica
et Cosmochimica Acta 70(21): 5336-5355.
Norman, J., C. Kucharik, S. Gower, D. Baldocchi, P. Crill, M. Rayment, K. Savage, and R. Striegl. 1997.
A comparison of six methods for measuring soil-surface carbon dioxide fluxes. Journal of Geophysical
Research 102D: 28771-28777.
Patro, R., I. Leifer, and P. Bowyer. 2001. Better bubble process modeling: Improved bubble
hydrodynamics parameterization. In: M.A. Donelan, W.M. Drennan, E.S. Saltzman, and R. Wanninkof
(eds.). Gas Transfer at Water Surfaces. AGU.
Quigley, B.C., J.S. Hornafius, B.P. Luyendyk, R.D. Francis, J. Clark, and L. Washburn. 1999. Decrease
in natural marine hydrocarbon seepage near Coal Oil Point, California, associated with offshore oil
production. Geology 27(11): 1047-1050.
Rahmstorf, S. 2007. A semi-empirical approach to projecting future sea-level rise. Science 315: 368-370.
Ryan, S., E.J. Dlugokencky, P.P. Tans, and M.E. Trudeau. 2006. Mauna Loa volcano is not a methane
source: Implications for Mars. Geophysical Research Letters 33(12): Art. No. L12301.
Schaefer, H., and M.J. Whiticar. 2008. Potential glacial-interglacial changes in stable carbonisotope ratios
of methane sources and sink fractionation. Global Biogeochemical Cycles 22:
doi: 10.1029/2006GB002889.
Schmale, O., J. Greinert, and G. Rehder. 2005. Methane emission from high-intensity marine gas seeps in
the Black Sea into the atmosphere. Geophysical Research Letters 32(7): L07609.
Simpson, D., W. Winiwarter, G. Borjesson, S. Cinderby, A. Ferreiro, A. Guenther, C.N. Hewitt, R.
Janson, M.A.K. Khalil, S. Owen, T.E. Pierce, H. Puxbaum, M. Shearer, U. Skiba, R. Steinbrecher, L.
Tarrason, and M.G. Oquist. 1999. Inventorying emissions from nature in Europe. Journal of
Geophysical Research-Atmospheres 104(D7): 8113-8152.
Voitov, G.I. 1975. Gas breath of Earth. Nature 3: 91-98 (in Russian).
Wallmann, K., M. Drews, G. Aloisi, and G. Bohrmann. 2006. Methane discharge into the Black Sea and
the global ocean via fluid flow through submarine mud volcanoes. Earth and Planetary Science Letters
248(1-2): 545-560.
Washburn, L., J.F. Clark, and P. Kyriakidis. 2005. The spatial scales, distribution, and intensity of natural
marine hydrocarbon seeps near Coal Oil Point, California. Marine and Petroleum Geology 22(4): 569-
578.
Wilson, R.D., P.H. Monaghan, A. Osanik, L.C. Price, and M.A. Rogers. 1973. Estimate of annual input of
petroleum to the marine environment from natural marine seepage. Transactions, Gulf Coast
Association of Geological Societies. 23rd Annual Convention, Houston, TX.
Wilson, R.D., P.H. Monaghan, A. Osanik, L.C. Price, and M.A. Rogers. 1974. Natural marine oil
seepage. Science 184: 857-865.
Yang, T.F., G.H. Yeh, C.C. Fu, C.C. Wang, T.F., Lan, H.F. Lee, C.H. Chen, V. Walia, and Q.C. Sung.
2004. Composition and exhalation flux of gases from mud volcanoes in Taiwan. Environmental
Geology 46(8): 1003-1011.
8-22
-------
Chapter 9. Wildfires
Wildfires are fires in unmanaged forests, grasslands, savannas, and shrublands (Aalde et al., 2006a).
These fires can be ignited by lightning strikes or started accidentally by humans, but do not include
deliberate controlled burns for land clearing activities. As they burn, wildfires release a number of
greenhouse gases, particulates, and other air pollutants. Incomplete combustion of biomass, consisting of
both living and dead organic matter, is the primary source for emissions of QrU from wildfires. In
contrast to CH4 emissions, N2O emissions from wildfires depend not only on the combustion conditions
but also on the nitrogen content in the biomass being burnt. Wildfires are typically a very small source of
N2O, as they are fairly low-temperature fires (in comparison to fossil fuel combustion), thus limiting the
conversion of atmospheric nitrogen (N2) to fixed nitrogen.
In most studies of the global CH4 or N2O budget, natural sources such as wildfire emissions are either
assumed to remain constant over time or lumped together with a larger biomass burning source, the
majority of which is deliberate or planned burning. Together, these assumptions have limited estimates of
historical changes in wildfire emissions, as well as projections of how they might be altered in the future
under conditions of changing climate.
The previous EPA report (U.S. EPA, 1993) does not discuss wildfires, perhaps because they were not
considered to be of sufficient magnitude or one of the primary natural sources most likely to be sensitive
to climate. However, a number of observational and modeling studies have now shown wildfire frequency
and intensity to be highly dependent on climate. Shifts in climate and weather patterns establish the
conditions necessary for wildfire ignition and spread.
Similarly, in earlier IPCC Assessment Reports (e.g., Ehhalt et al., 2001) a natural wildfire source was not
distinguished from the larger, primarily deliberate or planned anthropogenic biomass burning source. In
the most recent IPCC report (Denman et al., 2007), a table summarizing eight individual estimates of the
CH4 budget shows only two studies that provide values for a natural wildfire source: 5 Tg CtVyr
(Houweling et al., 2000), and 2 Tg CHVyr (Wuebbles and Hayhoe, 2002).
N2O emissions from wildfires were not considered in either of the latest IPCC reports (Ehhalt et al., 2001;
Denman et al., 2007), nor do there appear to be any peer-reviewed estimates of the contribution of
wildfire to the global N2O budget in the scientific literature.
9.1 Description of Emission Source
Although emissions from wildfires may be virtually indistinguishable from controlled burns, an important
distinction must be made between natural (accidental) and anthropogenic (deliberate) fires. According to
the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (Aalde et al., 2006b), "A wildfire is
any unplanned and uncontrolled wetland fire which, regardless of ignition source, may require
suppression response." Fires that are the result of accidental ignition that occur on unmanaged lands are
"natural" (i.e., not included in the anthropogenic inventory), while fires that result from either accidental
or deliberate ignition on managed lands are considered "anthropogenic." Almost 90 percent of all biomass
burning is considered to be deliberately human-initiated. Much of it occurs in the tropics, where savanna
and forest fires are driven by land clearing for agriculture and the need for fuelwood. Prescribed burning
for forest management and agricultural waste burning is also prevalent in temperate-boreal regions.
The remaining 10 percent of fires, classified as "natural wildfires," are attributed to natural causes, such
as lightning, and accidental human ignition. These natural fires occur predominantly in the mid- and high-
latitude temperate-boreal ecosystems (for example, boreal forest fires in Alaska and Canada in summer
2004 and Siberian forest fires in summer 2003; Levine, 1999, and references therein; Lavoue et al., 2000).
Natural fires are also prevalent in arid, heavily populated regions that are dominated by shrublands (for
9-1
-------
Chapter 9. Wildfires
example, the southern California wildfires in October 2007 and the Mediterranean wildfires in August
2007). In this report, EPA estimates emissions from natural lightning-ignited and uncontrolled human-
initiated biomass burning only, referred to from now on as wildfires.
As the general fire and emission characteristics of anthropogenic and natural fires are identical, the
discussion below pertains to both types of biomass burning. However, in this report EPA summarizes
estimates of wildfire emissions only.
Emission of various gases and particles in the smoke from fires depends on the composition of the
biomass burned, and on combustion conditions (Andreae and Merlet, 2001). Biomass typically contains
about 45 percent carbon by weight, with the remainder being hydrogen and oxygen (-55 percent)
(Levine, 2000; Levine et al., 2004; Andreae and Merlet, 2001). Trace amounts of nitrogen and sulfur are
also present depending on the nature of biomass. Under conditions of complete combustion, burning of
organic matter proceeds via oxidation that mainly produces water vapor (H2O) and carbon dioxide (CO2),
according to the following reaction:
(CH2O) + O2 -» CO2 + H2O,
where CH2O represents the average composition of organic matter (Levine et al., 2000; Levine, 2004).
However, combustion is not a single-stage process but proceeds in three stages that determine the ultimate
proportions of the chemical species emitted (Levine et al., 2000; Levine, 2004). The first stage involves
pyrolysis accompanied by the emission of organic volatile compounds in the form of white smoke. In the
second stage, high-temperature flaming combustion occurs that converts reduced substances to simple
oxidized species, including CO2, H2O, and N2O. Other intermediate species such as CO and CFL are also
released at this stage, depending on the chemical and physical interactions in the flame. Finally, after the
flaming combustion stops, low-temperature smoldering begins, which emits large amounts of
incompletely oxidized volatile organic compounds. The majority of non-CO2 chemical species including
CO, CH4, NO*, non-methane hydrocarbons, and various volatile organic compounds and particles are
emitted during this final smoldering stage of combustion. Experimental data indicate a linear relationship
between N2O emissions and the nitrogen content of the biomass being burned (Lobert et al., 1991).
Open vegetation fires are dynamic fires; hence, all three combustion types are present at any given time.
However, their proportions vary overtime, with flaming dominating in the earlier part of the fire and
smoldering during the later part. For forest fires, the flaming phase in given location typically lasts for an
hour or less, while the smoldering phase may last up to a day or longer depending on the fuel type,
moisture content, weather conditions, etc. (Andreae and Merlet, 2001). For savanna grassland and
agricultural waste fires, the flaming phase lasts only a few minutes, while the smoldering phase lasts up to
an hour (Levine et al., 2000; Andreae and Merlet, 2001). Thus, the type of ecosystem burning, the amount
of biomass burned, the phase of combustion, and the evolution of emissions with changing fire conditions
in various ecosystems are all important pieces of information required to determine the contribution of
wildfires emissions to the total global CFL budget. Knowledge of the nitrogen content in the biomass
being burned is also required, in addition to all of these factors, to estimate wildfire N2O emissions.
Factors influencing emissions from wildfires are described in Section 9.2. Estimates of current and future
wildfire CFL. and N2O emissions are described in Sections 9.3 and 9.4, respectively. Finally, areas of
further research are outlined in Section 9.5.
9.2 Factors That Influence Emissions
Wildfire emissions depend on the frequency and strength of wildfires, which are in turn determined by
several factors, including:
• The type of vegetation present in a region.
• The frequency of lightning (not associated with heavy rainfall) that triggers the fires.
9-2
-------
Chapter 9. Wildfires
• The frequency of dry periods or droughts, which would create optimal conditions for wildfire
(this is particularly important for arid regions when drought follows a wet period that has allowed
vegetation to grow).
• The frequency of other weather conditions—such as extreme wind events—that would help the
fires to spread.
• The presence of humans in the region, whose inadvertent activities—from dropping a burning
match to neglecting to extinguish sparks blown from a campfire—can begin a wildfire.
• Other human factors that can limit the area burned, including firefighting, breaks in vegetation
and flammable structures due to human clearing, etc.
Sections 9.2.1 to 9.2.3 discuss how these factors modulate wildfires and their emissions under broad
categories.
9.2.1 Type of Vegetation
The type of vegetation burned determines the dominant combustion process, which in turn controls the
level of CH4 and N2O emitted. Forest fires dominated by smoldering emit more CFLt and less N2O than do
savanna and grassland fires that are mostly flaming. For example, the estimated average amount of CFi4
emitted per area burned of tropical forest fires is three times that of the CFi4 emitted when burning
savanna and grassland (Andreae and Merlet, 2001).
Emissions from types of vegetation burned are estimated in terms of emission factors. Due to the large
difference in combustion processes or fire characteristics from one ecosystem to another, emission factors
are usually estimated specifically for the type of vegetation burned (Delmas, 1994; Andreae and Merlet,
2001). Emission factors are discussed in detail in Section 9.3.1.
9.2.2 Influence of Weather and Climate
Weather and climate strongly affect forest wildfires, specifically the area burned by fires (Swetnam, 1993;
Flannigan and Wotton, 2001). Weather is defined here as short-term processes that result in variations in
atmospheric conditions ranging from minutes to the length of a fire season, while climate refers to
processes that influence the atmosphere over time periods of years to decades, longer than a fire season.
The connection between weather and climate and the area burned by forest fires is discussed below.
9.2.2.1 Weather
Once lightning triggers a fire in a forest with dry flammable biomass, synoptic weather conditions that
produce strong winds, low relative humidities (dry conditions), and above-normal temperatures can cause
a fire to rapidly spread. Forest fires tend to be concentrated in summer months, when these weather
conditions are most prevalent.
Early studies of synoptic-scale weather conditions associated with large wildfires in the eastern half of the
United States indicate that nearly 80 percent of fires are associated with a cold front, either before or after
passage of a dry cold front. Analysis of the relationship between meteorological variables and monthly
area burned by wildfires in Canada suggest that severe fire months were independent of rainfall amount
but significantly dependent on rainfall frequency, temperature, and relative humidity.
Furthermore, since large forest fires are not constrained to the surface but extend many kilometers into the
atmosphere, the atmosphere's vertical structure can also have a significant impact on fire growth and
behavior. A dry and unstable atmosphere enhances the growth of forest fires by promoting erratic fire
behavior and fire spread. Thus, day-to-day weather can dramatically influence fire behavior and area
burned.
9-3
-------
Chapter 9. Wildfires
9.2.2.2 Climate
Variations in large-scale climatic patterns affect the corresponding weather variables and thus also
influence wildfires. Climate primarily interacts with fire through its direct effect on fuel moisture. Larger
fires usually occur during warm, dry years. However, climate also affects the geographic distribution of
vegetation types and site productivity, and thus indirectly influences the intensity, frequency, and size of
fires through the types of fuels that are made available and the rates at which those fuels accumulate
(Miller and Urban, 1999). Numerous studies have shown that wet years that contribute to fuel
accumulation often promote burning in the subsequent dry years (Swetnam and Betancourt, 1990;
Heyerdahl et al., 2002; Westerling et al., 2003).
More recent work has linked spatial and temporal
variability in fire activity to inter-annual and inter-
decadal climate variations caused by variations in
hemispheric and even global-scale atmospheric
circulation patterns. The most well-known source of
climatic fluctuations is the El Nino Southern
Oscillation, or ENSO (see text box).
Several researchers have studied the relationship
between fire occurrence in various parts of the
world and variations in weather patterns associated
with ENSO or ENSO-like atmospheric circulation.
Following are results from recent studies that
highlight the influences of variations in large-scale
climatic patterns on fire regimes:
• Correlation of the 300-year record of fire
activity for the southwestern United States
with an index for the Southern Oscillation
showed that large areas tend to burn after
dry springs associated with La Nina events,
while smaller areas burn after wet springs
associated with El Nino events that increase
vegetation growth (Swetnam and
Betancourt, 1990).
• El Nino years are characterized by dry
winters and warm spring temperatures in
the Pacific Northwest United States.
Heyerdahl et al. (2002) assessed the
influence of annual variations in climate on fire regimes of pine forests in eastern Oregon and
Washington using tree-ring reconstructions of annual total area burned for the years 1687 to
1994. They found that large fires burned during dry years and El Nino years in all watersheds,
while small fires burned regardless of variation in these climate parameters. However, large fires
also burned during relatively wet years and La Nina years in one watershed they examined,
indicating that local factors can override regional climate controls in some locations.
• The direct impact of El Nino was most recently observed in the forested peatlands of Indonesia
during the El Nino event of 1997, which caused widespread fires. The fires were initially started
to clear forests for agriculture, but soon became uncontrolled because of severe drought
conditions in the region owing to El Nino (Siegert et al., 2001).
El Nino Southern Oscillation
The Southern Oscillation (SO) reflects the
monthly or seasonal fluctuations in the air
pressure difference between Tahiti and Darwin.
It occurs at irregular intervals, from two to seven
years. The warm phase of SO, referred to as El
Nino, is characterized by warm sea surface
temperatures in the eastern and central
equatorial Pacific Ocean, high pressure over the
tropical western Pacific, and low pressure on
the southeastern Pacific near the coast of South
America. The opposite phase of SO, known as
La Nina, is characterized by higher surface
pressures in the eastern Pacific and lower sea
surface temperatures along the equator.
