1990-2019
Inventory of
U.S. Greenhouse Gas
Emissions and Sinks
United States
Environmental Protection
Agency
EPA 430-R-21-005

-------
Front cover photo credit for cow and digester: Vanguard Renewables.

-------
HOW TO OBTAIN COPIES
You can electronically download this document on the U.S. EPA's homepage at
.
All data tables of this document for the full time series 1990 through 2019, inclusive, will be made available with
the final report published on April 14, 2021 at the internet site mentioned above.
FOR FURTHER INFORMATION
Contact Ms. Mausami Desai, Environmental Protection Agency, (202) 343-9381, desai.mausami@epa.gov,
or Mr. Vincent Camobreco, Environmental Protection Agency, (202) 564-9043, camobreco.vincent@epa.gov.
For more information regarding climate change and greenhouse gas emissions, see the EPA web site at
.

-------
Acknowledgments
The Environmental Protection Agency would like to acknowledge the many individual and organizational
contributors to this document, without whose efforts this report would not be complete. Although the complete
list of researchers, government employees, and consultants who have provided technical and editorial support is
too long to list here, EPA would like to thank some key contributors and reviewers whose work has significantly
improved this year's report.
Within EPA's Office of Atmospheric Programs, development and compilation of emissions from fuel combustion
was led by Vincent Camobreco. Sarah Roberts and Justine Geidosch directed the work to compile estimates of
emissions from mobile sources. Work on fugitive methane emissions from the Energy sector was directed by
Melissa Weitz and Chris Sherry with support from Erin McDuffie. Development and compilation of emissions
estimates for the Waste sector were led by Rachel Schmeltz and Lauren Aepli. Tom Wirth and John Steller directed
work to compile estimates for the Agriculture and the Land Use, Land-Use Change, and Forestry chapters.
Development and compilation of Industrial Processes and Product Use (IPPU) C02, CH4, and N20 emissions was
directed by Amanda Chiu and Vincent Camobreco. Development and compilation of emissions of HFCs, PFCs, SF6,
and NF3 from the IPPU sector was directed by Deborah Ottinger, Dave Godwin, and Stephanie Bogle. Cross-cutting
work was directed by Mausami Desai with support from Erin McDuffie on the uncertainty analysis.
Other EPA offices and programs also contributed data, analysis, and technical review for this report. The Office of
Atmospheric Program's Greenhouse Gas Reporting Program staff facilitated aggregation and review of facility-level
data for use in the Inventory, in particular confidential business information data. The Office of Transportation and
Air Quality and the Office of Air Quality Planning and Standards provided analysis and review for several of the
source categories addressed in this report. The Office of Land and Emergency Management and the Office of
Research and Development also contributed analysis and research.
The Energy Information Administration and the Department of Energy contributed invaluable data and analysis on
numerous energy-related topics. William Sanchez at EIA provided annual energy data that are used in fossil fuel
combustion estimates. Other government agencies have contributed data as well, including the U.S. Geological
Survey, the Federal Highway Administration, the Department of Transportation, the Bureau of Transportation
Statistics, the Department of Commerce, the Mine Safety and Health Administration, and the National Agricultural
Statistics Service.
We thank the Department of Defense (David Asiello, DoD and Matthew Cleaver of Leidos) for compiling the data
on military bunker fuel use.
We thank the Federal Aviation Administration (Ralph Lovinelli, Maryalice Locke, and Jeetendra Upadhyay) for
compiling the inventory of emissions from commercial aircraft jet fuel consumption.
We thank the U.S. Department of Agriculture's Forest Service (Grant Domke, Brian Walters, Jim Smith and Mike
Nichols) for compiling the inventories for C02, CH4, and N20 fluxes associated with forest land.
We thank the Department of Agriculture's Agricultural Research Service (Stephen Del Grosso) and the Natural
Resource Ecology Laboratory and Department of Statistics at Colorado State University (Stephen Ogle, Bill Parton,
F. Jay Breidt, Shannon Spencer, Ram Gurung, Ernie Marx, Stephen Williams and Guhan Dheenadayalan Sivakami)

-------
for compiling the inventories for CH4 emissions, N20 emissions, and C02 fluxes associated with soils in croplands,
grasslands, and settlements.
We thank Silvestrum Climate Associates (Stephen Crooks, Lisa Schile Beers), National Oceanic and Atmospheric
Administration (Nate Herold, Ben DeAngelo and Meredith Muth), the Smithsonian Environmental Research Center
(J. Patrick Megonigal, James Holmquist and Meng Lu), and Florida International University (Tiffany Troxler) as well
as members of the U.S. Coastal Wetland Carbon Working Group for compiling inventories of land use change, soil
carbon stocks and stock change, CH4 emissions, and N20 emissions from aquaculture in coastal wetlands.
We would also like to thank Marian Martin Van Pelt, Leslie Chinery, Alexander Lataille, Erin Asher and the full
Inventory team at ICF including Diana Pape, Robert Lanza, Mollie Averyt, Larry O'Rourke, Deborah Harris, Rebecca
Ferenchiak, Kasey Knoell, Max Kaffel, Deep Shah, Lou Browning, Emily Golla, Katie O'Malley, Neha Vaingankar,
Grace Tamble, Mary Francis McDaniel, Mollie Carroll, Tyler Brewer, Johanna Garfinkel, Madeleine Pearce, Carolyn
Pugh, Claire Trevisan, Zeyu Hu, Alex Da Silva, and Eliza Puritz for technical support in compiling synthesis
information across the report and preparing many of the individual analyses for specific report chapters including
fluorinated emissions and fuel combustion.
We thank Eastern Research Group for their analytical support. Deborah Bartram, Kara Edquist and Tara Stout
support the development of emissions estimates for wastewater. Cortney Itle, Kara Edquist, Amber Allen, Tara
Stout, and Spencer Sauter support the development of emission estimates for Manure Management, Enteric
Fermentation, Wetlands Remaining Wetlands, and Landfilled Yard Trimmings and Food Scraps (included in
Settlements Remaining Settlements). Brandon Long, Bryan Lange, Gopi Manne, Sarah Downes, Marty Wolf, and
Casey Pickering develop estimates for Natural Gas and Petroleum Systems. Cortney Itle, Gopi Manne and Tara
Stout support the development of emission estimates for coal mine methane.
Finally, we thank the RTI International team: Kate Bronstein, Meaghan McGrath, Jeff Coburn, and Keith Weitz for
their analytical support in development of the estimates of emissions from landfills; Jason Goldsmith, Melissa
Icenhour, Michael Laney, Carson Moss, David Randall, Gabrielle Raymond, and Karen Schaffner for their analytical
support in development of IPPU C02, CH4, and N20 emissions; and Tiffany Moore and Cassy Becker for their
analytical support on disaggregating industrial sector fossil fuel combustion emissions.

-------
Preface
The United States Environmental Protection Agency (EPA) prepares the official U.S. Inventory of Greenhouse Gas
Emissions and Sinks to fulfill annual existing commitments under the United Nations Framework Convention on
Climate Change (UNFCCC). Under Article 4 and decisions at the First, Second, Fifth and Nineteenth Conference of
Parties, national inventories for UNFCCC Annex I parties should be provided to the UNFCCC Secretariat each year
by April 15.
In an effort to engage the public and researchers across the country, EPA conducts an annual public review and
comment process for this document. The document was made available on the EPA Greenhouse Gas Emissions
website and announced via Federal Register Notice for 30 days; comments received after the 30-day public
comment period were accepted and will be considered for the next edition of this annual report. Public review of
this year's report occurred from February 12 to March 15, 2021 and comments received were posted to the docket
EPA-HQ-OAR-2021-0008. Responses to comments received will be posted to EPA's website within 2-4 weeks
following publication of this report.

-------
Table of Contents
TABLE OF CONTENTS	VI
LIST OF TABLES, FIGURES, AND BOXES	IX
EXECUTIVE SUMMARY	ES-1
ES.l Background Information	ES-3
ES.2 Recent Trends in U.S. Greenhouse Gas Emissions and Sinks	ES-4
ES.3 Overview of Sector Emissions and Trends	ES-19
ES.4 Other Information	ES-26
1.	INTRODUCTION	1-1
1.1	Background Information	1-3
1.2	National Inventory Arrangements	1-11
1.3	Inventory Process	1-13
1.4	Methodology and Data Sources	1-16
1.5	Key Categories	1-17
1.6	Quality Assurance and Quality Control (QA/QC)	1-23
1.7	Uncertainty Analysis of Emission Estimates	1-25
1.8	Completeness	1-28
1.9	Organization of Report	1-29
2.	TRENDS IN GREENHOUSE GAS EMISSIONS	2-1
2.1	Trends in U.S. Greenhouse Gas Emissions and Sinks	2-1
2.2	Emissions by Economic Sector	2-29
2.3	Precursor Greenhouse Gas Emissions (CO, NOx, NMVOCs, and S02)	2-41
3.	ENERGY	3-1
3.1	Fossil Fuel Combustion (CRF Source Category 1A)	3-6
3.2	Carbon Emitted from Non-Energy Uses of Fossil Fuels (CRF Source Category 1A)	3-49
3.3	Incineration of Waste (CRF Source Category 1A5)	3-57
3.4	Coal Mining (CRF Source Category lBla)	3-61
3.5	Abandoned Underground Coal Mines (CRF Source Category lBla)	3-66
vi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
3.6	Petroleum Systems (CRF Source Category lB2a)	3-70
3.7	Natural Gas Systems (CRF Source Category lB2b)	3-89
3.8	Abandoned Oil and Gas Wells (CRF Source Categories lB2a and lB2b)	3-110
3.9	Energy Sources of Precursor Greenhouse Gas Emissions	3-115
3.10	International Bunker Fuels (CRF Source Category 1: Memo Items)	3-117
3.11	Wood Biomass and Biofuels Consumption (CRF Source Category 1A)	3-122
4.	INDUSTRIAL PROCESSES AND PRODUCT USE	4-1
4.1	Cement Production (CRF Source Category 2A1)	4-10
4.2	Lime Production (CRF Source Category 2A2)	4-14
4.3	Glass Production (CRF Source Category 2A3)	4-20
4.4	Other Process Uses of Carbonates (CRF Source Category 2A4)	4-24
4.5	Ammonia Production (CRF Source Category 2B1)	4-29
4.6	Urea Consumption for Non-Agricultural Purposes	4-34
4.7	Nitric Acid Production (CRF Source Category 2B2)	4-37
4.8	Adipic Acid Production (CRF Source Category 2B3)	4-42
4.9	Caprolactam, Glyoxal and Glyoxylic Acid Production (CRF Source Category 2B4)	4-46
4.10	Carbide Production and Consumption (CRF Source Category 2B5)	4-49
4.11	Titanium Dioxide Production (CRF Source Category 2B6)	4-52
4.12	Soda Ash Production (CRF Source Category 2B7)	4-56
4.13	Petrochemical Production (CRF Source Category 2B8)	4-59
4.14	HCFC-22 Production (CRF Source Category 2B9a)	4-67
4.15	Carbon Dioxide Consumption (CRF Source Category 2B10)	4-70
4.16	Phosphoric Acid Production (CRF Source Category 2B10)	4-74
4.17	Iron and Steel Production (CRF Source Category 2C1) and Metallurgical Coke Production	4-78
4.18	Ferroalloy Production (CRF Source Category 2C2)	4-88
4.19	Aluminum Production (CRF Source Category 2C3)	4-92
4.20	Magnesium Production and Processing (CRF Source Category 2C4)	4-99
4.21	Lead Production (CRF Source Category 2C5)	4-104
4.22	Zinc Production (CRF Source Category 2C6)	4-108
4.23	Electronics Industry (CRF Source Category 2E)	4-113
4.24	Substitution of Ozone Depleting Substances (CRF Source Category 2F)	4-129
4.25	Electrical Transmission and Distribution (CRF Source Category 2G1)	4-138
4.26	Nitrous Oxide from Product Uses (CRF Source Category 2G3)	4-146
4.27	Industrial Processes and Product Use Sources of Precursor Gases	4-149
5.	AGRICULTURE	5-1
vii

-------
5.1	Enteric Fermentation (CRF Source Category 3A)	5-4
5.2	Manure Management (CRF Source Category 3B)	5-11
5.3	Rice Cultivation (CRF Source Category 3C)	5-20
5.4	Agricultural Soil Management (CRF Source Category 3D)	5-27
5.5	Liming (CRF Source Category 3G)	5-45
5.6	Urea Fertilization (CRF Source Category 3H)	5-48
5.7	Field Burning of Agricultural Residues (CRF Source Category 3F)	5-51
6.	LAND USE, LAND-USE CHANGE, AND FORESTRY	6-1
6.1	Representation of the U.S. Land Base	6-8
6.2	Forest Land Remaining Forest Land (CRF Category 4A1)	6-23
6.3	Land Converted to Forest Land (CRF Source Category 4A2)	6-46
6.4	Cropland Remaining Cropland (CRF Category 4B1)	6-53
6.5	Land Converted to Cropland (CRF Category 4B2)	6-65
6.6	Grassland Remaining Grassland (CRF Category 4C1)	6-71
6.7	Land Converted to Grassland (CRF Category 4C2)	6-82
6.8	Wetlands Remaining Wetlands (CRF Category 4D1)	6-89
6.9	Land Converted to Wetlands (CRF Source Category 4D2)	6-115
6.10	Settlements Remaining Settlements (CRF Category 4E1)	6-122
6.11	Land Converted to Settlements (CRF Category 4E2)	6-141
6.12	Other Land Remaining Other Land (CRF Category 4F1)	6-147
6.13	Land Converted to Other Land (CRF Category 4F2)	6-148
7.	WASTE	7-1
7.1	Landfills (CRF Source Category 5A1)	7-4
7.2	Wastewater Treatment and Discharge (CRF Source Category 5D)	7-20
7.3	Composting (CRF Source Category 5B1)	7-53
7.3	Stand-Alone Anaerobic Digestion (CRF Source Category 5B2)	7-57
7.4	Waste Incineration (CRF Source Category 5C1)	7-63
7.5	Waste Sources of Precursor Greenhouse Gases	7-63
8.	OTHER	8-1
9.	RECALCULATIONS AND IMPROVEMENTS	9-1
10.	REFERENCES	10-1
viii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
List of Tables, Figures, and Boxes
Tables
Table ES-1: Global Warming Potentials (100-Year Time Horizon) Used in this Report	ES-3
Table ES-2: RecentTrends in U.S. Greenhouse Gas Emissions and Sinks (MMT C02 Eq.)	ES-7
Table ES-3: C02 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT C02 Eq.)	ES-13
Table ES-4: RecentTrends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCCSector (MMTC02 Eq.)
	ES-20
Table ES-5: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and
Forestry (MMT C02 Eq.)	ES-24
Table ES-6: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT C02 Eq.)	ES-26
Table ES-7: U.S. Greenhouse Gas Emissions by Economic Sector with Electricity-Related Emissions Distributed
(MMT C02 Eq.)	ES-27
Table ES-8: RecentTrends in Various U.S. Data (Index 1990 = 100)	ES-28
Table 1-1: Global Atmospheric Concentration, Rate of Concentration Change, and Atmospheric Lifetime of
Selected Greenhouse Gases	1-5
Table 1-2: Global Warming Potentials and Atmospheric Lifetimes (Years) Used in this Report	1-10
Table 1-3: Comparison of 100-Year GWP values	1-11
Table 1-4: Key Categories for the United States (1990 and 2019)	1-17
Table 1-5: Estimated Overall Inventory Quantitative Uncertainty for 1990 (MMT C02 Eq. and Percent)	1-26
Table 1-6: Estimated Overall Inventory Quantitative Uncertainty for 2019 (MMT C02 Eq. and Percent)	1-27
Table 1-7: Quantitative Assessment of Trend Uncertainty (MMT C02 Eq. and Percent)	1-28
Table 1-8: IPCC Sector Descriptions	1-29
Table 1-9: List of Annexes	1-30
Table 2-1: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT C02 Eq.)	2-3
Table 2-2: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (kt)	2-6
Table 2-3:	Recent Trends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT C02 Eq.)... 2-9
Table 2-4: Emissions from Energy (MMT C02 Eq.)	2-12
Table 2-5: C02 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT C02 Eq.)	2-14
Table 2-6: Emissions from Industrial Processes and Product Use (MMT C02 Eq.)	2-19
Table 2-7: Emissions from Agriculture (MMT C02 Eq.)	2-22
Table 2-8: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and Forestry
(MMT C02 Eq.)	2-25
Table 2-9: Emissions from Waste (MMT C02 Eq.)	2-28
Table 2-10: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT C02 Eq. and Percent of Total in
2019)	2-30
Table 2-11: Electric Power-Related Greenhouse Gas Emissions (MMT C02 Eq.)	2-34

-------
Table 2-12: U.S. Greenhouse Gas Emissions by Economic Sector and Gas with Electricity-Related Emissions
Distributed (MMT C02 Eq.) and Percent of Total in 2019	2-35
Table 2-13: Transportation-Related Greenhouse Gas Emissions (MMT C02 Eq.)	2-38
Table 2-14: Recent Trends in Various U.S. Data (Index 1990 = 100)	2-41
Table 2-15: Emissions of NOx, CO, NMVOCs, and S02 (kt)	2-42
Table 3-1: C02, CH4, and N20 Emissions from Energy (MMT C02 Eq.)	3-3
Table 3-2: C02, CH4, and N20 Emissions from Energy (kt)	3-4
Table 3-3: C02, CH4, and N20 Emissions from Fossil Fuel Combustion (MMT C02 Eq.)	3-7
Table 3-4: C02, CH4, and N20 Emissions from Fossil Fuel Combustion (kt)	3-7
Table 3-5: C02 Emissions from Fossil Fuel Combustion by Fuel Type and Sector (MMT C02 Eq.)	3-7
Table 3-6: Annual Change in C02 Emissions and Total 2019 C02 Emissions from Fossil Fuel Combustion for Selected
Fuels and Sectors (MMT C02 Eq. and Percent)	3-9
Table 3-7: C02 Emissions from Stationary Fossil Fuel Combustion (MMT C02 Eq.)	3-13
Table 3-8: CH4 Emissions from Stationary Combustion (MMT C02 Eq.)	3-14
Table 3-9: N20 Emissions from Stationary Combustion (MMT C02 Eq.)	3-14
Table 3-10: C02, CH4, and N20 Emissions from Fossil Fuel Combustion by Sector (MMT C02 Eq.)	3-15
Table 3-11: C02, CH4, and N20 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT C02 Eq.)	3-16
Table 3-12: Electric Power Generation by Fuel Type (Percent)	3-17
Table 3-13: C02 Emissions from Fossil Fuel Combustion in Transportation End-Use Sector (MMT C02 Eq.)	3-27
Table 3-14: CH4 Emissions from Mobile Combustion (MMT C02 Eq.)	3-30
Table 3-15: N20 Emissions from Mobile Combustion (MMT C02 Eq.)	3-31
Table 3-16: Carbon Intensity from Direct Fossil Fuel Combustion by Sector (MMT C02 Eq./QBtu)	3-35
Table 3-17: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Energy-Related Fossil Fuel
Combustion by Fuel Type and Sector (MMT C02 Eq. and Percent)	3-38
Table 3-18: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 Emissions from Energy-Related
Stationary Combustion, Including Biomass (MMT C02 Eq. and Percent)	3-44
Table 3-19: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 Emissions from Mobile Sources (MMT
C02 Eq. and Percent)	3-48
Table 3-20: C02 Emissions from Non-Energy Use Fossil Fuel Consumption (MMT C02 Eq. and Percent)	3-50
Table 3-21: Adjusted Consumption of Fossil Fuels for Non-Energy Uses (TBtu)	3-51
Table 3-22: 2019 Adjusted Non-Energy Use Fossil Fuel Consumption, Storage, and Emissions	3-51
Table 3-23: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Non-Energy Uses of Fossil Fuels
(MMT C02 Eq. and Percent)	3-54
Table 3-24: Approach 2 Quantitative Uncertainty Estimates for Storage Factors of Non-Energy Uses of Fossil Fuels
(Percent)	3-54
Table 3-25: C02, CH4, and N20 Emissions from the Incineration of Waste (MMT C02 Eq.)	3-58
Table 3-26: C02, CH4, and N20 Emissions from the Incineration of Waste (kt)	3-58
Table 3-27: Municipal Solid Waste Generation (Metric Tons) and Percent Combusted (BioCycle dataset)	3-59
x Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 3-28: Approach 2 Quantitative Uncertainty Estimates for C02 and N20 from the Incineration of Waste (MMT
C02 Eq. and Percent)	3-60
Table 3-29: Coal Production (kt)	3-61
Table 3-30: CH4 Emissions from Coal Mining (MMT C02 Eq.)	3-62
Table 3-31: CH4 Emissions from Coal Mining (kt)	3-62
Table 3-32: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Coal Mining (MMT C02 Eq. and
Percent)	3-65
Table 3-33: CH4 Emissions from Abandoned Coal Mines (MMT C02 Eq.)	3-67
Table 3-34: CH4 Emissions from Abandoned Coal Mines (kt)	3-67
Table 3-35: Number of Gassy Abandoned Mines Present in U.S. Basins in 2019, Grouped by Class According to
Post-Abandonment State	3-69
Table 3-36: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Abandoned Underground Coal
Mines (MMT C02 Eq. and Percent)	3-70
Table 3-37: Total Greenhouse Gas Emissions (C02, CH4, and N20) from Petroleum Systems (MMT C02 Eq.)	3-72
Table 3-38: CH4 Emissions from Petroleum Systems (MMT C02 Eq.)	3-72
Table 3-39: CH4 Emissions from Petroleum Systems (kt CH4)	3-73
Table 3-40: C02 Emissions from Petroleum Systems (MMT C02)	3-73
Table 3-41: C02 Emissions from Petroleum Systems (kt C02)	3-73
Table 3-42: N20 Emissions from Petroleum Systems (Metric Tons C02 Eq.)	3-74
Table 3-43: N20 Emissions from Petroleum Systems (Metric Tons N20)	3-74
Table 3-44: Approach 2 Quantitative Uncertainty Estimates for CH4 and C02 Emissions from Petroleum Systems
(MMT C02 Eq. and Percent)	3-77
Table 3-45: Recalculations of C02 in Petroleum Systems (MMT C02)	3-80
Table 3-46: Recalculations of CH4 in Petroleum Systems (MMT C02 Eq.)	3-80
Table 3-47: HF Oil Well Completions National C02 Emissions (kt C02)	3-80
Table 3-48: Produced Water National CH4 Emissions (Metric Tons CH4)	3-82
Table 3-49: Tanks National C02 Emissions (kt C02)	3-82
Table 3-50: Pneumatic Controller National CH4 Emissions (Metric Tons CH4)	3-82
Table 3-51: Associated Gas Flaring National C02 Emissions (kt C02)	3-83
Table 3-52 Associated Gas Flaring National CH4 Emissions (Metric Tons CH4)	3-83
Table 3-53: Miscellaneous Production Flaring National CH4 Emissions (Metric Tons CH4)	3-84
Table 3-54: Miscellaneous Production Flaring National C02 Emissions (kt C02)	3-84
Table 3-55: Chemical Injection Pump National CH4 Emissions (Metric Tons CH4)	3-84
Table 3-56: Oil Wellhead National CH4 Emissions (Metric Tons CH4)	3-85
Table 3-57: Gas Engine National CH4 Emissions (Metric Tons CH4)	3-85
Table 3-58: National Oil Well Counts	3-85
Table 3-59: Draft Mud Degassing National CH4 Emissions—Not Included in Totals (Metric Tons CH4)	3-87
xi

-------
Table 3-60: Quantity of C02 Captured and Extracted for EOR Operations (kt C02)	3-88
Table 3-61: Geologic Sequestration Information Reported Under GHGRP Subpart RR	3-88
Table 3-62: Total Greenhouse Gas Emissions (CH4, C02, and N20) from Natural Gas Systems (MMT C02 Eq.)	3-91
Table 3-63: CH4 Emissions from Natural Gas Systems (MMT C02 Eq.)a	3-91
Table 3-64: CH4 Emissions from Natural Gas Systems (kt)a	3-92
Table 3-65: Non-combustion C02 Emissions from Natural Gas Systems (MMT)	3-92
Table 3-66: Non-combustion C02 Emissions from Natural Gas Systems (kt)	3-92
Table 3-67: N20 Emissions from Natural Gas Systems (Metric Tons C02 Eq.)	3-93
Table 3-68: N20 Emissions from Natural Gas Systems (Metric Tons N20)	3-93
Table 3-69: Approach 2 Quantitative Uncertainty Estimates for CH4 and Non-combustion C02 Emissions from
Natural Gas Systems (MMT C02 Eq. and Percent)	3-96
Table 3-70: Recalculations of C02 in Natural Gas Systems (MMT C02)	3-98
Table 3-71: Recalculations of CH4 in Natural Gas Systems (MMT C02 Eq.)	3-98
Table 3-72: HF Gas Well Completions National Emissions (Metric Tons CH4)	3-99
Table 3-73: HF Gas Well Completions National Emissions (kt C02)	3-100
Table 3-74: Non-HF Gas Well Completions National Emissions (Metric Tons CH4)	3-100
Table 3-75: Produced Water National CH4 Emissions (Metric Tons CH4)	3-101
Table 3-76: Gathering Stations Sources National CH4 Emissions (Metric Tons CH4)	3-102
Table 3-77: Gathering Stations Flare Stacks National C02 Emissions (Metric Tons C02)	3-102
Table 3-78: Production Segment Pneumatic Controller National Emissions (Metric Tons CH4)	3-102
Table 3-79: Liquids Unloading National Emissions (MetricTons CH4)	3-103
Table 3-80: HF Gas Well Workovers National Emissions (MetricTons CH4)	3-103
Table 3-81: Gas Engine National Emissions (MetricTons CH4)	3-103
Table 3-82: Chemical Injection Pump National Emissions (MetricTons CH4)	3-103
Table 3-83: Kimray Pumps National Emissions (Metric Tons CH4)	3-104
Table 3-84: Compressors National Emissions (Metric Tons CH4)	3-104
Table 3-85: National Gas Well Counts	3-104
Table 3-86: AGR National C02 Emissions (kt C02)	3-105
Table 3-87: Processing Segment Flares National Emissions (kt C02)	3-105
Table 3-88: Processing Segment Reciprocating Compressors National Emissions (Metric Tons CH4)	3-105
Table 3-89: Transmission Station Reciprocating Compressors National Emissions (Metric Tons CH4)	3-106
Table 3-90: Storage Segment Pneumatic Controller National Emissions (Metric Tons CH4)	3-106
Table 3-91: LNG Export Terminal National Emissions (Metric Tons C02)	3-106
Table 3-92: Commercial and Industrial Meter National Emissions (MetricTons CH4)	3-107
Table 3-93: Commercial and Industrial Meter National Emissions (MetricTons C02)	3-107
xii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 3-94: Draft Underground Storage Wells National Emissions (Metric Tons CH4) and Well Counts - Not Included
in Totals	3-108
Table 3-95: Draft Mud Degassing National CH4 Emissions - Not Included in Totals (Metric Tons CH4)	3-110
Table 3-96: CH4 Emissions from Abandoned Oil and Gas Wells (MMT C02 Eq.)	3-111
Table 3-97: CH4 Emissions from Abandoned Oil and Gas Wells (kt)	3-111
Table 3-98: C02 Emissions from Abandoned Oil and Gas Wells (MMT C02)	3-111
Table 3-99: C02 Emissions from Abandoned Oil and Gas Wells (kt)	3-112
Table 3-100: Abandoned Oil Wells Activity Data, CH4 and C02 Emissions (kt)	3-112
Table 3-101: Abandoned Gas Wells Activity Data, CH4 and C02 Emissions (kt)	3-113
Table 3-102: Approach 2 Quantitative Uncertainty Estimates for CH4 and C02 Emissions from Petroleum and
Natural Gas Systems (MMT C02 Eq. and Percent)	3-114
Table 3-103: NOx, CO, and NMVOC Emissions from Energy-Related Activities (kt)	3-116
Table 3-104: C02, CH4, and N20 Emissions from International Bunker Fuels (MMT C02 Eq.)	3-118
Table 3-105: C02, CH4, and N20 Emissions from International Bunker Fuels (kt)	3-118
Table 3-106: Aviation Jet Fuel Consumption for International Transport (Million Gallons)	3-120
Table 3-107: Marine Fuel Consumption for International Transport (Million Gallons)	3-120
Table 3-108: C02 Emissions from Wood Consumption by End-Use Sector (MMT C02 Eq.)	3-122
Table 3-109: C02 Emissions from Wood Consumption by End-Use Sector (kt)	3-122
Table 3-110: C02 Emissions from Ethanol Consumption (MMT C02 Eq.)	3-123
Table 3-111: C02 Emissions from Ethanol Consumption (kt)	3-123
Table 3-112: C02 Emissions from Biodiesel Consumption (MMT C02 Eq.)	3-123
Table 3-113: C02 Emissions from Biodiesel Consumption (kt)	3-124
Table 3-114: Woody Biomass Consumption by Sector (Trillion Btu)	3-124
Table 3-115: Ethanol Consumption by Sector (Trillion Btu)	3-124
Table 3-116: Biodiesel Consumption by Sector (Trillion Btu)	3-125
Table 4-1: Emissions from Industrial Processes and Product Use (MMT C02 Eq.)	4-4
Table 4-2: Emissions from Industrial Processes and Product Use (kt)	4-5
Table 4-3: C02 Emissions from Cement Production (MMT C02 Eq. and kt)	4-11
Table 4-4: Clinker Production (kt)	4-12
Table 4-5: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Cement Production (MMT C02
Eq. and Percent)	4-13
Table 4-6: C02 Emissions from Lime Production (MMT C02 Eq. and kt)	4-16
Table 4-7: Gross, Recovered, and Net C02 Emissions from Lime Production (kt)	4-16
Table 4-8: High-Calcium- and Dolomitic-Quicklime, High-Calcium- and Dolomitic-Hydrated, and Dead-Burned-
Dolomite Lime Production (kt)	4-17
Table 4-9: Adjusted Lime Production (kt)	4-17
xiii

-------
Table 4-10: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Lime Production (MMT C02 Eq.
and Percent)	4-19
Table 4-11: C02 Emissions from Glass Production (MMT C02 Eq. and kt)	4-21
Table 4-12: Limestone, Dolomite, and Soda Ash Consumption Used in Glass Production (kt)	4-23
Table 4-13: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Glass Production (MMT C02 Eq.
and Percent)	4-23
Table 4-14: C02 Emissions from Other Process Uses of Carbonates (MMT C02 Eq.)	4-25
Table 4-15: C02 Emissions from Other Process Uses of Carbonates (kt)	4-26
Table 4-16: Limestone and Dolomite Consumption (kt)	4-27
Table 4-17: Soda Ash Consumption Not Associated with Glass Manufacturing (kt)	4-27
Table 4-18: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Other Process Uses of
Carbonates (MMT C02 Eq. and Percent)	4-28
Table 4-19: C02 Emissions from Ammonia Production (MMT C02 Eq.)	4-30
Table 4-20: C02 Emissions from Ammonia Production (kt)	4-30
Table 4-21: Ammonia Production, Recovered C02 Consumed for Urea Production, and Urea Production (kt).... 4-32
Table 4-22: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Ammonia Production (MMT
C02 Eq. and Percent)	4-33
Table 4-23: C02 Emissions from Urea Consumption for Non-Agricultural Purposes (MMT C02 Eq.)	4-35
Table 4-24: C02 Emissions from Urea Consumption for Non-Agricultural Purposes (kt)	4-35
Table 4-25: Urea Production, Urea Applied as Fertilizer, Urea Imports, and Urea Exports (kt)	4-36
Table 4-26: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Urea Consumption for Non-
Agricultural Purposes (MMT C02 Eq. and Percent)	4-36
Table 4-27: N20 Emissions from Nitric Acid Production (MMT C02 Eq. and kt N20)	4-38
Table 4-28: Nitric Acid Production (kt)	4-40
Table 4-29: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from Nitric Acid Production (MMT
C02 Eq. and Percent)	4-41
Table 4-30: N20 Emissions from Adipic Acid Production (MMT C02 Eq. and kt N20)	4-43
Table 4-31: Adipic Acid Production (kt)	4-44
Table 4-32: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from Adipic Acid Production (MMT
C02 Eq. and Percent)	4-45
Table 4-33: N20 Emissions from Caprolactam Production (MMT C02 Eq. and kt N20)	4-47
Table 4-34: Caprolactam Production (kt)	4-48
Table 4-35: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from Caprolactam, Glyoxal and
Glyoxylic Acid Production (MMT C02 Eq. and Percent)	4-48
Table 4-36
Table 4-37
Table 4-38
C02 and CH4 Emissions from Silicon Carbide Production and Consumption (MMT C02 Eq.)	4-50
C02 and CH4 Emissions from Silicon Carbide Production and Consumption (kt)	4-50
Production and Consumption of Silicon Carbide (Metric Tons)	4-51
xiv Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-39: Approach 2 Quantitative Uncertainty Estimates for CH4 and C02 Emissions from Silicon Carbide
Production and Consumption (MMT C02 Eq. and Percent)	4-52
Table 4-40: C02 Emissions from Titanium Dioxide (MMT C02 Eq. and kt)	4-53
Table 4-41: Titanium Dioxide Production (kt)	4-54
Table 4-42: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Titanium Dioxide Production
(MMT C02 Eq. and Percent)	4-55
Table 4-43: C02 Emissions from Soda Ash Production (MMT C02 Eq. and kt C02)	4-57
Table 4-44: Trona Ore Use (kt)	4-58
Table 4-45: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Soda Ash Production (MMT
C02 Eq. and Percent)	4-58
Table 4-46: C02 and CH4 Emissions from Petrochemical Production (MMT C02 Eq.)	4-61
Table 4-47: C02 and CH4 Emissions from Petrochemical Production (kt)	4-61
Table 4-48: Production of Selected Petrochemicals (kt)	4-63
Table 4-49: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Petrochemical Production and
C02 Emissions from Petrochemical Production (MMT C02 Eq. and Percent)	4-64
Table 4-50: HFC-23 Emissions from HCFC-22 Production (MMT C02 Eq. and kt HFC-23)	4-68
Table 4-51: HCFC-22 Production (kt)	4-69
Table 4-52: Approach 2 Quantitative Uncertainty Estimates for HFC-23 Emissions from HCFC-22 Production (MMT
C02 Eq. and Percent)	4-69
Table 4-53: C02 Emissions from C02 Consumption (MMT C02 Eq. and kt)	4-70
Table 4-54: C02 Production (kt C02) and the Percent Used for Non-EOR Applications	4-72
Table 4-55: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from C02 Consumption (MMT C02
Eq. and Percent)	4-73
Table 4-56: C02 Emissions from Phosphoric Acid Production (MMT C02 Eq. and kt)	4-74
Table 4-57: Phosphate Rock Domestic Consumption, Exports, and Imports (kt)	4-76
Table 4-58: Chemical Composition of Phosphate Rock (Percent by Weight)	4-76
Table 4-59: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Phosphoric Acid Production
(MMT C02 Eq. and Percent)	4-77
Table 4-60:	C02 Emissions from Metallurgical Coke Production (MMT C02 Eq.)	4-79
Table 4-61:	C02 Emissions from Metallurgical Coke Production (kt)	4-79
Table 4-62:	C02 Emissions from Iron and Steel Production (MMT C02 Eq.)	4-79
Table 4-63:	C02 Emissions from Iron and Steel Production (kt)	4-80
Table 4-64:	CH4 Emissions from Iron and Steel Production (MMT C02 Eq.)	4-80
Table 4-65:	CH4 Emissions from Iron and Steel Production (kt)	4-80
Table 4-66:	Material Carbon Contents for Metallurgical Coke Production	4-82
Table 4-67: Production and Consumption Data for the Calculation of C02 Emissions from Metallurgical Coke
Production (Thousand Metric Tons)	4-82
xv

-------
Table 4-68: Production and Consumption Data for the Calculation of C02 Emissions from Metallurgical Coke
Production (Million ft3)	4-82
Table 4-69: Material Carbon Contents for Iron and Steel Production	4-83
Table 4-70: CH4 Emission Factors for Sinter and Pig Iron Production	4-84
Table 4-71: C02 Emission Factors for Sinter Production, Direct Reduced Iron Production, and Pellet Production 4-84
Table 4-72: Production and Consumption Data for the Calculation of C02 and CH4 Emissions from Iron and Steel
Production (Thousand Metric Tons)	4-85
Table 4-73: Production and Consumption Data for the Calculation of C02 Emissions from Iron and Steel Production
(Million ft3 unless otherwise specified)	4-85
Table 4-74: Approach 2 Quantitative Uncertainty Estimates for C02 and CH4 Emissions from Iron and Steel
Production and Metallurgical Coke Production (MMT C02 Eq. and Percent)	4-87
Table 4-75: C02 and CH4 Emissions from Ferroalloy Production (MMT C02 Eq.)	4-89
Table 4-76: C02 and CH4 Emissions from Ferroalloy Production (kt)	4-89
Table 4-77: Production of Ferroalloys (Metric Tons)	4-90
Table 4-78: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Ferroalloy Production (MMT
C02 Eq. and Percent)	4-91
Table 4-79: C02 Emissions from Aluminum Production (MMT C02 Eq. and kt)	4-92
Table 4-80: PFC Emissions from Aluminum Production (MMT C02 Eq.)	4-93
Table 4-81: PFC Emissions from Aluminum Production (kt)	4-94
Table 4-82: Production of Primary Aluminum (kt)	4-97
Table 4-83: Approach 2 Quantitative Uncertainty Estimates for C02 and PFC Emissions from Aluminum Production
(MMT C02 Eq. and Percent)	4-98
Table 4-84: SF6, HFC-134a, FK 5-1-12 and C02 Emissions from Magnesium Production and Processing (MMT C02
Eq.)	4-99
Table 4-85: SF6, HFC-134a, FK 5-1-12 and C02 Emissions from Magnesium Production and Processing (kt)	4-100
Table 4-86: SF6 Emission Factors (kg SF6 per metric ton of magnesium)	4-102
Table 4-87: Approach 2 Quantitative Uncertainty Estimates for SF6, HFC-134a and C02 Emissions from Magnesium
Production and Processing (MMT C02 Eq. and Percent)	4-103
Table 4-88: C02 Emissions from Lead Production (MMT C02 Eq. and kt)	4-105
Table 4-89: Lead Production (Metric Tons)	4-106
Table 4-90: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Lead Production (MMT C02 Eq.
and Percent)	4-107
Table 4-91: C02 Emissions from Zinc Production (MMT C02 Eq. and kt)	4-109
Table 4-92: Zinc Production (Metric Tons)	4-109
Table 4-93: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Zinc Production (MMT C02 Eq.
and Percent)	4-112
Table 4-94
Table 4-95
Table 4-96
PFC, HFC, SF6, NF3, and N20 Emissions from Electronics Manufacture (MMT C02 Eq.)	4-115
PFC, HFC, SF6, NF3, and N20 Emissions from Electronics Manufacture (Metric Tons)	4-116
F-HTF Emissions from Electronics Manufacture by Compound Group (Metric Tons)	4-116
xvi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-97: F-GHGa Emissions from PV and MEMS manufacturing (MMT C02 Eq.)	4-116
Table 4-98: Approach 2 Quantitative Uncertainty Estimates for HFC, PFC, SF6, NF3 and N20 Emissions from
Semiconductor Manufacture (MMT C02 Eq. and Percent)3	4-127
Table 4-99: Emissions of HFCs and PFCs from ODS Substitutes (MMT C02 Eq.)	4-129
Table 4-100: Emissions of HFCs and PFCs from ODS Substitution (Metric Tons)	4-130
Table 4-101: Emissions of HFCs and PFCs from ODS Substitutes (MMT C02 Eq.) by Sector	4-130
Table 4-102: Approach 2 Quantitative Uncertainty Estimates for HFC and PFC Emissions from ODS Substitutes
(MMT C02 Eq. and Percent)	4-133
Table 4-103
Table 4-104
Table 4-105
139
Table 4-106
Table 4-107
Table 4-108
U.S. HFC Supply (MMT C02 Eq.)	4-135
Averaged U.S. HFC Demand (MMTC02 Eq.)	4-136
SF6 Emissions from Electric Power Systems and Electrical Equipment Manufacturers (MMT C02 Eq.) 4-
SF6 Emissions from Electric Power Systems and Electrical Equipment Manufacturers (kt)	4-139
Transmission Mile Coverage (Percent) and Regression Coefficients (kg per mile)	4-143
Approach 2 Quantitative Uncertainty Estimates for SF6 Emissions from Electrical Transmission and
Distribution (MMT C02 Eq. and Percent)	4-145
Table 4-109: N20 Production (kt)	4-146
Table 4-110: N20 Emissions from N20 Product Usage (MMT C02 Eq. and kt)	4-147
Table 4-111: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from N20 Product Usage (MMT
C02 Eq. and Percent)	4-148
Table 4-112: NOx, CO, and NMVOC Emissions from Industrial Processes and Product Use (kt)	4-150
Table 5-1: Emissions from Agriculture (MMT C02 Eq.)	5-3
Table 5-2: Emissions from Agriculture (kt)	5-3
Table 5-3: CH4 Emissions from Enteric Fermentation (MMT C02 Eq.)	5-4
Table 5-4: CH4 Emissions from Enteric Fermentation (kt)	5-5
Table 5-5: Cattle Sub-Population Categories for 2018 Population Estimates	5-8
Table 5-6: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Enteric Fermentation (MMT C02
Eq. and Percent)	5-9
Table 5-7: CH4 and N20 Emissions from Manure Management (MMT C02 Eq.)	5-13
Table 5-8: CH4 and N20 Emissions from Manure Management (kt)	5-14
Table 5-9: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 (Direct and Indirect) Emissions from
Manure Management (MMT C02 Eq. and Percent)	5-18
Table 5-10: IPCC (2006) Implied Emission Factor Default Values Compared with Calculated Values for CH4 from
Manure Management (kg/head/year)	5-19
Table 5-11
Table 5-12
Table 5-13
Table 5-14
CH4 Emissions from Rice Cultivation (MMT C02 Eq.)	5-21
CH4 Emissions from Rice Cultivation (kt)	5-22
Rice Area Harvested (1,000 Hectares)	5-23
Average Ratooned Area as Percent of Primary Growth Area (Percent)	5-24
xvii

-------
Table 5-15: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Rice Cultivation (MMT C02 Eq.
and Percent)	5-26
Table 5-16: N20 Emissions from Agricultural Soils (MMT C02 Eq.)	5-29
Table 5-17: N20 Emissions from Agricultural Soils (kt)	5-29
Table 5-18: Direct N20 Emissions from Agricultural Soils by Land Use Type and N Input Type (MMT C02 Eq.).... 5-29
Table 5-19: Indirect N20 Emissions from Agricultural Soils (MMT C02 Eq.)	5-30
Table 5-20: Quantitative Uncertainty Estimates of N20 Emissions from Agricultural Soil Management in 2019
(MMT C02 Eq. and Percent)	5-44
Table 5-21: Emissions from Liming (MMT C02 Eq.)	5-45
Table 5-22: Emissions from Liming (MMT C)	5-46
Table 5-23: Applied Minerals (MMT)	5-47
Table 5-24: Approach 2 Quantitative Uncertainty Estimates for C02 Emissions from Liming (MMT C02 Eq. and
Percent)	5-48
Table 5-25: C02 Emissions from Urea Fertilization (MMT C02 Eq.)	5-48
Table 5-26: C02 Emissions from Urea Fertilization (MMT C)	5-48
Table 5-27: Applied Urea (MMT)	5-49
Table 5-28: Quantitative Uncertainty Estimates for C02 Emissions from Urea Fertilization (MMT C02 Eq. and
Percent)	5-50
Table 5-29: CH4 and N20 Emissions from Field Burning of Agricultural Residues (MMT C02 Eq.)	5-51
Table 5-30: CH4, N20, CO, and NOx Emissions from Field Burning of Agricultural Residues (kt)	5-52
Table 5-31: Agricultural Crop Production (kt of Product)	5-55
Table 5-32: U.S. Average Percent Crop Area Burned by Crop (Percent)	5-56
Table 5-33: Parameters for Estimating Emissions from Field Burning of Agricultural Residues	5-57
Table 5-34: Greenhouse Gas Emission Ratios and Conversion Factors	5-58
Table 5-35: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 Emissions from Field Burning of
Agricultural Residues (MMT C02 Eq. and Percent)	5-58
Table 6-1: Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and Forestry (MMT C02 Eq.).... 6-3
Table 6-2: Emissions and Removals from Land Use, Land-Use Change, and Forestry by Gas (MMT C02 Eq.)	6-5
Table 6-3: Emissions and Removals from Land Use, Land-Use Change, and Forestry by Gas (kt)	6-6
Table 6-4: Managed and Unmanaged Land Area by Land-Use Categories for All 50 States (Thousands of Hectares)
	6-10
Table 6-5: Land Use and Land-Use Change for the U.S. Managed Land Base for All 50 States (Thousands of
Hectares)	6-10
Table 6-6: Data Sources Used to Determine Land Use and Land Area for the Conterminous United States, Hawaii,
and Alaska	6-16
Table 6-7: Total Land Area (Hectares) by Land-Use Category for U.S. Territories	6-22
Table 6-8: Net C02 Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and Harvested Wood
Pools (MMT CQ2 Eq.)	6-27
xviii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 6-9: Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and Harvested Wood
Pools (MMTC)	6-28
Table 6-10: Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and Harvested Wood Pools
(MMT C)	6-29
Table 6-11: Estimates of C02 (MMT per Year) Emissions from Forest Fires in the Conterminous 48 States and
Alaska3	6-31
Table 6-12: Quantitative Uncertainty Estimates for Net C02 Flux from Forest Land Remaining Forest Land: Changes
in Forest C Stocks (MMT C02 Eq. and Percent)	6-35
Table 6-13: Recalculations of Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C)	6-36
Table 6-14: Recalculations of Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C)	6-37
Table 6-15: Non-C02 Emissions from Forest Fires (MMT C02 Eq.)a	6-38
Table 6-16: Non-C02 Emissions from Forest Fires (kt)a	6-38
Table 6-17: Quantitative Uncertainty Estimates of Non-C02 Emissions from Forest Fires (MMT C02 Eq. and
Percent)3	6-39
Table 6-18: N20 Fluxes from Soils in Forest Land Remaining Forest Land and Land Converted to Forest Land (MMT
C02 Eq. and kt N20)	6-40
Table 6-19: Quantitative Uncertainty Estimates of N20 Fluxes from Soils in Forest Land Remaining Forest Land and
Land Converted to Forest Land (MMT C02 Eq. and Percent)	6-42
Table 6-20: Non-C02 Emissions from Drained Organic Forest Soilsa b (MMT C02 Eq.)	6-43
Table 6-21: Non-C02 Emissions from Drained Organic Forest Soilsa b (kt)	6-43
Table 6-22: States identified as having Drained Organic Soils, Area of Forest on Drained Organic Soils, and
Sampling Error	6-44
Table 6-23: Quantitative Uncertainty Estimates for Non-C02 Emissions on Drained Organic Forest Soils (MMT C02
Eq. and Percent)3	6-45
Table 6-24: Net C02 Flux from Forest C Pools in Land Converted to Forest Land by Land Use Change Category (MMT
C02 Eq.)	6-47
Table 6-25: Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use Change Category (MMT
C)	6-48
Table 6-26: Quantitative Uncertainty Estimates for Forest C Pool Stock Changes (MMT C02 Eq. per Year) in 2019
from Land Converted to Forest Land by Land Use Change	6-50
Table 6-27: Recalculations of the Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use
Change Category (MMT C)	6-52
Table 6-28: Net C02 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT C02 Eq.)	6-54
Table 6-29: Net C02 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT C)	6-55
Table 6-30: Approach 2 Quantitative Uncertainty Estimates for Soil C Stock Changes occurring within Cropland
Remaining Cropland (MMT C02 Eq. and Percent)	6-62
Table 6-31: Area of Managed Land in Cropland Remaining Cropland that is not included in the current Inventory
(Thousand Hectares)	6-64
xix

-------
Table 6-32: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Land Converted to
Cropland by Land Use Change Category (MMT C02 Eq.)	6-66
Table 6-33: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Land Converted to
Cropland (MMT C)	6-66
Table 6-34: Approach 2 Quantitative Uncertainty Estimates for Soil, Dead Organic Matter and Biomass C Stock
Changes occurring within Land Converted to Cropland (MMT C02 Eq. and Percent)	6-69
Table 6-35: Area of Managed Land in Land Converted to Cropland that is not included in the current Inventory
(Thousand Hectares)	6-71
Table 6-36: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Grassland Remaining
Grassland (MMT C02 Eq.)	6-72
Table 6-37: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Grassland Remaining
Grassland (MMT C)	6-72
Table 6-38: Approach 2 Quantitative Uncertainty Estimates for C Stock Changes Occurring Within Grassland
Remaining Grassland (MMT C02 Eq. and Percent)	6-77
Table 6-39: Area of Managed Land in Grassland Remaining Grassland in Alaska that is not included in the current
Inventory (Thousand Hectares)	6-78
Table 6-40: CH4 and N20 Emissions from Biomass Burning in Grassland (MMT C02 Eq.)	6-79
Table 6-41: CH4, N20, CO, and NOx Emissions from Biomass Burning in Grassland (kt)	6-79
Table 6-42: Thousands of Grassland Hectares Burned Annually	6-80
Table 6-43: Uncertainty Estimates for Non-C02 Greenhouse Gas Emissions from Biomass Burning in Grassland
(MMT C02 Eq. and Percent)	6-81
Table 6-44: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Grassland (MMT C02 Eq.)	6-83
Table 6-45: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Grassland (MMT C)	6-83
Table 6-46: Approach 2 Quantitative Uncertainty Estimates for Soil, Dead Organic Matter and Biomass C Stock
Changes occurring within Land Converted to Grassland (MMT C02 Eq. and Percent)	6-87
Table 6-47: Area of Managed Land in Land Converted to Grassland in Alaska that is not included in the current
Inventory (Thousand Hectares)	6-88
Table 6-48: Emissions from Peatlands Remaining Peatlands (MMT C02 Eq.)	6-91
Table 6-49: Emissions from Peatlands Remaining Peatlands (kt)	6-91
Table 6-50: Peat Production of Lower 48 States (kt)	6-92
Table 6-51: Peat Production of Alaska (Thousand Cubic Meters)	6-92
Table 6-52: Peat Production Area of Lower 48 States (Hectares)	6-93
Table 6-53: Peat Production Area of Alaska (Hectares)	6-93
Table 6-54: Peat Production (Hectares)	6-94
Table 6-55: Approach 2 Quantitative Uncertainty Estimates for C02, CH4, and N20 Emissions from Peatlands
Remaining Peatlands (MMT C02 Eq. and Percent)	6-95
Table 6-56: Emissions and Removals from Coastal Wetlands Remaining Coastal Wetlands (MMT C02 Eq.)	6-98
xx Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 6-57: Net C02 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands (MMT C02 Eq.)	6-99
Table 6-58: Net C02 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands (MMT C)	6-99
Table 6-59: CH4 Emissions from Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands (MMT C02
Eq. and kt CH4)	6-99
Table 6-60: Area of Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands, Vegetated Coastal
Wetlands Converted to Unvegetated Open Water Coastal Wetlands, and Unvegetated Open Water Coastal
Wetlands Converted to Vegetated Coastal Wetlands (ha)	6-100
Table 6-61: Aboveground Biomass Carbon Stocks for Vegetated Coastal Wetlands (t C ha	6-100
Table 6-62: Root to Shoot Ratios for Vegetated Coastal Wetlands	6-100
Table 6-63: Annual Soil Carbon Accumulation Rates for Vegetated Coastal Wetlands (t C ha 1 yr-1)	6-101
Table 6-64: IPCC Approach 1 Quantitative Uncertainty Estimates for C Stock Changes and CH4 Emissions occurring
within Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands in 1990 (MMT C02 Eq. and Percent). 6-
102
Table 6-65: IPCC Approach 1 Quantitative Uncertainty Estimates for C Stock Changes and CH4 Emissions occurring
within Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands in 2019 (MMT C02 Eq. and Percent). 6-
102
Table 6-66: Net C02 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to Unvegetated Open
Water Coastal Wetlands (MMT C02 Eq.)	6-105
Table 6-67: Net C02 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to Unvegetated Open
Water Coastal Wetlands (MMT C)	6-105
Table 6-68: Approach 1 Quantitative Uncertainty Estimates for C02 Flux Occurring within Vegetated Coastal
Wetlands Converted to Unvegetated Open Water Coastal Wetlands in 1990 (MMT C02 Eq. and Percent)	6-107
Table 6-69: Approach 1 Quantitative Uncertainty Estimates for C02 Flux Occurring within Vegetated Coastal
Wetlands Converted to Unvegetated Open Water Coastal Wetlands in 2019 (MMT C02 Eq. and Percent)	6-107
Table 6-70: C02 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands Converted to
Vegetated Coastal Wetlands (MMT C02 Eq.)	6-109
Table 6-71: C02 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands Converted to
Vegetated Coastal Wetlands (MMT C)	6-109
Table 6-72: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes Occurring within Unvegetated
Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands in 1990 (MMT C02 Eq. and Percent)6-lll
Table 6-73: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes Occurring within Unvegetated
Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands in 2019 (MMT C02 Eq. and Percent) 6-112
Table 6-74: N20 Emissions from Aquaculture in Coastal Wetlands (MMT C02 Eq. and kt N20)	6-113
Table 6-75: Approach 1 Quantitative Uncertainty Estimates for N20 Emissions from Aquaculture Production in
Coastal Wetlands in 1990 (MMT C02 Eq. and Percent)	6-114
Table 6-76: Approach 1 Quantitative Uncertainty Estimates for N20 Emissions from Aquaculture Production in
Coastal Wetlands in 2019 (MMT C02 Eq. and Percent)	6-114
Table 6-77: Net C02 Flux from C Stock Changes in Land Converted to Vegetated Coastal Wetlands (MMT C02 Eq.) 6-
116
Table 6-78: Net C02 Flux from C Stock Changes in Land Converted to Vegetated Coastal Wetlands (MMT C)... 6-116

-------
Table 6-79: CH4 Emissions from Land Converted to Vegetated Coastal Wetlands (MMT C02 Eq. and kt CH4).... 6-117
Table 6-80: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes occurring within Land Converted
to Vegetated Coastal Wetlands (MMT C02 Eq. and Percent)	6-120
Table 6-81: Net C02 Flux from Soil C Stock Changes in Settlements Remaining Settlements (MMT C02 Eq.)	6-122
Table 6-82: Net C02 Flux from Soil C Stock Changes in Settlements Remaining Settlements (MMT C)	6-123
Table 6-83: Thousands of Hectares of Drained Organic Soils in Settlements Remaining Settlements	6-123
Table 6-84: Uncertainty Estimates for C02 Emissions from Drained Organic Soils in Settlements Remaining
Settlements (MMT C02 Eq. and Percent)	6-124
Table 6-85: Area of Managed Land in Settlements Remaining Settlements that is not included in the current
Inventory (Thousand Hectares)	6-124
Table 6-86: Net Flux from Trees in Settlements Remaining Settlements (MMT C02 Eq. and MMT C)a	6-125
Table 6-87: Carbon Storage (kg C/m2 tree cover), Gross and Net Sequestration (kg C/m2 tree cover/year) and Tree
Cover (percent) among Sampled U.S. Cities (see Nowak et al. 2013)	6-128
Table 6-88: Estimated Annual C Sequestration (Metric Tons C/Year), Tree Cover (Percent), and Annual C
Sequestration per Area of Tree Cover (kg C/m2/year) for settlement areas in United States by State and the District
of Columbia (2019)	6-130
Table 6-89: Approach 2 Quantitative Uncertainty Estimates for Net C02 Flux from Changes in C Stocks in
Settlement Trees (MMT C02 Eq. and Percent)	6-132
Table 6-90: N20 Emissions from Soils in Settlements Remaining Settlements (MMT C02 Eq. and kt N20)	6-133
Table 6-91: Quantitative Uncertainty Estimates of N20 Emissions from Soils in Settlements Remaining Settlements
(MMT C02 Eq. and Percent)	6-135
Table 6-92: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills (MMT C02 Eq.)	6-136
Table 6-93: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills (MMT C)	6-136
Table 6-94: Moisture Contents, C Storage Factors (Proportions of Initial C Sequestered), Initial C Contents, and
Decay Rates for Yard Trimmings and Food Scraps in Landfills	6-139
Table 6-95: C Stocks in Yard Trimmings and Food Scraps in Landfills (MMT C)	6-139
Table 6-96: Approach 2 Quantitative Uncertainty Estimates for C02 Flux from Yard Trimmings and Food Scraps in
Landfills (MMT C02 Eq. and Percent)	6-140
Table 6-97: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Settlements (MMT C02 Eq.)	6-142
Table 6-98: Net C02 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Settlements (MMT C)	6-143
Table 6-99: Approach 2 Quantitative Uncertainty Estimates for Soil, Dead Organic Matter and Biomass C Stock
Changes occurring within Land Converted to Settlements (MMT C02 Eq. and Percent)	6-145
Table 6-100: Area of Managed Land in Settlements Remaining Settlements that is not included in the current
Inventory (Thousand Hectares)	6-147
Table 7-1
Table 7-2
Table 7-3
Table 7-4
Emissions from Waste (MMT C02 Eq.)	7-2
Emissions from Waste (kt)	7-2
CH4 Emissions from Landfills (MMT C02 Eq.)	7-6
CH4 Emissions from Landfills (kt)	7-7
xxii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 7-5: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Landfills (MMT C02 Eq. and
Percent)	7-15
Table 7-6: Materials Discarded in the Municipal Waste Stream by Waste Type from 1990 to 2018 (Percent)	7-19
Table 7-7: CH4and N20 Emissions from Domestic and Industrial Wastewater Treatment (MMT C02 Eq.)	7-22
Table 7-8: CH4and N20 Emissions from Domestic and Industrial Wastewater Treatment (kt)	7-23
Table 7-9: Domestic Wastewater CH4 Emissions from Septic and Centralized Systems (2019, kt, MMT C02 Eq. and
Percent)	7-24
Table 7-10:	Variables and Data Sources for CH4 Emissions from Septic Systems	7-24
Table 7-11:	Variables and Data Sources for Organics in Domestic Wastewater	7-25
Table 7-12:	U.S. Population (Millions) and Domestic Wastewater BOD5 Produced (kt)	7-26
Table 7-13:	Variables and Data Sources for Organics in Centralized Domestic Wastewater	7-26
Table 7-14: Variables and Data Sources for CH4 Emissions from Centrally Treated Aerobic Systems (Other than
Constructed Wetlands)	7-27
Table 7-15: Variables and Data Sources for CH4 Emissions from Centrally Treated Aerobic Systems (Constructed
Wetlands)	7-28
Table 7-16: Variables and Data Sources for CH4 Emissions from Centrally Treated Anaerobic Systems	7-29
Table 7-17: Variables and Data Sources for Emissions from Anaerobic Sludge Digesters	7-30
Table 7-18: Variables and Data Sources for CH4 Emissions from Centrally Treated Systems Discharge	7-31
Table 7-19: Total Industrial Wastewater CH4 Emissions by Sector (2019, MMT C02 Eq. and Percent)	7-32
Table 7-20: U.S. Pulp and Paper, Meat, Poultry, Vegetables, Fruits and Juices, Ethanol, Breweries, and Petroleum
Refining Production (MMT)	7-34
Table 7-21: U.S. Industrial Wastewater Characteristics Data (2019)	7-34
Table 7-22: U.S. Industrial Wastewater Treatment Activity Data	7-35
Table 7-23: Sludge Variables for Aerobic Treatment Systems	7-35
Table 7-24: Fraction of TOW Removed During Treatment by Industry	7-36
Table 7-25: Wastewater Outflow (m3/ton) for Pulp, Paper, and Paperboard Mills	7-37
Table 7-26: Wastewater Outflow (m3/ton) and BOD Production (g/L) for U.S. Vegetables, Fruits, and Juices
Production	7-37
Table 7-27: Domestic Wastewater N20 Emissions from Septic and Centralized Systems (2019, kt, MMT C02 Eq. and
Percent)	7-40
Table 7-28: Variables and Data Sources for Protein Consumed	7-40
Table 7-29: Variables and Data Sources for N20 Emissions from Septic System	7-41
Table 7-30: Variables and Data Sources for Non-Consumed Protein and Nitrogen Entering Centralized Systems 7-42
Table 7-31: Variables and Data Sources for N20 Emissions from Centrally Treated Aerobic Systems (Other than
Constructed Wetlands)	7-43
Table 7-32: Variables and Data Sources for N20 Emissions from Centrally Treated Aerobic Systems (Constructed
Wetlands)	7-43
Table 7-33: Variables and Data Sources for N20 Emissions from Centrally Treated Anaerobic Systems	7-44
xxiii

-------
Table 7-34: U.S. Population (Millions) Fraction of Population Served by Centralized Wastewater Treatment
(percent), Protein Supply (kg/person-year), and Protein Consumed (kg/person-year)	7-45
Table 7-35: Variables and Data Sources for N20 Emissions from Centrally Treated Systems Discharge	7-46
Table 7-36: Total Industrial Wastewater N20 Emissions by Sector (2019, MMT C02 Eq. and Percent)	7-47
Table 7-37: U.S. Industrial Wastewater Nitrogen Data	7-48
Table 7-38: Industrial Wastewater Nitrogen Discharged in 2018 by Sector (kg N)	7-49
Table 7-39: Approach 2 Quantitative Uncertainty Estimates for 2019 Emissions from Wastewater Treatment (MMT
C02 Eq. and Percent)	7-50
Table 7-40: Approach 2 Quantitative Uncertainty Estimates for 1990 Emissions from Wastewater Treatment (MMT
C02 Eq. and Percent)	7-50
Table 7-41: CH4 and N20 Emissions from Composting (MMT C02 Eq.)	7-54
Table 7-42: CH4 and N20 Emissions from Composting (kt)	7-54
Table 7-43: U.S. Waste Composted (kt)	7-55
Table 7-44: Tier 1 Quantitative Uncertainty Estimates for Emissions from Composting (MMT C02 Eq. and Percent)
	7-56
Table 7-45: CH4 Emissions from Stand-Alone Anaerobic Digestion (MMT C02 Eq.)	7-58
Table 7-46: CH4 Emissions from Stand-Alone Anaerobic Digestion (kt)	7-58
Table 7-47: U.S. Waste Digested (kt)	7-60
Table 7-48: Estimated Number of Stand-Alone AD Facilities Operating from 1990-20191	7-61
Table 7-49: Estimated Biogas Produced and Methane Recovered from Stand-Alone AD Facilities Operating from
1990-20191	7-61
Table 7-50: Tier 1 Quantitative Uncertainty Estimates for Emissions from Digestion (MMT C02 Eq. and Percent)7-62
Table 7-51: Emissions of NOx, CO, and NMVOC from Waste (kt)	7-64
Table 9-1: Revisions to U.S. Greenhouse Gas Emissions (MMT C02 Eq.)	9-5
Table 9-2: Revisions to U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change,
and Forestry (MMT C02 Eq.)	9-7
Figures
Figure ES-1: U.S. Greenhouse Gas Emissions by Gas	ES-5
Figure ES-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the Previous Year	ES-5
Figure ES-3: Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990 (1990=0)	ES-6
Figure ES-4: 2019 U.S. Greenhouse Gas Emissions by Gas (Percentages based on MMT C02 Eq.)	ES-10
Figure ES-5: 2019 Sources of C02 Emissions	ES-11
Figure ES-6: 2019 C02 Emissions from Fossil Fuel Combustion by Sector and Fuel Type	ES-12
Figure ES-7: 2019 End-Use Sector Emissions of C02 from Fossil Fuel Combustion	ES-13
Figure ES-8: Electric Power Generation and Emissions	ES-15
Figure ES-9: 2019 Sources of CH4 Emissions	ES-16
Figure ES-10: 2019 Sources of N20 Emissions	ES-17
xxiv Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure ES-11: 2019 Sources of HFCs, PFCs, SF6, and NF3 Emissions	ES-19
Figure ES-12: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector	ES-20
Figure ES-13: 2019 U.S. Energy Consumption by Energy Source (Percent)	ES-22
Figure ES-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors	ES-26
Figure ES-15: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed to Economic Sectors
	ES-28
Figure ES-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic Product (GDP)	ES-29
Figure ES-17: 2019 Key Categories	ES-30
Figure 1-1: National Inventory Arrangements and Process Diagram	1-12
Figure 1-2: U.S. QA/QC Plan Summary	1-24
Figure 2-1: U.S. Greenhouse Gas Emissions by Gas	2-2
Figure 2-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the Previous Year	2-2
Figure 2-3: Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990 (1990=0)	2-3
Figure 2-4: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector	2-8
Figure 2-5: 2019 Energy Chapter Greenhouse Gas Sources	2-11
Figure 2-6: Trends in Energy Chapter Greenhouse Gas Sources	2-11
Figure 2-7: 2019 C02 Emissions from Fossil Fuel Combustion by Sector and Fuel Type	2-15
Figure 2-8: 2019 End-Use Sector Emissions of C02 from Fossil Fuel Combustion	2-15
Figure 2-9: Electric Power Generation (Billion kWh) and Emissions	2-16
Figure 2-10: 2019 Industrial Processes and Product Use Chapter Greenhouse Gas Source	2-18
Figure 2-11: Trends in Industrial Processes and Product Use Chapter Greenhouse Gas Sources	2-19
Figure 2-12: 2019 Agriculture Chapter Greenhouse Gas Sources	2-21
Figure 2-13: Trends in Agriculture Chapter Greenhouse Gas Sources	2-22
Figure 2-14: 2019 LULUCF Chapter Greenhouse Gas Sources and Sinks	2-24
Figure 2-15: Trends in Emissions and Removals (Net C02 Flux) from Land Use, Land-Use Change, and Forestry3. 2-25
Figure 2-16: 2019 Waste Sector Greenhouse Gas Sources	2-27
Figure 2-17: Trends in Waste Chapter Greenhouse Gas Sources	2-28
Figure 2-18: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors	2-30
Figure 2-19: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed to Economic Sectors.. 2-
35
Figure 2-20: Trends in Transportation-Related Greenhouse Gas Emissions	2-38
Figure 2-21: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic Product	2-41
Figure 3-1: 2019 Energy Chapter Greenhouse Gas Sources	3-2
Figure 3-2: Trends in Energy Chapter Greenhouse Gas Sources	3-2
Figure 3-3: 2019 U.S. Fossil Carbon Flows	3-3
Figure 3-4: 2019 U.S. Energy Use by Energy Source	3-10
xxv

-------
Figure 3-5: Annual U.S. Energy Use	3-10
Figure 3-6: 2019 C02 Emissions from Fossil Fuel Combustion by Sector and Fuel Type	3-11
Figure 3-7: Annual Deviations from Normal Heating Degree Days for the United States (1950-2019, Index Normal =
100)	3-12
Figure 3-8: Annual Deviations from Normal Cooling Degree Days for the United States (1950-2019, Index Normal =
100)	3-12
Figure 3-9: Fuels Used in Electric Power Generation and Total Electric Power Sector C02 Emissions	3-18
Figure 3-10: Electric Power Retail Sales by End-Use Sector	3-19
Figure 3-11: Industrial Production Indices (Index 2012=100)	3-20
Figure 3-12: Fuels and Electricity Used in Industrial Sector, Industrial Output, and Total Sector C02 Emissions
(Including Electricity)	3-21
Figure 3-13: Fuels and Electricity Used in Residential and Commercial Sectors, Heating and Cooling Degree Days,
and Total Sector C02 Emissions (Including Electricity)	3-22
Figure 3-14:	Fuels Used in Transportation Sector, Onroad VMT, and Total Sector C02 Emissions	3-24
Figure 3-15:	Sales-Weighted Fuel Economy of New Passenger Cars and Light-Duty Trucks, 1990-2019	3-26
Figure 3-16:	Sales of New Passenger Cars and Light-Duty Trucks, 1990-2019	3-27
Figure 3-17:	Mobile Source CH4 and N20 Emissions	3-30
Figure 3-18:	U.S. Energy Consumption and Energy-Related C02 Emissions Per Capita and Per Dollar GDP	3-36
Figure 4-1
Figure 4-2
Figure 4-3
Figure 5-1
Figure 5-2
Figure 5-3
Figure 5-4
Figure 5-5
Figure 5-6
Figure 5-7
Figure 5-8
33
2019 Industrial Processes and Product Use Chapter Greenhouse Gas Sources	4-2
Trends in Industrial Processes and Product Use Chapter Greenhouse Gas Sources	4-3
U.S. HFC Consumption (MMT C02 Eq.)	4-135
2019 Agriculture Chapter Greenhouse Gas Emission Sources	5-1
Trends in Agriculture Chapter Greenhouse Gas Emission Sources	5-2
Annual CH4 Emissions from Rice Cultivation, 2015	5-22
Sources and Pathways of N that Result in N20 Emissions from Agricultural Soil Management	5-28
Croplands, 2015 Annual Direct N20 Emissions Estimated Using the Tier 3 DayCent Model	5-31
Grasslands, 2015 Annual Direct N20 Emissions Estimated Using the Tier 3 DayCent Model	5-32
Croplands, 2015 Annual Indirect N20 Emissions from Volatilization Using the Tier 3 DayCent Model 5-33
Grasslands, 2015 Annual Indirect N20 Emissions from Volatilization Using the Tier 3 DayCent Model.. 5-
Figure 5-9: Croplands, 2015 Annual Indirect N20 Emissions from Leaching and Runoff Using the Tier 3 DayCent
Model	5-34
Figure 5-10: Grasslands, 2015 Annual Indirect N20 Emissions from Leaching and Runoff Using the Tier 3 DayCent
Model	5-34
Figure 6-1: 2019 LULUCF Chapter Greenhouse Gas Sources and Sinks	6-2
Figure 6-2: Trends in Emissions and Removals (Net C02 Flux) from Land Use, Land-Use Change, and Forestry3	6-3
Figure 6-3: Percent of Total Land Area for Each State in the General Land-Use Categories for 2019	6-12
xxvi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 6-4: Changes in Forest Area by Region for Forest Land Remaining Forest Land in the conterminous United
States and Alaska (1990-2019)	6-26
Figure 6-5: Estimated Net Annual Changes in C Stocks for All C Pools in Forest Land Remaining Forest Land in the
Conterminous U.S. and Alaska (1990-2019)	6-30
Figure 6-6: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural Management within States,
2015, Cropland Remaining Cropland	6-56
Figure 6-7: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural Management within States,
2015, Cropland Remaining Cropland	6-57
Figure 6-8: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural Management within States,
2015, Grassland Remaining Grassland	6-73
Figure 6-9: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural Management within States,
2015, Grassland Remaining Grassland	6-74
Figure 7-1: 2019 Waste Chapter Greenhouse Gas Sources	7-1
Figure 7-2: Trends in Waste Chapter Greenhouse Gas Sources	7-2
Figure 7-3: Methodologies Used Across the Time Series to Compile the U.S. Inventory of Emission Estimates for
MSW Landfills	7-9
Figure 7-4
Figure 7-5
Figure 7-6
Figure 9-1
Management of Municipal Solid Waste in the United States, 2018	7-18
MSW Management Trends from 1990 to 2018	7-19
Percent of Degradable Materials Diverted from Landfills from 1990 to 2018 (Percent)	7-20
Impacts from Recalculations to U.S. Greenhouse Gas Emissions by Sector	9-4
Box ES-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program	ES-2
Box ES-2: Improvements and Recalculations Relative to the Previous Inventory	ES-6
Box ES-3: Trends in Various U.S. Greenhouse Gas Emissions-Related Data	ES-28
Box ES-4: Use of Ambient Measurements Systems for Validation of Emission Inventories	ES-31
Box 1-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program	1-2
Box 1-2: The IPCC Fifth Assessment Report and Global Warming Potentials	1-10
Box 1-3: Use of IPCC Reference Approach to support Verification of Emissions from Fossil Fuel Combustion	1-25
Box 2-1: Methodology for Aggregating Emissions by Economic Sector	2-32
Box 2-2: Trends in Various U.S. Greenhouse Gas Emissions-Related Data	2-40
Box 2-3: Sources and Effects of Sulfur Dioxide	2-43
Box 3-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program	3-5
Box 3-2: Weather and Non-Fossil Energy Effects on C02 Emissions from Fossil Fuel Combustion Trends	3-11
Box 3-3: Uses of Greenhouse Gas Reporting Program Data and Improvements in Reporting Emissions from
Industrial Sector Fossil Fuel Combustion	3-21
xxvii

-------
Box 3-4: Carbon Intensity of U.S. Energy Consumption	3-35
Box 3-5: Reporting of Lubricants, Waxes, and Asphalt and Road Oil Product Use in Energy Sector	3-52
Box 3-6: Carbon Dioxide Transport, Injection, and Geological Storage	3-87
Box 4-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals	4-7
Box 4-2: Industrial Process and Product Use Data from EPA's Greenhouse Gas Reporting Program	4-9
Box 5-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals	5-3
Box 5-2: Surrogate Data Method	5-25
Box 5-3: Tier 1 vs. Tier 3 Approach for Estimating N20 Emissions	5-36
Box 5-4: Surrogate Data Method	5-37
Box 5-5: Comparison of the Tier 2 U.S. Inventory Approach and IPCC (2006) Default Approach	5-46
Box 5-6: Comparison of Tier 2 U.S. Inventory Approach and IPCC (2006) Default Approach	5-54
Box 6-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals	6-8
Box 6-2: Preliminary Estimates of Land Use in U.S. Territories	6-22
Box 6-3: C02 Emissions from Forest Fires	6-30
Box 6-4: Surrogate Data Method	6-58
Box 6-5: Tier 3 Approach for Soil C Stocks Compared to Tier 1 or 2 Approaches	6-59
Box 7-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to Greenhouse Gas Reporting Data	7-3
Box 7-2: Description of a Modern, Managed Landfill in the United States	7-4
Box 7-3: Nationwide Municipal Solid Waste Data Sources	7-12
Box 7-4: Overview of U.S. Solid Waste Management Trends	7-18
xxviii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Executive Summary
An emissions inventory that identifies and quantifies a country's anthropogenic1 sources and sinks of greenhouse
gases is essential for addressing climate change. This inventory adheres to both (1) a comprehensive and detailed
set of methodologies for estimating national sources and sinks of anthropogenic greenhouse gases, and (2) a
common and consistent format that enables Parties to the United Nations Framework Convention on Climate
Change (UNFCCC) to compare the relative contribution of different emission sources and greenhouse gases to
climate change.
In 1992, the United States signed and ratified the UNFCCC. As stated in Article 2 of the UNFCCC, "The ultimate
objective of this Convention and any related legal instruments that the Conference of the Parties may adopt is to
achieve, in accordance with the relevant provisions of the Convention, stabilization of greenhouse gas
concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the
climate system. Such a level should be achieved within a time-frame sufficient to allow ecosystems to adapt
naturally to climate change, to ensure that food production is not threatened and to enable economic
development to proceed in a sustainable manner."2
As a signatory to the UNFCCC, consistent with Article 43 and decisions at the First, Second, Fifth, and Nineteenth
Conference of Parties,4 the United States is committed to submitting a national inventory of anthropogenic
sources and sinks of greenhouse gases to the UNFCCC by April 15 of each year. The United States views this report,
in conjunction with Common Reporting Format (CRF) reporting tables that accompany this report, as an
opportunity to fulfill this annual commitment under the UNFCCC.
This executive summary provides the latest information on U.S. anthropogenic greenhouse gas emission trends
from 1990 through 2019. The structure of this report is consistent with the UNFCCC guidelines for inventory
reporting, as discussed in Box ES-1.5
1	The term "anthropogenic," in this context, refers to greenhouse gas emissions and removals that are a direct result of human
activities or are the result of natural processes that have been affected by human activities (IPCC 2006).
2	Article 2 of the Framework Convention on Climate Change published by the UNEP/WMO Information Unit on Climate Change.
See .
3	Article 4(l)(a) of the United Nations Framework Convention on Climate Change (also identified in Article 12) and subsequent
decisions by the Conference of the Parties elaborated the role of Annex I Parties in preparing national inventories. Article 4
states "Parties to the Convention, by ratifying, shall develop, periodically update, publish and make available...national
inventories of anthropogenic emissions by sources and removals by sinks of all greenhouse gases not controlled by the
Montreal Protocol, using comparable methodologies..." See  for more information.
4	See UNFCCC decisions 3/CP.l, 9/CP.2, 3/CP.5, and 24/CP.19 at .
5	See .
Executive Summary ES-1