El Nino events tend to produce increased
rainfall and widespread flooding along coasts of
northern Peru and Ecuador, and severe drought
conditions in parts of Southeastern Asia and
Northern Australia. In the United States, El Nino
events are associated with warm and wet
winters in upper Midwestern states, the
Northeast, and Canada, while the central and
southern California, northwest Mexico, and the
southwestern United States are wetter and
cooler than normal. La Nina events are typically
associated with weather conditions opposite to
those of El Nino events.
9-4
-------
Chapter 9. Wildfires
Using satellite-derived burn areas,
Baltzer et al. (2005) also suggests a
relationship between fire severity and
., ... , ,,., , .. pressure patterns in northern middle and high
the positive phase of the Arctic |gtjtudes_ Jhe osd||ation js characterized by a
Oscillation (see text box) in Siberia.
Fauria and Johnson (2006) studied
the large lightning fires in Canada
and Alaska during 1959-1999 and pattern is reversed'
found a relationship between their
occurrence and both the Arctic
Oscillation and ENSO on an
interannual scale, with a frequency of
2 to 6 years.
The Arctic Oscillation
"Arctic Oscillation" refers to opposing atmospheric
"negative phase" with relatively high pressure over the
polar region and low pressure at mid-latitudes (about
45 degrees north), and a "positive phase" in which the
The positive phase is characterized by wetter weather
in Alaska, Scotland, and Scandinavia; drier conditions
in the western United States and the Mediterranean;
and warmer weather in Eurasia. Weather patterns in
the negative phase are generally opposite to those of
the positive phase.
While the above analyses illustrate important
relationships between wildfire severity,
frequency, and large-scale climate variations, these studies are constrained by the fire data being
influenced by human interventions as well.
9.2.3 Human Influence
Human activities also influence the frequency and intensity of wildfires, and therefore impact wildfire
emissions as well. Human settlement, transportation, and recreation patterns determine where people live
and where they travel to. This in turn sets the spatial pattern for accidental ignitions. The higher the
human presence in an area already predisposed for fire conditions by climate and weather patterns, the
more likely an accidental ignition is.
Residential development also affects wildfire probability. Human settlement along the urban-forest
perimeter limits the possibility for controlled burns or natural fires to remove highly flammable dead
material and underbrush. Fuel builds up; when a wildfire eventually does occur, more fuel is available and
so the fire often ends up spreading further than it would have if allowed to happen naturally earlier.
Residential areas also affect the path of wildfires. They provide breaks in natural vegetation as well as
additional flammable materials. Here in the United States and in many other industrial nations, the
presence of homes—particularly vacation expensive homes—is also a determining factor whether or not
resources are spent on combating a given fire, or if it is allowed to burn itself out.
Prescribed burning or controlled burning in unmanaged lands is sometimes used for forest management.
Benefits include protecting trees from insects and disease, improving habitat for wildlife species, and
decreasing the risk of future larger, uncontrolled fires. In dry and windy conditions with high flammable
fuel loads, controlled fires can sometimes get out of control, causing a wildfire event.
Given the complexity of these interactions, it is difficult to evaluate and quantify the human influence on
wildfire emissions on a global scale. However, it is important to note that they do exist and are likely
important.
9.3 Current Global Emissions
Wildfire emissions are generally estimated using the relationship first given by Seiler and Crutzen (1980)
and described in detail in more recent studies (Delmas, 1994; Levine, 2000, 2004; Levine et al., 2000;
Liousse et al., 2004). This relationship is also used to calculate emissions from prescribed fires and
wildfires in managed lands (Aalde et al., 2006a). Gas or aerosol emissions (Mx) are calculated as:
Mx = EFX x Mbiomass,
9-5
-------
Chapter 9. Wildfires
where EFX is the emission factor defined as the amount of any species x released per amount of dry matter
consumed in units of g/kg, andMbi0mass is the amount of biomass burned in mass units. Sections 9.2.1 and
9.2.2 discuss the methods to calculate these parameters and the factors that influence them.
9.3.1 Estimating Emission Factors
According to Andreae and Merlet (2001), the emission factor of carbon species from biomass burning can
be expressed as:
EF - M* -M
, ,
I , r^r T7T^ I , \r, 1
J + [CCO \ + [CCH4 \ + [CyoC \
where Mx is the mass of species emitted, Me is the mass of carbon emitted, \C\biomass is the carbon content
in the biomass burned, [x] is the concentration of species in the smoke, and [CcoJ, etc., are the
concentrations of the various carbon species in the smoke. Both the carbon content of the biomass burned
and the carbon budget of the fire are difficult to measure in the field but can be easily determined in the
laboratory. A fuel carbon content of 45 percent is usually assumed when fuel and residue data at the
ground are not available.
The emission ratio relates the emission of a species to the emission of a reference gas, for example, CO2
or CO, and can be used to derive species-specific emission factors. Emission ratios are given by dividing
the excess species concentration in a fire plume by the excess concentration of a simultaneously measured
reference gas. The "excess" concentrations are obtained by subtracting the ambient background
concentrations from the concentrations measured in the burn plume. For example, following Andreae and
Merlet (2001), the emission ratio of QrU relative to CC>2 in molar units is
= =
CHJCO, . ~~ /~~ \ _(rn }
LV^U2 VV ^2 ) smoke ^^2 ) ambient
and the emission ratio of N2O relative to CC>2 is given by
_
2 smoke V 2 ambient
The calculation of emission ratios requires only simultaneous measurements of the species of interest and
the reference gas in the smoke plume and in the background air. Furthermore, knowledge of the fuel
composition and the amounts of fuel burned is not needed, thus making emission ratios suitable for field
studies, particularly airborne plume measurements.
Emission factors, on the other hand, allow the most straightforward calculation of trace gas and aerosol
emissions when the amount of biomass burned is known. Therefore, experimental values of emission
ratios are usually converted to emission factors to determine regional or global-scale fire emissions. The
following equation is used to convert molar emission ratios (Andreae and Merlet, 2001):
EF = ERi YiY \ - EFY
(X'Y) MWY 7'
where ER(X/Y) is the emission ratio of species x relative to the reference species Y, A4Wxar\dA4WY are the
molecular weights of the species x, and 7, and EFy is the emission factor of the reference species. The
emission factor for N2O calculated in this way therefore represents the fraction of burned fuel nitrogen
emitted as N2O, and is independent of the fuel nitrogen content (Delmas et al., 1995), which would
otherwise represent a key uncertainty in the estimate.
9-6
-------
Chapter 9. Wildfires
Emission factors are highly dependent on the fire regime (type of vegetation burned, combustion
conditions). Fires dominated by smoldering, such as forest fires, generally have higher emission factors
for reduced species such as ClrU. In contrast, dominantly flaming fires, such as savanna, grassland, and
shrubland fires, have fairly low emission factors for reduced species and high emission factors for
oxidized species including N2O (contingent on nitrogen content of the biomass). For example, the
estimated average CFI4 emission factor for tropical forest fires is three times that of the CFI4 emission
factor for burning savanna and grassland (Andreae and Merlet, 2001).
For this reason, emission factors are usually estimated specifically for the different types of vegetation
burned (Delmas, 1994; Andreae and Merlet, 2001). Uncertainties in emission ratios (factors), however,
still arise due to lack of sufficient field data, sampling conditions with a tendency to over-represent one
mode of combustion over the other (for example, ground-based sampling tends to over-represent
smoldering emissions that remain closer to the ground while airborne sampling may be biased towards
emission from the flaming phase that are lofted to higher altitudes), and differences in the types of
measurements (tower vs. ground-based vs. aircraft measurements). Furthermore, emission factors vary as
the fire season progresses, due to changing moisture conditions. The seasonality in emission factors is
usually not reported in the literature. Instead, only annual average emission factors are tabulated.
Early estimates of emission factors for CFLt from biomass burning and wildfires were synthesized by
Delmas (1994) from various experimental studies. Extremely limited information exists on the emission
factors for N2O (Lobert et al., 1991; Andreae and Merlet (2001). Andreae and Merlet (2001) updated the
previous compilation of emission factors for CFLt along with emission factors for approximately 100 other
relevant species, including that for N2O, with information from recent large- and small-scale experimental
studies. Emission factors from Andreae and Merlet (2001) have also been recommended for calculating
biomass burning emissions by Aalde et al. (2006a). The most recent estimates of CH4 and N2O emission
factors for various types of wildfires and biomass burning are shown in Table 9. 1 . To calculate global
wildfire emissions, the emission factors for savanna and grassland, tropical forest, and extratropical forest
fires are combined with estimates of the amount of biomass burned by natural fires in each of these
categories. The methodologies to estimate the amount of biomass burned are discussed below.
Table 9-1. Methane and Nitrous Oxide Emission Factors for Different Types of
Wildfires and Biomass Burning (from Andreae and Merlet, 2001)
Type of Biomass Burning
Savanna and grassland
Tropical forest
Extratropical forest
Biofuel burning*
Charcoal making
Charcoal burning*
Agricultural residues*
Emission Factor (g/kg)
Methane
2. 3 ±0.9
6.8 ±2.0
4.7 ±1.9
6.1 ±2.2
10.7
6.2 ±3. 3
2.7
Nitrous Oxide
0.21 ±0.10
0.20
0.26 ±0.07
0.06
0.03
0.20
0.07
* Not relevant for wildfires
9.3.2 Estimating Amount of Biomass Burned
The amount of biomass burned annually (M) is by far the most difficult quantity to determine, and hence
the most uncertain. Annual values are usually obtained by a relationship originally given by Seiler and
9-7
-------
Chapter 9. Wildfires
Crutzen (1980) and applied in several other studies (Delmas, 1994; Levine, 1999, 2004; Levine et al.,
2000; and Liousse et al., 2004):
M = AxB xax /?,
where A is the area burned (m2), B is the average biomass per unit area in a particular ecosystem (kg m"2),
a is the fraction of the average above-ground biomass burned relative to the total average biomass B, and
/? is the burning efficiency of the above-ground
biomass.
GLOBSCAR and GBA-2000
The GLObal Burn SCARs (GLOBSCAR) project
was initiated in 2001 as part of the European
Space Agency (ESA) Data User Programme, for
production of global incremental monthly maps of
burnt areas using daytime data from year 2000 of
the Along Track Scanning Radiometer (ATSR-2)
instrument onboard the ESA ERS-2 satellite. The
final GLOBSCAR products are distributed
through VITO's GEOSUCCESS web server
(http://www.geosuccess.net/).
The Global Burnt Area-2000 initiative (GBA2000)
was launched by the Global Vegetation
Monitoring unit of the European Commission's
Joint Research Centre in partnership with several
other institutions around the world, with the
specific objective of producing a map of the
areas burnt globally for the year 2000, using the
medium resolution (1 km) satellite imagery
provided by the SPOT-Vegetation system, and to
derive statistics of area burnt per type of
vegetation cover. The GBA-2000 product can be
downloaded from http://www.grid.unep.ch/
activities/earlywarning/preview/ims/gba/.
Large uncertainties exist in the four parameters
required to calculate M. Burn area (^4), a
critical parameter, is particularly difficult to
estimate because of the high spatial and
interannual variability in this factor at
continental to global scales (Giglio et al.,
2006).
Early estimates of burn area were based on
statistics from the Food and Agriculture
Organization or national survey data (Hao and
Liu, 1994). Although remote sensing products
(i.e., satellite observations) have been used to
monitor fires since the late 1970s, it is only
recently that remote sensing has been used to
generate maps of annual burned area at the
global and regional scales (Gregoire et al.,
2003; Liousse et al., 2004; Simon et al., 2004;
Giglio et al., 2006). The two most evaluated
and frequently used satellite-derived burn area
products are the GLOB SCAR product
collected by the Along Track Scanning
Radiometer (ATSR) (Simon et al., 2004) and
the GBA-2000 product generated from data
collected by the SPOT-VEGETATION instrument (Gregoire et al., 2003; see text box).
While these efforts demonstrate the feasibility of using satellite products to determine the global area
burned, there are still uncertainties and shortcomings in the satellite-derived burned-area information
(Hoelzemann et al., 2004; Simon et al., 2004). Specifically, two major shortcomings are: (1) the area
estimates do not differentiate between fires in different ecosystems, and (2) the area estimates do not
make a distinction between anthropogenic biomass burning and natural wildfires. The second
shortcoming makes it particularly difficult to assess CFL emissions from wildfires (as highlighted below).
Future efforts by the remote sensing community are directed towards addressing these shortcomings and
further refining the data (van der Werf et al., 2004, 2006).
Climate/weather parameters (discussed in Section 9.2.3) induce significant temporal (season-to-season,
year-to-year) variability in burned areas. Most of the temporal variability occurs in fires in the tropics,
followed by the boreal and temperate forests (Dwyer et al., 2000; van der Werf et al., 2006).
Early studies incorporated seasonal variability in emissions using surrogates such as seasonal rainfall
patterns, cultural practices, vegetation types, and surface ozone for the presence of biomass burning.
For example, Hao and Liu (1994) derived monthly distributions of the amount of biomass burned in the
tropics by assuming that high surface ozone concentrations during the dry season were a result of these
fires. Cooke et al. (1996) investigated the seasonality in biomass burning emissions for Africa using
9-8
-------
Chapter 9. Wildfires
satellite data and compared their results with that of Hao and Liu (1994). Their analysis showed that the
inventory of Hao and Liu (1994) tended to underestimate the intensity of the peak months of burning, and
indicated that seasonal variation in surface ozone concentrations is not a good proxy for seasonality in
emissions on a regional and continental scale.
The first global assessment of the seasonal variability of fire occurrence using satellite data was
performed by Dwyer et al. (2000). Their analysis concluded that the strongest variability in biomass
burning is that associated with the movement of the dry season in the tropics from the northern to the
southern hemisphere. Intense burning in Africa, north of the equator, occurs from October to February,
transitioning to July through September for regions south of the equator. The most intense burning in
Central America and southern Asia occurs in April and May. The burning season runs from May to
August in most regions of the northern hemisphere' Is temperate and boreal biomes, while in the
temperate regions of the southern hemisphere, the burning season is from December to March.
Biomass density or the available fuel load is an equally important parameter required to estimate the
amount of biomass burned. Early estimates of fuel load were based on compilations of ecosystem-specific
biomass load obtained from field measurements (Hao and Liu, 1994; Lavoue et al., 2000). Recently,
vegetation models that simulate the global carbon cycle have been used to derive the biomass density of
vegetation susceptible to fire on a global scale (Hoelzemann et al., 2004; van der Werf et al., 2004). The
accuracy of simulated fuel loads depends on how well the vegetation model simulates the global carbon
cycle. The simulated ecosystem-specific fuel loads are compared with published values to evaluate and
validate the models. However, direct comparison of modeled fuel loads with literature values is difficult
because it is usually unclear what value (total biomass load, aboveground biomass density or the available
fuel load) is being reported in the published literature. Furthermore, the coarse resolution of vegetation
models (typically 0.5° x 0.5° or 1° x 1° grid cell) is unable to capture the spatial heterogeneity in fuel
loads within a grid cell.
Finally, combustion efficiency is estimated based on field observations in different ecosystems (Delmas,
1994). The biomass density and the combustion efficiency are only known to within ±50 percent (Liousse
et al., 2004).
9.3.3 Estimating Current Methane and Nitrous Oxide Emissions
As stated previously, it is extremely difficult to distinguish between burned areas from natural wildfires
and those burned by anthropogenic fires. This makes it difficult to isolate wildfire emissions relative to
the global total from all biomass burning. Estimates of wildfire CFL. emissions are currently based on the
assumption that about 10 percent of the global biomass burning is natural (Levine, 1999), although of
course this proportion is likely to vary year to year. No published estimates of global wildfire N2O
emissions are available. Therefore, EPA has applied the same assumption to provide a "best-guess"
estimate for global wildfire N2O emission. This estimated value is likely to be highly uncertain and
should therefore be used with caution.
Globally, wildfires are estimated to emit between 2 and 5 Tg CFL/yr (Denman et al., 2007; Houweling et
al., 2000; Wuebbles and Hayhoe, 2002). These emissions are not for a specific year but are climatological
averages. The estimate of Houweling et al. (2000) is based on the work of Chappellaz et al. (1993), who
assumed that of the 50 Tg CFL/yr (based on Fung et al., 1991) emitted from biomass burning, 10 percent
is released from wildfires.
Lobert et al. (1999) give estimates of carbon emissions from different types of vegetation feedstock,
including forest wildfires (mainly in temperate and boreal forests) representative of the early 1990s. The
database of the area and amount of biomass burned in each category was compiled from various sources
(Logan and Yevich, unpublished manuscripts, 1998, cited in Lobert et al., 1999). Total global carbon
emissions from forest wildfires are estimated to be 265 Tg C/yr. Applying a factor of 1/0.45 to convert
9-9
-------
Chapter 9. Wildfires
this value to mass of fuel burned and then combining with CFL emission factors for extratropical forests
from Table 9-1 suggests annual global wildfire CH4 emissions of ~3 Tg CFL/yr (Table 9-2). This estimate
coincides with the CH4 emissions from extratropical forests derived by Andreae and Merlet (2001) based
on the same source of biomass burning estimates. Based on the similar methodology using the N2O
emission factors summarized above, EPA also estimates global wildfire N2O emissions to be
approximately 0.1 Tg N/yr, although no reference for this value is available in the peer-reviewed
literature. Given the methodological problems associated with estimating the amount of biomass burned
in wildfires, a valid statistical error analysis of the emission estimates cannot be performed.
Global burned area derived from sophisticated satellite products combined with updated/refined data for
other parameters has been used recently to estimate global fire emissions (Hoelzemann et al., 2004; Ito
and Penner, 2004; van der Werf et al., 2004, 2006). As noted previously, no distinction is made between
natural and anthropogenic emissions in these studies; however, we can again apply the assumption that 10
percent of global biomass burning emissions are from wildfires. Ito and Penner (2004) used the GBA-
2000 data product supplemented by ATSR hot spot data from the World Fire Atlas as inputs into a global
emissions model to calculate global biomass burning emissions for various chemical species including
CH4 (32.2-55.2 Tg CFL/yr) for the year 2000. Wildfire CfL emissions estimated from their study would
therefore range from 3.2 to 5.5 Tg CFL/yr, which is within the range of emissions provided in previous
studies (Table 9-2). Hoelzemann et al. (2004) used the GLOBSCAR burned-area product supplemented
with data from the World Fire Atlas to produce estimates of emissions from global wildland fires for the
year 2000. Their estimate for global biomass burning CFI4 emissions was 12 Tg CFL/yr, resulting in
global wildfire CFI4 emissions of 1.2 Tg CFL/yr. These estimated emissions are significantly lower than
others, a fact attributed primarily to discrepancies in the burned area estimated using GLOBSCAR.
Finally, van der Werf et al. (2006) used a satellite-driven global vegetation model to estimate global fire
emissions for 1997 through 2004. Using their estimate of global CFL emissions of 15 Tg CFL/yr for the
year 2000 yields wildfire emissions of 1.5 Tg CFL/yr. Like the Hoelzemann et al. study, the van der Werf
et al. study produces a CFL emissions estimate that appears to be too low; the reason for this discrepancy
is not clear.
Because the above studies did not provide any estimates for N2O emissions from biomass burning, it is
not possible to apply the same assumptions to estimate wildfire N2O emissions as done for CFL.
Table 9-2. Estimates of Wildfire Methane and Nitrous Oxide Emissions
Reference
Houweling et al. (2000)
Wuebbles and Hayhoe (2002)
Lobertetal. (1999)
Hoelzemann et al. (2004)*
Ito and Penner (2004)*
van der Werf et al. (2006)
Base Year
1990s
2000
2000
2000
Methane Emissions
(Tg CH4/yr)
5.0
2.0
3.0
1.2
3.2-5.5
1.5
Nitrous Oxide
Emissions
(Tg N/yr)
0.1
Derived from these studies as 10 percent of global total biomass burning emissions
Because of the large temporal variability in burned areas, significant variability is introduced in year-to-
year wildfire emissions as well. Duncan et al. (2003) performed the first comprehensive analysis of the
interannual and seasonal variability of fire emissions on the global scale. They applied remotely sensed
data for fire counts and for smoke-related aerosols to estimate both the seasonal and the interannual
9-10
-------
Chapter 9. Wildfires
variation of biomass burning for six regions of the world: Southeast Asia, Indonesia and Malaysia, Brazil,
Central America and Mexico, Canada and Alaska, and Asiatic Russia. They found significant interannual
variability in carbon monoxide emissions from biomass burning, especially in Indonesia and Malaysia,
Brazil, Southeast Asia, and the boreal regions. The largest CO emissions resulted from uncontrolled forest
wildfires during the last two decades and were often associated with ENSO-induced droughts. Their
conclusions can also be applied to CFL, as both CFU and CO are emitted under similar combustion
conditions, particularly during the smoldering phase of the fire (see Section 9.1).
Significant interannual variability in fire CFL emissions was also demonstrated by van der Werf et al.
(2006), who investigated the temporal variability in fire emissions during the eight-year period from 1997
to 2004 using satellite data coupled with a biogeochemical model. Their analysis showed that burning in
forests, as opposed to savanna ecosystems, was highly variable from year to year and that this variability
combined with high fuel loads contributed to high variability in observed emissions.
Given the large spread in CFU emission values and the significant uncertainties in their estimates, EPA
estimates a present-day source to be bounded by the estimates of 2 and 5 Tg CFL/yr as summarized by
Denman et al. (2007).
9.4 Future Emission Scenarios
As discussed in Section 9.2.3, wildfires and climate/weather are intimately linked. Hence, any changes in
climate conditions are likely to influence the fire regime (Swetnam, 1993; Westerling et al., 2006). In
particular, warmer, drier, and windier conditions resulting from future climate changes are predicted to
increase the frequency and severity of wildfires, particularly temperate-boreal fires (Price and Rind, 1994;
Torn and Fried, 1992; Pinol et al., 1998; Brown et al., 2004; Fried et al., 2004, 2008; Levine, 2004;
Lenihan et al., 2006; Miller and Schlegel, 2006; Westerling and Bryant, 2006; Pitman et al., 2007), in
many locations. This would result in greater release of trace gases including CFL and possibly N2O as
well, depending on the nitrogen content of the vegetation.
Studies of the potential impact of climate change on wildfire relate fire characteristics, such as fire
severity, frequency, and burn areas, with climate variables simulated for different climate scenarios by
general circulation models (GCMs). For example, Torn and Fried (1992) examined the impact of a
doubling of atmospheric CO2 on wildfire area burned and the frequencies of escaped fires in northern
California using climate variables from three GCMs combined with daily weather records and
mechanistic models of fire behavior and fire suppression. They found increased burn areas and fire
frequencies in grasslands of northern California in response to a doubled CO2 climate. The magnitude of
those increases, however, strongly depended on the vegetation type, choice of the GCM scenario, and
choice of climate forcing variables. Recent research applying either an advanced version of the modeling
system of Torn and Fried (Fried et al., 2004) or other sophisticated models using GCM projections of
climate variables (Torn et al., 1998; Lenihan, 2006; Miller and Schlegel, 2006; Westerling and Bryant,
2006; Fried et al., 2008) all suggests increases in fire frequency and intensity in various parts of
California with warmer temperatures as well as drier and windier conditions.
Using a fire model combined with the Goddard Institute for Space Studies GCM, Price and Rind (1994)
show a possible 44 percent increase in wildfires ignited by lightning in the continental United States for
an equilibrium doubling of carbon dioxide, with a 78 percent increase in the area burned by these fires.
Their study also found that in the tropics, where most fires are human-initiated today, future climate
change could potentially result in large increases in tropical wildfires. This was recently observed in the
case of extensive and widespread tropical forest and peat fires that swept throughout Kalimantan and
Sumatra, Indonesia, between August and December 1997. The fires initially resulted from burning from
land clearing and land-use change. However, severe drought conditions following El Nino caused the
initially controlled, anthropogenic fires to become large uncontrolled wildfires (Levine, 2004, and
9-11
-------
Chapter 9. Wildfires
references therein). About 2.5 Tg ClrU was emitted from the 1997 Indonesian fires alone, which is
equivalent to the estimated total global wildfire emissions (Table 9-2).
Stocks et al. (1998) investigated the fire danger levels in Russian and Canadian boreal forests under
equilibrium-doubled CC>2 climate using climate outputs from four GCMs. Results from all four models
indicated large increases in the areal extent of extreme fire danger in both the countries under doubled
CC>2 climate. Applying two different approaches, Beer and Williams (1995) and Pitman et al. (2007)
concluded that fire risk over Australia is primarily driven by warming and reductions in relative humidity
(dry conditions), thus making the region very vulnerable to fires under future climate changes.
In summary, future climate change is likely to increase the frequency of weather conditions associated
with high wildfire risks in many regions of the world. Climate change has the potential to affect multiple
elements of wildfires, including fire behavior, ignition, fire management and vegetation fuels. The
complex interactions between each of these factors will determine future spatial and temporal distribution
of wildfires and their emissions in response to climate change.
Currently, no scenarios for future ClrU emissions from global wildfires exist in the literature. Efforts are,
however, directed toward developing models that can predict/forecast wildfire events and can therefore be
used to estimate emissions. Two types of models are currently being applied to predict wildfires.
• Statistical modeling approach: This approach uses statistical correlations to relate wildfire
activity to various fire-danger predictors. These predictors may be climate factors (for example,
seasonal temperature, precipitation, or relative humidity) or indices of large-scale climate
variations (for example, the Southern Oscillation Index). For example, Preisler and Westerling
(2007) have developed a statistical model to estimate the probability of a large fire event given a
list of fire danger predictors. Using historical fire occurrence and fire weather predictors, the
authors demonstrate the utility of this statistical tool for estimating one-month-ahead forecasts for
large wildfire events in the western United States. Additionally, most of the studies described
above apply statistical approaches to relate changes in climate variables to fire activity. While
useful in forecasting/predicting fire danger for a region, these models cannot be used directly to
predict gas or particle emissions from wildfires, either on a regional or a global scale.
• Fire module within a dynamic vegetation model: This approach involves including
representation of fire dynamics within a global dynamic vegetation model that can be run using
climate outputs from a GCM (offline) or coupled to a GCM (online). Fire dynamics are included
in the vegetation models using simple and robust parameterizations of relevant processes (fuel
availability, susceptibility to fire, and an ignition source). Fire-vegetation models are capable of
simulating the inter-related effects of climate change on vegetation dynamics, biomass,
hydrology, fire frequency and emissions on regional to global scales (Lenihan et al., 1998, 2003,
2006; Bachelet et al., 2001; Thonicke et al., 2001; Mouillot et al., 2002; Arora and Boer, 2005).
These models provide a consistent framework for the simulation of changes in the amount of
biomass burned and, therefore, emissions in response to current and future climates.
9.5 Areas for Further Research
With the advent of the satellite era, significant progress has been made—particularly over the last
decade—toward understanding wildfire emissions. Current understanding of CFLj emissions from
wildfires is still limited, while that for N2O is still primitive. Progress, however, can be made with further
research.
Below are highlighted the areas that need to be addressed in order to significantly improve estimates of
current and future CFLt emissions from wildfires.
9-12
-------
Chapter 9. Wildfires
• The area and amount of biomass burned are the most important parameters required to accurately
estimate wildfire CFL, and N2O emissions. However, these remain poorly constrained. A further
complication is the difficulty in distinguishing between biomass subjected to natural, uncontrolled
fires and human-initiated fires for land clearing. Further work on this topic is therefore a high
priority.
• Emission factors are in general uncertain because of many experimental, instrumental, and
methodological factors discussed in Section 9.3.1. However, emission factors forN2O are at best
"best guesses" at this point. Field and laboratory studies are needed to better constrain emission
factors for both CH4 and N2O.
• Burned area estimates from different satellite sensors using different approaches have recently
become available. These estimates demonstrate the strong potential of satellite-derived
information that can be used to derive fire emissions. However, there are significant differences
between estimates of burn areas derived from different approaches as evident from the studies of
Hoelzemann et al. (2004), Ito and Penner (2004), and van der Werf et al. (2004) for the year 2000.
These differences highlight the need for continued validation and intercomparison of different
burned area products.
• Uncertainties in biomass density or available fuel load also induce uncertainties in emission
estimates. Traditionally, compilations of ecosystem-specific fuel loads from different field
experiments were used to derive global-scale emissions. More recently, global vegetation models
have been used to simulate carbon/nitrogen pools within the vegetation on a global scale.
However, large differences exist between simulated and observed fuel loads, possibly due to
mismatch in scale between the measurements at the plot level and the much coarser model grid, as
well as lack of data and biases in the literature values. This highlights the need for more
observations on a global scale to validate the model results.
• Human activities, such as population growth and infrastructure development, are likely to
influence wildfires. For example, human settlements near forested areas could increase
anthropogenic ignition, resulting in high fire frequencies and high emissions. The impact of
human activities on wildfires needs to be assessed in conjunction with the influence of climate
change to better constrain the response of natural fire regimes to future global changes.
9.6 References
Aalde, H., P. Gonzalez, M. Gytarsky, T. Krug, W.A. Kurz, R.D. Lasco, D.L. Martino, E.G. McConkey,
S. Ogle, K. Paustian, J. Raison, N.H. Ravindranath, D. Schoene, P. Smith, Z. Somogyi, A. van Amstel,
and L. Verchot. 2006a. Generic methodologies applicable to multiple land-use categories. In: H.S.
Eggleston, L. Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.). 2006IPCC Guidelines for National
Greenhouse Gas Inventories. Volume 4: Agriculture, Forestry and Other Land Use. Prepared by the
National Greenhouse Gas Inventories Programme, IGES, Japan.
Aalde, H., P. Gonzalez, M. Gytarsky, T. Krug, W.A. Kurz, S. Ogle, J. Raison, D. Schoene, N.H.
; ; ; j j ; o* ? O ? ? ?
Ravindranath, N.G. Elhassan, L.S. Heath, N. Higuchi, S. Kainja, M. Matsumoto, M.J.S. Sanchez, and
Z. Somogyi. 2006b. Forest land. In: H.S. Eggleston H.S., L. Buendia, K. Miwa, T. Ngara, and K.
Tanabe (eds.). 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Volume 4: Agriculture,
Forestry and Other Land Use. Prepared by the National Greenhouse Gas Inventories Programme, IGES,
Japan.
Andreae, M.O., and P. Merlet. 2001. Emission of trace gases and aerosols from biomass burning. Global
Biogeochemical Cycles 15(4): 955-966.
9-13
-------
Chapter 9. Wildfires
Arora, V., and G.J. Boer. 2005. Fire as an interactive component of dynamic vegetation models. Journal
of Geophysical Research 110(G02008): doi: 10.1029/2005JG000042.
Bachelet, D., R.P. Neilson, J.M. Lenihan, and R.J. Drapek. 2001. Climate change effects on vegetation
distribution and carbon budget in the United States. Ecosystems 4: 164-185.
Baltzer, H., F.F. Gerard, C.T. George, C.S. Rowland, T.E. Jupp, I. McCallum, A. Shvidenko, S. Nilsson,
A. Sukhinin, A. Onuchin, and C. Schmullius. 2005. Impact of the Arctic oscillation pattern on
interannual forest fire variability in Central Siberia. Geophysical Research Letters 32(L4709):
doi:1029/2005GL022526.
Beer, T., and A. Williams. 1995. Estimating Australian forest fire danger under conditions of doubled
carbon dioxide concentrations. Climatic Change 29:169-188.
Brown, T.J., B.L. Hall, and A.L. Westerling. 2004. The impact of twenty-first century climate change on
wildland fire danger in the western United States: An applications perspective. Climatic Change 62 (1-
3): 365-388.
Chappellaz, J., I.Y. Fung, A.M. Thompson. 1993. The atmospheric CH4 increase since the last glacial
maximum I: source estimates. Tellus 45B: 228-241.
Cooke, W.F., B. Koffi, and J.-M. Gregoire. 1996. Seasonality of vegetation fires in Africa from remote
sensing data and application to a global chemistry model. Journal of Geophysical Research 101(D15):
21051-21065.
Delmas, R. 1994. An overview of present knowledge on methane emission from biomass burning.