-------
Box ES-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program
In following the UNFCCC requirement under Article 4.1 and related decisions to develop and submit annual
national greenhouse gas emission inventories, the emissions and removals presented in this report and this
chapter are organized by source and sink categories and calculated using internationally-accepted methods
provided by the IPCC in the 2006IPCC Guidelines for National Greenhouse Gas Inventories (2006IPCC
Guidelines) and where appropriate, its supplements and refinements. Additionally, the calculated emissions and
removals in a given year for the United States are presented in a common manner in line with the UNFCCC
reporting guidelines for the reporting of inventories under this international agreement. The use of consistent
methods to calculate emissions and removals by all nations providing their inventories to the UNFCCC ensures
that these reports are comparable. The presentation of emissions and removals provided in this Inventory does
not preclude alternative examinations, but rather this Inventory presents emissions and removals in a common
format consistent with how countries are to report Inventories under the UNFCCC. The report itself, and this
chapter, follows this standardized format, and provides an explanation of the application of methods used to
calculate emissions and removals.
EPA also collects greenhouse gas emissions data from individual facilities and suppliers of certain fossil fuels and
industrial gases through its Greenhouse Gas Reporting Program (GHGRP).6 The GHGRP applies to direct
greenhouse gas emitters, fossil fuel suppliers, industrial gas suppliers, and facilities that inject carbon dioxide
(C02) underground for sequestration or other reasons and requires reporting by over 8,000 sources or suppliers
in 41 industrial categories.7 Annual reporting is at the facility level, except for certain suppliers of fossil fuels and
industrial greenhouse gases. In general, the threshold for reporting is 25,000 metric tons or more of C02 Eq. per
year. Facilities in most source categories subject to GHGRP began reporting for the 2010 reporting year while
additional types of industrial operations began reporting for reporting year 2011. While the GHGRP does not
provide full coverage of total annual U.S. greenhouse gas emissions and sinks (e.g., the GHGRP excludes
emissions from the agricultural, land use, and forestry sectors), it is an important input to the calculations of
national-level emissions in the Inventory.
Data presented in this Inventory report and EPA's Greenhouse Gas Reporting Program (GHGRP) are
complementary. The GHGRP dataset continues to be an important resource for the Inventory, providing not
only annual emissions information, but also other annual information such as activity data and emission factors
that can improve and refine national emission estimates and trends over time. Methodologies used in EPA's
GHGRP are consistent with the 2006 IPCC Guidelines (e.g., higher tier methods). GHGRP data also allow EPA to
disaggregate national inventory estimates in new ways that can highlight differences across regions and sub-
categories of emissions, along with enhancing application of QA/QC procedures and assessment of
uncertainties. EPA uses annual GHGRP data in a number of categories to improve the national estimates
presented in this Inventory consistent with IPCC methodological guidance. See Annex 9 for more information on
specific uses of GHGRP data in the Inventory (e.g., natural gas systems).
6	On October 30, 2009 the EPA promulgated a rule requiring annual reporting of greenhouse gas data from large greenhouse
gas emissions sources in the United States. Implementation of the rule, codified at 40 CFR Part 98, is referred to as EPA's
Greenhouse Gas Reporting Program (GHGRP).
7	See  and .
ES-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
ES.l Background Information
Greenhouse gases absorb infrared radiation, thereby trapping heat in the atmosphere and making the planet
warmer. The most important greenhouse gases directly emitted by humans include carbon dioxide (C02), methane
(CH4), nitrous oxide (N20), and several fluorine-containing halogenated substances (HFCs, PFCs, SF6 and NF3).
Although C02, CH4, and N20 occur naturally in the atmosphere, human activities have changed their atmospheric
concentrations. From the pre-industrial era (i.e., ending about 1750) to 2019, concentrations of these greenhouse
gases have increased globally by 47,167, and 23 percent, respectively (IPCC 2013; NOAA/ESRL 2021a, 2021b,
2021c). This annual report estimates the total national greenhouse gas emissions and removals associated with
human activities across the United States.
Global Warming Potentials
Gases in the atmosphere can contribute to climate change both directly and indirectly. Direct effects occur when
the gas itself absorbs radiation. Indirect radiative forcing occurs when chemical transformations of the substance
produce other greenhouse gases, when a gas influences the atmospheric lifetimes of other gases, and/or when a
gas affects atmospheric processes that alter the radiative balance of the earth (e.g., affect cloud formation or
albedo).8 The IPCC developed the Global Warming Potential (GWP) concept to compare the ability of a greenhouse
gas to trap heat in the atmosphere relative to another gas.
The GWP of a greenhouse gas is defined as the ratio of the accumulated radiative forcing within a specific time
horizon caused by emitting 1 kilogram of the gas, relative to that of the reference gas C02 (IPCC 2013). Therefore
GWP-weighted emissions are provided in million metric tons of C02 equivalent (MMT C02 Eq.).9,10 Estimates for all
gases in this Executive Summary are presented in units of MMT C02 Eq. Emissions by gas in unweighted mass
kilotons are provided in the Trends chapter of this report and in the Common Reporting Format (CRF) tables that
are also part of the submission to the UNFCCC.
UNFCCC reporting guidelines for national inventories require the use of GWP values from the IPCC Fourth
Assessment Report (AR4) (IPCC 2007).11 All estimates are provided throughout the report in both C02 equivalents
and unweighted units. A comparison of emission values using the AR4 GWP values versus the IPCC Second
Assessment Report (SAR) (IPCC 1996), and the IPCC Fifth Assessment Report (AR5) (IPCC 2013) GWP values can be
found in Chapter 1 and, in more detail, in Annex 6.1 of this report. The GWP values used in this report are listed
below in Table ES-1.
Table ES-1: Global Warming Potentials (100-Year Time Horizon) Used in this Report
Gas	GWP
C02	1
CH4a	25
N20	298
HFC-23	14,800
HFC-32	675
HFC-41	92
HFC-125	3,500
8	Albedo is a measure of the Earth's reflectivity and is defined as the fraction of the total solar radiation incident on a body that
is reflected by it.
9	Carbon comprises 12/44 of carbon dioxide by weight.
10	One million metric ton is equal to 1012 grams or one teragram.
11	See .
Executive Summary ES-3

-------
HFC-134a
HFC-143a
HFC-152a
HFC-227ea
HFC-236fa
HFC-43-10mee
HFC-245fa
HFC-365mfc
CF4
c2f6
C3Fs
c-CsFs
C4F10
c-C4Fs
C5F12
C6F14
sf6
NFs
Other Fluorinated Gases
22,800
17,200
See Annex 6
10,300
12,200
1,430
4,470
124
3,220
9,810
1,640
1,030
794
7,390
8,830
1.97
8,860
9,160
9,300
a The GWP of CH4 includes the direct effects
and those indirect effects due to the
production of tropospheric ozone and
stratospheric water vapor. The indirect
effect due to production of C02 is not
included. See Annex 6 for additional
information.
Source: IPCC(2007).
ES.2 Recent Trends in U.S. Greenhouse Gas
Emissions and Sinks
In 2019, total gross U.S. greenhouse gas emissions were 6,558.3 million metric tons of carbon dioxide equivalent
(MMT C02 Eq).12 Total U.S. emissions have increased by 1.8 percent from 1990 to 2019, down from a high of 15.6
percent above 1990 levels in 2007. Emissions decreased from 2018 to 2019 by 1.7 percent (113.1 MMT C02 Eq.).
Net emissions (including sinks) were 5,769.1 MMT C02 Eq. Overall, net emissions decreased 1.7 percent from 2018
to 2019 and decreased 13.0 percent from 2005 levels as shown in Table ES-2. The decline reflects the combined
impacts of many long-term trends, including population, economic growth, energy market trends, technological
changes including energy efficiency, and carbon intensity of energy fuel choices. Between 2018 and 2019, the
decrease in total greenhouse gas emissions was largely driven by the decrease in C02 emissions from fossil fuel
combustion. The decrease in C02 emissions from fossil fuel combustion was a result of a 1 percent decrease in
total energy use and reflects a continued shift from coal to less carbon intensive natural gas and renewables in the
electric power sector.
Figure ES-1 through Figure ES-3 illustrate the overall trends in total U.S. emissions by gas, annual percent changes,
and relative change since 1990 for each year of the time series, and Table ES-2 provides information on trends in
12 The gross emissions total presented in this report for the United States excludes emissions and removals from Land Use,
Land-Use Change, and Forestry (LULUCF). The net emissions total presented in this report for the United States includes
emissions and removals from LULUCF.
ES-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
gross U.S. greenhouse gas emissions and sinks for 1990 through 2019. Unless otherwise stated, all tables and
figures provide total gross emissions and exclude the greenhouse gas fluxes from the Land Use, Land-Use Change,
and Forestry (LULUCF) sector. For more information about the LULUCF sector see Section ES.3 Overview of Sector
Emissions and Trends.
Figure ES-1: U.S. Greenhouse Gas Emissions by Gas
9 000 HFCs, PFCs, SF6 and NF3
I Nitrous Oxide
¦	Methane
8,000 _ Carbon Dioxide
¦	Net COz Flux from LULUCF3
7,000
6,000
I Net Emissions (including sinks)
3,000
2,000
1,000
5,000
u 4,000
l-
z
2:
-1,000
¦Sl"
cr»
cn
00
en
cn
O i-H
o o
o o
¦sJ"
O
o
00
o
o
rNr\ir\JrNrMrsJrMrNrMrsjrMr\ir\ir\Jf\ir\Jr\irMr\lcM
a The term "flux" is used to describe the exchange of CO2 to and from the atmosphere, with net flux being either positive or
negative depending on the overall balance. Removal and long-term storage of CO2 from the atmosphere is also referred to as
"carbon sequestration."
Figure ES-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the
Previous Year
4%
2%
0%
-2%
-4%
3.0%
3.2%
2.6%
2.7%
2.9%
1.8% 1.8%
-2.3% -2.3%
3.0%
-6.3%
Executive Summary ES-5

-------
Figure ES-3: Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990
(1990=0)
Box ES-2: Improvements and Recalculations Relative to the Previous Inventory
Each year, some emission and sink estimates in the Inventory are recalculated and revised to incorporate
improved methods and/or data. The most common reason for recalculating U.S. greenhouse gas emission
estimates is to update recent historical data. Changes in historical data are generally the result of changes in
data supplied by other U.S. government agencies or organizations, as they continue to make refinements and
improvements. These improvements are implemented consistently across the previous Inventory's time series
(i.e., 1990 to 2018) to ensure that the trend is accurate.
Below are categories with recalculations resulting in an average change over the time series of greater than 10
MMT C02 Eq.
•	Forest Land Remaining Forest Land: Changes in Forest Carbon Stocks (C02)
•	Wastewater Treatment (N20)
•	Land Converted to Forest Land: Changes in all Ecosystem Carbon Stocks (C02)
•	Non-Energy Use of Fuels (C02)
In each Inventory, the results of all methodological changes and historical data updates are summarized in the
Recalculations and Improvements chapter (Chapter 9). For more detailed descriptions of each recalculation
including references for data, please see the respective source or sink category description(s) within the
relevant report chapter (i.e., Energy chapter (Chapter 3), the Industrial Process and Product Use (IPPU) chapter
(Chapter 4) the Agriculture chapter (Chapter 5), the Land Use, Land Use Change and Forestry (LULUCF) chapter
(Chapter 6), and the Waste chapter (Chapter 7)). In implementing improvements, the United States follows the
2006IPCC Guidelines (IPCC 2006), which states, "Both methodological changes and refinements over time are
an essential part of improving inventory quality. It is good practice to change or refine methods when: available
data have changed; the previously used method is not consistent with the IPCC guidelines for that category; a
category has become key; the previously used method is insufficient to reflect mitigation activities in a
transparent manner; the capacity for inventory preparation has increased; new inventory methods become
available; and for correction of errors."
ES-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table ES-2: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 Eq.)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
CO?
5,113.5
6,134.5
5,371.8
5,248.0
5,207.8
5,375.5
5,255.8
Fossil Fuel Combustion
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Transportation
1,469.1
1,858.6
1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
Electric Power
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Industrial
853.8
852.9
797.3
792.5
790.1
813.6
822.5
Residential
338.6
358.9
317.3
292.8
293.4
338.1
336.8
Commercial
228.3
227.1
244.6
231.6
232.0
245.7
249.7
U.S. Territories
21.7
55.9
29.2
26.0
24.6
24.6
24.6
Non-Energy Use of Fuels
112.8
129.1
108.5
99.8
113.5
129.7
128.8
Petroleum Systems
9.7
12.1
32.4
21.8
25.0
37.1
47.3
Iron and Steel Production &







Metallurgical Coke Production
104.7
70.1
47.9
43.6
40.6
42.6
41.3
Cement Production
33.5
46.2
39.9
39.4
40.3
39.0
40.9
Natural Gas Systems
32.0
25.2
29.1
30.1
31.2
33.9
37.2
Petrochemical Production
21.6
27.4
28.1
28.3
28.9
29.3
30.8
Ammonia Production
13.0
9.2
10.6
10.2
11.1
12.2
12.3
Lime Production
11.7
14.6
13.3
12.6
12.9
13.1
12.1
Incineration of Waste
8.1
12.7
11.5
11.5
11.5
11.5
11.5
Other Process Uses of Carbonates
6.3
7.6
12.2
11.0
9.9
7.5
7.5
Urea Consumption for Non-







Agricultural Purposes
3.8
3.7
4.6
5.1
5.0
5.9
6.2
Urea Fertilization
2.4
3.5
4.7
4.9
5.1
5.2
5.3
Carbon Dioxide Consumption
1.5
1.4
4.9
4.6
4.6
4.1
4.9
Liming
4.7
4.3
3.7
3.1
3.1
2.2
2.4
Aluminum Production
6.8
4.1
2.8
1.3
1.2
1.5
1.9
Soda Ash Production
1.4
1.7
1.7
1.7
1.8
1.7
1.8
Ferroalloy Production
2.2
1.4
2.0
1.8
2.0
2.1
1.6
Titanium Dioxide Production
1.2
1.8
1.6
1.7
1.7
1.5
1.5
Glass Production
1.5
1.9
1.3
1.2
1.3
1.3
1.3
Zinc Production
0.6
1.0
0.9
0.8
0.9
1.0
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
0.9
0.9
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Carbide Production and Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Magnesium Production and







Processing
+
+
+
+
+
+
+
Wood Biomass, Ethanol, and Biodiesel







Consumptiona
219.4
230.7
317.7
316.6
312.3
319.6
316.2
International Bunker Fuelsb
103.5
113.2
110.9
116.6
120.1
122.1
116.1
CH4c
776.9
686.1
651.5
642.4
648.4
655.9
659.7
Enteric Fermentation
164.7
169.3
166.9
172.2
175.8
178.0
178.6
Natural Gas Systems
186.9
164.2
149.8
147.3
148.7
152.5
157.6
Landfills
176.6
131.4
111.4
108.0
109.4
112.1
114.5
Manure Management
37.1
51.6
57.9
59.6
59.9
61.7
62.4
Coal Mining
96.5
64.1
61.2
53.8
54.8
sin
47.4
Petroleum Systems
48.9
39.5
41.5
39.2
39.3
37.3
39.1
Wastewater Treatment
20.2
20.1
18.8
18.7
18.5
18.4
18.4
Executive Summary ES-7

-------
Rice Cultivation
16.0
18.0
16.2
15.8
14.9
15.6
15.1
Stationary Combustion
8.6
7.8
8.5
7.9
7.6
8.5
8.7
Abandoned Oil and Gas Wells
6.8
7.2
7.4
7.4
7.2
7.3
6.6
Abandoned Underground Coal Mines
7.2
6.6
6.4
6.7
6.4
6.2
5.9
Mobile Combustion
6.4
4.0
2.6
2.5
2.5
2.4
2.4
Composting
0.4
1.9
2.1
2.3
2.4
2.3
2.3
Field Burning of Agricultural Residues
0.4
0.4
0.4
0.4
0.4
0.4
0.4
Petrochemical Production
0.2
0.1
0.2
0.2
0.3
0.3
0.3
Anaerobic Digestion at Biogas







Facilities
+
0.1
0.2
0.2
0.2
0.2
0.2
Ferroalloy Production
+
+
+
+
+
+
+
Carbide Production and Consumption
+
+
+
+
+
+
+
Iron and Steel Production &







Metallurgical Coke Production
+
+
+
+
+
+
+
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
N2Oc
452.7
455.8
468.2
450.8
446.3
459.2
457.1
Agricultural Soil Management
315.9
313.4
348.5
330.1
327.6
338.2
344.6
Wastewater Treatment
18.7
23.0
25.4
25.9
26.4
26.1
26.4
Stationary Combustion
25.1
34.4
30.5
30.0
28.4
28.2
24.9
Manure Management
14.0
16.4
17.5
18.1
18.7
19.4
19.6
Mobile Combustion
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Nitric Acid Production
12.1
11.3
11.6
10.1
9.3
9.6
10.0
AdipicAcid Production
15.2
7.1
4.3
7.0
7.4
10.3
5.3
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Composting
0.3
1.7
1.9
2.0
2.2
2.0
2.0
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Electronics Industry
+
0.1
0.2
0.2
0.3
0.3
0.2
Field Burning of Agricultural Residues
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
1.0
1.0
1.1
1.1
1.0
HFCs
46.5
127.5
168.3
168.1
170.3
169.8
174.6
Substitution of Ozone Depleting







Substancesd
0.2
107.3
163.6
164.9
164.7
166.0
170.5
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
Electronics Industry
0.2
0.2
0.3
0.3
0.4
0.4
0.3
Magnesium Production and







Processing
+
+
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.2
4.4
4.1
4.7
4.5
Electronics Industry
2.8
3.3
3.1
2.9
2.9
3.0
2.7
Aluminum Production
21.5
3.4
2.1
1.4
1.1
1.6
1.8
Substitution of Ozone Depleting







Substances
+
+
+
+
+
0.1
0.1
sf6
28.8
11.8
5.5
6.0
5.9
5.7
5.9
Electrical Transmission and







Distribution
23.2
8.4
3.8
4.1
4.2
3.9
4.2
Magnesium Production and







Processing
5.2
2.7
1.0
1.1
1.0
1.0
0.9
Electronics Industry
0.5
0.7
0.7
0.8
0.7
0.8
0.8
ES-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
nf3
+
0.5
0.6
0.6
0.6
0.6
0.6
Electronics Industry
+
0.5
0.6
0.6
0.6
0.6
0.6
Unspecified Mix of HFCs, PFCs, SF6, and







nf3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total Emissions (Sources)
6,442.7
7,423.0
6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
LULUCF Emissionsc
7.9
16.8
27.8
13.2
26.0
23.4
23.5
LULUCF CH4 Emissions
5.0
9.3
16.6
7.7
15.3
13.8
13.8
LULUCF N20 Emissions
3.0
7.5
11.3
5.5
10.6
9.7
9.7
LULUCF Carbon Stock Change8
(908.7)
(804.8)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
LULUCF Sector Net Total'
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Net Emissions (Sources and Sinks)
5,541.9
6,635.0
5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
Notes: Total emissions presented without LULUCF. Net emissions presented with LULUCF. Totals may not sum due to
independent rounding. Parentheses indicate negative values or sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
a Emissions from Wood Biomass, Ethanol, and Biodiesel Consumption are not included specifically in summing Energy
sector totals. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for Land
Use, Land-Use Change, and Forestry.
b Emissions from International Bunker Fuels are not included in totals.
c LULUCF emissions of CH4 and N20 are reported separately from gross emissions totals. LULUCF emissions include the CH4
and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils, Grassland Fires, and
Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal Wetlands; and N20
emissions from Forest Soils and Settlement Soils.
d Small amounts of PFC emissions also result from this source.
6 LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining
Grassland, Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements
Remaining Settlements, and Land Converted to Settlements.
f The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net C stock
changes.
Figure ES-4 illustrates the relative contribution of the greenhouse gases to total U.S. emissions in 2019, weighted
by global warming potential. The primary greenhouse gas emitted by human activities in the United States was
C02, representing approximately 80.1 percent of total greenhouse gas emissions. The largest source of C02, and of
overall greenhouse gas emissions, was fossil fuel combustion primarily from transportation and power generation.
Methane emissions (CH4) account for approximately 10.1 percent of emissions. The major sources of methane
include enteric fermentation associated with domestic livestock, natural gas systems, and decomposition of wastes
in landfills. Agricultural soil management, wastewater treatment, stationary sources of fuel combustion, and
manure management were the major sources of N20 emissions. Ozone depleting substance substitute emissions
and emissions of HFC-23 during the production of HCFC-22 were the primary contributors to aggregate
hydrofluorocarbon (HFC) emissions. Perfluorocarbon (PFC) emissions were primarily attributable to electronics
manufacturing and primary aluminum production. Electrical transmission and distribution systems accounted for
most sulfur hexafluoride (SF6) emissions. The electronics industry is the only source of nitrogen trifluoride (NF3)
emissions.
Executive Summary ES-9

-------
Figure ES-4: 2019 U.S. Greenhouse Gas Emissions by Gas (Percentages based on MMT CO2
Eq.)
2.8%
HFCs, PFCs, SFe and NFs Subtotal
7.0%
NzO
10.1%
ChU
80.1%
COz
Overall, from 1990 to 2019, total emissions of C02 increased by 142.4 MMT C02 Eq. (2.8 percent), while total
emissions of CH4 decreased by 117.2 MMT C02 Eq. (15.1 percent) and emissions of N20 increased by 4.5 MMT C02
Eq. (1.0 percent). During the same period, aggregate weighted emissions of hydrofluorocarbons (HFCs),
perfluorocarbons (PFCs), sulfur hexafluoride (SF6), and nitrogen trifluoride (NF3) rose by 86.0 MMT C02 Eq. (86.3
percent). From 1990 to 2019, HFCs increased by 128.1 MMT C02 Eq. (275.4 percent), PFCs decreased by 19.8 MMT
C02 Eq. (81.5 percent), SF6 decreased by 22.9 MMT C02 Eq. (79.5 percent), and NF3 increased by 0.6 MMT C02 Eq.
(1,162.7 percent). Despite being emitted in smaller quantities relative to the other principal greenhouse gases,
emissions of HFCs, PFCs, SF6 and NF3 are significant because many of these gases have extremely high global
warming potentials and, in the cases of PFCs and SF6, long atmospheric lifetimes. Conversely, U.S. greenhouse gas
emissions were partly offset by carbon (C) sequestration in forests, trees in urban areas, agricultural soils,
landfilled yard trimmings and food scraps, and coastal wetlands, which, in aggregate, offset 12.4 percent of total
emissions in 2019 (as reflected in Figure ES-1). The following sections describe each gas's contribution to total U.S.
greenhouse gas emissions in more detail.
Carbon Dioxide Emissions
The global carbon cycle is made up of large carbon flows and reservoirs. Billions of tons of carbon in the form of
C02 are absorbed by oceans and living biomass (i.e., sinks) and are emitted to the atmosphere annually through
natural processes (i.e., sources). When in equilibrium, global carbon fluxes among these various reservoirs are
roughly balanced.13
Since the Industrial Revolution (i.e., about 1750), global atmospheric concentrations of C02 have risen
approximately 47 percent (IPCC 2013; NOAA/ESRL 2021a), principally due to the combustion of fossil fuels for
13 The term "flux" is used to describe the exchange of C02 to and from the atmosphere, with net flux being either positive or
negative depending on the overall balance. Removal and long-term storage of C02 from the atmosphere is also referred to as
"carbon sequestration."
ES-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
energy. Globally, an estimated 33,513 MMT of C02 were added to the atmosphere through the combustion of
fossil fuels in 2018, of which the United States accounted for approximately 15 percent.14
Within the United States, fossil fuel combustion accounted for 92.4 percent of C02 emissions in 2019.
Transportation was the largest emitter of C02 in 2019 followed by electric power generation. There are 25
additional sources of C02 emissions included in the Inventory (see Table ES-5). Although not illustrated in the Table
ES-5, changes in land use and forestry practices can also lead to net C02 emissions (e.g., through conversion of
forest land to agricultural or urban use) or to a net sink for C02 (e.g., through net additions to forest biomass). See
more on these emissions and removals in Table ES-5.
Figure ES-5: 2019 Sources of CO2 Emissions
Fossil Fuel Combustion
Non-Energy Use of Fuels
Petroleum Systems
Iron and Steel Prod. & Metallurgical Coke Prod.
Cement Production
Natural Gas Systems
Petrochemical Production
Ammonia Production
Lime Production
Incineration of Waste
Other Process Uses of Carbonates
Urea Consumption for Non-Agricultural Purposes
Urea Fertilization
Carbon Dioxide Consumption
Liming
Aluminum Production
Soda Ash Production
Ferroalloy Production
Titanium Dioxide Production
Glass Production
Zinc Production
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
Abandoned Oil and Gas Wells
Magnesium Production and Processing
< 0.5
4,857
CO2 as a Portion of All
Emissions
CO2
ICH4
N2O
HFCs, PFCs, SFe and NF3
25	50	75	100
MMT COz Eq.
125
150
As the largest source of U.S. greenhouse gas emissions, C02 from fossil fuel combustion has accounted for
approximately 76 percent of GWP-weighted total U.S. gross emissions across the time series. Between 1990 and
2019, C02 emissions from fossil fuel combustion increased from 4,731.5 MMT C02 Eq. to 4,856.7 MMT C02 Eq., a
2.6 percent total increase. Conversely, C02 emissions from fossil fuel combustion decreased by 896.8 MMT C02 Eq.
from 2005 levels, a decrease of approximately 15.6 percent. From 2018 to 2019, these emissions decreased by
134.7 MMT C02 Eq. (2.7 percent).
Historically, changes in emissions from fossil fuel combustion have been the driving factor affecting U.S. emission
trends. Changes in C02 emissions from fossil fuel combustion are influenced by many long-term and short-term
factors. Important drivers include: (1) changes in demand for energy; and (2) a general decline in the carbon
14 Global C02 emissions from fossil fuel combustion were taken from International Energy Agency C02 Emissions from Fossil
Fuels Combustion Overview. See  (IEA 2020).
The publication has not yet been updated to include complete global 2019 data.
Executive Summary ES-11

-------
intensity of fuels combusted for energy in recent years by non-transport sectors of the economy. Long-term
factors affecting energy demand include population and economic trends, technological changes including energy
efficiency, shifting energy fuel choices, and various policies at the national, state, and local level. In the short term,
the overall consumption and mix of fossil fuels in the United States fluctuates primarily in response to changes in
general economic conditions, overall energy prices, the relative price of different fuels, weather, and the
availability of non-fossil alternatives.
The five major fuel-consuming economic sectors are transportation, electric power, industrial, residential, and
commercial. Carbon dioxide emissions are produced by the electric power sector as fossil fuel is consumed to
provide electricity to one of the other four sectors, or "end-use" sectors, see Figure ES-6. Note that this Figure
reports emissions from U.S. Territories as their own end-use sector due to incomplete data for their individual end-
use sectors. Fossil fuel combustion for electric power also includes emissions of less than 0.5 MMT C02 Eq. from
geothermal-based generation.
Figure ES-6: 2019 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type
2,500
2,000
iS" 1,500
(M
0
u
t-
1	1,000
500
0
U.S. Territories Commercial	Residential	Industrial	Electric Power Transportation
Relative Contribution by Fuel Type
<0.05%
¦	Coal
¦	Geothermal
Natural Gas
I Petroleum
Figure ES-7 and Table ES-3 summarize C02 emissions from fossil fuel combustion by end-use sector including
electric power emissions. For Figure ES-7 below, electric power emissions have been distributed to each end-use
sector on the basis of each sector's share of aggregate electricity use (i.e., indirect fossil fuel combustion). This
method of distributing emissions assumes that each end-use sector uses electricity that is generated from the
national average mix of fuels according to their carbon intensity. Emissions from electric power are also addressed
separately after the end-use sectors are discussed.
ES-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure ES-7: 2019 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion
2,000 g Direct Fossil Fuel Combustion
I Indirect Fossil Fuel Combustion
U.S. Territories	Commercial	Residential	Industrial	Transportation
Table ES-3: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990

2005

2015
2016
2017
2018
2019
Transportation
1,472.2

1,863.4

1,723.5
1,764.1
1,786.8
1,821.2
1,821.9
Combustion
1,469.1

1,858.6

1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
Electricity
3.0

4.7

4.3
4.2
4.3
4.7
4.7
Industrial
1,540.2

1,589.2

1,346.8
1,310.1
1,294.5
1,314.9
1,287.8
Combustion
853.8

852.9

797.3
792.5
790.1
813.6
822.5
Electricity
686.4

736.3

549.5
517.6
504.4
501.3
465.3
Residential
931.3

1,214.9

1,001.1
946.2
910.5
980.2
920.3
Combustion
338.6

358.9

317.3
292.8
293.4
338.1
336.8
Electricity
592.7

856.0

683.8
653.5
617.1
642.1
583.5
Commercial
766.0

1,030.1

907.6
865.2
838.2
850.6
802.1
Combustion
228.3

227.1

244.6
231.6
232.0
245.7
249.7
Electricity
537.7

803.0

663.0
633.6
606.2
604.8
552.4
U.S. Territories3
21.7

55.9

29.2
26.0
24.6
24.6
24.6
Total
4,731.5

5,753.5

5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Electric Power
1,820.0

2,400.1

1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Notes: Combustion-related emissions from electric power are allocated based on aggregate national
electricity use by each end-use sector and represent indirect fossil fuel combustion for each end-use
sector. Totals may not sum due to independent rounding.
a Fuel consumption by U.S. Territories (i.e., American Samoa, Guam, Puerto Rico, U.S. Virgin Islands,
Wake Island, and other U.S. Pacific Islands) is included in this report.
Transportation End-Use Sector. Transportation activities accounted for 37.5 percent of U.S. C02 emissions from
fossil fuel combustion in 2019. The largest sources of transportation C02 emissions in 2019 were passenger cars
(40.5 percent); freight trucks (23.6 percent); light-duty trucks, which include sport utility vehicles, pickup trucks,
and minivans (17.2 percent); commercial aircraft (7.2 percent); pipelines (2.9 percent); other aircraft (2.4 percent);
rail (2.2 percent); and ships and boats (2.1 percent). Annex 3.2 presents the total emissions from all transportation
and mobile sources, including C02, CH4, N20, and HFCs.
In terms of the overall trend, from 1990 to 2019, total transportation C02 emissions increased due, in large part, to
increased demand for travel. The number of vehicle miles traveled (VMT) by light-duty motor vehicles (i.e.,
Executive Summary ES-13

-------
passenger cars and light-duty trucks) increased 47.5 percent from 1990 to 2019,15 as a result of a confluence of
factors including population growth, economic growth, urban sprawl, and low fuel prices during the beginning of
this period. While an increased demand for travel has led to increasing C02 emissions since 1990, improvements in
average new vehicle fuel economy since 2005 has slowed the rate of increase of C02 emissions. Petroleum-based
products supplied 98.0 percent of the energy consumed for transportation, with 56.5 percent being related to
gasoline consumption in automobiles and other highway vehicles. Diesel fuel for freight trucks and jet fuel for
aircraft, accounted for 24.3 and 13.3 percent, respectively. The remaining 3.9 percent of petroleum-based energy
consumed for transportation was supplied by natural gas, residual fuel, aviation gasoline, and liquefied petroleum
gases.
Industrial End-Use Sector. Industrial C02 emissions, resulting both directly from the combustion of fossil fuels and
indirectly from the generation of electricity that is used by industry, accounted for 27 percent of C02 emissions
from fossil fuel combustion in 2019. Approximately 64 percent of these emissions resulted from direct fossil fuel
combustion to produce steam and/or heat for industrial processes. The remaining emissions resulted from the use
of electricity for motors, electric furnaces, ovens, lighting, and other applications. Total direct and indirect
emissions from the industrial sector have declined by 16.4 percent since 1990. This decline is due to structural
changes in the U.S. economy (i.e., shifts from a manufacturing-based to a service-based economy), fuel switching,
and efficiency improvements.
Residential and Commercial End-Use Sectors. The residential and commercial end-use sectors accounted for 19 and
17 percent, respectively, of C02 emissions from fossil fuel combustion in 2019. The residential and commercial
sectors relied heavily on electricity for meeting energy demands, with 63 and 69 percent, respectively, of their
emissions attributable to electricity use for lighting, heating, cooling, and operating appliances. The remaining
emissions were due to the consumption of natural gas and petroleum for heating and cooking. Total direct and
indirect emissions from the residential sector have decreased by 1 percent since 1990. Total direct and indirect
emissions from the commercial sector have increased by 4.7 percent since 1990.
Electric Power. The United States relies on electricity to meet a significant portion of its energy demands.
Electricity generators used 31 percent of U.S. energy from fossil fuels and emitted 33 percent of the C02 from fossil
fuel combustion in 2019. The type of energy source used to generate electricity is the main factor influencing
emissions.16 For example, some electricity is generated through non-fossil fuel options such as nuclear,
hydroelectric, wind, solar, or geothermal energy. The mix of fossil fuels used also impacts emissions. The electric
power sector is the largest consumer of coal in the United States. The coal used by electricity generators
accounted for 93 percent of all coal consumed for energy in the United States in 2019.17 However, the amount of
coal and the percent of total electricity generation from coal has been decreasing over time. Coal-fired electric
generation (in kilowatt-hours [kWh]) decreased from 54 percent of generation in 1990 to 28 percent in 2019.18
This corresponded with an increase in natural gas generation and renewable energy generation, largely from wind
and solar energy. Natural gas generation (in kWh) represented 11 percent of electric power generation in 1990 and
increased over the thirty-year period to represent 34 percent of electric power generation in 2019. Wind and solar
15	VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2019). In 2007 and 2008
light-duty VMT decreased 3.0 percent and 2.3 percent, respectively. Note that the decline in light-duty VMT from 2006 to 2007
is due at least in part to a change in FHWA's methods for estimating VMT. In 2011, FHWA changed its methods for estimating
VMT by vehicle class, which led to a shift in VMT and emissions among on-road vehicle classes in the 2007 to 2019 time period.
In absence of these method changes, light-duty VMT growth between 2006 and 2007 would likely have been higher.
16	In line with the reporting requirements for inventories submitted under the UNFCCC, C02 emissions from biomass
combustion have been estimated separately from fossil fuel C02 emissions and are not included in the electricity sector totals
and trends discussed in this section. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the
estimates for Land Use, Land-Use Change, and Forestry.
17	See Table 6.2 Coal Consumption by Sector of EIA (2020a).
18	Values represent electricity net generation from the electric power sector. See Table 7.2b Electricity Net Generation: Electric
Power Sector of EIA (2020a).
ES-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
generation (in kWh) represented 0.1 percent of electric power generation in 1990 and increased over the thirty-
year period to represent 9 percent of electric power generation in 2019.
Across the time series, changes in electricity generation and the carbon intensity of fuels used for electric power
have a significant impact on C02 emissions. While C02 emissions from the electric power sector have decreased by
approximately 12 percent since 1990, the carbon intensity of the electric power sector, in terms of C02 Eq. per
QBtu input, has significantly decreased- by 27 percent-during that same timeframe. This decoupling of the level of
electric power generation and the resulting C02 emissions is shown in Figure ES-8.
Figure ES-8: Electric Power Generation and Emissions
4 500 ' Nuclear Generation (Billion kWh)
Renewable Generation (Billion kWh)
¦ Petroleum Generation (Billion kWh)
4 000 Coal Generation (Billion kWh)
' I Natural Gas Generation (Billion kWh)
. 3,500
I Total Emissions (MMT CO2 Eq.) [Right Axis]
= 3,000
ro 2,500
oj 2,000
Q 1,500
3,500
3,000
2,500
2,000
1,500
1,000
500
Other significant C02 trends included the following:
•	Carbon dioxide emissions from natural gas and petroleum systems increased by 42.8 MMT C02 Eq. (102.4
percent) from 1990 to 2019. This increase is due primarily to increases in the production segment, where
flaring emissions from associated gas flaring, tanks, and miscellaneous production flaring have increased
over time.
•	Carbon dioxide emissions from iron and steel production and metallurgical coke production have
decreased by 63.4 MMT C02 Eq. (60.6 percent) from 1990 through 2019, due to restructuring of the
industry, technological improvements, and increased scrap steel utilization.
•	Total C stock change (i.e., net C02 removals) in the LULUCF sector decreased by approximately 10.6
percent between 1990 and 2019. This decrease was primarily due to a decrease in the rate of net C
accumulation in forest C stocks and Cropland Remaining Cropland, as well as an increase in emissions
from Land Converted to Settlements.
Methane Emissions
Methane (CH4) is significantly more effective than C02 at trapping heat in the atmosphere-by a factor of 25 over a
100-year time frame based on the IPCC Fourth Assessment Report estimate (IPCC 2007). Over the last two hundred
and fifty years, the concentration of CH4 in the atmosphere increased by 167 percent (IPCC 2013; NOAA/ESRL
Executive Summary ES-15

-------
2021b). Within the United States, the main anthropogenic sources of CH4 include enteric fermentation from
domestic livestock, natural gas systems, landfills, domestic livestock manure management, coal mining, and
petroleum systems (see Figure ES-9).
Figure ES-9: 2019 Sources of ChU Emissions
Enteric Fermentation
Natural Gas Systems
Landfills
Manure Management
Coal Mining
Petroleum Systems
Wastewater Treatment
Rice Cultivation
Stationary Combustion
Abandoned Oil and Gas Wells
Abandoned Underground Coal Mines
Mobile Combustion
Composting
Field Burning of Agricultural Residues
Petrochemical Production
Anaerobic Digestion at Biogas Facilities
Ferroalloy Production
Carbide Production and Consumption
Iron and Steel Production & Metallurgical Coke Production
Incineration of Waste
Note: Emissions of CH4 from LULUCF are reported separately from gross emissions totals and are not included in Figure ES-9.
Refer to Table ES-5 for a breakout of LULUCF emissions by gas.
Significant trends for the largest sources of U.S. CH4 emissions include the following:
•	Enteric fermentation was the largest anthropogenic source of CH4 emissions in the United States. In 2019,
enteric fermentation CH4 emissions were 178.6 MMT C02 Eq. (27.1 percent of total CH4 emissions), which
represents an increase of 13.9 MMT C02 Eq. (8.4 percent) since 1990. This increase in emissions from
1990 to 2019 generally follows the increasing trends in cattle populations.
•	Natural gas systems were the second largest anthropogenic source category of CH4 emissions in the
United States in 2019 with 157.6 MMT C02 Eq. of CH4 emitted into the atmosphere. Those emissions have
decreased by 29.3 MMT C02 Eq. (15.7 percent) since 1990. The decrease in CH4 emissions is largely due to
decreases in emissions from distribution, transmission, and storage. The decrease in distribution
emissions is due to decreased emissions from pipelines and distribution station leaks, and the decrease in
transmission and storage emissions is largely due to reduced compressor station emissions (including
emissions from compressors and equipment leaks).
•	Landfills were the third largest anthropogenic source of CH4 emissions in the United States (114.5 MMT
C02 Eq.), accounting for 17.4 percent of total CH4 emissions in 2019. From 1990 to 2019, CH4 emissions
from landfills decreased by 62.1 MMT C02 Eq. (35.2 percent), with small year-to-year increases. This
downward trend in emissions coincided with increased landfill gas collection and control systems, and a
reduction of decomposable materials (i.e., paper and paperboard, food scraps, and yard trimmings)
CH4 as a Portion of All
Emissions
< 0.5
<	0.5
<	0.5
<	0.5
<	0.5
<	0.5
<	0.5
10.1%
CO2
ICH«
NzO
HFCs, PFCs, SFe and NFa
0 20 40 60 80 100 120 140 160 180
MMT CO2 Eq.
ES-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
discarded in MSW landfills over the time series.19 While the amount of landfill gas collected and
combusted continues to increase, the rate of increase in collection and combustion no longer exceeds the
rate of additional CH4 generation from the amount of organic MSW landfilled as the U.S. population
grows.
Nitrous Oxide Emissions
Nitrous oxide (N20) is produced by biological processes that occur in soil and water and by a variety of
anthropogenic activities in the agricultural, energy, industrial, and waste management fields. While total N20
emissions are much lower than C02 emissions, N20 is nearly 300 times more powerful than C02 at trapping heat in
the atmosphere over a 100-year time frame (IPCC 2007). Since 1750, the global atmospheric concentration of N20
has risen by approximately 23 percent (IPCC 2013; NOAA/ESRL 2021c). The main anthropogenic activities
producing N20 in the United States are agricultural soil management, wastewater treatment, stationary fuel
combustion, manure management, fuel combustion in motor vehicles, and nitric acid production (see Figure ES-
10).
Figure ES-10: 2019 Sources of N2O Emissions
Agricultural Soil Management
Wastewater Treatment
Stationary Combustion
Manure Management
Mobile Combustion
Nitric Acid Production
Adipic Acid Production
N2O from Product Uses
Composting
Caprolactam Production
Incineration of Waste
Electronics Industry
Field Burning of Agricultural Residues
Petroleum Systems
Natural Gas Systems
N2O as a Portion of All
Emissions
< 0.5
345
CO2
CH4
N2O
HFCs, PFCs, SFe and NF3
15
20
MMT CO2 Eq.
25
30
35
40
Note: Emissions of N20 from LULUCF are reported separately from gross emissions totals and are not included in Figure ES-10.
Refer to Table ES-5 for a breakout of LULUCF emissions by gas.
Significant trends for the largest sources of U.S. emissions of N20 include the following:
• Agricultural soils accounted for 75.4 percent of N20 emissions and 5.3 percent of total greenhouse gas
emissions in the United States in 2019. Estimated emissions from this source in 2019 were 344.6 MMT
C02 Eq. Annual N20 emissions from agricultural soils fluctuated between 1990 and 2019, although overall
emissions were 9.1 percent higher in 2019 than in 1990. Year-to-year fluctuations are largely a reflection
of annual variation in weather patterns, synthetic fertilizer use, and crop production.
19 Carbon dioxide emissions from landfills are not included specifically in summing waste sector totals. Net carbon fluxes from
changes in biogenic carbon reservoirs and decay of disposed wood products are accounted for in the estimates for LULUCF.
Executive Summary ES-17

-------
•	Wastewater treatment, both domestic and industrial, accounted for 5.8 percent of N20 emissions and 0.4
percent of total greenhouse gas emissions in the United States in 2019. Emissions from wastewater
treatment increased by 41.0 percent (7.7 MMT C02 Eq.) since 1990. Nitrous oxide emissions from
wastewater treatment processes gradually increased across the time series as a result of growing U.S.
population and protein consumption. Nitrous oxide emissions from industrial wastewater treatment
sources, included for the first time in the current (i.e., 1990 to 2019) Inventory, fluctuated throughout the
time series with production changes associated with the treatment of wastewater from the pulp and
paper manufacturing, meat and poultry processing, fruit and vegetable processing, starch-based ethanol
production, petroleum refining, and brewery industries. Industrial wastewater emissions have increased
since 2017.
•	Nitrous oxide emissions from manure management accounted for 4.3 percent of N20 emissions in 2019
and increased by 40.2 percent (5.6 MMT C02 Eq.) from 1990 to 2019. While the industry trend has been a
shift toward liquid systems, driving down the emissions per unit of nitrogen excreted (dry manure
handling systems have greater aerobic conditions that promote N20 emissions), increases in specific
animal populations have driven an increase in overall manure management N20 emissions over the time
series.
•	Nitrous oxide emissions from mobile combustion decreased by 26.8 MMT C02 Eq. (59.8 percent) from
1990 to 2019, primarily as a result of national vehicle emissions standards and emission control
technologies for on-road vehicles.
HFC, RFC, SFg, and NF3 Emissions
Hydrofluorocarbons (HFCs) are synthetic chemicals that are used as alternatives to ozone depleting substances
(ODS), which are being phased out under the Montreal Protocol and Clean Air Act Amendments of 1990.
Hydrofluorocarbons do not deplete the stratospheric ozone layer and therefore have been used as alternatives
under the Montreal Protocol on Substances that Deplete the Ozone Layer.
Perfluorocarbons (PFCs) are emitted from the production of electronics and aluminum and also (in smaller
quantities) from their use as alternatives to ozone depleting substances. Sulfur hexafluoride (SF6) is emitted from
the production of electronics and magnesium and from the manufacturing and use of electrical transmission and
distribution equipment. NF3 is also emitted from electronics production. One HFC, HFC-23, is emitted during
production of HCFC-22 and electronics (see Figure ES-11).
HFCs, PFCs, SF6, and NF3 are potent greenhouse gases. In addition to having very high global warming potentials,
SF6 and PFCs have extremely long atmospheric lifetimes, resulting in their essentially irreversible accumulation in
the atmosphere once emitted. Sulfur hexafluoride is the most potent greenhouse gas the IPCC has evaluated (IPCC
2013).
ES-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure ES-11: 2019 Sources of HFCs, PFCs, SFe, and NF3 Emissions
Substitution of Ozone Depleting Substances |	171
Electronics Industry
Electrical Transmission and Distribution
HCFC-22 Production
Aluminum Production
Magnesium Production and Processing
MMT CCh Eq.
HFCs, PFCs, SF6, and NF3 as a Portion
of All Emissions
HFCs, PFCs, SF6 and NF3
Some significant trends for the largest sources of U.S. HFC, PFC, SF6, and NF3 emissions include the following:
•	Hydrofluorocarbon and perfluorocarbon emissions resulting from their use as substitutes for ODS (e.g.,
chlorofluorocarbons [CFCs]) are the largest share of fluorinated emissions (92 percent) and have been
consistently increasing, from small amounts in 1990 to 170.6 MMT C02 Eq. in 2019. This increase was in
large part the result of efforts to phase out CFCs and other ODS in the United States.
•	Emissions from HCFC-22 production were 3.7 MMT C02 Eq. in 2019, a 92 percent decrease from 1990
emissions. The decrease from 1990 emissions was caused primarily by a reduction in the HFC-23 emission
rate (kg HFC-23 emitted/kg HCFC-22 produced). The emission rate was lowered by optimizing the
production process and capturing much of the remaining HFC-23 for use or destruction.
•	GWP-weighted PFC, HFC, SF6, and NF3 emissions from the electronics industry have increased by 23.7
percent from 1990 to 2019, reflecting the competing influences of industrial growth and the adoption of
emission reduction technologies. Within that time span, emissions peaked at 9.0 MMT C02 Eq. in 1999,
the initial year of EPA's PFC Reduction/Climate Partnership for the Semiconductor Industry, but have since
declined to 4.4 MMT C02 Eq. in 2019 (a 51.3 percent decrease relative to 1999).
•	Sulfur hexafluoride emissions from electric power transmission and distribution systems decreased by
81.7 percent (18.9 MMT C02 Eq.) from 1990 to 2019. There are two factors contributing to this decrease:
(1) a sharp increase in the price of SF6 during the 1990s and (2) a growing awareness of the environmental
impact of SF6 emissions through programs such as EPA's SF6 Emission Reduction Partnership for Electric
Power Systems.
ES.3 Overview of Sector Emissions and Trends
Figure ES-12 and Table ES-4 aggregate emissions and sinks by the sectors defined by the UNFCCC reporting
guidelines to promote comparability across countries. Over the thirty-year period of 1990 to 2019, total emissions
from the Energy, Industrial Processes and Product Use, and Agriculture sectors grew by 66.7 MMT C02 Eq. (1.3
percent), 28.2 MMT C02 Eq. (8.1 percent), and 73.3 MMT C02 Eq. (13.2 percent), respectively. Emissions from the
Waste sector decreased by 52.4 MMT C02 Eq. (24.2 percent). Over the same period, net C sequestration in the
LULUCF sector decreased by 96.0 MMT C02 (10.6 percent decrease in total net C sequestration), while emissions
from the LULUCF sector (i.e., CH4 and N20) increased by 15.5 MMT C02 Eq. (196.1 percent).
Executive Summary ES-19

-------
Figure ES-12: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector
8,000
¦	Net Emissions (Including Sinks)
LULUCF (emissions)
¦	Waste
Industrial Processes and Product Use
¦	Agriculture
Energy
¦	LULUCF (removals)
Table ES-4: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC
Sector (MMT COz Eq.)
Chapter/IPCC Sector
1990

2005

2015
2016
2017
2018
2019
Energy
5,325.6

6,302.3

5,519.8
5,390.9
5,351.0
5,518.1
5,392.3
Fossil Fuel Combustion
4,731.5

5,753.5

5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Natural Gas Systems
219.0

189.4

179.0
177.4
179.9
186.4
194.9
Non-Energy Use of Fuels
112.8

129.1

108.5
99.8
113.5
129.7
128.8
Petroleum Systems
58.6

51.5

73.9
61.1
64.4
74.5
86.4
Coal Mining
96.5

64.1

61.2
53.8
54.8
52.7
47.4
Stationary Combustion
33.7

42.2

39.0
37.9
36.1
36.8
33.5
Mobile Combustion
51.1

45.5

24.4
23.4
22.3
21.3
20.3
Incineration of Waste
8.5

13.1

11.8
11.8
11.8
11.9
11.8
Abandoned Oil and Gas Wells
6.8

7.2

7.4
7.4
7.2
7.3
6.6
Abandoned Underground Coal Mines
7.2

6.6

6.4
6.7
6.4
6.2
5.9
Industrial Processes and Product Use
345.6

365.7

375.4
368.0
367.7
371.3
373.7
Substitution of Ozone Depleting









Substances
0.2

107.3

163.6
164.9
164.7
166.1
170.6
Iron and Steel Production &









Metallurgical Coke Production
104.8

70.1

47.9
43.6
40.6
42.6
41.3
Cement Production
33.5

46.2

39.9
39.4
40.3
39.0
40.9
Petrochemical Production
21.8

27.5

28.2
28.6
29.2
29.6
31.1
Ammonia Production
13.0

9.2

10.6
10.2
11.1
12.2
12.3
Lime Production
11.7

14.6

13.3
12.6
12.9
13.1
12.1
Nitric Acid Production
12.1

11.3

11.6
10.1
9.3
9.6
10.0
Other Process Uses of Carbonates
6.3

7.6

12.2
11.0
9.9
7.5
7.5
Urea Consumption for Non-









Agricultural Purposes
3.8

3.7

4.6
5.1
5.0
5.9
6.2
AdipicAcid Production
15.2

7.1

4.3
7.0
7.4
10.3
5.3
Carbon Dioxide Consumption
1.5

1.4

4.9
4.6
4.6
4.1
4.9
Electronics Industry
3.6

4.8

5.0
5.0
4.9
5.1
4.6
Electrical Transmission and









Distribution
23.2

8.4

3.8
4.1
4.2
3.9
4.2
N,0 from Product Uses
4.2

4.2

4.2
4.2
4.2
4.2
4.2
ES-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
Aluminum Production
28.3
7.6
4.9
2.7
2.3
3.1
3.6
Soda Ash Production
1.4
1.7
1.7
1.7
1.8
1.7
1.8
Ferroalloy Production
2.2
1.4
2.0
1.8
2.0
2.1
1.6
Titanium Dioxide Production
1.2
1.8
1.6
1.7
1.7
1.5
1.5
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
Glass Production
1.5
1.9
1.3
1.2
1.3
1.3
1.3
Zinc Production
0.6
1.0
0.9
0.8
0.9
1.0
1.0
Magnesium Production and







Processing
5.2
2.7
1.1
1.2
1.1
1.1
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
0.9
0.9
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Carbide Production and Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Agriculture
555.3
577.1
616.1
604.4
605.5
621.0
628.6
Agricultural Soil Management
315.9
313.4
348.5
330.1
327.6
338.2
344.6
Enteric Fermentation
164.7
169.3
166.9
172.2
175.8
178.0
178.6
Manure Management
51.1
67.9
75.4
77.7
78.5
81.1
82.0
Rice Cultivation
16.0
18.0
16.2
15.8
14.9
15.6
15.1
Urea Fertilization
2.4
3.5
4.7
4.9
5.1
5.2
5.3
Liming
4.7
4.3
3.7
3.1
3.1
2.2
2.4
Field Burning of Agricultural Residues
0.5
0.6
0.6
0.6
0.6
0.6
0.6
Waste
216.2
178.0
159.8
157.1
159.0
161.1
163.7
Landfills
176.6
131.4
111.4
108.0
109.4
112.1
114.5
Wastewater Treatment
38.9
43.0
44.2
44.6
44.9
44.6
44.8
Composting
0.7
3.5
4.0
4.3
4.6
4.3
4.3
Anaerobic Digestion at Biogas







Facilities
+
0.1
0.2
0.2
0.2
0.2
0.2
Total Emissions3 (Sources)
6,442.7
7,423.0
6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
LULUCF Sector Net Totalb
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Forest land
(884.1)
(751.4)
(749.5)
(814.7)
(740.0)
(781.4)
(774.6)
Cropland
28.6
23.2
43.2
31.7
32.3
37.7
39.7
Grassland
2.2
(29.4)
(10.1)
(13.7)
(12.5)
(11.9)
(8.0)
Wetlands
(2.8)
(1.9)
(3.9)
(3.9)
(3.8)
(3.9)
(3.9)
Settlements
(44.7)
(28.5)
(43.5)
(42.2)
(42.1)
(42.0)
(42.4)
Net Emission (Sources and Sinks)c
5,541.9
6,635.0
5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
Notes: Total emissions presented without LULUCF. Net emissions presented with LULUCF. Totals may not sum due to
independent rounding. Parentheses indicate negative values or sequestration.
a Total emissions without LULUCF.
b The LULUCF Sector Net Total is the sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes in units of MMT C02 Eq.
c Net emissions with LULUCF.
Energy
The Energy chapter contains emissions of all greenhouse gases resulting from stationary and mobile energy
activities including fuel combustion and fugitive fuel emissions, and the use of fossil fuels for non-energy purposes.
Energy-related activities, primarily fossil fuel combustion, accounted for the vast majority of U.S. C02 emissions for
the period of 1990 through 2019. Energy-related activities are also responsible for CH4 and N20 emissions (40.6
percent and 9.5 percent of total U.S. emissions of each gas, respectively). Overall, emission sources in the Energy
chapter account for a combined 82.2 percent of total U.S. greenhouse gas emissions in 2019.
In 2019, approximately 80 percent of the energy used in the United States (on a Btu basis) was produced through
the combustion of fossil fuels. The remaining 20 percent came from other energy sources, such as hydropower,
biomass, nuclear, wind, and solar energy (see Figure ES-13).
Executive Summary ES-21

-------
Figure ES-13: 2019 U.S. Energy Consumption by Energy Source (Percent)
Nuclear Electric Power
8.4%
Renewable Energy
11.3%
ergy
Jb.87o

Coal
11.3%
Natural Gas
32.1%
Industrial Processes and Product Use
The Industrial Processes and Product Use (IPPU) chapter contains information on greenhouse gas emissions
generated and emitted as the byproducts of non-energy-related industrial processes, which involve the chemical
or physical transformation of raw materials and can release waste gases such as C02, CH4, N20, and fluorinated
gases (e.g., HFC-23). These processes include iron and steel production and metallurgical coke production, cement
production, petrochemical production, lime production, ammonia production, nitric acid production, other process
uses of carbonates (e.g., flue gas desulfurization), urea consumption for non-agricultural purposes, adipic acid
production, HCFC-22 production, aluminum production, soda ash production and use, ferroalloy production,
titanium dioxide production, caprolactam production, glass production, zinc production, phosphoric acid
production, lead production, and silicon carbide production and consumption. Most of these industries also emit
C02 from fossil fuel combustion which, in line with IPCC sectoral definitions, is included in the Energy Sector.
This chapter also contains information on the release of HFCs, PFCs, SF6, and NF3 and other fluorinated compounds
used in industrial manufacturing processes and by end-consumers (e.g., residential and mobile air conditioning).
These industries include electronics industry, electric power transmission and distribution, and magnesium metal
production and processing. In addition, N20 is used in and emitted by electronics industry and anesthetic and
aerosol applications, and C02 is consumed and emitted through various end-use applications. In 2019, emissions
resulting from use of the substitution of ODS (e.g., chlorofluorocarbons [CFCs]) by end-consumers was the largest
source of IPPU emissions and accounted for 170.6 MMT C02 Eq, or 45.6 percent of total IPPU emissions.
IPPU activities are responsible for 3.2, 0.1, and 4.6 percent of total U.S. C02, CH4, and N20 emissions respectively as
well as for all U.S. emissions of fluorinated gases such as HFCs, PFCs, SF6 and NF3. Overall, emission sources in the
IPPU chapter accounted for 5.7 percent of U.S. greenhouse gas emissions in 2019.
The Agriculture chapter contains information on anthropogenic emissions from agricultural activities (except fuel
combustion, which is addressed in the Energy chapter, and some agricultural C02, CH4, and N20 fluxes, which are
addressed in the Land Use, Land-Use Change, and Forestry chapter).
Agricultural activities contribute directly to emissions of greenhouse gases through a variety of processes,
including the following source categories: agricultural soil management, enteric fermentation in domestic
livestock, livestock manure management, rice cultivation, urea fertilization, liming, and field burning of agricultural
residues.
Agriculture
ES-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
In 2019, agricultural activities were responsible for emissions of 628.6 MMT C02 Eq., or 9.6 percent of total U.S.
greenhouse gas emissions. Methane, N20, and C02 are greenhouse gases emitted by agricultural activities.
Methane emissions from enteric fermentation and manure management represented approximately 27.1 percent
and 9.5 percent of total CH4 emissions from anthropogenic activities, respectively, in 2019. Agricultural soil
management activities, such as application of synthetic and organic fertilizers, deposition of livestock manure, and
growing N-fixing plants, were the largest contributors to U.S. N20 emissions in 2019, accounting for 75.4 percent of
total N20 emissions. Carbon dioxide emissions from the application of crushed limestone and dolomite (i.e., soil
liming) and urea fertilization represented 0.1 percent of total C02 emissions from anthropogenic activities.
Land Use, Land-Use Change, and Forestry
The LULUCF chapter contains emissions and removals of C02 and emissions of CH4 and N20 from managed lands in
the United States. Consistent with the 2006IPCC Guidelines, emissions and removals from managed lands are
considered to be anthropogenic, while emissions and removals from unmanaged lands are considered to be
natural.20 The share of managed land in the U.S. is approximately 95 percent of total land included in the
Inventory.21 More information on the definition of managed land used in the Inventory is provided in Chapter 6.
Overall, the Inventory results show that managed land is a net sink for C02 (C sequestration). The primary drivers
of fluxes on managed lands include forest management practices, tree planting in urban areas, the management of
agricultural soils, landfilling of yard trimmings and food scraps, and activities that cause changes in C stocks in
coastal wetlands. The main drivers for forest C sequestration include forest growth and increasing forest area (i.e.,
afforestation), as well as a net accumulation of C stocks in harvested wood pools. The net sequestration in
Settlements Remaining Settlements, which occurs predominantly from urban forests (i.e., Settlement Trees) and
landfilled yard trimmings and food scraps, is a result of net tree growth and increased urban forest area, as well as
long-term accumulation of yard trimmings and food scraps carbon in landfills.
The LULUCF sector in 2019 resulted in a net increase in C stocks (i.e., net C02 removals) of 812.7 MMT C02 Eq.
(Table ES-5).22 This represents an offset of 12.3 percent of total (i.e., gross) greenhouse gas emissions in 2019.
Emissions of CH4 and N20 from LULUCF activities in 2019 were 23.5 MMT C02 Eq. and represent 0.4 percent of
total greenhouse gas emissions.23 Between 1990 and 2019, total C sequestration in the LULUCF sector decreased
by 10.6 percent, primarily due to a decrease in the rate of net C accumulation in forests and Cropland Remaining
Cropland, as well as an increase in C02 emissions from Land Converted to Settlements. The overall net flux from
LULUCF (i.e., net sum of all CH4 and N20 emissions to the atmosphere plus net carbon stock changes in units of
MMT C02 eq.) resulted in a removal of 789.2 MMT C02 Eq. in 2019.
Forest fires were the largest source of CH4 emissions from the LULUCF sector in 2019, totaling 9.5 MMT C02 Eq.
(379 kt of CH4). Coastal Wetlands Remaining Coastal Wetlands resulted in CH4 emissions of 3.8 MMT C02 Eq. (153
kt of CH4). Grassland fires resulted in CH4 emissions of 0.3 MMT C02 Eq. (12 kt of CH4). Land Converted to Wetlands
resulted in CH4 emissions of 0.2 MMT C02 Eq (7 kt of CH4). Drained Organic Soils and Peatlands Remaining
Peatlands resulted in CH4 emissions of less than 0.05 MMT C02 Eq. each.
20	See .
21	The current land representation does not include land in U.S. Territories, but there are planned improvements to include
these regions in future Inventories. U.S. Territories represent approximately 0.1 percent of the total land base for the United
States. See Box 6-2 in Chapter 6 of this report.
22	LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining
Settlements, and Land Converted to Settlements.
23	LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
Executive Summary ES-23

-------
Forest fires were also the largest source of N20 emissions from the LULUCF sector in 2019, totaling 6.2 MMT C02
Eq. (21 kt of N20). Nitrous oxide emissions from fertilizer application to settlement soils in 2019 totaled to 2.4
MMT C02 Eq. (8 kt of N20). Additionally, the application of synthetic fertilizers to forest soils in 2019 resulted in
N20 emissions of 0.5 MMT C02 Eq. (2 kt of N20). Grassland fires resulted in N20 emissions of 0.3 MMT C02 Eq. (1 kt
of N20). Coastal Wetlands Remaining Coastal Wetlands and Drained Organic Soils resulted in N20 emissions of 0.1
MMT C02 Eq. each (less than 0.5 kt of N20). Peatlands Remaining Peatlands resulted in N20 emissions of less than
0.05 MMT C02 Eq.
Carbon dioxide removals from C stock changes are presented in Table ES-5 along with CH4 and N20 emissions for
LULUCF source categories.
Table ES-5: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-
Use Change, and Forestry (MMT CO2 Eq.)
Land-Use Category
1990
2005
2015
2016
2017
2018
2019
Forest Land Remaining Forest Land
(785.9)
(652.8)
(650.6)
(715.7)
(640.9)
(682.4)
(675.5)
Changes in Forest Carbon Stocks3
(787.6)
(661.5)
(671.4)
(721.9)
(659.7)
(698.6)
(691.8)
Non-C02 Emissions from Forest Firesb
1.5
8.2
20.3
5.6
18.3
15.7
15.7
N20 Emissions from Forest Soilsc
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Non-C02 Emissions from Drained







Organic Soilsd
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Land Converted to Forest Land
(98.2)
(98.7)
(98.9)
(99.0)
(99.1)
(99.1)
(99.1)
Changes in Forest Carbon Stocks0
(98.2)
(98.7)
(98.9)
(99.0)
(99.1)
(99.1)
(99.1)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.8)
(22.7)
(22.3)
(16.6)
(14.5)
Changes in Mineral and Organic Soil







Carbon Stocks
(23.2)
(29.0)
(12.8)
(22.7)
(22.3)
(16.6)
(14.5)
Land Converted to Cropland
51.8
52.2
56.1
54.4
54.6
54.3
54.2
Changes in all Ecosystem Carbon Stocks'
51.8
52.2
56.1
54.4
54.6
54.3
54.2
Grassland Remaining Grassland
8.5
10.7
13.8
10.4
11.9
12.3
15.1
Changes in Mineral and Organic Soil







Carbon Stocks
8.3
10.0
13.1
9.8
11.3
11.7
14.5
Non-C02 Emissions from Grassland







Fires5
0.2
0.7
0.7
0.6
0.6
0.6
0.6
Land Converted to Grassland
(6.2)
(40.1)
(23.9)
(24.0)
(24.4)
(24.1)
(23.2)
Changes in all Ecosystem Carbon Stocks'
(6.2)
(40.1)
(23.9)
(24.0)
(24.4)
(24.1)
(23.2)
Wetlands Remaining Wetlands
(3.5)
(2.6)
(4.1)
(4.1)
(4.0)
(4.0)
(4.0)
Changes in Organic Soil Carbon Stocks in







Peatlands
1.1
1.1
0.8
0.7
0.8
0.8
0.8
Changes in Biomass, DOM, and Soil







Carbon Stocks in Coastal Wetlands
(8.5)
(7.6)
(8.8)
(8.8)
(8.8)
(8.8)
(8.8)
CH4 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
3.7
3.8
3.8
3.8
3.8
3.8
3.8
N20 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
0.1
0.2
0.1
0.1
0.1
0.1
0.1
Non-C02 Emissions from Peatlands







Remaining Peatlands
+
+
+
+
+
+
+
Land Converted to Wetlands
0.7
0.7
0.2
0.2
0.2
0.2
0.2
Changes in Biomass, DOM, and Soil







Carbon Stocks
0.4
0.4
(0.1)
(+)
(+)
(+)
(+)
CH4 Emissions from Land Converted to







Coastal Wetlands
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Settlements Remaining Settlements
(107.6)
(113.5)
(123.7)
(121.5)
(121.4)
(121.2)
(121.7)
Changes in Organic Soil Carbon Stocks
11.3
12.2
15.7
16.0
16.0
15.9
15.9
Changes in Settlement Tree Carbon







Stocks
(96.4)
(117.4)
(130.4)
(129.8)
(129.8)
(129.8)
(129.8)
Changes in Yard Trimming and Food







Scrap Carbon Stocks in Landfills
(24.5)
(11.4)
(11.1)
(10.0)
(9.8)
(9.8)
(10.2)
N20 Emissions from Settlement Soilsh
2.0
3.1
2.2
2.2
2.3
2.4
2.4
ES-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Land Converted to Settlements	62.9	85.0	80.1 79.4 79.3 79.3 79.2
Changes in all Ecosystem Carbon Stocks' 62.9	85.0	80.1 79.4 79.3 79.3 79.2
LULUCF Carbon Stock Change'
(908.7)
(804.8)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
LULUCF Emissions1
7.9
16.8
27.8
13.2
26.0
23.4
23.5
LULUCF CH4 Emissions
5.0
9.3
16.6
7.7
15.3
13.8
13.8
LULUCF N20 Emissions
3.0
7.5
11.3
5.5
10.6
9.7
9.7
LULUCF Sector NetTotalk
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
a Includes the net changes to carbon stocks stored in all forest ecosystem pools and harvested wood products.
b Estimates include emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
c Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
d Estimates include CH4 and N20 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
6 Includes the net changes to carbon stocks stored in all forest ecosystem pools.
f Includes changes in mineral and organic soil carbon stocks for all land use conversions to cropland, grassland, and settlements,
respectively. Also includes aboveground/belowground biomass, dead wood, and litter carbon stock changes for conversion of
forest land to cropland, grassland, and settlements, respectively.
5 Estimates include CH4 and N20 emissions from fires on both Grassland Remaining Grassland and Land Converted to Grass/and.
h Estimates include N20 emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted to
Settlements because it is not possible to separate the activity data at this time.
' LULUCF Carbon Stock Change includes any C stock gains and losses from all land use and land use conversion categories.
> LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
k The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes in units of MMT C02 Eq.
Waste
The Waste chapter contains emissions from waste management activities (except incineration of waste, which is
addressed in the Energy chapter). Landfills were the largest source of anthropogenic greenhouse gas emissions
from waste management activities, generating 114.5 MMT C02 Eq. and accounting for 69.9 percent of total
greenhouse gas emissions from waste management activities, and 17.4 percent of total U.S. CH4 emissions.24
Additionally, wastewater treatment generated emissions of 44.8 MMT C02 Eq. and accounted for 27.3 percent of
total Waste sector greenhouse gas emissions, 2.8 percent of U.S. CH4 emissions, and 5.8 percent of U.S. N20
emissions in 2019. Emissions of CH4 and N20 from composting are also accounted for in this chapter, generating
emissions of 2.3 MMT C02 Eq. and 2.0 MMT C02 Eq., respectively. Anaerobic digestion at biogas facilities
generated CH4 emissions of 0.2 MMT C02 Eq., accounting for 0.1 percent of emissions from the waste sector.
Overall, emission sources accounted for in the Waste chapter generated 163.7 MMT C02 Eq., or 2.5 percent of total
U.S. greenhouse gas emissions in 2019.
24 Landfills also store carbon, due to incomplete degradation of organic materials such as harvest wood products, yard
trimmings, and food scraps, as described in the Land-Use, Land-Use Change, and Forestry chapter of the Inventory report.
Executive Summary ES-25

-------
ES.4 Other Information
Emissions by Economic Sector
Throughout the Inventory of U.S. Greenhouse Gas Emissions and Sinks report, emission estimates are grouped into
five sectors (i.e., chapters) defined by the IPCC: Energy; IPPU; Agriculture; LULUCF; and Waste. It is also useful to
characterize emissions according to commonly used economic sector categories: residential, commercial, industry,
transportation, electric power, and agriculture. Emissions from U.S. Territories are reported as their own end-use
sector due to a lack of specific consumption data for the individual end-use sectors within U.S. Territories. For
more information on trends in the Land use, Land Use Change and Forestry sector, see section ES.2 Recent Trends
in U.S. Greenhouse Gas Emissions and Sinks.
Figure ES-14 shows the trend in emissions by economic sector from 1990 to 2019, and Table ES-6 summarizes
emissions from each of these economic sectors.
Figure ES-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors
2,500
Electric Power Industry (Purple)
2,000
Transportation (Green)
1,500
Industry
f 1,000
Agriculture
Commercial (Orange)
500
Residential (Blue)
Note: Emissions and removals from Land Use, Land-Use Change, and Forestry are excluded from figure above. Excludes U.S.
Territories.
Table ES-6: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)
Economic Sectors
1990

2005

2015
2016
2017
2018
2019
Transportation
1,526.6

1,975.6

1,794.1
1,830.0
1,847.3
1,878.2
1,875.7
Electric Power Industry
1,875.7

2,456.3

1,950.0
1,857.6
1,778.9
1,798.0
1,648.1
Industry
1,640.7

1,518.8

1,441.6
1,402.2
1,423.4
1,483.3
1,504.8
Agriculture
600.2

629.7

658.5
645.8
646.6
662.0
669.5
Commercial
429.2

407.9

445.4
430.1
431.9
447.3
455.3
Residential
345.1

371.0

351.5
327.8
329.9
377.3
379.5
U.S. Territories
25.2

63.7

30.0
26.8
25.4
25.4
25.4
Total Emissions (Sources)
6,442.7

7,423.0

6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
LULUCF Sector Net Total3
(900.8)

(788.1)

(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Net Emissions (Sources and Sinks)
5,541.9

6,635.0

5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
Notes: Total emissions presented without LULUCF. Total net emissions presented with LULUCF. Totals may not sum
due to independent rounding. Parentheses indicate negative values or sequestration.
a The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net
carbon stock changes.
ES-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Using this categorization, emissions from transportation activities, in aggregate, accounted for the largest portion
(28.6 percent) of total U.S. greenhouse gas emissions in 2019. Electric power accounted for the second largest
portion (25.1 percent) of U.S. greenhouse gas emissions in 2019, while emissions from industry accounted for the
third largest portion (22.9 percent). Emissions from industry have in general declined over the past decade, due to
a number of factors, including structural changes in the U.S. economy (i.e., shifts from a manufacturing-based to a
service-based economy), fuel switching, and energy efficiency improvements.
The remaining 23.3 percent of U.S. greenhouse gas emissions were contributed by, in order of magnitude, the
agriculture, commercial, and residential sectors, plus emissions from U.S. Territories. Activities related to
agriculture accounted for 10.2 percent of U.S. emissions; unlike other economic sectors, agricultural sector
emissions were dominated by N20 emissions from agricultural soil management and CH4 emissions from enteric
fermentation. An increasing amount of carbon is stored in agricultural soils each year, but this C02 sequestration is
assigned to the LULUCF sector rather than the agriculture economic sector. The commercial and residential sectors
accounted for 6.9 percent and 5.8 percent of emissions, respectively, and U.S. Territories accounted for 0.4
percent of emissions; emissions from these sectors primarily consisted of C02 emissions from fossil fuel
combustion. Carbon dioxide was also emitted and sequestered by a variety of activities related to forest
management practices, tree planting in urban areas, the management of agricultural soils, landfilling of yard
trimmings, and changes in C stocks in coastal wetlands.
Electricity is ultimately used in the economic sectors described above. Table ES-7 presents greenhouse gas
emissions from economic sectors with emissions related to electric power distributed into end-use categories (i.e.,
emissions from electric power are allocated to the economic sectors in which the electricity is used). To distribute
electricity emissions among end-use sectors, emissions from the source categories assigned to electric power were
allocated to the residential, commercial, industry, transportation, and agriculture economic sectors according to
retail sales of electricity for each end-use sector (EIA 2020a and Duffield 20 06).25 These source categories include
C02 from fossil fuel combustion and the use of limestone and dolomite for flue gas desulfurization, C02 and N20
from incineration of waste, CH4 and N20 from stationary sources, and SF6 from electrical transmission and
distribution systems.
When emissions from electricity use are distributed among these end-use sectors, industrial activities and
transportation account for the largest shares of U.S. greenhouse gas emissions (29.7 percent and 28.7 percent,
respectively) in 2019. The commercial and residential sectors contributed the next largest shares of total U.S.
greenhouse gas emissions in 2019 (15.6 and 14.9 percent, respectively). Emissions from the commercial and
residential sectors increase substantially when emissions from electricity use are included, due to their relatively
large share of electricity use for energy (e.g., lighting, cooling, appliances). In all sectors except agriculture, C02
accounts for more than 79.0 percent of greenhouse gas emissions, primarily from the combustion of fossil fuels.
Figure ES-15 shows the trend in these emissions by sector from 1990 to 2019.
Table ES-7: U.S. Greenhouse Gas Emissions by Economic Sector with Electricity-Related
Emissions Distributed (MMT CO2 Eq.)
Economic Sectors
1990
2005
2015
2016
2017
2018
2019
Industry
2,313.1
2,234.1
1,964.2
1,894.6
1,902.7
1,958.3
1,947.2
Transportation
1,529.8
1,980.4
1,798.4
1,834.3
1,851.8
1,883.0
1,880.6
Commercial
983.4
1,229.8
1,125.7
1,080.8
1,054.5
1,067.8
1,022.3
Residential
956.0
1,247.1
1,053.1
998.9
963.7
1,035.9
978.3
Agriculture
635.3
668.0
699.7
684.9
685.3
701.1
704.6
U.S. Territories
25.2
63.7
30.0
26.8
25.4
25.4
25.4
Total Emissions (Sources)
6,442.7
7,423.0
6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
LULUCF Sector Net Total3
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Net Emissions (Sources and Sinks)
5,541.9
6,635.0
5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
25 U.S. Territories consumption data that are obtained from EIA are only available at the aggregate level and cannot be broken
out by end-use sector. The distribution of emissions to each end-use sector for the 50 states does not apply to territories data.
Executive Summary ES-27

-------
Notes: Emissions from electric power are allocated based on aggregate electricity use in each end-use sector. Totals
may not sum due to independent rounding. Parentheses indicate negative values or sequestration.
a The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon
stock changes.
Figure ES-15: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed
to Economic Sectors
2,500
Industry
2,000
Transportation
1,500
Commercial (Orange)
2 1,000
Residential (Blue)
Agriculture
500
Ol o
O tH
o o
Note: Emissions and removals from Land Use, Land-Use Change, and Forestry are excluded from figure above. Excludes U.S.
Territories.
Box ES-3: Trends in Various U.S. Greenhouse Gas Emissions-Related Data
Total greenhouse gas emissions can be compared to other economic and social indices to highlight changes over
time. These comparisons include: (1) emissions per unit of aggregate energy use, because energy-related
activities are the largest sources of emissions; (2) emissions per unit of fossil fuel consumption, because almost
all energy-related emissions involve the combustion of fossil fuels; (3) emissions per unit of total gross domestic
product as a measure of national economic activity; and (4) emissions per capita.
Table ES-8 provides data on various statistics related to U.S. greenhouse gas emissions normalized to 1990 as a
baseline year. These values represent the relative change in each statistic since 1990. Greenhouse gas emissions
in the United States have grown at an average annual rate of 0.1 percent since 1990, although changes from
year to year have been significantly larger. This growth rate is slightly slower than that for total energy use and
fossil fuel consumption, and overall gross domestic product (GDP), and national population (see Figure ES-16).
The direction of these trends started to change after 2005, when greenhouse gas emissions, total energy use
and fossil fuel consumption began to peak. Greenhouse gas emissions in the United States have decreased at an
average annual rate of 0.8percent since 2005. Fossil fuel consumption has also decreased at a slower rate than
emissions since 2005, while total energy use, GDP, and national population continued to increase.
Table ES-8: Recent Trends in Various U.S. Data (Index 1990 = 100)
Variable
1990
2005
2015
2016
2017
2018
2019
Avg. Annual
Growth Rate
Since 1990a
Avg. Annual
Growth Rate
Since 2005a
Greenhouse Gas Emissions'5
100

115
104
101
101
104
102
0.1%
-0.8%
Energy Usec
100

119
116
116
116
120
119
0.6%
0.0%
GDPd
100

159
186
189
194
200
204
2.5%
1.8%
ES-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Population6	100	118	128 129 130 131 132	L0%	0.8%
a Average annual growth rate.
b GWP-weighted values.
c Energy content-weighted values (EIA 2020a).
d GDP in chained 2009 dollars (BEA 2020).
0 U.S. Census Bureau (2020).
Figure ES-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic
Product (GDP)
Real GDP
200
180
160
Population
Energy Use
Emissions
Energy Use Per Capita
Emissions per Capita
Emissions per GDP
Source: BEA (2019), U.S. Census Bureau (2020), and emission estimates in this report.
Key Categories
The 2006IPCC Guidelines (IPCC 2006) defines a key category as a "[category] that is prioritized within the national
inventory system because its estimate has a significant influence on a country's total inventory of greenhouse
gases in terms of the absolute level, the trend, or the uncertainty in emissions and removals."26 A key category
analysis identifies priority source or sink categories for focusing efforts to improve overall Inventory quality. In
addition, a qualitative review of key categories and non-key categories can also help identify additional source and
sink categories to consider for improvement efforts, including reducing uncertainty.
Figure ES-17 presents the key categories identified by Approach 1 and Approach 2 level assessments including the
LULUCF sector for 2019. A level assessment using Approach 1 identifies all source and sink categories that
cumulatively account for 95 percent of total (i.e., gross) emissions in a given year when assessed in descending
order of absolute magnitude. An Approach 2 level assessment incorporates the results of the uncertainty analysis
for each category and identifies all sources and sink categories that cumulatively account for 90 percent of the sum
of all level assessments when sorted in descending order of magnitude.
26 See Chapter 4 "Methodological Choice and Identification of Key Categories" in IPCC (2006). See .
Executive Summary ES-29

-------
For a complete list of key categories and more information regarding the overall key category analysis, including
approaches accounting for the influence of trends of individual source and sink categories, see the Introduction
chapter, Section 1.5 - Key Categories and Annex 1.
Figure ES-17: 2019 Key Categories3
CO2 Emissions from Stationary Combustion - Coal - Electricity Generation
CO2 Emissions from Mobile Combustion: Road
Net Carbon Stock Change from Forest Land Remaining Forest Landb
CO2 Emissions from Stationary Combustion - Gas - Industrial
CO2 Emissions from Stationary Combustion - Oil - Industrial
Direct N2O Emissions from Agricultural Soil Management
CO2 Emissions from Stationary Combustion - Gas - Residential
CO2 Emissions from Mobile Combustion: Aviation
CH4 Emissions from Natural Gas Systems
CH4 Emissions from Landfills
CO2 Emissions from Stationary Combustion - Gas - Electricity Generation
CH4 Emissions from Enteric Fermentation: Cattle
CO2 Emissions from Stationary Combustion - Coal - Industrial
CO2 Emissions from Stationary Combustion - Gas - Commercial
CO2 Emissions from Non-Energy Use of Fuels
Net Carbon Stock Change from Settlements Remaining Settlements'"
CO2 Emissions from Iron and Steel Production & Metallurgical Coke Production
Net Carbon Stock Change from Land Converted to Forest Landb
CO2 Emissions from Stationary Combustion - Oil - Residential
CO2 Emissions from Stationary Combustion - Oil - Electricity Generation
Fugitive Emissions from Coal Mining
CO2 Emissions from Stationary Combustion - Oil - Commercial
Net Carbon Stock Change from Land Converted to Settlementsb
Net Carbon Stock Change from Land Converted to Croplandb
CH4 Emissions from Petroleum Systems
HFC-23 Emissions from HCFC-22 Production
Indirect N2O Emissions from Applied Nitrogen
CO2 Emissions from Mobile Combustion: Marine
N2O Emissions from Mobile Combustion: Road
CO2 Emissions from Mobile Combustion: Other
CO2 Emissions from Mobile Combustion: Railways
CO2 Emissions from Cement Production
CO2 Emissions from Natural Gas Systems
Net Carbon Stock Change from Cropland Remaining Croplandb
SF6 Emissions from Electrical Transmission and Distribution
CO2 Emissions from Petrochemical Production
PFC Emissions from Aluminum Production
N2O Emissions from Stationary Combustion - Coal - Electricity Generation
NzO Emissions from Wastewater Treatment
CH4 Emissions from Rice Cultivation
Net Carbon Stock Change from Grassland Remaining Grasslandb
CH4 Emissions from Abandoned Oil and Gas Wells
CH4 Emissions from Stationary Combustion - Residential
Key Categories as a Portion of
All Emissions
95.3%
Key Categories
I Key Categories LULUCF
Other Categories
400
800
1,200 1,600
2019 Emissions
a For a complete list of key categories and detailed discussion of the underlying key category analysis, see Annex 1. Bars indicate
key categories identified using Approach 1 and Approach 2 level assessment including the LULUCF sector.
b The absolute values of net C02 emissions from LULUCF are presented in this figure but reported separately from gross
emissions totals. Refer to Table ES-5 for a breakout of emissions and removals for LULUCF by gas and source/sink category.
Quality Assurance and Quality Control (QA/QC)
The United States seeks to continually improve the quality, transparency, and usability of the Inventory of U.S.
Greenhouse Gas Emissions and Sinks. To assist in these efforts, the United States implemented a systematic
approach to QA/QC. The procedures followed for the Inventory have been formalized in accordance with the U.S.
Inventory QA/QC plan for the Inventory, and the UNFCCC reporting guidelines and 2006IPCC Guidelines. The QA
process includes expert and public reviews for both the Inventory estimates and the Inventory report.
ES-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Box ES-4: Use of Ambient Measurements Systems for Validation of Emission Inventories
In following the UNFCCC requirement under Article 4.1 to develop and submit national greenhouse gas emission
inventories, the emissions and sinks presented in this report are organized by source and sink categories and
calculated using internationally-accepted methods provided by the IPCC.27 Several recent studies have
estimated emissions at the national or regional level with estimated results that sometimes differ from EPA's
estimate of emissions. EPA has engaged with researchers on how remote sensing, ambient measurement, and
inverse modeling techniques for estimating greenhouse gas emissions could assist in improving the
understanding of inventory estimates. In working with the research community on ambient measurement and
remote sensing techniques to improve national greenhouse gas inventories, EPA follows guidance from the IPCC
on the use of measurements and modeling to validate emission inventories.28 An area of particular interest in
EPA's outreach efforts is how ambient measurement data can be used in a manner consistent with this
Inventory report's transparency of its calculation methodologies, and the ability of these techniques to attribute
emissions and removals from remote sensing to anthropogenic sources, as defined by the IPCC for this report,
versus natural sources and sinks.
In an effort to improve the ability to compare the national-level greenhouse gas inventory with measurement
results that may be at other scales, a team at Harvard University along with EPA and other coauthors developed
a gridded inventory of U.S. anthropogenic methane emissions with 0.1° x 0.1° spatial resolution, monthly
temporal resolution, and detailed scale-dependent error characterization. The gridded inventory is designed to
be consistent with the 1990 to 2014 U.S. EPA Inventory of U.S. Greenhouse Gas Emissions and Sinks estimates
for the year 2012, which presents national totals for different source types.29 This gridded inventory is
consistent with the recommendations contained in two National Academies of Science reports examining
greenhouse gas emissions data (National Research Council 2010; National Academies of Sciences, Engineering,
and Medicine 2018).
Uncertainty Analysis of Emission Estimates
Uncertainty estimates are an essential element of a complete inventory of greenhouse gas emissions and removals
because they help to inform and prioritize inventory improvements. Recognizing the benefit of conducting an
uncertainty analysis, the UNFCCC reporting guidelines follow the recommendations of the 2006 IPCC Guidelines
(IPCC 2006), Volume 1, Chapter 3 and require that countries provide single estimates of uncertainty for source and
sink categories. In addition to quantitative uncertainty assessments, a qualitative discussion of uncertainty is
presented for each source and sink category identifying specific factors affecting the uncertainty surrounding the
estimates provided in accordance with UNFCCC reporting guidelines. Some of the current estimates, such as those
for C02 emissions from energy-related combustion activities, are considered to have low uncertainties. This is
because the amount of C02 emitted from energy-related combustion activities is directly related to the amount of
fuel consumed, the fraction of the fuel that is oxidized, and the carbon content of the fuel, and for the United
States, the uncertainties associated with estimating those factors is believed to be relatively small. For some other
categories of emissions, however, inherent variability or a lack of data increases the uncertainty or systematic
error associated with the estimates presented. Finally, an analysis is conducted to assess uncertainties associated
with the overall emissions, sinks and trends estimates. The overall uncertainty surrounding total net greenhouse
gas emissions is estimated to be -6 to + 6 percent in 1990 and -5 to +5 percent in 2019. When the LULUCF sector is
excluded from the analysis the uncertainty is estimated to be -2 to +5 percent in 1990 and -2 to +4 percent in 2019.
27	See .
28	See .
29	See .
Executive Summary ES-31

-------
1. Introduction
This report presents estimates by the United States government of U.S. anthropogenic greenhouse gas emissions
and sinks for the years 1990 through 2019. A summary of these estimates is provided in Table 2-1 and Table 2-2 by
gas and source category in the Trends in Greenhouse Gas Emissions chapter. The emission estimates in these
tables are presented on both a full mass basis and on a global warming potential (GWP) weighted basis1 in order to
show the relative contribution of each gas to global average radiative forcing. This report also discusses the
methods and data used to calculate these emission estimates.
In 1992, the United States signed and ratified the United Nations Framework Convention on Climate Change
(UNFCCC). As stated in Article 2 of the UNFCCC, 'The ultimate objective of this Convention and any related legal
instruments that the Conference of the Parties may adopt is to achieve, in accordance with the relevant provisions
of the Convention, stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent
dangerous anthropogenic interference with the climate system. Such a level should be achieved within a time-
frame sufficient to allow ecosystems to adapt naturally to climate change, to ensure that food production is not
threatened and to enable economic development to proceed in a sustainable manner."2'3
As a signatory to the UNFCCC, consistent with Article 44 and decisions at the First, Second, Fifth, and Nineteenth
Conference of Parties,5 the U.S. is committed to submitting a national inventory of anthropogenic sources and
sinks of greenhouse gases to the UNFCCC by April 15 of each year. The United States views this report, in
conjunction with Common Reporting Format (CRF) reporting tables that accompany this report, as an opportunity
to fulfill this annual commitment under the UNFCCC.
In 1988, preceding the creation of the UNFCCC, the World Meteorological Organization (WMO) and the United
Nations Environment Programme (UNEP) jointly established the Intergovernmental Panel on Climate Change
(IPCC). The role of the IPCC is to assess on a comprehensive, objective, open and transparent basis the scientific,
1	More information provided in the Global Warming Potentials section of this chapter on the use of IPCC Fourth Assessment
Report (AR4) GWP values.
2	The term "anthropogenic," in this context, refers to greenhouse gas emissions and removals that are a direct result of human
activities or are the result of natural processes that have been affected by human activities (IPCC 2006).
3	Article 2 of the Framework Convention on Climate Change published by the UNEP/WMO Information Unit on Climate Change
(UNEP/WMO 2000). See .
4	Article 4(l)(a) of the United Nations Framework Convention on Climate Change (also identified in Article 12) and subsequent
decisions by the Conference of the Parties elaborated the role of Annex I Parties in preparing national inventories. Article 4
states "Parties to the Convention, by ratifying, shall develop, periodically update, publish and make available...national
inventories of anthropogenic emissions by sources and removals by sinks of all greenhouse gases not controlled by the
Montreal Protocol, using comparable methodologies..." See  for more information.
5	See UNFCCC decisions 3/CP.l, 9/CP.2, 3/CP.5, and 24/CP.19 at https://unfccc.int/documents.
Introduction 1-1

-------
technical and socio-economic information relevant to understanding the scientific basis of risk of human-induced
climate change, its potential impacts and options for adaptation and mitigation (IPCC 2014). Under Working Group
1 of the IPCC, nearly 140 scientists and national experts from more than thirty countries collaborated in the
creation of the Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC/UNEP/OECD/IEA 1997)
to ensure that the emission inventories submitted to the UNFCCC are consistent and comparable between nations.
The IPCC Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories and the
IPCC Good Practice Guidance for Land Use, Land-Use Change, and Forestry further expanded upon the
methodologies in the Revised 1996 IPCC Guidelines. In 2006, the IPCC accepted the 2006 Guidelines for National
Greenhouse Gas Inventories at its Twenty-Fifth Session (Mauritius, April 2006). The 2006 IPCC Guidelines built upon
the previous bodies of work and include new sources and gases, "...as well as updates to the previously published
methods whenever scientific and technical knowledge have improved since the previous guidelines were issued."
The UNFCCC adopted the 2006 IPCC Guidelines as the standard methodological approach for Annex I countries and
encouraged countries to gain experience in using the 2013 Supplement to the 2006 IPCC Guidelines for National
Greenhouse Gas Inventories: Wetlands at the Nineteenth Conference of the Parties (Warsaw, November 11-23,
2013). The IPCC has recently released the 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse
Gas Inventories to clarify and elaborate on the existing guidance in the 2006 IPCC Guidelines, along with providing
updates to default values of emission factors and other parameters based on updated science. This report does
make use of both the 2013 Supplement and updated guidance in the 2019 Refinement to improve accuracy and
completeness of the Inventory. For more information on specific uses see Section 1.4 of this chapter on
Methodology and Data Sources.
Overall, this Inventory of anthropogenic greenhouse gas emissions and sinks provides a common and consistent
mechanism through which Parties to the UNFCCC can estimate emissions and compare the relative contribution of
individual sources, gases, and nations to climate change. The Inventory provides a national estimate of sources and
sinks for the United States, including all states, the District of Columbia and U.S. Territories.6 The structure of this
report is consistent with the current UNFCCC Guidelines on Annual Inventories (UNFCCC 2014) for Parties included
in Annex I of the Convention.
Box 1-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program
In following the UNFCCC requirement under Article 4.1 and decision 24/CP.19 to develop and submit annual
national greenhouse gas emission inventories, the emissions and removals presented in this report and this
chapter are organized by source and sink categories and calculated using internationally-accepted methods
provided by the IPCC in the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (2006 IPCC
Guidelines) and where appropriate, its supplements and refinements. Additionally, the calculated emissions and
removals in a given year for the United States are presented in a common format in line with the UNFCCC
reporting guidelines for the reporting of inventories under this international agreement. The use of consistent
methods to calculate emissions and removals by all nations providing their inventories to the UNFCCC ensures
that these reports are comparable. The presentation of emissions and removals provided in this Inventory does
not preclude alternative examinations, but rather this Inventory presents emissions and removals in a common
format consistent with how countries are to report Inventories under the UNFCCC. The report itself, and this
chapter, follows this standardized format, and provides an explanation of the application of methods used to
calculate emissions and removals.
EPA also collects greenhouse gas emissions data from individual facilities and suppliers of certain fossil fuels and
6 U.S. Territories include American Samoa, Guam, Commonwealth of the Northern Mariana Islands, Puerto Rico, U.S. Virgin
Islands, and other U.S. Pacific Islands which are not permanently inhabited such as Wake Island. See
.
1-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
industrial gases through its Greenhouse Gas Reporting Program (GHGRP).7 The GHGRP applies to direct
greenhouse gas emitters, fossil fuel suppliers, industrial gas suppliers, and facilities that inject carbon dioxide
(C02) underground for sequestration or other reasons and requires reporting by over 8,000 sources or suppliers
in 41 industrial categories.8 Annual reporting is at the facility level, except for certain suppliers of fossil fuels and
industrial greenhouse gases. In general, the threshold for reporting is 25,000 metric tons or more of C02 Eq. per
year. Facilities in most source categories subject to GHGRP began reporting for the 2010 reporting year while
additional types of industrial operations began reporting for reporting year 2011. While the GHGRP does not
provide full coverage of total annual U.S. greenhouse gas emissions and sinks (e.g., the GHGRP excludes
emissions from the agricultural, land use, and forestry sectors), it is an important input to the calculations of
national-level emissions in the Inventory.
Data presented in this Inventory report and EPA's GHGRP are complementary. The GHGRP dataset continues to
be an important resource for the Inventory, providing not only annual emissions information, but also other
annual information such as activity data and emission factors that can improve and refine national emission
estimates and trends over time. Methodologies used in EPA's GHGRP are consistent with the 2006IPCC
Guidelines (e.g., higher tier methods). GHGRP data also allow EPA to disaggregate national inventory estimates
in new ways that can highlight differences across regions and sub-categories of emissions, along with enhancing
the application of QA/QC procedures and assessment of uncertainties. EPA uses annual GHGRP data in a
number of categories to improve the national estimates presented in this Inventory consistent with IPCC
methodological guidance. See Annex 9 for more information on specific uses of GHGRP data in the Inventory
(e.g., natural gas systems).
1.1 Background Information
Science
For over the past 200 years, the burning of fossil fuels such as coal and oil, along with deforestation, land-use
changes, and other activities have caused the concentrations of heat-trapping "greenhouse gases" to increase
significantly in our atmosphere (NOAA 2017). These gases in the atmosphere absorb some of the energy being
radiated from the surface of the Earth that would otherwise be lost to space, essentially acting like a blanket that
makes the Earth's surface warmer than it would be otherwise.
Greenhouse gases are necessary to life as we know it. Without greenhouse gases to create the natural heat-
trapping properties of the atmosphere, the planet's surface would be about 60 degrees Fahrenheit cooler than
present (USGCRP 2017). Carbon dioxide is also necessary for plant growth. With emissions from biological and
geological sources, there is a natural level of greenhouse gases that is maintained in the atmosphere. Human
emissions of greenhouse gases and subsequent changes in atmospheric concentrations alter the balance of energy
transfers between space and the earth system (IPCC 2013). A gauge of these changes is called radiative forcing,
which is a measure of a substance's total net effect on the global energy balance for which a positive number
represents a warming effect and a negative number represents a cooling effect (IPCC 2013). IPCC concluded in its
most recent scientific assessment report that it is extremely likely that human influences have been the dominant
cause of warming since the mid-20th century (IPCC 2013).
7	On October 30, 2009 the EPA promulgated a rule requiring annual reporting of greenhouse gas data from large greenhouse
gas emissions sources in the United States. Implementation of the rule, codified at 40 CFR Part 98, is referred to as EPA's
Greenhouse Gas Reporting Program (GHGRP).
8	See  and .
Introduction 1-3

-------
As concentrations of greenhouse gases continue to increase in from man-made sources, the Earth's temperature is
climbing above past levels. The Earth's average land and ocean surface temperature has increased by about 1.8
degrees Fahrenheit from 1901 to 2016 (USGCRP 2017). The last three decades have each been the warmest
decade successively at the Earth's surface since 1850 (IPCC 2013). Other aspects of the climate are also changing,
such as rainfall patterns, snow and ice cover, and sea level. If greenhouse gas concentrations continue to increase,
climate models predict that the average temperature at the Earth's surface is likely to increase from 0.5 to 8.6
degrees Fahrenheit above 1986 through 2005 levels by the end of this century, depending on future emissions and
the responsiveness of the climate system (IPCC 2013).
For further information on greenhouse gases, radiative forcing, and implications for climate change, see the recent
scientific assessment reports from the IPCC,9 the U.S. Global Change Research Program (USGCRP),10 and the
National Academies of Sciences, Engineering, and Medicine (NAS).11
Greenhf
Although the Earth's atmosphere consists mainly of oxygen and nitrogen, neither plays a significant role in
enhancing the greenhouse effect because both are essentially transparent to terrestrial radiation. The greenhouse
effect is primarily a function of the concentration of water vapor, carbon dioxide (C02), methane (CH4), nitrous
oxide (N20), and other trace gases in the atmosphere that absorb the terrestrial radiation leaving the surface of
the Earth (IPCC 2013).
Naturally occurring greenhouse gases include water vapor, C02, CH4, N20, and ozone (03). Several classes of
halogenated substances that contain fluorine, chlorine, or bromine are also greenhouse gases, but they are, for the
most part, solely a product of industrial activities. Chlorofluorocarbons (CFCs) and hydrochlorofluorocarbons
(HCFCs) are halocarbons that contain chlorine, while halocarbons that contain bromine are referred to as
bromofluorocarbons (i.e., halons). As stratospheric ozone depleting substances, CFCs, HCFCs, and halons are
covered under the Montreal Protocol on Substances that Deplete the Ozone Layer. The UNFCCC defers to this
earlier international treaty. Consequently, Parties to the UNFCCC are not required to include these gases in
national greenhouse gas inventories.12 Some other fluorine-containing halogenated substances—
hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SF6), and nitrogen trifluoride (NF3)—do
not deplete stratospheric ozone but are potent greenhouse gases. These latter substances are addressed by the
UNFCCC and accounted for in national greenhouse gas inventories.
There are also several other substances that influence the global radiation budget but are short-lived and
therefore not well-mixed, leading to spatially variable radiative forcing effects. These substances include carbon
monoxide (CO), nitrogen dioxide (N02), sulfur dioxide (S02), and tropospheric (ground level) ozone (03).
Tropospheric ozone is formed from chemical reactions in the atmosphere of precursor pollutants, which include
volatile organic compounds (VOCs, including CH4) and nitrogen oxides (NOx), in the presence of ultraviolet light
(sunlight).
Aerosols are extremely small particles or liquid droplets suspended in the Earth's atmosphere that are often
composed of sulfur compounds, carbonaceous combustion products (e.g., black carbon), crustal materials (e.g.,
dust) and other human-induced pollutants. They can affect the absorptive characteristics of the atmosphere (e.g.,
scattering incoming sunlight away from the Earth's surface, or, in the case of black carbon, absorb sunlight) and
can play a role in affecting cloud formation and lifetime, as well as the radiative forcing of clouds and precipitation
9	See .
10	See .
11	See .
12	Emissions estimates of CFCs, HCFCs, halons and other ozone-depleting substances are included in this document for
informational purposes.
1-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
patterns. Comparatively, however, while the understanding of aerosols has increased in recent years, they still
account for the largest contribution to uncertainty estimates in global energy budgets (IPCC 2013).
Carbon dioxide, CH4, and N20 are continuously emitted to and removed from the atmosphere by natural processes
on Earth. Anthropogenic activities (such as fossil fuel combustion, cement production, land-use, land-use change,
and forestry, agriculture, or waste management), however, can cause additional quantities of these and other
greenhouse gases to be emitted or sequestered, thereby changing their global average atmospheric
concentrations. Natural activities such as respiration by plants or animals and seasonal cycles of plant growth and
decay are examples of processes that only cycle carbon or nitrogen between the atmosphere and organic biomass.
Such processes, except when directly or indirectly perturbed out of equilibrium by anthropogenic activities,
generally do not alter average atmospheric greenhouse gas concentrations over decadal timeframes. Climatic
changes resulting from anthropogenic activities, however, could have positive or negative feedback effects on
these natural systems. Atmospheric concentrations of these gases, along with their rates of growth and
atmospheric lifetimes, are presented in Table 1-1.
Table 1-1: Global Atmospheric Concentration, Rate of Concentration Change, and
Atmospheric Lifetime of Selected Greenhouse Gases
Atmospheric Variable
C02
ch4
N20
sf6
cf4
Pre-industrial atmospheric concentration
Atmospheric concentration
Rate of concentration change
Atmospheric lifetime (years)
280 ppm
411ppma
2.3 ppm/yrf
See footnote11
0.700 ppm
1.866 ppmb
7 ppb/yrf'g
12.4'
0.270 ppm
0.331 ppmc
0.8 ppb/yrf
121'
Oppt
9.9 pptd
0.27 ppt/yrf
3,200
40 ppt
79 ppt0
0.7 ppt/yrf
50,000
a The atmospheric C02 concentration is the 2019 annual average at the Mauna Loa, HI station (NOAA/ESRL 2021a). The
concentration in 2019 at Mauna Loa was 411 ppm. The global atmospheric C02 concentration, computed using an average of
sampling sites across the world, was 409 ppm in 2019.
b The values presented are global 2019 annual average mole fractions (NOAA/ESRL 2021b).
c The values presented are global 2019 annual average mole fractions (NOAA/ESRL 2021c).
d The values presented are global 2019 annual average mole fractions (NOAA/ESRL 2021d).
6 The 2011 CF4 global mean atmospheric concentration is from the Advanced Global Atmospheric Gases Experiment (IPCC 2013).
f The rate of concentration change for C02 and CH4 is the average rate of change between 2007 and 2019 (NOAA/ESRL 2021a).
The rate of concentration change for N20, SF6, and CF4 is the average rate of change between 2005 and 2011 (IPCC 2013).
g The growth rate for atmospheric CH4 decreased from over 10 ppb/year in the 1980s to nearly zero in the early 2000s; recently,
the growth rate has been about 7 ppb/year.
h For a given amount of C02 emitted, some fraction of the atmospheric increase in concentration is quickly absorbed by the
oceans and terrestrial vegetation, some fraction of the atmospheric increase will only slowly decrease over a number of years,
and a small portion of the increase will remain for many centuries or more.
1 This lifetime has been defined as an "adjustment time" that takes into account the indirect effect of the gas on its own
residence time.
Source: Pre-industrial atmospheric concentrations, atmospheric lifetime, and rate of concentration changes for CH4, N20, SF6,
and CF4 are from IPCC (2013). The rate of concentration change for C02 is an average of the rates from 2007 through 2019 and
has fluctuated between 1.5 to 3.0 ppm per year over this period (NOAA/ESRL 2021a).
A brief description of each greenhouse gas, its sources, and its role in the atmosphere is given below. The following
section then explains the concept of GWPs, which are assigned to individual gases as a measure of their relative
average global radiative forcing effect.
Water Vapor (H20). Water vapor is the largest contributor to the natural greenhouse effect. Water vapor is
fundamentally different from other greenhouse gases in that it can condense and rain out when it reaches high
concentrations, and the total amount of water vapor in the atmosphere is in part a function of the Earth's
temperature. While some human activities such as evaporation from irrigated crops or power plant cooling release
water vapor into the air, these activities have been determined to have a negligible effect on global climate (IPCC
2013). The lifetime of water vapor in the troposphere is on the order of 10 days. Water vapor can also contribute
to cloud formation, and clouds can have both warming and cooling effects by either trapping or reflecting heat.
Because of the relationship between water vapor levels and temperature, water vapor and clouds serve as a
feedback to climate change, such that for any given increase in other greenhouse gases, the total warming is
Introduction 1-5

-------
greater than would happen in the absence of water vapor. Aircraft emissions of water vapor can create contrails,
which may also develop into contrail-induced cirrus clouds, with complex regional and temporal net radiative
forcing effects that currently have a low level of scientific certainty (IPCC 2013).
Carbon Dioxide (C02). In nature, carbon is cycled between various atmospheric, oceanic, land biotic, marine biotic,
and mineral reservoirs. The largest fluxes occur between the atmosphere and terrestrial biota, and between the
atmosphere and surface water of the oceans. In the atmosphere, carbon predominantly exists in its oxidized form
as C02. Atmospheric C02 is part of this global carbon cycle, and therefore its fate is a complex function of
geochemical and biological processes. Carbon dioxide concentrations in the atmosphere increased from
approximately 280 parts per million by volume (ppmv) in pre-industrial times to 411 ppmv in 2019, a 47 percent
increase (IPCC 2013; NOAA/ESRL 2021a).1314 The IPCC definitively states that "the increase of C02... is caused by
anthropogenic emissions from the use of fossil fuel as a source of energy and from land use and land use changes,
in particular agriculture" (IPCC 2013). The predominant source of anthropogenic C02 emissions is the combustion
of fossil fuels. Forest clearing, other biomass burning, and some non-energy production processes (e.g., cement
production) also emit notable quantities of C02. In its Fifth Assessment Report, the IPCC stated "it is extremely
likely that more than half of the observed increase in global average surface temperature from 1951 to 2010 was
caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings
together," of which C02 is the most important (IPCC 2013).
Methane (CH4). Methane is primarily produced through anaerobic decomposition of organic matter in biological
systems. Agricultural processes such as wetland rice cultivation, enteric fermentation in animals, and the
decomposition of animal wastes emit CH4, as does the decomposition of municipal solid wastes and treatment of
wastewater. Methane is also emitted during the production and distribution of natural gas and petroleum, and is
released as a byproduct of coal mining and incomplete fossil fuel combustion. Atmospheric concentrations of CH4
have increased by about 167 percent since 1750, from a pre-industrial value of about 700 ppb to 1,866 ppb in
201915 although the rate of increase decreased to near zero in the early 2000s, and has recently increased again to
about 7 ppb/year. The IPCC has estimated that slightly more than half of the current CH4 flux to the atmosphere is
anthropogenic, from human activities such as agriculture, fossil fuel production and use, and waste disposal (IPCC
2007).
Methane is primarily removed from the atmosphere through a reaction with the hydroxyl radical (OH) and is
ultimately converted to C02. Minor removal processes also include reaction with chlorine in the marine boundary
layer, a soil sink, and stratospheric reactions. Increasing emissions of CH4 reduce the concentration of OH, a
feedback that increases the atmospheric lifetime of CH4 (IPCC 2013). Methane's reactions in the atmosphere also
lead to production of tropospheric ozone and stratospheric water vapor, both of which also contribute to climate
change.
Nitrous Oxide (N20). Anthropogenic sources of N20 emissions include agricultural soils, especially production of
nitrogen-fixing crops and forages, the use of synthetic and manure fertilizers, and manure deposition by livestock;
fossil fuel combustion, especially from mobile combustion; adipic (nylon) and nitric acid production; wastewater
treatment and waste incineration; and biomass burning. The atmospheric concentration of N20 has increased by
23 percent since 1750, from a pre-industrial value of about 270 ppb to 331 ppb in 2019,16 a concentration that has
not been exceeded during the last 800 thousand years. Nitrous oxide is primarily removed from the atmosphere by
the photolytic action of sunlight in the stratosphere (IPCC 2013).
13	The pre-industrial period is considered as the time preceding the year 1750 (IPCC 2013).
14	Carbon dioxide concentrations during the last 1,000 years of the pre-industrial period (i.e., 750 to 1750), a time of relative
climate stability, fluctuated by about +10 ppmv around 280 ppmv (IPCC 2013).
15	This value is the global 2019 annual average mole fraction (NOAA/ESRL 2021b).
16	This value is the global 2019 annual average (NOAA/ESRL 2021c).
1-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Ozone (O3). Ozone is present in both the upper stratosphere,17 where it shields the Earth from harmful levels of
ultraviolet radiation, and at lower concentrations in the troposphere,18 where it is the main component of
anthropogenic photochemical "smog." During the last two decades, emissions of anthropogenic chlorine and
bromine-containing halocarbons, such as CFCs, have depleted stratospheric ozone concentrations. This loss of
ozone in the stratosphere has resulted in negative radiative forcing, representing an indirect effect of
anthropogenic emissions of chlorine and bromine compounds (IPCC 2013). The depletion of stratospheric ozone
and its radiative forcing remains relatively unchanged since 2000 and recovery is expected to start occurring in the
middle of the twenty-first century (WMO/UNEP 2014; WMO 2015).
The past increase in tropospheric ozone, which is also a greenhouse gas, is estimated to provide the fourth largest
increase in direct radiative forcing since the pre-industrial era, behind C02, black carbon, and CH4. Tropospheric
ozone is produced from complex chemical reactions of volatile organic compounds (including CH4) mixing with NOx
in the presence of sunlight. The tropospheric concentrations of ozone and these other pollutants are short-lived
and, therefore, spatially variable (IPCC 2013).
Halocarbons, Sulfur Hexafluoride, and Nitrogen Trifluoride. Halocarbons are, for the most part, man-made
chemicals that have direct radiative forcing effects and could also have an indirect effect. Halocarbons that contain
chlorine (CFCs, HCFCs, methyl chloroform, and carbon tetrachloride) and bromine (halons, methyl bromide, and
hydrobromofluorocarbons) result in stratospheric ozone depletion and are therefore controlled under the
Montreal Protocol on Substances that Deplete the Ozone Layer. Although most CFCs and HCFCs are potent global
warming gases, their net radiative forcing effect on the atmosphere is reduced because they cause stratospheric
ozone depletion, which itself is a greenhouse gas but which also shields the Earth from harmful levels of ultraviolet
radiation. Under the Montreal Protocol, the United States phased out the production and importation of halons by
1994 and of CFCs by 1996. Under the Copenhagen Amendments to the Protocol, a cap was placed on the
production and importation of HCFCs by non-Article 5 countries, including the United States,19 beginning in 1996,
and then followed by intermediate requirements and a complete phase-out by the year 2030. While ozone
depleting gases covered under the Montreal Protocol and its Amendments are not covered by the UNFCCC, they
are reported in this Inventory under Annex 6.2 for informational purposes.
Hydrofluorocarbons, PFCs, SF6, and NF3 are not ozone depleting substances. The most common HFCs are, however,
powerful greenhouse gases. Hydrofluorocarbons are primarily used as replacements for ozone depleting
substances but also emitted as a byproduct of the HCFC-22 (chlorodifluoromethane) manufacturing process.
Currently, they have a small aggregate radiative forcing impact, but it is anticipated that without further controls
their contribution to overall radiative forcing will increase (IPCC 2013). In 2020, the U.S. Congress passed
legislation designed to phase down the production and consumption of HFCs in the U.S., which would lead to
lower emissions over time.
Perfluorocarbons, SF6, and NF3 are predominantly emitted from various industrial processes including aluminum
smelting, semiconductor manufacturing, electric power transmission and distribution, and magnesium casting.
Currently, the radiative forcing impact of PFCs, SF6, and NF3 is also small, but they have a significant growth rate,
17	The stratosphere is the layer from the troposphere up to roughly 50 kilometers. In the lower regions the temperature is
nearly constant but in the upper layer the temperature increases rapidly because of sunlight absorption by the ozone layer. The
ozone-layer is the part of the stratosphere from 19 kilometers up to 48 kilometers where the concentration of ozone reaches
up to 10 parts per million.
18	The troposphere is the layer from the ground up to 11 kilometers near the poles and up to 16 kilometers in equatorial
regions (i.e., the lowest layer of the atmosphere where people live). It contains roughly 80 percent of the mass of all gases in
the atmosphere and is the site for most weather processes, including most of the water vapor and clouds.
19	Article 5 of the Montreal Protocol covers several groups of countries, especially developing countries, with low consumption
rates of ozone depleting substances. Developing countries with per capita consumption of less than 0.3 kg of certain ozone
depleting substances (weighted by their ozone depleting potential) receive financial assistance and a grace period of ten
additional years in the phase-out of ozone depleting substances.
Introduction 1-7

-------
extremely long atmospheric lifetimes, and are strong absorbers of infrared radiation, and therefore have the
potential to influence climate far into the future (IPCC 2013).
Carbon Monoxide (CO). Carbon monoxide has an indirect radiative forcing effect by elevating concentrations of CH4
and tropospheric ozone through chemical reactions with other atmospheric constituents (e.g., the hydroxyl radical,
OH) that would otherwise assist in destroying CH4 and tropospheric ozone. Carbon monoxide is created when
carbon-containing fuels are burned incompletely. Through natural processes in the atmosphere, it is eventually
oxidized to C02. Carbon monoxide concentrations are both short-lived in the atmosphere and spatially variable.
Nitrogen Oxides (NOx). The primary climate change effects of nitrogen oxides (i.e., NO and N02) are indirect.
Warming effects can occur due to reactions leading to the formation of ozone in the troposphere, but cooling
effects can occur due to the role of NOx as a precursor to nitrate particles (i.e., aerosols) and due to destruction of
stratospheric ozone when emitted from very high-altitude aircraft.20 Additionally, NOx emissions are also likely to
decrease CH4 concentrations, thus having a negative radiative forcing effect (IPCC 2013). Nitrogen oxides are
created from lightning, soil microbial activity, biomass burning (both natural and anthropogenic fires) fuel
combustion, and, in the stratosphere, from the photo-degradation of N20. Concentrations of NOx are both
relatively short-lived in the atmosphere and spatially variable.
Non-methane Volatile Organic Compounds (NMVOCs). Non-methane volatile organic compounds include
substances such as propane, butane, and ethane. These compounds participate, along with NOx, in the formation
of tropospheric ozone and other photochemical oxidants. NMVOCs are emitted primarily from transportation and
industrial processes, as well as biomass burning and non-industrial consumption of organic solvents.
Concentrations of NMVOCs tend to be both short-lived in the atmosphere and spatially variable.
Aerosols. Aerosols are extremely small particles or liquid droplets found in the atmosphere that are either directly
emitted into or are created through chemical reactions in the Earth's atmosphere. Aerosols or their chemical
precursors can be emitted by natural events such as dust storms, biogenic or volcanic activity, or by anthropogenic
processes such as transportation, coal combustion, cement manufacturing, waste incineration, or biomass burning.
Various categories of aerosols exist from both natural and anthropogenic sources, such as soil dust, sea salt,
biogenic aerosols, sulfates, nitrates, volcanic aerosols, industrial dust, and carbonaceous21 aerosols (e.g., black
carbon, organic carbon). Aerosols can be removed from the atmosphere relatively rapidly by precipitation or
through more complex processes under dry conditions.
Aerosols affect radiative forcing differently than greenhouse gases. Their radiative effects occur through direct and
indirect mechanisms: directly by scattering and absorbing solar radiation (and to a lesser extent scattering,
absorption, and emission of terrestrial radiation); and indirectly by increasing cloud droplets and ice crystals that
modify the formation, precipitation efficiency, and radiative properties of clouds (IPCC 2013). Despite advances in
understanding of cloud-aerosol interactions, the contribution of aerosols to radiative forcing are difficult to
quantify because aerosols generally have short atmospheric lifetimes, and have number concentrations, size
distributions, and compositions that vary regionally, spatially, and temporally (IPCC 2013).
The net effect of aerosols on the Earth's radiative forcing is believed to be negative (i.e., net cooling effect on the
climate). In fact, "despite the large uncertainty ranges on aerosol forcing, there is high confidence that aerosols
have offset a substantial portion of greenhouse gas forcing" (IPCC 2013).22 Although because they remain in the
atmosphere for only days to weeks, their concentrations respond rapidly to changes in emissions.23 Not all
aerosols have a cooling effect. Current research suggests that another constituent of aerosols, black carbon, has a
20	NOx emissions injected higher in the stratosphere, primarily from fuel combustion emissions from high altitude supersonic
aircraft, can lead to stratospheric ozone depletion.
21	Carbonaceous aerosols are aerosols that are comprised mainly of organic substances and forms of black carbon (or soot)
(IPCC 2013).
22	The IPCC (2013) defines high confidence as an indication of strong scientific evidence and agreement in this statement.
23	Volcanic activity can inject significant quantities of aerosol producing sulfur dioxide and other sulfur compounds into the
stratosphere, which can result in a longer negative forcing effect (i.e., a few years) (IPCC 2013).
1-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
positive radiative forcing by heating the Earth's atmosphere and causing surface warming when deposited on ice
and snow (IPCC 2013). Black carbon also influences cloud development, but the direction and magnitude of this
forcing is an area of active research.
A global warming potential is a quantified measure of the globally averaged relative radiative forcing impacts of a
particular greenhouse gas (see Table 1-2). It is defined as the accumulated radiative forcing within a specific time
horizon caused by emitting 1 kilogram (kg) of the gas, relative to that of the reference gas C02 (IPCC 2014). Direct
radiative effects occur when the gas itself absorbs radiation. Indirect radiative forcing occurs when chemical
transformations involving the original gas produce a gas or gases that are greenhouse gases, or when a gas
influences other radiatively important processes such as the atmospheric lifetimes of other gases. The reference
gas used is C02, and therefore GWP-weighted emissions are measured in million metric tons of C02 equivalent
(MMT C02 Eq.).24 The relationship between kilotons (kt) of a gas and MMT C02 Eq. can be expressed as follows:
MMT C02 Eq. = Million metric tons of C02 equivalent
kt = kilotons (equivalent to a thousand metric tons)
GWP = Global warming potential
MMT = Million metric tons
GWP values allow for a comparison of the impacts of emissions and reductions of different gases.
According to the IPCC, GWPs typically have an uncertainty of ±35 percent. Parties to the UNFCCC have
also agreed to use GWPs based upon a 100-year time horizon, although other time horizon values are
available.
...the global warming potential values used by Parties included in Annex I to the Convention (Annex I Parties)
to calculate the carbon dioxide equivalence of anthropogenic emissions by sources and removals by sinks of
greenhouse gases shall be those listed in the column entitled "Global warming potential for given time
horizon" in table 2.14 of the errata to the contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change, based on the effects of greenhouse gases over a 100-year
time horizon...25
Greenhouse gases with relatively long atmospheric lifetimes (e.g., C02, CH4, N20, HFCs, PFCs, SF6, NF3) tend to be
evenly distributed throughout the atmosphere, and consequently global average concentrations can be
determined. The short-lived gases such as water vapor, carbon monoxide, tropospheric ozone, ozone precursors
(e.g., NOx, and NMVOCs), and tropospheric aerosols (e.g., S02 products and carbonaceous particles), however, vary
regionally, and consequently it is difficult to quantify their global radiative forcing impacts. Parties to the UNFCCC
have not agreed upon GWP values for these gases that are short-lived and spatially inhomogeneous in the
atmosphere.
24	Carbon comprises 12/44ths of carbon dioxide by weight.
25	Framework Convention on Climate Change; Available online at:
; 31 January 2014; Report of the Conference of the Parties at its
nineteenth session; held in Warsaw from 11 to 23 November 2013; Addendum; Part two: Action taken by the Conference of the
Parties at its nineteenth session; Decision 24/CP.19; Revision of the UNFCCC reporting guidelines on annual inventories for
Parties included in Annex I to the Convention; p. 2. (UNFCCC 2014).
Global Warming Potentials
s	s / MMT \
Eq. = (kt of gas) x (GWP) x [1QQQkt)
MMT CO-
where,
Introduction 1-9

-------
Table 1-2: Global Warming Potentials and Atmospheric Lifetimes (Years) Used in this Report
Gas
Atmospheric Lifetime
GWPa
C02
See footnote15
1
ch4c
12
25
n2o
114
298
HFC-23
270
14,800
HFC-32
4.9
675
HFC-41d
3.7
92
HFC-125
29
3,500
HFC-134a
14
1,430
HFC-143a
52
4,470
HFC-152a
1.4
124
HFC-227ea
34.2
3,220
HFC-236fa
240
9,810
cf4
50,000
7,390
c2f6
10,000
12,200
C3Fs
2,600
8,830
c-C4Fs
3,200
10,300
sf6
3,200
22,800
nf3
740
17,200
Other Fluorinated Gases

See Annex 6
a 100-year time horizon.
b For a given amount of C02 emitted, some fraction of the
atmospheric increase in concentration is quickly absorbed by the
oceans and terrestrial vegetation, some fraction of the atmospheric
increase will only slowly decrease over a number of years, and a
small portion of the increase will remain for many centuries or
more.
c The GWP of CH4 includes the direct effects and those indirect
effects due to the production of tropospheric ozone and
stratospheric water vapor. The indirect effect due to the production
of C02 is not included.
d See Table A-l of 40 CFR Part 98
Source: IPCC 2013.
Box 1-2: The IPCC Fifth Assessment Report and Global Warming Potentials
In 2014, the IPCC published its Fifth Assessment Report (AR5), which updated its comprehensive scientific
assessment of climate change. Within the AR5 report, the GWP values of gases were revised relative to previous
IPCC reports, namely the IPCC Second Assessment Report (SAR) (IPCC 1996), the IPCC Third Assessment Report
(TAR) (IPCC 2001), and the IPCC Fourth Assessment Report (AR4) (IPCC 2007). Although the AR4 GWP values are
used throughout this report, consistent with UNFCCC reporting requirements, it is straight-forward to review
the changes to the GWP values and their impact on estimates of the total GWP-weighted emissions of the
United States. In the AR5, the IPCC applied an improved calculation of C02 radiative forcing and an improved C02
response function in presenting updated GWP values. Additionally, the atmospheric lifetimes of some gases
have been recalculated, and updated background concentrations were used. In addition, the values for radiative
forcing and lifetimes have been recalculated for a variety of halocarbons, and the indirect effects of methane on
ozone have been adjusted to match more recent science. Table 1-3 presents the new GWP values, relative to
those presented in the AR4 and using the 100-year time horizon common to UNFCCC reporting.
For consistency with international reporting standards under the UNFCCC, official emission estimates are
reported by the United States using AR4 GWP values, as required by the 2013 revision to the UNFCCC reporting
1-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
guidelines for national inventories.26 All estimates provided throughout this report are also presented in
unweighted units. For informational purposes, emission estimates that use GWPs from other IPCC Assessment
Reports are presented in detail in Annex 6.1 of this report.
Table 1-3: Comparison of 100-Year GWP values
100-Year GWP Values
Comparison to AR4




AR5 with


AR5 with
Gas
SAR
AR4
AR5a
feedbacks'1
SAR
AR5
feedbacks'1
C02
1
1
1
1
NC
NC
NC
ch4c
21
25
28
34
(4)
3
9
n2o
310
298
265
298
12
(33)
NC
HFC-23
11,700
14,800
12,400
13,856
(3,100)
(2,400)
(944)
HFC-32
650
675
677
817
(25)
2
142
HFC-41
NA
92
116
NA
NA
24
NA
HFC-125
2,800
3,500
3,170
3,691
(700)
(330)
191
HFC-134a
1,300
1,430
1,300
1,549
(130)
(130)
119
HFC-143a
3,800
4,470
4,800
5,508
(670)
330
1,038
HFC-152a
140
124
138
167
16
14
43
HFC-227ea
2,900
3,220
3,350
3,860
(320)
130
640
HFC-236fa
6,300
9,810
8,060
8,998
(3,510)
(1,750)
(812)
cf4
6,500
7,390
6,630
7,349
(890)
(760)
(41)
c2f6
9,200
12,200
11,100
12,340
(3,000)
(1,100)
140
C3Fs
7,000
8,830
8,900
9,878
(1,830)
70
(1,048)
c-C4Fs
8,700
10,300
9,540
10,592
(1,600)
(760)
292
sf6
23,900
22,800
23,500
26,087
1,100
700
3,287
nf3
NA
17,200
16,100
17,885
NA
(1,100)
685
Source: IPCC 2013, IPCC 2007, IPCC 2001, IPCC 1996.
Note: Parentheses indicate negative values.
NA (Not Applicable)
NC (No Change)
a The GWPs presented here are the ones most consistent with the methodology used in the AR4
report.
b The GWP values presented here from the AR5 report include climate-carbon feedbacks for the non-
C02 gases in order to be consistent with the approach used in calculating the C02 lifetime.
c The GWP of CH4 includes the direct effects and those indirect effects due to the production of
tropospheric ozone and stratospheric water vapor. Including the indirect effect due to the production
of C02 resulting from methane oxidation would lead to an increase in AR5 methane GWP values by 2
for fossil methane and is not shown in this table.
1.2 National Inventory Arrangements
The U.S. Environmental Protection Agency (EPA), in cooperation with other U.S. government agencies, prepares
the Inventory of U.S. Greenhouse Gas Emissions and Sinks. A wide range of agencies and individuals are involved in
supplying data to, planning methodological approaches and improvements, reviewing, or preparing portions of the
U.S. Inventory—including federal and state government authorities, research and academic institutions, industry
associations, and private consultants.
26 See .
Introduction 1-11

-------
Within EPA, the Office of Atmospheric Programs (OAP) is the lead office responsible for the emission calculations
provided in the Inventory, as well as the completion of the National Inventory Report and the Common Reporting
Format (CRF) tables. EPA's Office of Transportation and Air Quality (OTAQ) is also involved in calculating emissions
for the Inventory. The U.S. Department of State serves as the overall focal point to the UNFCCC, and EPA's OAP
serves as the National Inventory Focal Point for this report, including responding to technical questions and
comments on the U.S. Inventory. EPA staff coordinate the annual methodological choice, activity data collection,
emission calculations, QA/QC processes, and improvement planning at the individual source and sink category
level. EPA, the inventory coordinator, compiles the entire Inventory into the proper reporting format for
submission to the UNFCCC, and is responsible for the synthesis of information and for the consistent application of
cross-cutting IPCC good practice across the Inventory.
Several other government agencies contribute to the collection and analysis of the underlying activity data used in
the Inventory calculations via formal (e.g., interagency agreements) and informal relationships, in addition to the
calculation of estimates integrated in the report (e.g., U.S. Department of Agriculture's U.S. Forest Service and
Agricultural Service, National Oceanic and Atmospheric Administration, Federal Aviation Administration, and
Department of Defense). Other U.S. agencies provide official data for use in the Inventory. The U.S. Department of
Energy's Energy Information Administration provides national fuel consumption data and the U.S. Department of
Defense provides data on military fuel consumption and use of bunker fuels. Other U.S. agencies providing activity
data for use in EPA's emission calculations include: the U.S. Department of Agriculture, National Oceanic and
Atmospheric Administration, the U.S. Geological Survey, the Federal Highway Administration, the Department of
Transportation, the Bureau of Transportation Statistics, the Department of Commerce, and the Federal Aviation
Administration. Academic and research centers also provide activity data and calculations to EPA, as well as
individual companies participating in voluntary outreach efforts with EPA. Finally, EPA as the National Inventory
Focal Point, in coordination with the U.S. Department of State, officially submits the Inventory to the UNFCCC each
April.
Figure 1-1: National Inventory Arrangements and Process Diagram
United States Greenhouse Gas Inventory Institutional Arrangements
1. Data Collection
Energy Data Sources
Agriculture and
LULUCF Data Sources
Industrial Processes
and Product Use Data
Waste Data Sources
2. Emissions
Calculations
U.S. Environmental
Protection Agency
Government Agencies
(USF5, NOAA,
DOD, U5GS, FAA)
3. Inventory
Compilation
U.S. Environmental
Protection Agency
Inventory Compiler
4. Inventory
Submission
U.S. Department
of State
United Nations
Framework
Convention on
Climate Change
Overview of Inventory Data Sources by Source and Sink Category.
Energy
Agriculture and LULUCF IPPU Waste
Energy Information
Administration
EPA Office of Land and EmergencyEPA Greenhouse Gas Reporting EPA Greenhouse Gas
Management Program (GHGRP) Reporting Program (GHGRP)
U.S. Department of Commerce Alaska Department of Natural American Chemistry Council EPA Office of Land and
- Bureau of the Census	Resources	(ACC)	Emergency Management
1-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
U.S. Department of Defense -
Defense Logistics Agency
Federal Highway
Administration
EPA Acid Rain Program
EPA Office of Transportation
and Air Quality MOVES Model
EPA Greenhouse Gas Reporting
Program (GHGRP)
U.S. Department of Labor -
Mine Safety and Health
Administration
American Association of
Railroads
American Public Transportation
Association
U.S. Department of Homeland
Security
U.S. Department of Energy and
its National Laboratories
National Oceanic and
Atmospheric Administration
(NOAA)
Association of American Plant
Food Control Officials (AAPFCO)
U.S. Census Bureau
USDA Animal and Plant Health
Inspection Service (APHIS)
EPA Office of Air and Radiation
U.S. Department of Agriculture
(USDA) National Agricultural
Statistics Service and Agricultural
Research Service
USDA U.S. Forest Service Forest
Inventory and Analysis Program
USDA Natural Resource
Conservation Service (NRCS)
USDA Economic Research Service
(ERS)
USDA Farm Service Agency (FSA)
U.S. Geological Survey (USGS)
National Minerals Information
Center
American Iron and Steel
Institute (AISI)
U.S. Aluminum Association
U.S. International Trade
Commission (USITC)
Air-Conditioning, Heating, and
Refrigeration Institute
Data from other U.S.
government agencies, research
studies, trade publications, and
industry associations
Data from research studies,
trade publications, and
industry associations
Federal Aviation Administration U.S. Geological Survey (USGS)
U.S. Department of	U.S. Department of the Interior
Transportation & Bureau of (DOE), Bureau of Land
Transportation Statistics	Management (BLM)
Data from research studies, Data from research studies, trade
trade publications, and industrypublications, and industry
associations	associations
Note: This table is not an exhaustive list of all data sources.
1.3 Inventory Process
This section describes EPA's approach to preparing the annual U.S. Inventory, which consists of a National
Inventory Report (NIR) and Common Reporting Format (CRF) tables. The inventory coordinator at EPA, with
support from the cross-cutting compilation staff, is responsible for aggregating all emission estimates, conducting
the overall uncertainty analysis of Inventory emissions and trends over time, and ensuring consistency and quality
throughout the NIR and CRF tables. Emission calculations, including associated uncertainty analysis for individual
sources and/or sink categories are the responsibility of individual source and sink category leads, who are most
familiar with each category, underlying data, and the unique national circumstances relevant to its emissions or
removals profile. Using IPCC good practice guidance, the individual leads determine the most appropriate
methodology and collect the best activity data to use in the emission and removal calculations, based upon their
expertise in the source or sink category, as well as coordinating with researchers and contractors familiar with the
sources. Each year, the coordinator overseas a multi-stage process for collecting information from each individual
source and sink category lead to compile all information and data for the Inventory.
Methodology Development, Data Collection, and Emissions
and Sink Estimation
Source and sink category leads at EPA collect input data and, as necessary, evaluate or develop the estimation
methodology for the individual source and/or sink categories. Because EPA has been preparing the Inventory for
many years, for most source and sink categories, the methodology for the previous year is applied to the new
Introduction 1-13

-------
"current" year of the Inventory, and inventory analysts collect any new data or update data that have changed
from the previous year. If estimates for a new source or sink category are being developed for the first time, or if
the methodology is changing for an existing category (e.g., the United States is implementing improvement efforts
to apply a higher tiered approach for that category), then the source and/or sink category lead will develop and
implement the new or refined methodology, gather the most appropriate activity data and emission factors (or in
some cases direct emission measurements) for the entire time series, and conduct any further category-specific
review with involvement of relevant experts from industry, government, and universities (see Box ES-4 on EPA's
approach to recalculations).
Once the methodology is in place and the data are collected, the individual source and sink category leads
calculate emission and removal estimates. The individual leads then update or create the relevant text and
accompanying annexes for the Inventory. Source and sink category leads are also responsible for completing the
relevant sectoral background tables of the CRF, conducting quality assurance and quality control (QA/QC) checks,
and category-level uncertainty analyses.
The treatment of confidential business information (CBI) in the Inventory is based on EPA internal guidelines, as
well as regulations27 applicable to the data used. EPA has specific procedures in place to safeguard CBI during the
inventory compilation process. When information derived from CBI data is used for development of inventory
calculations, EPA procedures ensure that these confidential data are sufficiently aggregated to protect
confidentiality while still providing useful information for analysis. For example, within the Energy and Industrial
Processes and Product Use (IPPU) sectors, EPA has used aggregated facility-level data from the Greenhouse Gas
Reporting Program (GHGRP) to develop, inform, and/or quality-assure U.S. emission estimates. In 2014, EPA's
GHGRP, with industry engagement, compiled criteria that would be used for aggregating its confidential data to
shield the underlying CBI from public disclosure.28 In the Inventory, EPA is publishing only data values that meet
the GHGRP aggregation criteria.29 Specific uses of aggregated facility-1 eve I data are described in the respective
methodological sections within those chapters. In addition, EPA uses historical data reported voluntarily to EPA via
various voluntary initiatives with U.S. industry (e.g., EPA Voluntary Aluminum Industrial Partnership (VAIP)) and
follows guidelines established under the voluntary programs for managing CBI.
Summary Data Compilation and Storage
The inventory coordinator at EPA with support from the data/document manager collects the source and sink
categories' descriptive text and annexes, and also aggregates the emission estimates into a summary data file that
links the individual source and sink category data files together. This summary data file contains all of the essential
data in one central location, in formats commonly used in the Inventory document. In addition to the data from
each source and sink category, national trend and related data are also gathered in the summary sheet for use in
the Executive Summary, Introduction, and Trends sections of the Inventory report. Similarly, uncertainty estimates
for each source and sink category are also aggregated into uncertainty summary data files that are used to conduct
the overall Inventory uncertainty analysis (see Chapter 1.7). Electronic copies of each year's summary data files,
which contain all the emission and sink estimates for the United States, are kept on a central server at EPA under
the jurisdiction of the inventory coordinator.
27	40 CFR part 2, Subpart B titled "Confidentiality of Business Information" which is the regulation establishing rules governing
handling of data entitled to confidentiality treatment. See .
28	Federal Register Notice on "Greenhouse Gas Reporting Program: Publication of Aggregated Greenhouse Gas Data." See pp.
79 and 110 of notice at .
29	U.S. EPA Greenhouse Gas Reporting Program. Developments on Publication of Aggregated Greenhouse Gas Data, November
25, 2014. See .
1-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
National Inventory Report Preparation
The NIR is compiled from the sections developed by each individual source or sink category lead. In addition, the
inventory coordinator prepares a brief overview of each chapter that summarizes the emissions from all sources
discussed in the chapters. The inventory coordinator then carries out a key category analysis for the Inventory,
consistent with the 2006IPCC Guidelines for National Greenhouse Gas Inventories, and in accordance with the
reporting requirements of the UNFCCC. Also at this time, the Introduction, Executive Summary, and Trends in
Greenhouse Gas Emissions chapters are drafted, to reflect the trends for the most recent year of the current
Inventory. The analysis of trends necessitates gathering supplemental data, including weather and temperature
conditions, economic activity and gross domestic product, population, atmospheric conditions, and the annual
consumption of electricity, energy, and fossil fuels. Changes in these data are used to explain the trends observed
in greenhouse gas emissions in the United States. Furthermore, specific factors that affect individual sectors are
researched and discussed. Many of the factors that affect emissions are included in the Inventory document as
separate analyses or side discussions in boxes within the text. Text boxes are also created to examine the data
aggregated in different ways than in the remainder of the document, such as a focus on transportation activities or
emissions from electricity generation. The document is prepared to match the specification of the UNFCCC
reporting guidelines for National Inventory Reports.
Common Reporting Format Table Compilation
The CRF tables are compiled from individual tables completed by each individual source or sink category lead,
which contain emissions and/or removals and activity data. The inventory coordinator integrates the category data
into the UNFCCC's "CRF Reporter" for the United States, assuring consistency across all sectoral tables. The
summary reports for emissions, methods, and emission factors used, the overview tables for completeness of
estimates, the recalculation tables, the notation key completion tables, and the emission trends tables are then
completed by the inventory coordinator. Internal automated quality checks on the CRF Reporter, as well as
reviews by the category leads, are completed for the entire time series of CRF tables before submission.
QA/QC and Uncertainty
QA/QC and uncertainty analyses are guided by the QA/QC and Inventory coordinators, who help maintain the
QA/QC plan and the overall uncertainty analysis procedures (see sections on QA/QC and Uncertainty, below). This
coordinator works closely with the Inventory coordinator and source and sink category leads to ensure that a
consistent QA/QC plan and uncertainty analysis is implemented across all inventory sources. The inventory QA/QC
plan, outlined in Section 1.7 and Annex 8, is consistent with the quality assurance procedures outlined by EPA and
IPCC good practices. The QA/QC and uncertainty findings also inform overall improvement planning, and specific
improvements are noted in the Planned Improvements sections of respective categories. QA processes are
outlined below.
L*pert, Public, and UNFCCC Reviews
The compilation of the inventory includes a two-stage review process, in addition to international technical expert
review following submission of the report. During the first stage (the 30-day Expert Review period), a first draft of
sectoral chapters of the document are sent to a select list of technical experts outside of EPA who are not directly
involved in preparing estimates. The purpose of the Expert Review is to provide an objective review, encourage
feedback on the methodological and data sources used in the current Inventory, especially for sources which have
experienced any changes since the previous Inventory.
Once comments are received and addressed, the second stage, or second draft of the document is released for
public review by publishing a notice in the U.S. Federal Register and posting the entire draft Inventory document
on the EPA website. The Public Review period allows for a 30-day comment period and is open to the entire U.S.
public. Comments may require further discussion with experts and/or additional research, and specific Inventory
Introduction 1-15

-------
improvements requiring further analysis as a result of comments are noted in the relevant category's Planned
Improvement section. EPA publishes responses to comments received during both reviews with the publication of
the final report on its website.
Following completion and submission of the report to the UNFCCC, the report also undergoes review by an
independent international team of experts for adherence to UNFCCC reporting guidelines and IPCC guidance.30
Feedback from all review processes that contribute to improving inventory quality over time are described further
in Annex 8.
Final Submittal to UNFCCC and Document Publication
After the final revisions to incorporate any comments from the Expert Review and Public Review periods, EPA
prepares the final NIR and the accompanying CRF tables for electronic reporting. EPA, as the National Inventory
focal point, sends the official submission of the U.S. Inventory to the UNFCCC using the CRF Reporter software,
coordinating with the U.S. Department of State, the overall UNFCCC focal point. Concurrently, for timely public
access, the report is also published on EPA's website.31
1.4 Methodology and Data Sources
Emissions of greenhouse gases from various source and sink categories have been estimated using methodologies
that are consistent with the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2006). To a great
extent, this report makes use of published official economic and physical statistics for activity data and emission
factors. Depending on the emission source category, activity data can include fuel consumption or deliveries,
vehicle-miles traveled, raw material processed, etc. Emission factors are factors that relate quantities of emissions
to an activity. For more information on data sources see Section 1.2 above, Box 1-1 on use of GHGRP data, and
categories' methodology sections for more information on other data sources. In addition to official statistics, the
report utilizes findings from academic studies, trade association surveys and statistical reports, along with expert
judgment, consistent with the 2006 IPCC Guidelines.
The methodologies provided in the 2006 IPCC Guidelines represent foundational methodologies for a variety of
source categories, and many of these methodologies continue to be improved and refined as new research and
data become available. This report uses the IPCC methodologies when applicable, and supplements them with
other available country-specific methodologies and data where possible. For examples, as noted earlier in this
chapter, this report does use supplements and refinements to 2006 IPCC Guidelines in estimating emissions from
wastewater, Low Voltage Anode Effects (LVAE) during aluminum production, drained organic soils, and
management of wetlands. Choices made regarding the methodologies and data sources used are provided in
conjunction with the discussion of each source category in the main body of the report. Where additional detail is
helpful and necessary to explain methodologies and data sources used to estimate emissions, complete
documentation is provided in the annexes as indicated in the methodology sections of those respective source
categories (e.g., Coastal Wetlands).
30	See .
31	See .
1-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
1.5 Key Categories
The 2006IPCC Guidelines (IPCC 2006) defines a key category as a "[category] that is prioritized within the national
inventory system because its estimate has a significant influence on a country's total inventory of greenhouse
gases in terms of the absolute level, the trend, or the uncertainty in emissions and removals."32 A key category
analysis identifies source or sink categories for focusing efforts to improve overall Inventory quality.
The 2006 IPCC Guidelines (IPCC 2006) defines several approaches, both quantitative and qualitative, to conduct a
key category analysis and identify key categories both in terms of absolute level and trend, along with
consideration of uncertainty. This report employs all approaches to identify key categories for the United States.
The first method, Approach 1, was implemented to identify the key categories without considering uncertainty in
its calculations. A level assessment using Approach 1 identifies all source and sink categories that cumulatively
account for 95 percent of total (i.e., gross) emissions in a given year when assessed in descending order of absolute
magnitude. This analysis was performed twice; one analysis included sources and sinks from the Land Use, Land-
Use Change, and Forestry (LULUCF) sector, the other analysis did not include the LULUCF categories. The second
method, Approach 2, was then implemented to identify any additional key categories not already identified in
Approach 1 level assessment. This analysis differs from Approach 1 by incorporating each category's uncertainty
assessments (or proxies) in its calculations and was also performed twice to include or exclude LULUCF categories.
An Approach 2 level assessment identifies all sources and sink categories that cumulatively account for 90 percent
of the sum of all level assessments when sorted in descending order of magnitude. Similarly, a trend analysis can
identify categories that significantly influence trends since 1990 by identifying all source and sink categories that
cumulatively account for 95 percent of the sum all the trend assessments (e.g., percent change to trend) when
sorted in descending order of absolute magnitude.
In addition to conducting Approach 1 and 2 level and trend assessments as described above, a qualitative
assessment of the source categories was conducted to capture any additional key categories that were not
identified using the previously described quantitative approaches. For this inventory, no additional categories were
identified using qualitative criteria recommend by IPCC, but EPA continues to review its qualitative assessment on
an annual basis. Find more information regarding the overall key category analysis in Annex 1 to this report.
Table 1-4: Key Categories for the United States (1990 and 2019)








2019








Emissions








(MMT



Approach 1

Approach 2 (includes uncertainty)
C02 Eq.)


Level
Trend Level
Trend
Level
Trend
Level Trend

CRF Source/Sink

Without
Without With
With
Without
Without
With With

Categories
Gas
LULUCF
LULUCF LULUCF
LULUCF
LULUCF
LULUCF
LULUCF LULUCF

Energy
l.A.3.b C02








Emissions from
C02






1,510.5
Mobile Combustion:






Road








l.A.l C02 Emissions








from Stationary








Combustion - Coal -
co2
•
• •
•
•
•
• •
973.5
Electricity








Generation








32 See Chapter 4 Volume 1, "Methodological Choice and Identification of Key Categories" in IPCC (2006). See .
Introduction 1-17

-------
CRF Source/Sink
Categories
Gas
Approach 1
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Approach 2 (includes uncertainty)
Level Trend
Without Without
LULUCF LULUCF
Level Trend
With With
LULUCF LULUCF
2019
Emissions
(MMT
C02 Eq.)
l.A.l C02 Emissions
from Stationary
Combustion - Gas -
Electricity
Generation
1.A.2 C02 Emissions
from Stationary
Combustion - Gas -
Industrial
l.A.4.b C02
Emissions from
Stationary
Combustion - Gas -
Residential
1.A.2 C02 Emissions
from Stationary
Combustion - Oil -
Industrial
l.A.4.a C02
Emissions from
Stationary
Combustion - Gas -
Commercial
l.A.3.a C02
Emissions from
Mobile Combustion:
Aviation
1.A.5 C02 Emissions
from Non-Energy
Use of Fuels
l.A.4.b C02
Emissions from
Stationary
Combustion - Oil -
Residential
l.A.4.a C02
Emissions from
Stationary
Combustion - Oil -
Commercial
l.A.3.e C02
Emissions from
Mobile Combustion:
Other
1.A.2 C02 Emissions
from Stationary
Combustion - Coal -
Industrial
l.B.2.a C02
Emissions from
Petroleum Systems
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
616.0
503.3
275.3
269.7
192.8
178.5
128.8
61.5
55.3
53.7
49.5
47.3
1-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
CRF Source/Sink
Categories
Gas
Approach 1
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Approach 2 (includes uncertainty)
Level Trend
Without Without
LULUCF LULUCF
Level Trend
With With
LULUCF LULUCF
2019
Emissions
(MMT
C02 Eq.)
l.B.2.b C02
Emissions from
Natural Gas Systems
1.A.3.C CO;
Emissions from
Mobile Combustion:
Railways
l.A.3.d C02
Emissions from
Mobile Combustion:
Marine
1.A.5 C02 Emissions
from Stationary
Combustion - Oil -
U.S. Territories
l.A.l C02 Emissions
from Stationary
Combustion - Oil -
Electricity
Generation
l.A.5.b C02
Emissions from
Mobile Combustion:
Military
1.A.5 C02 Emissions
from Stationary
Combustion - Gas -
U.S. Territories
l.A.4.a C02
Emissions from
Stationary
Combustion - Coal -
Commercial
l.A.4.b C02
Emissions from
Stationary
Combustion - Coal -
Residential
l.B.2.b CH4
Emissions from
Natural Gas Systems
l.B.l Fugitive
Emissions from Coal
Mining
l.B.2.a CH4
Emissions from
Petroleum Systems
1.B.2 CH4 Emissions
from Abandoned Oil
and Gas Wells
C02
C02
C02
C02
C02
C02
C02
C02
C02
CH4
ch4
ch4
ch4
37.2
37.1
32.1
19.5
16.2
5.3
2.5
1.6
NO
157.6
47.4
39.1
6.6
Introduction 1-19

-------
CRF Source/Sink
Categories
Gas
Approach 1
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Approach 2 (includes uncertainty)
Level Trend
Without Without
LULUCF LULUCF
Level Trend
With With
LULUCF LULUCF
2019
Emissions
(MMT
C02 Eq.)
l.A.4.b CH4
Emissions from
Stationary
Combustion -
Residential
l.A.3.b CH4
Emissions from
Mobile Combustion:
Road
l.A.l N20 Emissions
from Stationary
Combustion - Coal -
Electricity
Generation
l.A.3.b N20
Emissions from
Mobile Combustion:
Road
CH4
ch4
n2o
n2o
4.6
0.9
16.7
Industrial Processes and Product Use
2.C.1 C02 Emissions
from Iron and Steel
Production &
Metallurgical Coke
Production
2.A.1 C02 Emissions
from Cement
Production
2.B.8 C02 Emissions
from Petrochemical
Production
2.B.3 N20 Emissions
from Adipic Acid
Production
2.F.1 Emissions from
Substitutes for
Ozone Depleting
Substances:
Refrigeration and Air
Conditioning
2.F.4 Emissions from
Substitutes for
Ozone Depleting
Substances: Aerosols
2.F.2 Emissions from
Substitutes for
Ozone Depleting
Substances: Foam
Blowing Agents
C02
C02
C02
N20
HFCs,
PFCs
HFCs,
PFCs
HFCs,
PFCs
41.3
40.9
30.8
5.3
133.4
16.3
16.1
1-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
CRF Source/Sink
Categories
Gas
Approach 1
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Approach 2 (includes uncertainty)
Level Trend
Without Without
LULUCF LULUCF
Level Trend
With With
LULUCF LULUCF
2.F.3 Emissions from
Substitutes for
Ozone Depleting
Substances: Fire
Protection
2.F.5 Emissions from
Substitutes for
Ozone Depleting
Substances: Solvents
2.G SF6 Emissions
from Electrical
Transmission and
Distribution
2.B.9 HFC-23
Emissions from
HCFC-22 Production
2.C.3 PFC Emissions
from Aluminum
Production
HFCs,
PFCs
HFCs,
PFCs
SF6
HFCs
PFCs
Agriculture
3.G C02 Emissions
from Liming
C02





•


2.4
3.A.1 CH4 Emissions










from Enteric
ch4
•
•
•
•
•

•

172.3
Fermentation: Cattle










3.B.1 CH4 Emissions










from Manure
ch4
•
•
•
•

•

•
35.4
Management: Cattle










3.B.4 CH4 Emissions










from Manure
Management: Other
ch4
•

•





26.9
Livestock










3.C CH4 Emissions
from Rice Cultivation
ch4




•

•

15.1
3.D.1 Direct N20










Emissions from
Agricultural Soil
n2o
•
•
•
•
•
•
•
•
290.4
Management










3.D.2 Indirect N20










Emissions from
n2o
•
•
•
•
•
•
•
•
54.2
Applied Nitrogen










Waste
5.A CH4 Emissions
from Landfills
ch4
•
•
•
•
•
•
•
•
114.5
5.D CH4 Emissions










from Wastewater
ch4
•

•

•



18.4
Treatment










Introduction 1-21

-------










2019










Emissions










(MMT



Approach 1

Approach 2 (includes uncertainty)
C02 Eq.)