Fertilizer Research 37: 181-190.
Delmas, R., J.P. Lacaux, and D. Brocard. 1995. Determination of biomass burning emission factors:
methods and results. Environmental Monitoring and Assessment 38: 181-204.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Dias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press, pp. 499-587.
Duncan, B.N., R.V. Martin, A.C. Staudt, R. Yevich, and J. A. Logan. 2003. Interannual and seasonal
variability of biomass burning emissions constrained by satellite observations. Journal of Geophysical
Research 108(D2): doi:10.1029/2002JD002378.
Dwyer, E., J.M.C. Pereira, J.-M. Gregoire, C.C. DaCamara. 2000. Characterization of the spatio-temporal
patterns of global fire activity using satellite imagery for the period April 1992 to March 1993. Journal
ofBiogeography 27: 57-69.
Ehhalt, D., M. Prather, F. Dentener, R. Derwent, E. Dlugokencky, E. Holland, I. Isaksen, J. Katima, V.
Kirchhoff, P. Matson, P. Midgley, and M. Wang. 2001. Atmospheric chemistry and greenhouse gases.
In: J.T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, K. Maskell, and C.A.
Johnson. Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge
University Press, pp. 239-287.
Fauria, M.M., and E.A. Johnson. 2006. Large-scale climatic patterns control large lightning fire
occurrence in Canada and Alaska forest regions. Journal of Geophysical Research 111(G04008):
doi: 10.1029/2006 JG000181.
9-14
-------
Chapter 9. Wildfires
Flannigan, M.D., and B.M. Wotton. 2001. Climate, weather, and area burned. In: E.A. Johnson and L.
Miyanishi (eds.). Forest Fires, Behavior and Ecological Effects. Academic Press, pp. 351-373.
Fried, J.S., M.S. Torn, and E. Mills. 2004. The impact of climate change on wildfire severity: A regional
forecast for northern California. Climatic Change 64: 169-191.
Fried, J.S., J.K. Gilless, W.J. Riley, T.J. Moody, C.S. de Bias, K. Hayhoe, M. Moritz, S. Stephens, and M.
Torn. 2008. Predicting the effect of climate change on wildfire severity and outcomes in California:
preliminary analysis. Climatic Change 87: 251.
Fung, I., J. John, J. Lerner, E. Matthews, M. Prather, L. P. Steele, and P. J. Fraser. 1991. Three-
dimensional model synthesis of the global methane cycle. Journal of Geophysical Research 96(D7):
13033-13065.
Giglio, L., G.R. van der Werf, J.T. Randerson, G.J. Collatz, and P. Kasibhatla. 2006. Global estimation of
burned area using MODIS active fire observations. Atmospheric Chemistry and Physics 6: 957-974.
Gregoire, J.-M., K. Tansey, and J.M.N. Silva. 2003. The GBA-2000 initiative: Developing a global
burned area database from SPOT-VEGETATION imagery. InternationalJournal of Remote Sensing
24: 1369-1376.
Hao, W.M., and M.-H. Liu. 1994. Spatial and temporal distribution of tropical biomass burning. Global
Biogeochemical Cycles 8(4): 495-503.
Heyerdahl, E.K., L.B. Brubaker, and J.K. Agee. 2002. Annual and decadal climate forcing of historical
fire regimes in the interior Pacific Northwest, USA. The Holocene 12: 597-604.
Hoelzemann, J.J., M.G. Schultz, G.P. Brasseur, C. Granier, and M. Simon. 2004. Global wildland fire
emission model (GWEM): Evaluating the use of global area burned satellite data. Journal of
Geophysical Research 109(D9): doi:10.1029/2003JD003666.
Houweling, S., F. Dentener, and J. Lelieveld. 2000. Simulation of preindustrial atmospheric methane to
constrain the global source strength of natural wetlands. Journal of Geophysical Research 105(D13):
17243-17255.
Ito, A., and J. Penner. 2004. Global estimates of biomass burning emissions based on satellite imagery for
the year 2000. Journal of Geophysical Research 109(D14S05): doi: 10.1029/2003 JD004423.
Lavoue, D., C. Liousse, H. Cachier, B. J. Stocks, and J. G. Goldammer. 2000. Modeling of carbonaceous
particles emitted by boreal and temperate wildfires at northern latitudes. Journal of Geophysical
Research 105(D22): 26871-26890.
Lenihan, J. M., C. Daly, D. Bachelet, and R. P. Neilson. 1998. Simulating broad-scale fire severity in a
dynamic global vegetation model. Northwest Science 72: 91-103.
Lenihan, J. M., R. Drapek, D. Bachelet, and R. P. Neilson. 2003. Climate change effects on vegetation
distribution, carbon, and fire in California. Ecological Applications 13(6): 1667-1681.
Lenihan, J. M., D. Bachelet, R. Drapek, and R. P. Neilson. 2006. The response of vegetation distribution,
ecosystem productivity, and fire in California to future climate scenarios simulated by the MC1
dynamic vegetation model. In: A Report From the California Climate Change Center. CEC-500-2005-
191-SF: 1-19.
Levine, J.S. 1999. The 1997 fires in Kalimantan and Sumatra, Indonesia: Gaseous and particulate
emissions. Geophysical Research Letters 26(7): 815-818.
Levine, J.S. 2000. Gaseous and particulate emissions released to the atmosphere from vegetation fires. In:
K.-T. Goh, D. Schwela, J.G. Goldammer, and O. Simpson (eds.). Health Guidelines for Vegetation Fire
9-15
-------
Chapter 9. Wildfires
Events: Background Papers. United Nations Environmental Program, World Health Organization,
World Meteorological Organization, and Institute for Environmental Epidemiology, pp. 284-298.
Levine, J.S. 2004. Biomass burning: The cycling of gases and particulates from the biosphere to the
atmosphere. In: R.F. Keeling (ed.). The Atmosphere, Treatise on Geochemistry Volume 4. Amsterdam:
Elsevier Pergamon. pp. 143-158.
Levine, J.S., W.R. Cofer III, and J.P. Pinto. 2000. Biomass burning. In: M.A.K. Khalil (ed.). Atmospheric
Methane: Its Role in the Global Environment. Berlin: Springer-Verlag. pp. 190-201.
Liousse, C., M.O. Andreae, P. Artaxo, P. Barbosa, H. Cachier, J. M. Gregoire, P. Hobbs, D. Lavoue, F.
Mouillot, J. Penner, M. Scholes, and M. G. Schultz. 2004. Deriving global quantitative estimates for
spatial and temporal distributions of biomass burning emissions. In: C. Granier, P. Artaxo, and C.
Reeves. Emissions of Atmospheric Trace Gases. Netherlands: Kluwer Academic Publishers, pp. 71-
113.
Lobert, J.M., D.H. Scharffe, W.-M. Hao, T.A. Kuhllbusch, R Seuwen, P. Warneck, and P. J. Crutzen.
1991. Experimental evaluation of biomass burning emissions: Nitrogen and carbon containing
compounds. In: J.S. Levine (ed.). Global Biomass Burning: Atmospheric, Climatic, andBiospheric
Implications. Cambridge, Massachusetts: MIT Press, pp. 289-304.
Lobert, J.M., W.C. Keene, J.A. Logan, and R. Yevich. 1999. Global chlorine emissions from biomass
burning: Reactive chlorine emissions inventory. Journal of Geophysical Research 104(D7): 8373-8389.
Miller, N.L., and N.J. Schlegel. 2006. Climate change projected fire weather activity: California Santa
Ana wind occurrence. Geophysical Research Letters 33(L15711): doi:10.1029/2006GL025808.
Miller, C., and D.L. Urban. 1999. A model of surface fire, climate and forest pattern in the Sierra Nevada,
California. Ecological Modelling 114: 113-135.
Mouillot, F., S. Rambal, and R. Joffre. 2002. Simulating climate change impacts on fire frequency and
vegetation dynamics in a Mediterranean-type ecosystem. Global Change Biology 8: 423-437.
Pinol, J., J. Terradas, and F. Lloret. 1998. Climate warming, wildfire hazard, and wildfire occurrence in
coastal eastern Spain. Climatic Change 38(3): 345-357.
Pitman, A. J., G.T. Narisma, and J. McAneney. 2007. The impact of climate change in the risk of forest
and grassland fires in Australia. Climatic Change 84: 383-401.
Preisler, H.K., and A.L. Westerling. 2007. Statistical model for forecasting monthly large wildfire events
in western United States. Journal of Applied Meteorology 46: 1020-1030
Price, C., and D. Rind. 1994. The impact of 2xCO2 climate on lightning-caused fires. Journal of Climate
7: 1484-1494.
Schultz, M.G. 2002. On the use of ATSR fire count data to estimate the seasonal and interannual
variability of vegetation fire emissions. Atmospheric Chemistry and Physics 2:387-395.
Seiler, W., and P.J. Crutzen. 1980. Estimates of gross and net fluxes of carbon between the biosphere and
the atmosphere from biomass burning. Climatic Change 2: 207-247.
Siegert, F., G. Riicker, A. Hinrichs, and A. Hoffmann. 2001. Increased fire impacts in logged over forests
during El Nino driven fires. Nature 414: 437-440.
Simon, M., S. Plummer, F. Fierens, J.J. Hoelzemann, and O. Arino. 2004. Burned area detection at global
scale using ATSR-2: The GLOBSCAR products and their qualification. Journal of Geophysical
Research 109(D14S02): doi:10.1029/2003JD003622.
9-16
-------
Chapter 9. Wildfires
Stocks, B.J., M.A. Fosberg, T.J. Lynham, L. Mearns, B.M. Wotton, Q. Yang, J.-Z. Lin, K. Lawrence,
G.R. Hartley, J.A. Mason, and D.W. McKenney. 1998. Climate change and forest fire potential in
Russian and Canadian Boreal Forests. Climatic Change 38: 1-13.
Swetnam, T.W., and J.L. Betancourt. 1990. Fire-southern oscillations in the southwestern United States.
Science 249: 1017-1020.
Swetnam, T.W. 1993. Fire history and climate change in giant sequoia groves. Science 262: 885-889.
Thonicke, K., S. Venevsky, S. Sitch, and W. Cramer. 2001. The role of fire disturbance for global
vegetation dynamics: Coupling fire into a dynamic global vegetation model. Global Ecology and
Biogeography 10: 661-677.
Torn, M., and J.S. Fried. 1992. Predicting the impacts of global warming on wildland fires. Climatic
Change 21: 257-27'4.
Torn, M., E. Mills, and J. Fried. 1998. Will climate change spark more wildfire damages? Contingencies:
Journal of the American Academy of Actuaries July/August issue: 34-43.
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
van der Werf, G.R., J.T. Randerson, G.J. Collatz, L. Giglio, P.S. Kasibhatla, A.F. Arellano Jr., S.C.
Olsen, and E.S. Kasischke. 2004. Continental-scale partitioning of fire emissions during the 1997 to
2001 El Nino/La Nina Period. Science 303: 73-76.
van der Werf, G.R., J.T. Randerson, L. Giglio, G.J. Collatz, P.S. Kasibhatla, and A.F. Arellano Jr. 2006.
Interannual variability in global biomass burning emissions from 1997 to 2004. Atmospheric Chemistry
andPhysics 6: 3423-3441.
Westerling, A.L., A. Gershunov, T.J. Brown, D.R. Cayan, and M.D. Dettinger. 2003. Climate and
wildfire in the western United States. Bulletin of the American Meteorological Society 84: 595-604.
Westerling, A.L. and B. Bryant. 2006. Climate change and wildfire in and around California: Fire
modeling and loss modeling. In: A report From the California Climate Change Center, pp. 1-28.
Westerling, A.L., H.G. Hidalgo, D.R. Cayan, and T.W. Swetnam. 2006. Warming and earlier spring
increase western U.S. forest wildfire activity. Science 313: 940-943.
Wuebbles, D.J., and K. Hayhoe. 2002. Atmospheric methane and global change. Earth-Science Reviews
57: 177-210.
World Fire Atlas, http://dup.esrin.esa.it/ionia/wfa/index.asp
9-17
-------
Chapter 9. Wildfires
9-18
-------
Chapter 10. Vegetation
This chapter addresses recent evidence of aerobic CH4 emissions from plants and vegetation types that
have not generally been included in emissions inventories to date. Unflooded ecosystems, such as upland
tropical forest, tropical savannas, and well-drained boreal forests, are not generally considered potential
sources of CH4, because drier soils act as an oxidative sink for CH4. Neither the previous EPA report
(U.S. EPA, 1993) nor the most recent IPCC assessment (Denman et al, 2007) includes estimates of CFU
emissions from well-drained ecosystems in its emissions inventory. However, recent findings suggest that
such ecosystems may be a significant unrecognized source of CH4. Frankenberg et al. (2005) first
reported that satellite-based estimates of CFU concentrations were higher over tropical forests than would
be expected from current inventories of tropical CFU sources. The following year, Keppler et al. (2006)
reported direct measurements of CFU emissions from living plants and plant litter under aerobic
conditions, and proposed that forests and other vegetation might be a significant overlooked source of
CFU to the atmosphere. Together, these findings have prompted recent efforts to explain the high CFLi
concentrations over tropical forests (Frankenberg et al., 2005, 2008; Bergamaschi et al., 2007; Meirink et
al., 2008; Schneising et al., 2009), verify the measurements of plant emissions (Dueck et al., 2007; Fedele
et al., 2007; Wang et al., 2008; Beerling et al., 2008; Kirschbaum et al., 2008; Bruggemann et al., 2009),
identify a mechanism for aerobic plant emissions (Keppler et al., 2008; Vigano et al., 2008; McLeod et
al., 2008; Messenger et al., 2009; Nisbet et al., 2009; Wang et al., 2009), refine the estimates of global
plant emissions (Houweling et al., 2006; Kirschbaum et al., 2006; Parsons et al., 2006; Butenhoff and
Khalil, 2007; Ferretti et al., 2007), and reexamine the potential for ecosystem CFi4 fluxes in forests and
other dryland vegetation (Carmo et al., 2006; Crutzen et al., 2006; Sanhueza and Donoso, 2006;
Sanhueza, 2007; Sinha et al., 2007).
10.1 Description of Emission Source
Plants have long been recognized as important conduits for CFi4 emissions, transporting CFi4 from
anaerobic soils and sediments to the atmosphere (Dacey and Klug, 1979). However, it is only recently
that plants themselves have been considered a possible source of CFU production (Keppler et al., 2006).
Most existing studies of CFU flux in unflooded ecosystems focus on the balance between CFU production
by methanogenic bacteria and CFU consumption by CFLpOxidizing bacteria in soils. Because oxidative
consumption exceeds CFU production in well-drained soils, unflooded ecosystems are generally
inventoried as CFLt sinks, rather than sources.
Based on their large land area and the strong oxidative capacity of their soils, unflooded tropical forests,
tropical savannas, tropical steppes, and boreal forests are among the largest estimated soil sinks for
atmospheric CFU (Potter et al., 1996). However, recent studies suggest that these unflooded ecosystems
may emit CFLt at the ecosystem level (Carmo et al., 2006; Crutzen et al., 2006; Sanhueza and Donoso,
2006; Sanhueza, 2007; Sinha et al., 2007), and that plants themselves may emit CFU through an as-yet-
unidentified aerobic process (Keppler et al., 2006; Wang et al., 2008; Bruggemann et al., 2009). The
following sections describe evidence of CH4 emissions from well-drained ecosystems from three sources:
indications that bottom-up inventories are underestimating CFU emissions over tropical regions (Hein et
al., 1997; Houweling et al., 1999; Frankenberg et al., 2005, 2006; Bergamaschi et al., 2007);
measurements of direct CFLt emissions from plants under aerobic conditions (Keppler et al., 2006; Wang
et al., 2008, 2009; Bruggemann et al., 2009; but see Dueck et al., 2007; Beerling et al., 2008; Kirschbaum
et al., 2008; Nisbet et al., 2009); and direct measurements of ecosystem CFU flux (Carmo et al., 2006;
Crutzen et al., 2006; Sanhueza and Donoso, 2006; Sanhueza, 2007; Sinha et al., 2007).