Level
Trend
Level
Trend
Level
Trend
Level
Trend

CRF Source/Sink

Without
Without
With
With
Without
Without
With
With

Categories
Gas
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF

5.D N20 Emissions










from Wastewater
N20
•

•

•
•
•
•
26.4
Treatment










Land Use, Land-Use Change, and Forestry
4.E.2 Net C02







Emissions from Land
Converted to
C02

•
•
•
•
79.2
Settlements







4.B.2 Net C02







Emissions from Land
Converted to
co2

•

•

54.
Cropland







4.C.1 Net C02







Emissions from
Grassland Remaining
co2



•
•
14.5
Grassland







4.B.1 Net C02







Emissions from
Cropland Remaining
co2

•
•
•
•
(14.
Cropland







4.C.2 Net C02







Emissions from Land
Converted to
co2

•
•
•
•
(23.2)
Grassland







4.A.2 Net C02







Emissions from Land
Converted to Forest
co2

•

•

(99.
Land







4.E.1 Net C02







Emissions from







Settlements
co2

•
•
•
•
(124.1)
Remaining







Settlements







4.A.1 Net C02







Emissions from







Forest Land
co2

•
•
•
•
(691.3)
Remaining Forest







Land







4.A.1 CH4 Emissions
from Forest Fires
ch4
•

9.5
Subtotal Without LULUCF
6,398.6
Total Emissions Without LULUCF




6,558.3
Percent of Total Without LULUCF




98%
Subtotal With LULUCF
5,598.0
Total Emissions With LULUCF
5,769.1
Percent of Total With LULUCF
97%
NO (Not Occurring)
1-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
1.6 Quality Assurance and Quality Control
(QA/QC)
As part of efforts to achieve its stated goals for inventory quality, transparency, and credibility, the United States
has developed a quality assurance and quality control plan designed to check, document, and improve the quality
of its inventory over time. QA/QC activities on the Inventory are undertaken within the framework of the U.S.
Quality Assurance/Quality Control and Uncertainty Management Plan (QA/QC plan) for the U.S. Greenhouse Gas
Inventory: Procedures Manual for QA/QC and Uncertainty Analysis.
Key attributes of the QA/QC plan are summarized in Figure 1-2. These attributes include:
•	Procedures and Forms: detailed and specific systems that serve to standardize the process of
documenting and archiving information, as well as to guide the implementation of QA/QC and the
analysis of uncertainty
•	Implementation of Procedures: application of QA/QC procedures throughout the whole inventory
development process from initial data collection, through preparation of the emission estimates, to
publication of the Inventory
•	Quality Assurance (QA): expert and public reviews for both the inventory estimates and the Inventory
report (which is the primary vehicle for disseminating the results of the inventory development process).
The expert technical review conducted by the UNFCCC supplements these QA processes, consistent with
the QA good practice and the 2006IPCC Guidelines (IPCC 2006)
•	Quality Control (QC): application of General (Tier 1) and Category-specific (Tier 2) quality controls and
checks, as recommended by 2006 IPCC Guidelines (IPCC 2006), along with consideration of secondary data
and category-specific checks (additional Tier 2 QC) in parallel and coordination with the uncertainty
assessment; the development of protocols and templates, which provides for more structured
communication and integration with the suppliers of secondary information
•	General (Tier 1) and Category-specific (Tier 2) Checks: quality controls and checks, as recommended by
IPCC Good Practice Guidance and 2006 IPCC Guidelines (IPCC 2006)
•	Record Keeping: provisions to track which procedures have been followed, the results of the QA/QC,
uncertainty analysis, and feedback mechanisms for corrective action based on the results of the
investigations which provide for continual data quality improvement and guided research efforts.
•	Multi-Year Implementation: a schedule for coordinating the application of QA/QC procedures across
multiple years, especially for category-specific QC, prioritizing key categories
•	Interaction and Coordination: promoting communication within the EPA, across Federal agencies and
departments, state government programs, and research institutions and consulting firms involved in
supplying data or preparing estimates for the Inventory. The QA/QC Management Plan itself is intended
to be revised and reflect new information that becomes available as the program develops, methods are
improved, or additional supporting documents become necessary.
Introduction 1-23

-------
Figure 1-2: U.S. QA/QC Plan Summary
Data
Gathering

Data
^Documentation

Calculating
Emissions
Obtain data in

Contact reports

P* Clearly label
electronic

for non-electronic

parameters, units,
format (if

communications

and conversion
possible)

• Provide cell

factors
Review

references for

• Review spreadsheet
spreadsheet

primary data

integrity
construction

elements

o Equations
o Avoid

• Obtain copies of

o Units
hardwiring

all data sources

o Inputs and
o Use data

• List and location

outputs
validation

of any

• Develop automated
o Protect cells

working/external

checkers for:
Develop

spreadsheets

o Input ranges
automatic

• Document

o Calculations
checkers for;

assumptions

o Emission
o Outliers,

• Complete QA/QC

aggregation
negative

checklists

o Trend and IEF
values, or

• CRF and summary

checks
missing data

tab links


o Variable




types match




values




o Time series




consistency




Maintain




tracking tab for




status of




gathering




efforts




Check input

• Check citations in

• Reproduce
data for

spreadsheet and

calculations
transcription

text for accuracy

• Review time
errors

and style

series consistency
Inspect

• Check reference

• Review changes
automatic

docket for new

in
checkers

citations

data/consistency
Identify

• Review

with IPCC
spreadsheet

documentation

methodology
modifications

for any data /


that could

methodology


provide

changes


additional

• Complete QA/QC


QA/QC checks

checklists




• CRF and summary




tab links


Cross-Cutting
Coordination
•	Common starting
versions for each
inventory year
•	Utilize
unalterable
summary and
CRF tab for each
source
spreadsheet for
linking to a
master summary
spreadsheet
•	Follow strict
version control
procedures
•	Document
QA/QC
procedures
1-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Box 1-3: Use of IPCC Reference Approach to support Verification of Emissions from Fossil Fuel Combustion
The UNFCCC reporting guidelines require countries to complete a "top-down" reference approach for
estimating C02 emissions from fossil fuel combustion in addition to their "bottom-up" sectoral methodology for
purposes of verification. This estimation method uses alternative methodologies and different data sources
than those contained in that section of the Energy chapter. The reference approach estimates fossil fuel
consumption by adjusting national aggregate fuel production data for imports, exports, and stock changes
rather than relying on end-user consumption surveys (see Annex 4 of this report). The reference approach
assumes that once carbon-based fuels are brought into a national economy, they are either saved in some way
(e.g., stored in products, kept in fuel stocks, or left unoxidized in ash) or combusted, and therefore the carbon in
them is oxidized and released into the atmosphere. Accounting for actual consumption of fuels at the sectoral
or sub-national level is not required.
In addition, based on the national QA/QC plan for the Inventory, some sector, subsector and category-specific
QA/QC and verification checks have been developed. These checks follow the procedures outlined in the national
QA/QC plan, tailoring the procedures to the specific documentation and data files associated with individual
sources. For each greenhouse gas emissions source or sink included in this Inventory, a minimum of general or Tier
1 QC analysis has been undertaken. Where QC activities for a particular category go beyond the minimum Tier 1
level, and include category-specific checks (Tier 2) or include verification, further explanation is provided within the
respective source or sink category text. Similarly, responses or updates based on comments from the expert, public
and the international technical expert reviews (e.g., UNFCCC) are also addressed within the respective source or
sink category sections in each sectoral chapter and Annex 8.
The quality control activities described in the U.S. QA/QC plan occur throughout the inventory process; QA/QC is
not separate from, but is an integral part of, preparing the Inventory. Quality control—in the form of both good
practices (such as documentation procedures) and checks on whether good practices and procedures are being
followed—is applied at every stage of inventory development and document preparation. In addition, quality
assurance occurs during the Expert Review and the Public Review, in addition to the UNFCCC expert technical
review. While all phases significantly contribute to improving inventory quality, the public review phase is also
essential for promoting the openness of the inventory development process and the transparency of the inventory
data and methods.
The QA/QC plan guides the process of ensuring inventory quality by describing data and methodology checks,
developing processes governing peer review and public comments, and developing guidance on conducting an
analysis of the uncertainty surrounding the emission and removal estimates. The QA/QC procedures also include
feedback loops and provide for corrective actions that are designed to improve the inventory estimates over time.
1.7 Uncertainty Analysis of Emission Estimates
Emissions calculated for the U.S. Inventory reflect best estimates for greenhouse gas source and sink categories in
the United States and are continuously revised and improved as new information becomes available. Uncertainty
estimates are an essential element of a complete and transparent emissions inventory that help inform and
prioritize Inventory improvements. For the U.S. Inventory, uncertainty analyses are conducted for each source and
sink category as well as for the uncertainties associated with the overall emission (current and base year) and
trends estimates. These analyses reflect the quantitative uncertainty in the emission (and removal) estimates
associated with uncertainties in their input parameters (e.g., activity data and EFs) and serve to evaluate the
relative contribution of individual input parameter uncertainties to the overall Inventory, its trends, and each
source and sink category.
Introduction 1-25

-------
The overall uncertainty estimate for total U.S. greenhouse gas emissions was developed using the IPCC Approach 2
uncertainty estimation methodology, which employs a Monte Carlo Stochastic Simulation technique. The IPCC
provides good practice guidance on two approaches—Approach 1 and Approach 2—to estimating uncertainty for
both individual and combined source categories. Approach 2 quantifies uncertainties based on a distribution of
emissions (or removals), built-up from repeated calculations of emission estimation models and the underlying
input parameters, randomly selected according to their known distributions. Approach 2 methodology is applied to
each individual source and sink category wherever data and resources are permitted and is also used to quantify
the uncertainty in the overall Inventory and its Trends. Source and sink chapters in this report provide additional
details on the uncertainty analysis conducted for each source and sink category. See Annex 7 of this report for
further details on the U.S. process for estimating uncertainty associated with the overall emission (base and
current year) and trends estimates. Consistent with IPCC (IPCC 2006), the United States has ongoing efforts to
continue to improve the overall Inventory uncertainty estimates presented in this report.
The United States has also implemented many improvements over the last several years to reduce uncertainties
across the source and sink categories and improve Inventory estimates. These improvements largely result from
new data sources that provide more accurate data and/or increased data coverage, as well as methodological
improvements. Following IPCC good practice, additional efforts to reduce Inventory uncertainties can occur
through efforts to incorporate excluded emission and sink categories (see Annex 5), improve emission estimation
methods, and collect more detailed, measured, and representative data. Individual source chapters and Annex 7
both describe current ongoing and planned Inventory and uncertainty analysis improvements. Consistent with
IPCC (2006), the United States has ongoing efforts to continue to improve the category-specific uncertainty
estimates presented in this report, largely prioritized by considering improvements categories identified as
significant by the Key Category Analysis.
Estimates of quantitative uncertainty for the total U.S. greenhouse gas emissions in 1990 (base year) and 2019 are
shown below in Table 1-5 and Table 1-6, respectively. The overall uncertainty surrounding the Total Net Emissions
is estimated to be -6 to +6 percent in 1990 and -5 to +5 percent in 2019. When the LULUCF sector is excluded from
the analysis the uncertainty is estimated to be -2 to +5 percent in 1990 and -2 to +4 percent in 2019.
Table 1-5: Estimated Overall Inventory Quantitative Uncertainty for 1990 (MMT CO2 Eq. and
Percent)
1990 Emission
Estimate Uncertainty Range Relative to Emission Estimate3
Gas	(MMTCOz
Eq.)	(MMT C02 Eq.)	(%)


Lower
Upper
Lower
Upper




Boundc
Boundc
Bound
Bound


co2
5,113.5
5,008.6
5,349.7
-2%
5%
5,177.3
88.5
CH4d
776.9
710.7
863.0
-9%
11%
785.8
38.7
N2Od
452.7
368.3
581.4
-19%
28%
461.0
54.8
PFC, HFC, SFe, and NF3d
99.7
90.2
112.2
-9%
13%
100.3
5.6
Total
6,442.7
6,311.7
6,748.8
-2%
5%
6,524.4
111.0
LULUCF Emissions0
7.9
6.0
10.0
-24%
26%
8.0
1.0
LULUCF Carbon Stock Change Fluxf
(908.7)
(1,221.6)
(741.6)
34%
-18%
(982.4)
122.3
LULUCF Sector Net Total8
(900.8)
(1,213.8)
(733.6)
35%
-19%
(974.4)
122.3
Net Emissions (Sources and Sinks)
5,541.9
5,232.4
5,877.3
-6%
6%
5,550.0
164.7
Notes: Total emissions (excluding emissions for which uncertainty was not quantified) are presented without LULUCF. Net
emissions are presented with LULUCF. Totals may not sum due to independent rounding. Parentheses indicate net
sequestration.
+ Does not exceed 0.5 percent.
a The lower and upper bounds for emission estimates correspond to a 95 percent confidence interval, with the lower bound
corresponding to 2.5th percentile and the upper bound corresponding to 97.5th percentile.
Standard
Meanb Deviation'5
(MMTCOz Eq.)
1-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
b Mean value indicates the arithmetic average of the simulated emission estimates; standard deviation indicates the extent of
deviation of the simulated values from the mean.
c The lower and upper bound emission estimates for the sub-source categories do not sum to total emissions because the low
and high estimates for total emissions were calculated separately through simulations.
d The overall uncertainty estimates did not take into account the uncertainty in the GWP values for CH4, N20 and high GWP
gases used in the Inventory emission calculations for 1990.
6 LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
f LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining
Settlements, and Land Converted to Settlements. Since the resulting flux is negative the signs of the resulting lower and upper
bounds are reversed.
g The LULUCF Sector Net Total is the net sum of all CH4 and N20 emissions to the atmosphere plus net carbon stock changes.
Table 1-6: Estimated Overall Inventory Quantitative Uncertainty for 2019 (MMT CO2 Eq. and
Percent)

2019



Emission
Uncertainty Range Relative to Emission
Standard

Estimate
Estimate3
Meanb Deviation'5
Gas
(mmtco2



Eq.)
(MMT C02 Eq.) (%)
(MMT C02 Eq.)


Lower
Upper
Lower
Upper




Boundc
Boundc
Bound
Bound


C02
5,255.8
5,129.9
5,461.8
-2%
4%
5,295.1
85.4
CH4d
659.7
608.3
732.5
-8%
11%
670.1
31.6
N2Od
457.1
367.5
598.2
-20%
31%
468.7
59.1
PFC, HFC, SF6, and NF3d
185.6
179.4
208.1
-3%
12%
193.1
7.4
Total
6,558.3
6,417.7
6,845.6
-2%
4%
6,627.0
108.9
LULUCF Emissions0
23.5
20.1
27.9
-14%
19%
23.9
2.0
LULUCF Carbon Stock Change Fluxf
(812.7)
(1,089.1)
(664.0)
34%
-18%
(878.2)
108.4
LULUCF Sector Net Total6
(789.2)
(1,064.9)
(640.0)
35%
-19%
(854.3)
108.4
Net Emissions (Sources and Sinks)
5,769.0
5,471.4
6,074.3
-5%
5%
5,772.8
153.7
Notes: Total emissions (excluding emissions for which uncertainty was not quantified) are presented without LULUCF. Net
emissions are presented with LULUCF. Totals may not sum due to independent rounding. Parentheses indicate net
sequestration.
+ Does not exceed 0.5 percent.
a The lower and upper bounds for emission estimates correspond to a 95 percent confidence interval, with the lower bound
corresponding to 2.5th percentile and the upper bound corresponding to 97.5th percentile.
b Mean value indicates the arithmetic average of the simulated emission estimates; standard deviation indicates the extent of
deviation of the simulated values from the mean.
c The lower and upper bound emission estimates for the sub-source categories do not sum to total emissions because the low
and high estimates for total emissions were calculated separately through simulations.
d The overall uncertainty estimates did not take into account the uncertainty in the GWP values for CH4, N20 and high GWP
gases used in the Inventory emission calculations for 2019.
6 LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
f LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land, Land
Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland, Land
Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining Settlements, and
Land Converted to Settlements. Since the resulting flux is negative the signs of the resulting lower and upper bounds are
reversed.
g The LULUCF Sector Net Total is the net sum of all CH4 and N20 emissions to the atmosphere plus net carbon stock changes.
Introduction 1-27

-------
In addition to the estimates of uncertainty associated with the current and base year emission estimates, Table 1-7
presents the estimates of inventory trend uncertainty. The 2006IPCC Guidelines defines trend as the difference in
emissions between the base year (i.e., 1990) and the current year (i.e., 2019) Inventory estimates. However, for
purposes of understanding the concept of trend uncertainty, the emission trend is defined in this Inventory as the
percentage change in the emissions (or removal) estimated for the current year, relative to the emission (or
removal) estimated for the base year. The uncertainty associated with this emission trend is referred to as trend
uncertainty and is reported as between -3 and 13 percent between Inventory estimates between 1990 and 2019.
See Annex 7 for trend uncertainty estimates for individual source and sink categories by gas.
Table 1-7: Quantitative Assessment of Trend Uncertainty (MMT CO2 Eq. and Percent)

Base Year
2019
Emissions


Gas/Source
Emissions3 Emissions
Trend
Trend Rangeb

(MMT C02
Eq.)
(%)
(%)





Lower
Upper




Bound
Bound
C02
5,113.5
5,255.8
3%
-2%
7%
ch4
776.9
659.7
-15%
-26%
-3%
n2o
452.7
457.1
1%
-27%
43%
HFCs, PFCs, SF6, and NF3
99.7
185.6
41%
68%
119%
Total Emissionsc
6,442,7
6,558.3
1%
-3%
7%
LULUCF Emissions'1
7.9
23.5
196%
124%
306%
LULUCF Carbon Stock Change Flux0
(908.7)
-812.7
-11%
-37%
27%
LULUCF Sector Net Total'
(900.8)
(789.2)
-12%
-38%
25%
Net Emissions (Sources and Sinks)c
5,541.9
5,769.0
4%
-3%
13%
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration. Total emissions (excluding emissions
for which uncertainty was not quantified) are presented without LULUCF. Net emissions are presented with LULUCF.
+ Does not exceed 0.05 MMT C02 Eq. or 0.5 percent.
a Base Year is 1990 for all sources.
bThe trend range represents a 95 percent confidence interval for the emission trend, with the lower bound corresponding to 2.5th
percentile value and the upper bound corresponding to 97.5th percentile value.
c Emissions from Wood Biomass and Biofuel Consumption are not included specifically in the energy sector totals.
c Totals exclude emissions for which uncertainty was not quantified.
dLULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils,
Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal Wetlands; and N20
emissions from Forest Soils and Settlement Soils.
0 LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining Settlements,
and Land Converted to Settlements.
1.8 Completeness
This report, along with its accompanying CRF tables, serves as a thorough assessment of the anthropogenic
sources and sinks of greenhouse gas emissions for the United States for the time series 1990 through 2019. This
report is intended to be comprehensive and includes the vast majority of emissions and removals identified as
anthropogenic, consistent with IPCC and UNFCCC guidelines. In general, sources or sink categories not accounted
for in this Inventory are excluded because they are not occurring in the United States, or because data are
1-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
unavailable to develop an estimate and/or the categories were determined to be insignificant33 in terms of overall
national emissions per UNFCCC reporting guidelines.
The United States is continually working to improve upon the understanding of such sources and sinks currently
not included and seeking to find the data required to estimate related emissions and removals, focusing on
categories that are anticipated to be significant. As such improvements are implemented, new emission and
removal estimates are quantified and included in the Inventory, improving completeness of national estimates. For
a list of sources and sink categories not included and more information on significance of these categories, see
Annex 5 and the respective category sections in each sectoral chapter of this report.
1.9 Organization of Report
In accordance with the revision of the UNFCCC reporting guidelines agreed to at the nineteenth Conference of the
Parties (UNFCCC 2014), this Inventory of U.S. Greenhouse Gas Emissions and Sinks is grouped into five sector-
specific chapters consistent with the UN Common Reporting Framework, listed below in Table 1-8. In addition,
chapters on Trends in Greenhouse Gas Emissions, Other information, and Recalculations and Improvements to be
considered as part of the U.S. Inventory submission are included.
Table 1-8: IPCC Sector Descriptions
Chapter/IPCC Sector
Activities Included
Energy
Emissions of all greenhouse gases resulting from stationary and mobile energy

activities including fuel combustion and fugitive fuel emissions, and non-energy

use of fossil fuels.
Industrial Processes and
Emissions resulting from industrial processes and product use of greenhouse
Product Use
gases.
Agriculture
Emissions from agricultural activities except fuel combustion, which is

addressed under Energy.
Land Use, Land-Use
Emissions and removals of C02, and emissions of CH4, and N20 from land use,
Change, and Forestry
land-use change and forestry.
Waste
Emissions from waste management activities.
Within each chapter, emissions are identified by the anthropogenic activity that is the source or sink of the
greenhouse gas emissions being estimated (e.g., coal mining). Overall, the following organizational structure is
consistently applied throughout this report:
Chapter/IPCC Sector: Overview of emissions and trends for each IPCC defined sector.
CRF Source or Category: Description of category pathway and emission/removal trends based on IPCC
methodologies, consistent with UNFCCC reporting guidelines.
Methodology: Description of analytical methods (e.g., from 2006 IPCC Guidelines, or country-specific methods)
employed to produce emission estimates and identification of data references, primarily for activity data and
emission factors.
Uncertainty and Time Series Consistency: A discussion and quantification of the uncertainty in emission estimates
and a discussion of time-series consistency.
33 See paragraph 32 of Decision 24/CP.19, the UNFCCC reporting guidelines on annual inventories for Parties included in Annex
1 to the Convention. Paragraph notes that "...An emission should only be considered insignificant if the likely level of emissions
is below 0.05 per cent of the national total GHG emissions, and does not exceed 500 kt C02 Eq. The total national aggregate of
estimated emissions for all gases and categories considered insignificant shall remain below 0.1 percent of the national total
GHG emissions."
Introduction 1-29

-------
QA/QC and Verification: A discussion on steps taken to QA/QC and verify the emission estimates, consistent with
the U.S. QA/QC plan, and any key QC findings.
Recalculations Discussion: A discussion of any data or methodological changes that necessitate a recalculation of
previous years' emission estimates, and the impact of the recalculation on the emission estimates, if applicable.
Planned Improvements: A discussion on any category-specific planned improvements, if applicable.
Special attention is given to C02 from fossil fuel combustion relative to other sources because of its share of
emissions and its dominant influence on emission trends. For example, each energy consuming end-use sector
(i.e., residential, commercial, industrial, and transportation), as well as the electricity generation sector, is
described individually. Additional information for certain source categories and other topics is also provided in
several Annexes listed in Table 1-9.
Table 1-9: List of Annexes	
ANNEX 1 Key Category Analysis
ANNEX 2 Methodology and Data for Estimating C02 Emissions from Fossil Fuel Combustion
2.1.	Methodology for Estimating Emissions of C02 from Fossil Fuel Combustion
2.2.	Methodology for Estimating the Carbon Content of Fossil Fuels
2.3.	Methodology for Estimating Carbon Emitted from Non-Energy Uses of Fossil Fuels
ANNEX 3 Methodological Descriptions for Additional Source or Sink Categories
3.1.	Methodology for Estimating Emissions of CH4, N20, and Indirect Greenhouse Gases from Stationary
Combustion
3.2.	Methodology for Estimating Emissions of CH4, N20, and Indirect Greenhouse Gases from Mobile
Combustion and Methodology for and Supplemental Information on Transportation-Related Greenhouse Gas
Emissions
3.3.	Methodology for Estimating Emissions from Commercial Aircraft Jet Fuel Consumption
3.4.	Methodology for Estimating CH4 Emissions from Coal Mining
3.5.	Methodology for Estimating CH4 and C02 Emissions from Petroleum Systems
3.6.	Methodology for Estimating CH4 Emissions from Natural Gas Systems
3.7.	Methodology for Estimating C02 and N20 Emissions from Incineration of Waste
3.8.	Methodology for Estimating Emissions from International Bunker Fuels used by the U.S. Military
3.9.	Methodology for Estimating HFC and PFC Emissions from Substitution of Ozone Depleting Substances
3.10.	Methodology for Estimating CH4 Emissions from Enteric Fermentation
3.11.	Methodology for Estimating CH4 and N20 Emissions from Manure Management
3.12.	Methodology for Estimating N20 Emissions, CH4 Emissions and Soil Organic C Stock Changes from
Agricultural Lands (Cropland and Grassland)
3.13.	Methodology for Estimating Net Carbon Stock Changes in Forest Land Remaining Forest Land and Land
Converted to Forest Land
3.14.	Methodology for Estimating CH4 Emissions from Landfills
ANNEX 4 IPCC Reference Approach for Estimating C02 Emissions from Fossil Fuel Combustion
ANNEX 5 Assessment of the Sources and Sinks of Greenhouse Gas Emissions Not Included
ANNEX 6 Additional Information
6.1.	Global Warming Potential Values
6.2.	Ozone Depleting Substance Emissions
6.3.	Sulfur Dioxide Emissions
6.4.	Complete List of Source Categories
6.5.	Constants, Units, and Conversions
6.6.	Abbreviations
6.7.	Chemical Formulas
ANNEX 7 Uncertainty
7.1.	Overview
7.2.	Methodology and Results
7.3.	Reducing Uncertainty
7.4.	Planned Improvements
7.5.	Additional Information on Uncertainty Analyses by Source
ANNEX 8 QA/QC Procedures
8.1.	Background
8.2.	Purpose
1-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
8.3.	Assessment Factors
8.4.	Responses During the Review Process
ANNEX 9 Use of Greenhouse Gas Reporting Program (GHGRP) in Inventory
Introduction 1-31

-------
2. Trends in Greenhouse Gas Emissions
2.1 Trends in U.S. Greenhouse Gas Emissions
and Sinks
In 2019, total gross U.S. greenhouse gas emissions were 6,558.3 million metric tons carbon dioxide equivalent
(MMT C02 Eq).1 Total U.S. emissions have increased by 1.8 percent from 1990 to 2019, down from a high of 15.6
percent above 1990 levels in 2007. Emissions decreased from 2018 to 2019 by 1.7 percent (113.1 MMT C02 Eq.).
Net emissions (i.e., including sinks) were 5,769.1 MMT C02 Eq. Overall, net emissions decreased 1.7 percent from
2018 to 2019 and decreased 13.0 percent from 2005 levels, as shown in Table 2-1. The decline reflects the
combined impacts of many long-term trends, including population, economic growth, energy market trends,
technological changes including energy efficiency, and carbon intensity of energy fuel choices. Between 2018 and
2019, the decrease in total greenhouse gas emissions was driven largely by a decrease in C02 emissions from fossil
fuel combustion. The decrease in C02 emissions from fossil fuel combustion was a result of a 1 percent decrease in
total energy use and reflects a continued shift from coal to less carbon intensive natural gas and renewables in the
electric power sector.
Figure 2-1 through Figure 2-3 illustrate the overall trend in total U.S. emissions by gas, annual changes, and relative
changes since 1990.
1 The gross emissions total presented in this report for the United States excludes emissions and sinks from removals from Land
Use, Land-Use Change, and Forestry (LULUCF). The net emissions total presented in this report for the United States includes
emissions and sinks from removals from LULUCF.
Trends 2-1

-------
Figure 2-1: U.S. Greenhouse Gas Emissions by Gas
9 000 HFCs, PFCs, SF6 and NF3
I Nitrous Oxide
¦	Methane
8,000 _ carbon Dioxide
¦	Net CO2 Flux from LULUCF3
7,000
I Net Emissions (including sinks)
cn 0 C7» 0
a>	C7> CTi
ci cn	o-* o** o-*
fNrM(Nf\rMfNtNfSJ(Nf\irMrNt\(Nf\(Nf\(NfN|(N
a The term "flux" is used to describe the exchange of C02 to and from the atmosphere, with net flux being either positive or
negative depending on the overall balance. Removal and long-term storage of C02 from the atmosphere is also referred to as
"carbon sequestration."
Figure 2-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the
Previous Year
4%
3.0%
3.2%
2.7%
2.9%
1.8%
2.6%
1.3%
1.6%
1.8%
0.8%
2%
1.3%
0.8%
1.4%
0.9%
0.7%
0.6%
0.6%
0.1%
0%
-0.6%
-1.1%
¦1.1%
-2%
-1.7%
-2.3%-2.3%
-2.3%
-3.5%
-4%
-6%
-6.3%
¦8%
1—1
i-H fM
cri cn
o
T—i
rsi
i—i
oo
1—1
1—1
o
1—1
1—1
2-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 2-3: Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990
(1990=0)
1,200
^HfMnsj-in^DPsooa>Oi-irMm^LnkDrNOOCT
CnOtOtOtCTiOlOliOlCTiOOOOOOOOOOiHiHiHi-li-li-li-li-li-li-l
CTicricncTicncricncTiaioooooooooooooooooooo
Overall, from 1990 to 2019, total emissions of C02 increased by 142.4 MMT C02 Eq. (2.8 percent), while total
emissions of methane (CH4) decreased by 117.2 MMT C02 Eq. (15.1 percent), and total emissions of nitrous oxide
(N20) remained constant despite fluctuations throughout the time series. During the same period, aggregate
weighted emissions of hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SF6), and nitrogen
trifluoride (NF3) rose by 86.0 MMT C02 Eq. (86.3 percent). Despite being emitted in smaller quantities relative to
the other principal greenhouse gases, emissions of HFCs, PFCs, SF6, and NF3 are significant because many of them
have extremely high global warming potentials (GWPs), and, in the cases of PFCs, SF6, and NF3, long atmospheric
lifetimes. Conversely, U.S. greenhouse gas emissions were partly offset by carbon (C) sequestration in managed
forests, trees in urban areas, agricultural soils, landfilled yard trimmings, and coastal wetlands. These were
estimated to offset 12.4 percent (812.7 MMT C02 Eq.) of total emissions in 2019.
Table 2-1 provides information on trends in emissions and sinks from all U.S. anthropogenic sources in weighted
units of MMT C02 Eq., while unweighted gas emissions and sinks in kilotons (kt) are provided in Table 2-2.
Table 2-1: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 Eq.)
Gas/Source
1990
2005

2015
2016
2017
2018
2019
C02
5,113.5
6,134.5

5,371.8
5,248.0
5,207.8
5,375.5
5,255.8
Fossil Fuel Combustion
4,731.5
5,753.5

5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Transportation
1,469.1
1,858.6

1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
Electric Power
1,820.0
2,400.1

1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Industrial
853.8
852.9

797.3
792.5
790.1
813.6
822.5
Residential
338.6
358.9

317.3
292.8
293.4
338.1
336.8
Commercial
228.3
227.1

244.6
231.6
232.0
245.7
249.7
U.S. Territories
21.7
55.9

29.2
26.0
24.6
24.6
24.6
Non-Energy Use of Fuels
112.8
129.1

108.5
99.8
113.5
129.7
128.8
Petroleum Systems
9.7
12.1

32.4
21.8
25.0
37.1
47.3
Iron and Steel Production &








Metallurgical Coke Production
104.7
70.1

47.9
43.6
40.6
42.6
41.3
Cement Production
33.5
46.2

39.9
39.4
40.3
39.0
40.9
Natural Gas Systems
32.0
25.2

29.1
30.1
31.2
33.9
37.2
Petrochemical Production
21.6
27.4

28.1
28.3
28.9
29.3
30.8
Ammonia Production
13.0
9.2

10.6
10.2
11.1
12.2
12.3
Lime Production
11.7
14.6

13.3
12.6
12.9
13.1
12.1
Incineration of Waste
8.1
12.7

11.5
11.5
11.5
11.5
11.5
Other Process Uses of Carbonates
6.3
7.6

12.2
11.0
9.9
7.5
7.5
Urea Consumption for Non-








Agricultural Purposes
3.8
3.7

4.6
5.1
5.0
5.9
6.2
Trends 2-3

-------
Urea Fertilization
2.4
3.5
4.7
4.9
5.1
5.2
5.3
Carbon Dioxide Consumption
1.5
1.4
4.9
4.6
4.6
4.1
4.9
Liming
4.7
4.3
3.7
3.1
3.1
2.2
2.4
Aluminum Production
6.8
4.1
2.8
1.3
1.2
1.5
1.9
Soda Ash Production
1.4
1.7
1.7
1.7
1.8
1.7
1.8
Ferroalloy Production
2.2
1.4
2.0
1.8
2.0
2.1
1.6
Titanium Dioxide Production
1.2
1.8
1.6
1.7
1.7
1.5
1.5
Glass Production
1.5
1.9
1.3
1.2
1.3
1.3
1.3
Zinc Production
0.6
1.0
0.9
0.8
0.9
1.0
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
0.9
0.9
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Carbide Production and







Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Magnesium Production and







Processing
+
+
+
+
+
+
+
Wood Biomass, Ethanol, and







Biodiesel Consumption0
219.4
230.7
317.7
316.6
312.3
319.6
316.2
International Bunker Fuelsb
103.5
113.2
110.9
116.6
120.1
122.1
116.1
CH4c
776.9
686.1
651.5
642.4
648.4
655.9
659.7
Enteric Fermentation
164.7
169.3
166.9
172.2
175.8
178.0
178.6
Natural Gas Systems
186.9
164.2
149.8
147.3
148.7
152.5
157.6
Landfills
176.6
131.4
111.4
108.0
109.4
112.1
114.5
Manure Management
37.1
51.6
57.9
59.6
59.9
61.7
62.4
Coal Mining
96.5
64.1
61.2
53.8
54.8
sin
47.4
Petroleum Systems
48.9
39.5
41.5
39.2
39.3
37.3
39.1
Wastewater Treatment
20.2
20.1
18.8
18.7
18.5
18.4
18.4
Rice Cultivation
16.0
18.0
16.2
15.8
14.9
15.6
15.1
Stationary Combustion
8.6
7.8
8.5
7.9
7.6
8.5
8.7
Abandoned Oil and Gas Wells
6.8
7.2
7.4
7.4
7.2
7.3
6.6
Abandoned Underground Coal







Mines
7.2
6.6
6.4
6.7
6.4
6.2
5.9
Mobile Combustion
6.4
4.0
2.6
2.5
2.5
2.4
2.4
Composting
0.4
1.9
2.1
2.3
2.4
2.3
2.3
Field Burning of Agricultural







Residues
0.4
0.4
0.4
0.4
0.4
0.4
0.4
Petrochemical Production
0.2
0.1
0.2
0.2
0.3
0.3
0.3
Anaerobic Digestion at Biogas







Facilities
+
0.1
0.2
0.2
0.2
0.2
0.2
Ferroalloy Production
+
+
+
+
+
+
+
Carbide Production and







Consumption
+
+
+
+
+
+
+
Iron and Steel Production &







Metallurgical Coke Production
+
+
+
+
+
+
+
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
N2Oc
452.7
455.8
468.2
450.8
446.3
459.2
457.1
Agricultural Soil Management
315.9
313.4
348.5
330.1
327.6
338.2
344.6
Wastewater Treatment
18.7
23.0
25.4
25.9
26.4
26.1
26.4
Stationary Combustion
25.1
34.4
30.5
30.0
28.4
28.2
24.9
Manure Management
14.0
16.4
17.5
18.1
18.7
19.4
19.6
Mobile Combustion
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Nitric Acid Production
12.1
11.3
11.6
10.1
9.3
9.6
10.0
AdipicAcid Production
15.2
7.1
4.3
7.0
7.4
10.3
5.3
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Composting
0.3
1.7
1.9
2.0
2.2
2.0
2.0
2-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Electronics Industry
+
0.1
0.2
0.2
0.3
0.3
0.2
Field Burning of Agricultural







Residues
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
1.0
1.0
1.1
1.1
1.0
HFCs
46.5
127.5
168.3
168.1
170.3
169.8
174.6
Substitution of Ozone Depleting







Substancesd
0.2
107.3
163.6
164.9
164.7
166.0
170.5
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
Electronics Industry
0.2
0.2
0.3
0.3
0.4
0.4
0.3
Magnesium Production and







Processing
+
+
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.2
4.4
4.1
4.7
4.5
Electronics Industry
2.8
3.3
3.1
2.9
2.9
3.0
2.7
Aluminum Production
21.5
3.4
2.1
1.4
1.1
1.6
1.8
Substitution of Ozone Depleting







Substancesd
+
+
+
+
+
0.1
0.1
sf6
28.8
11.8
5.5
6.0
5.9
5.7
5.9
Electrical Transmission and







Distribution
23.2
8.4
3.8
4.1
4.2
3.9
4.2
Magnesium Production and







Processing
5.2
2.7
1.0
1.1
1.0
1.0
0.9
Electronics Industry
0.5
0.7
0.7
0.8
0.7
0.8
0.8
nf3
+
0.5
0.6
0.6
0.6
0.6
0.6
Electronics Industry
+
0.5
0.6
0.6
0.6
0.6
0.6
Unspecified Mix of HFCs, PFCs, SF6,







and NF3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total Emissions (Sources)
6,442.7
7,423.0
6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
LULUCF Emissions (Sources)c
7.9
16.8
27.8
13.2
26.0
23.4
23.5
LULUCF CH4 Emissions
5.0
9.3
16.6
7.7
15.3
13.8
13.8
LULUCF N20 Emissions
3.0
7.5
11.3
5.5
10.6
9.7
9.7
LULUCF Carbon Stock Change8
(908.7)
(804.8)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
LULUCF Sector Net Total'
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Net Emissions (Sources and Sinks)
5,541.9
6,635.0
5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
Notes: Total emissions presented without LULUCF. Net emissions presented with LULUCF. Totals may not sum due to
independent rounding. Parentheses indicate negative values or sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
a Emissions from Wood Biomass, Ethanol, and Biodiesel Consumption are not included specifically in summing Energy
sector totals. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for
LULUCF.
b Emissions from International Bunker Fuels are not included in totals.
c LULUCF emissions of CH4 and N20 are reported separately from gross emissions totals. LULUCF emissions include the
CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils, Grassland
Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal Wetlands;
and N20 emissions from Forest Soils and Settlement Soils. Refer to Table 2-8 for a breakout of emissions and removals
for LULUCF by gas and source category.
d Small amounts of PFC emissions also result from this source.
0 LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest
Land, Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining
Grassland, Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements
Remaining Settlements, and Land Converted to Settlements. Refer to Table 2-8 for a breakout of emissions and
removals for LULUCF by gas and source category.
Trends 2-5

-------
f The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon
stock changes.
Table 2-2: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (kt)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
co2
5,113,455
6,134,521
5,371,771
5,248,024
5,207,751
5,375,491
5,255,816
Fossil Fuel Combustion
4,731,466
5,753,507
5,008,270
4,911,532
4,854,480
4,991,420
4,856,702
Transportation
1,469,116
1,858,648
1,719,230
1,759,866
1,782,441
1,816,563
1,817,209
Electric Power
1,819,951
2,400,057
1,900,637
1,808,871
1,732,031
1,752,936
1,606,024
Industrial
853,808
852,895
797,270
792,496
790,069
813,569
822,470
Residential
338,578
358,898
317,304
292,764
293,397
338,058
336,752
Commercial
228,298
227,130
244,596
231,552
231,989
245,738
249,691
U.S. Territories
21,715
55,879
29,232
25,983
24,552
24,555
24,556
Non-Energy Use of Fuels
112,766
129,135
108,476
99,840
113,539
129,728
128,763
Petroleum Systems
9,709
12,059
32,412
21,847
24,979
37,115
47,269
Iron and Steel Production &







Metallurgical Coke







Production
104,732
70,076
47,941
43,621
40,566
42,627
41,310
Cement Production
33,484
46,194
39,907
39,439
40,324
38,971
40,896
Natural Gas Systems
32,042
25,179
29,127
30,054
31,200
33,885
37,234
Petrochemical Production
21,611
27,383
28,062
28,310
28,910
29,314
30,792
Ammonia Production
13,047
9,177
10,616
10,245
11,112
12,163
12,272
Lime Production
11,700
14,552
13,342
12,630
12,882
13,106
12,112
Incineration of Waste
8,062
12,713
11,533
11,525
11,537
11,547
11,471
Other Process Uses of







Carbonates
6,297
7,644
12,182
10,972
9,933
7,469
7,457
Urea Consumption for Non-







Agricultural Purposes
3,784
3,653
4,578
5,132
5,028
5,857
6,222
Urea Fertilization
2,417
3,504
4,728
4,877
5,051
5,192
5,341
Carbon Dioxide Consumption
1,472
1,375
4,940
4,640
4,580
4,130
4,870
Liming
4,667
4,349
3,737
3,081
3,080
2,248
2,442
Aluminum Production
6,831
4,142
2,767
1,334
1,205
1,451
1,880
Soda Ash Production
1,431
1,655
1,714
1,723
1,753
1,714
1,792
Ferroalloy Production
2,152
1,392
1,960
1,796
1,975
2,063
1,598
Titanium Dioxide Production
1,195
1,755
1,635
1,662
1,688
1,541
1,474
Glass Production
1,535
1,928
1,299
1,249
1,296
1,305
1,280
Zinc Production
632
1,030
886
838
900
999
1,026
Phosphoric Acid Production
1,529
1,342
999
998
1,028
940
891
Lead Production
516
553
473
500
513
513
540
Carbide Production and







Consumption
370
213
176
170
181
184
175
Abandoned Oil and Gas Wells
6
7
7
7
7
7
7
Magnesium Production and







Processing
1
3
3
3
3
1
1
Wood Biomass, Ethanol, and







Biodiesel Consumption0
219,413
230,700
317,742
316,610
312,304
319,647
316,191
International Bunker Fuelsb
103,463
113,232
110,908
116,611
120,121
122,112
116,064
CH4c
31,075
27,445
26,061
25,696
25,935
26,237
26,389
Enteric Fermentation
6,588
6,772
6,675
6,890
7,032
7,119
7,142
Natural Gas Systems
7,478
6,567
5,994
5,894
5,949
6,101
6,305
Landfills
7,063
5,255
4,456
4,321
4,375
4,482
4,580
Manure Management
1,485
2,062
2,316
2,385
2,395
2,467
2,495
Coal Mining
3,860
2,565
2,449
2,154
2,191
2,109
1,895
2-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Petroleum Systems
1,955
1,579
1,659
1,568
1,574
1,492
1,563
Wastewater Treatment
806
803
753
747
739
737
736
Rice Cultivation
640
720
648
631
596
623
602
Stationary Combustion
344
313
339
315
306
342
346
Abandoned Oil and Gas Wells
271
287
294
296
288
290
263
Abandoned Underground







Coal Mines
288
264
256
268
257
247
237
Mobile Combustion
256
158
105
102
100
98
95
Composting
15
75
85
91
98
90
91
Field Burning of Agricultural







Residues
15
17
18
17
17
17
17
Petrochemical Production
9
3
7
10
10
12
13
Anaerobic Digestion at Biogas







Facilities
1
2
7
7
7
7
7
Ferroalloy Production
1
+
1
1
1
1
+
Carbide Production and







Consumption
1
+
+
+
+
+
+
Iron and Steel Production &







Metallurgical Coke







Production
1
1
+
+
+
+
+
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
7
5
4
4
4
4
4
N2Oc
1,519
1,530
1,571
1,513
1,498
1,541
1,534
Agricultural Soil Management
1,060
1,052
1,169
1,108
1,099
1,135
1,156
Wastewater Treatment
63
77
85
87
89
88
88
Stationary Combustion
84
115
102
101
95
95
84
Manure Management
47
55
59
61
63
65
66
Mobile Combustion
150
139
73
70
67
63
60
Nitric Acid Production
41
38
39
34
31
32
34
Adipic Acid Production
51
24
14
23
25
35
18
N20 from Product Uses
14
14
14
14
14
14
14
Composting
1
6
6
7
7
7
7
Caprolactam, Glyoxal, and







Glyoxylic Acid Production
6
7
6
6
5
5
5
Incineration of Waste
2
1
1
1
1
1
1
Electronics Industry
+
+
1
1
1
1
1
Field Burning of Agricultural







Residues
1
1
1
1
1
1
1
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
3
3
3
3
4
4
3
HFCs
M
M
M
M
M
M
M
Substitution of Ozone







Depleting Substancesd
M
M
M
M
M
M
M
HCFC-22 Production
3
1
+
+
+
+
+
Electronics Industry
M
M
M
M
M
M
M
Magnesium Production and







Processing
0
0
+
+
+
+
+
PFCs
M
M
M
M
M
M
M
Electronics Industry
M
M
M
M
M
M
M
Aluminum Production
M
M
M
M
M
M
M
Substitution of Ozone







Depleting Substancesd
+
+
+
+
+
+
+
sf6
1
1
+
+
+
+
+
Trends 2-7

-------
Electrical Transmission and
Distribution
Magnesium Production and
Processing
Electronics Industry
NF3
Electronics Industry
Unspecified Mix of HFCs, PFCs,
SF6, and NF3
Electronics Industry
+

+
+
+
+
+
+

+
+
+
+
+
+

+
+
+
+
+
+

+
+
+
+
+
+

+
+
+
+
+
M

M
M
M
M
M
M

M
M
M
M
M
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or sequestration.
+ Does not exceed 0.5 kt.
M - Mixture of multiple gases
a Emissions from Wood Biomass, Ethanol, and Biodiesel Consumption are not included specifically in summing Energy sector
totals. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for LULUCF.
b Emissions from International Bunker Fuels are not included in totals.
c LULUCF emissions of LULUCF CH4 and N20 are reported separately from gross emissions totals. Refer to Table 2-8 for a
breakout of emissions and removals for LULUCF by gas and source category.
d Small amounts of PFC emissions also result from this source.
Emissions of all gases can be summed from each source category into a set of five sectors defined by the
Intergovernmental Panel on Climate Change (IPCC). Figure 2-4 and Table 2-3 illustrate that over the thirty-year
period of 1990 to 2019, total emissions from the Energy, Industrial Processes and Product Use, and Agriculture
sectors grew by 66.7 MMT C02 Eq. (1.3 percent), 28.2 MMT C02 Eq. (8.1 percent), and 73.3 MMT C02 Eq. (13.2
percent), respectively. Emissions from the Waste sector decreased by 52.4 MMT C02 Eq. (24.2 percent). Over the
same period, total C sequestration in the Land Use, Land-Use Change, and Forestry (LULUCF) sector decreased by
96.0 MMT C02 (10.6 percent decrease in total C sequestration), and emissions from the LULUCF sector increased
by 15.5 MMT C02 Eq. (196.1 percent).
Figure 2-4: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector
8,000
¦	Net Emissions (Including Sinks)
LULUCF (emissions)
¦	Waste
Industrial Processes and Product Use
¦	Agriculture
Energy
¦	LULUCF (removals)
2-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 2-3: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC
Sector (MMT COz Eq.)
Chapter/IPCC Sector
1990
2005
2015
2016
2017
2018
2019
Energy
5,325.6
6,302.3
5,519.8
5,390.9
5,351.0
5,518.1
5,392.3
Fossil Fuel Combustion
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Natural Gas Systems
219.0
189.4
179.0
177.4
179.9
186.4
194.9
Non-Energy Use of Fuels
112.8
129.1
108.5
99.8
113.5
129.7
128.8
Petroleum Systems
58.6
51.5
73.9
61.1
64.4
74.5
86.4
Coal Mining
96.5
64.1
61.2
53.8
54.8
52.7
47.4
Stationary Combustion
33.7
42.2
39.0
37.9
36.1
36.8
33.5
Mobile Combustion
51.1
45.5
24.4
23.4
22.3
21.3
20.3
Incineration of Waste
8.5
13.1
11.8
11.8
11.8
11.9
11.8
Abandoned Oil and Gas Wells
6.8
7.2
7.4
7.4
7.2
7.3
6.6
Abandoned Underground Coal Mines
7.2
6.6
6.4
6.7
6.4
6.2
5.9
Industrial Processes and Product Use
345.6
365.7
375.4
368.0
367.7
371.3
373.7
Substitution of Ozone Depleting







Substances
0.2
107.3
163.6
164.9
164.7
166.1
170.6
Iron and Steel Production &







Metallurgical Coke Production
104.8
70.1
47.9
43.6
40.6
42.6
41.3
Cement Production
33.5
46.2
39.9
39.4
40.3
39.0
40.9
Petrochemical Production
21.8
27.5
28.2
28.6
29.2
29.6
31.1
Ammonia Production
13.0
9.2
10.6
10.2
11.1
12.2
12.3
Lime Production
11.7
14.6
13.3
12.6
12.9
13.1
12.1
Nitric Acid Production
12.1
11.3
11.6
10.1
9.3
9.6
10.0
Other Process Uses of Carbonates
6.3
7.6
12.2
11.0
9.9
7.5
7.5
Urea Consumption for Non-







Agricultural Purposes
3.8
3.7
4.6
5.1
5.0
5.9
6.2
AdipicAcid Production
15.2
7.1
4.3
7.0
7.4
10.3
5.3
Carbon Dioxide Consumption
1.5
1.4
4.9
4.6
4.6
4.1
4.9
Electronics Industry
3.6
4.8
5.0
5.0
4.9
5.1
4.6
Electrical Transmission and







Distribution
23.2
8.4
3.8
4.1
4.2
3.9
4.2
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
Aluminum Production
28.3
7.6
4.9
2.7
2.3
3.1
3.6
Soda Ash Production
1.4
1.7
1.7
1.7
1.8
1.7
1.8
Ferroalloy Production
2.2
1.4
2.0
1.8
2.0
2.1
1.6
Titanium Dioxide Production
1.2
1.8
1.6
1.7
1.7
1.5
1.5
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
Glass Production
1.5
1.9
1.3
1.2
1.3
1.3
1.3
Zinc Production
0.6
1.0
0.9
0.8
0.9
1.0
1.0
Magnesium Production and







Processing
5.2
2.7
1.1
1.2
1.1
1.1
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
0.9
0.9
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Carbide Production and Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Agriculture
555.3
577.1
616.1
604.4
605.5
621.0
628.6
Agricultural Soil Management
315.9
313.4
348.5
330.1
327.6
338.2
344.6
Enteric Fermentation
164.7
169.3
166.9
172.2
175.8
178.0
178.6
Manure Management
51.1
67.9
75.4
77.7
78.5
81.1
82.0
Rice Cultivation
16.0
18.0
16.2
15.8
14.9
15.6
15.1
Urea Fertilization
2.4
3.5
4.7
4.9
5.1
5.2
5.3
Liming
4.7
4.3
3.7
3.1
3.1
2.2
2.4
Trends 2-9

-------
Field Burning of Agricultural Residues
0.5
0.6
0.6
0.6
0.6
0.6
0.6
Waste
216.2
178.0
159.8
157.1
159.0
161.1
163.7
Landfills
176.6
131.4
111.4
108.0
109.4
112.1
114.5
Wastewater Treatment
38.9
43.0
44.2
44.6
44.9
44.6
44.8
Composting
0.7
3.5
4.0
4.3
4.6
4.3
4.3
Anaerobic Digestion at Biogas







Facilities
+
0.1
0.2
0.2
0.2
0.2
0.2
Total Emissions (Sources)3
6,442.7
7,423.0
6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
LULUCF Sector Net Totalb
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Forest land
(884.1)
(751.4)
(749.5)
(814.7)
(740.0)
(781.4)
(774.6)
Cropland
28.6
23.2
43.2
31.7
32.3
37.7
39.7
Grassland
2.2
(29.4)
(10.1)
(13.7)
(12.5)
(11.9)
(8.0)
Wetlands
(2.8)
(1.9)
(3.9)
(3.9)
(3.8)
(3.9)
(3.9)
Settlements
(44.7)
(28.5)
(43.5)
(42.2)
(42.1)
(42.0)
(42.4)
Net Emission (Sources and Sinks)c
5,541.9
6,635.0
5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
Notes: Total emissions presented without LULUCF. Net emissions presented with LULUCF. Totals may not sum due to
independent rounding. Parentheses indicate negative values or sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
a Total emissions without LULUCF.
b LULUCF emissions of CH4 and N20 are reported separately from gross emissions totals. LULUCF emissions include the CH4
and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils, Grassland Fires, and
Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal Wetlands; and N20
emissions from Forest Soils and Settlement Soils. Refer to Table 2-8 for a breakout of emissions and removals for LULUCF
by gas and source category.
c Net emissions with LULUCF.
Energy
Energy-related activities, primarily fossil fuel combustion, accounted for the vast majority of U.S. C02 emissions for
the period of 1990 through 2019. Fossil fuel combustion is the largest source of energy-related emissions, with C02
being the primary gas emitted (see Figure 2-5 and Figure 2-6). Due to their relative importance, fossil fuel
combustion-related C02 emissions are considered in detail in the Energy chapter (see Energy chapter).
In 2019, approximately 80 percent of the energy used in the United States (on a Btu basis) was produced through
the combustion of fossil fuels. The remaining 20 percent came from other energy sources such as hydropower,
biomass, nuclear, wind, and solar energy. A discussion of specific trends related to C02 as well as other greenhouse
gas emissions from energy use is presented here with more detail in the Energy chapter. Energy-related activities
are also responsible for CH4 and N20 emissions (40.6 percent and 9.5 percent of total U.S. emissions of each gas,
respectively). Table 2-4 presents greenhouse gas emissions from the Energy chapter, by source and gas.
2-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 2-5: 2019 Energy Chapter Greenhouse Gas Sources
CO2 Emissions from Fossil Fuel Combustion
Natural Gas Systems
Non-Energy Use of Fuels
Petroleum Systems
Energy as a Portion of All
Emissions
Non-C02 Emissions from Stationary Combustion
Non-C02 Emissions from Mobile Combustion
Incineration of Waste
¦	Waste 1
I Energy
IPPU
¦	Agriculture
Abandoned Oil and Gas Wells
Abandoned Underground Coal Mines
0	50	100	150	200	250 300
MMT COz Eq.
Figure 2-6: Trends in Energy Chapter Greenhouse Gas Sources
LTJ JN
00 S
00
CT1
O

Incineration of Waste
U.S Territories Fossil Fuel Combustion
Non-Energy Use of Fuels
Commerical Fossil Fuel Combustion
I Fugitive Emissions
Residential Fossil Fuel Combustion
Industrial Fossil Fuel Combustion
I Transportation Fossil Fuel Combustion
Electric Power Fossil Fuel Combustion
Oi-HfMfo^-LDvDrsoocnoi-i(Nn^-LnvDrsoocnoi-i(Nfn^-LnvDrNoocn
CTicncncn^cncnCTicncnooooooooooi-ii-i *—i ¦ *—¦ i—
oicncncna^cncnc^cncnoooooooooooooooooooo
i—ii—ii-ii—ii—ii-Hi—ii—ii—11-irMrMfMfMfMrMfMrMrMfMrMrMfMrMrMfMfMrNjrMfM
00
cn °° o
id m n
cn	o
ai ko vo rM
T in 1/1

s >a g
to m pri.
Trends 2-11

-------
Table 2-4: Emissions from Energy (MMT CO2 Eq.)2
Gas/Source
1990
2005
2015
2016
2017
2018
2019
CO?
4,894.1
5,932.6
5,189.8
5,074.8
5,035.7
5,203.7
5,081.4
Fossil Fuel Combustion
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Transportation
1,469.1
1,858.6
1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
Electric Power
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Industrial
853.8
852.9
797.3
792.5
790.1
813.6
822.5
Residential
338.6
358.9
317.3
292.8
293.4
338.1
336.8
Commercial
228.3
227.1
244.6
231.6
232.0
245.7
249.7
U.S. Territories
21.7
55.9
29.2
26.0
24.6
24.6
24.6
Non-Energy Use of Fuels
112.8
129.1
108.5
99.8
113.5
129.7
128.8
Petroleum Systems
9.7
12.1
32.4
21.8
25.0
37.1
47.3
Natural Gas Systems
32.0
25.2
29.1
30.1
31.2
33.9
37.2
Incineration of Waste
8.1
12.7
11.5
11.5
11.5
11.5
11.5
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Biomass-WoodP
215.2
206.9
224.7
215.7
211.5
219.8
216.5
International Bunker Fuelsb
103.5
113.2
110.9
116.6
120.1
122.1
116.1
Biofuels-Ethanola
4.2
22.9
78.9
81.2
82.1
81.9
82.6
Biofuels-Bio diesela
0.0
0.9
14.1
19.6
18.7
17.9
17.1
ch4
361.3
293.3
277.4
264.9
266.6
267.0
267.6
Natural Gas Systems
186.9
164.2
149.8
147.3
148.7
152.5
157.6
Coal Mining
96.5
64.1
61.2
53.8
54.8
sin
47.4
Petroleum Systems
48.9
39.5
41.5
39.2
39.3
37.3
39.1
Stationary Combustion
8.6
7.8
8.5
7.9
7.6
8.5
8.7
Abandoned Oil and Gas Wells
6.8
7.2
7.4
7.4
7.2
7.3
6.6
Abandoned Underground Coal
7.2
6.6
6.4
6.7
6.4
6.2
5.9
Mines







Mobile Combustion
6.4
4.0
2.6
2.5
2.5
2.4
2.4
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
n2o
70.3
76.3
52.6
51.2
48.6
47.4
43.2
Stationary Combustion
25.1
34.4
30.5
30.0
28.4
28.2
24.9
Mobile Combustion
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
1.0
1.0
1.1
1.1
1.0
Total
5,325.6
6,302.3
5,519.8
5,390.9
5,351.0
5,518.1
5,392.3
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Emissions from Wood Biomass and Biofuel Consumption are not included specifically in summing energy sector totals. Net
carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for LULUCF.
b Emissions from International Bunker Fuels are not included in totals.
CO2 Emissions from Fossil Fuel Combustion
As the largest contributor to U.S. greenhouse gas emissions, C02 from fossil fuel combustion has accounted for
approximately 74 percent of gross emissions across the time series. Within the United States, fossil fuel
combustion accounted for 92.4 percent of C02 emissions in 2019. Emissions from this source category grew by 2.6
2 The full time-series data is available in Common Reporting Format (CRF) Tables included in the U.S. UNFCCC submission and in
CSV format available at .
2-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
percent (125.2 MMT C02 Eq.) from 1990 to 2019 and were responsible for most of the increase in national
emissions during this period. Conversely, C02 emissions from fossil fuel combustion decreased by 896.8 MMT C02
Eq. from 2005 and by 489.6 MMT C02 Eq. from 2010, representing decreases of approximately 15.6 percent
between 2005 and 2019 and 9.2 percent between 2010 and 2019. From 2018 to 2019, these emissions decreased
by 2.7 percent (134.7 MMT C02 Eq.). Historically, changes in emissions from fossil fuel combustion have been the
main factor influencing U.S. emission trends.
Changes in C02 emissions from fossil fuel combustion since 1990 are affected by many long-term and short-term
factors, including population and economic growth, energy price fluctuations and market trends, technological
changes, carbon intensity of energy fuel choices, and seasonal temperatures. C02 emissions from coal combustion
gradually increased between 1990 and 2007, then began to decrease at a faster rate from 2008 to 2019. C02
emissions from natural gas combustion remained relatively constant, with a slight increase between 1990 and
2009, then began to consistently increase between 2010 and 2019. The replacement of coal combustion with
natural gas combustion was largely driven by new discoveries of natural gas fields and advancements in drilling
technologies, which led to more competitive natural gas prices. On an annual basis, the overall consumption and
mix of fossil fuels in the United States fluctuates primarily in response to changes in general economic conditions,
overall energy prices, the relative price of different fuels, weather, and the availability of non-fossil alternatives.
For example, coal consumption for electric power is influenced by a number of factors, including the relative price
of coal and alternative sources, the ability to switch fuels, and longer-term trends in coal markets. Likewise,
warmer winters lead to a decrease in heating degree days and result in a decreased demand for heating fuel and
electricity for heat in the residential and commercial sectors, which leads to a decrease in emissions from reduced
fuel consumption.
Fossil fuel combustion C02 emissions also depend on the type of fuel consumed or energy used and its carbon
intensity. Producing a unit of heat or electricity using natural gas instead of coal, for example, reduces C02
emissions because of the lower C content of natural gas (see Table A-28 in Annex 2.1 for more detail on the C
Content Coefficient of different fossil fuels).
Recent trends in C02 emissions from fossil fuel combustion have been strongly influenced by trends in the electric
power sector, which from 1990 to 2017 accounted for the largest share of emissions from this source (see Figure
2-18). Electric power sector emissions are driven by the total amount of electricity generated to meet electricity
demand and the carbon intensity of the energy mix used to produce the electricity. From 1990 to 2005, power
sector C02 emissions increased 32 percent with a 34 percent increase in generation (see Figure 2-9). From 2005 to
2019, power sector C02 emissions dropped 33 percent while generation remained relatively flat (a 2 percent
increase). The types of fuel consumed to produce electricity have shifted over time, impacting emission trends.
Electricity generation from lower carbon intensity renewable energy sources increased by 115 percent from 2005
to 2019 and natural gas generation increased by 116 percent while coal generation decreased by 52 percent over
the same time period (see Table 3-12 for more detail on electricity generation by source). The decrease in coal-
powered electricity generation and increase in natural gas and renewable energy electricity generation have
contributed to the 33 percent decrease in overall C02 emissions from electric power generation from 2005 to 2019
(see Figure 2-9). Between 2018 and 2019, emissions from the electric power sector decreased 8.4 percent due to a
decrease in electric power generation of 1.4 percent and a decrease in the carbon intensity of the electric power
energy mix reflecting the continued shift in the share of electric power generation from coal to natural gas and
renewable energy.
Petroleum use is another major driver of C02 emissions from fossil fuel combustion, particularly in the
transportation sector, which represents the largest source of C02 emissions from fossil fuel combustion in 2019.
Emissions from petroleum consumption for transportation (including bunker fuels) have increased by 4.9 percent
since 2015; this trend can be primarily attributed to a 5.4 percent increase in vehicle miles traveled (VMT) over the
same time period. Fuel economy of light-duty vehicles is another important factor. The decline in new light-duty
vehicle fuel economy between 1990 and 2004 reflected the increasing market share of light-duty trucks, which
grew from about 30 percent of new vehicle sales in 1990 to 48 percent in 2004. Starting in 2005, average new
vehicle fuel economy began to increase while light-duty VMT grew only modestly for much of the period and has
slowed the rate of increase of C02 emissions.
Trends 2-13

-------
Overall, across all sectors, there was a 2.7 percent decrease in total C02 emissions from fossil fuel combustion from
2018 to 2019 and a 3.0 percent reduction since 2015. Carbon dioxide emissions from fossil fuel combustion,
separated by end-use sector, are presented in Table 2-5 and Figure 2-7 based on the underlying U.S. energy
consumer data collected by the U.S. Energy Information Administration (EIA). Figure 2-8 further describes direct
and indirect C02 emissions from fossil fuel combustion, separated by end-use sector. Estimates of C02 emissions
from fossil fuel combustion are calculated from these EIA "end-use sectors" based on total fuel consumption and
appropriate fuel properties described below. (Any additional analysis and refinement of the EIA data is further
explained in the Energy chapter of this report.)
•	Transportation. ElA's fuel consumption data for the transportation sector consists of all vehicles whose
primary purpose is transporting people and/or goods from one physical location to another.
•	Industry. EIA statistics for the industrial sector include fossil fuel consumption that occurs in the fields of
manufacturing, agriculture, mining, and construction. ElA's fuel consumption data for the industrial sector
consist of all facilities and equipment used for producing, processing, or assembling goods. (EIA includes
generators that produce electricity and/or useful thermal output primarily to support on-site industrial
activities in this sector.)
•	Electric Power. ElA's fuel consumption data for the electric power sector are comprised of electricity-only
and combined-heat-and-power (CHP) plants within the North American Industry Classification System
(NAICS) 22 category whose primary business is to sell electricity, or electricity and heat, to the public.
(Non-utility power producers are included in this sector as long as they meet the electric power sector
definition.)
•	Residential. ElA's fuel consumption data for the residential sector consist of living quarters for private
households.
•	Commercial. ElA's fuel consumption data for the commercial sector consist of service-providing facilities
and equipment from private and public organizations and businesses. (EIA includes generators that
produce electricity and/or useful thermal output primarily to support the activities at commercial
establishments in this sector.)
Table 2-5: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation
1,472.2
1,863.4
1,723.5
1,764.1
1,786.8
1,821.2
1,821.9
Combustion
1,469.1
1,858.6
1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
Electricity
3.0
4.7
4.3
4.2
4.3
4.7
4.7
Industrial
1,540.2
1,589.2
1,346.8
1,310.1
1,294.5
1,314.9
1,287.8
Combustion
853.8
852.9
797.3
792.5
790.1
813.6
822.5
Electricity
686.4
736.3
549.5
517.6
504.4
501.3
465.3
Residential
931.3
1,214.9
1,001.1
946.2
910.5
980.2
920.3
Combustion
338.6
358.9
317.3
292.8
293.4
338.1
336.8
Electricity
592.7
856.0
683.8
653.5
617.1
642.1
583.5
Commercial
766.0
1,030.1
907.6
865.2
838.2
850.6
802.1
Combustion
228.3
227.1
244.6
231.6
232.0
245.7
249.7
Electricity
537.7
803.0
663.0
633.6
606.2
604.8
552.4
U.S. Territories3
21.7
55.9
29.2
26.0
24.6
24.6
24.6
Total
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Electric Power
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Notes: Combustion-related emissions from electric power are allocated based on aggregate national
electricity use by each end-use sector. Totals may not sum due to independent rounding.
a Fuel consumption by U.S. Territories (i.e., American Samoa, Guam, Puerto Rico, U.S. Virgin Islands, Wake
Island, and other U.S. Pacific Islands) is included in this report.
2-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 2-
7:
2019 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type
2,500
2,000
iS" 1,500

-------
Figure 2-9: Electric Power Generation (Billion kWh) and Emissions
Electric power C02 emissions can also be allocated to the end-use sectors that use electricity, as presented in Table
2-5. With electricity C02 emissions allocated to end-use sectors, the transportation end-use sector represents the
largest source of fossil fuel combustion emissions accounting for 1,821.9 MMT C02 Eq. in 2019 or approximately 38
percent of total C02 emissions from fossil fuel combustion. The industrial end-use sector accounted for 27 percent
of C02 emissions from fossil fuel combustion when including allocated electricity emissions. The residential and
commercial end-use sectors accounted for 19 and 17 percent, respectively, of C02 emissions from fossil fuel
combustion when including allocated electricity emissions. Both of these end-use sectors were heavily reliant on
electricity for meeting energy needs, with electricity use for lighting, heating, air conditioning, and operating
appliances contributing 63 and 69 percent of emissions from the residential and commercial end-use sectors,
respectively.
4 500 ® Nuclear Generation (Billion kWh)
¦	Renewable Generation (Billion kWh)
¦	Petroleum Generation (Billion kWh)
4 000 ® Coal Generation (Billion kWh)
¦	Natural Gas Generation (Billion kWh)
] Total Emissions (MMT CO2 Eq.) [Right Axis]	3,500
2,500
2,000
1,500 e
~ 3,500
i
c
1	3,000
c
o
2	2,500

-------
•	Methane emissions from coal mining decreased by 49.1 MMT C02 Eq. (50.9 percent) from 1990 through
2019, primarily due to a decrease in the number of active mines and annual coal production over the time
period.
•	Nitrous oxide emissions from mobile combustion decreased by 26.8 MMT C02 Eq. (58.9 percent) from
1990 through 2019, primarily as a result of national vehicle criteria pollutant emissions standards and
emission control technologies for on-road vehicles.
•	Carbon dioxide emissions from non-energy uses of fossil fuels increased by 16.0 MMT C02 Eq. (14.2
percent) from 1990 through 2019. Emissions from non-energy uses of fossil fuels were 128.8 MMT C02
Eq. in 2019, which constituted 2.4 percent of total national C02 emissions, approximately the same
proportion as in 1990.
•	Carbon dioxide emissions from incineration of waste (11.5 MMT C02 Eq. in 2019) increased by 3.4 MMT
C02 Eq. (42.3 percent) from 1990 through 2019, as the volume of scrap tires and other fossil C-containing
materials in waste increased.
Industrial Processes and Product Use
In many cases, greenhouse gas emissions are generated and emitted in two different ways. First, they are
generated and emitted as the byproducts of many non-energy-related industrial activities. For example, industrial
processes can chemically or physically transform raw materials, which often release waste gases such as C02, CH4,
N20, and fluorinated gases (e.g., HFC-23). In the case of byproduct emissions, the emissions are generated by an
industrial process itself, and are not directly a result of energy consumed during the process.
Second, industrial manufacturing processes and use by end-consumers also release HFCs, PFCs, SF6, and NF3 and
other fluorinated compounds. In addition to the use of HFCs and some PFCs as substitutes for ozone depleting
substances (ODS), fluorinated compounds such as HFCs, PFCs, SF6, NF3, and others are also emitted through use by
a number of other industrial sources in the United States. These industries include the electronics industry, electric
power transmission and distribution, and magnesium metal production and processing. In addition, N20 is used in
and emitted by the electronics industry and anesthetic and aerosol applications. Figure 2-10 and Table 2-6
presents greenhouse gas emissions from industrial processes and product use by source category. Overall,
emission sources in the Industrial Processes and Product Use (IPPU) chapter account for 5.7 percent of U.S.
greenhouse gas emissions in 2019.
Trends 2-17

-------
Figure 2-10: 2019 Industrial Processes and Product Use Chapter Greenhouse Gas Source
Substitution of Ozone Depleting Substances
Iron and Steel Production & Metallurgical Coke Production
Cement Production
Petrochemical Production
Ammonia Production
Lime Production
Nitric Acid Production
Other Process Uses of Carbonates
Urea Consumption for Non-Agricultural Purposes
Adipic Acid Production
Carbon Dioxide Consumption
Electronics Industry
Electrical Transmission and Distribution
N2O from Product Uses
HCFC-22 Production
Aluminum Production
Soda Ash Production
Ferroalloy Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Glass Production
Zinc Production
Magnesium Production and Processing
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
171
Industrial Processes and Product Use
as a Portion of All Emissions
5.7%
\
Energy
I Agriculture
IPPU
Waste
< 0.5
10
20
30 40
MMT COz Eq.
50
60
70
2-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 2-11: Trends in Industrial Processes and Product Use Chapter Greenhouse Gas
Sources
400
350
300
250
O
u
200
150
100
50
Ln
iv
ro
vo
ro
Ln
vo
m
vo
ro
ld
r*»
m
Electronics Industry
Other Product Manufacture and Use
Mineral Products
Metal Production
Chemical Industry
Substitution of Ozone Depleting Substances
o^rNro^-LniorvcocnoT-HfNro^-LnvDrvooCTiOT-HrMro^j-LnkorvooCTi
0^0^0*0^0^0*0^0^0*0^0000000000i—I i—i i—i i—I i—I i—l i—II iH tHI i—I
cric^o^cna^c^cna^c^cnoooooooooooooooooooo
cn
vo
ro
oo
vo
ro
co co
VO vo
ro ro
ro
iv
ro
Table 2-6: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)
Gas/Source
1990

2005

2015
2016
2017
2018
2019
C02
212.3

194.1

173.5
165.3
163.9
164.3
166.6
Iron and Steel Production & Metallurgical Coke









Production
104.7

70.1

47.9
43.6
40.6
42.6
41.3
Iron and Steel Production
99.1

66.2

43.5
41.0
38.6
41.3
39.9
Metallurgical Coke Production
5.6

3.9

4.4
2.6
2.0
1.3
1.4
Cement Production
33.5

46.2

39.9
39.4
40.3
39.0
40.9
Petrochemical Production
21.6

27.4

28.1
28.3
28.9
29.3
30.8
Ammonia Production
13.0

9.2

10.6
10.2
11.1
12.2
12.3
Lime Production
11.7

14.6

13.3
12.6
12.9
13.1
12.1
Other Process Uses of Carbonates
6.3

7.6

12.2
11.0
9.9
7.5
7.5
Urea Consumption for Non-Agricultural









Purposes
3.8

3.7

4.6
5.1
5.0
5.9
6.2
Carbon Dioxide Consumption
1.5

1.4

4.9
4.6
4.6
4.1
4.9
Aluminum Production
6.8

4.1

2.8
1.3
1.2
1.5
1.9
Soda Ash Production
1.4

1.7

1.7
1.7
1.8
1.7
1.8
Ferroalloy Production
2.2

1.4

2.0
1.8
2.0
2.1
1.6
Titanium Dioxide Production
1.2

1.8

1.6
1.7
1.7
1.5
1.5
Glass Production
1.5

1.9

1.3
1.2
1.3
1.3
1.3
Zinc Production
0.6

1.0

0.9
0.8
0.9
1.0
1.0
Phosphoric Acid Production
1.5

1.3

1.0
1.0
1.0
0.9
0.9
Lead Production
0.5

0.6

0.5
0.5
0.5
0.5
0.5
Carbide Production and Consumption
0.4

0.2

0.2
0.2
0.2
0.2
0.2
Trends 2-19

-------
Magnesium Production and Processing
+
+
+
+
+
+
+
ch4
0.3
0.1
0.2
0.3
0.3
0.3
0.4
Petrochemical Production
0.2
0.1
0.2
0.2
0.3
0.3
0.3
Ferroalloy Production
+
+
+
+
+
+
+
Carbide Production and Consumption
+
+
+
+
+
+
+
Iron and Steel Production & Metallurgical Coke







Production
+
+
+
+
+
+
+
Iron and Steel Production
+
+
+
+
+
+
+
Metallurgical Coke Production
0.0
0.0
0.0
0.0
0.0
0.0
0.0
n2o
33.3
24.9
22.2
23.3
22.7
25.8
21.1
Nitric Acid Production
12.1
11.3
11.6
10.1
9.3
9.6
10.0
AdipicAcid Production
15.2
7.1
4.3
7.0
7.4
10.3
5.3
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Caprolactam, Glyoxal, and Glyoxylic Acid







Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
Electronics Industry
+
0.1
0.2
0.2
0.3
0.3
0.2
HFCs
46.5
127.5
168.3
168.1
170.3
169.8
174.6
Substitution of Ozone Depleting Substances3
0.2
107.3
163.6
164.9
164.7
166.0
170.5
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
Electronics Industry
0.2
0.2
0.3
0.3
0.4
0.4
0.3
Magnesium Production and Processing
0.0
0.0
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.2
4.4
4.1
4.7
4.5
Electronics Industry
2.8
3.3
3.1
2.9
2.9
3.0
2.7
Aluminum Production
21.5
3.4
2.1
1.4
1.1
1.6
1.8
Substitution of Ozone Depleting Substances
0.0
+
+
+
+
0.1
0.1
sf6
28.8
11.8
5.5
6.0
5.9
5.7
5.9
Electrical Transmission and Distribution
23.2
8.4
3.8
4.1
4.2
3.9
4.2
Magnesium Production and Processing
5.2
2.7
1.0
1.1
1.0
1.0
0.9
Electronics Industry
0.5
0.7
0.7
0.8
0.7
0.8
0.8
nf3
+
0.5
0.6
0.6
0.6
0.6
0.6
Electronics Industry
+
0.5
0.6
0.6
0.6
0.6
0.6
Unspecified Mix of HFCs, NF3, PFCs and SF6
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total
345.6
365.7
375.4
368.0
367.7
371.3
373.7
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Small amounts of PFC emissions also result from this source.
Overall, emissions from the IPPU sector increased by 8.1 percent from 1990 to 2019. Significant trends in emissions
from IPPU source categories (Figure 2-11) over the thirty-year period from 1990 through 2019 included the
following:
•	HFC and PFC emissions resulting from the substitution of ODS (e.g., chlorofluorocarbons [CFCs]) have
been increasing from small amounts in 1990 to 170.6 MMT C02 Eq. in 2019 and accounted for 45.6
percent of total IPPU emissions.
•	Combined C02 and CH4 emissions from iron and steel production and metallurgical coke production
decreased by 3.1 percent to 41.3 MMT C02 Eq. from 2018 to 2019, and have declined overall by 63.4
MMT C02 Eq. (60.6 percent) from 1990 through 2019, due to restructuring of the industry. The trend in
the United States has been a shift towards fewer integrated steel mills and more EAFs. EAFs use scrap
steel as their main input and generally have less on-site emissions.
•	Carbon dioxide emissions from petrochemicals increased by 42.5 percent between 1990 and 2019 from
21.6 MMT C02 Eq. to 30.8 MMT C02 Eq. The increase in emissions is largely driven by an almost doubling
of production of ethylene over that time period.
2-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Carbon dioxide emissions from ammonia production (12.3 MMT C02 Eq. in 2019) decreased by 5.9
percent (0.8 MMT C02 Eq.) since 1990. Ammonia production relies on natural gas as both a feedstock and
a fuel, and as such, market fluctuations and volatility in natural gas prices affect the production of
ammonia from year to year. Emissions from ammonia production have increased steadily since 2016, due
to the addition of new ammonia production facilities and new production units at existing facilities.
Agricultural demands continue to drive demand for nitrogen fertilizers and the need for new ammonia
production capacity.
•	Carbon dioxide emissions from cement production increased by 22.1 percent (7.4 MMT C02 Eq.) from
1990 through 2019. They rose from 1990 through 2006 and then fell until 2009, due to a decrease in
demand for construction materials during the economic recession. Since 2010, C02 emissions from
cement production have risen 30.0 percent (9.4 MMT C02 Eq.).
•	PFC emissions from aluminum production decreased by 91.8 percent (19.7 MMT C02 Eq.) from 1990 to
2019, due to both industry emission reduction efforts and lower domestic aluminum production.
Agricultural activities contribute directly to emissions of greenhouse gases through a variety of processes,
including the following source categories: enteric fermentation in domestic livestock, livestock manure
management, rice cultivation, agricultural soil management, liming, urea fertilization, and field burning of
agricultural residues. Methane, N20, and C02 were the primary greenhouse gases emitted by agricultural activities.
Carbon stock changes from agricultural soils are included in the LULUCF sector.
In 2019, agricultural activities were responsible for emissions of 628.6 MMT C02 Eq., or 9.6 percent of total U.S.
greenhouse gas emissions. Methane emissions from enteric fermentation and manure management represented
approximately 27.1 percent and 9.5 percent of total CH4 emissions from anthropogenic activities, respectively, in
2019. Agricultural soil management activities, such as application of synthetic and organic fertilizers, deposition of
livestock manure, and growing N-fixing plants, were the largest contributors to U.S. N20 emissions in 2019,
accounting for 75.4 percent. Carbon dioxide emissions from the application of crushed limestone and dolomite
(i.e., soil liming) and urea fertilization represented 0.1 percent of total C02 emissions from anthropogenic
activities. Figure 2-12 and Table 2-7 illustrate agricultural greenhouse gas emissions by source.
Figure 2-12: 2019 Agriculture Chapter Greenhouse Gas Sources
Agriculture
Field Burning of Agricultural Residues
Agricultural Soil Management
Manure Management
Enteric Fermentation
Urea Fertilization
Rice Cultivation
Liming
Agriculture as a Portion of
All Emissions
9.6% 	
I Energy
¦ Agriculture
IPPU
Waste
| 345
0 20 40 60 80 100 120 140 160 180 200
MMT COz Eq.
Trends 2-21

-------
Figure 2-13: Trends in Agriculture Chapter Greenhouse Gas Sources
650
600
550
500
450
400
350
300
250
200
150
100
50
Field Burning of Agricultural Residues
Urea Fertilization
Liming
Rice Cultivation
I Manure Management
I Enteric Fermentation
I Agricultural Soil Management
^ lO lO
° <3
KD
0
1—1
f\i
m

LD
VO
r>
00
cn
0
1—1
rM
m

LD
10
rv
00
cn
O
1—1
rM
m

m
VD
rv
00
cn
cn
cn

cn
cn
cn
cn
cn

cn
0
0
O
O
0
O
0
0
0
0
1—1
i—l
1—1
1—1
1—1
1—1
1—1
1—1
T—1
1—1

cn
cr>
cn
cn
cn
cn
cn
cn
cn
0
0
0
O
0
O
0
0
0
0
O
O
0
0
0
0
O
0
O
0
1—1
1—1
1—1
1—1
1—1
1—1
1—1
1—1
1—1
1—1
rM
pm
fM
CM
CM
rM
rM
rM
rM
CM
CM
CM
rM
rM
rM
CM
rM
rM
rM
rM
Table 2-7: Emissions from Agriculture (MMT CO2 Eq.)
Gas/Source
1990

2005

2015
2016
2017
2018
2019
C02
7.1

7.9

8.5
8.0
8.1
7.4
7.8
Urea Fertilization
2.4

3.5

4.7
4.9
5.1
5.2
5.3
Liming
4.7

4.3

3.7
3.1
3.1
2.2
2.4
ch4
218.2

239.3

241.4
248.1
251.0
255.7
256.4
Enteric Fermentation
164.7

169.3

166.9
172.2
175.8
178.0
178.6
Manure Management
37.1

51.6

57.9
59.6
59.9
61.7
62.4
Rice Cultivation
16.0

18.0

16.2
15.8
14.9
15.6
15.1
Field Burning of Agricultural









Residues
0.4

0.4

0.4
0.4
0.4
0.4
0.4
n2o
330.1

329.9

366.2
348.4
346.4
357.9
364.4
Agricultural Soil Management
315.9

313.4

348.5
330.1
327.6
338.2
344.6
Manure Management
14.0

16.4

17.5
18.1
18.7
19.4
19.6
Field Burning of Agricultural









Residues
0.2

0.2

0.2
0.2
0.2
0.2
0.2
Total
555.3

577.1

616.1
604.4
605.5
621.0
628.6
Note: Totals may not sum due to independent rounding.
2-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Some significant trends in U.S. emissions from Agriculture source categories (Figure 2-13) over the thirty-year
period from 1990 through 2019 included the following:
•	Agricultural soils are the largest anthropogenic source of N20 emissions in the United States, accounting
for approximately 75.4 percent of N20 emissions in 2019 and 5.3 percent of total emissions in the United
States in 2019. Estimated emissions from this source in 2019 were 344.6 MMT C02 Eq. Annual N20
emissions from agricultural soils fluctuated between 1990 and 2019, although overall emissions were 28.7
MMT C02 Eq. or 9.1 percent higher in 2019 than in 1990. Year-to-year fluctuations are largely a reflection
of annual variation in weather patterns, synthetic fertilizer use, and crop production.
•	Enteric fermentation is the largest anthropogenic source of CH4 emissions in the United States. In 2019,
enteric fermentation CH4 emissions were 27.1 percent of total CH4 emissions (178.6 MMT C02 Eq.), which
represents an increase of 13.9 MMT C02 Eq. (8.4 percent) since 1990. This increase in emissions from
1990 to 2019 in enteric fermentation generally follows the increasing trends in cattle populations. For
example, from 1990 to 1995, emissions increased and then generally decreased from 1996 to 2004,
mainly due to fluctuations in beef cattle populations and increased digestibility of feed for feedlot cattle.
Emissions increased from 2005 to 2007, as both dairy and beef populations increased. Research indicates
that the feed digestibility of dairy cow diets decreased during this period. Emissions decreased again from
2008 to 2014 as beef cattle populations again decreased. Emissions increased from 2014 to 2019,
consistent with an increase in beef cattle population over those same years.
•	Overall, emissions from manure management increased 60.3 percent between 1990 and 2019. This
encompassed an increase of 67.9 percent for CH4, from 37.1 MMT C02 Eq. in 1990 to 62.4 MMT C02 Eq. in
2019; and an increase of 40.2 percent for N20, from 14.0 MMT C02 Eq. in 1990 to 19.6 MMT C02 Eq. in
2019. The majority of the increase observed in CH4 resulted from swine and dairy cattle manure, where
emissions increased 48.6 and 117.3 percent, respectively, from 1990 to 2019. From 2018 to 2019, there
was a 1.1 percent increase in total CH4 emissions from manure management, mainly due to minor shifts in
the animal populations and the resultant effects on manure management system allocations.
•	Liming and urea fertilization are the only sources of C02 emissions reported in the Agriculture sector. All
other C02 emissions and removals are characterized in the LULUCF sector. Estimated emissions from
these sources were 2.4 and 5.3 MMT C02 Eq., respectively. Liming emissions increased by 8.6 percent
relative to 2018 and decreased 2.2 MMT C02 Eq. or 47.7 percent relative to 1990, while urea fertilization
emissions increased by 2.9 percent relative to 2018 and 2.9 MMT C02 Eq. or 121.0 percent relative to
1990.
Land Use, Land-Use Change, and Forestry
When humans alter the terrestrial biosphere through land use, changes in land use, and land management
practices, they also influence the carbon (C) stock fluxes on these lands and cause emissions of CH4 and N20.
Overall, managed land is a net sink for C02 (C sequestration) in the United States. The primary driver of fluxes on
managed lands is from management of forest lands, but also includes trees in settlements (i.e., urban areas),
afforestation, conversion of forest lands to settlements and croplands the management of croplands and
grasslands, and the landfilling of yard trimmings and food scraps. The main drivers for net forest sequestration
include net forest growth, increasing forest area, and a net accumulation of C stocks in harvested wood pools. The
net sequestration in Settlements Remaining Settlements, is driven primarily by C stock gains in urban forests (i.e.,
Settlement Trees) through net tree growth and increased urban area, as well as long-term accumulation of C in
landfills from additions of yard trimmings and food scraps.
Trends 2-23

-------
The LULUCF sector in 2019 resulted in a net increase in C stocks (i.e., net C02 removals) of 812.7 MMT C02 Eq.
(Table 2-8).3 This represents an offset of approximately 12.3 percent of total (i.e., gross) greenhouse gas emissions
in 2019. Emissions of CH4 and N20 from LULUCF activities in 2019 were 23.5 MMT C02 Eq. and represent 0.4
percent of total greenhouse gas emissions.4 Between 1990 and 2019, total C sequestration in the LULUCF sector
decreased by 10.6 percent, primarily due to a decrease in the rate of net C accumulation in forests and Cropland
Remaining Cropland, as well as an increase in C02 emissions from Land Converted to Settlements.
Forest fires were the largest source of CH4 emissions from LULUCF in 2019, totaling 9.5 MMT C02 Eq. (379 kt of
CH4). Coastal Wetlands Remaining Coastal Wetlands resulted in CH4 emissions of 3.8 MMT C02 Eq. (153 kt of CH4).
Grassland fires resulted in CH4 emissions of 0.3 MMT C02 Eq. (12 kt of CH4). Land Converted to Wetlands, Drained
Organic Soils, and Peatlands Remaining Peatlands resulted in CH4 emissions of less than 0.05 MMT C02 Eq. each.
Forest fires were also the largest source of N20 emissions from LULUCF in 2019, totaling 6.2 MMT C02 Eq. (21 kt of
N20). Nitrous oxide emissions from fertilizer application to settlement soils in 2019 totaled to 2.4 MMT C02 Eq. (8
kt of N20). Additionally, the application of synthetic fertilizers to forest soils in 2019 resulted in N20 emissions of
0.5 MMT C02 Eq. (2 kt of N20). Grassland fires resulted in N20 emissions of 0.3 MMT C02 Eq. (1 kt of N20). Coastal
Wetlands Remaining Coastal Wetlands and Drained Organic Soils resulted in N20 emissions of 0.1 MMT C02 Eq.
each (0.5 kt of N20). Peatlands Remaining Peatlands resulted in N20 emissions of less than 0.05 MMT C02 Eq.
Carbon dioxide removals from C stock changes are presented (green) in Figure 2-14. Figure 2-15 and Table 2-8
along with CH4 and N20 emissions (purple) for LULUCF source categories.
Figure 2-14: 2019 LULUCF Chapter Greenhouse Gas Sources and Sinks
Forest Land Remaining Forest Land


Settlements Remaining Settlements


Land Converted to Forest Land


Land Converted to Grassland
¦

Cropland Remaining Cropland
¦

Wetlands Remaining Wetlands
I

Land Converted to Wetlands

l< 0.5|
Non-COz Emissions from Peatlands Remaining Peatlands

l< 0.5|
Non-C02 Emissions from Drained Organic Soils

l< 0.5|
CH4 Emissions from Land Converted to Coastal Wetlands

l< 0.5|
N2O Emissions from Forest Soils

l< 0.5|
Non-C02 Emissions from Grassland Fires


N2O Emissions from Settlement Soils


Non-C02 Emissions from Coastal Wetlands Remaining Coastal Wetlands


Grassland Remaining Grassland

1
Non-C02 Emissions from Forest Fires

¦
Land Converted to Cropland
Land Converted to Settlements
Carbon Stock Change
¦ Non-C02 Emissions


(250) (200) (150) (100) (50) 0 50 100
MMT COz Eq.
3	LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining Settlements,
and Land Converted to Settlements.
4	LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
2-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 2-15: Trends in Emissions and Removals (Net CO2 Flux) from Land Use, Land-Use
Change, and Forestry3
400
300
200
100
-100
-200
« -300
O
u
Z -400
z
-500
-600
-700
-800
-900
-1,000
I Land Converted to Settlements
I Land Converted to Cropland
Grassland Remaining Grassland
Land Converted to Wetlands
Wetlands Remaining Wetlands
Cropland Remaining Cropland
Land Converted to Grassland
Land Converted to Forest Land
Settlements Remaining Settlements
Forest Land Remaining Forest Land
I Net Emission (Sources and Sinks)
HHitimt
Oi-HfMn'fl-LnvDrsoocnoi-irsjfn^rLovDrN.oocnoi-irMro^rmiDr-soo
cncncncnchcncncricncnooooooooooi-ii—ii—ii—1 1 i •*—1 ¦»—1 ¦»—1
cncncncncncncncncnoiooooooooooooooooooo
1—1 1—1 1—ii—ii—ii—1 vH 1—1 1—1 *Hf>jpsjr\jfNjrsjrNjrMfMfMr\irMfMrMfMfMfMf\IfMfM
a In Figure 2-15, the values above stacked bars represent only non-C02 LULUCF emission. LULUCF emissions include the CH4 and
N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils, Grassland Fires, and Coastal
Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal Wetlands; and N20 emissions from
Forest Soils and Settlement Soils
Table 2-8: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-
Use Change, and Forestry (MMT CO2 Eq.)
Land-Use Category
1990
Forest Land Remaining Forest Land	(785.9)
Changes in Forest Carbon Stocks3	(787.6)
Non-C02 Emissions from Forest Firesb	1.5
N20 Emissions from Forest Soilsc	0.1
Non-C02 Emissions from Drained Organic
Soilsd	0.1
Land Converted to Forest Land	(98.2)
Changes in Forest Carbon Stocks0	(98.2)
Cropland Remaining Cropland	(23.2)
Changes in Mineral and Organic Soil
Carbon Stocks	(23.2)
Land Converted to Cropland	51.8
Changes in all Ecosystem Carbon Stocks'	51.8
Grassland Remaining Grassland	8.5
2005
(652.8)
(661.5)
2015
2016
2017
2018
2019
0.5
0.1
(98.7)
(98.7)
(29.0)
(29.0)
52.2
52.2
10.7
(650.6)
(671.4)
20.3
0.5
0.1
(98.9)
(98.9)
(12.8)
(12.8)
56.1
56.1
13.8
(715.7)
(721.9)
5.6
0.5
0.1
(99.0)
(99.0)
(22.7)
(22.7)
54.4
54.4
10.4
(640.9)
(659.7)
18.3
0.5
0.1
(99.1)
(99.1)
(22.3)
(22.3)
54.6
54.6
11.9
(682.4)
(698.6)
15.7
0.5
0.1
(99.1)
(99.1)
(16.6)
(16.6)
54.3
54.3
12.3
(675.5)
(691.8)
15.7
0.5
0.1
(99.1)
(99.1)
(14.5)
(14.5)
54.2
54.2
15.1
Trends 2-25

-------
Changes in Mineral and Organic Soil
Carbon Stocks
8.3
10.0
13.1
9.8
11.3
11.7
14.5
Non-C02 Emissions from Grassland Fires5
0.2
0.7
0.7
0.6
0.6
0.6
0.6
Land Converted to Grassland
(6.2)
(40.1)
(23.9)
(24.0)
(24.4)
(24.1)
(23.2)
Changes in all Ecosystem Carbon Stocks'
(6.2)
(40.1)
(23.9)
(24.0)
(24.4)
(24.1)
(23.2)
Wetlands Remaining Wetlands
(3.5)
(2.6)
(4.1)
(4.1)
(4.0)
(4.0)
(4.0)
Changes in Organic Soil Carbon Stocks in







Peatlands
1.1
1.1
0.8
0.7
0.8
0.8
0.8
Changes in Aboveground and Soil Carbon







Stocks in Coastal Wetlands
(8.5)
(7.6)
(8.8)
(8.8)
(8.8)
(8.8)
(8.8)
CH4 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
3.7
3.8
3.8
3.8
3.8
3.8
3.8
N20 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
0.1
0.2
0.1
0.1
0.1
0.1
0.1
Non-C02 Emissions from Peatlands







Remaining Peatlands
+
+
+
+
+
+
+
Land Converted to Wetlands
0.7
0.7
0.2
0.2
0.2
0.2
0.2
Changes in Aboveground and Soil Carbon







Stocks
0.4
0.4
(0.1)
(+)
(+)
(+)
(+)
CH4 Emissions from Land Converted to







Coastal Wetlands
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Settlements Remaining Settlements
(107.6)
(113.5)
(123.7)
(121.5)
(121.4)
(121.2)
(121.7)
Changes in Organic Soil Carbon Stocks
11.3
12.2
15.7
16.0
16.0
15.9
15.9
Changes in Settlement Tree Carbon







Stocks
(96.4)
(117.4)
(130.4)
(129.8)
(129.8)
(129.8)
(129.8)
Changes in Yard Trimming and Food







Scrap Carbon Stocks in Landfills
(24.5)
(11.4)
(11.1)
(10.0)
(9.8)
(9.8)
(10.2)
N20 Emissions from Settlement Soilsh
2.0
3.1
2.2
2.2
2.3
2.4
2.4
Land Converted to Settlements
62.9
85.0
80.1
79.4
79.3
79.3
79.2
Changes in all Ecosystem Carbon Stocks'
62.9
85.0
80.1
79.4
79.3
79.3
79.2
LULUCF Carbon Stock Change'
(908.7)
(804.8)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
LULUCF Emissions1
7.9
16.8
27.8
13.2
26.0
23.4
23.5
LULUCF CH4 Emissions
5.0
9.3
16.6
7.7
15.3
13.8
13.8
LULUCF N20 Emissions
3.0
7.5
11.3
5.5
10.6
9.7
9.7
LULUCF Sector NetTotalk
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
a Includes the net changes to carbon stocks stored in all forest ecosystem pools and harvested wood products.
b Estimates include emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
c Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
d Estimates include CH4 and N20 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
6 Includes the net changes to carbon stocks stored in all forest ecosystem pools.
' Includes changes in mineral and organic soil carbon stocks for all land use conversions to cropland, grassland, and
settlements, respectively. Also includes aboveground/belowground biomass, dead wood, and litter carbon stock changes
for conversion of forest land to cropland, grassland, and settlements, respectively.
5 Estimates include CH4 and N20 emissions from fires on both Grassland Remaining Grassland and Land Converted to
Grass/and.
h Estimates include N20 emissions from N fertilizer additions on both Settlements Remaining Settlements and Land
Converted to Settlements because it is not possible to separate the activity data at this time.
1 LULUCF Carbon Stock Change includes any C stock gains and losses from all land use and land use conversion categories.
> LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
k The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes in units of MMT C02 Eq.
2-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Other significant trends from 1990 to 2019 in emissions from LULUCF categories (Figure 2-15) over the thirty-year
period from 1990 through 2019 included the following:
•	Annual carbon (C) sequestration by forest land (i.e., annual C stock accumulation in the five ecosystem C
pools and harvested wood products for Forest Land Remaining Forest Land and Land Converted to Forest
Land) has decreased by approximately 10.7 percent since 1990. This is primarily due to decreased C stock
gains in Land Converted to Forest Land and the harvested wood products pools within Forest Land
Remaining Forest Land.
•	Annual C sequestration from Settlements Remaining Settlements (which includes organic soils, settlement
trees, and landfilled yard trimmings and food scraps) has increased by 13.2 percent over the period from
1990 to 2019. This is primarily due to an increase in urbanized land area in the United States with trees
growing on it.
•	Annual emissions from Land Converted to Settlements increased by approximately 26.0 percent from
1990 to 2019 due primarily to C stock losses from Forest Land Converted to Settlements and mineral soils
C stocks from Grassland Converted to Settlements.
Waste
Waste management and treatment activities are sources of greenhouse gas emissions (see Figure 2-16). In 2019,
landfills were the third-largest source of U.S. anthropogenic CH4 emissions, generating 114.5 MMT C02 Eq. and
accounting for 17.4 percent of total U.S. CH4 emissions.5 Additionally, wastewater treatment generates emissions
of 44.8 MMT C02 Eq. and accounts for 27.3 percent of waste emissions, 2.8 percent of U.S. CH4 emissions, and 5.8
percent of U.S. N20 emissions. Emissions of CH4 and N20 from composting are also accounted for in this chapter,
generating emissions of 2.3 MMT C02 Eq. and 2.0 MMT C02 Eq., respectively. Overall, emission sources accounted
for in the Waste chapter generated 163.7 MMT C02Eq., or 2.5 percent of total U.S. greenhouse gas emissions in
2019. A summary of greenhouse gas emissions from the Waste chapter is presented in Table 2-9.
Figure 2-16: 2019 Waste Sector Greenhouse Gas Sources
Landfills
Wastewater Treatment
Composting
Anaerobic Digestion at
Biogas Facilities
Waste as a Portion of AN
Emissions
Energy
Agriculture
IPPU
Waste
114
10 20 30 40 50 60 70
MMT CO2 Eq.
80
90
100
110
120
5 Landfills also store carbon, due to incomplete degradation of organic materials such as wood products and yard trimmings, as
described in the Land Use, Land-Use Change, and Forestry chapter.
Trends 2-27

-------
Figure 2-17: Trends in Waste Chapter Greenhouse Gas Sources
50 ¦ Anaerobic Digestion at Biogas Facilities
¦	Composting
¦	Wastewater Treatment
. ¦ Landfills
Table 2-9: Emissions from Waste (MMT CO2 Eq.)
Gas/Source
1990

2005

2015
2016
2017
2018
2019
ch4
197.1

153.4

132.5
129.2
130.5
132.9
135.3
Landfills
176.6

131.4

111.4
108.0
109.4
112.1
114.5
Wastewater Treatment
20.2

20.1

18.8
18.7
18.5
18.4
18.4
Composting
0.4

1.9

2.1
2.3
2.4
2.3
2.3
Anaerobic Digestion at









Biogas Facilities
+

0.1

0.2
0.2
0.2
0.2
0.2
n2o
19.0

24.6

27.3
27.9
28.6
28.2
28.4
Wastewater Treatment
18.7

23.0

25.4
25.9
26.4
26.1
26.4
Composting
0.3

1.7

1.9
2.0
2.2
2.0
2.0
Total
216.2

178.0

159.8
157.1
159.0
161.1
163.7
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Some significant trends in U.S. emissions from waste source categories (Figure 2-17) over the thirty-year period
from 1990 through 2019 included the following:
•	From 1990 to 2019, net CH4 emissions from landfills decreased by 62.1 MMT C02 Eq. (35.2 percent), with
small increases occurring in interim years. This downward trend in emissions coincided with increased
landfill gas collection and control systems, and a reduction of decomposable materials (i.e., paper and
paperboard, food scraps, and yard trimmings) discarded in municipal solid waste (MSW) landfills over the
time series.
•	From 1990 to 2019, CH4 and N20 emissions from wastewater treatment decreased by 1.8 MMT C02 Eq.
(8.7 percent) and increased by 7.7 MMT C02 Eq. (41.0 percent), respectively. Methane emissions from
domestic wastewater treatment have decreased since 1999 due to decreasing percentages of wastewater
being treated in anaerobic systems, including reduced use of on-site septic systems and central anaerobic
treatment systems. Nitrous oxide emissions from wastewater treatment processes gradually increased
across the time series as a result of increasing U.S. population and protein consumption.
•	Combined CH4 and N20 emissions from composting have generally increased approximately 3.6 MMT C02
Eq. since 1990, from 0.7 MMT C02 Eq. to 4.3 MMT C02 Eq. in 2019, which represents more than a five-fold
2-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
increase over the time series. The growth in composting since the 1990s is attributable to primarily four
factors: (1) the enactment of legislation by state and local governments that discouraged the disposal of
yard trimmings and food waste in landfills; (2) yard trimming collection and yard trimming drop off sites
provided by local solid waste management districts; (3) an increased awareness of the environmental
benefits of composting; and (4) loans or grant programs to establish or expand composting infrastructure.
2.2 Emissions by Economic Sector
Throughout this report, emission estimates are grouped into five sectors (i.e., chapters) defined by the IPCC and
detailed above: Energy, IPPU, Agriculture, LULUCF, and Waste. It is also useful to characterize emissions according
to commonly used economic sector categories: residential, commercial, industry, transportation, electric power,
and agriculture. Emissions from U.S. Territories are reported as their own end-use sector due to a lack of specific
consumption data for the individual end-use sectors within U.S. Territories. See Box 2-1 for more information on
how economic sectors are defined. For more information on trends in the Land Use, Land Use Change, and
Forestry sector, see Section 2.1.
Using this categorization, transportation activities, in aggregate, accounted for the largest portion (28.6 percent) of
total U.S. greenhouse gas emissions in 2019. Emissions from electric power accounted for the second largest
portion (25.1 percent), while emissions from industry accounted for the third largest portion (22.9 percent) of total
U.S. greenhouse gas emissions in 2019. Emissions from industry have in general declined over the past decade due
to a number of factors, including structural changes in the U.S. economy (i.e., shifts from a manufacturing-based to
a service-based economy), fuel switching, and efficiency improvements.
The remaining 23.3 percent of U.S. greenhouse gas emissions were contributed by, in order of magnitude, the
agriculture, commercial, and residential sectors, plus emissions from U.S. Territories. Activities related to
agriculture accounted for roughly 10.2 percent of emissions; unlike other economic sectors, agricultural sector
emissions were dominated by N20 emissions from agricultural soil management and CH4 emissions from enteric
fermentation, rather than C02 from fossil fuel combustion. An increasing amount of carbon is stored in agricultural
soils each year, but this C02 sequestration is assigned to the LULUCF sector rather than the agriculture economic
sector. The commercial and residential sectors accounted for roughly 6.9 percent and 5.8 percent of greenhouse
gas emissions, respectively, and U.S. Territories accounted for 0.4 percent of emissions; emissions from these
sectors primarily consisted of C02 emissions from fossil fuel combustion. Carbon dioxide was also emitted and
sequestered (in the form of C) by a variety of activities related to forest management practices, tree planting in
urban areas, the management of agricultural soils, landfilling of yard trimmings, and changes in C stocks in coastal
wetlands. Table 2-10 presents a detailed breakdown of emissions from each of these economic sectors by source
category, as they are defined in this report. Figure 2-18 shows the trend in emissions by sector from 1990 to 2019.
Trends 2-29

-------
Figure 2-18: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors
2,500
Electric Power Industry (Purple)
2,000
CT
Transportation (Green)
LU
1,500
Industry
1,000
Agriculture
Commercial (Orange)
500
Residential (Blue)
0
01
o
o
LO VD IN. 00
o o o o
o o o o

-------
(Industrial)
5.8
6.4
6.4
6.6
6.7
6.8
6.9
0.1%
Abandoned Oil and Gas Wells
6.8
7.2
7.4
7.4
7.2
7.3
6.6
0.1%
Mobile Combustion
4.0
6.2
5.6
5.7
6.0
6.1
6.3
0.1%
Urea Consumption for Non-








Agricultural Purposes
3.8
3.7
4.6
5.1
5.0
5.9
6.2
0.1%
Abandoned Underground Coal








Mines
7.2
6.6
6.4
6.7
6.4
6.2
5.9
0.1%
AdipicAcid Production
15.2
7.1
4.3
7.0
7.4
10.3
5.3
0.1%
Carbon Dioxide Consumption
1.5
1.4
4.9
4.6
4.6
4.1
4.9
0.1%
Electronics Industry
3.6
4.8
5.0
5.0
4.9
5.1
4.6
0.1%
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
0.1%
Stationary Combustion
4.9
4.7
4.2
4.2
4.1
4.1
4.0
0.1%
Other Process Uses of Carbonates
3.1
3.8
6.1
5.5
5.0
3.7
3.7
0.1%
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
0.1%
Aluminum Production
28.3
7.6
4.9
2.7
2.3
3.1
3.6
0.1%
Soda Ash Production
1.4
1.7
1.7
1.7
1.8
1.7
1.8
+%
Ferroalloy Production
2.2
1.4
2.0
1.8
2.0
2.1
1.6
+%
Titanium Dioxide Production
1.2
1.8
1.6
1.7
1.7
1.5
1.5
+%
Caprolactam, Glyoxal, and








Glyoxylic Acid Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
+%
Glass Production
1.5
1.9
1.3
1.2
1.3
1.3
1.3
+%
Zinc Production
0.6
1.0
0.9
0.8
0.9
1.0
1.0
+%
Magnesium Production and








Processing
5.2
2.7
1.1
1.2
1.1
1.1
1.0
+%
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
0.9
0.9
+%
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
+%
Carbide Production and








Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
+%
Agriculture
600.2
629.7
658.5
645.8
646.6
662.0
669.5
10.2%
N20 from Agricultural Soil








Management
315.9
313.4
348.5
330.1
327.6
338.2
344.6
5.3%
Enteric Fermentation
164.7
169.3
166.9
172.2
175.8
178.0
178.6
2.7%
Manure Management
51.1
67.9
75.4
77.7
78.5
81.1
82.0
1.3%
C02 from Fossil Fuel Combustion
43.4
50.8
41.1
40.2
39.8
39.8
39.7
0.6%
Rice Cultivation
16.0
18.0
16.2
15.8
14.9
15.6
15.1
0.2%
Urea Fertilization
2.4
3.5
4.7
4.9
5.1
5.2
5.3
0.1%
Liming
4.7
4.3
3.7
3.1
3.1
2.2
2.4
+%
Mobile Combustion
1.5
1.8
1.2
1.2
1.2
1.2
1.2
+%
Field Burning of Agricultural








Residues
0.5
0.6
0.6
0.6
0.6
0.6
0.6
+%
Stationary Combustion
+
+
+
+
+
+
+
+%
Commercial
429.2
407.9
445.4
430.1
431.9
447.3
455.3
6.9%
C02 from Fossil Fuel Combustion
228.3
227.1
244.6
231.6
232.0
245.7
249.7
3.8%
Landfills (Municipal)
165.7
117.0
96.4
93.1
94.4
97.0
99.4
1.5%
Substitution of Ozone Depleting








Substances
+
22.1
60.8
61.5
61.0
60.8
62.3
0.9%
Wastewater Treatment








(Domestic)
33.0
36.6
37.8
38.0
38.2
37.8
37.9
0.6%
Composting
0.7
3.5
4.0
4.3
4.6
4.3
4.3
0.1%
Stationary Combustion
1.5
1.4
1.6
1.5
1.5
1.6
1.6
+%
Anaerobic Digestion at Biogas








Facilities
+
0.1
0.2
0.2
0.2
0.2
0.2
+%
Residential
345.1
371.0
351.5
327.8
329.9
377.3
379.5
5.8%
Trends 2-31

-------
Substitution of Ozone Depleting
Substances
0.2

7.2

28.9
30.4
32.0
33.8
37.2
0.6%
Stationary Combustion
6.3

4.9

5.3
4.7
4.5
5.4
5.5
0.1%
U.S. Territories
25.2

63.7

30.0
26.8
25.4
25.4
25.4
0.4%
C02 from Fossil Fuel Combustion
21.7

55.9

29.2
26.0
24.6
24.6
24.6
0.4%
Non-Energy Use of Fuels
3.4

7.6

0.7
0.7
0.7
0.7
0.7
+%
Stationary Combustion
0.1

0.2

0.1
0.1
0.1
0.1
0.1
+%
Total Emissions (Sources)
6,442.7

7,423.0

6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
100.0%
LULUCF Sector Net Totalb
(900.8)

788.1)

(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
(12.0%)
Net Emissions (Sources and Sinks)
5,541.9

6,635.0

5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
88.0%
Notes: Total emissions presented without LULUCF. Total net emissions presented with LULUCF. Totals may not sum due to
independent rounding. Parentheses indicate negative values or sequestration.
+ Does not exceed 0.05 MMT C02 Eq. or 0.05 percent.
a Percent of total (gross) emissions excluding emissions from LULUCF for 2019.
b The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
Box 2-1: Methodology for Aggregating Emissions by Economic Sector
In presenting the Economic Sectors in the annual Inventory of U.S. Greenhouse Gas Emissions and Sinks, the
Inventory expands upon the standard IPCC sectors common for UNFCCC reporting. Discussing greenhouse gas
emissions relevant to U.S.-specific economic sectors improves communication of the report's findings.
The Electric Power economic sector includes C02, CH4 and N20 emissions from the combustion of fossil fuels that
are included in the EIA electric power sector. Carbon dioxide, CH4i and N20 emissions from waste incineration
are included in the Electric Power economic sector, as the majority of municipal solid waste is combusted in
plants that produce electricity. The Electric Power economic sector also includes SF6 from Electrical
Transmission and Distribution, and a portion of C02 from Other Process Uses of Carbonates (from pollution
control equipment installed in electric power plants).
The Transportation economic sector includes C02 emissions from the combustion of fossil fuels that are
included in the EIA transportation fuel-consuming sector. (Additional analyses and refinement of the EIA data
are further explained in the Energy chapter of this report.) Emissions of CH4 and N20 from mobile combustion
are also apportioned to the Transportation economic sector based on the EIA transportation fuel-consuming
sector. Substitution of Ozone Depleting Substances emissions are apportioned to the Transportation economic
sector based on emissions from refrigerated transport and motor vehicle air-conditioning systems. Finally, C02
emissions from Non-Energy Uses of Fossil Fuels identified as lubricants for transportation vehicles are included
in the Transportation economic sector.
The Industry economic sector includes C02 emissions from the combustion of fossil fuels that are included in the
EIA industrial fuel-consuming sector, minus the agricultural use of fuel explained below. The CH4 and N20
emissions from stationary and mobile combustion are also apportioned to the Industry economic sector based
on the EIA industrial fuel-consuming sector, minus emissions apportioned to the Agriculture economic sector.
Substitution of Ozone Depleting Substances emissions are apportioned based on their specific end-uses within
the source category, with most emissions falling within the Industry economic sector. Finally, CH4 emissions
from industrial landfills and CH4 and N20 from industrial wastewater treatment are included in the Industry
economic sector.
Additionally, all process-related emissions from sources with methods considered within the IPCC IPPU sector
are apportioned to the Industry economic sector. This includes the process-related emissions (i.e., emissions
from the actual process to make the material, not from fuels to power the plant) from activities such as Cement
Production, Iron and Steel Production and Metallurgical Coke Production, and Ammonia Production.
Additionally, fugitive emissions from energy production sources, such as Natural Gas Systems, Coal Mining, and
Petroleum Systems are included in the Industry economic sector. A portion of C02 from Other Process Uses of
Carbonates (from pollution control equipment installed in large industrial facilities) is also included in the
Industry economic sector. Finally, all remaining C02 emissions from Non-Energy Uses of Fossil Fuels are assumed
2-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
to be industrial in nature (besides the lubricants for transportation vehicles specified above) and are attributed
to the Industry economic sector.
The Agriculture economic sector includes C02 emissions from the combustion of fossil fuels that are based on
supplementary sources of agriculture fuel use data, because EIA does not include an agriculture fuel-consuming
sector. Agriculture equipment is included in the EIA industrial fuel-consuming sector. Agriculture fuel use
estimates are obtained from U.S. Department of Agriculture survey data, in combination with separate EIA fuel
sales reports (USDA 2019; EIA 2020a). These supplementary data are subtracted from the industrial fuel use
reported by EIA to obtain agriculture fuel use. C02 emissions from fossil fuel combustion, and CH4 and N20
emissions from stationary and mobile combustion, are then apportioned to the Agriculture economic sector
based on agricultural fuel use.
The other IPCC Agriculture emission source categories apportioned to the Agriculture economic sector include
N20 emissions from Agricultural Soils, CH4 from Enteric Fermentation, CH4 and N20 from Manure Management,
CH4 from Rice Cultivation, C02 emissions from Liming and Urea Application, and CH4 and N20 from Field Burning
of Agricultural Residues.
The Residential economic sector includes C02 emissions from the combustion of fossil fuels that are included in
the EIA residential fuel-consuming sector. Stationary combustion emissions of CH4 and N20 are also based on
the EIA residential fuel-consuming sector. Substitution of Ozone Depleting Substances are apportioned to the
Residential economic sector based on emissions from residential air-conditioning systems. Nitrous oxide
emissions from the application of fertilizers to developed land (termed "settlements" by the IPCC) are also
included in the Residential economic sector.
The Commercial economic sector includes C02 emissions from the combustion of fossil fuels that are included in
the EIA commercial fuel-consuming sector. Emissions of CH4 and N20 from Mobile Combustion are also
apportioned to the Commercial economic sector based on the EIA commercial fuel-consuming sector.
Substitution of Ozone Depleting Substances emissions are apportioned to the Commercial economic sector
based on emissions from commercial refrigeration/air-conditioning systems. Public works sources, including
direct CH4 from municipal landfills, CH4 from anaerobic digestion at biogas facilities, CH4 and N20 from domestic
wastewater treatment, and composting, are also included in the Commercial economic sector.
Emissions with Electricity Distributed to Economic Sectors
It is also useful to view greenhouse gas emissions from economic sectors with emissions related to electric power
distributed into end-use categories (i.e., emissions from electric power are allocated to the economic sectors in
which the electricity is used).
The generation, transmission, and distribution of electricity accounted for 25.1 percent of total U.S. greenhouse
gas emissions in 2019. Electric power-related emissions decreased by 12.1 percent since 1990 and by 8.3 percent
from 2018 to 2019, primarily due to a significantly colder winter and a hotter summer in 2019 compared to 2018,
which increased the amount of energy required for heating and cooling. Between 2018 to 2019, the consumption
of natural gas for electric power generation increased by 6.7, while the consumption of coal and petroleum
decreased by 15.5 and 27.6 percent, respectively, reflecting a continued shift from coal to natural gas for electricity
generation.
From 2018 to 2019, electricity sales to the residential and commercial end-use sectors decreased by 2.0 percent
and 1.5 percent, respectively. Electricity sales to the industrial sector increased by approximately 0.2 percent.
Overall, from 2018 to 2019, the amount of electricity retail sales (in kWh) decreased by 1.2 percent. Table 2-11
provides a detailed summary of emissions from electric power-related activities.
Trends 2-33

-------
Table 2-11: Electric Power-Related Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Fuel Type or Source
1990
2005
2015
2016
2017
2018
2019
co2
1,831.2
2,416.6
1,918.3
1,825.9
1,748.5
1,768.2
1,621.2
Fossil Fuel Combustion
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Coal
1,546.5
1,982.8
1,351.4
1,242.0
1,207.1
1,152.9
973.5
Natural Gas
175.4
318.9
525.2
545.0
505.6
577.4
616.0
Petroleum
97.5
98.0
23.7
21.5
18.9
22.2
16.2
Geo thermal
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Incineration of Waste
8.1
12.7
11.5
11.5
11.5
11.5
11.5
Other Process Uses of







Carbonates
3.1
3.8
6.1
5.5
5.0
3.7
3.7
ch4
0.4
0.9
1.2
1.2
1.1
1.2
1.3
Stationary Sources3
0.4
0.9
1.2
1.2
1.1
1.2
1.3
Incineration of Waste
+
+
+
+
+
+
+
n2o
21.0
30.4
26.8
26.5
25.1
24.7
21.4
Stationary Sources3
20.5
30.1
26.5
26.2
24.8
24.4
21.1
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
sf6
23.2
8.4
3.8
4.1
4.2
3.9
4.2
Electrical Transmission and







Distribution
23.2
8.4
3.8
4.1
4.2
3.9
4.2
Total
1,875.7
2,456.3
1,950.0
1,857.6
1,778.9
1,798.0
1,648.1
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Includes only stationary combustion emissions related to the generation of electricity.
To distribute electricity emissions among economic end-use sectors, emissions from the source categories
assigned to the electric power sector were allocated to the residential, commercial, industry, transportation, and
agriculture economic sectors according to each economic sector's share of retail sales of electricity (EIA 2020b;
Duffield 2006). These source categories include C02 from Fossil Fuel Combustion, CH4 and N20 from Stationary
Combustion, Incineration of Waste, Other Process Uses of Carbonates, and SF6 from Electrical Transmission and
Distribution Systems. Note that only 50 percent of the Other Process Uses of Carbonates emissions were
associated with electric power and distributed as described; the remainder of Other Process Uses of Carbonates
emissions were attributed to the industrial processes economic end-use sector.6
When emissions from electricity use are distributed among these economic end-use sectors, industrial activities
account for the largest share of total U.S. greenhouse gas emissions (29.7 percent), followed closely by emissions
from transportation (28.7 percent). Emissions from the commercial and residential sectors also increase
substantially when emissions from electricity are included (15.6 and 14.9 percent, respectively). In all economic
end-use sectors except agriculture, C02 accounts for more than 79.0 percent of greenhouse gas emissions,
primarily from the combustion of fossil fuels. Table 2-12 presents a detailed breakdown of emissions from each of
these economic sectors, with emissions from electric power distributed to them. Figure 2-19 shows the trend in
these emissions by sector from 1990 to 2019.
6 Emissions were not distributed to U.S. Territories, since the electric power sector only includes emissions related to the
generation of electricity in the 50 states and the District of Columbia.
2-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 2-19: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed
to Economic Sectors
2,500
Industry
2,000
Transportation
1,500
Commercial (Orange)
Residential (Blue)
Agriculture
500
rsjro^-m^rvcocTiOT-iojro^-Ln^rvco
c7itT»CT>CT»CT>criooooooooo
rMro^Lnvx)rvQOcr>
1—I 1—I 1—I T—I 1—I 1—I 1—I •»—I
1—1
1—1
Note: Emissions and removals from Land Use, Land Use Change, and Forestry are excluded from figure above. Excludes U.S.
Territories.
Table 2-12: U.S. Greenhouse Gas Emissions by Economic Sector and Gas with Electricity-
Related Emissions Distributed (MMT CO2 Eq.) and Percent of Total in 2019
Sector/Gas
1990

2005

2015
2016
2017
2018
2019
Percent3
Industry
2,313.1

2,234.1

1,964.2
1,894.6
1,902.7
1,958.3
1,947.2
29.7%
Direct Emissions
1,640.7

1,518.8

1,441.6
1,402.2
1,423.4
1,483.3
1,504.8
22.9%
C02
1,158.9

1,140.9

1,081.9
1,052.8
1,068.6
1,125.2
1,149.4
17.5%
ch4
365.2

304.4

289.7
278.1
280.2
280.0
280.7
4.3%
n2o
40.3

33.8

30.2
31.5
30.9
34.2
29.6
0.5%
HFCs, PFCs,SF6,and NF3
76.3

39.6

39.8
39.9
43.6
43.9
45.2
0.7%
Electricity-Related
672.4

715.3

522.6
492.4
479.3
475.0
442.4
6.7%
C02
656.4

703.7

514.1
484.0
471.1
467.2
435.2
6.6%
ch4
0.2

0.3

0.3
0.3
0.3
0.3
0.3
+%
n2o
7.5

8.9

7.2
7.0
6.8
6.5
5.7
0.1%
sf6
8.3

2.4

1.0
1.1
1.1
1.0
1.1
+%
Transportation
1,529.8

1,980.4

1,798.4
1,834.3
1,851.8
1,883.0
1,880.6
28.7%
Direct Emissions
1,526.6

1,975.6

1,794.1
1,830.0
1,847.3
1,878.2
1,875.7
28.6%
C02
1,481.0

1,868.8

1,730.2
1,770.2
1,792.0
1,825.8
1,826.1
27.8%
ch4
5.7

3.0

1.8
1.7
1.6
1.5
1.4
+%
n2o
39.9

34.5

15.8
14.8
13.6
12.4
11.5
0.2%
HFCsb
+

69.3

46.3
43.3
40.1
38.5
36.7
0.6%
Electricity-Related
3.1

4.8

4.4
4.3
4.4
4.8
4.9
0.1%
C02
3.1

4.8

4.3
4.2
4.4
4.7
4.8
0.1%
ch4
+

+

+
+
+
+
+
+%
n2o
+

0.1

0.1
0.1
0.1
0.1
0.1
+%
sf6
+

+

+
+
+
+
+
+%
Commercial
983.4

1,229.8

1,125.7
1,080.8
1,054.5
1,067.8
1,022.3
15.6%
Direct Emissions
429.2

407.9

445.4
430.1
431.9
447.3
455.3
6.9%
C02
228.3

227.1

244.6
231.6
232.0
245.7
249.7
3.8%
ch4
181.9

134.2

112.9
109.3
110.5
112.9
115.2
1.8%
n2o
19.0

24.5

27.1
27.7
28.4
27.9
28.2
0.4%
HFCs
+

22.1

60.8
61.5
61.0
60.8
62.3
0.9%
Trends 2-35

-------
Electricity-Related
554.2
821.8
680.3
650.7
622.6
620.4
566.9
8.6%
C02
541.0
808.5
669.2
639.6
612.0
610.1
557.7
8.5%
ch4
0.1
0.3
0.4
0.4
0.4
0.4
0.4
+%
n2o
6.2
10.2
9.4
9.3
00
00
8.5
7.4
0.1%
sf6
6.8
2.8
1.3
1.4
1.5
1.3
1.5
+%
Residential
956.0
1,247.1
1,053.1
998.9
963.7
1,035.9
978.3
14.9%
Direct Emissions
345.1
371.0
351.5
327.8
329.9
377.3
379.5
5.8%
C02
338.6
358.9
317.3
292.8
293.4
338.1
336.8
5.1%
ch4
5.2
4.1
4.5
3.9
3.8
4.5
4.6
0.1%
n2o
1.0
0.9
0.9
0.8
0.7
0.9
0.9
+%
HFCs
0.2
7.2
28.9
30.4
32.0
33.8
37.2
0.6%
Electricity-Related
610.9
876.1
701.6
671.1
633.8
658.7
598.8
9.1%
C02
596.4
861.9
690.1
659.6
623.0
647.7
589.0
9.0%
ch4
0.1
0.3
0.4
0.4
0.4
0.4
0.5
+%
n2o
6.8
10.9
9.7
9.6
8.9
9.1
7.8
0.1%
sf6
7.5
3.0
1.4
1.5
1.5
1.4
1.5
+%
Agriculture
635.3
668.0
699.7
684.9
685.3
701.1
704.6
10.7%
Direct Emissions
600.2
629.7
658.5
645.8
646.6
662.0
669.5
10.2%
C02
50.5
58.7
49.6
48.1
47.9
47.3
47.4
0.7%
ch4
218.3
239.5
241.5
248.2
251.1
255.8
256.5
3.9%
n2o
331.4
331.5
367.3
349.5
347.5
358.9
365.5
5.6%
Electricity-Related
35.1
38.3
41.3
39.1
38.7
39.1
35.2
0.5%
C02
34.2
37.7
40.6
38.5
38.1
38.5
34.6
0.5%
ch4
+
+
+
+
+
+
+
+%
n2o
0.4
0.5
0.6
0.6
0.5
0.5
0.5
+%
sf6
0.4
0.1
0.1
0.1
0.1
0.1
0.1
+%
U.S. Territories
25.2
63.7
30.0
26.8
25.4
25.4
25.4
0.4%
Total Emissions (Sources)
6,442.7
7,423.0
6,671.1
6,520.3
6,483.3
6,671.4
6,558.3
100.0%
LULUCF Sector Net Totalc
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
(+)%
Net Emissions (Sources and








Sinks)
5,541.9
6,635.0
5,907.3
5,677.5
5,717.2
5,870.0
5,769.1
88.0%
Notes: Total emissions presented without LULUCF. Net emissions presented with LULUCF. Emissions from electric power are
allocated based on aggregate electricity use in each end-use sector. Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq. or 0.05 percent.
a Percent of total (gross) emissions excluding emissions from LULUCF for year 2019.
b Includes primarily HFC-134a.
c The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
Industry
The industry end-use sector includes C02 emissions from fossil fuel combustion from all manufacturing facilities, in
aggregate, and with the distribution of electricity-related emissions, accounts for 29.7 percent of U.S. greenhouse
gas emissions in 2019. This end-use sector also includes emissions that are produced as a byproduct of the non-
energy-related industrial process activities. The variety of activities producing these non-energy-related emissions
includes CH4 emissions from petroleum and natural gas systems, fugitive CH4 emissions from coal mining,
byproduct C02 emissions from cement manufacture, and HFC, PFC, SF6, and NF3 byproduct emissions from the
electronics industry, to name a few.
Since 1990, industrial sector emissions have declined by 15.8 percent. The decline has occurred both in direct
emissions and indirect emissions associated with electricity use. Structural changes within the U.S. economy
thatled to shifts in industrial output away from energy-intensive manufacturing products to less energy-intensive
products (e.g., from steel to computer equipment) have had a significant effect on industrial emissions.
2-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Transportation
When electricity-related emissions are distributed to economic end-use sectors, transportation activities
accounted for 28.7 percent of U.S. greenhouse gas emissions in 2019. The largest sources of transportation
greenhouse gas emissions in 2019 were passenger cars (40.5 percent); freight trucks (23.6 percent); light-duty
trucks, which include sport utility vehicles, pickup trucks, and minivans (17.2 percent); commercial aircraft (7.2
percent); pipelines (2.9 percent); other aircraft (2.4 percent); rail (2.2 percent); and ships and boats (2.1 percent).
These figures include direct C02, CH4, and N20 emissions from fossil fuel combustion used in transportation,
indirect emissions from electricity use and emissions from non-energy use (i.e., lubricants) used in transportation,
as well as HFC emissions from mobile air conditioners and refrigerated transport allocated to these vehicle types.
In terms of the overall trend, from 1990 to 2019, total transportation emissions increased due, in large part, to
increased demand for travel. The number of VMT by light-duty motor vehicles (passenger cars and light-duty
trucks) increased 47.5 percent from 1990 to 2019, as a result of a confluence of factors including population
growth, economic growth, urban sprawl, and periods of low fuel prices.
The decline in new light-duty vehicle fuel economy between 1990 and 2004 reflected the increasing market share
of light-duty trucks, which grew from about 30 percent of new vehicle sales in 1990 to 48 percent in 2004. Starting
in 2005, average new vehicle fuel economy began to increase while light-duty VMT grew only modestly for much
of the period. Light-duty VMT grew by less than one percent or declined each year between 2005 and 2013,7 then
grew at a faster rate until 2016 (2.6 percent from 2014 to 2015, and 2.5 percent from 2015 to 2016). Since 2016,
the rate of light-duty VMT growth has slowed to less than one percent each year. Average new vehicle fuel
economy has increased almost every year since 2005, while light-duty truck market share decreased to about 33
percent in 2009 and has since varied from year to year between 36 and 56 percent. Light-duty truck market share
was about 56 percent of new vehicles in model year 2019 (EPA 2020a).
Table 2-13 provides a detailed summary of greenhouse gas emissions from transportation-related activities with
electricity-related emissions included in the totals. Historically, the majority of electricity use in the transportation
sector was for rail transport. However, more recently there has been increased electricity use in on-road electric
and plug-in hybrid vehicles. For a more detailed breakout of emissions by fuel type by vehicle see Table A-104 in
Annex 3.
Almost all of the energy used for transportation was supplied by petroleum-based products, with more than half
being related to gasoline consumption in automobiles and other highway vehicles. Other fuel uses, especially
diesel fuel for freight trucks and jet fuel for aircraft, accounted for the remainder. The primary driver of
transportation-related emissions was C02 from fossil fuel combustion, which increased by 24 percent from 1990 to
2019.8 This rise in C02 emissions, combined with an increase in HFCs from close to zero emissions in 1990 to 36.7
MMT C02 Eq. in 2019, led to an increase in overall greenhouse gas emissions from transportation activities of 23
percent.9
7	VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2018). In 2007 and 2008
light-duty VMT decreased 3.0 percent and 2.3 percent, respectively. Note that the decline in light-duty VMT from 2006 to 2007
is due at least in part to a change in FHWA's methods for estimating VMT. In 2011, FHWA changed its methods for estimating
VMT by vehicle class, which led to a shift in VMT and emissions among on-road vehicle classes in the 2007 to 2018 time period.
In absence of these method changes, light-duty VMT growth between 2006 and 2007 would likely have been higher.
8	See previous footnote.
9	See previous footnote.
Trends 2-37

-------
Figure 2-20: Trends in Transportation-Related Greenhouse Gas Emissions10
2,200
2,000


o
oi
00
VO
1—1
rM



i—i
1—1
1—1

o
i—i

hs
r*s
10



1—1
l—i
i—l


¦
| 1,000
800
600
400
200
Motorcycles
Buses
Pipelines
Rail
Ships and Boats
Aircraft
Medium- and Heavy-Duty Trucks
Light-Duty Trucks
I Passenger Cars
c>i-ifMm^Ln^or*«.cocnoi-ir\jfo^a-Lnkorvcocnoi-i
cricric^cncnaicncncncnooooooooooT-i*-i
cricric^cncnCT^cncncncriooooooooooooooa
vD


vO


00
CM
kD


-------
co2
8.4
11.8
18.9
18.4
19.9
21.2
21.4
ch4
+
0.2
0.2
0.2
0.2
0.2
0.2
n2o
+
+
0.1
0.1
0.1
0.1
0.1
HFCs
0.0
0.3
0.4
0.4
0.4
0.4
0.4
Motorcycles
1.7
1.6
3.7
3.9
3.8
3.8
3.6
C02
1.7
1.6
3.6
3.8
3.7
3.8
3.6
ch4
+
+
+
+
+
+
+
n2o
+
+
+
+
+
+
+
Commercial Aircraft3
110.9
134.0
120.1
121.5
129.2
130.8
135.4
C02
109.9
132.7
119.0
120.4
128.0
129.6
134.2
ch4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
n2o
1.0
1.2
1.1
1.1
1.2
1.2
1.2
Other Aircraftb
78.3
59.7
40.4
47.5
45.6
44.7
45.7
co2
77.5
59.1
40.0
47.0
45.2
44.3
45.2
ch4
0.1
0.1
+
+
+
+
+
n2o
0.7
0.5
0.4
0.4
0.4
0.4
0.4
Ships and Boatsc
47.0
45.4
33.8
40.8
43.9
41.2
40.4
co2
46.3
44.3
30.6
37.2
40.0
37.0
35.9
ch4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
n2o
0.3
0.3
0.2
0.2
0.2
0.2
0.2
HFCs
0.0
0.5
2.6
2.9
3.3
3.6
3.9
Rail
39.0
51.5
44.1
40.3
41.5
43.3
40.8
C02
38.5
50.8
43.5
39.7
40.9
42.7
40.2
ch4
0.1
0.1
0.1
0.1
0.1
0.1
0.1
n2o
0.3
0.4
0.4
0.3
0.4
0.4
0.3
HFCs
0.0
0.1
0.1
0.1
0.1
0.1
0.1
Other Emissions from







Electric Powerd
0.1
+
+
+
0.1
0.1
0.1
Pipelines8
36.0
32.4
38.5
39.2
41.3
49.9
53.7
C02
36.0
32.4
38.5
39.2
41.3
49.9
53.7
Total Transportation
1,517.9
1,970.2
1,787.5
1,823.9
1,842.2
1,873.8
1,871.7
International Bunker Fuels?
54.8
44.7
31.6
35.0
34.6
32.5
26.4
Ethanol C02a
4.1
21.6
74.2
76.9
77.7
78.6
78.7
Biodiesel C02a
0.0
0.9
14.1
19.6
18.7
17.9
17.1
Notes: Passenger cars and light-duty trucks include vehicles typically used for personal travel and less than
8,500 lbs; medium- and heavy-duty trucks include vehicles larger than 8,500 lbs. HFC emissions primarily
reflect HFC-134a. Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Consists of emissions from jet fuel consumed by domestic operations of commercial aircraft (no bunkers).
b Consists of emissions from jet fuel and aviation gasoline consumption by general aviation and military
aircraft.
c Fluctuations in emission estimates are associated with fluctuations in reported fuel consumption and may
reflect issues with data sources.
d Other emissions from electric power are a result of waste incineration (as the majority of municipal solid
waste is combusted in "trash-to-steam" electric power plants), electrical transmission and distribution, and a
portion of Other Process Uses of Carbonates (from pollution control equipment installed in electric power
plants).
0 C02 estimates reflect natural gas used to power pipelines, but not electricity. While the operation of pipelines
produces CH4 and N20, these emissions are not directly attributed to pipelines in the Inventory.
f Emissions from International Bunker Fuels include emissions from both civilian and military activities; these
emissions are not included in the transportation totals.
g Ethanol and biodiesel C02 estimates are presented for informational purposes only. See Section 3.11 and the
estimates in Land Use, Land-Use Change, and Forestry (see Chapter 6), in line with IPCC methodological
guidance and UNFCCC reporting obligations, for more information on ethanol and biodiesel.
Trends 2-39

-------
Commercial
The commercial end-use sector, with electricity-related emissions distributed, accounts for 15.6 percent of U.S.
greenhouse gas emissions in 2019 and is heavily reliant on electricity for meeting energy needs, with electricity use
for lighting, heating, air conditioning, and operating appliances. The remaining emissions were largely due to the
direct consumption of natural gas and petroleum products, primarily for heating and cooking needs. Energy-
related emissions from the commercial sector have generally been increasing since 1990, and annual variations are
often correlated with short-term fluctuations in energy use caused by weather conditions, rather than prevailing
economic conditions. Decreases in energy-related emissions in the commercial sector in recent years can be
largely attributed to an overall reduction in energy use driven by a reduction in heating degree days and increases
in energy efficiency.
Landfills and wastewater treatment are included in the commercial sector, with landfill emissions decreasing since
1990 and wastewater treatment emissions decreasing slightly.
Residential
The residential end-use sector, with electricity-related emissions distributed, accounts for 14.9 percent of U.S.
greenhouse gas emissions in 2019 and similarly, is heavily reliant on electricity for meeting energy needs, with
electricity use for lighting, heating, air conditioning, and operating appliances. The remaining emissions were
largely due to the direct consumption of natural gas and petroleum products, primarily for heating and cooking
needs. Emissions from the residential sector have generally been increasing since 1990, and annual variations are
often correlated with short-term fluctuations in energy use caused by weather conditions, rather than prevailing
economic conditions. In the long term, the residential sector is also affected by population growth, migration
trends toward warmer areas, and changes in housing and building attributes (e.g., larger sizes and improved
insulation). A shift toward energy-efficient products and more stringent energy efficiency standards for household
equipment has also contributed to recent trends in energy demand in households (EIA 2018).
Agriculture
The agriculture end-use sector accounts for 10.7 percent of U.S. greenhouse gas emissions in 2019 when
electricity-related emissions are distributed, and includes a variety of processes, including enteric fermentation in
domestic livestock, livestock manure management, and agricultural soil management. In 2019, agricultural soil
management was the largest source of N20 emissions, and enteric fermentation was the largest source of CH4
emissions in the United States. This sector also includes small amounts of C02 emissions from fossil fuel
combustion by motorized farm equipment such as tractors.
Box 2-2: Trends in Various U.S. Greenhouse Gas Emissions-Related Data
Total greenhouse gas emissions can be compared to other economic and social indices to highlight changes over
time. These comparisons include: (1) emissions per unit of aggregate energy use, because energy-related
activities are the largest sources of emissions; (2) emissions per unit of fossil fuel consumption, because almost
all energy-related emissions involve the combustion of fossil fuels; (3) emissions per unit of total gross domestic
product as a measure of national economic activity; and (4) emissions per capita.
Table 2-14 provides data on various statistics related to U.S. greenhouse gas emissions normalized to 1990 as a
baseline year. These values represent the relative change in each statistic since 1990. Greenhouse gas emissions
in the United States have grown at an average annual rate of 0.1 percent since 1990, although changes from
year to year have been significantly larger. This growth rate is slightly slower than that for total energy use,
overall gross domestic product (GDP) and national population (see Table 2-14 and Figure 2-21). The direction of
these trends started to change after 2005, when greenhouse gas emissions, total energy use and associated
fossil fuel consumption began to peak. Greenhouse gas emissions in the United States have decreased at an
average annual rate of 0.8 percent since 2005. Fossil fuel consumption has also decreased at a slower rate than
2-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
emissions since 2005, while total energy use, GDP, and national population continued to increase.
Table 2-14: Recent Trends in Various U.S. Data (Index 1990 = 100)	
Avg. Annual Avg. Annual
Variable
1990

2005

2015
2016
2017
2018
2019
Change
Since 1990a
Change
Since 2005a
Greenhouse Gas Emissions'5
100

115

104
101
101
104
102
0.1%
-0.8%
Energy Usec
100

119

116
116
116
120
119
0.6%
+%
GDPd
100

159

186
189
194
200
204
2.5%
1.8%
Population6
100

118

128
129
130
131
132
1.0%
0.8%
+ Does not exceed 0.05 percent.
a Average annual growth rate.
b GWP-weighted values.
c Energy-content-weighted values (EIA 2020b).
d GDP in chained 2009 dollars (BEA 2020).
6 U.S. Census Bureau (2020).
Figure 2-21: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic
Product
Real GDP
200
180
160
140
Population
Energy Use
120
Emissions
Energy Use Per Capita
Emissions per Capita
x 100
Emissions per GDP
Source: BEA (2020), U.S. Census Bureau (2020), and emission estimates in this report.
2.3 Precursor Greenhouse Gas Emissions (CO,
NOx, NMVOCs, and S02)	
The reporting requirements of the UNFCCC11 request that information be provided on precursor greenhouse
gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-CH4 volatile organic compounds
11 See .
Trends 2-41

-------
(NMVOCs), and sulfur dioxide (S02). These gases are not direct greenhouse gases, but indirectly affect terrestrial
radiation absorption by influencing the formation and destruction of tropospheric and stratospheric ozone, or, in
the case of S02, by affecting the absorptive characteristics of the atmosphere. Additionally, some of these gases
may react with other chemical compounds in the atmosphere to form compounds that are greenhouse gases.
Carbon monoxide is produced when carbon-containing fuels are combusted incompletely. Nitrogen oxides (i.e., NO
and N02) are created by lightning, fires, fossil fuel combustion, and in the stratosphere from N20. Non-methane
volatile organic compounds—which include hundreds of organic compounds that participate in atmospheric
chemical reactions (i.e., propane, butane, xylene, toluene, ethane, and many others)—are emitted primarily from
transportation, industrial processes, and non-industrial consumption of organic solvents. In the United States, S02
is primarily emitted from coal combustion for electric power generation and the metals industry. Sulfur-containing
compounds emitted into the atmosphere tend to exert a negative radiative forcing (i.e., cooling) and therefore are
discussed separately.
One important indirect climate change effect of NMVOCs and NOx is their role as precursors for tropospheric
ozone formation. They can also alter the atmospheric lifetimes of other greenhouse gases. Another example of
indirect greenhouse gas formation into greenhouse gases is the interaction of CO with the hydroxyl radical—the
major atmospheric sink for CH4 emissions—to form C02. Therefore, increased atmospheric concentrations of CO
limit the number of hydroxyl molecules (OH) available to destroy CH4.
Since 1970, the United States has published estimates of emissions of CO, NOx, NMVOCs, and S02 (EPA 2020b),12
which are regulated under the Clean Air Act. Table 2-15 shows that fuel combustion accounts for the majority of
emissions of these indirect greenhouse gases. Industrial processes—such as the manufacture of chemical and
allied products, metals processing, and industrial uses of solvents—are also significant sources of CO, NOx, and
NMVOCs.
Table 2-15: Emissions of NOx, CO, NMVOCs, and SO2 (kt)
Gas/Activity
1990
2005
2015
2016
2017
2018
2019
NOx
21,739
17,339
10,187
8,792
8,642
8,145
7,754
Mobile Fossil Fuel Combustion
10,862
10,295
5,634
4,739
4,563
4,123
3,862
Stationary Fossil Fuel Combustion
10,023
5,858
3,084
2,856
2,728
2,711
2,581
Oil and Gas Activities
139
321
622
594
565
565
565
Industrial Processes and Product Use
592
572
408
402
397
397
397
Forest Fires
22
126
312
87
281
242
242
Waste Combustion
82
128
88
80
71
71
71
Grassland Fires
5
21
21
19
21
20
20
Agricultural Burning
13
15
14
14
14
14
14
Waste
+
2
2
1
1
1
1
CO
130,969
71,781
51,525
39,287
45,314
42,355
41,524
Mobile Fossil Fuel Combustion
119,360
58,615
32,635
28,789
28,124
26,590
25,749
Forest Fires
800
4,511
11,136
3,080
10,036
8,626
8,626
Stationary Fossil Fuel Combustion
5,000
4,648
3,688
3,690
3,692
3,692
3,692
Waste Combustion
978
1,403
1,576
1,375
1,175
1,175
1,175
Industrial Processes and Product Use
4,129
1,557
1,163
1,075
1,006
1,006
1,006
Oil and Gas Activities
302
318
622
607
592
592
592
Grassland Fires
84
358
356
324
345
331
341
Agricultural Burning
315
363
342
340
339
338
337
Waste
1
7
7
6
5
5
5
NMVOCs
20,930
13,154
10,596
9,774
9,444
9,228
9,123
Industrial Processes and Product Use
7,638
5,849
3,796
3,776
3,767
3,767
3,767
Mobile Fossil Fuel Combustion
10,932
5,724
3,458
2,873
2,758
2,543
2,437
12 NOx and CO emission estimates from Field Burning of Agricultural Residues were estimated separately, and therefore not
taken from EPA (2019b).
2-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Oil and Gas Activities
554

510

2,656
2,459
2,262
2,262
2,262
Stationary Fossil Fuel Combustion
912

716

493
489
496
496
496
Waste Combustion
222

241

132
121
109
109
109
Waste
673

114

63
57
52
52
52
Agricultural Burning
NA

NA

NA
NA
NA
NA
NA
so2
20,935

13,196

3,578
2,906
2,313
2,233
1,966
Stationary Fossil Fuel Combustion
18,407

11,541

2,901
2,269
1,638
1,569
1,304
Industrial Processes and Product Use
1,307

831

482
466
509
509
509
Oil and Gas Activities
390

180

92
89
86
86
86
Mobile Fossil Fuel Combustion
793

619

78
57
58
47
45
Waste Combustion
38

25

26
24
22
22
22
Waste
+

1

1
1
1
1
1
Agricultural Burning
NA

NA

NA
NA
NA
NA
NA
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
NA (Not Available)
Source: (EPA 2020b) except for estimates from Forest Fires, Grassland Fires, and Field Burning of Agricultural
Residues.
Box 2-3: Sources and Effects of Sulfur Dioxide
Sulfur dioxide (S02) emitted into the atmosphere through natural and anthropogenic processes affects the
earth's radiative budget through its photochemical transformation into sulfate aerosols that can:
(1)	scatter radiation from the sun back to space, thereby reducing the radiation reaching the earth's
surface;
(2)	affect cloud formation; and
(3)	affect atmospheric chemical composition (e.g., by providing surfaces for heterogeneous chemical
reactions).
The indirect effect of sulfur-derived aerosols on radiative forcing can be considered in two parts. The first
indirect effect is the aerosols' tendency to decrease water droplet size and increase water droplet concentration
in the atmosphere. The second indirect effect is the tendency of the reduction in cloud droplet size to affect
precipitation by increasing cloud lifetime and thickness. Although still highly uncertain, the radiative forcing
estimates from both the first and the second indirect effect are believed to be negative, as is the combined
radiative forcing of the two (IPCC 2013).
Sulfur dioxide is also a major contributor to the formation of regional haze, which can cause significant increases
in acute and chronic respiratory diseases. Once S02 is emitted, it is chemically transformed in the atmosphere
and returns to the earth as the primary source of acid rain. Because of these harmful effects, the United States
has regulated S02 emissions in the Clean Air Act.
Electric power is the largest anthropogenic source of S02 emissions in the United States, accounting for 46.9
percent in 2019. Coal combustion contributes nearly all of those emissions (approximately 92 percent). Sulfur
dioxide emissions have decreased in recent years, primarily as a result of electric power generators switching
from high-sulfur to low-sulfur coal and installing flue gas desulfurization equipment.
Trends 2-43

-------
3. Energy
Energy-related activities were the primary sources of U.S. anthropogenic greenhouse gas emissions, accounting for
82.2 percent of total greenhouse gas emissions on a carbon dioxide (C02) equivalent basis in 2019.1 This included
96.7, 40.6, and 9.5 percent of the nation's C02, methane (CH4), and nitrous oxide (N20) emissions, respectively.
Energy-related C02 emissions alone constituted 77.5 percent of U.S. greenhouse gas emissions from all sources on
a C02-equivalent basis, while the non-C02 emissions from energy-related activities represented a much smaller
portion of total national emissions (4.7 percent collectively).
Emissions from fossil fuel combustion comprise the vast majority of energy-related emissions, with C02 being the
primary gas emitted (see Figure 3-1 and Figure 3-2). Globally, approximately 33,300 million metric tons (MMT) of
C02 were added to the atmosphere through the combustion of fossil fuels in 2019, of which the United States
accounted for approximately 15 percent.2 Due to their relative importance over time (see Figure 3-2), fossil fuel
combustion-related C02 emissions are considered separately and in more detail than other energy-related
emissions (see Figure 3-3).
Fossil fuel combustion also emits CH4 and N20. Stationary combustion of fossil fuels was the third largest source of
N20 emissions in the United States and mobile fossil fuel combustion was the fifth largest source. Energy-related
activities other than fuel combustion, such as the production, transmission, storage, and distribution of fossil fuels,
also emit greenhouse gases. These emissions consist primarily of fugitive CH4 emissions from natural gas systems,
coal mining, and petroleum systems.
1	Estimates are presented in units of million metric tons of carbon dioxide equivalent (MMT C02 Eq.), which weight each gas by
its global warming potential, or GWP, value. See section on global warming potentials in the Executive Summary.
2	Global C02 emissions from fossil fuel combustion were taken from International Energy Agency Energy related C02 emissions,
1990-2019 - Charts Available at: 
(IEA 2020).
Energy 3-1

-------
Figure 3-1: 2019 Energy Chapter Greenhouse Gas Sources
CO2 Emissions from Fossil Fuel Combustion
Natural Gas Systems
Non-Energy Use of Fuels
Petroleum Systems
Coal Mining
Non-C02 Emissions from Stationary Combustion
Non-C02 Emissions from Mobile Combustion
Incineration of Waste
Abandoned Oil and Gas Wells
Abandoned Underground Coal Mines
Energy as a Portion of All
Emissions
100	150	200
MMT CO2 Eq.
Figure 3-2: Trends in Energy Chapter Greenhouse Gas Sources
7,000
6'000 ¦» M ES 5 Sis
^ LT> Ł!
—I O OO
" m SI
„ o s 8 s S s. K. a s
J"1 2 r*» ,>0	( vd us «o	cl
t	10 ^ ^	10 j § a
5,000
iff 4,000
IN
O
u
Z 3,000
2,000
1,000
Incineration of Waste
U.S Territories Fossil Fuel Combustion
Non-Energy Use of Fuels
Commerical Fossil Fuel Combustion
I Fugitive Emissions
Residential Fossil Fuel Combustion
Industrial Fossil Fuel Combustion
I Transportation Fossil Fuel Combustion
Electric Power Fossil Fuel Combustion
i-HrMro^-uiiorvooo^
CT^ CT^	CT» 0*1
cn  CTi CX> CT> CT>	C>
" I 1 ! I 3 y -

m	[A w d
- ^ n in ^
lo	¦ Ln i m in ' i.
3-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 3-3: 2019 U.S. Fossil Carbon Flows
Internationa!
bunxers
11S
industrial
Processes
NEU Emissions 13
Fossil Fuel
Energy Exports
~-t NEU
i ar*
Balancing
Hem
Combuston Emissions
K645
Stock
Changes
^ 101
Natural Gas
Emissions
1,658
•NEU Emissions 11
Combustion Emissions
1,027
Natural Gas
1,842
*"Coal
Emissions 1,038
Atmospheric
Emissions
6,181
Apparent
Consumption
5,424
Combustion Emissions
2.185
Petroleum
1,888
Hyrdocarbon>
liquids and
Other Liquids
506
Fossil Fuel
Energy
Imports
1,620
Petroleum
1,215 -
NEU U S
Territories
Natural Gas 149-
Coal 16-
Note. Totals may not sum due to independent founding.
The "Balancing Item" accounts for the statistical imbalances
and unknowns in the reported data sets combined here
NEU - Non-Energy Use
W Fossil Fuei
NEU imports Consumption
22	U.S.
Territories
Other 241
Table 3-1 summarizes emissions from the Energy sector in units of MMT C02 Eq., while unweighted gas emissions
in kilotons (kt) are provided in Table 3-2. Overall, emissions due to energy-related activities were 5,392.3 MMT C02
Eq. in 2019,3 an increase of 1.3 percent since 1990 and a decrease of 2.3 percent since 2018.
Table 3-1: CO2, ChU, and N2O Emissions from Energy (MMT CO2 Eq.)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
C02
4,894.1
5,932.6
5,189.8
5,074.8
5,035.7
5,203.7
5,081.4
Fossil Fuel Combustion
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Transportation
1,469.1
1,858.6
1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
Electric Power
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Industrial
853.8
852.9
797.3
792.5
790.1
813.6
822.5
Residential
338.6
358.9
317.3
292.8
293.4
338.1
336.8
Commercial
228.3
227.1
244.6
231.6
232.0
245.7
249.7
U.S. Territories
21.7
55.9
29.2
26.0
24.6
24.6
24.6
Non-Energy Use of Fuels
112.8
129.1
108.5
99.8
113.5
129.7
128.8
Petroleum Systems
9.7
12.1
32.4
21.8
25.0
37.1
47.3
Natural Gas Systems
32.0
25.2
29.1
30.1
31.2
33.9
37.2
Incineration of Waste
8.1
12.7
11.5
11.5
11.5
11.5
11.5
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Biomass-Wooda
215.2
206.9
224.7
215.7
211.5
219.8
216.5
International Bunker Fuelsb
103.5
113.2
110.9
116.6
120.1
122.1
116.1
Biofuels-Ethanola
4.2
22.9
78.9
81.2
82.1
81.9
82.6
Biofuels-BiodieseP
0.0
0.9
14.1
19.6
18.7
17.9
17.1
ch4
361.3
293.3
277A
264.9
266.6
267.0
267.6
Natural Gas Systems
186.9
164.2
149.8
147.3
148.7
152.5
157.6
Coal Mining
96.5
64.1
61.2
53.8
54.8
52.7
47.4
Petroleum Systems
48.9
39.5
41.5
39.2
39.3
37.3
39.1
Stationary Combustion
8.6
7.8
8.5
7.9
7.6
8.5
8.7
Abandoned Oil and Gas Wells
6.8
7.2
7.4
7.4
7.2
7.3
6.6
3 Following the current reporting requirements under the UNFCCC, this Inventory report presents C02 equivalent values based
on the IPCC Fourth Assessment Report (AR4) GWP values. See the Introduction chapter for more information.
Energy 3-3

-------
Abandoned Underground Coal
Mines
7.2
6.6
6.4
6.7
6.4
6.2
5.9
Mobile Combustion
6.4
4.0
2.6
2.5
2.5
2.4
2.4
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
n2o
70.3
76.3
52.6
51.2
48.6
47.4
43.2
Stationary Combustion
25.1
34.4
30.5
30.0
28.4
28.2
24.9
Mobile Combustion
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
1.0
1.0
1.1
1.1
1.0
Total
5,325.6
6,302.3
5,519.8
5,390.9
5,351.0
5,518.1
5,392.3
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Emissions from Wood Biomass, Ethanol, and Biodiesel Consumption are not included specifically in summing Energy
sector totals. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for LULUCF.
b Emissions from International Bunker Fuels are not included in totals. These values are presented for informational
purposes only, in line with the 2006IPCC Guidelines and UNFCCC reporting obligations.
Table 3-2: CO2, ChU, and N2O Emissions from Energy (kt)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
CO?
4,894,051
5,932,600
5,189,826
5,074,805
5,035,743
5,203,702
5,081,445
Fossil Fuel Combustion
4,731,466
5,753,507
5,008,270
4,911,532
4,854,480
4,991,420
4,856,702
Non-Energy Use of