10-1
-------
Chapter 10. Vegetation
10.1.1 Identification of High Methane Concentrations Over Tropical Forests
Frankenberg et al. (2005) first compared global patterns of CH4 concentrations inferred from satellite-
based SCIAMACHY observations to values obtained using an atmospheric transport model and current
inventories of CFL sources, to show that SCIAMACHY retrievals over tropical forests were up to 4
percent higher than modeled values. This led them to suggest a missing source of 30 Tg CFL over the
four-month measurement period (August through November 2003). This finding supported earlier inverse
modeling studies based on ground-based measurements, which had indicated that source inventories
tended to underestimate CFL emissions in the tropics (Hein et al., 1997; Houweling et al., 1999).
Later reports (Frankenberg et al., 2006; Bergamaschi et al., 2007) showed that SCIAMACFIY retrievals
over tropical regions were higher than modeled values throughout the measurement period January 2003
through December 2004, and exhibited strong seasonality, with the greatest discrepancies occurring
during August through December. While SCIAMACFfY retrievals are prone to some bias (Frankenberg et
al., 2006), the high CFL, concentrations inferred for the Amazon basin are supported by recent aircraft-
based measurements (Miller et al., 2007).
Using a new wetland inventory in the transport model redistributed wetland emissions from higher
latitudes to the tropics, reducing the discrepancy between SCIAMACFfY observations and modeled
emissions over the Amazon, particularly during the first half of the year (Bergamaschi et al., 2007).
Subsequent analyses have further reduced the discrepancy, particularly the correction of an error that
amplified methane concentration estimates in areas with high water vapor abundances (Frankenberg et al.,
2008). However, the SCIAMACHY retrievals continue to indicate higher tropical emissions than
previously estimated from ground-based observations (Frankenberg et al., 2008; Meirink et al., 2008;
Schneising et al., 2009).
10.1.2 Measurement of Methane Emissions From Plants Under Aerobic
Conditions
Keppler et al. (2006) first reported CFL, emissions of 0
per hour from detached leaves from 19 different
species and 12 to 370 ng per g (dry weight) per hour
from intact plants representing nine herbaceous
species measured under aerobic conditions in the
dark. Under natural sunlight, emissions were 3 to 5
times higher, ranging from 1.6 to 15.8 ng per g (dry
weight) per hour for detached leaves and 198 to 598
ng per g (dry weight) per hour for intact plants.
Because the amount of CFL generated was small
relative to background CFL^ concentrations,
measurements were conducted in static chambers
that had been purged with CFLrfree air. Global
emissions were then estimated by scaling measured
emission rates to annual net primary productivity
(NPP), a measure of the total plant matter produced
over the course of the year, and taking into account
season length and day length for each biome (Table
10-1).
This approach yielded a global emissions estimate
of 62 to 236 Tg CFL/yr from living plants and 0.5 to
6.6 Tg CFL/yr from plant litter, which would be
equivalent to approximately 10 to 40 percent of the
2 to 3 nanograms per gram (ng per g) (dry weight)
Do plants produce methane?
Since Keppler et al. (2006) first proposed that
plants produce methane through a previously
unrecognized mechanism, a number of studies
have attempted to measure direct CH4
emissions from living plants under aerobic
conditions. The results have been mixed, with
some studies confirming aerobic plant
emissions (McLeod et al., 2008; Vigano et al.,
2008; Wang et al., 2008; Bruggemann et al.,
2009; Messenger et al., 2009) and others
finding no evidence of plant emissions (Dueck
et al., 2007; Beerling et al., 2008; Kirschbaum et
al., 2008; Nisbetetal., 2009).
The mechanism for aerobic CH4 production has
not been identified, but recent studies link
methane emissions to exposure to ultraviolet
light (Vigano et al., 2008; McLeod et al., 2008),
which may trigger chemical reactions that
produce methane from antioxidants commonly
found in the mitochondria of living cells (Ghyczy
et al., 2008; Messenger et al., 2009).
10-2
-------
Chapter 10. Vegetation
annual total emissions of 582 Tg QVyr estimated by Denman et al. (2007), with tropical forests
accounting for about half of the estimated source strength.
A number of studies have attempted to verify Keppler et al.'s (2006) measurements of direct CH4
emissions from living plants (Dueck et al., 2007; Beerling et al., 2008; Kirschbaum et al., 2008; Vigano et
al., 2008; Wang et al., 2008, 2009; Bruggemann et al., 2009; Nisbet et al., 2009) and plant litter (Fedele et
al., 2007), with mixed results. Dueck et al. (2007) grew six plant species in air containing 13C-labeled CO2
(99 atom percent 13C) for nine weeks, and then measured CH4 emissions from the plants using a high
sensitivity laser-based system. Because any CIL, emitted from the plants would be labeled with 13C, it
would be detectable even against background atmospheric CIL, concentrations. However, emission rates
measured from individual plants and from numerous plants grown together were not statistically different
from zero. The authors therefore suggested that the CIL, emissions measured by Keppler et al. (2006) may
have simply resulted from diffusion of CIL, out of air spaces in the plants and soil after the measurement
chambers had been flushed with ClrLrfree air. Beerling et al. (2008) also failed to detect methane
emissions from leaves of rice or corn plants incubated in flow-though chambers, either in the dark or
when exposed to photosynthetically active radiation. Nisbet et al. (2009) demonstrated that plants can
emit dissolved methane in the transpiration stream, and that small amounts of methane can be produced as
a byproduct of the breakdown of plant material under high stress conditions, but showed through a
genome analysis that plants do not contain genes linked to the known biological mechanisms for methane
formation.
In response, Vigano et al. (2008) and McLeod et al. (2008) conducted a series of follow-up experiments,
using both flow-through chambers and static chambers containing both ambient and QrLrfree air, which
confirmed CIL, emissions from plant tissues and structural compounds exposed to UV light and heating.
Methane emissions have also been detected by other groups, including Wang et al. (2008), whose
examination of 44 species from the Inner Mongolia steppe found that the leaves of seven out of nine
woody shrub species produced CH4 emissions as high as 3.39 ng CHVgdry weight/hr, although
herbaceous species did not emit measurable levels of plant-derived CH4. Low levels of methane emission
have also been detected in poplar shoot incubated under low-light conditions (Bruggemann et al., 2009).
While the precise mechanism for aerobic CtL, production by plants has not been identified, the
relationship between observed methane emissions and exposure to UV radiation and heat has led to the
suggestion that methane production may be linked to the production of reactive oxygen species in
response to physiological stress. Aerobic CH4 production had previously been demonstrated in animal
cells and mitochondria exposed to oxidative stress (Ghyczy et al., 2003, 2008). During oxidative stress,
overproduction of reactive oxygen species such as hydrogen peroxide (H2O2) and the hydroxyl radical
(OH") can damage vital cell components. Cells produce a number of antioxidant molecules, which protect
against cell damage by reacting with and neutralizing reactive oxygen species. Ghyczy et al. (2003, 2008)
have demonstrated that a group of antioxidant molecules containing methyl (-CH3) groups can produce
CH4 upon reaction with reactive oxygen species in mammalian cells. The isotopic signature of the Clr^
released in their original study led Keppler et al. (2006) to propose plant pectins as the source of aerobic
plant emissions. A later deuterium labeling study suggested that CH4 was generated from methyoxl
groups on the pectin molecules (Keppler et al., 2008). Messenger et al. (2009) recently demonstrated that
reactive oxygen species caused the release of methane from plant pectin, and proposed that reactive
oxygen species, particularly the -OH radical, causes the release of methane from methoxyl groups in the
pectin of plants exposed to UV radiation and other physiological stresses.
10.1.3 Measurement of Ecosystem Methane Flux
Because CIL, oxidation generally exceeds production in the well-drained soils of upland tropical forests
(Potter et al., 1996), tropical savannas (Castaldi et al., 2006), and boreal forests (Whalen et al., 1992),
such ecosystems are generally considered sinks for atmospheric CH4. However, recent studies suggest
10-3
-------
Chapter 10. Vegetation
these ecosystems may be significant sources of CH4. Static chamber measurements indicate that soil
disturbance from selective logging can cause upland forest soils to switch from acting as a CH4 sink to
producing QrU at rates as high as 98 to 531 mg CH4/m2/day (Keller et al., 2005). Measurements of
nighttime canopy concentration profiles yielded emission estimates of 2 to 21 mg QVmVday for four
upland forest sites in the Amazon basin (Carmo et al., 2006). Similarly, canopy profile measurements
indicated emissions of 8.7 mg CHVmVday for a tropical forest site in Suriname and 8.3 mg CHVm2/day
for a boreal forest site in Finland (Sinha et al., 2007). Reexamination of previously published
measurements from Venezuelan savanna sites indicates that the savanna ecosystem is an intermittent
source of CH4, with positive fluxes as high as 3.8 mg CHVm2/day (Crutzen et al., 2006; Sanhueza, 2007).
Using static chamber measurements, Sanhueza and Donoso (2006) reported higher CUt emissions from
Venezuelan savanna plots with undisturbed vegetation compared to plots from which live and dead grass
biomass had been clipped to just above the soil surface. While they interpret the difference as plant-
generated CFLt emissions, differences in CFLt flux might also result from differences in soil moisture, soil
carbon stocks, and plant-mediated CH4 transport between disturbed and undisturbed plots.
10.2 Factors That Influence Emissions
Because a possible plant CFi4 source has only recently been suggested and the mechanism for such
emissions has yet to be identified, little is known about factors that might control rates of CFi4 emission
from plants. Keppler et al. (2006) reported that emission rates increased with temperatures ranging from
30 to 70 degrees Celsius, and were significantly higher in the sunlight than in the dark. The observation
that CFi4 emissions continued to increase at such high temperatures suggests that the CFi4 was not
produced by enzymatic processes (Kirschbaum et al., 2006). The higher emission rates measured when
the incubation chambers were exposed to direct sunlight could indicate that CFi4 production is tied to
photosynthetic processes, or flux rate could have simply responded to heating of the incubation chambers
by sunlight (Kirschbaum et al., 2006). The relationship between CFLt emissions and UV exposure
suggests the mechanism may involve direct photochemical reactions (Vigano et al., 2008) or biochemical
reactions that occur under oxidative stress (Ghyczy et al., 2008; Messenger et al., 2009). As described
below, more research is needed to characterize the controls over CFi4 emissions from plants and variation
across species and ecosystems.
10.3 Current Global Emissions
Keppler et al. (2006) estimated that plants emit 62 to 236 Tg CHVyr, with tropical forests contributing
33.2 to 123 Tg CFiVyr, or about half of the total. These estimates were developed by scaling the measured
emission rates by the annual NPP of each biome, taking into account day and season length (Keppler et
al., 2006). The use of NPP as a scaling factor has been criticized for two reasons. First, since NPP
measures the accumulation of biomass over the length of the growing season, standing biomass is less
than NPP for much of the growing season (Butenhoff and Khalil, 2007; Parsons et al., 2006). Second,
NPP is partitioned between aboveground and belowground biomass, and these would likely differ in
production (Kirschbaum et al., 2006; Butenhoff and
Khalil, 2007). As a result, the findings of Keppler et al.
(2006) have prompted a number of attempts to better
constrain the estimate of global aerobic emissions from
plants through both top-down and bottom-up ^ other photosynthesizing organisms) per
f , . * ,• + , onn/; v uu + i umt time- NPP refers to a rate process, i.e.,
techniques (Houwelmg et al, 2006; Kirschbaum et al the gmount Qf bjomass duced (pet
2006; Parsons et al., 2006; Butenhoff and Khalil, 2007; prjmary production) per day week or year
Ferretti et al., 2007).
Specifically, Bousquet et al. (2006) found that adding a
plant source of 150 Tg CFiVyr to the emissions
Net primary productivity (NPP) is defined
as the net flux of carbon from the
atmosphere into primary producers (plants
Because annual NPP describes the biomass
accumulated over the course of the year, the
total standing biomass at a given point in
time will usually be less than annual NPP.
10-4
-------
Chapter 10. Vegetation
inventory used in inverse modeling simulations could be accommodated by reducing the strength of other
sources, such as wetlands and anthropogenic emissions, within the range of their uncertainties. Houweling
et al. (2006) found that including vegetation emissions of 125 Tg CHVyr in forward modeling simulations
explained up to 50 percent of the discrepancy between previous simulations and SCIAMACHY retrievals
over the Amazon basin (Frankenberg et al., 2005). However, including vegetation emissions of that
magnitude caused the model to overestimate pre -industrial CFLj concentrations, leading Houweling et al.
(2006) to estimate an upper limit of 85 Tg CHVyr for vegetation emissions.
Keppler et al. (2006) reported that the CFU emitted from plants was enriched in 13C relative to wetland
emissions, with mean 513C values of -52%o (per mil) for C3 plants and -46.5%o for C4 plants. The ratio of
13C, a stable isotope of carbon, to the more common 12C in CH4 can be related to different sources, which
have different characteristic isotopic signatures. Stable isotope ratios are expressed as the ratio of 13C to
12C in a sample, relative to that of a standard reference material, in parts per thousand. This suggested that
plant emissions could explain unexpectedly enriched atmospheric 513CH4 values obtained from ice cores
(Ferretti et al., 2005). However, Ferretti et al. (2007) found that including a large (34 to 121 Tg CHVyr)
pre-industrial aerobic plant source in mass balance calculations required using unrealistically low biomass
burning emissions estimates to accommodate variation in atmospheric 513CH4 over the period from 0 to
1700 A.D. (Ferretti et al., 2005). Instead, Ferretti et al. (2007) report "best estimates" of 0 to 46 Tg
CHVyr for pre-industrial plant emissions and 0 to 176 Tg CHVyr for modern plan emissions. Similarly,
Schaefer et al. (2006) proposed that plant emissions could help explain unexpectedly enriched 513CH4
values at the end of the last glacial period (about 12,000 years ago), when temperatures and atmospheric
CH4 concentrations increased rapidly. More recently, Schaefer and Whiticar (2008) were able to
accommodate a modern aerobic plant source of 42 Tg CHVyr with a 513CH4 value of -51%o in a detailed
budget based on changes in CH4 concentration and 513CH4 between the last glacial maximum and the
present.
Table 10-1. Bottom-Up Estimates of Plant Methane Emissions by Biome
Biome
Tropical forest
Temperate forest
Boreal forest
Mediterranean shrublands
Tropical savanna and
grasslands
Temperate grasslands
Deserts
Crops
Total
Keppler et al., 2006
(Tg CH4/yr)a'b
78.2(33.2-123)
17.7(7.1-28.4)
3(1.1-4.1)
2.7(1.2-4.3)
29.2(12.4-45.9)
7.4(2.9-11.8)
3.8(1.7-5.9)
7.2(2.9-11.5)
149 (62.3-236)
Kirschbaum et al., 2006
(Tg CH4/yr)c
18.8
3.4
2.8
1.0
6.6
1.6
0.6
1.6
36.4(15.1-60.3)
Parsons et al., 2006
(Tg CH4/yr)b
15.6
8.0
3.6
0.8
8.0
2.0
2.2
2.2
42
a Units are teragrams (Tg) of CH4 emitted per year; a Tg is equivalent to a megaton (Mt), or 1 million metric tons.
b Estimated by scaling mean (low to high) emissions measurements by biome NPP.
c Estimated by scaling mean emissions measurements from Keppler et al., 2006, by estimated standing leaf biomass.
In contrast to these "top-down" estimates, a number of studies have used "bottom-up" approaches to
extrapolate the Keppler et al. (2006) flux measurements to the global scale (Table 10-1). Kirschbaum et
al. (2006) used two approaches to estimate global plant emissions: the first based on estimates of standing
leaf mass in each biome, and the second based on estimates of photosynthetic productivity for each
10-5
-------
Chapter 10. Vegetation
biome. The leaf-mass approach yielded an estimate of 15 to 60 Tg CHVyr, with tropical forests
contributing about half the total, and the photosynthesis-based approach yielded an estimate of 9.6 Tg
CHVyr, with tropical forests contributing about a third of the total plant emissions. Parsons et al. (2006)
also used an approach that scaled CH4 emissions based on standing leaf biomass in each biome, yielding
an estimate of 42 Tg CHVyr from leafy biomass and an additional 10.7 Tg CHVyr from non-leafy
biomass, with tropical forests accounting for about half of the total plant source.
Butenhoff and Khalil (2007) used two methods to scale plant emissions: one based on estimates of
standing leaf biomass derived from satellite maps of leaf area index (LAI) and a second based on monthly
estimates of above-ground net primary production (ANPP). Using Keppler et al.'s (2006) mid-range
values of 374 ng CIL per gram (dry weight) per hour for sunlit tissue and 1 19 ng CH4 per gram (dry
weight) per hour for shaded tissue, Butenhoff and Khalil (2007) calculated global plant emissions of 36
Tg CHVyr for the LAI method and 20 Tg CHVyr for the ANPP method, with tropical forests contributing
about a third of the total emissions.
Given the limited number of studies that have attempted to measure QrU emissions from plants under
aerobic conditions, the conflicting results of those studies, and the fact that an actual physical mechanism
by which plants directly produce ClrU has yet to be identified, a best estimate of global plant emissions
must include the possibility of zero emissions (i.e., that plants are in fact not a direct source of CIL).
If an aerobic plant source is confirmed, additional work will still be required to identify the mechanism of
CH4 production and the environmental and biological controls over emissions rates before plant emissions
can be included in process-based models in a meaningful way. Based on the information currently
available, in this report EPA estimates a plant CIL source of either zero (if plants do not actually emit
CIL) or (if the source is confirmed) 20 to 60 Tg CHVyr, a range that captures the agreement between
bottom-up estimates. This range is on the low end for current top-down estimates, which generally
worked backward from the very large plant source proposed by Keppler et al. (2006) by reducing the
estimated contribution from other sources, particularly wetlands.
10.4 Future Emission Scenarios
The recently proposed aerobic plant QrU source has not yet been incorporated into simulations of future
CH4 emissions. However, future plant emissions would likely depend on changes in the distribution of
different vegetation types, as well as changes in environmental factors that might control emission rates.
Current estimates attribute 35 to 50 percent of global plant emissions to tropical forests, with the second
largest source, tropical savanna and grasslands, contributing about 20 percent. This suggests that future
plant emissions would depend largely on changes in climate and land use in the tropics.
10.5 Areas for Further Research
To date, few published studies have successfully measured direct CIL. emissions from plants under
aerobic conditions (Keppler et al., 2006; Vigano et al., 2008; Wang et al., 2008, 2009; Bruggemann et al.,
2009), while others have been unable to detect measurable emissions (Dueck et al., 2007; Beerling et al.,
2008; Kirschbaum et al., 2008; Nisbet et al., 2009). These studies' conflicting results indicate that the first
priorities for future research should be to verify the findings of Keppler et al. (2006) and identify a
mechanism for plant ClrU emissions. If an aerobic plant CH4 source is confirmed, experimental work is
needed to determine the environmental and physiological controls over emission rates. Measurements
should also be taken for a wider range of species to characterize the variation in plant CIL production at
both the species and ecosystem levels. Until more is understood about the seasonal patterns and controls
over emission rates, it will be difficult to accurately quantify the global significance of a possible plant
CH4 source.
10-6
-------
Chapter 10. Vegetation
10.6 References
Beerling, D.J., T. Gardiner, G. Leggett, A. McLeod, and W.P. Quick. 2008. Missing methane emissions
from leaves of terrestrial plants. Global Change Biology 14: 1821-1826.
Bergamaschi, P., C. Frankenberg, J.F. Meirink, M. Krol, F. Dentener, T. Wagner, U. Platt, J.O. Kaplan, S.
Korner, M. Hermann, E.J. Dlugokencky, and A. Goede. 2007. Satellite chartography of atmospheric
methane from SCIAMACFiY onboard ENVISAT: 2. Evaluation based on inverse model simulations.
Journal of Geophysical Research 112: D02304.
Bousquet, P., P. Ciais, J.B. Miller, E.J. Dlugokencky, D.A. Hauglustaine, C. Prigent, G.R. Van der Werf,
P. Peylin, E.-G. Brunke, C. Carouge, R.L. Langenfelds, J. Lathiere, F. Papa, M. Ramonet, M. Schmidt,
L.P. Steele, S.C. Tyler, and J. White. 2006. Contribution of anthropogenic and natural sources to
atmospheric methane variability. Nature 443: 439-443.
Bruggemann, N., R. Meier, D. Steigner, I. Zimmer, S. Louis, and J.P. Schnitzler. 2009. Nonmicrobial
aerobic methane emission from poplar shoot cultures under low-light conditions. New Phytologist 182:
912-918.
Butenhoff, C.K., and M.A.K. Khalil. 2007. Global methane emissions from terrestrial plants.
Environmental Science and Technology 41: 4032-4037.
Carmo, J.B., M. Keller, J.D. Bias, P.B. de Camargo, and P. Crill. 2006. A source of methane from upland
forests in the Brazilian Amazon. Geophysical Research Letters 33: L04809.
Castaldi, S., A. Ermice, and S. Strumia. 2006. Fluxes of N2O and CFL, from soils of savannas and
seasonally-dry ecosystems. Ecological Biogeography 33: 401-415.
Crutzen, P.J., E. Sanhueza, and C.A.M. Brenninkmeijer. 2006. Methane production from mixed tropical
savanna and forest vegetation in Venezuela. Atmospheric Chemistry and Physics Discussions 6: 3093-
3097.
Dacey, J.W.H., and M.J. Klug. 1979. Methane efflux from lake-sediments through water lilies. Science
203: 1253-1255.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Bias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press, pp. 499-587.
Dueck, T.A., R. de Visser, H. Poorter, S. Persijn, A. Gorissen, W. de Visser, A. Schapendonk, J.
Verstappen, H. Bouwmeester, L.A.C.J. Voesenek, and A. van der Werf. 2007. No evidence for
substantial aerobic methane emission by terrestrial plants: a 13C-labelling approach. New Phytologist
175: 29-35.
Fedele, R., I.E. Galbally, N. Porter, and LA. Weeks. 2007. Biogenic VOC emissions from fresh leaf
mulch and wood chips ofGrevillea robusta (Australian Silky Oak). Atmospheric Environment 41:
8736-8746.
Ferretti, D.F., J.B. Miller, J.W.C. White, D.M. Etheridge, K.R Lassey, D.C. Lowe, C.M. MacFarling
Meure, M.F. Dreier, C.M. Trudinger, T.D. van Ommen, R.L. Langenfelds. 2005. Unexpected changes
to the global methane budget over the past 2000 years. Science 309: 1714-1717.
10-7
-------
Chapter 10. Vegetation
Ferretti, D.F., J.B. Miller, J.W.C. White, K.R. Lassey, D.C. Lowe, and D.M. Etheridge. 2007. Stable
isotopes provide revised global limits of aerobic methane emissions from plants. Atmospheric
Chemistry and Physics 7:237-241.
Frankenberg, C., P. Bergamaschi, A. Butz, S. Houweling, J.F. Meirink, J. Notholt, A.K. Petersen, H.
Schrijver, T. Warneke, and I. Aben. 2008. Tropical methane emissions: A revised view from
SCIAMACHY onboard ENVISAT. Geophysical Research Letters 35: L15811.
Frankenberg, C., J.-F. Meirink, M. van Weele, U. Platt, and T. Wagner. 2005. Assessing methane
emissions from global space-borne observations. Science 308: 1010-1014.
Frankenberg, D., J.F. Meirink, P. Bergamaschi, A.P.H. Goede, M. Hermann, S. Korner, U. Platt, M. van
Weele, and T. Wagner. 2006. Satellite chartography of atmospheric methane from SCIAMACFiY on
borad ENVISAT: Analysis of the years 2003 and 2004. Journal of Geophysical Research 111: D07303.
Ghyczy, M., C. Torday, and M. Boros. 2003. Simultaneous generation of methane, carbon dioxide, and
carbon monoxide from choline and ascorbic acid—a defensive mechanism against reductive stress? The
FASEB Journal 17: 1124-1126.
Ghyczy, M., C. Torday, J. Kaszaki, A. Szabo, M. Czobel, and M. Boros. 2008. Hypoxia-induced
generation of methane in mitochondria and eukaryotic cells—an alternative approach to
methanogenesis. Cellular Physiology and Biochemistry 21: 251-258.
Hein, R., P.J. Crutzen, M. Heimann. 1997. An inverse modeling approach to investigate the global
atmospheric methane cycle. Global Biogeochemical Cycles 11: 43-76.
Houweling, S., T. Kaminski, F. Dentener, J. Lelieveld, and M. Heimann. 1999. Inverse modeling of
methane sources and sinks using the adjoint of a global transport model. Journal of Geophysical
Research 104: 26137-26160.
Houweling, S., T. Rockmann, I. Aben, F. Keppler, M. Krol, J.F. Meirink, E.J. Dlugokencky, and C.
Frankenberg. 2006. Atmospheric constraints on global emissions of methane from plants. Geophysical
Research Letters 33: LI5821.
Keller, M., R. Varner, J.D. Dias, H. Silva, P. Crill, R.C. de Oliveira, and G.P. Asner. 2005. Soil-
atmosphere exchange of nitrous oxide, nitric oxide, methane, and carbon dioxide in logged and
undisturbed forest in the Tapajos National Forest, Brazil. Earth Interactions 9: 1-28.
Keppler, F., J.T.G. Hamilton, W.C. McRoberts, I. Vigano, M. Brass, and T. Rockmann. 2008. Methoxyl
groups of plant pectin as a precursor of atmospheric methane: evidence from deuterium labeling
studies. New Phytologist 178: 808-814.
Keppler, F., J.T.G. Hamilton, M. BraB, and T. Rockmann. 2006. Methane emissions from terrestrial
plants under aerobic conditions. Nature 449: 187-191.
Kirschbaum, M.U.F., D. Bruhn, D.M. Etheridge, J.R. Evans, G.D. Farquhar, RM. Gifford, K.I. Paul, and
A.J. Winters. 2006. A comment on the quantitative significance of aerobic methane release by plants.
Functional Plant Biology 33:521-530.
Kirschbaum, M.U.F., and A. Walcroft. 2008. No detectable aerobic methane efflux from plant material,
nor from adsorption/desorption processes. Biogeosciences 5: 1551-1558.
McLeod, A.R, S.C. Fry, G.J. Loake, D.J. Messenger, D.S. Reay, K.A. Smith, and B.W. Yun. 2008.
Ultraviolet radiation drives methane emissions from terrestrial plant pectins. New Phytologist 180: 124-
132.
Meirink, J.F., P. Bergamaschi, C. Frankenberg, M.T.S. d'Amelio, E.J. Dlugokencky, L.V. Gatti, S.
Houweling, J.B. Miller, T. Rockmann, M.G. Villani, and M.C. Kroll. 2008. Four-dimensional
variational data assimilation for inverse modeling of atmospheric methane emissions: Analysis of
10-8
-------
Chapter 10. Vegetation
SCIAMACHY observations. Journal of Geophysical Research—Atmospheres 113: D17301.
Messenger, D.J., A.R. McLeod, and S.C. Fry. 2009. The role of ultraviolet radiation, photosynsitizers,
reactive oxygen species and ester groups in mechanisms of methane formation from pectin. Plant Cell
and Environment 32: 1-9.
Miller, J.B., L.V. Gatti, M.T.S. d'Amelio, A.M. Crotwell, E.J. Dlugokencky, P. Bakwin, P. Artaxo, and
P.P. Tans. 2007. Airborne measurements indicate large methane emissions from the eastern Amazon
basin. Geophysical Research Letters 34: L10809, doi:10.1029/2006GL029213.
Nisbet, R.E.R., R. Fisher, RH. Nimmo, D.S. Bendall, P.M. Crill, A.V. Gallego-Sala, E.R.C. Hornibrook,
E. Lopez-Juez, D. Lowry, P.B.R. Nisbet, E.F. Shuckburgh, S. Sriskantharajah, C.J. Jowe, and E.G.
Nisbet. 2009. Emission of methane from plants. Proceedings of the Royal Society B—Biological
Sciences 276: 1347-1354.
Parsons, A.J., P.C.D. Newton, H. Clark, and P.M. Kelliher. 2006. Scaling methane emissions from
vegetation. Trends in Ecology and Evolution 21(8): 423-424.
Potter, C.S., E.A. Davidson, L.V. Verchot. 1996. Estimation of global biogeochemical controls and
seasonality in soil methane consumption. Chemosphere 32: 2219-2246.
Sanhueza, E. 2007. Methane soil-vegetation-atmosphere fluxes in tropical ecosystems. Interciencia 32:
30-34.
Sanhueza, E., and L. Donoso. 2006. Methane emissions from tropical savanna Trachypogon sp. grasses.
Atmospheric Chemistry and Physics 6: 5315-5319.
Schaefer, H., and M.J. Whiticar. 2008. Potential glacial-interglacial changes in stable carbon isotope
ratios of methane sources and sink fractionation. Global Biogeochemical Cycles 22: GB1001, doi:
10.1029/2006GB002889.
Schaefer, H., M.J. Whiticar, E.J. Brook, V.V. Petrenko, D.F. Ferretti, and J.P. Severinghaus. 2006. Ice
Record of 513C for atmospheric CFL, across the Younger Dryas-Preboreal transition. Science 313: 1109-
1112.
Schneising, O., M. Buchwitz, J.P. Burrows, H. Bovensmann, P. Bergamaschi, and W. Peters. 2009. Three
years of greenhouse gas column-averaged dry air mole fractions retrieved from satellite—Part 2:
methane. Atmospheric Chemistry and Physics 9: 443-465.
Sinha, V., J. Williams, and J. Levieveld. 2007. Methane emissions from boreal and tropical forest
ecosystems derived from in-situ measurements. Atmospheric Chemistry and Physics Discussions 7:
14011-14039.
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Vigano, I., H. van Weelden, R. Holzinger, F. Keppler, A. McLeod, and T. Rockmann. 2008. Effect of UV
radiation and temperature on the emission of methane from plant biomass and structural components.
Biogeosciences 5: 937-947.
Wang, Z.P., J. Gulledge, J.Q. Zheng, W. Liu, L.H. Li, and X.G. Han. 2009. Physical injury stimulates
aerobic methane emissions from terrestrial plants. Biogeosciences 6: 615-621.
Wang, Z.P., X.G. Han, G.G. Wang, Y. Song, and J. Gulledge. 2008. Aerobic methane emission from
plants in the Inner Mongolia Steppe. Environmental Science and Technology 42(1): 62-68.
Whalen, S.C., W.S. Reeburgh, and V.A. Barber. 1992. Oxidation of methane in boreal forest soils: A
comparison of seven measures. Biogeochemistry 16: 181-211.
10-9
-------
Chapter 10. Vegetation
10-10
-------
Chapter 11. Terrestrial Arthropods and Wild Animals
The original 1993 report (U.S. EPA, 1993) cited termites as a contributor to natural QrU emissions,
estimating their global contribution at 20 Tg CHVyr (10 to 50 Tg CHVyr). However, the report did not
discuss this source, and it considered no other type
of terrestrial arthropod nor any contributions by Arthropods
wild animals. . .. . .. , . . , ,
Arthropods are the largest phylum of animals
and include insects, arachnids, crustaceans,
and others. Arthropods are characterized by
the possession of a segmented body with
appendages on at least one segment. All
. .. . , . , ., . ^TT arthropods are covered by a hard exoskeleton.
termite species, and suggested that Cft may Arthropods are common throughout marine,
In the years since the publication of the 1993
report, additional investigation of ClrU emissions
from termites has resulted in more refined
estimates of QrU emissions from the various
freshwater, terrestrial, and aerial
environments.
oxidize in termite mounds prior to atmospheric
release. In addition, other terrestrial arthropods
have been studied to assess whether they generate
CH4 and should be included in any estimates of
global emissions. From this new research, termites and other terrestrial arthropods continue to be a small
but not insignificant contributor to global QrU emissions.
Wild animals also contribute to global CFLj emissions, although far less research has been conducted on
this source. Much is known about enteric fermentation from domesticated animals, and from this area of
study (combined with estimates of wild animal populations), some conclusions can be drawn about the
relative contributions from ruminant wild animals.
This chapter reviews the current scientific understanding of terrestrial arthropods and wild animals as
sources of CFLt, including the factors that influence emissions and the most recent estimates of current
and future global emissions.