Fuels
112,766
129,135
108,476
99,840
113,539
129,728
128,763
Petroleum Systems
9,709
12,059
32,412
21,847
24,979
37,115
47,269
Natural Gas Systems
32,042
25,179
29,127
30,054
31,200
33,885
37,234
Incineration of Waste
8,062
12,713
11,533
11,525
11,537
11,547
11,471
Abandoned Oil and







Gas Wells
6
7
7
7
7
7
7
Biomass-WoodP
215,186
206,901
224,730
215,712
211,511
219,794
216,533
International Bunker







Fuelsb
103,463
113,232
110,908
116,611
120,121
122,112
116,064
Biofuels-Ethanola
4,227
22,943
78,934
81,250
82,088
81,917
82,578
Biofuels-Biodiesela
0
856
14,077
19,648
18,705
17,936
17,080
ch4
14,451
11,733
11,095
10,596
10,665
10,680
10,704
Natural Gas Systems
7,478
6,567
5,994
5,894
5,949
6,101
6,305
Coal Mining
3,860
2,565
2,449
2,154
2,191
2,109
1,895
Petroleum Systems
1,955
1,579
1,659
1,568
1,574
1,492
1,563
Stationary Combustion
344
313
339
315
306
342
346
Abandoned Oil and







Gas Wells
271
287
294
296
288
290
263
Abandoned







Underground Coal







Mines
288
264
256
268
257
247
237
Mobile Combustion
256
158
105
102
100
98
95
Incineration of Waste
+
+
+
+
+
+
+
International Bunker







Fuelsb
7
5
4
4
4
4
4
n2o
236
256
177
172
163
159
145
Stationary Combustion
84
115
102
101
95
95
84
3-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Mobile Combustion
150

139

73
70
67
63
60
Incineration of Waste
2

1

1
1
1
1
1
Petroleum Systems
+

+

+
+
+
+
+
Natural Gas Systems
+

+

+
+
+
+
+
International Bunker









Fuelsb
3

3

3
3
4
4
3
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
a Emissions from Wood Biomass, Ethanol, and Biodiesel Consumption are not included specifically in summing Energy sector
totals. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for LULUCF.
b Emissions from International Bunker Fuels are not included in totals. These values are presented for informational purposes
only, in line with the 2006IPCC Guidelines and UNFCCC reporting obligations.
Emissions estimates reported in the Energy chapter from fossil fuel combustion and fugitive sources include those
from all 50 states, including Hawaii and Alaska, and the District of Columbia. Emissions are also included from U.S.
Territories to the extent they are known to occur (e.g., coal mining does not occur in U.S. Territories). For some
sources there is a lack of detailed information on U.S. Territories including some non-C02 emissions from
combustion. As part of continuous improvement efforts, EPA reviews this on an ongoing basis to ensure emission
sources are included across all geographic areas including U.S. Territories if they are occurring. See Annex 5 for
more information on EPA's assessment of the sources not included in this Inventory.
Each year, some emission and sink estimates in the Inventory are recalculated and revised with improved methods
and/or data. In general, recalculations are made to the U.S. greenhouse gas emission estimates either to
incorporate new methodologies or, most commonly, to update recent historical data. These improvements are
implemented consistently across the previous Inventory's time series (i.e., 1990 to 2018) to ensure that the trend
is accurate. Key updates in this year's Inventory include updates to C02 emissions from Fossil Fuel Combustion
(e.g., updates to C02 emission factors for gasoline and diesel fuels, updates to CH4 and N20 emission factors for
newer non-road gasoline and diesel vehicles and changes to activity and carbon content coefficients), updates to
carbon emissions from non-energy uses of fossil fuels (e.g., heat contents for hydrocarbon gas liquids) and updates
to fugitive emission sources (e.g., CH4 and C02 emissions from natural gas systems distribution and production).
The combined impact of these recalculations averaged -10.5 MMT C02 Eq. (-0.2 percent) per year across the time
series. For more information on specific methodological updates, please see the Recalculations section for each
category in this chapter.
Box 3-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program
In following the United Nations Framework Convention on Climate Change (UNFCCC) requirement under Article
4.1 to develop and submit national greenhouse gas emission inventories, the emissions and removals presented
in this report and this chapter are organized by source and sink categories and calculated using internationally-
accepted methods provided by the Intergovernmental Panel on Climate Change (IPCC) in the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories (2006 IPCC Guidelines). Additionally, the calculated
emissions and removals in a given year for the United States are presented in a common format in line with the
UNFCCC reporting guidelines for the reporting of inventories under this international agreement. The use of
consistent methods to calculate emissions and removals by all nations providing their inventories to the
UNFCCC ensures that these reports are comparable. The presentation of emissions and removals provided in
the Energy chapter do not preclude alternative examinations, but rather, this chapter presents emissions and
removals in a common format consistent with how countries are to report Inventories under the UNFCCC. The
report itself, and this chapter, follows this standardized format, and provides an explanation of the application
of methods used to calculate emissions and removals from energy-related activities.
Energy Data from EPA's Greenhouse Gas Reporting Program
Energy 3-5

-------
EPA's Greenhouse Gas Reporting Program (GHGRP)4 dataset and the data presented in this Inventory are
complementary. The Inventory was used to guide the development of the GHGRP, particularly in terms of scope
and coverage of both sources and gases. The GHGRP dataset continues to be an important resource for the
Inventory, providing not only annual emissions information, but also other annual information, such as activity
data and emission factors that can improve and refine national emission estimates and trends over time.
GHGRP data also allow EPA to disaggregate national inventory estimates in new ways that can highlight
differences across regions and sub-categories of emissions, along with enhancing application of QA/QC
procedures and assessment of uncertainties.
EPA uses annual GHGRP data in a number of Energy sector categories to improve the national estimates
presented in this Inventory consistent with IPCC guidelines (see Box 3-3 of this chapter, and sections 3.4 Coal
Mining, 3.6 Petroleum Systems, and 3.6 Natural Gas Systems).5 Methodologies used in EPA's GHGRP are
consistent with IPCC guidelines, including higher tier methods. Under EPA's GHGRP, facilities collect detailed
information specific to their operations according to detailed measurement standards. It should be noted that
the definitions and provisions for reporting fuel types in EPA's GHGRP may differ from those used in the
Inventory in meeting the UNFCCC reporting guidelines. In line with the UNFCCC reporting guidelines, the
Inventory report is a comprehensive accounting of all emissions from fuel types identified in the IPCC guidelines
and provides a separate reporting of emissions from biomass.
In addition to using GHGRP data to estimate emissions (Section 3.4 Coal Mining, 3.6 Petroleum Systems, and 3.7
Natural Gas Systems), EPA also uses the GHGRP fuel consumption activity data in the Energy sector to
disaggregate industrial end-use sector emissions in the category of C02 Emissions from Fossil Fuel Combustion,
for use in reporting emissions in Common Reporting Format (CRF) tables (See Box 3-3). The industrial end-use
sector activity data collected for the Inventory (EIA 2020) represent aggregated data for the industrial end-use
sector. EPA's GHGRP collects industrial fuel consumption activity data by individual categories within the
industrial end-use sector. Therefore, GHGRP data are used to provide a more detailed breakout of total
emissions in the industrial end-use sector within that source category.
As indicated in the respective Planned Improvements sections for source categories in this chapter, EPA
continues to examine the uses of facility-level GHGRP data to improve the national estimates presented in this
Inventory. See Annex 9 for more information on use of EPA's GHGRP in the Inventory.
3.1 Fossil Fuel Combustion (CRF Source
Category 1A)
Emissions from the combustion of fossil fuels for energy include the greenhouse gases C02, CH4, and N20. Given
that C02 is the primary gas emitted from fossil fuel combustion and represents the largest share of U.S. total
emissions, C02 emissions from fossil fuel combustion are discussed at the beginning of this section. An overview of
CH4 and N20 emissions from the combustion of fuels in stationary sources is then presented, followed by fossil fuel
combustion emissions for all three gases by end-use sector: electric power, industrial, residential, commercial, U.S.
Territories, and transportation.
4	On October 30, 2009, the U.S. Environmental Protection Agency (EPA) published a rule requiring annual reporting of
greenhouse gas data from large greenhouse gas emission sources in the United States. Implementation of the rule, codified at
40 CFR Part 98, is referred to as EPA's Greenhouse Gas Reporting Program (GHGRP).
5	See .
3-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodologies for estimating C02 emissions from fossil fuel combustion differ from the estimation of CH4 and N20
emissions from stationary combustion and mobile combustion. Thus, three separate descriptions of
methodologies, uncertainties, recalculations, and planned improvements are provided at the end of this section.
Total C02, CH4, and N20 emissions from fossil fuel combustion are presented in Table 3-3 and Table 3-4.
Table 3-3: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion (MMT CO2 Eq.)
Gas
1990
2005
2015
2016
2017
2018
2019
C02
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
ch4
15.0
11.8
11.1
10.4
10.1
11.0
11.0
n2o
69.8
75.9
52.3
50.8
48.3
47.1
42.9
Total
4,816.3
5,841.2
5,071.6
4,972.8
4,912.9
5,049.5
4,910.6
Note: Totals may not sum due to independent rounding.
Table 3-4: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion (kt)
Gas	1990	2005	2015	2016	2017	2018	2019
CO^	4,731,466	5,753,507	5,008,270 4,911,532 4,854,480 4,991,420 4,856,702
CH4	600	471	444	417	406	440	441
N;0	234	255 / ¦<	175	171	162	158	144
C02 from Fossil Fuel Combustion
Carbon dioxide is the primary gas emitted from fossil fuel combustion and represents the largest share of U.S. total
greenhouse gas emissions. Carbon dioxide emissions from fossil fuel combustion are presented in Table 3-5. In
2019, C02 emissions from fossil fuel combustion decreased by 2.7 percent relative to the previous year (as shown
in Table 3-6). The decrease in C02 emissions from fossil fuel consumption was a result of a 1 percent decrease in
total energy use and reflects a continued shift from coal to less carbon intensive natural gas and renewables in the
electric power sector. Carbon dioxide emissions from natural gas consumption increased by 53.4 MMT C02 Eq. in
2019, a 3.4 percent increase from 2018, while C02 emissions from coal consumption decreased by 185.3 MMT C02
Eq., a 15.2 percent decrease. The increase in natural gas consumption and emissions in 2019 is observed across all
sectors and is primarily driven by a shift away from coal consumption in the Electric Power sector. In 2019, C02
emissions from fossil fuel combustion were 4,856.7 MMT C02 Eq., or 2.6 percent above emissions in 1990 (see
Table 3-5).6
Table 3-5: CO2 Emissions from Fossil Fuel Combustion by Fuel Type and Sector (MMT CO2
Eq.)
Fuel/Sector
1990
2005
2015
2016
2017
2018
2019
Coal
1,719.8
2,113.7
1,428.5
1,310.7
1,270.2
1,211.6
1,027.1
Residential
3.0
0.8
NO
NO
NO
NO
NO
Commercial
12.0
9.3
3.0
2.3
2.0
1.8
1.6
Industrial
157.8
117.8
70.0
63.2
58.7
54.4
49.5
Transportation
NO
NO
NO
NO
NO
NO
NO
Electric Power
1,546.5
1,982.8
1,351.4
1,242.0
1,207.1
1,152.9
973.5
U.S. Territories
0.5
3.0
4.1
3.2
2.5
2.5
2.5
Natural Gas
1,000.0
1,167.0
1,454.9
1,461.3
1,434.6
1,591.2
1,644.6
Residential
237.8
262.2
252.7
238.4
241.5
273.8
275.3
Commercial
142.0
162.9
175.4
170.5
173.2
192.5
192.8
Industrial
408.8
388.6
459.1
463.9
469.5
494.0
503.3
Transportation
36.0
33.1
39.4
40.1
42.3
50.9
54.8
Electric Power
175.4
318.9
525.2
545.0
505.6
577.4
616.0
6 An additional discussion of fossil fuel emission trends is presented in the Trends in U.S. Greenhouse Gas Emissions chapter.
Energy 3-7

-------
U.S. Territories
NO
1.3
3.0
3.4
2.5
2.5
2.5
Petroleum
2,011.2
2,472.3
2,124.5
2,139.2
2,149.2
2,188.2
2,184.6
Residential
97.8
95.9
64.6
54.4
51.9
64.2
61.5
Commercial
74.3
54.9
66.2
58.7
56.8
51.4
55.3
Industrial
287.2
346.4
268.2
265.4
261.9
265.2
269.7
Transportation
1,433.1
1,825.6
1,679.8
1,719.8
1,740.2
1,765.6
1,762.5
Electric Power
97.5
98.0
23.7
21.5
18.9
22.2
16.2
U.S. Territories
21.2
51.6
22.1
19.4
19.5
19.5
19.5
Geothermal3
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Electric Power
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Total
4,731.5
5,753.5
5,008.3
4,911.5
4,854.5
4,991.4
4,856.7
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
a Although not technically a fossil fuel, geothermal energy-related C02 emissions are included for reporting purposes.
The source of C02 is non-condensable gases in subterranean heated water.
Trends in C02 emissions from fossil fuel combustion are influenced by many long-term and short-term factors. On
a year-to-year basis, the overall demand for fossil fuels in the United States and other countries generally
fluctuates in response to changes in general economic conditions, energy prices, weather, and the availability of
non-fossil alternatives. For example, in a year with increased consumption of goods and services, low fuel prices,
severe summer and winter weather conditions, nuclear plant closures, and lower precipitation feeding
hydroelectric dams, there would likely be proportionally greater fossil fuel consumption than a year with poor
economic performance, high fuel prices, mild temperatures, and increased output from nuclear and hydroelectric
plants.
Longer-term changes in energy usage patterns, however, tend to be more a function of aggregate societal trends
that affect the scale of energy use (e.g., population, number of cars, size of houses, and number of houses), the
efficiency with which energy is used in equipment (e.g., cars, HVAC systems, power plants, steel mills, and light
bulbs), and social planning and consumer behavior (e.g., walking, bicycling, or telecommuting to work instead of
driving).
Carbon dioxide emissions also depend on the source of energy and its carbon (C) intensity. The amount of C in
fuels varies significantly by fuel type. For example, coal contains the highest amount of C per unit of useful energy.
Petroleum has roughly 75 percent of the C per unit of energy as coal, and natural gas has only about 55 percent.7
Table 3-6 shows annual changes in emissions during the last five years for coal, petroleum, and natural gas in
selected sectors.
7 Based on national aggregate carbon content of all coal, natural gas, and petroleum fuels combusted in the United States. See
Annex 2.2 for more details on fuel carbon contents.
3-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 3-6: Annual Change in CO2 Emissions and Total 2019 CO2 Emissions from Fossil Fuel
Combustion for Selected Fuels and Sectors (MMT CO2 Eq. and Percent)
Sector
Fuel Type
2015 to 2016
2016 to 2017
2017 to 2018
2018 to 2019
Total 2019
Electric Power
Coal
-109.4
-8.1%
-34.9
-2.8%
-54.2
-4.5%
-179.3
-15.6%
973.5
Electric Power
Natural Gas
19.8
3.8%
-39.4
-7.2%
71.8
14.2%
38.5
6.7%
616.0
Electric Power
Petroleum
-2.2
-9.4%
-2.5
-11.8%
3.3
17.4%
-6.1
-27.3%
16.2
Transportation
Petroleum
40.0
2.4%
20.4
1.2%
25.5
1.5%
-3.2
-0.2%
1,762.5
Residential
Natural Gas
-14.3
-5.7%
3.1
1.3%
32.3
13.4%
1.5
0.5%
275.3
Commercial
Natural Gas
-4.9
-2.8%
2.6
1.6%
19.3
11.2%
0.3
0.1%
192.8
Industrial
Natural Gas
4.8
1.0%
5.6
1.2%
24.5
5.2%
9.3
1.9%
503.3
Electric Power
All Fuels3
-91.8
-4.8%
-76.8
-4.2%
20.9
1.2%
-146.9
-8.4%
1,606.0
Transportation
All Fuels3
40.6
2.4%
22.6
1.3%
34.1
1.9%
0.6
+%
1,817.2
Residential
All Fuels3
-24.5
-7.7%
0.6
0.2%
44.7
15.2%
-1.3
-0.4%
336.8
Commercial
All Fuels3
-13.0
-5.3%
0.4
0.2%
13.7
5.9%
4.0
1.6%
249.7
Industrial
All Fuels3
-4.8
-0.6%
-2.4
-0.3%
23.5
3.0%
8.9
1.1%
822.5
All Sectors3
All Fuels3
-96.7
-1.9%
-57.1
-1.2%
136.9
2.8%
-134.7
-2.7%
4,856.7
a Includes sector and fuel combinations not shown in this table.
+ Does not exceed 0.05 percent.
As shown in Table 3-6, recent trends in C02 emissions from fossil fuel combustion show a 1.9 percent decrease
from 2015 to 2016, a 1.2 percent decrease from 2016 to 2017, a 2.8 percent increase from 2017 to 2018, and a 2.7
percent decrease from 2018 to 2019. These changes contributed to an overall 3.0 percent decrease in C02
emissions from fossil fuel combustion from 2015 to 2019.
Trends in C02 emissions from fossil fuel combustion over the past five years are largely driven by the electric
power sector, which until recently has accounted for the largest portion of these emissions. The types of fuels
consumed to produce electricity have changed in recent years. Electric power sector consumption of natural gas
primarily increased due to increased production capacity as natural gas-fired plants replaced coal-fired plants and
increased electricity demand related to heating and cooling needs (EIA 2018; EIA 2020f). Total electric power
generation increased by 0.01 percent from 2015 to 2016, decreased by 1.0 percent from 2016 to 2017, increased
by 3.6 percent from 2017 to 2018 and decreased by 1.4 percent from 2018 to 2019. Carbon dioxide emissions
decreased from 2018 to 2019 by 8.4 percent due to increasing electric power generation from natural gas and
decreasing generation from petroleum and coal. Carbon dioxide emissions from coal consumption for electric
power generation decreased by 28.0 percent since 2015, which can be largely attributed to a shift to the use of
less-C02-intensive natural gas to generate electricity and a rapid increase in renewable energy capacity additions in
the electric power sector in recent years.
The trends in C02 emissions from fossil fuel combustion over the past five years also follow changes in heating
degree days (see Box 3-2). Emissions from natural gas consumption in the residential and commercial sectors
increased by 8.2 percent and 9.0 percent from 2015 to 2019, respectively. This trend can be largely attributed to a
5.3 percent increase in heating degree days from 2015 to 2019, which led to an increased demand for heating fuel
and electricity for heat in these sectors. Industrial consumption of natural gas is dependent on market effects of
supply and demand in addition to weather-related heating needs.
Petroleum use in the transportation sector is another major driver of emissions, representing the largest source of
C02 emissions from fossil fuel combustion in 2019. Emissions from petroleum consumption for transportation have
increased by 4.9 percent since 2015 and are primarily attributed to a 5.4 percent increase in vehicle miles traveled
(VMT) over the same time period. Beginning with 2017, the transportation sector is the largest source of national
C02 emissions - whereas in prior years, electric power was the largest source sector.
In the United States, 80 percent of the energy used in 2019 was produced through the combustion of fossil fuels
such as petroleum, natural gas, and coal (see Figure 3-4 and Figure 3-5). Specifically, petroleum supplied the
largest share of domestic energy demands, accounting for 37 percent of total U.S. energy used in 2019. Natural gas
and coal followed in order of energy demand importance, accounting for approximately 32 percent and 11 percent
Energy 3-9

-------
of total U.S. energy used, respectively. Petroleum was consumed primarily in the transportation end-use sector
and the majority of coal was used in the electric power sector. Natural gas was broadly consumed in all end-use
sectors except transportation (see Figure 3-6) (EIA 2020c). The remaining portion of energy used in 2019 was
supplied by nuclear electric power (8 percent) and by a variety of renewable energy sources (11 percent), primarily
hydroelectric power, wind energy, and biofuels (EIA 2020c).8
Figure 3-4: 2019 U.S. Energy Use by Energy Source
Nuclear Electric Power
8.4%
Renewable Energy
11.3%
Petroleum
36.8%
Natural Gas
32.1%
Figure 3-5: Annual U.S. Energy Use
-120
CD
S ioo
Q.
E
80
60
o
u
>-
ai

-------
Figure 3-6: 2019 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type
2,500
2,000
iS" 1,500
(M
0
u
t-
1	1,000
500
0
Fossil fuels are generally combusted for the purpose of producing energy for useful heat and work. During the
combustion process, the C stored in the fuels is oxidized and emitted as C02 and smaller amounts of other gases,
including CH4, carbon monoxide (CO), and non-methane volatile organic compounds (NMVOCs).9 These other C-
containing non-C02 gases are emitted as a byproduct of incomplete fuel combustion, but are, for the most part,
eventually oxidized to C02 in the atmosphere. Therefore, as per IPCC guidelines it is assumed all of the C in fossil
fuels used to produce energy is eventually converted to atmospheric C02.
Box 3-2: Weather and Non-Fossil Energy Effects on C02 Emissions from Fossil Fuel Combustion Trends
Relative Contribution by Fuel Type
<0.05%
250
Petroleum
Coal
Natural Gas
Geothermal
337
25
1,817
1,606
U.S. Territories Commercial	Residential	Industrial	Electric Power Transportation
The United States in 2019 experienced a slightly colder winter overall compared to 2018, as heating degree days
increased 0.6 percent. Colder winter conditions compared to 2018 impacted the amount of energy required for
heating. However, in 2019 heating degree days in the United States were still 5.2 percent below normal (see Figure
3-7). Cooling degree days decreased by 5.4 percent compared to 2018, which reduced demand for air conditioning
in the residential and commercial sector. Cooler summer conditions compared to 2018 impacted the amount of
energy required for cooling, however, 2019 cooling degree days in the United States were still 22.2 percent above
normal (see Figure 3-8) (EIA 2020c).10 The combination of slightly colder winter and cooler summer conditions led
to overall residential and commercial energy consumption decreases of 0.4 and 1.6 percent, respectively relative
to 2018.
9	See the sections entitled Stationary Combustion and Mobile Combustion in this chapter for information on non-C02 gas
emissions from fossil fuel combustion.
10	Degree days are relative measurements of outdoor air temperature. Heating degree days are deviations of the mean daily
temperature below 65 degrees Fahrenheit, while cooling degree days are deviations of the mean daily temperature above 65
degrees Fahrenheit. Heating degree days have a considerably greater effect on energy demand and related emissions than do
cooling degree days. Excludes Alaska and Hawaii. Normals are based on data from 1981 through 2010. The variation in these
normals during this time period was +15 percent and +23 percent for heating and cooling degree days, respectively (99 percent
confidence interval).
Energy 3-11

-------
Figure 3-7: Annual Deviations from Normal Heating Degree Days for the United States
(1950-2019, Index Normal = 100)
30
Normal
(4,538 Heating Degree Days)
20
c -10 99% Confidence
-20
Note: Climatological normal data are highlighted in dark red. Statistical confidence interval for "normal" climatology period of 1981
through 2010.
THninivCT>THnini^o^THnL/irNO^THroL/irsCT*THnmivo^THnmivo^THroL/irsCT*
L/iinmi/iinvD^DiovD^DrsivivrNivoooooocoooo^o^oio^CTioooooTHTHTHTHTH
CTi Ol CT>	0"»	0"> C7i	O"*	O"*	0"» CT* CTi CT»	CT> O O O O O O O O O O
H1H1—H i-H t-H H iH H t—I 1—I 1—I T—I 1—I 1—I 1—I 1—I 1—I 1—I r\J fNJ fN r\J f>J r\j rsi fm rvj rM
Figure 3-8: Annual Deviations from Normal Cooling Degree Days for the United States
(1950-2019, Index Normal = 100)
Normal
(1,228 cooling degree days)
99% Confidence
E
P

c
o
4-1
ro
>
CD
O
oj
o_
Note: Climatological normal data are highlighted dark blue. Statistical confidence interval for "normal" climatology period of 1981
through 2010.
-40
3-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The carbon intensity of the electric power sector is impacted by the amount of non-fossil energy sources of
electricity. The utilization (i.e., capacity factors)11 of nuclear power plants in 2019 remained high at 94 percent. In
2019, nuclear power represented 20 percent of total electricity generation. Since 1990, the wind and solar power
sectors have shown strong growth (between an observed minimum of 89 percent annual electricity generation
growth to a maximum of 162 percent annual electricity generation growth) and have become relatively important
electricity sources. Between 1990 and 2019, renewable energy generation (in kWh) from solar and wind energy
have increased from 0.1 percent in 1990 to 9 percent in 2019 of total electricity generation, which helped drive the
decrease in the carbon intensity of the electricity supply in the United States.
Stationary Combustion
The direct combustion of fuels by stationary sources in the electric power, industrial, commercial, and residential
sectors represent the greatest share of U.S. greenhouse gas emissions. Table 3-7 presents C02 emissions from
fossil fuel combustion by stationary sources. The C02 emitted is closely linked to the type of fuel being combusted
in each sector (see Methodology section of C02 from Fossil Fuel Combustion). In addition to the C02 emitted from
fossil fuel combustion, CH4 and N20 are emitted as well. Table 3-8 and Table 3-9 present CH4 and N20 emissions
from the combustion of fuels in stationary sources. The CH4 and N20 emissions are estimated by applying a
"bottom-up" methodology that utilizes facility-specific technology and fuel use data reported to EPA's Acid Rain
Program (EPA 2020a) (see Methodology section for CH4 and N20 from Stationary Combustion).
Table 3-7: CO2 Emissions from Stationary Fossil Fuel Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2015
2016
2017
2018
2019
Electric Power
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
Coal
1,546.5
1,982.8
1,351.4
1,242.0
1,207.1
1,152.9
973.5
Natural Gas
175.4
318.9
525.2
545.0
505.6
577.4
616.0
Fuel Oil
97.5
98.0
23.7
21.5
18.9
22.2
16.2
Geothermal
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Industrial
853.8
852.9
797.3
792.5
790.1
813.6
822.5
Coal
157.8
117.8
70.0
63.2
58.7
54.4
49.5
Natural Gas
408.8
388.6
459.1
463.9
469.5
494.0
503.3
Fuel Oil
287.2
346.4
268.2
265.4
261.9
265.2
269.7
Commercial
228.3
227.1
244.6
231.6
232.0
245.7
249.7
Coal
12.0
9.3
3.0
2.3
2.0
1.8
1.6
Natural Gas
142.0
162.9
175.4
170.5
173.2
192.5
192.8
Fuel Oil
74.3
54.9
66.2
58.7
56.8
51.4
55.3
Residential
338.6
358.9
317.3
292.8
293.4
338.1
336.8
Coal
3.0
0.8
NO
NO
NO
NO
NO
Natural Gas
237.8
262.2
252.7
238.4
241.5
273.8
275.3
Fuel Oil
97.8
95.9
64.6
54.4
51.9
64.2
61.5
U.S. Territories
21.7
55.9
29.2
26.0
24.6
24.6
24.6
Coal
0.5
3.0
4.1
3.2
2.5
2.5
2.5
Natural Gas
NO
1.3
3.0
3.4
2.5
2.5
2.5
Fuel Oil
21.2
51.6
22.1
19.4
19.5
19.5
19.5
Total
3,262.4
3,894.9
3,289.0
3,151.7
3,072.0
3,174.9
3,039.5
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
11 The capacity factor equals generation divided by net summer capacity. Summer capacity is defined as "The maximum output
that generating equipment can supply to system load, as demonstrated by a multi-hour test, at the time of summer peak
demand (period of June 1 through September 30)." Data for both the generation and net summer capacity are from EIA (2019).
Energy 3-13

-------
Table 3-8: ChU Emissions from Stationary Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2015
2016
2017
2018
2019
Electric Power
0.4
0.9
1.2
1.2
1.1
1.2
1.3
Coal
0.3
0.4
0.3
0.2
0.2
0.2
0.2
Fuel Oil
+
+
+
+
+
+
+
Natural gas
0.1
0.5
0.9
0.9
0.9
1.0
1.1
Wood
+
+
+
+
+
+
+
Industrial
1.8
1.7
1.6
1.6
1.5
1.5
1.5
Coal
0.4
0.3
0.2
0.2
0.2
0.1
0.1
Fuel Oil
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Natural gas
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Wood
1.0
1.0
1.1
1.0
1.0
1.0
1.0
Commercial
1.1
1.1
1.2
1.2
1.2
1.2
1.2
Coal
+
+
+
+
+
+
+
Fuel Oil
0.3
0.2
0.2
0.2
0.2
0.2
0.2
Natural gas
0.3
0.4
0.4
0.4
0.4
0.4
0.4
Wood
0.5
0.5
0.6
0.6
0.6
0.6
0.6
Residential
5.2
4.1
4.5
3.9
3.8
4.5
4.6
Coal
0.2
0.1
0.0
0.0
0.0
0.0
0.0
Fuel Oil
0.3
0.3
0.2
0.2
0.2
0.2
0.2
Natural Gas
0.5
0.6
0.6
0.5
0.5
0.6
0.6
Wood
4.1
3.1
3.7
3.1
3.0
3.7
3.8
U.S. Territories
+
0.1
+
+
+
+
+
Coal
+
+
+
+
+
+
+
Fuel Oil
+
0.1
+
+
+
+
+
Natural Gas
NO
+
+
+
+
+
+
Wood
NO
NO
NO
NO
NO
NO
NO
Total
8.6
7.8
8.5
7.9
7.6
8.5
8.7
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
NO (Not Occurring)
Table 3-9: N2O Emissions from Stationary Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2015
2016
2017
2018
2019
Electric Power
20.5
30.1
26.5
26.2
24.8
24.4
21.1
Coal
20.1
28.0
22.8
22.4
21.2
20.3
16.7
Fuel Oil
0.1
0.1
+
+
+
+
+
Natural Gas
0.3
1.9
3.7
3.8
3.6
4.1
4.4
Wood
+
+
+
+
+
+
+
Industrial
3.1
2.9
2.6
2.6
2.5
2.5
2.5
Coal
0.7
0.6
0.3
0.3
0.3
0.3
0.2
Fuel Oil
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Natural Gas
0.2
0.2
0.2
0.2
0.3
0.3
0.3
Wood
1.6
1.6
1.7
1.7
1.6
1.6
1.6
Commercial
0.4
0.3
0.4
0.3
0.3
0.3
0.3
Coal
0.1
+
+
+
+
+
+
Fuel Oil
0.2
0.1
0.2
0.1
0.1
0.1
0.1
Natural Gas
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wood
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Residential
1.0
0.9
0.9
0.8
0.7
0.9
0.9
Coal
+
+
0.0
0.0
0.0
0.0
0.0
Fuel Oil
0.2
0.2
0.2
0.1
0.1
0.2
0.2
Natural Gas
0.1
0.1
0.1
0.1
0.1
0.1
0.1
3-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Wood
0.7
0.5
0.6
0.5
0.5
0.6
0.6
U.S. Territories
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Coal
+
+
+
+
+
+
+
Fuel Oil
0.1
0.1
0.1
+
+
+
+
Natural Gas
NO
+
+
+
+
+
+
Wood
NO
NO
NO
NO
NO
NO
NO
Total
25.1
34.4
30.5
30.0
28.4
28.2
24.9
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
NO (Not Occurring)
Fossil Fuel Combustion Emissions by Sector
Table 3-10 provides an overview of the C02, CH4, and N20 emissions from fossil fuel combustion by sector,
including transportation, electric power, industrial, residential, commercial, and U.S. territories.
Table 3-10: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion by Sector (MMT CO2
Eq.)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation
1,520.2
1,904.2
1,743.6
1,783.2
1,804.8
1,837.8
1,837.5
C02
1,469.1
1,858.6
1,719.2
1,759.9
1,782.4
1,816.6
1,817.2
ch4
6.4
4.0
2.6
2.5
2.5
2.4
2.4
n2o
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Electric Power
1,840.9
2,431.0
1,928.3
1,836.2
1,757.9
1,778.5
1,628.4
C02
1,820.0
2,400.1
1,900.6
1,808.9
1,732.0
1,752.9
1,606.0
ch4
0.4
0.9
1.2
1.2
1.1
1.2
1.3
n2o
20.5
30.1
26.5
26.2
24.8
24.4
21.1
Industrial
858.7
857.6
801.5
796.7
794.2
817.6
826.5
co2
853.8
852.9
797.3
792.5
790.1
813.6
822.5
ch4
1.8
1.7
1.6
1.6
1.5
1.5
1.5
n2o
3.1
2.9
2.6
2.6
2.5
2.5
2.5
Residential
344.9
363.8
322.6
297.4
297.9
343.5
342.3
C02
338.6
358.9
317.3
292.8
293.4
338.1
336.8
ch4
5.2
4.1
4.5
3.9
3.8
4.5
4.6
n2o
1.0
0.9
0.9
0.8
0.7
0.9
0.9
Commercial
229.8
228.6
246.2
233.1
233.5
247.3
251.3
co2
228.3
227.1
244.6
231.6
232.0
245.7
249.7
ch4
1.1
1.1
1.2
1.2
1.2
1.2
1.2
n2o
0.4
0.3
0.4
0.3
0.3
0.3
0.3
U.S. Territories3
21.8
56.1
29.3
26.1
24.6
24.7
24.7
Total
4,816.3
5,841.2
5,071.6
4,972.8
4,912.9
5,049.5
4,910.6
Note: Totals may not sum due to independent rounding.
a U.S. Territories are not apportioned by sector, and emissions shown in the table are total greenhouse gas
emissions from all fuel combustion sources.
Other than greenhouse gases C02, CH4, and N20, gases emitted from stationary combustion include the
greenhouse gas precursors nitrogen oxides (NOx), CO, and NMVOCs.12 Methane and N20 emissions from stationary
combustion sources depend upon fuel characteristics, size and vintage of combustion device, along with
combustion technology, pollution control equipment, ambient environmental conditions, and operation and
maintenance practices. Nitrous oxide emissions from stationary combustion are closely related to air-fuel mixes
12 Sulfur dioxide (S02) emissions from stationary combustion are addressed in Annex 6.3.
Energy 3-15

-------
and combustion temperatures, as well as the characteristics of any pollution control equipment that is employed.
Methane emissions from stationary combustion are primarily a function of the CH4 content of the fuel and
combustion efficiency.
Mobile combustion also produces emissions of CH4, N20, and greenhouse gas precursors including NOx, CO, and
NMVOCs. As with stationary combustion, N20 and NOx emissions from mobile combustion are closely related to
fuel characteristics, air-fuel mixes, combustion temperatures, and the use of pollution control equipment. Nitrous
oxide from mobile sources, in particular, can be formed by the catalytic processes used to control NOx, CO, and
hydrocarbon emissions. Carbon monoxide emissions from mobile combustion are significantly affected by
combustion efficiency and the presence of post-combustion emission controls. Carbon monoxide emissions are
highest when air-fuel mixtures have less oxygen than required for complete combustion. These emissions occur
especially in vehicle idle, low speed, and cold start conditions. Methane and NMVOC emissions from motor
vehicles are a function of the CH4 content of the motor fuel, the amount of hydrocarbons passing uncombusted
through the engine, and any post-combustion control of hydrocarbon emissions (such as catalytic converters).
An alternative method of presenting combustion emissions is to allocate emissions associated with electric power
to the sectors in which it is used. Four end-use sectors are defined: transportation, industrial, residential, and
commercial. In Table 3-11 below, electric power emissions have been distributed to each end-use sector based
upon the sector's share of national electricity use, with the exception of CH4 and N20 from transportation
electricity use.13 Emissions from U.S. Territories are also calculated separately due to a lack of end-use-specific
consumption data.14 This method assumes that emissions from combustion sources are distributed across the four
end-use sectors based on the ratio of electricity use in that sector. The results of this alternative method are
presented in Table 3-11.
Table 3-11: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion by End-Use Sector
(MMT COz Eq.)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation
1,523.3
1,908.9
1,747.9
1,787.4
1,809.1
1,842.5
1,842.3
C02
1,472.2
1,863.4
1,723.5
1,764.1
1,786.8
1,821.2
1,821.9
ch4
6.4
4.0
2.6
2.5
2.5
2.4
2.4
n2o
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Industrial
1,553.0
1,603.4
1,359.1
1,322.1
1,306.1
1,326.3
1,298.3
C02
1,540.2
1,589.2
1,346.8
1,310.1
1,294.5
1,314.9
1,287.8
ch4
2.0
2.0
1.9
1.9
1.9
1.9
1.9
n2o
10.8
12.2
10.3
10.1
9.8
9.5
8.6
Residential
944.4
1,230.9
1,016.4
960.8
924.2
995.0
933.9
C02
931.3
1,214.9
1,001.1
946.2
910.5
980.2
920.3
ch4
5.4
4.4
4.9
4.3
4.2
5.0
5.1
n2o
7.7
11.6
10.4
10.3
9.6
9.9
8.6
Commercial
773.7
1,041.9
918.9
876.3
848.8
861.0
811.4
C02
766.0
1,030.1
907.6
865.2
838.2
850.6
802.1
ch4
1.2
1.4
1.6
1.6
1.6
1.6
1.7
n2o
6.5
10.4
9.6
9.5
9.0
8.8
7.6
U.S. Territories3
21.8
56.1
29.3
26.1
24.6
24.7
24.7
Total
4,816.3
5,841.2
5,071.6
4,972.8
4,912.9
5,049.5
4,910.6
Notes: Totals may not sum due to independent rounding. Emissions from fossil fuel combustion by electric
power are allocated based on aggregate national electricity use by each end-use sector.
13	Separate calculations are performed for transportation-related CH4 and N20. The methodology used to calculate these
emissions is discussed in the Mobile Combustion section.
14	U.S. Territories consumption data that are obtained from EIA are only available at the aggregate level and cannot be broken
out by end-use sector. The distribution of emissions to each end-use sector for the 50 states does not apply to territories data.
3-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
a U.S. Territories are not apportioned by sector, and emissions are total greenhouse gas emissions from all
fuel combustion sources.
Electric Power Sector
The process of generating electricity is the largest stationary source of C02 emissions in the United States,
representing 30.6 percent of total C02 emissions from all C02 emissions sources across the United States. Methane
and N20 accounted for a small portion of total greenhouse gas emissions from electric power, representing 0.1
percent and 1.3 percent, respectively. Electric power also accounted for 33.1 percent of C02 emissions from fossil
fuel combustion in 2019. Methane and N20 from electric power represented 11.4 and 49.3 percent of total CH4
and N20 emissions from fossil fuel combustion in 2019, respectively.
For the underlying energy data used in this chapter, the Energy Information Administration (EIA) places electric
power generation into three functional categories: the electric power sector, the commercial sector, and the
industrial sector. The electric power sector consists of electric utilities and independent power producers whose
primary business is the production of electricity. This includes both regulated utilities and non-utilities (e.g.,
independent power producers, qualifying co-generators, and other small power producers). Electric generation is
reported as occurring in other sectors where the producer of the power indicates that its primary business is
something other than the production of electricity.15
Total greenhouse gas emissions from the electric power sector have decreased by 11.5 percent since 1990. From
1990 to 2007, electric power sector emissions increased by 32 percent, driven by a significant increase in electricity
demand (37 percent) while the carbon intensity of electricity generated showed a minor increase (0.3 percent).
From 2008 to 2019, as electricity demand increased by only 2 percent, electric power sector emissions decreased
by 13 percent, driven by a significant drop (26 percent) in the carbon intensity of electricity generated. Overall, the
carbon intensity of the electric power sector, in terms of C02 Eq. per QBtu, decreased by 16 percent from 1990 to
2019 with additional trends detailed in Box 3-4. This decoupling of electric power generation and the resulting C02
emissions is shown in Figure 3-9. This recent decarbonization of the electric power sector is a result of several key
drivers.
Coal-fired electric generation (in kilowatt-hours [kWh]) decreased from 54 percent of generation in 1990 to 24
percent in 2019.16 This corresponded with an increase in natural gas generation and renewable energy generation,
largely from wind and solar energy. Natural gas generation (in kWh) represented 11 percent of electric power
generation in 1990 and increased over the 30-year period to represent 37 percent of electric power sector
generation in 2019 (see Table 3-12). Natural gas has a much lower carbon content than coal and is generated in
power plants that are generally more efficient in terms of kWh produced per Btu of fuel combusted, which has led
to lower emissions as natural gas replaces coal-powered electricity generation. Natural gas and coal used in the
U.S. in 2019 had an average carbon content of 14.43 MMT C/QBtu and 26.08 MMT C/QBtu respectively.
Table 3-12: Electric Power Generation by Fuel Type (Percent)
Fuel Type
1990
2005
2015
2016
2017
2018
2019
Coal
54.1%
51.1%
34.2%
31.4%
30.9%
28.4%
24.2%
Natural Gas
10.7%
17.5%
31.6%
32.7%
30.9%
34.0%
37.3%
Nuclear
19.9%
20.0%
20.4%
20.6%
20.8%
20.1%
20.4%
Renewables
11.3%
8.3%
13.0%
14.7%
16.8%
16.8%
17.6%
Petroleum
4.1%
3.0%
0.7%
0.6%
0.5%
0.6%
0.4%
Other Gases3
+%
0.1%
0.1%
0.1%
0.1%
0.1%
0.1%
15	Utilities primarily generate power for the U.S. electric grid for sale to retail customers. Non-utilities typically generate
electricity for sale on the wholesale electricity market (e.g., to utilities for distribution and resale to retail customers). Where
electricity generation occurs outside the ElA-defined electric power sector, it is typically for the entity's own use.
16	Values represent electricity net generation from the electric power sector (EIA 2020c).
Energy 3-17

-------
Net Electricity Generation







(Billion kWh)b
2,905 |
3,902 |
3,917
3,917
3,877
4,017
3,962
+ Does not exceed 0.05 percent.
a Other gases include blast furnace gas, propane gas, and other manufactured and waste gases derived from fossil
fuels.
b Represents net electricity generation from the electric power sector. Excludes net electricity generation from
commercial and industrial combined-heat-and-power and electricity-only plants. Does not include electricity
generation from purchased steam as the fuel used to generate the steam cannot be determined.
In 2019, C02 emissions from the electric power sector decreased by 8.4 percent relative to 2018. This decrease in
C02 emissions was primarily driven by a decrease in coal and petroleum consumed to produce electricity in the
electric power sector as well as a decrease in electricity demand (1.2 percent reduction in retail sales).
Consumption of coal for electric power decreased by 15.5 percent while consumption of natural gas increased 6.7
percent from 2018 to 2019. There has also been a rapid increase in renewable energy electricity generation in the
electric power sector in recent years. Electricity generation from renewable sources increased by 3 percent from
2018 to 2019 (see Table 3-12). The decrease in coal-powered electricity generation and increase in natural gas and
renewable energy electricity generation contributed to a decoupling of emissions trends from electric power
generation trends over the recent time series (see Figure 3-9).
Decreases in natural gas prices and the associated increase in natural gas generation, particularly between 2005
and 2019, was one of the main drivers of the recent fuel switching and decrease in electric power sector carbon
intensity. During this time period, the cost of natural gas (in $/MMBtu) decreased by 56 percent while the cost of
coal (in $/MMBtu) increased by 74 percent (EIA 2020c). Also, between 1990 and 2019, renewable energy
generation (in kWh) from wind and solar energy increased from 0.1 percent of total generation in 1990 to 9
percent in 2019, which also helped drive the decrease in electric power sector carbon intensity. This decrease in
carbon intensity occurred even as total electricity retail sales increased 41 percent, from 2,713 billion kWh in 1990
to 3,811 billion kWh in 2019.
Figure 3-9: Fuels Used in Electric Power Generation and Total Electric Power Sector CO2
Emissions
50,000
40,000
CO
_ 30,000
>
Id 20,000
10,000
Nuclear (TBtu)
Renewable Energy Sources (TBtu)
Petroleum (TBtu)
Natural Gas (TBtu)
Coal (TBtu)
I Net Generation (Index from 1990) [Right Axis]
I Sector CO2 Emissions (Index from 1990) [Right Axis]
160
140
120
100
80
20
o -i-i csj n	m
O"!	O"! CT* G"»
O"!	CT* CT*
 r*s 00
O	CT>
O	CT>
o
CTt
cn
X

-------
Figure 3-10: Electric Power Retail Sales by End-Use Sector
1,600
1,500
Residential
1,400
Commercial
1=0 1,100
1,000
Industrial
900
800
o o o o o o o
In 2019, electricity sales to the residential and commercial end-use sectors, as presented in Figure 3-10, decreased
by 2.0 percent and 1.5 percent relative to 2018, respectively. Electricity sales to the industrial sector in 2019
increased approximately 0.2 percent relative to 2018. The sections below describe end-use sector energy use in
more detail. Overall, in 2019, the amount of electricity retail sales (in kWh) decreased by 1.2 percent relative to
2018.
Industrial Sector
Industrial sector C02, CH4, and N20 emissions accounted for 17,14, and 6 percent of C02, CH4, and N20 emissions
from fossil fuel combustion, respectively in 2019. Carbon dioxide, CH4, and N20 emissions resulted from the direct
consumption of fossil fuels for steam and process heat production.
The industrial end-use sector, per the underlying energy use data from EIA, includes activities such as
manufacturing, construction, mining, and agriculture. The largest of these activities in terms of energy use is
manufacturing, of which six industries—Petroleum Refineries, Chemicals, Paper, Primary Metals, Food, and
Nonmetallic Mineral Products—represent the majority of the energy use (EIA 2020c; EIA 2009b).
There are many dynamics that impact emissions from the industrial sector including economic activity, changes in
the make-up of the industrial sector, changes in the emissions intensity of industrial processes, and weather-
related impacts on heating and cooling of industrial buildings.17 Structural changes within the U.S. economy that
lead to shifts in industrial output away from energy-intensive manufacturing products to less energy-intensive
products (e.g., from steel to computer equipment) have had a significant effect on industrial emissions.
From 2018 to 2019, total industrial production and manufacturing output increased by 0.2 percent (FRB 2019).
Over this period, output increased across production indices for Food, and Nonmetallic Mineral Products, and
decreased slightly for Paper, Petroleum Refineries, Chemicals, and Primary Metals (see Figure 3-11). From 2018 to
2019,	total energy use in the industrial sector increased by 1.5 percent. Due to the relative increases and decreases
of individual indices there was an increase in natural gas and a decrease in electricity used by the sector (see Figure
3-12). In 2019, C02, CH4, and N20 emissions from fossil fuel combustion and electricity use within the industrial
end-use sector totaled 1,298.3 MMT C02 Eq., a 2.1 percent decrease from 2018 emissions.
17 Some commercial customers are large enough to obtain an industrial price for natural gas and/or electricity and are
consequently grouped with the industrial end-use sector in U.S. energy statistics. These misclassifications of large commercial
customers likely cause the industrial end-use sector to appear to be more sensitive to weather conditions.
Energy 3-19

-------
Through EPA's Greenhouse Gas Reporting Program (GHGRP), specific industrial sector trends can be discerned
from the overall total EIA industrial fuel consumption data used for these calculations. For example, from 2018 to
2019, the underlying EIA data showed decreased consumption of coal, and increase of natural gas in the industrial
sector. The GHGRP data highlights that several industries contributed to these trends, including chemical
manufacturing; pulp, paper and print; food processing, beverages and tobacco; minerals manufacturing; and
agriculture-forest-fisheries.18
Figure 3-11: Industrial Production Indices (Index 2012=100)
140
Total Industrial excluding Computers, Communications Equipment, and
Semiconductors
120
100
Total Industrial
140
Paper
120
100
Food
140
Stone, Clay, and Glass Products
120
100
Chemicals
140
Primary Metals
120
100
Petroleum Refineries
l—l
O iH
o o
o o
CO
l—l
LH
18 Further details on industrial sector combustion emissions are provided by EPA's GHGRP. See
.
3-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 3-12: Fuels and Electricity Used in Industrial Sector, Industrial Output, and Total
Sector CO2 Emissions (Including Electricity)
Renewable Energy Sources (TBtu) I Industrial Output (Index vs. 1990) [Right Axis]
Coal (TBtu)	¦ Sector CO2 Emissions (Index vs. 1990) [Right Axis]
Petroleum (TBtu)
Natural Gas
Electricity Use (TBtu)
35,000
30,000
25,000
m
^ 20,000
 CT* O"!
Qi	C7> CT*	CT*
Despite the growth in industrial output (70 percent) and the overall U.S. economy (104 percent) from 1990 to
2019, direct C02 emissions from fossil fuel combustion in the industrial sector decreased by 3.7 percent over the
same time series (see Figure 3-12). A number of factors are assumed to result in decoupling of growth in industrial
output from industrial greenhouse gas emissions, for example: (1) more rapid growth in output from less energy-
intensive industries relative to traditional manufacturing industries, and (2) energy-intensive industries such as
steel are employing new methods, such as electric arc furnaces, that are less carbon intensive than the older
methods.
Box 3-3: Uses of Greenhouse Gas Reporting Program Data and Improvements in Reporting Emissions from
Industrial Sector Fossil Fuel Combustion
As described in the calculation methodology, total fossil fuel consumption for each year is based on aggregated
end-use sector consumption published by the EIA. The availability of facility-level combustion emissions through
EPA's GHGRP has provided an opportunity to better characterize the industrial sector's energy consumption and
emissions in the United States, through a disaggregation of ElA's industrial sector fuel consumption data from
select industries.
For GHGRP 2010 through 2019 reporting years, facility-level fossil fuel combustion emissions reported through
EPA's GHGRP were categorized and distributed to specific industry types by utilizing facility-reported NAICS
codes (as published by the U.S. Census Bureau). As noted previously in this report, the definitions and provisions
for reporting fuel types in EPA's GHGRP include some differences from the Inventory's use of EIA national fuel
statistics to meet the UNFCCC reporting guidelines. The IPCC has provided guidance on aligning facility-level
reported fuels and fuel types published in national energy statistics, which guided this exercise.19
As with previous Inventory reports, the current effort represents an attempt to align, reconcile, and coordinate
the facility-level reporting of fossil fuel combustion emissions under EPA's GHGRP with the national-level
approach presented in this report. Consistent with recommendations for reporting the Inventory to the
19 See Section 4 "Use of Facility-Level Data in Good Practice National Greenhouse Gas Inventories" of the IPCC meeting report,
and specifically the section on using facility-level data in conjunction with energy data, at .
Energy 3-21

-------
UNFCCC, progress was made on certain fuel types for specific industries and has been included in the CRF tables
that are submitted to the UNFCCC along with this report.20 The efforts in reconciling fuels focus on standard,
common fuel types (e.g., natural gas, distillate fuel oil) where the fuels in ElA's national statistics aligned well
with facility-level GHGRP data. For these reasons, the current information presented in the Common Reporting
Format (CRF) tables should be viewed as an initial attempt at this exercise. Additional efforts will be made for
future Inventory reports to improve the mapping of fuel types and examine ways to reconcile and coordinate
any differences between facility-level data and national statistics. The current analysis includes the full time
series presented in the CRF tables. Analyses were conducted linking GHGRP facility-level reporting with the
information published by EIA in its MECS data in order to disaggregate the full 1990 through 2019 time period in
the CRF tables. It is believed that the current analysis has led to improvements in the presentation of data in the
Inventory, but further work will be conducted, and future improvements will be realized in subsequent
Inventory reports. This includes incorporating the latest MECS data as it becomes available.
Residential and Commercial Sectors
Emissions from the residential and commercial sectors have generally decreased since 1990. Short-term trends are
often correlated with seasonal fluctuations in energy use caused by weather conditions, rather than prevailing
economic conditions. Population growth and a trend towards larger houses has led to increasing energy use over
the time series, while population migration to warmer areas and improved energy efficiency and building
insulation have slowed the increase in energy use in recent years. Starting in around 2014, energy use and
emissions begin to decouple due to decarbonization of the electric power sector (see Figure 3-13).
Figure 3-13: Fuels and Electricity Used in Residential and Commercial Sectors, Heating and
Cooling Degree Days, and Total Sector CO2 Emissions (Including Electricity)
25,000
20,000
m
.
pi
15,000
=y 10,000
5,000
Coal (TBtu)
Renewable Energy Sources (TBtu)
Petroleum (TBtu)
Natural Gas
I Electricity Use (TBtu)
I Sector CO2 Emissions (Index vs. 1990) [Right Axis]
I Heating and Cooling Degree Days (Index vs. 1990) [Right Axis]
180
160
140
120
100
80
60
40
20
0
o
cn
CT1
X
ooooooooooooooo
i-ii—11—ii—11—ii—ii-HHr-ir>jr\jrsjr>jrMrMr\irMfMf\irMf\ijr\i
0000
fM M fM fM
In 2019 the residential and commercial sectors accounted for 7 and 5 percent of C02 emissions from fossil fuel
combustion, respectively; 42 and 11 percent of CH4 emissions from fossil fuel combustion, respectively; and 2 and
1 percent of N20 emissions from fossil fuel combustion, respectively. Emissions from these sectors were largely
due to the direct consumption of natural gas and petroleum products, primarily for heating and cooking needs.
20 See .
3-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Coal consumption was a minor component of energy use in the commercial sector and did not contribute to any
energy use in the residential sector. In 2019, total emissions (C02, CH4, and N20) from fossil fuel combustion and
electricity use within the residential and commercial end-use sectors were 933.9 MMT C02 Eq. and 811.4 MMT C02
Eq., respectively. Total C02, CH4, and N20 emissions from combined fossil fuel combustion and electricity use
within the residential and commercial end-use sectors decreased by 6.1 and 5.8 percent from 2018 to 2019,
respectively. A slight increase in heating degree days (0.6 percent) impacted energy demand for heating in the
residential and commercial sectors. This was partially offset by a 5.4 percent decrease in cooling degree days
compared to 2018, which reduced demand for air conditioning in the residential and commercial sectors. In
addition, a shift toward energy efficient products and more stringent energy efficiency standards for household
equipment has contributed to a decrease in energy demand in households (EIA 2020g), resulting in a decrease in
energy-related emissions. In the long term, the residential sector is also affected by population growth, migration
trends toward warmer areas, and changes in total housing units and building attributes (e.g., larger sizes and
improved insulation).
In 2019, combustion emissions from natural gas consumption represented 82 and 77 percent of the direct fossil
fuel C02 emissions from the residential and commercial sectors, respectively. Carbon dioxide emissions from
natural gas combustion in the residential and commercial sectors in 2019 increased by 0.5 percent and 0.1 percent
from 2018 to 2019, respectively.
U.S. Territories
Emissions from U.S. Territories are based on the fuel consumption in American Samoa, Guam, Puerto Rico, U.S.
Virgin Islands, Wake Island, and other U.S. Pacific Islands. As described in the Methodology section of C02 from
Fossil Fuel Combustion, this data is collected separately from the sectoral-level data available for the general
calculations. As sectoral information is not available for U.S. Territories, C02, CH4, and N20 emissions are not
presented for U.S. Territories in the tables above by sector, though the emissions will occur across all sectors and
sources including stationary, transportation and mobile combustion sources.
Transportation Sector and Mobile Combustion
This discussion of transportation emissions follows the alternative method of presenting combustion emissions by
allocating emissions associated with electricity generation to the transportation end-use sector, as presented in
Table 3-11. Table 3-10 presents direct C02, CH4, and N20 emissions from all transportation sources (i.e., excluding
emissions allocated to electricity consumption in the transportation end-use sector).
The transportation end-use sector and other mobile combustion accounted for 1,842.3 MMT C02 Eq. in 2019,
which represented 36 percent of C02 emissions, 22 percent of CH4 emissions, and 42 percent of N20 emissions
from fossil fuel combustion, respectively.21 Fuel purchased in the United States for international aircraft and
marine travel accounted for an additional 117.2 MMT C02 Eq. in 2019; these emissions are recorded as
international bunkers and are not included in U.S. totals according to UNFCCC reporting protocols.
Transportation End-Use Sector
From 1990 to 2019, transportation emissions from fossil fuel combustion rose by 21 percent due, in large part, to
increased demand for travel (see Figure 3-14). The number of vehicle miles traveled (VMT) by light-duty motor
21 Note that these totals include C02, CH4 and N20 emissions from some sources in the U.S. Territories (ships and boats,
recreational boats, non-transportation mobile sources) and CH4 and N20 emissions from transportation rail electricity.
Energy 3-23

-------
vehicles (passenger cars and light-duty trucks) increased 47 percent from 1990 to 2019,22 as a result of a
confluence of factors including population growth, economic growth, urban sprawl, and periods of low fuel prices.
From 2018 to 2019, C02 emissions from the transportation end-use sector increased by 0.04 percent. The small
increase in emissions is primarily attributed to an increase in non-road fuel use, particularly jet fuel consumption.
Commercial aircraft emissions increased by 3.5 percent between 2018 and 2019, but have decreased 4 percent
since 2007 (FAA 2021).23 Decreases in jet fuel emissions (excluding bunkers) since 2007 are due in part to
improved operational efficiency that results in more direct flight routing, improvements in aircraft and engine
technologies to reduce fuel burn and emissions, and the accelerated retirement of older, less fuel-efficient aircraft.
Almost all of the energy consumed for transportation was supplied by petroleum-based products, with more than
half being related to gasoline consumption in automobiles and other highway vehicles. Other fuel uses, especially
diesel fuel for freight trucks and jet fuel for aircraft, accounted for the remainder. The primary driver of
transportation-related emissions was C02 from fossil fuel combustion, which increased by 24 percent from 1990 to
2019. Annex 3.2 presents the total emissions from all transportation and mobile sources, including C02, N20, CH4,
and HFCs.
Figure 3-14: Fuels Used in Transportation Sector, Onroad VMT, and Total Sector CO2
Emissions
40,000
35,000
30,000
-3- 25,000
m
b
5 20,000
>S
Ol
q3
c
m 15,000
10,000
5,000
I Other Fuels (TBtu)
Residual Fuel (TBtu)
Natural Gas (TBtu)
Renewable Energy (TBtu)
Jet Fuel (TBtu)
Distillate Fuel (TBtu)
I Motor Gasoline (TBtu)
Onroad VMT (Index vs. 1990) [Right Axis]
I Sector CO2 Emissions (Index vs. 1990) [Right Axis]
200
180
160
140
Notes: Distillate fuel, residual fuel, and jet fuel include adjustments for international bunker fuels. Distillate fuel and motor
gasoline include adjustments for the sectoral allocation of these fuels. Other Fuels includes aviation gasoline and propane.
Source: Information on fuel consumption was obtained from EIA (2019a).
22	VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2019). In 2011, FHWA
changed its methods for estimating VMT by vehicle class, which led to a shift in VMT and emissions among on-road vehicle
classes in the 2007 to 2019 time period. In absence of these method changes, light-duty VMT growth between 1990 and 2019
would likely have been even higher.
23	Commercial aircraft, as modeled in FAA's AEDT (FAA 2021), consists of passenger aircraft, cargo, and other chartered flights.
3-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Transportation Fossil Fuel Combustion CO2 Emissions
Domestic transportation C02 emissions increased by 24 percent (349.8 MMT C02) between 1990 and 2019, an
annualized increase of 0.8 percent. Among domestic transportation sources in 2019, light-duty vehicles (including
passenger cars and light-duty trucks) represented 58 percent of C02 emissions from fossil fuel combustion,
medium- and heavy-duty trucks and buses 25 percent, commercial aircraft 7 percent, and other sources 10
percent. See Table 3-13 for a detailed breakdown of transportation C02 emissions by mode and fuel type.
Almost all of the energy consumed by the transportation sector is petroleum-based, including motor gasoline,
diesel fuel, jet fuel, and residual oil. Carbon dioxide emissions from the combustion of ethanol and biodiesel for
transportation purposes, along with the emissions associated with the agricultural and industrial processes
involved in the production of biofuel, are captured in other Inventory sectors.24 Ethanol consumption by the
transportation sector has increased from 0.7 billion gallons in 1990 to 13.6 billion gallons in 2019, while biodiesel
consumption has increased from 0.01 billion gallons in 2001 to 1.8 billion gallons in 2019. For additional
information, see Section 3.11 on biofuel consumption at the end of this chapter and Table A-81 in Annex 3.2.
Carbon dioxide emissions from passenger cars and light-duty trucks totaled 1,052.6 MMT C02 in 2019. This is an
increase of 14 percent (128.1 MMT C02) from 1990 due, in large part, to increased demand for travel as fleet-wide
light-duty vehicle fuel economy was relatively stable (average new vehicle fuel economy declined slowly from 1990
through 2004 and then increased more rapidly from 2005 through 2019). Carbon dioxide emissions from
passenger cars and light-duty trucks peaked at 1,154.7 MMT C02 in 2004, and since then have declined about 9
percent. The decline in new light-duty vehicle fuel economy between 1990 and 2004 (Figure 3-15) reflects the
increasing market share of light-duty trucks, which grew from about 30 percent of new vehicle sales in 1990 to 48
percent in 2004. Starting in 2005, average new vehicle fuel economy began to increase while light-duty vehicle
VMT grew only modestly for much of the period. Light-duty vehicle VMT grew by less than one percent or declined
each year between 2005 and 2013,25 then grew at a faster rate until 2016 (2.6 percent from 2014 to 2015, and 2.5
percent from 2015 to 2016). Since 2017, the rate of light-duty vehicle VMT growth slowed to less than one percent
each year. Average new vehicle fuel economy has increased almost every year since 2005, while the light-duty
truck share decreased to about 33 percent in 2009 and has since varied from year to year between 36 and 56
percent. Since 2014, light-duty truck share has slowly increased and is about 56 percent of new vehicles sales in
model year 2019 (EPA 2019b). See Annex 3.2 for data by vehicle mode and information on VMT and the share of
new vehicles (in VMT).
Medium- and heavy-duty truck C02 emissions increased by 90 percent from 1990 to 2019. This increase was largely
due to a substantial growth in medium- and heavy-duty truck VMT, which increased by 109 percent between 1990
and 2019.26 Carbon dioxide from the domestic operation of commercial aircraft increased by 22 percent (24.3
MMT C02) from 1990 to 2019.27 Across all categories of aviation, excluding international bunkers, C02 emissions
24	Biofuel estimates are presented in the Energy chapter for informational purposes only, in line with IPCC methodological
guidance and UNFCCC reporting obligations. Net carbon fluxes from changes in biogenic carbon reservoirs in croplands are
accounted for in the estimates for Land Use, Land-Use Change, and Forestry (see Chapter 6). More information and additional
analyses on biofuels are available at EPA's Renewable Fuels Standards website. See .
25	VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2019). In 2007 and 2008
light-duty VMT decreased 3.0 percent and 2.3 percent, respectively. Note that the decline in light-duty VMT from 2006 to 2007
is due at least in part to a change in FHWA's methods for estimating VMT. In 2011, FHWA changed its methods for estimating
VMT by vehicle class, which led to a shift in VMT and emissions among on-road vehicle classes in the 2007 to 2019 time period.
In absence of these method changes, light-duty VMT growth between 2006 and 2007 would likely have been higher.
26	While FHWA data shows consistent growth in medium- and heavy-duty truck VMT over the 1990 to 2019 time period, part of
the growth reflects a method change for estimating VMT starting in 2007. This change in methodology in FHWA's VM-1 table
resulted in large changes in VMT by vehicle class, thus leading to a shift in VMT and emissions among on-road vehicle classes in
the 2007 to 2019 time period. During the time period prior to the method change (1990 to 2006), VMT for medium- and heavy-
duty trucks increased by 51 percent.
27	Commercial aircraft, as modeled in FAA's AEDT, consists of passenger aircraft, cargo, and other chartered flights.
Energy 3-25

-------
decreased by 4 percent (7.9 MMT C02) between 1990 and 20 19.28 This includes a 66 percent (23.1 MMT C02)
decrease in C02 emissions from domestic military operations.
Transportation sources also produce CH4 and N20; these emissions are included in Table 3-14 and Table 3-15 and
in the CH4 and N20 from Mobile Combustion section. Annex 3.2 presents total emissions from all transportation
and mobile sources, including C02, CH4, N20, and HFCs.
Figure 3-15: Sales-Weighted Fuel Economy of New Passenger Cars and Light-Duty Trucks,
1990-2019
? 22
rH
Source: EPA (2020a).
28 Includes consumption of jet fuel and aviation gasoline. Does not include aircraft bunkers, which are not included in national
emission totals, in line with IPCC methodological guidance and UNFCCC reporting obligations.
3-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 3-16: Sales of New Passenger Cars and Light-Duty Trucks, 1990-2019
100%
90%
80%
J8 70%
ra
in
15 60%
50%
m 40%
 vd rv co cn
i—i i—i i—i i—i i—i
Source: EPA (2019b).
Table 3-13: CO2 Emissions from Fossil Fuel Combustion in Transportation End-Use Sector
(MMT COz Eq.)
Fuel/Vehicle Type
1990

2005

2015a
2016a
2017a
2018a
2019a
Gasolineb
958.9

1,150.1

1,058.6
1,084.8
1,081.8
1,097.1
1,086.8
Passenger Cars
604.3

637.1

724.3
737.8
737.4
748.8
742.3
Light-Duty Trucks
300.6

463.5

280.5
291.8
288.2
290.9
289.1
Medium- and Heavy-Duty









Trucks0
37.7

33.8

38.9
40.0
40.9
41.9
40.1
Buses
0.3

0.4

0.9
0.9
0.9
1.0
1.0
Motorcycles
1.7

1.6

3.6
3.8
3.7
3.8
3.6
Recreational Boatsd
14.3

13.7

10.5
10.6
10.6
10.7
10.7
Distillate Fuel Oil (Diesel)b
262.9

462.6

457.5
454.2
468.3
480.2
481.1
Passenger Cars
7.9

4.3

4.3
4.3
4.3
4.4
4.6
Light-Duty Trucks
11.5

26.1

13.8
14.1
14.1
14.2
14.9
Medium- and Heavy-Duty









Trucks0
190.5

364.2

366.8
369.3
381.6
390.9
394.8
Buses
8.0

10.7

17.0
16.6
17.9
19.1
19.3
Rail
35.5

46.1

39.8
36.3
37.5
39.4
37.1
Recreational Boatsd
2.7

2.9

2.6
2.7
2.8
2.9
2.9
Ships and Non-Recreational









Boats0
6.8

8.4

13.2
10.9
10.1
9.4
7.6
International Bunker Fuels?
11.7

9.5

8.4
8.7
9.0
10.0
10.1
Jet Fuel
184.2

189.3

157.6
166.0
171.8
172.3
177.8
Commercial Aircraft5
109.9

132.7

119.0
120.4
128.0
129.6
134.2
Military Aircraft
35.0

19.4

13.5
12.3
12.2
11.8
11.9
General Aviation Aircraft
39.4

37.3

25.1
33.4
31.5
30.9
31.7
International Bunker Fuels?
38.0

60.1

71.9
74.1
77.7
80.8
80.7
International Bunker Fuels from









Commercial Aviation
30.0

55.6

68.6
70.8
74.5
77.7
77.6
Aviation Gasoline
3.1

2.4

1.5
1.4
1.4
1.5
1.6
General Aviation Aircraft
3.1

2.4

1.5
1.4
1.4
1.5
1.6
Energy 3-27

-------
Residual Fuel Oil
22.6
19.3
4.2
12.9
16.5
14.0
14.7
Ships and Boats0
22.6
19.3
4.2
12.9
16.5
14.0
14.7
International Bunker Fueld
53.7
43.6
30.6
33.8
33.4
31.4
25.2
Natural GasJ
36.0
33.1
39.4
40.1
42.3
50.9
54.8
Passenger Cars
+
+
+
+
+
+
+
Light-Duty Trucks
+
+
+
+
+
+
+
Medium- and Heavy-Duty







Trucks
+
+
+
+
+
+
+
Buses
+
0.6
0.9
0.8
0.9
0.9
1.0
Pipeline11
36.0
32.4
38.5
39.2
41.3
49.9
53.7
LPGj
1.4
1.8
0.4
0.4
0.4
0.5
0.5
Passenger Cars
+
+
+
+
+
+
+
Light-Duty Trucks
0.2
0.3
0.1
0.1
0.1
0.1
0.1
Medium- and Heavy-Duty







Trucks0
1.1
1.3
0.3
0.3
0.3
0.3
0.3
Buses
0.1
0.1
+
0.1
0.1
0.1
0.1
Electricity1
3.0
4.7
4.3
4.2
4.3
4.7
4.7
Passenger Cars
+
+
0.5
0.6
0.8
1.2
1.4
Light-Duty Trucks
+
+
+
0.1
0.1
0.2
0.2
Buses
+
+
+
+
+
+
+
Rail
3.0
4.7
3.7
3.5
3.4
3.3
3.1
Totalk
1,472.2
1,863.4
1,723.5
1,764.1
1,786.8
1,821.2
1,821.9
Total (Including Bunkers)'
1,575.6
1,976.6
1,834.4
1,880.7
1,906.9
1,943.3
1,938.0
Biofuels-Ethanol'
4.1
21.6
74.2
76.9
77.7
78.6
78.7
Biofuels-Biodiesel'
+
0.9
14.1
19.6
18.7
17.9
17.1
Notes: This table does not include emissions from non-transportation mobile sources, such as agricultural equipment and
construction/mining equipment; it also does not include emissions associated with electricity consumption by pipelines or
lubricants used in transportation. In addition, this table does not include C02 emissions from U.S. Territories, since these are
covered in a separate chapter of the Inventory. Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a In 2011, FHWA changed its methods for estimating vehicle miles traveled (VMT) and related data. These methodological
changes included how vehicles are classified, moving from a system based on body-type to one that is based on wheelbase.
These changes were first incorporated for the 1990 through 2010 Inventory and apply to the 2007 through 2019 time
period. This resulted in large changes in VMT and fuel consumption data by vehicle class, thus leading to a shift in emissions
among on-road vehicle classes.
b Gasoline and diesel highway vehicle fuel consumption estimates are based on data from FHWA Highway Statistics Table MF-
27 and VM-1 (FHWA 1996 through 2019). Data from Table VM-1 is used to estimate the share of consumption between each
on-road vehicle class. These fuel consumption estimates are combined with estimates of fuel shares by vehicle type from
DOE's TEDB Annex Tables A.l through A.6 (DOE 1993 through 2018). TEDB data for 2019 has not been published yet,
therefore 2018 data are used as a proxy.
c Includes medium- and heavy-duty trucks over 8,500 lbs.
d In 2014, EPA incorporated the NONROAD2008 model into MOVES2014. The current Inventory uses the Nonroad component
of MOVES2014b for years 1999 through 2019.
0 Note that large year over year fluctuations in emission estimates partially reflect nature of data collection for these sources.
f Official estimates exclude emissions from the combustion of both aviation and marine international bunker fuels; however,
estimates including international bunker fuel-related emissions are presented for informational purposes.
5 Commercial aircraft, as modeled in FAA's Aviation Environmental Design Tool (AEDT), consists of passenger aircraft, cargo,
and other chartered flights.
h Pipelines reflect C02 emissions from natural gas-powered pipelines transporting natural gas.
' Ethanol and biodiesel estimates are presented for informational purposes only. See Section 3.11 of this chapter and the
estimates in Land Use, Land-Use Change, and Forestry (see Chapter 6), in line with IPCC methodological guidance and
UNFCCC reporting obligations, for more information on ethanol and biodiesel.
' Transportation sector natural gas and LPG consumption are based on data from EIA (2019b). Prior to the 1990 to 2015
Inventory, data from DOE TEDB were used to estimate each vehicle class's share of the total natural gas and LPG
consumption. Since TEDB does not include estimates for natural gas use by medium and heavy-duty trucks or LPG use by
3-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
passenger cars, EIA Alternative Fuel Vehicle Data (Browning 2017) is now used to determine each vehicle class's share of the
total natural gas and LPG consumption. These changes were first incorporated in the 1990 to 2016 Inventory and apply to
the 1990 to 2019 time period.
k Includes emissions from rail electricity.
1 Electricity consumption by passenger cars, light-duty trucks (SUVs), and buses is based on plug-in electric vehicle sales and
engine efficiency data, as outlined in Browning (2018a). In prior Inventory years, C02 emissions from electric vehicle
charging were allocated to the residential and commercial sectors. They are now allocated to the transportation sector.
These changes apply to the 2010 through 2019 time period.
Mobile Fossil Fuel Combustion CH4 and N2O Emissions
Mobile combustion includes emissions of CH4 and N20 from all transportation sources identified in the U.S.
Inventory with the exception of pipelines and electric locomotives;29 mobile sources also include non-
transportation sources such as construction/mining equipment, agricultural equipment, vehicles used off-road,
and other sources (e.g., snowmobiles, lawnmowers, etc.).30 Annex 3.2 includes a summary of all emissions from
both transportation and mobile sources. Table 3-14 and Table 3-15 provide mobile fossil fuel CH4 and N20 emission
estimates in MMT C02 Eq.31
Mobile combustion was responsible for a small portion of national CH4 emissions (0.4 percent) and was the fifth
largest source of national N20 emissions (4.4 percent). From 1990 to 2019, mobile source CH4 emissions declined
by 63 percent, to 2.4 MMT C02 Eq. (95 kt CH4), due largely to emissions control technologies employed in on-road
vehicles since the mid-1990s to reduce CO, NOx, NMVOC, and CH4 emissions. Mobile source emissions of N20
decreased by 60 percent from 1990 to 2019, to 18.0 MMT C02 Eq. (60 kt N20). Earlier generation emissions control
technologies initially resulted in higher N20 emissions, causing a 29 percent increase in N20 emissions from mobile
sources between 1990 and 1997. Improvements in later-generation emissions control technologies have reduced
N20 emissions, resulting in a 69 percent decrease in mobile source N20 emissions from 1997 to 2019 (Figure 3-17).
Overall, CH4 and N20 emissions were predominantly from gasoline-fueled passenger cars and light-duty trucks and
non-highway sources. See Annex 3.2 for data by vehicle mode and information on VMT and the share of new
vehicles (in VMT).
29	Emissions of CH4 from natural gas systems are reported separately. More information on the methodology used to calculate
these emissions are included in this chapter and Annex 3.4.
30	See the methodology sub-sections of the C02 from Fossil Fuel Combustion and CH4 and N20 from Mobile Combustion
sections of this chapter. Note that N20 and CH4 emissions are reported using different categories than C02. C02 emissions are
reported by end-use sector (Transportation, Industrial, Commercial, Residential, U.S. Territories), and generally adhere to a top-
down approach to estimating emissions. C02 emissions from non-transportation sources (e.g., lawn and garden equipment,
farm equipment, construction equipment) are allocated to their respective end-use sector (i.e., construction equipment C02
emissions are included in the Industrial end-use sector instead of the Transportation end-use sector). CH4 and N20 emissions
are reported using the "Mobile Combustion" category, which includes non-transportation mobile sources. CH4 and N20
emission estimates are bottom-up estimates, based on total activity (fuel use, VMT) and emissions factors by source and
technology type. These reporting schemes are in accordance with IPCC guidance. For informational purposes only, C02
emissions from non-transportation mobile sources are presented separately from their overall end-use sector in Annex 3.2.
31	See Annex 3.2 for a complete time series of emission estimates for 1990 through 2019.
Energy 3-29

-------
Figure 3-17: Mobile Source ChU and N2O Emissions
60
50
40
30
20
10
0
Table 3-14: ChU Emissions from Mobile Combustion (MMT CO2 Eq.)
Fuel Type/Vehicle Type3
1990