11.1 Description of Emission Source
Termites and other terrestrial arthropods produce and emit ClrUas a result of microbial degradation of the
organic matter they ingest. In wild animals, CFLj emissions are caused by enteric fermentation in
ruminants such as bison, deer, elk, mountain goats, and sheep, and also in some smaller rodent species.
CFLj production from termites was first observed in the 1930s. As research progressed, it became clear
that the amount of CFLt generated varied substantially from species to species. Termites can be divided
into two groups: (1) lower termites, which live in a mutualistic relationship with one or more protozoan
flagellates that live inside their guts and digest the wood ingested by the termites, and (2) higher termites,
which possess anaerobic (methanogenic) bacteria in their guts (Sanderson, 1996). Higher termites have
been found to have higher ClrU emission rates and larger biomasses than lower termites, thereby
contributing more to the global QrU emissions from this source (Sugimoto and Inoue, 1998).
Certain termites build nests or mounds in which to live. Some groups simply excavate areas in dead wood
or make underground nests. The more advanced termites build huge mounds, largely from soil excavated
from their underground chambers and cemented with saliva. These are especially common in Africa and
Australia. Some nests are open and emit all CH4 generated by the termites; others are closed (such as
subterranean nests) and have a rate of emission close to zero. Some nests have ventilation vents.
CH4 emitted by the termites in their nests may be partially oxidized by the mound material during
emission into the atmosphere. Research conducted by Sugimoto and Inoue (1998) evaluated the emission
of CH4 from various types of termite mounds and developed a set of emission factors defined as CH4
emitted from the mound divided by ClrU produced by the termites. These factors were used in
11-1
-------
Chapter 11. Terrestrial Arthropods and Wild Animals
combination with estimates of termite populations and biomass to estimate global emissions. For those
colonies with mounds, over half (and in some cases up to more than 80 percent) of CH4 emissions were
oxidized by the mound.
Research by Hackstein and Stumm (1994) evaluated whether CFLus emitted from other types of
arthropods, in addition to termites (Isopterd). This research showed that three other groups exhibited
production: Diplopoda (millipedes), Blattaria (cockroaches), and Cetonidae (flower beetles). Although
research is still preliminary, it suggests that these other species, particularly flower beetles, may be
significant sources
In animals, CFLj is produced as part of normal digestive processes. During digestion, microbes resident in
the digestive system ferment food consumed by the animal. This microbial fermentation process, referred
to as enteric fermentation, produces CH4as a byproduct, which can be exhaled or eructated by the animal.
The amount of QrU produced and excreted by an individual animal depends primarily upon the animal's
digestive system, and the amount and type of feed it consumes.
Ruminant animals are the major emitters of ClrU because of their unique digestive system. Ruminants
possess a rumen, or large "fore -stomach," in which microbial fermentation breaks down the feed they
consume into products that can be absorbed and metabolized. The microbial fermentation that occurs in
the rumen enables them to digest coarse plant material that non-ruminant animals cannot. Ruminant
animals, consequently, have the highest QrU emissions among all animal types.
11.2 Factors That Influence Emissions
The factors that determine the magnitude of emissions of QrU from terrestrial arthropods include the
species of arthropod, including the specific type of termite. The highest rates of QrU are produced by
arthropods with methanogenic bacteria, typically found in higher termites. Environmental conditions such
as temperature and humidity also affect the rate of CFLj generation, as does the population density, termite
activity, time of day, and size and type of termite mounds. However, even though different species
produce widely different amounts of CFLj, the overall total QrU emission remains largely constant. In
addition, according to a study by Martius et al. (1996), rainforest clearing and the conversion of primary
forest to pasture land did not significantly change the amount of CFLj emitted by termites.
Emissions of QrUfrom wild animals are influenced by feed quality and feed intake. Because a ruminant
digests food in two steps, its food intake is affected by the quality, or digestibility, of the food. The
higher-quality food passes through the rumen more quickly and leads to higher intake. However, the
lower-quality food tends to generate higher QrU emissions (Owensby et al., 1996). Therefore, as available
forage changes, the type and quantity of emissions from wild animals is likely to change.
11.3 Current Global Emissions
Table 11-1 presents a summary of global emission estimates for termites, other arthropods, and wild
animals. Estimating global emissions of QrU from terrestrial arthropods and wild animals has primarily
been based on bottom-up calculations using estimates of insect and animal populations, combined with an
emission factor based on available measurements of emissions from these systems. These methods
contain high levels of uncertainty, since there are limited data available for these calculations.
The AR4 provided a global emissions total for termites of 20 Tg CHVyr (Denman et al., 2007), which has
been generally agreed upon, with a variation of 50 percent (Sanderson, 1996; Houweling et al., 2000).
However, this estimate does not account for the oxidation of ClrU in the soil mounds surrounding certain
termite nests. Research by Sugimoto and Inoue estimated the global QrU emitted by termites as 2 to 7 Tg
CHVyr. The research of Sanderson (1996) and Sugimoto and Inoue (1998) suggests that a range of 2 to 22
Tg CHVyr is appropriate.
11-2
-------
Chapter 11. Terrestrial Arthropods and Wild Animals
As mentioned earlier, other research (Hackstein and Stumm, 1994) suggests that there are additional
contributions by other types of arthropods that contribute to global CH4 and these emissions may be as
large as 100 Tg CHVyr. However, this research is preliminary and has not yet been corroborated by other
sources.
Estimates of QrU emissions from wild animals range from 2 to 6 Tg CWyr (Leng, 1993, adapted from
Crutzen et al., 1986), to more recent estimates of 15 Tg CHVyr (Houweling et al., 2000). The global
distribution of ClrU emissions from wild ruminants has been most often approximated using the method
described by Bouwman et al. (1997), which assumed that wild animals consume a certain percentage of
vegetation, of which a constant fraction is assumed to be emitted as CH4. The methodology includes
assumptions regarding how much vegetation in forested ecosystems consists of consumable grass or
leaves, combined with assigning vegetation types on the basis of the land cover.
Table 11-1. Summary of Global Methane Emissions
From Arthropods and Wild Animals
Source
Termites
Other arthropods
Wild animals
Emissions
(Tg CH4/Yr)
20
20
20
20
5
15
Range of Estimate
(Tg CH4/Yr)
0-40
10-30
2-22
18-22
1.5-7.4
Up to 100
0-10
2-6
Reference
Houweling, 1999
Houweling et al., 2000
Wuebbles and Hayhoe, 2002
Sanderson, 1996
Sugimoto and Inoue, 1998
Hackstein and Stumm, 1994
Houweling, 1999
Leng, 1993 (adapted from Crutzen et al., 1986)
Houweling et al., 2000
11.4 Future Emission Scenarios
Emissions from terrestrial arthropods (including termites) and wild animals are not expected to change
significantly in the future. Changes to land use, which alter the type of plants available for wild
ruminants, could affect the diets of these animals and subsequently their rate of enteric fermentation. As
human activity encroaches on wildlife ecosystems, the reduced habitat availability will likely cause a
decrease in wild animal populations. The habitats for terrestrial arthropods and wild animals are also
linked to climate effects resulting in shifting ecosystems (in more northern environments) or drought,
which are again likely to decrease populations.
Currently, no scenarios for future QrU emissions from this source exist in the literature. However,
Owensby et al. (1996) reviewed the impact of increased atmospheric CO2 on forage quality and the effect
on ruminant ClrU emissions. They found that increased CO2 reduced nitrogen in the vegetation, and
overall negatively affected the quality of forages. Consequently, ruminant intake declined as forage
quality decreased. Owensby et al. concluded that wild ruminant diet quality will be affected and growth
and reproduction will likely be reduced.
1 1 .5 Areas for Further Research
Much research is needed to improve the bottom-up calculations of QrU emissions from terrestrial
arthropods and wild animals. Investigation of other arthropods and their potentially significant
contributions to global ClrU is needed, as well as improvements in estimates of current and future
populations of these arthropods and whether all ClrU generated is emitted to the atmosphere.
11-3
-------
Chapter 11. Terrestrial Arthropods and Wild Animals
For wild animals, improvements to the inventories of animal populations and the data available on their
diets would contribute to the development of more specific emission factors and confirm the rate of
enteric fermentation.
11.6 References
Bouwman A.F., Lee D.S., Asman W.A.H., Dentener F.J., Van Der Hoek K.W. and J.G.J. Olivier. 1997. A
Global High-Resolution Emission Inventory for Ammonia. Global Biogeochem Cycles 11(4): 561-587.
Crutzen, P.J., I. Aselmann, and W. Seiler. 1986. Methane production by domestic animals, wild
ruminants, other herbivorous fauna, and humans. Tellus, Ser. B 38: 271-284.
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Bias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press, pp. 499-587.
Hackstein, J.H.P., and C.K. Stumm. 1994. Methane production in terrestrial arthropods. Proceedings of
the National Academy of Sciences of the United States of America 91(12): 5441-5445.
Houweling, S. 1999. Global Modeling of Atmospheric Methane Sources and Sinks. Netherlands:
Universal Press.
Houweling, S., F. Dentener, and J. Lelieveld. 2000. Simulation of preindustrial atmospheric methane to
constrain the global source strength of natural wetlands. J. Geophys. Res. 105(D13): 17,243-17,255.
Leng, R.A. 1993. The impact of livestock development on environmental change. In: Food and
Agriculture Organization of the United Nations. Strategies for Sustainable Animal Agriculture in
Developing Countries. Rome: FAO.
Martius, C., P.M. Fearside, A.G. Bandeira, and R. Wassmann. 1996. Deforestation and methane release
from termites in Amazonia. Chemosphere 33: 517-536.
Owensby, C.E., R.M. Cochran, and L.M. Auen, 1996. Effects of elevated carbon dioxide on forage
quality for ruminants. In: Carbon Dioxide, Populations, and Communities. Physicologic Ecology
Series. Academic Press, pp. 363-371. http://spuds.agron.ksu.edu/fq3.html. Date accessed: August 2007.
Sanderson, M.G. 1996. Biomass of termites and their emissions of methane and carbon dioxide: A global
database. Global Biogeochemical Cycles 10(4): 543-557.
Sugimoto, A., and T. Inoue. 1998. Methane oxidation by termite mounds estimated by the carbon isotopic
composition of methane. Global Biogeochemical Cycles 12(4): 595-605.
U.S. EPA (United States Environmental Protection Agency). 1993. Current and Future Methane
Emissions From Natural Sources. EPA-430-R-93-011. Washington: U.S. Environmental Protection
Agency.
Wuebbles, D.J., and K. Hayhoe. 2002. Atmospheric methane and global change. Earth-Science Reviews
57(2002): 177-210.
11-4
-------
Chapter 12. Summary and Conclusions
Table 12-1 summarizes the current global estimates of QrU and N2O emissions from natural sources.
When compared to the total global sources of ClrU and N2O (anthropogenic emissions as identified in the
IPCC's Fourth Assessment Report plus natural emissions from this report), natural sources of QrU are
estimated to contribute about 37 percent of the annual flux of CFLj to the atmosphere, while natural
sources of N2O are estimated to represent about 64 percent of the global total.
12.1 Summary of Methane Emissions
For some natural sources, such as wetlands, CFi4 emissions are reasonably well understood and have been
quantified over some time. Other sources, such as vegetation and terrestrial and marine geologic sources,
are potentially significant but are either newly identified or early in the research stages of quantification.
Before the Industrial Revolution, natural wetlands were the dominant source of CFI4 to the atmosphere
(Brook et al., 2000; Etheridge et al., 1998). The increase in human population has decreased the relative
importance of wetland sources both by increasing anthropogenic sources and by decreasing wetland
acreage through drainage and land use change. Both direct and indirect changes in wetland fluxes will
continue, because many of the environmental variables such as temperature, rainfall, and vegetation type
that control wetland CFI4 emissions are associated with climate. A number of studies have attempted to
model wetlands' response to climate change. They calculate that emissions from wet soils will be
enhanced more than oxidation in dry tundra and forest soils. As a result, projected CFLj emissions from
northern wetlands are expected to nearly double by the end of the century. Changes in land use,
particularly in the tropics, are also likely to significantly alter emissions. Since models point to tropical
wetlands as contributing the majority of emissions, understanding change in these regions is critical.
Another potentially significant natural source of CFI4 is natural seeps from geologic sources deep within
the Earth's crust. Previous estimates of natural sources have either ignored this source or only evaluated
marine seeps (ignoring terrestrial seeps, such as magmatic volcanoes). Current estimates include
emissions from mud volcanoes, other macroseepage locations, terrestrial microseepage, and marine seeps.
There continue to be large uncertainties in these estimates, the largest of which lies in estimating sub-
oceanic emissions. In contrast, estimates of onshore emissions are based on direct measurements and up-
scaling procedures based on standard emission factor concepts applied to point sources (for individual
macroseepage features such as mud volcanoes) and homogeneous area sources (for diffuse sources such
as microseepage). Relatively few climate- or human-related factors are hypothesized to be capable of
influencing emissions of CFLj from geologic sources. Some hypotheses indicate decreased emissions
associated with large-scale extraction of oil and gas, and increased emissions following deglaciation
events, as seismic activity increases. While geologic CH4 emissions have likely changed in the past and
are likely to continue to change in the future, these mechanisms are too speculative to use as a basis to
estimate even the general direction of future changes in geologic CFLt emissions.
Recent evidence has suggested the possibility of a significant contribution of CFLt emissions from
vegetation, a natural source omitted from previous budget estimates. Previous research classified
unflooded ecosystems, such as upland tropical forest and well-drained boreal forests, as CFLj sinks.
However, more recent results suggest that plants may produce methane under aerobic conditions.
While the mechanism of plant methane production has not been identified, a number of independent
studies have demonstrated measurable methane production, particularly under stressful conditions such as
ultraviolet radiation, high temperature, or tissue damage. In contrast, some laboratory and field studies
have failed to confirm measurable methane emissions from plants. Given these conflicting results, and the
fact that an actual physical mechanism by which plants directly produce CUt has yet to be
12-1
-------
Chapter 12. Summary and Conclusions
identified, a best estimate of global plant emissions must also include the possibility of zero emissions
(i.e., that plants are in fact not a direct source o
Other notable natural sources of ClrU include microbial activity in oceans, estuaries, and rivers, and lakes,
as well as in the digestive systems of certain arthropods and wild mammals. Exact estimates still vary
widely, however.
12.2 Summary of Nitrous Oxide Emissions
The primary natural sources of N2O are upland soils and riparian areas, oceans, estuaries, and rivers.
Upland soils are well-aerated and generally oxic (Conrad, 1996), and the dry soil conditions favor
microbial processes which make dry upland soils a sink for QrU and a source of N2O. In addition, riparian
zones have saturated soil conditions and microbially available carbon which contribute to higher rates of
production of N2O than dry upland soils. The vast majority of studies in the past have focused on N2O
emissions from agricultural, not natural, soil sources. Recently, the number of N2O emissions
measurements has increased steadily, allowing for improvements in emission models and budgets,
although there are still significant model uncertainties. The prediction of future emissions of N2O
production in soils depends on the changing human activities on these soils, as well as on climate patterns
that are shifting as a result of global climate change. The clearing of land for agricultural use has been
shown to lead to increased N2O emissions and a decreased capacity for QrU oxidation, for example.
Global climate change models show patterns of temperature and precipitation changes worldwide.
Because soil moisture is a key determinant of the microbial processes that consume or produce N2O and
CH4, these shifting climate patterns will determine the fluxes of these greenhouse gases into the future.
The oceans are another major natural source of N2O to the atmosphere, with N2O produced primarily in
the water column. Emissions of N2O from other aquatic environments (e.g., estuaries and rivers) are
typically classified as largely anthropogenic because the majority of nitrogen entering these systems is
believed to be associated with human activities such as agriculture. However, a small portion of these
emissions actually reflect the natural movement of nitrogen.