2005

2015
2016
2017
2018
2019
Gasoline On-Roadb
5.2

2.2

1.0
0.9
0.8
0.7
0.7
Passenger Cars
3.2

1.3

0.6
0.6
0.5
0.5
0.4
Light-Duty Trucks
1.7

0.8

0.2
0.2
0.2
0.2
0.2
Medium- and Heavy-Duty









Trucks and Buses
0.3

0.1

0.1
+
+
+
+
Motorcycles
+

+

+
+
+
+
+
Diesel On-Roadb
+

+

0.1
0.1
0.1
0.1
0.1
Passenger Cars
+

+

+
+
+
+
+
Light-Duty Trucks
+

+

+
+
+
+
+
Medium- and Heavy-Duty









Trucks and Buses
+

+

+
0.1
0.1
0.1
0.1
Alternative Fuel On-Road
+

0.2

0.2
0.2
0.2
0.2
0.2
Non-Roadc
1.2

1.5

1.4
1.4
1.4
1.4
1.4
Ships and Boats
0.4

0.4

0.4
0.4
0.4
0.4
0.4
Rail
0.1

0.1

0.1
0.1
0.1
0.1
0.1
Aircraft
0.1

0.1

+
+
+
+
+
Agricultural Equipment
0.1

0.2

0.1
0.1
0.1
0.1
0.1
Construction/Mining









Equipment6
0.1

0.2

0.2
0.2
0.2
0.2
0.2
Other'
0.4

0.6

0.6
0.6
0.6
0.6
0.6
Total
6.4

4.0

2.6
2.5
2.5
2.4
2.4
Notes: In 2011, FHWA changed its methods for estimating vehicle miles traveled (VMT) and related data. These
methodological changes included how vehicles are classified, moving from a system based on body type to one
that is based on wheelbase. These changes were first incorporated for the 1990 through 2010 Inventory and
apply to the 2007 through 2019 time period. This resulted in large changes in VMT and fuel consumption data
by vehicle class, thus leading to a shift in emissions among on-road vehicle classes. Totals may not sum due to
independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a See Annex 3.2 for definitions of on-road vehicle types.
b Gasoline and diesel highway vehicle mileage estimates are based on data from FHWA Highway Statistics Table
VM-1.
c Rail emissions do not include emissions from electric powered locomotives. Class II and Class III diesel
consumption data for 2014 to 2017 is estimated by applying the historical average fuel usage per carload factor
to the annual number of carloads.
3-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
d Includes equipment, such as tractors and combines, as well as fuel consumption from trucks that are used off-
road in agriculture.
0 Includes equipment, such as cranes, dumpers, and excavators, as well as fuel consumption from trucks that
are used off-road in construction.
f "Other" includes snowmobiles and other recreational equipment, logging equipment, lawn and garden
equipment, railroad equipment, airport equipment, commercial equipment, and industrial equipment, as well
as fuel consumption from trucks that are used off-road for commercial/industrial purposes.
Table 3-15: N2O Emissions from Mobile Combustion (MMT CO2 Eq.)
Fuel Type/Vehicle Type3
1990
2005
2015
2016
2017
2018
2019
Gasoline On-Roadb
37.5
31.8
11.6
10.2
8.7
7.3
6.2
Passenger Cars
24.1
17.3
8.0
7.0
6.0
5.1
4.3
Light-Duty Trucks
12.8
13.6
3.1
2.7
2.3
1.9
1.6
Medium- and Heavy-Duty







Trucks and Buses
0.5
0.9
0.4
0.4
0.3
0.3
0.2
Motorcycles
+
+
+
+
+
+
+
Diesel On-Roadb
0.2
0.3
2.1
2.4
2.6
2.8
3.0
Passenger Cars
+
+
+
0.1
0.1
0.1
0.1
Light-Duty Trucks
+
+
0.1
0.1
0.1
0.1
0.1
Medium- and Heavy-Duty







Trucks and Buses
0.2
0.3
2.0
2.2
2.5
2.7
2.8
Alternative Fuel On-Road
+
+
0.1
0.2
0.2
0.2
0.2
Non-Road
7.1
9.4
7.9
8.1
8.4
8.5
8.7
Ships and Boats
0.3
0.3
0.2
0.2
0.2
0.2
0.2
Railc
0.3
0.3
0.3
0.3
0.3
0.3
0.3
Aircraft
1.7
1.8
1.5
1.5
1.6
1.6
1.6
Agricultural Equipment
1.3
1.6
1.1
1.1
1.1
1.1
1.1
Construction/Mining







Equipment0
1.3
2.1
1.5
1.6
1.7
1.8
1.9
Other'
2.2
3.3
3.3
3.4
3.4
3.5
3.6
Total
44.7
41.6
21.7
20.8
19.8
18.8
18.0
Notes: In 2011, FHWA changed its methods for estimating vehicle miles traveled (VMT) and related data. These
methodological changes included how vehicles are classified, moving from a system based on body type to one that is
based on wheelbase. These changes were first incorporated for the 1990 through 2010 Inventory and apply to the
2007 through 2019 time period. This resulted in large changes in VMT and fuel consumption data by vehicle class,
thus leading to a shift in emissions among on-road vehicle classes. Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a See Annex 3.2 for definitions of on-road vehicle types.
b Gasoline and diesel highway vehicle mileage estimates are based on data from FHWA Highway Statistics Table VM-
1.
c Rail emissions do not include emissions from electric powered locomotives. Class II and Class III diesel consumption
data for 2014-2017 is estimated by applying the historical average fuel usage per carload factor to the annual number
of carloads.
d Includes equipment, such as tractors and combines, as well as fuel consumption from trucks that are used off-road
in agriculture.
0 Includes equipment, such as cranes, dumpers, and excavators, as well as fuel consumption from trucks that are used
off-road in construction.
f "Other" includes snowmobiles and other recreational equipment, logging equipment, lawn and garden equipment,
railroad equipment, airport equipment, commercial equipment, and industrial equipment, as well as fuel
consumption from trucks that are used off-road for commercial/industrial purposes.
Energy 3-31

-------
C02 from Fossil Fuel Combustion
Methodology
C02 emissions from fossil fuel combustion are estimated in line with a Tier 2 method described by the IPCC in the
2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2006) with some exceptions as discussed
below.32 A detailed description of the U.S. methodology is presented in Annex 2.1, and is characterized by the
following steps:
1.	Determine total fuel consumption by fuel type and sector. Total fossil fuel consumption for each year is
estimated by aggregating consumption data by end-use sector (e.g., commercial, industrial), primary fuel
type (e.g., coal, petroleum, gas), and secondary fuel category (e.g., motor gasoline, distillate fuel oil). Fuel
consumption data for the United States were obtained directly from the EIA of the U.S. Department of
Energy (DOE), primarily from the Monthly Energy Review (EIA 2020c). EIA data include fuel consumption
statistics from the 50 U.S. states and the District of Columbia, including tribal lands. The EIA does not
include territories in its national energy statistics, so fuel consumption data for territories were collected
separately from ElA's International Energy Statistics (EIA 2020e).33
For consistency of reporting, the IPCC has recommended that countries report energy data using the
International Energy Agency (IEA) reporting convention and/or IEA data. Data in the IEA format are
presented "top down"—that is, energy consumption for fuel types and categories are estimated from
energy production data (accounting for imports, exports, stock changes, and losses). The resulting
quantities are referred to as "apparent consumption." The data collected in the United States by EIA on
an annual basis and used in this Inventory are predominantly from mid-stream or conversion energy
consumers such as refiners and electric power generators. These annual surveys are supplemented with
end-use energy consumption surveys, such as the Manufacturing Energy Consumption Survey, that are
conducted on a periodic basis (every four years). These consumption datasets help inform the annual
surveys to arrive at the national total and sectoral breakdowns for that total.34
Also, note that U.S. fossil fuel energy statistics are generally presented using gross calorific values (GCV)
(i.e., higher heating values). Fuel consumption activity data presented here have not been adjusted to
correspond to international standards, which are to report energy statistics in terms of net calorific values
(NCV) (i.e., lower heating values).35
2.	Subtract uses accounted for in the Industrial Processes and Product Use chapter. Portions of the fuel
consumption data for seven fuel categories—coking coal, distillate fuel, industrial other coal, petroleum
coke, natural gas, residual fuel oil, and other oil—were reallocated to the Industrial Processes and Product
Use chapter, as they were consumed during non-energy-related industrial activity. To make these
adjustments, additional data were collected from AISI (2004 through 2020), Coffeyville (2012), U.S. Census
Bureau (2001 through 2011), EIA (2020a, 2020c, 2020d), USAA (2008 through 2020), USGS (1991 through
2017), (USGS 2019), USGS (2014 through 2020a), USGS (2014 through 2020b), USGS (1995 through 2013),
USGS (1995, 1998, 2000, 2001, 2002, 2007), USGS (2020a), USGS (1991 through 2015a), USGS (1991
32	The IPCC Tier 3B methodology is used for estimating emissions from commercial aircraft.
33	Fuel consumption by U.S. Territories (i.e., American Samoa, Guam, Puerto Rico, U.S. Virgin Islands, Wake Island, and other
U.S. Pacific Islands) is included in this report and contributed total emissions of 24.6 MMT C02 Eq. in 2019.
34	See IPCC Reference Approach for Estimating C02 Emissions from Fossil Fuel Combustion in Annex 4 for a comparison of U.S.
estimates using top-down and bottom-up approaches.
35	A crude convention to convert between gross and net calorific values is to multiply the heat content of solid and liquid fossil
fuels by 0.95 and gaseous fuels by 0.9 to account for the water content of the fuels. Biomass-based fuels in U.S. energy
statistics, however, are generally presented using net calorific values.
3-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
through 2017a), USGS (2014 through 2020a), USGS (1991 through 2015b), USGS (2020b), USGS (1991
through 2017).36
3.	Adjust for biofuels and petroleum denaturant. Fossil fuel consumption estimates are adjusted downward
to exclude fuels with biogenic origins and avoid double counting in petroleum data statistics. Carbon
dioxide emissions from ethanol added to motor gasoline and biodiesel added to diesel fuel are not
included specifically in summing energy sector totals. Net carbon fluxes from changes in biogenic carbon
reservoirs are accounted for in the estimates for LULUCF, therefore, fuel consumption estimates are
adjusted to remove ethanol and biodiesel.37 For the years 1993 through 2008, petroleum denaturant is
currently included in EIA statistics for both natural gasoline and finished motor gasoline. To avoid double
counting, petroleum denaturant is subtracted from finished motor gasoline for these years.38
4.	Adjust for exports ofC02. Since October 2000, the Dakota Gasification Plant has been exporting C02
produced in the coal gasification process to Canada by pipeline. Because this C02 is not emitted to the
atmosphere in the United States, the associated fossil fuel (lignite coal) that is gasified to create the
exported C02 is subtracted from EIA (2020d) coal consumption statistics that are used to calculate
greenhouse gas emissions from the Energy Sector. The associated fossil fuel is the total fossil fuel burned
at the plant with the C02 capture system multiplied by the fraction of the plant's total site-generated C02
that is recovered by the capture system. To make these adjustments, data for C02 exports were collected
from Environment and Climate Change Canada (2020). A discussion of the methodology used to estimate
the amount of C02 captured and exported by pipeline is presented in Annex 2.1.
5.	Adjust sectoral allocation of distillate fuel oil and motor gasoline. EPA conducted a separate bottom-up
analysis of transportation fuel consumption based on data from the Federal Highway Administration that
indicated that the amount of distillate and motor gasoline consumption allocated to the transportation
sector in the EIA statistics should be adjusted. Therefore, for these estimates, the transportation sector's
distillate fuel and motor gasoline consumption were adjusted to match the value obtained from the
bottom-up analysis. As the total distillate and motor gasoline consumption estimate from EIA are
considered to be accurate at the national level, the distillate and motor gasoline consumption totals for
the residential, commercial, and industrial sectors were adjusted proportionately. The data sources used
in the bottom-up analysis of transportation fuel consumption include AAR (2008 through 2018), Benson
(2002 through 2004), DOE (1993 through 2017), EIA (2007), EIA (1991 through 2019), EPA (2018), and
FHWA (1996 through 2018).39
6.	Adjust for fuels consumed for non-energy uses. U.S. aggregate energy statistics include consumption of
fossil fuels for non-energy purposes. These are fossil fuels that are manufactured into plastics, asphalt,
lubricants, or other products. Depending on the end-use, this can result in storage of some or all of the C
contained in the fuel for a period of time. As the emission pathways of C used for non-energy purposes
are vastly different than fuel combustion (since the C in these fuels ends up in products instead of being
combusted), these emissions are estimated separately in Section 3.2 - Carbon Emitted and Stored in
Products from Non-Energy Uses of Fossil Fuels. Therefore, the amount of fuels used for non-energy
purposes was subtracted from total fuel consumption. Data on non-fuel consumption were provided by
EIA (2020c).
7.	Subtract consumption of international bunker fuels. According to the UNFCCC reporting guidelines
emissions from international transport activities, or bunker fuels, should not be included in national
totals. U.S. energy consumption statistics include these bunker fuels (e.g., distillate fuel oil, residual fuel
oil, and jet fuel) as part of consumption by the transportation end-use sector, however, so emissions from
36	See sections on Iron and Steel Production and Metallurgical Coke Production, Ammonia Production and Urea Consumption,
Petrochemical Production, Titanium Dioxide Production, Ferroalloy Production, Aluminum Production, and Silicon Carbide
Production and Consumption in the Industrial Processes and Product Use chapter.
37	Natural gas energy statistics from EIA (2020g) are already adjusted downward to account for biogas in natural gas.
38	These adjustments are explained in greater detail in Annex 2.1.
39	Bottom-up gasoline and diesel highway vehicle fuel consumption estimates are based on data from FHWA Highway Statistics
Table MF-21, MF-27, and VM-1 (FHWA 1996 through 2019).
Energy 3-33

-------
international transport activities were calculated separately following the same procedures used to
calculate emissions from consumption of all fossil fuels (i.e., estimation of consumption, and
determination of Carbon content).40 The Office of the Under Secretary of Defense (Installations and
Environment) and the Defense Logistics Agency Energy (DLA Energy) of the U.S. Department of Defense
(DoD) (DLA Energy 2020) supplied data on military jet fuel and marine fuel use. Commercial jet fuel use
was obtained from FAA (2021); residual and distillate fuel use for civilian marine bunkers was obtained
from DOC (1991 through 2019) for 1990 through 2001 and 2007 through 2018, and DHS (2008) for 2003
through 2006.41 Consumption of these fuels was subtracted from the corresponding fuels totals in the
transportation end-use sector. Estimates of international bunker fuel emissions for the United States are
discussed in detail in Section 3.10 - International Bunker Fuels.
8.	Determine the total Carbon content of fuels consumed. Total C was estimated by multiplying the amount
of fuel consumed by the amount of C in each fuel. This total C estimate defines the maximum amount of C
that could potentially be released to the atmosphere if all of the C in each fuel was converted to C02. A
discussion of the methodology and sources used to develop the C content coefficients are presented in
Annexes 2.1 and 2.2.
9.	Estimate C02 Emissions. Total C02 emissions are the product of the adjusted energy consumption (from
the previous methodology steps 1 through 6), the Carbon content of the fuels consumed, and the fraction
of C that is oxidized. The fraction oxidized was assumed to be 100 percent for petroleum, coal, and
natural gas based on guidance in IPCC (2006) (see Annex 2.1). Carbon emissions were multiplied by the
molecular-to-atomic weight ratio of C02 to C (44/12) to obtain total C02 emitted from fossil fuel
combustion in million metric tons (MMT).
10.	Allocate transportation emissions by vehicle type. This report provides a more detailed accounting of
emissions from transportation because it is such a large consumer of fossil fuels in the United States. For
fuel types other than jet fuel, fuel consumption data by vehicle type and transportation mode were used
to allocate emissions by fuel type calculated for the transportation end-use sector. Heat contents and
densities were obtained from EIA (2020c) and USAF (1998).42
•	For on-road vehicles, annual estimates of combined motor gasoline and diesel fuel consumption by
vehicle category were obtained from FHWA (1996 through 2019); for each vehicle category, the
percent gasoline, diesel, and other (e.g., CNG, LPG) fuel consumption are estimated using data from
DOE (1993 through 2018). 43 44
•	For non-road vehicles, activity data were obtained from AAR (2008 through 2019), APTA (2007
through 2018), APTA (2006), BEA (2020), Benson (2002 through 2004), DLA Energy (2019), DOC (1991
through 2019), DOE (1993 through 2017), DOT (1991 through 2019), EIA (2009a), EIA (2020c), EIA
40	See International Bunker Fuels section in this chapter for a more detailed discussion.
41	Data for 2002 were interpolated due to inconsistencies in reported fuel consumption data.
42	For a more detailed description of the data sources used for the analysis of the transportation end use sector see the Mobile
Combustion (excluding C02) and International Bunker Fuels sections of the Energy chapter, Annex 3.2, and Annex 3.8,
respectively.
43	Data from FHWA's Table VM-1 is used to estimate the share of fuel consumption between each on-road vehicle class. These
fuel consumption estimates are combined with estimates of fuel shares by vehicle type from DOE's TEDB Annex Tables A.l
through A.6 (DOE 1993 through 2017). In 2011, FHWA changed its methods for estimating data in the VM-1 table. These
methodological changes included how vehicles are classified, moving from a system based on body-type to one that is based on
wheelbase. These changes were first incorporated for the 1990 through 2010 Inventory and apply to the time period from 2007
through 2019. This resulted in large changes in VMT and fuel consumption data by vehicle class, thus leading to a shift in
emissions among on-road vehicle classes.
44	Transportation sector natural gas and LPG consumption are based on data from EIA (2020g). In previous Inventory years,
data from DOE (1993 through 2017) TEDB was used to estimate each vehicle class's share of the total natural gas and LPG
consumption. Since TEDB does not include estimates for natural gas use by medium- and heavy-duty trucks or LPG use by
passenger cars, EIA Alternative Fuel Vehicle Data (Browning 2017) is now used to determine each vehicle class's share of the
total natural gas and LPG consumption. These changes were first incorporated in the previous Inventory and apply to the time
period from 1990 to 2015.
3-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
(2019f), EIA (1991 through 2019), EPA (2018),45 and Gaffney (2007).
• For jet fuel used by aircraft, C02 emissions from commercial aircraft were developed by the U.S.
Federal Aviation Administration (FAA) using a Tier 3B methodology, consistent IPCC (2006) (see
Annex 3.3). Carbon dioxide emissions from other aircraft were calculated directly based on reported
consumption of fuel as reported by EIA. Allocation to domestic military uses was made using DoD
data (see Annex 3.8). General aviation jet fuel consumption is calculated as the remainder of total jet
fuel use (as determined by EIA) nets all other jet fuel use as determined by FAA and DoD. For more
information, see Annex 3.2.
Box 3-4: Carbon Intensity of U.S. Energy Consumption
The amount of C emitted from the combustion of fossil fuels is dependent upon the carbon content of the fuel
and the fraction of that C that is oxidized. Fossil fuels vary in their average carbon content, ranging from about
53 MMT C02 Eq./QBtu for natural gas to upwards of 95 MMT C02 Eq./QBtu for coal and petroleum coke (see
Tables A-42 and A-43 in Annex 2.1 for carbon contents of all fuels). In general, the carbon content per unit of
energy of fossil fuels is the highest for coal products, followed by petroleum, and then natural gas. The overall
carbon intensity of the U.S. economy is thus dependent upon the quantity and combination of fuels and other
energy sources employed to meet demand.
Table 3-16 provides a time series of the carbon intensity of direct emissions for each sector of the U.S. economy.
The time series incorporates only the energy from the direct combustion of fossil fuels in each sector. For
example, the carbon intensity for the residential sector does not include the energy from or emissions related to
the use of electricity for lighting, as it is instead allocated to the electric power sector. For the purposes of
maintaining the focus of this section, renewable energy and nuclear energy are not included in the energy totals
used in Table 3-16 in order to focus attention on fossil fuel combustion as detailed in this chapter. Looking only
at this direct consumption of fossil fuels, the residential sector exhibited the lowest carbon intensity, which is
related to the large percentage of its energy derived from natural gas for heating. The carbon intensity of the
commercial sector has predominantly declined since 1990 as commercial businesses shift away from petroleum
to natural gas. The industrial sector was more dependent on petroleum and coal than either the residential or
commercial sectors, and thus had higher C intensities over this period. The Carbon intensity of the
transportation sector was closely related to the Carbon content of petroleum products (e.g., motor gasoline and
jet fuel, both around 70 MMT C02 Eq./QBtu), which were the primary sources of energy. Lastly, the electric
power sector had the highest Carbon intensity due to its heavy reliance on coal for generating electricity.
Table 3-16: Carbon Intensity from Direct Fossil Fuel Combustion by Sector (MMT CO2
Eq./QBtu)
Sector
1990

2005

2015
2016
2017
2018
2019
Residential3
57.4

56.8

55.5
55.2
55.1
55.3
55.1
Commercial3
59.7

57.8

57.1
56.7
56.5
56.0
56.1
Industrial3
64.5

64.6

61.4
61.0
60.8
60.5
60.3
Transportation3
71.1

71.5

71.1
71.1
71.2
71.0
71.0
Electric Powerb
87.3

85.8

78.1
76.8
77.3
75.5
73.0
U.S. Territories0
72.3

72.6

72.0
71.0
71.3
71.3
71.3
All Sectors0
73.1

73.6

69.6
69.2
69.1
68.3
67.3
Note: Excludes non-energy fuel use emissions and consumption.
a Does not include electricity or renewable energy consumption.
b Does not include electricity produced using nuclear or renewable energy.
c Does not include nuclear or renewable energy consumption.
45 In 2014, EPA incorporated the NONROAD2008 model into MOVES2014. The current Inventory uses the Nonroad component
of MOVES2014b for years 1999 through 2019.
Energy 3-35

-------
For the time period of 1990 through about 2008, the carbon intensity of U.S. energy consumption was fairly
constant, as the proportion of fossil fuels used by the individual sectors did not change significantly over that
time. Starting in 2008 the carbon intensity has decreased, reflecting the shift from coal to natural gas in the
electric power sector during that time period. Per capita energy consumption fluctuated little from 1990 to
2007, but then started decreasing after 2007 and, in 2019, was approximately 9.8 percent below levels in 1990
(see Figure 3-18). To differentiate these estimates from those of Table 3-16, the carbon intensity trend shown in
Figure 3-18 and described below includes nuclear and renewable energy EIA data to provide a comprehensive
economy-wide picture of energy consumption. Due to a general shift from a manufacturing-based economy to a
service-based economy, as well as overall increases in efficiency, energy consumption and energy-related C02
emissions per dollar of gross domestic product (GDP) have both declined since 1990 (BEA 2018).
Figure 3-18: U.S. Energy Consumption and Energy-Related CO2 Emissions Per Capita and
Per Dollar GDP
110
C02/Energy Consumption (green)
100
Energy Consumption/capita (dark blue)
(/)
>
x
 o
O iH
^r
1—1
i-H f\l
o o
m
1—1
T—I
7—1
Carbon intensity estimates were developed using nuclear and renewable energy data from EIA (2020c), EPA
(2010), and fossil fuel consumption data as discussed above and presented in Annex 2.1.
Uncertainty and Time-Series Consistency
For estimates of C02 from fossil fuel combustion, the amount of C02 emitted is directly related to the amount of
fuel consumed, the fraction of the fuel that is oxidized, and the carbon content of the fuel. Therefore, a careful
accounting of fossil fuel consumption by fuel type, average carbon contents of fossil fuels consumed, and
production of fossil fuel-based products with long-term carbon storage should yield an accurate estimate of C02
emissions.
Nevertheless, there are uncertainties in the consumption data, carbon content of fuels and products, and carbon
oxidation efficiencies. For example, given the same primary fuel type (e.g., coal, petroleum, or natural gas), the
amount of carbon contained in the fuel per unit of useful energy can vary. For the United States, however, the
impact of these uncertainties on overall C02 emission estimates is believed to be relatively small. See, for example,
Marland and Pippin (1990). See also Annex 2.2 for a discussion of uncertainties associated with fuel carbon
contents. Recent updates to carbon factors for natural gas and coal utilized the same approach as previous
Inventories with updated recent data, therefore, the uncertainty estimates around carbon contents of the
different fuels as outlined in Annex 2.2 were not impacted and the historic uncertainty ranges still apply.
Although national statistics of total fossil fuel and other energy consumption are relatively accurate, the allocation
of this consumption to individual end-use sectors (i.e., residential, commercial, industrial, and transportation) is
3-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
less certain. For example, for some fuels the sectoral allocations are based on price rates (i.e., tariffs), but a
commercial establishment may be able to negotiate an industrial rate or a small industrial establishment may end
up paying an industrial rate, leading to a misallocation of emissions. Also, the deregulation of the natural gas
industry and the more recent deregulation of the electric power industry have likely led to some minor challenges
in collecting accurate energy statistics as firms in these industries have undergone significant restructuring.
To calculate the total C02 emission estimate from energy-related fossil fuel combustion, the amount of fuel used in
non-energy production processes were subtracted from the total fossil fuel consumption. The amount of C02
emissions resulting from non-energy related fossil fuel use has been calculated separately and reported in the
Carbon Emitted from Non-Energy Uses of Fossil Fuels section of this report (Section 3.2). These factors all
contribute to the uncertainty in the C02 estimates. Detailed discussions on the uncertainties associated with C
emitted from Non-Energy Uses of Fossil Fuels can be found within that section of this chapter.
Various sources of uncertainty surround the estimation of emissions from international bunker fuels, which are
subtracted from the U.S. totals (see the detailed discussions on these uncertainties provided in Section 3.10 -
International Bunker Fuels). Another source of uncertainty is fuel consumption by U.S. Territories. The United
States does not collect energy statistics for its territories at the same level of detail as for the fifty states and the
District of Columbia. Therefore, estimating both emissions and bunker fuel consumption by these territories is
difficult.
Uncertainties in the emission estimates presented above also result from the data used to allocate C02 emissions
from the transportation end-use sector to individual vehicle types and transport modes. In many cases, bottom-up
estimates of fuel consumption by vehicle type do not match aggregate fuel-type estimates from EIA. Further
research is planned to improve the allocation into detailed transportation end-use sector emissions.
The uncertainty analysis was performed by primary fuel type for each end-use sector, using the IPCC-
recommended Approach 2 uncertainty estimation methodology, Monte Carlo Stochastic Simulation technique,
with @RISK software. For this uncertainty estimation, the inventory estimation model for C02 from fossil fuel
combustion was integrated with the relevant variables from the inventory estimation model for International
Bunker Fuels, to realistically characterize the interaction (or endogenous correlation) between the variables of
these two models. About 170 input variables were modeled for C02 from energy-related Fossil Fuel Combustion
(including about 20 for non-energy fuel consumption and about 20 for International Bunker Fuels).
In developing the uncertainty estimation model, uniform distributions were assumed for all activity-related input
variables and emission factors, based on the SAIC/EIA (2001) report.46 Triangular distributions were assigned for
the oxidization factors (or combustion efficiencies). The uncertainty ranges were assigned to the input variables
based on the data reported in SAIC/EIA (2001) and on conversations with various agency personnel.47
The uncertainty ranges for the activity-related input variables were typically asymmetric around their inventory
estimates; the uncertainty ranges for the emissions factors were symmetric. Bias (or systematic uncertainties)
associated with these variables accounted for much of the uncertainties associated with these variables (SAIC/EIA
2001).48 For purposes of this uncertainty analysis, each input variable was simulated 10,000 times through Monte
46	SAIC/EIA (2001) characterizes the underlying probability density function for the input variables as a combination of uniform
and normal distributions (the former to represent the bias component and the latter to represent the random component).
However, for purposes of the current uncertainty analysis, it was determined that uniform distribution was more appropriate to
characterize the probability density function underlying each of these variables.
47	In the SAIC/EIA (2001) report, the quantitative uncertainty estimates were developed for each of the three major fossil fuels
used within each end-use sector; the variations within the sub-fuel types within each end-use sector were not modeled.
However, for purposes of assigning uncertainty estimates to the sub-fuel type categories within each end-use sector in the
current uncertainty analysis, SAIC/EIA (2001)-reported uncertainty estimates were extrapolated.
48	Although, in general, random uncertainties are the main focus of statistical uncertainty analysis, when the uncertainty
estimates are elicited from experts, their estimates include both random and systematic uncertainties. Hence, both these types
of uncertainties are represented in this uncertainty analysis.
Energy 3-37

-------
Carlo sampling.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-17. Fossil fuel
combustion C02 emissions in 2019 were estimated to be between 4,757.7 and 5,073.0 MMT C02 Eq. at a 95
percent confidence level. This indicates a range of 2 percent below to 4 percent above the 2019 emission estimate
of 4,856.7 MMT C02 Eq.
Table 3-17: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Energy-
Related Fossil Fuel Combustion by Fuel Type and Sector (MMT CO2 Eq. and Percent)
2019 Emission Estimate Uncertainty Range Relative to Emission Estimate3
Fuel/Sector	(MMTCQ2 Eq.)	(MMTCQ2 Eq.)	(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Coalb
1,027.1
991.9
1,125.1
-3%
10%
Residential
NO
NO
NO
NO
NO
Commercial
1.6
1.5
1.8
-5%
15%
Industrial
49.5
47.1
57.2
-5%
16%
Transportation
NO
NO
NO
NO
NO
Electric Power
973.5
935.8
1,068.4
-4%
10%
U.S. Territories
2.5
2.2
3.0
-12%
19%
Natural Gasb
1,644.6
1,625.8
1,720.0
-1%
5%
Residential
275.3
267.5
294.6
-3%
7%
Commercial
192.8
187.4
206.3
-3%
7%
Industrial
503.3
486.5
540.5
-3%
7%
Transportation
54.8
53.2
58.6
-3%
7%
Electric Power
616.0
598.2
647.4
-3%
5%
U.S. Territories
2.5
2.2
3.0
-12%
17%
Petroleumb
2,184.6
2,054.4
2,313.2
-6%
6%
Residential
61.5
58.0
64.8
-6%
5%
Commercial
55.3
52.3
58.1
-5%
5%
Industrial
269.7
215.7
324.3
-20%
20%
Transportation
1,762.5
1,649.9
1,873.9
-6%
6%
Electric Power
16.2
15.4
17.5
-5%
8%
U.S. Territories
19.5
18.0
21.7
-8%
11%
Total (excluding Geothermal)b
4,856.3
4,757.1
5,072.4
-2%
4%
Geothermal
0.4
NE
NE
NE
NE
Electric Power
0.4
NE
NE
NE
NE
Total (including Geothermal)b'c
4,856.7
4,757.7
5,073.0
-2%
4%
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
NE (Not Estimated)
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
b The low and high estimates for total emissions were calculated separately through simulations and, hence, the low and
high emission estimates for the sub-source categories do not sum to total emissions.
c Geothermal emissions added for reporting purposes, but an uncertainty analysis was not performed for C02 emissions
from geothermal production.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above. As discussed in Annex 5, data are unavailable to include estimates of C02 emissions from any liquid
3-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
fuel used in pipeline transport or non-hazardous industrial waste incineration, but those emissions are assumed to
be insignificant.
QA/QC and Verification
In order to ensure the quality of the C02 emission estimates from fossil fuel combustion, general (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved
checks specifically focusing on the activity data and methodology used for estimating C02 emissions from fossil fuel
combustion in the United States. Emission totals for the different sectors and fuels were compared and trends
were investigated to determine whether any corrective actions were needed. Minor corrective actions were taken.
The UNFCCC reporting guidelines require countries to complete a "top-down" reference approach for estimating
C02 emissions from fossil fuel combustion in addition to their "bottom-up" sectoral methodology. The reference
approach (detailed in Annex 4) uses alternative methodologies and different data sources than those contained in
this section of the report. The reference approach estimates fossil fuel consumption by adjusting national
aggregate fuel production data for imports, exports, and stock changes rather than relying on end-user
consumption surveys. The reference approach assumes that once carbon-based fuels are brought into a national
economy, they are either saved in some way (e.g., stored in products, kept in fuel stocks, or left unoxidized in ash)
or combusted, and therefore the carbon in them is oxidized and released into the atmosphere. In the reference
approach, accounting for actual consumption of fuels at the sectoral or sub-national level is not required. One
difference between the two approaches is that emissions from carbon that was not stored during non-energy use
of fuels are subtracted from the sectoral approach and reported separately (see Section 3.2). These emissions,
however, are not subtracted in the reference approach. As a result, the reference approach emission estimates are
comparable to those of the sectoral approach, with the exception that the Non-Energy Use (NEU) source category
emissions are included in the reference approach (see Annex 4 for more details).
Recalculations Discussion
Several updates to activity data and emission factors lead to recalculations of previous year results. The major
updates are as follows:
•	EIA (2020c) updated energy consumption statistics across the time series relative to the previous
Inventory. As a result of revised natural gas heat contents, EIA updated natural gas consumption in the
residential, commercial, and industrial sectors for 2018. Approximate heat rates for electricity and the
heat content of electricity were revised for natural gas and noncombustible renewable energy, which
impacted electric power energy consumption by sector. EIA also revised sector allocations for distillate
fuel oil, residual fuel oil, and kerosene for 2018, and for propane for 2010 through 2012, 2014, 2017, and
2018, which impacted LPG by sector. EIA revised product supplied totals for crude oil and petroleum
products, which impacted the nonfuel sequestration statistics, particularly for lubricants for 2018 and LPG
for 2010 through 2018 relative to the previous Inventory. This resulted in a slight decrease in energy used
in the industrial sector.
•	To align with ElA's methodology for calculating motor gasoline consumption, petroleum denaturant
adjustments to motor gasoline consumption for the period 1993 through 2008 were corrected. This
resulted in an average annual decrease of 6.2 TBtu in motor gasoline consumption for the period 1993
through 2008, which led to a decrease in emissions from gasoline consumption in those years because
denaturant emissions were previously being double counted.
•	Newly published U.S. Territories data from EIA (2020e) was integrated, which impacted total estimates for
U.S. Territories across the time series. This resulted in the following observed changes:
o average annual decrease of 0.3 MMT C02 Eq. (21.3 percent) in coal use across the time series;
o decrease of 0.01 MMT C02 Eq. (0.48 percent) in natural gas use across the time series; and
o decrease in petroleum use from 1990 through 1999, increase in petroleum use from 2000
through 2008, then a decrease from 2009 through 2018, resulting in an average annual decrease
of 5.8 MMT CQ2 eq. (17.9 percent) in petroleum use across the time series.
Energy 3-39

-------
•	Updated MECS data for 2018 resulted in an increase in natural gas used in non-energy use. This resulted
in a decrease in natural gas used in the industrial sector as part of fossil fuel combustion estimates. The
updates mainly impacted years 2014-2018. See Section 3.2 for more details on NEU emissions and
adjustments.
•	EPA (2020c) revised distillate fuel oil and motor gasoline carbon contents, which impacted petroleum
consumption in the transportation, residential, commercial, and industrial sectors. The combined effect of
both the diesel fuel and gasoline emission factor update was an increase in emissions early in the time
series and then decreases in emissions in more recent years. For years 1990 through 2005, the average
annual increase in total emissions was approximately 7 MMT C02 (0.1 percent of emissions). For the years
2006 to 2018 the average annual decrease in total emissions is about 5 MMT C02 (less than 0.1 percent of
emissions).
•	EPA also revised HGL C contents to align with ElA's revised heat contents and HGL fuel type categorization
(EIA 2020c; ICF 2020). A discussion of the methodology used to develop the C content coefficients is
presented in Annex 2.2. This resulted in an average annual increase of 3.0 percent in the weighted
industrial HGL C contents.
•	To account for coal consumed during the production of coke oven gas (COG) and blast furnace gas (BFG)
for energy purposes (e.g., as an input to the natural gas distribution system), consumption of COG and
BFG was included in industrial coal consumption estimates in the energy sector, in alignment with ElA's
methodology (EIA 2020c). Previously, COG and BFG consumption that enters the natural gas distribution
system was removed from industrial natural gas consumption estimates in the energy sector. These
adjustments are explained in greater detail in Annex 2.1. This resulted in an average annual increase of
1.5 TBtu (48 percent) in coal use between 1990 through 1992, no change between 1993 through 1999, an
average annual decrease of 0.5 TBtu (146 percent) between 2000 and 2001, and no change from 2002
forward.
•	The Dakota Gasification Plant uses a coal gasification process that produces synthetic natural gas (SNG)
from lignite coal. Coal consumption at this plant is included in ElA's statistics for industrial coal
consumption, which is used to estimate C02 emissions from coal combustion in the U.S. Inventory.
Previously, coal consumption for the production of SNG was subtracted from industrial coal consumption
statistics. However, SNG is not included in industrial natural gas consumption data in ElA's MER and
rather, SNG is accounted for in its primary energy category (e.g., gasification of coal). To account for SNG
from coal gasification, the adjustment to industrial coal consumption to subtract the quantity of SNG
produced was removed. These adjustments are explained in greater detail in Annex 2.1. This resulted in
an average annual increase in coal use across the time series of 31.9 Tbtu (3 percent).
•	The Dakota Gasification Plant also produces C02 as a byproduct. A fraction of the plant's total site-
generated C02 that is captured by the plant's C02 capture system is exported by pipeline to Canada. The
remainder of the byproduct C02 is emitted to the atmosphere. Because the exported C02 is not emitted
to the atmosphere in the United States, the amount of associated fossil fuel (lignite coal) that is gasified to
create the exported C02 is subtracted from the EIA industrial coal consumption statistics used in the
Inventory, so that the amount of C02 exported is not included in the reported greenhouse gas emissions
from the Energy Sector. Previously, the amount of C02 captured and exported by pipeline annually was
estimated from publicly available data for plant operations, including historical C02 export data and the
publicly reported transport (C02 gas compressor) capacity of the C02 pipeline, assuming that the C02
pipeline operates at 100 percent of the pipeline's transport capacity. To ensure consistency in reporting
between the Inventory and the Canadian National Greenhouse Gas Inventory, the amount of associated
fossil fuel (lignite coal) that is gasified to create the exported C02 has been revised to align with the
Canadian National Greenhouse Gas Inventory (Environment and Climate Change Canada 2020). These
adjustments are explained in greater detail in Annex 2.1. This resulted in an average decrease of 0.7 MMT
C02 Eq. (10 percent) in the amount of C02 exported each year between 2000 and 2018 and therefore an
increase in coal use across the time series.
All of the revisions discussed above resulted in the following impacts on emissions over time by fuel type:
3-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Coal emissions increased by an average annual amount of 2.8 MMT C02 Eq. (0.2 percent increase of
emissions from coal) across the entire time series. This is primarily due to the update to the adjustment
for COG and BFG and the change in C02 export data.
•	There was a slight average annual increase in natural gas emissions from 1990 to 2006 of 0.2 MMT C02
Eq. (less than 0.1 percent of natural gas emissions). This is mainly due to the removal of the COG and BFG
adjustment from industrial natural gas consumption. There was a bigger average annual decrease in
emissions of 5.0 MMT C02 Eq. (0.3 percent) from 2007 to 2018. The decrease is much larger in the latter
years due to the update of the 2018 MECS data, which increases natural gas use as NEU in the industrial
sector.
•	Petroleum emissions decreased by an average annual amount of 15.3 MMT C02 Eq. (0.7 percent of
petroleum emissions) from 1990 to 1999, which is mainly due to decreased emissions in the industrial
sector as a result of the update in the weighted industrial HGL C contents and the decrease in petroleum
use from the updated data for U.S. Territories.
•	Petroleum emissions increased by an average annual amount of 14.5 MMT C02 Eq. (0.6 percent) from
2000 to 2007. This is mainly due to an increase in petroleum emissions in U.S. Territories from the newly
integrated data and increased emissions in the Transportation sector due to changes in accounting for
denaturants and updates in the distillate fuel oil and motor gasoline emissions factors.
•	Finally, petroleum emissions decreased at the end of the time series by an average annual amount of 15.8
MMT C02 Eq. (0.7 percent) from 2008 to 2018. This is mainly due to the decrease in petroleum use from
the newly integrated data for U.S. Territories and decreases across the other sectors based on updated
gasoline and diesel fuel emission factors.
Overall, these changes resulted in an average annual decrease of 6.4 MMT C02 Eq. (0.1 percent) in C02 emissions
from fossil fuel combustion for the period 1990 through 2018, relative to the previous Inventory. However, there
were bigger absolute changes across the time series as discussed above by fuel type. The changes in petroleum
emissions drive the overall change in emissions from the recalculations across time.
Planned Improvements
To reduce uncertainty of C02 from fossil fuel combustion estimates for U.S. Territories, further expert elicitation
may be conducted to better quantify the total uncertainty associated with emissions from U.S. Territories.
The availability of facility-level combustion emissions through EPA's GHGRP will continue to be examined to help
better characterize the industrial sector's energy consumption in the United States and further classify total
industrial sector fossil fuel combustion emissions by business establishments according to industrial economic
activity type. Most methodologies used in EPA's GHGRP are consistent with IPCC methodologies, though for EPA's
GHGRP, facilities collect detailed information specific to their operations according to detailed measurement
standards, which may differ with the more aggregated data collected for the Inventory to estimate total, national
U.S. emissions. In addition, and unlike the reporting requirements for this chapter under the UNFCCC reporting
guidelines, some facility-level fuel combustion emissions reported under the GHGRP may also include industrial
process emissions.49 In line with UNFCCC reporting guidelines, fuel combustion emissions are included in this
chapter, while process emissions are included in the Industrial Processes and Product Use chapter of this report. In
examining data from EPA's GHGRP that would be useful to improve the emission estimates for the C02 from fossil
fuel combustion category, particular attention will also be made to ensure time-series consistency, as the facility-
level reporting data from EPA's GHGRP are not available for all inventory years as reported in this Inventory.
Additional analyses will be conducted to align reported facility-level fuel types and IPCC fuel types per the national
energy statistics. For example, efforts will be taken to incorporate updated industrial fuel consumption data from
ElA's Manufacturing Energy Consumption Survey (MECS), with updated data for 2018. Additional work will look at
C02 emissions from biomass to ensure they are separated in the facility-1 eve I reported data and maintaining
consistency with national energy statistics provided by EIA. In implementing improvements and integration of data
49 See .
Energy 3-41

-------
from EPA's GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories will
continue to be relied upon.50
An ongoing planned improvement is to develop improved estimates of domestic waterborne fuel consumption.
The Inventory estimates for residual and distillate fuel used by ships and boats is based in part on data on bunker
fuel use from the U.S. Department of Commerce. Domestic fuel consumption is estimated by subtracting fuel sold
for international use from the total sold in the United States. It may be possible to more accurately estimate
domestic fuel use and emissions by using detailed data on marine ship activity. The feasibility of using domestic
marine activity data to improve the estimates will continue to be investigated.
EPA is also evaluating the methods used to adjust for conversion of fuels and exports of C02. EPA is exploring the
approach used to account for C02 transport, injection, and geologic storage, as part of this there may be changes
made to accounting for C02 exports. EPA is also exploring the data provided by EIA in terms of tracking
supplemental natural gas which may impact the treatment of adjustments for synthetic fuels.
CH4and N20 from Stationary Combustion
Methodology
Methane and N20 emissions from stationary combustion were estimated by multiplying fossil fuel and wood
consumption data by emission factors (by sector and fuel type for industrial, residential, commercial, and U.S.
Territories; and by fuel and technology type for the electric power sector). The electric power sector utilizes a Tier
2 methodology, whereas all other sectors utilize a Tier 1 methodology. The activity data and emission factors used
are described in the following subsections.
More detailed information on the methodology for calculating emissions from stationary combustion, including
emission factors and activity data, is provided in Annex 3.1.
Industrial, Residential, Commercial, and U.S. Territories
National coal, natural gas, fuel oil, and wood consumption data were grouped by sector: industrial, commercial,
residential, and U.S. Territories. For the CH4 and N20 emission estimates, consumption data for each fuel were
obtained from ElA's Monthly Energy Review (EIA 2020a). Because the United States does not include territories in
its national energy statistics, fuel consumption data for territories were provided separately by ElA's International
Energy Statistics (EIA 2020b).51 Fuel consumption for the industrial sector was adjusted to subtract out mobile
source construction and agricultural use, which is reported under mobile sources. Construction and agricultural
mobile source fuel use was obtained from EPA (2019) and FHWA (1996 through 2019). Estimates for wood biomass
consumption for fuel combustion do not include municipal solid waste, tires, etc., that are reported as biomass by
EIA. Non-C02 emissions from combustion of the biogenic portion of municipal solid waste and tires is included
under waste incineration (Section 3.2). Estimates for natural gas combustion do not include biogas, and therefore
non-C02 emissions from biogas are not included (see the Planned Improvements section, below). Tier 1 default
emission factors for the industrial, commercial, and residential end-use sectors were provided by the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories (IPCC 2006). U.S. Territories' emission factors were estimated
using the U.S. emission factors for the primary sector in which each fuel was combusted.
50	See .
51	U.S. Territories data also include combustion from mobile activities because data to allocate territories' energy use were
unavailable. For this reason, CH4 and N20 emissions from combustion by U.S. Territories are only included in the stationary
combustion totals.
3-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Electric Power Sector
The electric power sector uses a Tier 2 emission estimation methodology as fuel consumption for the electric
power sector by control-technology type was based on EPA's Acid Rain Program Dataset (EPA 2021). Total fuel
consumption in the electric power sector from EIA (2020a) was apportioned to each combustion technology type
and fuel combination using a ratio of fuel consumption by technology type derived from EPA (2020a) data. The
combustion technology and fuel use data by facility obtained from EPA (2020a) were only available from 1996 to
2019, so the consumption estimates from 1990 to 1995 were estimated by applying the 1996 consumption ratio by
combustion technology type from EPA (2020a) to the total EIA (2020a) consumption for each year from 1990 to
1995.
Emissions were estimated by multiplying fossil fuel and wood consumption by technology-, fuel-, and country-
specific Tier 2 emission factors. The Tier 2 emission factors used are based in part on emission factors published by
EPA, and EPA's Compilation of Air Pollutant Emission Factors, AP-42 (EPA 1997) for coal wall-fired boilers, residual
fuel oil, diesel oil and wood boilers, natural gas-fired turbines, and combined cycle natural gas units.52
Uncertainty and Time-Series Consistency
Methane emission estimates from stationary sources exhibit high uncertainty, primarily due to difficulties in
calculating emissions from wood combustion (i.e., fireplaces and wood stoves). The estimates of CH4 and N20
emissions presented are based on broad indicators of emissions (i.e., fuel use multiplied by an aggregate emission
factor for different sectors), rather than specific emission processes (i.e., by combustion technology and type of
emission control).
An uncertainty analysis was performed by primary fuel type for each end-use sector, using the IPCC-recommended
Approach 2 uncertainty estimation methodology, Monte Carlo Stochastic Simulation technique, with @ RISK
software.
The uncertainty estimation model for this source category was developed by integrating the CH4 and N20
stationary source inventory estimation models with the model for C02 from fossil fuel combustion to realistically
characterize the interaction (or endogenous correlation) between the variables of these three models. About 55
input variables were simulated for the uncertainty analysis of this source category (about 20 from the C02
emissions from fossil fuel combustion inventory estimation model and about 35 from the stationary source
inventory models).
In developing the uncertainty estimation model, uniform distribution was assumed for all activity-related input
variables and N20 emission factors, based on the SAIC/EIA (2001) report.53 For these variables, the uncertainty
ranges were assigned to the input variables based on the data reported in SAIC/EIA (2001).54 However, the CH4
emission factors differ from those used by EIA. These factors and uncertainty ranges are based on IPCC default
uncertainty estimates (IPCC 2006).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-18. Stationary
combustion CH4 emissions in 2019 (including biomass) were estimated to be between 5.5 and 20.2 MMT C02 Eq. at
a 95 percent confidence level. This indicates a range of 36 percent below to 133 percent above the 2019 emission
52	Several of the U.S. Tier 2 emission factors were used in IPCC (2006) as Tier 1 emission factors. See Table A-75 in Annex 3.1 for
emission factors by technology type and fuel type for the electric power sector.
53	SAIC/EIA (2001) characterizes the underlying probability density function for the input variables as a combination of uniform
and normal distributions (the former distribution to represent the bias component and the latter to represent the random
component). However, for purposes of the current uncertainty analysis, it was determined that uniform distribution was more
appropriate to characterize the probability density function underlying each of these variables.
54	In the SAIC/EIA (2001) report, the quantitative uncertainty estimates were developed for each of the three major fossil fuels
used within each end-use sector; the variations within the sub-fuel types within each end-use sector were not modeled.
However, for purposes of assigning uncertainty estimates to the sub-fuel type categories within each end-use sector in the
current uncertainty analysis, SAIC/EIA (2001)-reported uncertainty estimates were extrapolated.
Energy 3-43

-------
estimate of 8.7 MMT C02 Eq.55 Stationary combustion N20 emissions in 2019 (including biomass) were estimated
to be between 18.7 and 37.7 MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 25 percent
below to 51 percent above the 2019 emission estimate of 24.9 MMT C02 Eq.
Table 3-18: Approach 2 Quantitative Uncertainty Estimates for ChU and N2O Emissions from
Energy-Related Stationary Combustion, Including Biomass (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTCOz Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Stationary Combustion
Stationary Combustion
ch4
n2o
8.7
24.9
5.5
18.7
20.2
37.7
-36% +133%
-25% +51%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
The uncertainties associated with the emission estimates of CH4 and N20 are greater than those associated with
estimates of C02 from fossil fuel combustion, which mainly rely on the carbon content of the fuel combusted.
Uncertainties in both CH4 and N20 estimates are due to the fact that emissions are estimated based on emission
factors representing only a limited subset of combustion conditions. For the indirect greenhouse gases,
uncertainties are partly due to assumptions concerning combustion technology types, age of equipment, emission
factors used, and activity data projections.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019 as discussed below. Details on the emission trends through time are described in more detail in the
Methodology section, above. As discussed in Annex 5, data are unavailable to include estimates of CH4 and N20
emissions from biomass use in Territories, but those emissions are assumed to be insignificant.
QA/QC and Verification
In order to ensure the quality of the non-C02 emission estimates from stationary combustion, general (IPCC Tier 1)
and category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved
checks specifically focusing on the activity data and emission factor sources and methodology used for estimating
CH4, N20, and the greenhouse gas precursors from stationary combustion in the United States. Emission totals for
the different sectors and fuels were compared and trends were investigated.
Recalculations Discussion
Methane and N20 emissions from stationary sources (excluding C02) across the entire time series were revised due
to revised data from EIA (2020a) and EPA (2020a) relative to the previous Inventory. Most notably, newly
published U.S. Territories data from EIA (2020b) was integrated, which impacted coal, fuel oil, and natural gas
estimates for U.S. Territories across the time series. EIA (2020a) revised approximate heat rates for electricity and
the heat content of electricity for natural gas and noncombustible renewable energy, which impacted electric
power energy consumption by sector. As a result of revised natural gas heat contents, EIA updated natural gas
consumption in the residential, commercial, and industrial sectors for 2018.
EIA also revised sector allocations for distillate fuel oil, residual fuel oil, and kerosene for 2018, and for propane for
2010 through 2012, 2014, 2017, and 2018, which impacted LPG by sector. EPA (2020a) revised coal, fuel oil,
natural gas, and wood consumption statistics for 2018 in the electric power sector. EPA revised distillate fuel oil
55 The low emission estimates reported in this section have been rounded down to the nearest integer values and the high
emission estimates have been rounded up to the nearest integer values.
3-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
carbon contents and LPG heat contents and carbon contents, which affect petroleum consumption in the
residential, commercial, and industrial sectors (EPA 2020; ICF 2020). The historical data changes and methodology
updates resulted in an average annual decrease of less than 0.05 MMT C02 Eq. (0.2 percent) in CH4 emissions, and
an average annual decrease of less than 0.05 MMT C02 Eq. (0.1 percent) in N20 emissions for the 1990 through
2018 period.
Planned Improvements
Several items are being evaluated to improve the CH4 and N20 emission estimates from stationary combustion and
to reduce uncertainty for U.S. Territories. Efforts will be taken to work with EIA and other agencies to improve the
quality of the U.S. Territories data. Because these data are not broken out by stationary and mobile uses, further
research will be aimed at trying to allocate consumption appropriately. In addition, the uncertainty of biomass
emissions will be further investigated because it was expected that the exclusion of biomass from the estimates
would reduce the uncertainty; and in actuality the exclusion of biomass increases the uncertainty. These
improvements are not all-inclusive but are part of an ongoing analysis and efforts to continually improve these
stationary combustion estimates from U.S. Territories.
Other forms of biomass-based gas consumption include biogas. EPA will examine EIA and GHGRP data on biogas
collected and burned for energy use and determine if CH4 and N20 emissions from biogas can be included in future
inventories. EIA (2020a) natural gas data already deducts biogas used in the natural gas supply, so no adjustments
are needed to the natural gas fuel consumption data to account for biogas.
CH4 and N20 from Mobile Combustion
Methodology
Estimates of CH4 and N20 emissions from mobile combustion were calculated by multiplying emission factors by
measures of activity for each fuel and vehicle type (e.g., light-duty gasoline trucks). Activity data included vehicle
miles traveled (VMT) for on-road vehicles and fuel consumption for non-road mobile sources. The activity data and
emission factors used in the calculations are described in the subsections that follow. A complete discussion of the
methodology used to estimate CH4 and N20 emissions from mobile combustion and the emission factors used in the
calculations is provided in Annex 3.2.
On-Road Vehicles
Estimates of CH4 and N20 emissions from gasoline and diesel on-road vehicles are based on VMT and emission
factors (in grams of CH4 and N20 per mile) by vehicle type, fuel type, model year, and emission control technology.
Emission estimates for alternative fuel vehicles (AFVs) are based on VMT and emission factors (in grams of CH4 and
N20 per mile) by vehicle and fuel type.56
CH4 and N20 emissions factors for newer (starting with model year 2004) on-road gasoline vehicles were calculated
by Browning (2019) from annual vehicle certification data compiled by EPA. CH4 and N20 emissions factors for
older (model year 2003 and earlier) on-road gasoline vehicles were developed by ICF (2004). These emission
factors were derived from EPA, California Air Resources Board (CARB) and Environment and Climate Change
Canada laboratory test results of different vehicle and control technology types. The EPA, CARB and Environment
and Climate Change Canada tests were designed following the Federal Test Procedure (FTP). The procedure covers
three separate driving segments, since vehicles emit varying amounts of greenhouse gases depending on the
driving segment. These driving segments are: (1) a transient driving cycle that includes cold start and running
emissions, (2) a cycle that represents running emissions only, and (3) a transient driving cycle that includes hot
56 Alternative fuel and advanced technology vehicles are those that can operate using a motor fuel other than gasoline or
diesel. This includes electric or other bi-fuel or dual-fuel vehicles that may be partially powered by gasoline or diesel.
Energy 3-45

-------
start and running emissions. For each test run, a bag was affixed to the tailpipe of the vehicle and the exhaust was
collected; the content of this bag was then analyzed to determine quantities of gases present. The emissions
characteristics of driving segment 2 tests were used to define running emissions. Running emissions were
subtracted from the total FTP emissions to determine start emissions. These were then recombined to
approximate average driving characteristics, based upon the ratio of start to running emissions for each vehicle
class from MOBILE6.2, an EPA emission factor model that predicts gram per mile emissions of C02, CO, HC, NOx,
and PM from vehicles under various conditions.57
Diesel on-road vehicle emission factors were developed by ICF (2006a). CH4 and N20 emissions factors for newer
(starting at model year 2007) on-road diesel vehicles (those using engine aftertreatment systems) were calculated
from annual vehicle certification data compiled by EPA.
CH4 and N20 emission factors for AFVs were developed based on the 2019 Greenhouse gases, Regulated
Emissions, and Energy use in Transportation (GREET) model (ANL 2020). For light-duty trucks, EPA used a curve fit
of 1999 through 2011 travel fractions for LDT1 and LDT2 (MOVES Source Type 31 for LDT1 and MOVES Source Type
32 for LDT2). For medium-duty vehicles, EPA used emission factors for light heavy-duty vocational trucks. For
heavy-duty vehicles, EPA used emission factors for long-haul combination trucks. For buses, EPA used emission
factors for transit buses. These values represent vehicle operations only (tank-to-wheels); upstream well-to-tank
emissions are calculated elsewhere in the Inventory. Biodiesel CH4 emission factors were corrected from GREET
values to be the same as CH4 emission factors for diesel vehicles. GREET overestimated biodiesel CH4 emission
factors based upon an incorrect CH4-to-THC ratio for diesel vehicles with aftertreatment technology.
Annual VMT data for 1990 through 2019 were obtained from the Federal Highway Administration's (FHWA)
Highway Performance Monitoring System database as reported in Highway Statistics (FHWA 1996 through 2019).58
VMT estimates were then allocated from FHWA's vehicle categories to fuel-specific vehicle categories using the
calculated shares of vehicle fuel use for each vehicle category by fuel type reported in DOE (1993 through 2018)
and information on total motor vehicle fuel consumption by fuel type from FHWA (1996 through 2019). VMT for
AFVs were estimated based on Browning (2017 and 2018a). The age distributions of the U.S. vehicle fleet were
obtained from EPA (2019a, 2000), and the average annual age-specific vehicle mileage accumulation of U.S.
vehicles were obtained from EPA (2019a).
Control technology and standards data for on-road vehicles were obtained from EPA's Office of Transportation and
Air Quality (EPA 2019a, 2020c, 2000,1998, and 1997) and Browning (2005). These technologies and standards are
defined in Annex 3.2, and were compiled from EPA (1994a, 1994b, 1998,1999a) and IPCC (2006) sources.
Non-Road Mobile Sources
The non-road mobile category for CH4 and N20 includes ships and boats, aircraft, locomotives and off-road sources
(e.g., construction or agricultural equipment). For non-road sources, fuel-based emission factors are applied to
data on fuel consumption, following the IPCC Tier 1 approach, for locomotives, aircraft, ships and boats. The Tier 2
approach would require separate fuel-based emissions factors by technology for which data are not available. For
57	Additional information regarding the MOBILE model can be found online at .
58	The source of VMT data is FHWA Highway Statistics Table VM-1. In 2011, FHWA changed its methods for estimating data in
the VM-1 table. These methodological changes included how vehicles are classified, moving from a system based on body-type
to one that is based on wheelbase. These changes were first incorporated for the 1990 through 2010 Inventory and apply to
the 2007 through 2019 time period. This resulted in large changes in VMT by vehicle class, thus leading to a shift in emissions
among on-road vehicle classes. For example, the category "Passenger Cars" has been replaced by "Light-duty Vehicles-Short
Wheelbase" and "Other 2 axle-4 Tire Vehicles" has been replaced by "Light-duty Vehicles, Long Wheelbase." This change in
vehicle classification has moved some smaller trucks and sport utility vehicles from the light truck category to the passenger
vehicle category in the current Inventory. These changes are reflected in a large drop in light-truck emissions between 2006
and 2007.
3-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
some of the non-road categories, 2-stroke and 4-stroke technologies are broken out and have separate emission
factors; those cases could be considered a Tier 2 approach.
To estimate CH4 and N20 emissions from non-road mobile sources, fuel consumption data were employed as a
measure of activity, and multiplied by fuel-specific emission factors (in grams of N20 and CH4 per kilogram of fuel
consumed).59 Activity data were obtained from AAR (2008 through 2019), APTA (2007 through 2019), Raillnc (2014
through 2019), APTA (2006), BEA (1991 through 2015), Benson (2002 through 2004), DLA Energy (2020), DOC
(1991 through 2019), DOE (1993 through 2018), DOT (1991 through 2019), EIA (2002, 2007, 2020a), EIA (2020f),
EIA (1991 through 2019), EPA (2019a), Esser (2003 through 2004), FAA (2021), FHWA (1996 through 2019),60
Gaffney (2007), and Whorton (2006 through 2014). Emission factors for non-road modes were taken from IPCC
(2006) and Browning (2020 and 2018b).
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis was conducted for the mobile source sector using the IPCC-recommended
Approach 2 uncertainty estimation methodology, Monte Carlo Stochastic Simulation technique, using @ RISK
software. The uncertainty analysis was performed on 2019 estimates of CH4 and N20 emissions, incorporating
probability distribution functions associated with the major input variables. For the purposes of this analysis, the
uncertainty was modeled for the following four major sets of input variables: (1) VMT data, by on-road vehicle and
fuel type, (2) emission factor data, by on-road vehicle, fuel, and control technology type, (3) fuel consumption,
data, by non-road vehicle and equipment type, and (4) emission factor data, by non-road vehicle and equipment
type.
Uncertainty analyses were not conducted for NOx, CO, or NMVOC emissions. Emission factors for these gases have
been extensively researched because emissions of these gases from motor vehicles are regulated in the United
States, and the uncertainty in these emission estimates is believed to be relatively low. For more information, see
Section 3.9 - Uncertainty Analysis of Emission Estimates. However, a much higher level of uncertainty is associated
with CH4 and N20 emission factors due to limited emission test data, and because, unlike C02 emissions, the
emission pathways of CH4 and N20 are highly complex.
Based on the uncertainty analysis, mobile combustion CH4 emissions from all mobile sources in 2019 were
estimated to be between 2.3 and 3.5 MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 2
percent below to 46 percent above the corresponding 2019 emission estimate of 2.4 MMT C02 Eq. Mobile
combustion N20 emissions from mobile sources in 2019 were estimated to be between 16.4 and 21.3 MMT C02
Eq. at a 95 percent confidence level. This indicates a range of 9 percent below to 19 percent above the
corresponding 2019 emission estimate of 18.0 MMT C02 Eq.
59	The consumption of international bunker fuels is not included in these activity data, but emissions related to the
consumption of international bunker fuels are estimated separately under the International Bunker Fuels source category.
60	This Inventory uses FHWA's Agriculture, Construction, and Commercial/Industrial MF-24 fuel volumes along with the MOVES
model gasoline volumes to estimate non-road mobile source CH4 and N20 emissions for these categories. For agriculture, the
MF-24 gasoline volume is used directly because it includes both non-road trucks and equipment. For construction and
commercial/industrial category gasoline estimates, the 2014 and older MF-24 volumes represented non-road trucks only;
therefore, the MOVES gasoline volumes for construction and commercial/industrial categories are added to the respective
categories in the Inventory. Beginning in 2015, this addition is no longer necessary since the FHWA updated its methods for
estimating on-road and non-road gasoline consumption. Among the method updates, FHWA now incorporates MOVES
equipment gasoline volumes in the construction and commercial/industrial categories.
Energy 3-47

-------
Table 3-19: Approach 2 Quantitative Uncertainty Estimates for ChU and N2O Emissions from
Mobile Sources (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate9
(MMTCOz Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTCOzEq.) (Percent)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Mobile Sources
ch4
2.4
2.3
3.5
-2% +46%
Mobile Sources
n2o
18.0
16.4
21.3
-9% +19%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
This uncertainty analysis is a continuation of a multi-year process for developing quantitative uncertainty estimates
for this source category using the IPCC Approach 2 uncertainty estimation methodology. As a result, as new
information becomes available, uncertainty characterization of input variables may be improved and revised. For
additional information regarding uncertainty in emission estimates for CH4 and N20 please refer to the Uncertainty
Annex. As discussed in Annex 5, data are unavailable to include estimates of CH4 and N20 emissions from any liquid
fuel used in pipeline transport or some biomass used in transportation sources, but those emissions are assumed
to be insignificant.
QA/QC and Verification
In order to ensure the quality of the emission estimates from mobile combustion, general (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The specific plan used for mobile combustion was
updated prior to collection and analysis of this current year of data. The Tier 2 procedures focused on the emission
factor and activity data sources, as well as the methodology used for estimating emissions. These procedures
included a qualitative assessment of the emission estimates to determine whether they appear consistent with the
most recent activity data and emission factors available. A comparison of historical emissions between the current
Inventory and the previous Inventory was also conducted to ensure that the changes in estimates were consistent
with the changes in activity data and emission factors.
Recalculations Discussion
Updates were made to CH4 and N20 emission factors for newer non-road gasoline and diesel vehicles. Previously,
these emission factors were calculated using the updated 2006 IPCC Tier 3 guidance and the nonroad component
EPA's MOVES2014b model. CH4 emission factors were calculated directly from MOVES. N20 emission factors were
calculated using MOVES-Nonroad activity and emission factors in g/kWh by fuel type from the European
Environment Agency. Updated emission factors were developed this year using EPA engine certification data for
non-road small and large spark-ignition (SI) gasoline engines and compression-ignition diesel engines (model year
2011 and newer), as well as non-road motorcycles (model year 2006 and newer), SI marine engines (model year
2011 and newer), and diesel marine engines (model year 2000 and newer).
The collective result of these changes was a net decrease in CH4 emissions and an increase in N20 emissions from
mobile combustion relative to the previous Inventory. Methane emissions decreased by 23.2 percent. Nitrous
oxide emissions increased by 23.6 percent.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019 with one recent notable exception. An update by FHWA to the method for estimating on-road VMT
created an inconsistency in on-road CH4 and N20 for the time periods 1990 to 2006 and 2007 to 2019. Details on
the emission trends and methodological inconsistencies through time are described in the Methodology section
above.
3-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Planned Improvements
While the data used for this report represent the most accurate information available, several areas for
improvement have been identified.
•	Update emission factors for motorcycles. The Inventory does not currently account for advanced
technology motorcycles. EPA certification data can be used to update motorcycle assumptions to better
capture the portion of the motorcycle fleet using advanced emissions controls.
•	Update emission factors for buses. The Inventory currently groups buses into the heavy-duty vehicle
category. New emission factors specific to buses can be developed from EPA certification data.
•	Update emission factors for ships and boats using residual fuel and distillate fuel, emission factors for
locomotives using ultra low sulfur diesel, and emission factors for aircraft using jet fuel. The Inventory is
currently using IPCC default values for these emissions factors.
•	Continue to explore potential improvements to estimates of domestic waterborne fuel consumption for
future Inventories. The Inventory estimates for residual and distillate fuel used by ships and boats is based
in part on data on bunker fuel use from the U.S. Department of Commerce. Domestic fuel consumption is
estimated by subtracting fuel sold for international use from the total sold in the United States. Since
2015, all ships travelling within 200 nautical miles of the US coastlines must use distillate fuels thereby
overestimating the residual fuel used by US vessels and underestimating distillate fuel use in these ships.
3.2 Carbon Emitted from Non-Energy Uses of
Fossil Fuels (CRF Source Category 1A)
In addition to being combusted for energy, fossil fuels are also consumed for non-energy uses (NEU) in the United
States. The fuels used for these purposes are diverse, including natural gas, hydrocarbon gas liquids (HGL),61
asphalt (a viscous liquid mixture of heavy crude oil distillates), petroleum coke (manufactured from heavy oil), and
coal (metallurgical) coke (manufactured from coking coal). The non-energy applications of these fuels are equally
diverse, including feedstocks for the manufacture of plastics, rubber, synthetic fibers and other materials; reducing
agents for the production of various metals and inorganic products; and products such as lubricants, waxes, and
asphalt (IPCC 2006). Emissions from non-energy use of lubricants, paraffin waxes, bitumen / asphalt, and solvents
are reported in the Energy sector, as opposed to the Industrial Processes and Product Use (IPPU) sector, to reflect
national circumstances in its choice of methodology and to increase transparency of this source category's unique
country-specific data sources and methodology (see Box 3-5). In addition, estimates of non-energy use emissions
included here do not include emissions already reflected in the IPPU sector, e.g., fuels used as reducing agents. To
avoid double counting, the "raw" non-energy fuel consumption data reported by EIA are reduced to account for
these emissions already included under IPPU.
Carbon dioxide emissions arise from non-energy uses via several pathways. Emissions may occur during the
manufacture of a product, as is the case in producing plastics or rubber from fuel-derived feedstocks. Additionally,
emissions may occur during the product's lifetime, such as during solvent use. Overall, throughout the time series
61 HGL (formerly referred to as liquefied petroleum gas, or LPG) are hydrocarbons that occur as gases at atmospheric pressure
and as liquids under higher pressures. HGLs include paraffins, such as ethane, propane, butanes, and pentanes plus, and HGLs
include olefins, such as ethylene, propylene, and butylene. Adjustments were made in the current Inventory report to HGL
activity data, carbon content coefficients, and heat contents HGL. For more information about the updated HGL data and
assumptions, see the Recalculations Discussion section below.
Energy 3-49

-------
and across all uses, about 62 percent of the total C consumed for non-energy purposes was stored in products
(e.g., plastics), and not released to the atmosphere; the remaining 38 percent was emitted.
There are several areas in which non-energy uses of fossil fuels are closely related to other parts of this Inventory.
For example, some of the non-energy use products release C02 at the end of their commercial life when they are
combusted after disposal; these emissions are reported separately within the Energy chapter in the Incineration of
Waste source category. There are also net exports of petrochemical intermediate products that are not completely
accounted for in the EIA data, and the Inventory calculations adjust for the effect of net exports on the mass of C in
non-energy applications.
As shown in Table 3-20, fossil fuel emissions in 2019 from the non-energy uses of fossil fuels were 128.8 MMT C02
Eq., which constituted approximately 2 percent of overall fossil fuel emissions. In 2019, the consumption of fuels
for non-energy uses (after the adjustments described above) was 5,635.0 TBtu (see Table 3-21). A portion of the C
in the 5,635.0 TBtu of fuels was stored (228.8 MMT C02 Eq.), while the remaining portion was emitted (128.8 MMT
C02 Eq.). Non-energy use emissions decreased by 0.7 percent from 2018 to 2019 mainly due to a decrease in the
ratio between C stored and potential emissions. See Annex 2.3 for more details.
Table 3-20: CO2 Emissions from Non-Energy Use Fossil Fuel Consumption (MMT CO2 Eq. and
Percent)
Year
1990
2005
2015
2016
2017
2018
2019
Potential Emissions
306.1
367.4
322.8
317.8
332.7
352.8
357.5
C Stored
193.3
238.3
214.4
218.0
219.2
223.1
228.8
Emissions as a % of Potential
37%
35%
34%
31%
34%
37%
36%
C Emitted
112.8
129.1
108.5
99.8
113.5
129.7
128.8
Methodology
The first step in estimating C stored in products was to determine the aggregate quantity of fossil fuels consumed
for non-energy uses. The C content of these feedstock fuels is equivalent to potential emissions, or the product of
consumption and the fuel-specific C content values. Both the non-energy fuel consumption and C content data
were supplied by the EIA (2020) (see Annex 2.1). Consumption values for industrial coking coal, petroleum coke,
other oils, and natural gas in Table 3-21 and Table 3-22 have been adjusted to subtract non-energy uses that are
included in the source categories of the Industrial Processes and Product Use chapter.62 Consumption of natural
gas, HGL, pentanes plus, naphthas, other oils, and special naphtha were adjusted to subtract out net exports of
these products that are not reflected in the raw data from EIA. Consumption values were also adjusted to subtract
net exports of HGL components (e.g., propylene, ethane).
For the remaining non-energy uses, the quantity of C stored was estimated by multiplying the potential emissions
by a storage factor.
• For several fuel types—petrochemical feedstocks (including natural gas for non-fertilizer uses, HGL,
pentanes plus, naphthas, other oils, still gas, special naphtha, and industrial other coal), asphalt and road
oil, lubricants, and waxes—U.S. data on C stocks and flows were used to develop C storage factors,
calculated as the ratio of (a) the C stored by the fuel's non-energy products to (b) the total C content of
the fuel consumed. A lifecycle approach was used in the development of these factors in order to account
for losses in the production process and during use. Because losses associated with municipal solid waste
management are handled separately in the Energy sector under the Incineration of Waste source
category, the storage factors do not account for losses at the disposal end of the life cycle.
62 These source categories include Iron and Steel Production, Lead Production, Zinc Production, Ammonia Manufacture, Carbon
Black Manufacture (included in Petrochemical Production), Titanium Dioxide Production, Ferroalloy Production, Silicon Carbide
Production, and Aluminum Production.
3-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	For industrial coking coal and distillate fuel oil, storage factors were taken from Marland and Rotty (1984).
•	For the remaining fuel types (petroleum coke, miscellaneous products, and other petroleum), IPCC (2006)
does not provide guidance on storage factors, and assumptions were made based on the potential fate of
C in the respective non-energy use products. Carbon dioxide emissions from carbide production are
implicitly accounted for in the storage factor calculation for the non-energy use of petroleum coke.
Table 3-21: Adjusted Consumption of Fossil Fuels for Non-Energy Uses (TBtu)
Year
1990
2005
2015
2016
2017
2018
2019
Industry
4,317.2
5,111.1
4,864.5
4,833.1
5,089.6
5,445.4
5,492.3
Industrial Coking Coal
NO
80.4
122.4
89.6
113.0
124.8
132.1
Industrial Other Coal
7.6
11.0
9.5
9.5
9.5
9.5
9.5
Natural Gas to Chemical Plants
282.4
260.9
418.9
496.4
588.0
675.9
664.6
Asphalt & Road Oil
1,170.2
1,323.2
831.7
853.4
849.2
792.8
843.9
HGL
1,218.4
1,610.4
2,160.2
2,128.0
2,193.5
2,505.1
2,545.1
Lubricants
186.3
160.2
142.1
135.1
124.9
121.9
117.6
Pentanes Plus
117.5
95.4
78.3
53.1
81.7
105.2
154.7
Naphtha (<401 °F)
327.0
679.5
418.1
398.2
413.0
421.0
368.8
Other Oil (>401 °F)
663.6
499.5
216.9
204.6
242.9
219.0
211.7
Still Gas
36.7
67.7
162.2
166.1
163.8
166.9
158.7
Petroleum Coke
28.1
106.2
NO
NO
NO
NO
NO
Special Naphtha
101.1
60.9
97.1
89.0
95.3
87.0
89.3
Distillate Fuel Oil
7.0
11.7
5.8
5.8
5.8
5.8
5.8
Waxes
33.3
31.4
12.4
12.8
10.2
12.4
10.4
Miscellaneous Products
137.8
112.8
188.9
191.3
198.8
198.0
180.2
Transportation
176.0
151.3
162.8
154.4
142.0
137.0
132.1
Lubricants
176.0
151.3
162.8
154.4
142.0
137.0
132.1
U.S. Territories
50.8
114.9
10.3
10.5
10.7
10.7
10.7
Lubricants
0.7
4.6
1.0
1.0
1.0
1.0
1.0
Other Petroleum (Misc. Prod.)
50.1
110.3
9.3
9.5
9.6
9.6
9.6
Total
4,544.0
5,377.3
5,037.7
4,998.0
5,242.3
5,593.0
5,635.0
NO (Not Occurring)
Table 3-22: 2019 Adjusted Non-Energy Use Fossil Fuel Consumption, Storage, and Emissions
Adjusted

Non-Energy Carbon Content
Potential
Storage
Carbon
Carbon
Carbon

Use3
Coefficient
Carbon
Factor
Stored
Emissions
Emissions
Sector/Fuel Type
(TBtu)
(MMT C/QBtu)
(MMT C)

(MMT C)
(MMT C) (MMTCOz Eq.)
Industry
5,492.3
NA
94.6
NA
62.1
32.5
119.2
Industrial Coking Coal
132.1
25.59
3.4
0.10
0.3
3.0
11.2
Industrial Other Coal
9.5
26.07
0.2
0.62
0.2
0.1
0.3
Natural Gas to