12-2
-------
Chapter 12. Summary and Conclusions
Table 12-1. Current Methane and Nitrous Oxide Emissions From Natural Sources
Source
Wetlands
- Northern/bogs
- Tropical/swamps
Upland soils and riparian
areas
Oceans, estuaries, and
rivers
Permafrost
Lakes
Gas hydrates
Terrestrial and marine
geologic sources
Wildfires
Vegetation
Terrestrial arthropodsf
Wild animals
All natural sources
All sources to the
atmosphere (anthropogenic
and natural)
Natural sources as a percent
of the total
Methane (Tg CH4/year)
Emissions
Estimate3
170.3
42.7
127.6
-30
9.1
0.5
30
20
8
208
566j
37%
Rangeb
24-72
81-206
Not
available
2.3-15.6
0-1
10-50
2-9e
42-64
2-5
Not a
source
or 20-60
2-22
2-15
See
note9
503-
610k
See
note9
513C (%0)c
-62
-58.9
-58
-53.8
-62.5
-41.8
-25
Not
available
-63
-60.5
-57h
-54.5C
n/a
Nitrous Oxide (Tg N/year)
Emissions
Estimate3
Rangeb
515N (%0)d
Negligible
6.6
5.4
3.3-9.0
1.5-9.1
-38 to +2
-2 to +12
Negligible
0.1
12.1
18.81
64%
0.004-
0.04
Not
available
See
note9
8.5-27.7
See
note9
8.8'
?m
n/a
a In some cases, a point estimate cannot be provided due to large uncertainty.
b Ranges presented here may reflect a compilation of several different estimates. Published estimates vary due in
part to uncertainty in estimating the global number of point and diffuse sources and the average annual emissions
from each individual source or source area.
c Mean value from Whiticar and Schaefer, 2007, and references therein.
d Range from Rahn and Wahlen, 2000, and references therein.
e The emission estimates for gas hydrates correspond to the flux of methane to the ocean, most of which is likely to
be oxidized in the ocean water column.
f Estimates for terrestrial arthropods include termites. It is estimated that other arthropods could contribute up to 100
Tg CH4/year.
12-3
-------
Chapter 12. Summary and Conclusions
g Because the relative contributions of emissions from each source to the total budget are not independent of each
other (i.e., if one source is at the lower end of its estimated range, another may be at the higher), the ranges
cannot be summed.
h Lasseyetal., 2007.
' Based on change from pre-industrial to present as estimated by Rockmann et al., 2003; assuming that pre-
industrial emissions are primarily natural.
1 Mean value for anthropogenic emissions from Wuebbles and Hayhoe, 2002; natural emissions from this work.
k Range in total anthropogenic and natural emissions from Denman et al., 2007, and references therein.
1 Estimates of anthropogenic emissions from Denman et al., 2007; natural emissions from this work.
m Observed tropospheric values from Rahn and Wahlen, 2000, and references therein.
12.3 Future Needs
It continues to be difficult to estimate contributions from natural sources, and uncertainties can be large,
as evidenced by the large ranges associated with the emissions estimates. Additional research focused on
improving our understanding of the processes that result in CH4 and N2O emissions should improve
current flux estimates and help refine future estimates under altered environmental conditions. High
uncertainty in some sources is a result of a lack of basic data - flux measurements may be sparse from
some geographic regions and/or seasons. For a number of sources such as wetlands, uncertainties are high
in part because these are highly dynamic systems that respond to short-term climate and weather
variability with changed emissions. This source of uncertainty will always be present. A number of
sources currently rely on inventory-type data to extrapolate small-scale measurements. While this is
reasonable for some sources (for example, the number of mud volcanoes is unlikely to change quickly or
drastically), this means that they are largely static estimates. Even if modeled, these flux estimates will be
limited by the spatial and temporal resolution of the data used for their extrapolation. Reliance on
inventory or long-term average data also means that it is difficult to fully take advantage of the
accumulating database of atmospheric mixing ratios and isotopic signatures. These data are highly
dynamic and this short-term variability is a crucial part of their utility in inverse modeling approaches.
These techniques have proven that they can both help to constrain "bottom up" estimates and provide a
way to integrate highly variable natural systems.
For wetlands, the major natural source contributing to CFI4 emissions, research in tropical areas remains
sparse and incomplete. Increased work linking emissions to environmental controls, long-term studies to
capture seasonality and inter-annual variability, and work on the importance of episodic emissions will
help resolve difficulties in modeling these systems. In addition, more work should examine the
relationships between CFLj flux and net primary productivity (the rate at which biomass is produced, for
example by photosynthesis), since these relationships appear to be habitat-specific. Because emissions to
the atmosphere are a function of the competing processes of CH4 production and consumption, both
processes and their responses to environmental controls must be understood across the landscape.
Episodic emissions, which may release a sizeable fraction of annual flux, remain difficult to measure and
include in models. Failure to adequately incorporate these fluxes, however, can yield inaccurate and
misleading results.
For upland soils and riparian zones, the major natural source contributing to N2O emissions, more field
measurements and improvements in global emissions models are needed. While field measurements of
N2O have increased steadily in the past several years, coverage of global vegetation zones remains
incomplete. More measurement data are needed, especially for the dry tropical forest, savanna, tundra,
and temperate ecosystems not affected by nitrogen deposition. These measurements should be carried out
over extended periods, to help improve our understanding of the complex factors that impact emissions as
well as to assess natural variability.
12-4
-------
Chapter 12. Summary and Conclusions
There are many additional areas where research would help improve flux estimates. These are discussed
in more detail in the source-specific chapters, but we briefly list a number of them here, where
uncertainties are notably high. They include:
• Data from tropical and southern latitude oceans, estuaries, and rivers, as well as estimates of
upwelling sources.
• Improvements to permafrost models to account for lateral water movement, dynamic vegetation
algorithms, and detailed soil physics.
• Data to quantify lake fluxes, particularly in the Arctic, boreal region, and tropics.
• Better quantification of CH4 reserves stored as gas hydrates, as well as better estimation of the
rate of QrU absorption into oceans and CH4 oxidation in the water column.
Rates of QrU from seeps and mud volcanoes oxidized in sediments, as well as better
quantification of the source locations (e.g., number of mud volcanoes, frequency of eruption).
Activity data for wildfires, including area and amount of biomass, burned area estimates
associated with natural wildfires, and additional research on emissions related to different "fuels"
(i.e., different types of vegetation).
Confirmation or rejection of vegetation as a source
• Research that better quantifies the oxidation of QrU through termite mounds, confirmation of CH4
from non-termite terrestrial arthropods, and activity data for arthropods and wild animals.
12.4 Summary
Natural sources make important contributions to the global atmospheric budgets of QrU and N2O.
Emissions from these sources will change as a result of increased human activities (e.g., decreasing CH4
from wetlands through land use changes) and as a consequence of climate change (e.g., increasing
frequency and severity of wildfires due to warmer and drier conditions). Although our understanding of
the scope of possible changes in emissions has increased significantly in the past few years through model
development and improvement, large unknowns remain. There is a potential for very large changes in
natural CH4 emissions and the possibility of positive feedbacks between these radiatively important gases
and climate means that research and model refinement must continue. As illustrated by recent work
suggesting that recent decreases in natural ClrU emissions from wetlands have temporarily masked
increases in anthropogenic emissions (Bosquet et al., 2006), it is impossible to understand the system as a
whole if its various components and their links are not understood.
12.5 References
Brook, E.J., S. Harder, J. Severinghaus, E.J. Steig, and C.M. Sucher. 2000. On the origin and timing of
rapid changes in atmospheric methane during the last glacial period. Global Biogeochem. Cycles 14:
559-572.
Bousquet, P., P. Ciais, J. B. Miller, E. J. Dlugokencky, D. A. Hauglustaine, C. Prigent, G. R. Van der
Werf, P. Peylin, E.-G. Brunke, C. Carougel, R. L. Langenfelds, J. Lathiere, F. Papa, M. Ramonet, M.
Schmidt, L. P. Steele, S. C. Tyler, and J. White. 2006. Contribution of anthropogenic and natural
sources to atmospheric methane variability. Nature 443(28): 439-443.
Conrad, R. 1996. Soil microorganisms as controllers of atmospheric trace gases (H2, CO, CFLj, OCS, N2O,
). Microbiological Review 60(4): 609-640.
12-5
-------
Chapter 12. Summary and Conclusions
Denman, K.L., G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S. Ramachandran, P.L. da Silva Bias, S.C. Wofsy, and X.
Zhang. 2007. Couplings between changes in the climate system and biogeochemistry. In: S. Solomon,
D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.). Climate
Change 2007: The Physical Basis. Contribution of Working Group I to the Fourth Assessment Report
of the Intergovernmental Panel on Climate Change. Cambridge, UK, and New York, NY: Cambridge
University Press, pp. 499-587.
Etheridge, D.M., L.P. Steele, R.J. Francey, and R.L. Langenfelds. 1998. Atmospheric methane between
1000 A.D. and present: Evidence of anthropogenic emissions and climatic variability. J. Geophys. Res.
103: 15,979-15,993.
12-6
-------
Appendix A: Glossary
albedo: The fraction of solar radiation reflected by a surface or object, often expressed as a percentage.
anoxic: without oxygen.
anthropogenic: Made by people or resulting from human activities (e.g., emissions that are produced as a
result of human activities.
apparent oxygen utilization (AOU): the difference between a measured dissolved O2 concentration and
that expected when at atmospheric equilibrium saturation.
aphotic: having no light.
Arctic Oscillation: refers to opposing atmospheric pressure patterns in northern middle and high
latitudes. The dominant pattern of sea-level pressure variation north of 20°N. Pressure variations cause
changes in wind patterns and ocean currents. The pattern of alternating high and low pressure thus creates
alternating warm and cool temperatures throughout this large region.
arthropods: the largest phylum of animals which includes insects, arachnids, crustaceans, and others.
Characterized by the possession of a segmented body with appendages on at least one segment. All
arthropods are covered by a hard exoskeleton.
biogeochemical cycle: the biology, geology, and chemistry of the global or regional cycles of the "life
elements" carbon, nitrogen, sulfur, and phosphorus with reservoirs including the atmosphere, oceans,
sediments, and living organisms.
biome: a distinct ecological community of plants and animals living together in a particular climate.
budget: a balance sheet of all sources and sinks of a reservoir (e.g., all methane into and out of the
atmosphere).
clathrate-hydrates: see "gas hydrates."
climate change: a change in the state of the climate that can be identified (e.g., by using statistical tests)
by changes in the mean and/or the variability of its properties, and that persists for an extended period,
typically decades or longer.
controls: variables that affect production and consumption of CH4 or N2O.
denitrification: reduction of nitrate or nitrite to molecular nitrogen or nitrogen oxides by microbial
activity or by chemical reactions involving nitrite.
ebullition: bubbling.
emissions: the release of a substance (usually a gas when referring to the subject of climate change) into
the atmosphere.
ephemeral: short-lived (e.g., a wetland, pond, or spring exists for only a brief period, usually following
precipitation or snowmelt).
A-1
-------
Appendix A. Glossary
euphotic: having light.
extratropical: occurring between 30° and 60° latitudes from the equator in both the hemispheres
flux: the amount of material transferred from one reservoir to another per unit time (e.g., methane
emissions to the atmosphere, or methane consumption of soils).
gas hydrates: ice-like compounds formed between water and a gas molecule such as CH4, under high
pressure and at temperatures near the freezing point of water.
global warming: an average increase in the temperature of the atmosphere near the Earth's surface and in
the troposphere, which can contribute to changes in global climate patterns.
global warming potential (GWP): the cumulative radiative forcing effects of a gas over a specified time
horizon resulting from the emission of a unit mass of gas relative to a reference gas.
greenhouse effect: trapping and build-up of heat in the atmosphere near the Earth's surface (the
troposphere).
greenhouse gas (GHG): any gas that absorbs infrared radiation in the atmosphere.
gyre: a relatively stationary region of the open ocean with a circular current created by the Coriolis effect.
Gyres are permanent large-scale water circulation features whose circulation tends to isolate them from
the rest of the ocean.
ice core: a cylindrical section of ice removed from a glacier or an ice sheet in order to study climate
patterns of the past.
isotope: Any two or more forms of an element having identical or very closely related chemical
properties and the same atomic number but different atomic weights or mass numbers.
macroseepage: relatively large, visibly detectable, localized emissions from identified geologic features
and events such as mud volcanoes and other seeps independent of mud volcanism .
methane (CH/t): a hydrocarbon that is a greenhouse gas with a global warming potential most recently
estimated at 25 times that of carbon dioxide.
methanogens: microorganisms (Archaea) that produce methane as a metabolic byproduct in anoxic
conditions. Methanogens are anaerobic; most are rapidly killed by the presence of oxygen.
methanotrophs: bacteria that can grow using methane as their only source of carbon and energy.
methylotrophs: bacteria that can grow using reduced one-carbon compounds, such as methanol or
methane, as the carbon source for their growth. Some methylotrophs can degrade the methane (see
methanotrophs).
microsite: small volume of soil where biological or chemical processes differ from those of the soil as a
whole.
microseepage: diffuse fluxes of gaseous hydrocarbons over wide areas.
A-2
-------
Appendix A. Glossary
net primary productivity (NPP): the net production of organic compounds from atmospheric or aquatic
carbon dioxide, principally by plants through the process of photosynthesis. It is the difference between
the amount of material or energy produced (gross productivity) and respiration (the costs of producing
it—cellular respiration and tissue maintenance).
nitrification: Biological oxidation of ammonium to nitrite and nitrate, or a biologically induced increase
in the oxidation state of nitrogen.
nitrous oxide (N2O): a powerful greenhouse gas with a global warming potential of 298 times that of
carbon dioxide.
oigotrophic: nutrient-poor, usually having low productivity. Its opposite is "eutrophic," or nutrient-rich.
oxic: containing oxygen.
oxidize: To chemically transform a substance by combining it with oxygen.
peatlands: An environment where partially decayed vegetation matter accumulates to form organic-rich
soils (peats), usually in wetlands.
permafrost: soil, sediment, or rock that is continuously frozen (temperature below 0°C) for at least two
consecutive years.
pycnocline: a water layer with a large change in density caused by temperature or salinity. When caused
by temperature, it is usually called a thermocline. Mixing is impeded across such a layer.
pyrolysis: chemical decomposition of organic materials by heating in the absence of oxygen or any other
chemical agents
radiative forcing: a measure of how the energy balance of the Earth-atmosphere system is influenced
when factors that affect climate are altered.
reactive N: forms of nitrogen that can be used by living organisms.
redox (short for reduction/oxidation): the relative oxidation status of a soil. Soils with a low redox status
have little available oxygen, which limits the types of reactions that can take place.
reduced species: term used to describe the degree of reduction (number of electrons or number of
hydrogen atoms) in atoms, molecules, or ions. For example, CF^ is a reduced compound with relatively
little available energy for microbial growth, while CO2 is oxidized compound that can yield greater
energy as it is broken down.
riparian areas: vegetated ecosystems along a waterbody through which energy, materials, and water
pass.
shrublands: ecosystems dominated by woody or herbaceous shrubs
sink: a flux of material out of a reservoir.
A-3
-------
Appendix A. Glossary
Southern Oscillation Index (SOI): the normalized pressure difference in surface pressure between
Tahiti, French Polynesia, and Darwin, Australia. Positive SOI values of indicate a La Nina event, while
negative SOI values indicate an El Nino event.
source: a flux of material into a reservoir.
suboxic: oxygen deficient; may describe the transition zone between the two extremes. Technically, it is
often defined as less than or equal to 10 uM (micromoles) O2.
synoptic scale: scale used with respect to weather systems ranging in size from several hundred
kilometers to several thousand kilometers, the scale of high and low pressure systems (frontal cyclones)
of the lower troposphere.
thermogenesis: production of a substance (e.g., methane) by thermal breakdown of organic matter.
thermokarst: a pitted land surface that forms as permafrost melts.
thermokarst lake: a body of freshwater formed in a depression by water from thawing.
upland soils: well-aerated soils with lower moisture content than wetland soils.
upwelling: a pattern of coastal and open water oceanic circulation. It is created by persistent winds
blowing across the ocean surface. As winds move surface waters, they are replaced by deeper waters that
are richer in nutrients and which can support increased phytoplankton growth, which in turn supports
higher populations offish and other consumers.
volatile organic compounds (VOCs): Carbon-containing chemical compounds that can vaporize easily
and can play an important role in atmospheric chemistry. Based on their molecular structure, these
organic species can be grouped in different classes of compounds, including, aldehydes, alcohols,
ketones, acids, etc.
wetlands: areas where water covers the soil, or is present either at or near the surface of the soil all year
or for varying periods of time during the year.
A-4
------- |