Chemical Plants
664.6
14.47
9.6
0.62
5.9
3.6
13.4
Asphalt & Road Oil
843.9
20.55
17.3
1.00
17.3
0.1
0.3
HGL
2,545.1
16.85
42.9
0.62
26.6
16.3
59.8
Lubricants
117.6
20.20
2.4
0.09
0.2
2.2
7.9
Pentanes Plus
154.7
18.24
2.8
0.62
1.7
1.1
3.9
Naphtha (<401° F)
368.8
18.55
6.8
0.62
4.2
2.6
9.5
Other Oil (>401° F)
211.7
20.17
4.3
0.62
2.6
1.6
6.0
Still Gas
158.7
17.51
2.8
0.62
1.7
1.1
3.9
Petroleum Coke
NO
27.85
NO
0.30
NO
NO
NO
Energy 3-51

-------
Special Naphtha
89.3
19.74
1.8
0.62
1.1
0.7
2.5
Distillate Fuel Oil
5.8
20.22
0.1
0.50
0.1
0.1
0.2
Waxes
10.4
19.80
0.2
0.58
0.1
0.1
0.3
Miscellaneous







Products
180.2
0.00
0.0
0.00
0.0
0.0
0.0
Transportation
132.1
NA
2.7
NA
0.2
2.4
8.9
Lubricants
132.1
20.20
2.7
0.09
0.2
2.4
8.9
U.S. Territories
10.7
NA
0.2
NA
+
0.2
0.7
Lubricants
1.0
20.20
+
0.09
+
+
0.1
Other Petroleum







(Misc. Prod.)
9.6
20.00
0.2
0.10
+
0.2
0.6
Total
5,635.0

97.5

62.4
35.1
128.8
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 TBtu, MMT C, or MMT C02 Eq.
NA (Not Applicable)
NO (Not Occurring)
a To avoid double counting, net exports have been deducted.
Lastly, emissions were estimated by subtracting the C stored from the potential emissions (see Table 3-20). More
detail on the methodology for calculating storage and emissions from each of these sources is provided in Annex
2.3.
Where storage factors were calculated specifically for the United States, data were obtained on (1) products such
as asphalt, plastics, synthetic rubber, synthetic fibers, cleansers (soaps and detergents), pesticides, food additives,
antifreeze and deicers (glycols), and silicones; and (2) industrial releases including energy recovery (waste gas from
chemicals), Toxics Release Inventory (TRI) releases, hazardous waste incineration, and volatile organic compound,
solvent, and non-combustion CO emissions. Data were taken from a variety of industry sources, government
reports, and expert communications. Sources include EPA reports and databases such as compilations of air
emission factors (EPA 2001), National Emissions Inventory (NEI)Air Pollutant Emissions Trends Data (EPA 2020),
Toxics Release Inventory, 1998 (EPA 2000b), Biennial Reporting System (EPA 2000a, 2009), Resource Conservation
and Recovery Act Information System (EPA 2013b, 2015, 2016b, 2018b, 2021), pesticide sales and use estimates
(EPA 1998,1999, 2002, 2004, 2011, 2017), and the Chemical Data Access Tool (EPA 2014b); the EIA Manufacturer's
Energy Consumption Survey (MECS) (EIA 1994,1997, 2001, 2005, 2010, 2013, 2017, 2021); the National
Petrochemical & Refiners Association (NPRA 2002); the U.S. Census Bureau (1999, 2004, 2009, 2014); Bank of
Canada (2012, 2013, 2014, 2016, 2017, 2018, 2019, 2020); Financial Planning Association (2006); INEGI (2006); the
United States International Trade Commission (1990 through 2018); Gosselin, Smith, and Hodge (1984); EPA's
Municipal Solid Waste (MSW) Facts and Figures (EPA 2013, 2014a, 2016a, 2018a, 2019); the Rubber
Manufacturers' Association (RMA 2009, 2011, 2014, 2016, 2018); the International Institute of Synthetic Rubber
Products (IISRP 2000, 2003); the Fiber Economics Bureau (FEB 2001, 2003, 2005, 2007, 2009, 2010, 2011, 2012,
2013); the Independent Chemical Information Service (ICIS 2008, 2016); the EPA Chemical Data Access Tool (CDAT)
(EPA 2014b); the American Chemistry Council (ACC 2003 through 2011, 2013, 2014, 2015, 2016, 2017, 2018, 2019,
2020b); and the Guide to the Business of Chemistry (ACC 2020a). Specific data sources are listed in full detail in
Annex 2.3.
Box 3-5: Reporting of Lubricants, Waxes, and Asphalt and Road Oil Product Use in Energy Sector
IPCC (2006) provides methodological guidance to estimate emissions from the first use of fossil fuels as a
product for primary purposes other than combustion for energy purposes (including lubricants, paraffin waxes,
bitumen / asphalt, and solvents) under the IPPU sector.63 In this Inventory, C storage and C emissions from
product use of lubricants, waxes, and asphalt and road oil are reported under the Energy sector in the Carbon
63 See for example Volume 3: Industrial Processes and Product Use, and Chapter 5: Non-Energy Products from Fuels and
Solvent Use of the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2006).
3-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Emitted from Non-Energy Uses of Fossil Fuels source category (CRF Source Category 1A5).64
The emissions are reported in the Energy sector, as opposed to the IPPU sector, to reflect national
circumstances in its choice of methodology and to increase transparency of this source category's unique
country-specific data sources and methodology. Although emissions from these non-energy uses are reported in
the Energy chapter the methodologies used to determine emissions are compatible with the 2006 IPCC
Guidelines. The country-specific methodology used for the Carbon Emitted from Non-Energy Uses of Fossil Fuels
source category is based on a carbon balance (i.e., C inputs-outputs) calculation of the aggregate amount of
fossil fuels used for non-energy uses, including inputs of lubricants, waxes, asphalt and road oil (see Table 3-22).
For those inputs, U.S. country-specific data on C stocks and flows are used to develop carbon storage factors,
which are calculated as the ratio of the C stored by the fossil fuel non-energy products to the total C content of
the fuel consumed, taking into account losses in the production process and during product use.65 The country-
specific methodology to reflect national circumstances starts with the aggregate amount of fossil fuels used for
non-energy uses and applies a C balance calculation, breaking out the C emissions from non-energy use of
lubricants, waxes, and asphalt and road oil. The emissions are reported under the Energy chapter to improve
transparency, report a more complete carbon balance and to avoid double counting. Due to U.S. national
circumstances, reporting these C emissions separately under IPPU would involve making artificial adjustments
to allocate both the C inputs and C outputs of the non-energy use C balance. For example, only the emissions
from the first use of lubricants and waxes are to be reported under the IPPU sector, emissions from use of
lubricants in 2-stroke engines and emissions from secondary use of lubricants and waxes in waste incineration
with energy recovery are to be reported under the Energy sector. Reporting these non-energy use emissions
from only first use of lubricants and waxes under IPPU would involve making artificial adjustments to the non-
energy use C carbon balance and could potentially result in double counting of emissions. These artificial
adjustments would also be required for asphalt and road oil and solvents (which are captured as part of
petrochemical feedstock emissions) and could also potentially result in double counting of emissions. To avoid
presenting an incomplete C balance and a less transparent approach for the Carbon Emitted from Non-Energy
Uses of Fossil Fuels source category calculation, the entire calculation of C storage and C emissions is therefore
conducted in the Non-Energy Uses of Fossil Fuels category calculation methodology, and both the C storage and
C emissions for lubricants, waxes, and asphalt and road oil are reported under the Energy sector.
However, emissions from non-energy uses of fossil fuels as feedstocks or reducing agents (e.g., petrochemical
production, aluminum production, titanium dioxide and zinc production) are reported in the IPPU chapter,
unless otherwise noted due to specific national circumstances.
Uncertainty and Time-Series Consistency
An uncertainty analysis was conducted to quantify the uncertainty surrounding the estimates of emissions and
storage factors from non-energy uses. This analysis, performed using @RISK software and the IPCC-recommended
Approach 2 methodology (Monte Carlo Stochastic Simulation technique), provides for the specification of
probability density functions for key variables within a computational structure that mirrors the calculation of the
inventory estimate. The results presented below provide the 95 percent confidence interval, the range of values
within which emissions are likely to fall, for this source category.
As noted above, the non-energy use analysis is based on U.S.-specific storage factors for (1) feedstock materials
(natural gas, HGL, pentanes plus, naphthas, other oils, still gas, special naphthas, and other industrial coal), (2)
64	Non-methane volatile organic compound (NMVOC) emissions from solvent use are reported separately in the IPPU sector,
following Chapter 5 of the 2006 IPCC Guidelines.
65	Data and calculations for lubricants and waxes and asphalt and road oil are in Annex 2.3 - Methodology for Estimating
Carbon Emitted from Non-Energy Uses of Fossil Fuels.
Energy 3-53

-------
asphalt, (3) lubricants, and (4) waxes. For the remaining fuel types (the "other" category in Table 3-21 and Table
3-22), the storage factors were taken directly from IPCC (2006), where available, and otherwise assumptions were
made based on the potential fate of carbon in the respective NEU products. To characterize uncertainty, five
separate analyses were conducted, corresponding to each of the five categories. In all cases, statistical analyses or
expert judgments of uncertainty were not available directly from the information sources for all the activity
variables; thus, uncertainty estimates were determined using assumptions based on source category knowledge.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-23 (emissions) and Table
3-24 (storage factors). Carbon emitted from non-energy uses of fossil fuels in 2019 was estimated to be between
81.0 and 187.2 MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 37 percent below to 45
percent above the 2019 emission estimate of 128.8 MMT C02 Eq. The uncertainty in the emission estimates is a
function of uncertainty in both the quantity of fuel used for non-energy purposes and the storage factor.
Table 3-23: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Non-
Energy Uses of Fossil Fuels (MMT CO2 Eq. and Percent)
2019 Emission Estimate Uncertainty Range Relative to Emission Estimate3
Source	Gas
(MMT CO? Eq.)	(MMT C02 Eq.)	(%)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Feedstocks
C02
99.3
56.4
160.4
-43%
+62%
Asphalt
C02
0.3
0.1
0.6
-58%
+118%
Lubricants
C02
16.9
14.0
19.6
-17%
+16%
Waxes
C02
0.3
0.2
0.6
-24%
+90%
Other
C02
12.0
2.5
13.9
-79%
+16%
Total
C02
128.8
81.0
187.2
-37%
+45%
Note: Totals may not sum due to independent rounding.
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Table 3-24: Approach 2 Quantitative Uncertainty Estimates for Storage Factors of Non-
Energy Uses of Fossil Fuels (Percent)
2019 Storage Factor Uncertainty Range Relative to Emission Estimate3
Source	Gas
(%)	(%)	(%, Relative)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Feedstocks
C02
62.0%
49.5%
72.8%
-20%
+18%
Asphalt
C02
99.6%
99.1%
99.8%
-0.5%
+0.2%
Lubricants
C02
9.2%
3.9%
17.5%
-58%
+91%
Waxes
C02
57.8%
47.5%
67.5%
-18%
+17%
Other
C02
11.3%
7.9%
81.0%
-30%
+618%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent
confidence interval, as a percentage of the inventory value (also expressed in percent terms).
As shown in Table 3-24, feedstocks and asphalt contribute least to overall storage factor uncertainty on a
percentage basis. Although the feedstocks category—the largest use category in terms of total carbon flows-
appears to have tight confidence limits, this is to some extent an artifact of the way the uncertainty analysis was
structured. As discussed in Annex 2.3, the storage factor for feedstocks is based on an analysis of six fates that
result in long-term storage (e.g., plastics production), and eleven that result in emissions (e.g., volatile organic
compound emissions). Rather than modeling the total uncertainty around all of these fate processes, the current
3-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
analysis addresses only the storage fates, and assumes that all C that is not stored is emitted. As the production
statistics that drive the storage values are relatively well-characterized, this approach yields a result that is
probably biased toward understating uncertainty.
As is the case with the other uncertainty analyses discussed throughout this document, the uncertainty results
above address only those factors that can be readily quantified. More details on the uncertainty analysis are
provided in Annex 2.3.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019 as discussed below. Details on the emission trends through time are described in more detail in the
Methodology section, above.
QA/QC and Verification
In order to ensure the quality of the emission estimates from non-energy uses of fossil fuels, general (IPCC Tier 1)
and category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. This effort included a general analysis, as well as portions
of a category specific analysis for non-energy uses involving petrochemical feedstocks and for imports and exports.
The Tier 2 procedures that were implemented involved checks specifically focusing on the activity data and
methodology for estimating the fate of C (in terms of storage and emissions) across the various end-uses of fossil
C. Emission and storage totals for the different subcategories were compared, and trends across the time series
were analyzed to determine whether any corrective actions were needed. Corrective actions were taken to rectify
minor errors and to improve the transparency of the calculations, facilitating future QA/QC.
For petrochemical import and export data, special attention was paid to NAICS numbers and titles to verify that
none had changed or been removed. Import and export totals were compared with 2018 totals as well as their
trends across the time series.
Some degree of double counting may occur between these estimates of non-energy use of fuels and process
emissions from petrochemical production presented in the Industrial Processes and Produce Use (IPPU) sector.
This was examined and is not considered to be a significant issue since the non-energy use industrial release data
includes different categories of sources than those included in the IPPU sector. Data integration is not feasible at
this time as feedstock data from EIA used to estimate non-energy uses of fuels are aggregated by fuel type, rather
than disaggregated by both fuel type and particular industries (e.g., petrochemical production) as currently
collected through EPA's GHGRP and used for the petrochemical production category.
Recalculations Discussion
The "miscellaneous products" category reported by EIA includes miscellaneous products that are not reported
elsewhere in the EIA data set. The EIA does not have firm data concerning the amounts of various products that
are being reported in the "miscellaneous products" category; however, EIA has indicated that recovered sulfur
compounds from petroleum and natural gas processing, and potentially also carbon black feedstock could be
reported in this category. Recovered sulfur has no carbon content and would not be reported in the NEU
calculation or elsewhere in the Inventory. Based on this information, the miscellaneous products category
reported by EIA was assumed to be mostly petroleum refinery sulfur compounds that do not contain carbon (EIA
2019). Therefore, the carbon content for miscellaneous products was updated to be zero across the time series.
In addition, adjustments were made to activity data, carbon content coefficients, and heat contents for HGLfor
1990 to 2018. Historical HGL activity data from 1990 to 2007 were adjusted to use ElA's Petroleum Supply Annual
tables for consistency with the rest of the entire time series (i.e., 2008 to 2019). In previous Inventory reports, HGL
activity data from 1990 to 2007 were extracted from the American Petroleum Institute's Sales of Natural Gas
Liquids and Liquefied Refinery Gases. Thus, the HGL data source for the 1990 to 2007 portion of the time series was
updated to align with the HGL activity data used for 2008 to 2019 as well as with data used in other Energy sector
source categories (e.g., ElA's Monthly Energy Review (EIA 2020a)). In addition, the HGL carbon content coefficient
Energy 3-55

-------
for NEU was updated by separating each fuel out by its natural gas liquid (NGL)66 and associated olefin to calculate
a more accurate and annually variable factor, and the heat contents for HGL and pentanes plus were updated
using updated data from ElA's Monthly Energy Review (EIA 2020a).
Natural Gas to Chemical Plants data were updated to reflect the 2018 MECS data. This resulted in an increase in
natural gas used for NEU of 120 percent in 2018 compared to previous reports. Adjustments were also made to
historical calculations to linearly interpolate between ElA's MECS data years. Previously, fuel consumption data for
years between MECS releases were assumed to be equal to the previous year of data.
Non-energy use of petroleum coke consumption was adjusted to account for leap years when converting from
barrels per day to barrels per year. The carbon factor used to determine the amount of petroleum coke used in
several IPPU categories was updated to be consistent with the factors used in the fossil fuel combustion estimates.
This update impacted the amount of petroleum coke subtracted from non-energy use calculations.
Overall, these changes resulted in an average annual decrease of 10.9 MMT C02 Eq. (8.7 percent) in carbon
emissions from non-energy uses of fossil fuels for the period 1990 through 2018, relative to the previous
Inventory. This decrease is primarily due to the removal of miscellaneous products, which previously constituted
an average of 8.2 percent of total emissions from 1990 to 2018.
Planned Improvements
There are several future improvements planned:
•	More accurate accounting of C in petrochemical feedstocks. EPA has worked with EIA to determine the
cause of input/output discrepancies in the C mass balance contained within the NEU model. In the future,
two strategies to reduce or eliminate this discrepancy will continue to be pursued as part of quality
control procedures. First, accounting of C in imports and exports will be improved. The import/export
adjustment methodology will be examined to ensure that net exports of intermediaries such as ethylene
and propylene are fully accounted for. Second, the use of top-down C input calculation in estimating
emissions will be reconsidered. Alternative approaches that rely more substantially on the bottom-up C
output calculation will be considered instead.
•	Improving the uncertainty analysis. Most of the input parameter distributions are based on professional
judgment rather than rigorous statistical characterizations of uncertainty.
•	Better characterizing flows of fossil C. Additional fates may be researched, including the fossil C load in
organic chemical wastewaters, plasticizers, adhesives, films, paints, and coatings. There is also a need to
further clarify the treatment of fuel additives and backflows (especially methyl tert-butyl ether, MTBE).
•	Reviewing the trends in fossil fuel consumption for non-energy uses. Annual consumption for several fuel
types is highly variable across the time series, including industrial coking coal and other petroleum. A
better understanding of these trends will be pursued to identify any mischaracterized or misreported fuel
consumption for non-energy uses.
•	Updating the average C content of solvents was researched, since the entire time series depends on one
year's worth of solvent composition data. The data on C emissions from solvents that were readily
available do not provide composition data for all categories of solvent emissions and also have conflicting
definitions for volatile organic compounds, the source of emissive C in solvents. Additional sources of
solvents data will be investigated in order to update the C content assumptions.
•	Updating the average C content of cleansers (soaps and detergents) was researched; although production
and consumption data for cleansers are published every 5 years by the Census Bureau, the composition (C
content) of cleansers has not been recently updated. Recently available composition data sources may
facilitate updating the average C content for this category.
•	Revising the methodology for consumption, production, and C content of plastics was researched;
66 NGL are defined by EIA as "a group of hydrocarbons including ethane, propane, normal butane, isobutane, and natural
gasoline. [NGL] generally include natural gas plant liquids and all liquefied refinery gases except olefins" (EIA 2020b).
3-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
because of recent changes to the type of data publicly available for plastics, the NEU model for plastics
applies data obtained from personal communications. Potential revisions to the plastics methodology to
account for the recent changes in published data will be investigated.
•	Although U.S.-specific storage factors have been developed for feedstocks, asphalt, lubricants, and waxes,
default values from IPCC are still used for two of the non-energy fuel types (industrial coking coal,
distillate oil), and broad assumptions are being used for miscellaneous products and other petroleum.
Over the long term, there are plans to improve these storage factors by analyzing C fate similar to those
described in Annex 2.3 or deferring to more updated default storage factors from IPCC where available.
•	Reviewing the storage of carbon black across various sectors in the Inventory; in particular, the carbon
black abraded and stored in tires.
3.3 Incineration of Waste (CRF Source
Category 1A5)
Incineration is used to manage about 7 to 19 percent of the solid wastes generated in the United States,
depending on the source of the estimate and the scope of materials included in the definition of solid waste (EPA
2000; EPA 2020; Goldstein and Madtes 2001; Kaufman et al. 2004; Simmons et al. 2006; van Haaren et al. 2010). In
the context of this section, waste includes all municipal solid waste (MSW) as well as scrap tires. In the United
States, incineration of MSW tends to occur at waste-to-energy facilities or industrial facilities where useful energy
is recovered, and thus emissions from waste incineration are accounted for in the Energy chapter. Similarly, scrap
tires are combusted for energy recovery in industrial and utility boilers, pulp and paper mills, and cement kilns.
Incineration of waste results in conversion of the organic inputs to C02. According to the 2006 IPCC Guidelines,
when the C02 emitted is of fossil origin, it is counted as a net anthropogenic emission of C02 to the atmosphere.
Thus, the emissions from waste incineration are calculated by estimating the quantity of waste combusted and the
fraction of the waste that is C derived from fossil sources.
Most of the organic materials in MSW are of biogenic origin (e.g., paper, yard trimmings), and have their net C
flows accounted for under the Land Use, Land-Use Change, and Forestry chapter. However, some components of
MSW and scrap tires—plastics, synthetic rubber, synthetic fibers, and carbon black—are of fossil origin. Plastics in
the U.S. waste stream are primarily in the form of containers, packaging, and durable goods. Rubber is found in
durable goods, such as carpets, and in non-durable goods, such as clothing and footwear. Fibers in MSW are
predominantly from clothing and home furnishings. As noted above, scrap tires (which contain synthetic rubber
and carbon black) are also considered a "non-hazardous" waste and are included in the waste incineration
estimate, though waste disposal practices for tires differ from MSW. Estimates on emissions from hazardous waste
incineration can be found in Annex 2.3 and are accounted for as part of the C mass balance for non-energy uses of
fossil fuels.
Approximately 20.8 million metric tons of MSW were incinerated in 2011 (van Haaren et al. 2010). Updated data
were not available for 2012 through 2019 from this source so the data were proxied to the 2011 estimate. Carbon
dioxide emissions from incineration of waste increased 42 percent since 1990, to an estimated 11.5 MMT C02
(11,471 kt) in 2019, as the volume of scrap tires and other fossil C-containing materials in waste increased (see
Table 3-25 and Table 3-26).
Waste incineration is also a source of CH4 and N20 emissions (De Soete 1993; IPCC 2006). Methane emissions from
the incineration of waste were estimated to be less than 0.05 MMT C02 Eq. (less than 0.05 kt CH4) in 2019 and
have decreased by 11 percent since 1990. Nitrous oxide emissions from the incineration of waste were estimated
to be 0.3 MMT C02 Eq. (1 kt N20) in 2019 and have decreased by 32 percent since 1990. This decrease is driven by
the decrease in total MSW incinerated.
Energy 3-57

-------
Table 3-25: CO2, ChU, and N2O Emissions from the Incineration of Waste (MMT CO2 Eq.)
Gas/Waste Product
1990
2005
2015
2016
2017
2018
2019
co2
8.1
12.7
11.5
11.5
11.5
11.5
11.5
Plastics
5.7
7.2
6.3
6.4
6.5
6.6
6.6
Synthetic Rubber in Tires
0.3
1.6
1.4
1.4
1.3
1.3
1.2
Carbon Black in Tires
0.4
2.0
1.8
1.7
1.5
1.5
1.5
Synthetic Rubber in MSW
0.9
0.8
0.7
0.7
0.7
0.7
0.7
Synthetic Fibers
0.8
1.2
1.3
1.4
1.4
1.4
1.4
ch4
+
+
+
+
+
+
+
n2o
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Total
8.5
13.1
11.8
11.8
11.8
11.9
11.8
+ Does not exceed 0.05 MMT C02 Eq.






Table 3-26: CO2, ChU, and N2O Emissions from the Incineration of Waste (kt)

Gas/Waste Product
1990
2005
2015
2016
2017
2018
2019
co2
8,062
12,713
11,533
11,525
11,537
11,547
11,471
Plastics
5,699
7,163
6,316
6,370
6,532
6,588
6,588
Synthetic Rubber in Tires
308
1,599
1,440
1,369
1,298
1,264
1,229
Carbon Black in Tires
385
1,958
1,755
1,670
1,585
1,544
1,503
Synthetic Rubber in MSW
854
766
703
717
731
739
739
Synthetic Fibers
816
1,227
1,319
1,399
1,392
1,412
1,410
ch4
+
+
+
+
+
+
+
n2o
2
1
1
1
1
1
1
+ Does not exceed 0.5 kt.
Methodology
Emissions of C02from the incineration of waste include C02 generated by the incineration of plastics, synthetic
fibers, and synthetic rubber in MSW, as well as the incineration of synthetic rubber and carbon black in scrap tires.
The emission estimates are calculated for all four sources on a mass-basis based on the data available. These
emissions were estimated by multiplying the mass of each material incinerated by the C content of the material
and the fraction oxidized (98 percent). Plastics incinerated in MSW were categorized into seven plastic resin types,
each material having a discrete C content. Similarly, synthetic rubber is categorized into three product types, and
synthetic fibers were categorized into four product types, each having a discrete C content. Scrap tires contain
several types of synthetic rubber, carbon black, and synthetic fibers. Each type of synthetic rubber has a discrete C
content, and carbon black is 100 percent C. Emissions of C02 were calculated based on the amount of scrap tires
used for fuel and the synthetic rubber and carbon black content of scrap tires. More detail on the methodology for
calculating emissions from each of these waste incineration sources is provided in Annex 3.7.
For each of the methods used to calculate C02 emissions from the incineration of waste, data on the quantity of
product combusted and the C content of the product are needed. For plastics, synthetic rubber, and synthetic
fibers in MSW, the amount of specific materials discarded as MSW (i.e., the quantity generated minus the quantity
recycled) was taken from Municipal Solid Waste Generation, Recycling, and Disposal in the United States: Facts and
Figures (EPA 2000 through 2003, 2005 through 2014), and Advancing Sustainable Materials Management: Facts
and Figures: Assessing Trends in Material Generation, Recycling and Disposal in the United States (EPA 2015; EPA
2016; EPA 2018a; EPA 2019; EPA 2020) and detailed unpublished backup data for some years not shown in the
reports (Schneider 2007). For 2012 through 2019 data on total waste incinerated were assumed to equal to the
2011 value from Shin (2014). For synthetic rubber and carbon black in scrap tires, information was obtained
biannually from U.S. Scrap Tire Management Summary for 2005 through 2019 data (RMA 2020). Average C
contents for the "Other" plastics category and synthetic rubber in MSW were calculated from 1998 and 2002
production statistics; C content for 1990 through 1998 is based on the 1998 value; C content for 1999 through
3-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2001 is the average of 1998 and 2002 values; and C content for 2002 through 2019 is based on the 2002 value.
Carbon content for synthetic fibers was calculated from a weighted average of production statistics from 1990
through 2019. Information about scrap tire composition was taken from the Rubber Manufacturers' Association
internet site (RMA 2012a). The mass of incinerated material is multiplied by its C content to calculate the total
amount of carbon stored.
The assumption that 98 percent of organic C is oxidized (which applies to all waste incineration categories for C02
emissions) was reported in EPA's life cycle analysis of greenhouse gas emissions and sinks from management of
solid waste (EPA 2006). This percentage is multiplied by the carbon stored to estimate the amount of carbon
emitted.
Incineration of waste, including MSW, also results in emissions of CH4and N20. These emissions were calculated as
a function of the total estimated mass of waste incinerated and emission factors. As noted above, CH4 and N20
emissions are a function of total waste incinerated in each year; for 1990 through 2008, these data were derived
from the information published in BioCycle (van Haaren et al. 2010). Data for 2009 and 2010 were interpolated
between 2008 and 2011 values. Data for 2011 were derived from Shin (2014). Data on total waste incinerated was
not available in the BioCycle data set for 2012 through 2019, so these values were assumed to equal the 2011
BioCycle dataset value.
Table 3-27 provides data on MSW discarded and percentage combusted for the total waste stream. The emission
factors of N20 and CH4 emissions per quantity of MSW combusted are default emission factors for the default
continuously-fed stoker unit MSW incineration technology type and were taken from IPCC (2006).
Table 3-27: Municipal Solid Waste Generation (Metric Tons) and Percent Combusted
(BioCycle dataset)
Incinerated (% of
Year	Waste Discarded	Waste Incinerated	Discards)
1990	235,733,657	30,632,057	13.0%
2005	259,559,787	25,973,520	10.0%
2015	273,116,704a	20,756,870	7.6%
2016	273,116,704a	20,756,870	7.6%
2017	273,116,704a	20,756,870	7.6%
2018	273,116,704a	20,756,870	7.6%
2019	273,116,704a	20,756,870	7.6%
a Assumed equal to 2011 value.
Source: van Haaren et al. (2010), Shin (2014).
Uncertainty and Time-Series Consistency
An Approach 2 Monte Carlo analysis was performed to determine the level of uncertainty surrounding the
estimates of C02 emissions and N20 emissions from the incineration of waste (given the very low emissions for
CH4, no uncertainty estimate was derived). IPCC Approach 2 analysis allows the specification of probability density
functions for key variables within a computational structure that mirrors the calculation of the Inventory estimate.
Uncertainty estimates and distributions for waste generation variables (i.e., plastics, synthetic rubber, and textiles
generation) were obtained through a conversation with one of the authors of the Municipal Solid Waste in the
United States reports. Statistical analyses or expert judgments of uncertainty were not available directly from the
information sources for the other variables; thus, uncertainty estimates for these variables were determined using
Energy 3-59

-------
assumptions based on source category knowledge and the known uncertainty estimates for the waste generation
variables.
The uncertainties in the waste incineration emission estimates arise from both the assumptions applied to the data
and from the quality of the data. Key factors include MSW incineration rate; fraction oxidized; missing data on
waste composition; average C content of waste components; assumptions on the synthetic/biogenic C ratio; and
combustion conditions affecting N20 emissions. The highest levels of uncertainty surround the variables that are
based on assumptions (e.g., percent of clothing and footwear composed of synthetic rubber); the lowest levels of
uncertainty surround variables that were determined by quantitative measurements (e.g., combustion efficiency, C
content of C black).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-28. Waste incineration
C02 emissions in 2019 were estimated to be between 8.6 and 14.5 MMT C02 Eq. at a 95 percent confidence level.
This indicates a range of 25 percent below to 27 percent above the 2019 emission estimate of 11.5 MMT C02 Eq.
Also at a 95 percent confidence level, waste incineration N20 emissions in 2019 were estimated to be between 0.2
and 1.3 MMT C02 Eq. This indicates a range of 50 percent below to 325 percent above the 2019 emission estimate
of 0.3 MMT C02 Eq.
Table 3-28: Approach 2 Quantitative Uncertainty Estimates for CO2 and N2O from the
Incineration of Waste (MMT CO2 Eq. and Percent)


2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
Source
Gas
(MMTCOz Eq.)
(MMTCOz
Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Incineration of Waste
C02
11.5
8.6
14.5
-25%
27%
Incineration of Waste
N20
0.3
0.2
1.3
-50%
325%
a Range of emission estimates predicted by Monte Carlo Simulation for a 95 percent confidence interval.
QA/QC and Verification
In order to ensure the quality of the emission estimates from waste incineration, general (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved
checks specifically focusing on the activity data and specifically focused on the emission factor and activity data
sources and methodology used for estimating emissions from incineration of waste. Trends across the time series
were analyzed to determine whether any corrective actions were needed. Corrective actions were taken to rectify
minor errors in the use of activity data.
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
The waste incineration estimates have recently relied on MSW mass flow (i.e., tonnage) data that has not been
updated since 2011. These values come from BioCycle (Shin 2014) and EPA Facts and Figures (EPA 2015). EPA
performed an examination of facility-level MSW tonnage data availability, primarily focusing on EPA's GHGRP data,
Energy Information Administration (EIA) waste-to-energy data, and other sources. EPA concluded that the MSW
mass flow of waste incinerated can be derived from GHGRP data and that the GHGRP dataset is the most complete
dataset (i.e., includes the most facilities), but does not contain data for all inventory years (1990 through 2010).
The EIA data can be used to supplement years not available in the GHGRP dataset and corroborate MSW mass flow
3-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
tonnage obtained for years in which GHGRP data are available. These MSW mass flow tonnages currently influence
calculations for C02 and non-C02 emissions.
Additional improvements will focus on investigating new methods and sources for C02 emission estimates. As part
of the Public Review process of this year's Inventory cycle, EPA proposed a new method for calculating emissions
associated with waste incineration. The proposed method relied on MSW tonnage estimates back calculated from
GHGRP reporting data and MSW assumed carbon content factors based on the EPA's Facts and Figures Reports.
Based on review and discussions with industry representatives it was felt that the proposed approach could lead to
an overestimate of fossil carbon content of waste combusted. Therefore, the approach used here reverts to the
existing methodology used in past calculations.
Future, proposed improvements to the current C02 emissions estimation methodology build off the work done for
the proposed approach and include the calculation of an overall carbon content for MSW incinerated. GHGRP and
EIA both provide emissions information for C02, which will allow EPA to calculate an overall carbon content of
MSW incinerated and apply this to MSW mass flows. Further research is required to compare the carbon contents
of MSW incinerated from GHGRP and EIA.
Currently, emission estimates for the biomass and biomass-based fuels source category included in this Inventory
are limited to woody biomass, ethanol, and biodiesel. EPA will incorporate emissions from biogenic components of
MSW to biomass and biomass-based fuels or waste incineration in future Inventory assessments.
3.4 Coal Mining (CRF Source Category lBla)
Three types of coal mining-related activities release CH4 to the atmosphere: underground mining, surface mining,
and post-mining (i.e., coal-handling) activities. While surface mines account for the majority of U.S. coal
production, underground coal mines contribute the largest share of CH4 emissions (see Table 3-30 and Table 3-31)
due to the higher CH4 content of coal in the deeper underground coal seams. In 2019, 226 underground coal mines
and 432 surface mines were operating in the United States (EIA 2020). In recent years, the total number of active
coal mines in the United States has declined. In 2019, the United States was the third-largest coal producer in the
world (640 MMT), after China (3,693 MMT) and India (769 MMT) (IEA 2020).
Table 3-29: Coal Production (kt)
Year
Underground
Surface

Total


Number of Mines
Production
Number of Mines
Production
Number of Mines
Production
1990
1,683
384,244
1,656
546,808
3,339
931,052

2005
586
334,399
789
691,447
1,398
1,025,846

2015
305
278,344
529
534,092
834
812,435
2016
251
228,707
439
431,282
690
659,989
2017
237
247,778
434
454,301
671
702,080
2018
236
249,804
430
435,521
666
685,325
2019
226
242,557
432
397,750
658
640,307
Underground mines liberate CH4 from ventilation systems and from degasification systems. Ventilation systems
pump air through the mine workings to dilute noxious gases and ensure worker safety; these systems can exhaust
significant amounts of CH4 to the atmosphere in low concentrations. Degasification systems are wells drilled from
the surface or boreholes drilled inside the mine that remove large, often highly concentrated volumes of CH4
before, during, or after mining. Some mines recover and use CH4 generated from ventilation and degasification
systems, thereby reducing emissions to the atmosphere.
Surface coal mines liberate CH4 as the overburden is removed and the coal is exposed to the atmosphere. Methane
emissions are normally a function of coal rank (a classification related to the percentage of carbon in the coal) and
Energy 3-61

-------
depth. Surface coal mines typically produce lower-rank coals and remove less than 250 feet of overburden, so their
level of emissions is much lower than from underground mines.
In addition, CH4 is released during post-mining activities, as the coal is processed, transported, and stored for use.
Total CH4 emissions in 2019 were estimated to be 1,895 kt (47.4 MMT C02 Eq.), a decline of approximately 51
percent since 1990 (see Table 3-30 and Table 3-31). In 2019, underground mines accounted for approximately 73
percent of total emissions, surface mines accounted for 13 percent, and post-mining activities accounted for 14
percent. In 2019, total CH4 emissions from coal mining decreased by approximately 10 percent relative to the
previous year. This decrease was due to a decrease in annual coal production and a decrease in reported annual
ventilation emissions.67 The amount of CH4 recovered and used in 2019 decreased by approximately 17 percent
compared to 2018 levels. In 2019, all but two mines reported lower levels of CH4 recovered and used compared to
2018 levels.
Table 3-30: ChU Emissions from Coal Mining (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Underground (UG) Mining
74.2
42.0
44.9
40.7
40.7
38.9
34.5
Liberated
80.8
59.7
61.2
57.1
58.1
57.7
50.1
Recovered & Used
(6.6)
(17.7)
(16.4)
(16.4)
(17.4)
(18.8)
(15.7)
Surface Mining
10.8
11.9
r>
CO
6.8
7.2
7.0
6.4
Post-Mining (UG)
9.2
7.6
5.8
4.8
5.3
5.3
5.2
Post-Mining (Surface)
2.3
2.6
1.9
1.5
1.6
1.5
1.4
Total
96.5
64.1
61.2
53.8
54.8
52.7
47.4
Table 3-31: ChU Emissions from Coal Mining (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Underground (UG) Mining
2,968
1,682
1,796
1,629
1,626
1,556
1,379
Liberated
3,231
2,388
2,450
2,283
2,324
2,307
2,005
Recovered & Used
(263)
(706)
(654)
(654)
(698)
(751)
(627)
Surface Mining
430
475
347
273
290
280
255
Post-Mining (UG)
368
306
231
193
213
212
206
Post-Mining (Surface)
93
103
75
59
63
61
55
Total
3,860
2,565
2,449
2,154
2,191
2,109
1,895
Methodology
EPA uses an IPCC Tier 3 method for estimating CH4 emissions from underground coal mining and an IPCC Tier 2
method for estimating CH4 emissions from surface mining and post-mining activities (for both coal production from
underground mines and surface mines). The methodology for estimating CH4 emissions from coal mining consists
of two steps:
•	Estimate CH4 emissions from underground mines. These emissions have two sources: ventilation systems
and degasification systems. They are estimated using mine-specific data, then summed to determine total
CH4 liberated. The CH4 recovered and used is then subtracted from this total, resulting in an estimate of
net emissions to the atmosphere.
•	Estimate CH4 emissions from surface mines and post-mining activities. Unlike the methodology for
underground mines, which uses mine-specific data, the methodology for estimating emissions from
67 This indicates lower underground mine activity, which is supported by EIA coal production data for 2019 (reduction in
production compared to 2018 and 2017).
3-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
surface mines and post-mining activities consists of multiplying basin-specific coal production by basin-
specific gas content and an emission factor.
Step 1: Estimate CH4 Liberated and CH4 Emitted from Underground Mines
Underground mines generate CH4 from ventilation systems and degasification systems. Some mines recover and
use the liberated CH4, thereby reducing emissions to the atmosphere. Total CH4 emitted from underground mines
equals the CH4 liberated from ventilation systems, plus the CH4 liberated from degasification systems, minus the
CH4 recovered and used.
Step 1.1: Estimate CH4 Liberated from Ventilation Systems
To estimate CH4 liberated from ventilation systems, EPA uses data collected through its Greenhouse Gas Reporting
Program (GHGRP)68 (Subpart FF, "Underground Coal Mines"), data provided by the U.S. Mine Safety and Health
Administration (MSHA) (MSHA 2020), and occasionally data collected from other sources on a site-specific level
(e.g., state gas production databases). Since 2011, the nation's "gassiest" underground coal mines—those that
liberate more than 36,500,000 actual cubic feet of CH4 per year (about 17,525 MT C02 Eq.)—have been required to
report to EPA's GHGRP (EPA 2020).69 Mines that report to EPA's GHGRP must report quarterly measurements of
CH4 emissions from ventilation systems; they have the option of recording and reporting their own measurements,
or using the measurements taken by MSHA as part of that agency's quarterly safety inspections of all mines in the
United States with detectable CH4 concentrations.70
Since 2013, ventilation CH4 emission estimates have been calculated based on both quarterly GHGRP data
submitted by underground mines and on quarterly measurement data obtained directly from MSHA. Because not
all mines report under EPA's GHGRP, the emissions of the mines that do not report must be calculated using MSHA
data. The MSHA data also serves as a quality assurance tool for validating GHGRP data. For GHGRP data, reported
quarterly ventilation methane emissions (metric tons) are summed for each mine to develop mine-specific annual
ventilation emissions. For MSHA data, the average daily CH4 emission rate for each mine is determined using the
CH4 total for all data measurement events conducted during the calendar year and total duration of all data
measurement events (in days). The calculated average daily CH4 emission rate is then multiplied by 365 days to
estimate annual ventilation CH4 emissions for the MSHA dataset.
Step 1.2: Estimate CH4 Liberated from Degasification Systems
Particularly gassy underground mines also use degasification systems (e.g., wells or boreholes) to remove CH4
before, during, or after mining. This CH4 can then be collected for use or vented to the atmosphere. Nineteen
mines used degasification systems in 2019 and 17 of these mines reported the CH4 removed through these
systems to EPA's GHGRP under Subpart FF (EPA 2020). Based on the weekly measurements reported to EPA's
GHGRP, degasification data summaries for each mine are added to estimate the CH4 liberated from degasification
systems. Thirteen of the 19 mines with degasification systems had operational CH4 recovery and use projects (see
step 1.3 below).71
68	In implementing improvements and integrating data from EPA's GHGRP, EPA followed the latest guidance from the IPCC on
the use of facility-level data in national inventories (IPCC 2011).
69	Underground coal mines report to EPA under Subpart FF of the GHGRP (40 CFR Part 98). In 2019, 65 underground coal mines
reported to the program.
70	MSHA records coal mine CH4 readings with concentrations of greater than 50 ppm (parts per million) CH4. Readings below
this threshold are considered non-detectable.
71	Several of the mines venting CH4 from degasification systems use a small portion of the gas to fuel gob well blowers in
remote locations where electricity is not available. However, this CH4 use is not considered to be a formal recovery and use
project.
Energy 3-63

-------
Degasification data reported to EPA's GHGRP by underground coal mines is the primary source of data used to
develop estimates of CH4 liberated from degasification systems. Data reported to EPA's GHGRP were used
exclusively to estimate CH4 liberated from degasification systems at 14 of the 19 mines that used degasification
systems in 2019. Data from state gas well production databases were used exclusively for two mines and state gas
well production data were used to supplement GHGRP degasification data for the remaining three mines (DMME
2020, GSA 2020, and WVGES 2020).
For pre-mining wells, cumulative degasification volumes that occur prior to the well being mined through are
attributed to the mine in the inventory year in which the well is mined through.72 EPA's GHGRP does not require
gas production from virgin coal seams (coalbed methane) to be reported by coal mines under Subpart FF.73 Most
pre-mining wells drilled from the surface are considered coalbed methane wells prior to mine-through and
associated CH4 emissions are reported under another subpart of the GHGRP (Subpart W, "Petroleum and Natural
Gas Systems"). As a result, GHGRP data must be supplemented to estimate cumulative degasification volumes that
occurred prior to well mine-through. There were five mines with degasification systems that include pre-mining
wells that were mined through in 2019. For three of these mines, GHGRP data were supplemented with historical
data from state gas well production databases (GSA 2020 and WVGES 2020), as well as with mine-specific
information regarding the locations and dates on which the pre-mining wells were mined through (JWR 2010; El
Paso 2009; ERG 2020). State gas well production data were exclusively used for the remaining two mines (DMME
2020 and GSA 2020).
Step 1.3: Estimate CH4 Recovered from Ventilation and Degasification Systems, and Utilized or
Destroyed (Emissions Avoided)
Thirteen mines had CH4 recovery and use projects in place in 2019, including one mine that had two recovery and
use projects. Thirteen of these projects involved degasification systems and one involved a ventilation air methane
abatement project (VAM). Eleven of these mines sold the recovered CH4 to a pipeline, including one that also used
CH4 to fuel a thermal coal dryer. One mine used recovered CH4to heat mine ventilation air (data were unavailable
for estimating CH4 recovery at this mine). One mine destroyed the recovered CH4 (VAM) using Regenerative
Thermal Oxidation (RTO) without energy recovery.
The CH4 recovered and used (or destroyed) at the twelve mines described above for which data were available are
estimated using the following methods:
•	EPA's GHGRP data was exclusively used to estimate the CH4 recovered and used from seven of the 12
mines that deployed degasification systems in 2019. Based on weekly measurements, the GHGRP
degasification destruction data summaries for each mine are added together to estimate the CH4
recovered and used from degasification systems.
•	State sales data were used to estimate CH4 recovered and used from the remaining five mines that
deployed degasification systems in 2019 (DMME 2020, GSA 2020). These five mines intersected pre-
mining wells in 2019. Supplemental information is used for these mines because estimating CH4 recovery
and use from pre-mining wells requires additional data not reported under Subpart FF of EPA's GHGRP
(see discussion in step 1.2 above) to account for the emissions avoided prior to the well being mined
through. The supplemental data is obtained from state gas production databases as well as mine-specific
information on the timing of mined-through pre-mining wells.
•	For the single mine that employed VAM for CH4 recovery and use, the estimates of CH4 recovered and
used were obtained from the mine's offset verification statement (OVS) submitted to the California Air
Resources Board (CARB) (McElroy OVS 2020).
72 A well is "mined through" when coal mining development or the working face intersects the borehole or well.
This applies for pre-drainage in years prior to the well being mined through. Beginning with the year the well is mined
through, the annual volume of CH4 liberated from a pre-drainage well is reported under Subpart FF of EPA's GHGRP.
3-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Step 2: Estimate CH4 Emitted from Surface Mines and Post-Mining Activities
Mine-specific data are not available for estimating CH4 emissions from surface coal mines or for post-mining
activities. For surface mines, basin-specific coal production obtained from the Energy Information Administration's
Annual Coal Report (EIA 2020) is multiplied by basin-specific CH4 contents (EPA 1996, 2005) and a 150 percent
emission factor (to account for CH4from over- and under-burden) to estimate CH4 emissions (King 1994, Saghafi
2013). For post-mining activities, basin-specific coal production is multiplied by basin-specific CH4 contents and a
mid-range 32.5 percent emission factor for CH4 desorption during coal transportation and storage (Creedy 1993).
Basin-specific in situ gas content data were compiled from AAPG (1984) and USBM (1986).
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis was conducted for the coal mining source category using the IPCC-
recommended Approach 2 uncertainty estimation methodology. Because emission estimates from underground
ventilation systems were based on actual measurement data from EPA's GHGRP or from MSHA, uncertainty is
relatively low. A degree of imprecision was introduced because the ventilation air measurements used were not
continuous but rather quarterly instantaneous readings that were used to determine the average annual emission
rates. Additionally, the measurement equipment used can be expected to have resulted in an average of 10
percent overestimation of annual CH4 emissions (Mutmansky & Wang 2000). Equipment measurement uncertainty
is applied to GHGRP data.
Estimates of CH4 liberated and recovered by degasification systems are relatively certain for utilized CH4 because of
the availability of EPA's GHGRP data and gas sales information. Many of the liberation and recovery estimates use
data on wells within 100 feet of a mined area. However, uncertainty exists concerning the radius of influence of
each well. The number of wells counted, and thus the liberated CH4 and avoided emissions, may vary if the
drainage area is found to be larger or smaller than estimated.
EPA's GHGRP requires weekly CH4 monitoring of mines that report degasification systems, and continuous CH4
monitoring is required for CH4 utilized on- or off-site. Since 2012, GHGRP data have been used to estimate CH4
emissions from vented degasification wells, reducing the uncertainty associated with prior MSHA estimates used
for this sub-source. Beginning in 2013, GHGRP data were also used for determining CH4 recovery and use at mines
without publicly available gas usage or sales records, which has reduced the uncertainty from previous estimation
methods that were based on information from coal industry contacts.
Surface mining and post-mining emissions are associated with considerably more uncertainty than underground
mines, because of the difficulty in developing accurate emission factors from field measurements. However, since
underground emissions constitute the majority of total coal mining emissions, the uncertainty associated with
underground emissions is the primary factor that determines overall uncertainty.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-32. Coal mining CH4
emissions in 2019 were estimated to be between 43.2 and 57.0 MMT C02 Eq. at a 95 percent confidence level. This
indicates a range of 8.8 percent below to 20.3 percent above the 2019 emission estimate of 47.4 MMT C02 Eq.
Table 3-32: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Coal
Mining (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Coal Mining
ch4
47.4
43.2 57.0
-8.8% +20.3%
a Range of emission estimates predicted by Monte Carlo stochastic simulation for a 95 percent confidence interval.
Energy 3-65

-------
QA/QC and Verification
In order to ensure the quality of the emission estimates for coal mining, general (IPCC Tier 1) and category-specific
(Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent with the U.S.
Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved checks
specifically focusing on the activity data and reported emissions data used for estimating emissions from coal
mining. Trends across the time series were analyzed to determine whether any corrective actions were needed.
Emission estimates for coal mining rely in large part on data reported by coal mines to EPA's GHGRP. EPA verifies
annual facility-level reports through a multi-step process to identify potential errors and ensure that data
submitted to EPA are accurate, complete, and consistent. All reports submitted to EPA are evaluated by electronic
validation and verification checks. If potential errors are identified, EPA will notify the reporter, who can resolve
the issue either by providing an acceptable response describing why the flagged issue is not an error or by
correcting the flagged issue and resubmitting their annual report. Additional QA/QC and verification procedures
occur for each GHGRP subpart.
Recalculations Discussion
State gas sales production values were updated for one mine for 2011 and 2012, and for 1994 to 2018 for another
mine, as part of normal updates. These changes resulted in slightly higher degasification CH4 emissions and CH4
emissions avoided from underground mining. The change in both the degasification emissions and emissions
avoided is less than 0.5 percent over the 1994 to 2018 time series, compared to the previous Inventory.
Annual coal production numbers were updated for 2001 to 2018 based on revised data from EIA. The previously
used coal production numbers were revised by EIA, primarily for the Appalachian basins. This update resulted in
changes to surface mining and post-surface mining emissions for 2001 to 2018. The change in emissions averaged
an increase of approximately 9 percent over the 2001 to 2018 time series. The highest change was in 2007 (11.4
percent) and the lowest change was in 2015 (5.3 percent), compared to the previous Inventory.
Planned Improvements
EPA intends to include estimating fugitive C02 emissions from underground and surface mining, based on methods
included in the 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories.
3.5 Abandoned Underground Coal Mines (CRF
Source Category lBla)
Underground coal mines contribute the largest share of coal mine methane (CMM) emissions, with active
underground mines the leading source of underground emissions. However, mines also continue to release CH4
after closure. As mines mature and coal seams are mined through, mines are closed and abandoned. Many are
sealed and some flood through intrusion of groundwater or surface water into the void. Shafts or portals are
generally filled with gravel and capped with a concrete seal, while vent pipes and boreholes are plugged in a
manner similar to oil and gas wells. Some abandoned mines are vented to the atmosphere to prevent the buildup
of CH4 that may find its way to surface structures through overburden fractures. As work stops within the mines,
CH4 liberation decreases but it does not stop completely. Following an initial decline, abandoned mines can
liberate CH4 at a near-steady rate over an extended period of time, or, if flooded, produce gas for only a few years.
The gas can migrate to the surface through the conduits described above, particularly if they have not been sealed
adequately. In addition, diffuse emissions can occur when CH4 migrates to the surface through cracks and fissures
in the strata overlying the coal mine. The following factors influence abandoned mine emissions:
3-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Time since abandonment;
•	Gas content and adsorption characteristics of coal;
•	CH4 flow capacity of the mine;
•	Mine flooding;
•	Presence of vent holes; and
•	Mine seals.
Annual gross abandoned mine CH4 emissions ranged from 7.2 to 10.8 MMT C02 Eq. from 1990 to 2019, varying, in
general, by less than 1 percent to approximately 19 percent from year to year. Fluctuations were due mainly to the
number of mines closed during a given year as well as the magnitude of the emissions from those mines when
active. Gross abandoned mine emissions peaked in 1996 (10.8 MMT C02 Eq.) due to the large number of gassy
mine74 closures from 1994 to 1996 (72 gassy mines closed during the three-year period). In spite of this rapid rise,
abandoned mine emissions have been generally on the decline since 1996. Since 2002, there have been fewer than
twelve gassy mine closures each year. In 2019 there were no gassy mine closures. Gross abandoned mine
emissions decreased slightly from 8.9 MMT C02 Eq. (355 kt CH4) in 2018 to 8.5 (341 kt CH4) MMT C02 Eq. in 2019
(see Table 3-33 and Table 3-34). Gross emissions are reduced by CH4 recovered and used at 45 mines, resulting in
net emissions in 2019 of 5.9 MMT C02 Eq.
Table 3-33: ChU Emissions from Abandoned Coal Mines (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Abandoned Underground Mines
7.2
8.4
9.0
9.5
9.2
8.9
8.5
Recovered & Used
0.0
(1.8)
(2.6)
(2.8)
(2.7)
(2.7)
(2.6)
Total
7.2
6.6
6.4
6.7
6.4
6.2
5.9
Note: Parentheses indicate negative values.
+ Does not exceed 0.05 MMT C02 Eq.
Table 3-34: ChU Emissions from Abandoned Coal Mines (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Abandoned Underground Mines
288
334
359
380
367
355
341
Recovered & Used
0.0
(70)
(102)
(112)
(109)
(107)
(104)
Total
288
264
256
268
257
247
237
+ Does not exceed 0.5 kt.
Methodology
Estimating CH4 emissions from an abandoned coal mine requires predicting the emissions of a mine from the time
of abandonment through the inventory year of interest. The flow of CH4 from the coal to the mine void is primarily
dependent on the mine's emissions when active and the extent to which the mine is flooded or sealed. The CH4
emission rate before abandonment reflects the gas content of the coal, the rate of coal mining, and the flow
capacity of the mine in much the same way as the initial rate of a water-free conventional gas well reflects the gas
content of the producing formation and the flow capacity of the well. A well or a mine that produces gas from a
coal seam and the surrounding strata will produce less gas through time as the reservoir of gas is depleted.
Depletion of a reservoir will follow a predictable pattern depending on the interplay of a variety of natural physical
conditions imposed on the reservoir. The depletion of a reservoir is commonly modeled by mathematical
equations and mapped as a type curve. Type curves, which are referred to as decline curves, have been developed
74 A mine is considered a "gassy" mine if it emits more than 100 thousand cubic feet of CH4 per day (100 mcfd).
Energy 3-67

-------
for abandoned coal mines. Existing data on abandoned mine emissions through time, although sparse, appear to
fit the hyperbolic type of decline curve used in forecasting production from natural gas wells.
In order to estimate CH4 emissions over time for a given abandoned mine, it is necessary to apply a decline
function, initiated upon abandonment, to that mine. In the analysis, mines were grouped by coal basin with the
assumption that they will generally have the same initial pressures, permeability and isotherm. As CH4 leaves the
system, the reservoir pressure (Pr) declines as described by the isotherm's characteristics. The emission rate
declines because the mine pressure (Pw) is essentially constant at atmospheric pressure for a vented mine, and the
productivity index (PI), which is expressed as the flow rate per unit of pressure change, is essentially constant at
the pressures of interest (atmospheric to 30 psia). The CH4 flow rate is determined by the laws of gas flow through
porous media, such as Darcy's Law. A rate-time equation can be generated that can be used to predict future
emissions. This decline through time is hyperbolic in nature and can be empirically expressed as:
q = qt (1 + bDity-W
where,
q	=	Gas flow rate at time t in million cubic feet per day (mmcfd)
q,	=	Initial gas flow rate at time zero (t0), mmcfd
b	=	The hyperbolic exponent, dimensionless
D,	=	Initial decline rate, 1/year
t	=	Elapsed time from t0 (years)
This equation is applied to mines of various initial emission rates that have similar initial pressures, permeability
and adsorption isotherms (EPA 2004).
The decline curves created to model the gas emission rate of coal mines must account for factors that decrease the
rate of emissions after mining activities cease, such as sealing and flooding. Based on field measurement data, it
was assumed that most U.S. mines prone to flooding will become completely flooded within eight years and
therefore will no longer have any measurable CH4 emissions. Based on this assumption, an average decline rate for
flooded mines was established by fitting a decline curve to emissions from field measurements. An exponential
equation was developed from emissions data measured at eight abandoned mines known to be filling with water
located in two of the five basins. Using a least squares, curve-fitting algorithm, emissions data were matched to
the exponential equation shown below. For this analysis of flooded abandoned mines, there was not enough data
to establish basin-specific equations, as was done with the vented, non-flooding mines (EPA 2004). This decline
through time can be empirically expressed as:
,(-Dt)
q = q^
where,
q	=	Gas flow rate at time t in mmcfd
q,	=	Initial gas flow rate at time zero (t0), mmcfd
D	=	Decline rate, 1/year
t	=	Elapsed time from t0 (years)
Seals have an inhibiting effect on the rate of flow of CH4 into the atmosphere compared to the flow rate that
would exist if the mine had an open vent. The total volume emitted will be the same, but emissions will occur over
a longer period of time. The methodology, therefore, treats the emissions prediction from a sealed mine similarly
to the emissions prediction from a vented mine, but uses a lower initial rate depending on the degree of sealing. A
computational fluid dynamics simulator was used with the conceptual abandoned mine model to predict the
decline curve for inhibited flow. The percent sealed is defined as 100 x (1 - [initial emissions from sealed mine /
emission rate at abandonment prior to sealing]). Significant differences are seen between 50 percent, 80 percent
and 95 percent closure. These decline curves were therefore used as the high, middle, and low values for
emissions from sealed mines (EPA 2004).
For active coal mines, those mines producing over 100 thousand cubic feet per day (mcfd) of CH4 account for about
98 percent of all CH4 emissions. This same relationship is assumed for abandoned mines. It was determined that
3-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
the 527 abandoned mines closed after 1972 produced CH4 emissions greater than 100 mcfd when active. Further,
the status of 304 of the 527 mines (or 58 percent) is known to be either: 1) vented to the atmosphere; 2) sealed to
some degree (either earthen or concrete seals); or 3) flooded (enough to inhibit CH4 flow to the atmosphere). The
remaining 42 percent of the mines whose status is unknown were placed in one of these three categories by
applying a probability distribution analysis based on the known status of other mines located in the same coal
basin (EPA 2004). Table 3-35 presents the count of mines by post-abandonment state, based on EPA's probability
distribution analysis.
Table 3-35: Number of Gassy Abandoned Mines Present in U.S. Basins in 2019, Grouped by
Class According to Post-Abandonment State
Total
Basin
Sealed
Vented
Flooded
Known
Unknown Total Mines
Central Appl.
42
26
52
120
146
266
Illinois
34
3
14
51
31
82
Northern Appl.
47
22
16
85
39
124
Warrior Basin
0
0
16
16
0
16
Western Basins
28
4
2
34
10
44
Total
151
55
100
306
226
532
Inputs to the decline equation require the average CH4 emission rate prior to abandonment and the date of
abandonment. Generally, these data are available for mines abandoned after 1971; however, such data are largely
unknown for mines closed before 1972. Information that is readily available, such as coal production by state and
county, is helpful but does not provide enough data to directly employ the methodology used to calculate
emissions from mines abandoned before 1972. It is assumed that pre-1972 mines are governed by the same
physical, geologic, and hydrologic constraints that apply to post-1971 mines; thus, their emissions may be
characterized by the same decline curves.
During the 1970s, 78 percent of CH4 emissions from coal mining came from seventeen counties in seven states.
Mine closure dates were obtained for two states, Colorado and Illinois, for the hundred-year period extending
from 1900 through 1999. The data were used to establish a frequency of mine closure histogram (by decade) and
applied to the other five states with gassy mine closures. As a result, basin-specific decline curve equations were
applied to the 145 gassy coal mines estimated to have closed between 1920 and 1971 in the U.S., representing 78
percent of the emissions. State-specific, initial emission rates were used based on average coal mine CH4 emissions
rates during the 1970s (EPA 2004).
Abandoned mine emission estimates are based on all closed mines known to have active mine CH4 ventilation
emission rates greater than 100 mcfd at the time of abandonment. For example, for 1990 the analysis included 145
mines closed before 1972 and 258 mines closed between 1972 and 1990. Initial emission rates based on MSHA
reports, time of abandonment, and basin-specific decline curves influenced by a number of factors were used to
calculate annual emissions for each mine in the database (MSHA 2020). Coal mine degasification data are not
available for years prior to 1990, thus the initial emission rates used reflect only ventilation emissions for pre-1990
closures. CH4 degasification amounts were added to the quantity of CH4 vented to determine the total CH4
liberation rate for all mines that closed between 1992 and 2019. Since the sample of gassy mines described above
is assumed to account for 78 percent of the pre-1972 and 98 percent of the post-1971 abandoned mine emissions,
the modeled results were multiplied by 1.22 and 1.02 to account for all U.S. abandoned mine emissions.
From 1993 through 2019, emission totals were downwardly adjusted to reflect CH4 emissions avoided from those
abandoned mines with CH4 recovery and use or destruction systems. The Inventory totals were not adjusted for
abandoned mine CH4 emissions avoided from 1990 through 1992, because no data was reported for abandoned
coal mine CH4 recovery and use or destruction projects during that time.
Energy 3-69

-------
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis was conducted to estimate the uncertainty surrounding the estimates of CH4
emissions from abandoned underground coal mines. The uncertainty analysis provides for the specification of
probability density functions for key variables within a computational structure that mirrors the calculation of the
Inventory estimate. The results provide the range within which, with 95 percent certainty, emissions from this
source category are likely to fall.
As discussed above, the parameters for which values must be estimated for each mine in order to predict its
decline curve are: 1) the coal's adsorption isotherm; 2) CH4 flow capacity as expressed by permeability; and 3)
pressure at abandonment. Because these parameters are not available for each mine, a methodological approach
to estimating emissions was used that generates a probability distribution of potential outcomes based on the
most likely value and the probable range of values for each parameter. The range of values is not meant to capture
the extreme values, but rather values that represent the highest and lowest quartile of the cumulative probability
density function of each parameter. Once the low, mid, and high values are selected, they are applied to a
probability density function.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-36. Annual abandoned
coal mine CH4 emissions in 2019 were estimated to be between 4.6 and 7.1 MMT C02 Eq. at a 95 percent
confidence level. This indicates a range of 22 percent below to 19 percent above the 2019 emission estimate of 5.9
MMT C02 Eq. One of the reasons for the relatively narrow range is that mine-specific data is available for use in the
methodology for mines closed in 1972 and later years. Emissions from mines closed prior to 1972 have the largest
degree of uncertainty because no mine-specific CH4 liberation rates exist.
Table 3-36: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from
Abandoned Underground Coal Mines (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate9
(MMT C02 Eq.) (%)



Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Abandoned Underground
Coal Mines
ch4
5.9
4.6
7.1
-22.4% +19.2%
a Range of emission estimates predicted by Monte Carlo Simulation for a 95 percent confidence interval.
QA/QC and Verification
In order to ensure the quality of the emission estimates for abandoned coal mines, general (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved
checks specifically focusing on the activity data and reported emissions data used for estimating emissions from
abandoned coal mines. Trends across the time series were analyzed to determine whether any corrective actions
were needed.
3.6 Petroleum Systems (CRF Source Category
lB2a)	
This IPCC category (lB2a) is for fugitive emissions, which per IPCC include emissions from leaks, venting, and
flaring. Methane emissions from petroleum systems are primarily associated with onshore and offshore crude oil
3-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
production, transportation, and refining operations. During these activities, CH4 is released to the atmosphere as
emissions from leaks, venting (including emissions from operational upsets), and flaring. Carbon dioxide emissions
from petroleum systems are primarily associated with onshore and offshore crude oil production and refining
operations. Note, C02 emissions in Petroleum Systems exclude all combustion emissions (e.g., engine combustion)
except for flaring C02 emissions. All combustion C02 emissions (except for flaring) are accounted for in the fossil
fuel combustion chapter (see Section 3). Emissions of N20 from petroleum systems are primarily associated with
flaring. Total greenhouse gas emissions (CH4, C02, and N20) from petroleum systems in 2019 were 86.4 MMT C02
Eq., an increase of 47 percent from 1990, primarily due to increases in C02 emissions. Since 2009, total emissions
increased by 64 percent and since 2018, total emissions increased by 16 percent. Total C02 emissions from
petroleum systems in 2019 were 47.3 MMT C02 (47,269 kt C02), 3.9 times higher than in 1990. Total C02 emissions
in 2019 were 2.5 times higher than in 2009 and 27 percent higher than in 2018. Total CH4 emissions from
petroleum systems in 2019 were 39.1 MMT C02 Eq. (1,563 kt CH4), a decrease of 20 percent from 1990. Since
2009, total CH4 emissions increased by less than 0.5 percent; and since 2018, CH4 emissions increased by 5
percent. Total N20 emissions from petroleum systems in 2019 were 0.05 MMT C02 Eq. (0.16 kt N20), 1.8 times
higher than in 1990,1.5 times higher than in 2009, and 13 percent higher than in 2018. Since 1990, U.S. oil
production has increased by 67 percent. In 2019, production was 129 percent higher than in 2009 and 12 percent
higher than in 2018.
Each year, some estimates in the Inventory are recalculated with improved methods and/or data. These
improvements are implemented consistently across the previous Inventory's time series (i.e., 1990 to 2018) to
ensure that the trend is accurate. Recalculations in petroleum systems in this year's Inventory include:
•	Incorporation of an estimate for produced water
•	Updates to well counts using the most recent data from Enverus
•	Recalculations due to GHGRP submission revisions
The Recalculations Discussion section below provides more details on the updated methods.
Exploration. Exploration includes well drilling, testing, and completions. Exploration accounted for approximately 1
percent of total CH4 emissions (including leaks, vents, and flaring) from petroleum systems in 2019. The
predominant sources of CH4 emissions from exploration are hydraulically fractured oil well completions. Other
sources include well testing, well drilling, and well completions without hydraulic fracturing. Since 1990,
exploration CH4 emissions have decreased 91 percent, and while the number of hydraulically fractured wells
completed increased by a factor of 2.9, there were decreases in the fraction of such completions without reduced
emissions completions (RECs) or flaring (from 90 percent in 1990 to less than 1 percent in 2019). Emissions of CH4
from exploration were highest in 2012, nearly 30 times higher than in 2019; and lowest in 2019. Emissions of CH4
from exploration decreased 30 percent from 2018 to 2019, due to a decrease in hydraulically fractured oil well
completions with flaring. Exploration accounts for 4 percent of total emissions (including leaks, vents, and flaring)
from petroleum systems in 2019. Emissions of C02 from exploration in 2019 were 7 times higher than in 1990, and
decreased by 28 percent from 2018, due to a decrease in hydraulically fractured oil well completions with flaring.
Emissions of C02 from exploration were highest in 2014, around 33 percent higher than in 2019. Exploration
accounts for nearly 2 percent of total N20 emissions from petroleum systems in 2019. Emissions of N20 from
exploration in 2019 are 4.3 times higher than in 1990, and 37 percent lower than in 2018, due to the
abovementioned changes in hydraulically fractured oil well completions with flaring.
Production. Production accounted for 96 percent of total CH4 emissions (including leaks, vents, and flaring) from
petroleum systems in 2019. The predominant sources of emissions from production field operations are pneumatic
controllers, offshore oil platforms, gas engines, equipment leaks, produced water, and associated gas flaring. These
six sources together accounted for 82 percent of the CH4 emissions from production. Since 1990, CH4 emissions
from production have decreased by 16 percent due to decreases in emissions from offshore platforms, tanks, and
pneumatic controllers. Overall, production segment CH4 emissions increased by 5 percent from 2018 levels due
primarily to increased associated gas venting and flaring emissions in the Gulf Coast and Williston basins.
Production emissions account for 85 percent of the total C02 emissions (including leaks, vents, and flaring) from
petroleum systems in 2019. The principal sources of C02 emissions are associated gas flaring, miscellaneous
production flaring, and oil tanks with flares. These three sources together account for 98 percent of the C02
Energy 3-71

-------
emissions from production. In 2019, C02 emissions from production were 3.9 times higher than in 1990, due to
increases in flaring emissions from associated gas flaring, miscellaneous production flaring, and tanks. Overall,
production segment C02 emissions increased by 32 percent from 2018 levels primarily due to an increase in
associated gas flaring in the Williston Basin. Production emissions accounted for 67 percent of the total N20
emissions from petroleum systems in 2019. The principal sources of N20 emissions are associated gas flaring, oil
tanks with flares, and miscellaneous production flaring. In 2019, N20 emissions from production were 3.2 times
higher than in 1990 and 2.5 times higher than in 2009, due primarily to increases in N20 from associated gas
flaring. In 2019, N20 emissions from production increased by 6 percent from 2018 levels.
Crude Oil Transportation. Emissions from crude oil transportation account for a very small percentage of the total
emissions (including leaks, vents, and flaring) from petroleum systems and have little impact on the overall
emissions. Crude oil transportation activities account for less than 1 percent of total CH4 emissions from petroleum
systems. Emissions from tanks, marine loading, and truck loading operations account for 76 percent of CH4
emissions from crude oil transportation. Since 1990, CH4 emissions from transportation have increased by 40
percent. In 2019, CH4 emissions from transportation increased by 8 percent from 2018 levels. Crude oil
transportation activities account for less than 0.01 percent of total C02 emissions from petroleum systems.
Emissions from tanks, marine loading, and truck loading operations account for 76 percent of C02 emissions from
crude oil transportation.
Crude Oil Refining. Crude oil refining processes and systems account for 2 percent of total fugitive (including leaks,
vents, and flaring) CH4 emissions from petroleum systems. This low share is because most of the CH4 in crude oil is
removed or escapes before the crude oil is delivered to the refineries. There is an insignificant amount of CH4 in all
refined products. Within refineries, flaring accounts for 47 percent of the CH4 emissions, while uncontrolled
blowdowns, delayed cokers, and process vents account for 15,12, and 10 percent, respectively. Fugitive CH4
emissions from refining of crude oil have increased by 32 percent since 1990, and increased 16 percent from 2018;
however, like the transportation subcategory, this increase has had little effect on the overall emissions of CH4
from petroleum systems. Crude oil refining processes and systems account for 11 percent of total fugitive
(including leaks, vents, and flaring) C02 emissions from petroleum systems. Of the total fugitive C02 emissions
from refining, almost all (about 99 percent) of it comes from flaring.75 Since 1990, refinery fugitive C02 emissions
increased by 53 percent and have increased by 34 percent from the 2018 levels, due to an increase in flaring.
Flaring occurring at crude oil refining processes and systems accounts for 31 percent of total fugitive N20
emissions from petroleum systems. Refinery fugitive N20 emissions increased by 61 percent from 1990 to 2019
and increased by 40 percent from 2018 levels.
Table 3-37: Total Greenhouse Gas Emissions (CO2, ChU, and N2O) from Petroleum Systems
(MMT COz Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Exploration
3.3
4.9
4.2
1.7
2.1
3.3
2.4
Production
51.2
42.0
64.6
54.4
57.5
66.4
77.9
Transportation
0.2
0.1
0.2
0.2
0.2
0.2
0.2
Crude Refining
4.0
4.5
4.9
4.8
4.6
4.5
5.9
Total
58.6
51.5
73.9
61.1
64.4
74.5
86.4
Note: Totals may not sum due to independent rounding.
Table 3-38: ChU Emissions from Petroleum Systems (MMT CO2 Eq.)
Activity	1990	2005	2015 2016 2017 2018 2019
Exploration	3.0	4.5	2.1	0.5	0.4	0.4	0.3
75 Petroleum Systems includes fugitive emissions (leaks, venting, and flaring). In many industries, including petroleum
refineries, the largest source of onsite C02 emissions is often fossil fuel combustion, which is covered in section 3.1 of this
chapter.
3-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Production
45.1
34.0
38.4
37.7
38.0
35.9
37.7
Pneumatic Controllers
19.8
16.8
18.8
19.6
20.0
17.3
17.5
Offshore Production
9.3
6.5
5.5
5.1
5.1
5.0
5.0
Gas Engines
2.2
1.8
2.5
2.4
2.4
2.4
2.4
Equipment Leaks
2.0
2.0
2.5
2.4
2.4
2.4
2.3
Produced Water
2.3
1.6
2.1
1.9
2.0
2.1
2.1
Assoc Gas Flaring
0.5
0.4
1.2
0.7
1.0
1.7
2.0
Other Sources
8.9
4.9
5.8
5.5
5.2
5.1
6.2
Crude Oil Transportation
0.2
0.1
0.2
0.2
0.2
0.2
0.2
Refining
0.7
0.8
0.8
0.8
0.8
0.8
0.9
Total
48.9
39.5
41.5
39.2
39.3
37.3
39.1
Note: Totals may not sum due to independent rounding.
Table 3-39: ChU Emissions from Petroleum Systems (kt ChU)
Activity
1990
2005
2015
2016
2017
2018
2019
Exploration
119
182
83
19
14
15
11
Production
1,802
1,361
1,535
1,508
1,519
1,438
1,507
Pneumatic Controllers
792
673
750
785
799
694
699
Offshore Production
374
261
221
206
205
199
201
Gas Engines
88
74
99
95
94
96
98
Equipment Leaks
82
81
101
97
96
95
94
Produced Water
91
62
82
77
79
83
85
Assoc Gas Flaring
20
15
49
29
38
68
82
Other Sources
355
196
233
219
207
203
249
Crude Oil Transportation
7
5
8
8
8
8
9
Refining
27
31
33
33
33
31
36
Total	1,955	1,579	1,659 1,568 1,574 1,492 1,563
Note: Totals may not sum due to independent rounding.
Table 3-40: CO2 Emissions from Petroleum Systems (MMT CO2)
Activity
1990
2005
2015
2016
2017
2018
2019
Exploration
0.3
0.3
2.2
1.2
1.7
2.9
2.1
Production
6.1
8.0
26.2
16.6
19.6
30.5
40.2
Transportation
+
+
+
+
+
+
+
Crude Refining
3.3
3.7
4.1
4.0
3.7
3.7
5.0
Total
9.7
12.1
32.4
21.8
25.0
37.1
47.3
Note: Totals may not sum due to independent rounding.
Table 3-41: CO2 Emissions from Petroleum Systems (kt CO2)
Activity
1990
2005
2015
2016
2017
2018
2019
Exploration
313
340
2,182
1,212
1,700
2,906
2,081
Production
6,111
7,991
26,163
16,643
19,564
30,473
40,168
Transportation
0.9
0.7
1.2
1.1
1.1
1.2
1.3
Crude Refining
3,284
3,728
4,067
3,991
3,714
3,735
5,019
Total
9,709
12,059
32,412
21,847
24,979
37,115
47,269
Note: Totals may not sum due to independent rounding.
Energy 3-73

-------
Table 3-42: N2O Emissions from Petroleum Systems (Metric Tons CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Exploration
162
173
1,118
617
744
1,370
860
Production
7,502
8,050
20,300
15,087
15,812
29,636
31,269
Transportation
NE
NE
NE
NE
NE
NE
NE
Crude Refining
9,138
10,372
11,656
11,575
10,796
10,557
14,749
Total
16,802
18,596
33,074
27,279
27,352
41,562
46,878
Note: Totals may not sum due to independent rounding.
NE (Not Estimated)
Table 3-43: N2O Emissions from Petroleum Systems (Metric Tons N2O)
Activity
1990
2005
2015
2016
2017
2018
2019
Exploration
0.5
0.6
3.8
2.1
2.5
4.6
2.9
Production
25.2
27.0
68.1
50.6
53.1
99.4
104.9
Transportation
NE
NE
NE
NE
NE
NE
NE
Crude Refining
30.7
34.8
39.1
38.8
36.2
35.4
49.5
Total
56.4
62.4
111.0
91.5
91.8
139.5
157.3
Note: Totals may not sum due to independent rounding.
NE (Not Estimated)
Methodology
See Annex 3.5 for the full time series of emissions data, activity data, emission factors, and additional information
on methods and data sources.
Petroleum systems includes emission estimates for activities occurring in petroleum systems from the oil wellhead
through crude oil refining, including activities for crude oil exploration, production field operations, crude oil
transportation activities, and refining operations. Generally, emissions are estimated for each activity by
multiplying emission factors (e.g., emission rate per equipment or per activity) by corresponding activity data (e.g.,
equipment count or frequency of activity). Certain sources within petroleum refineries are developed with a Tier 3
approach (i.e., all refineries in the nation report emissions data for these sources to the GHGRP, and they are
included in the estimates here). Other estimates are developed with a Tier 2 approach. Tier 1 approaches are not
used.
EPA received stakeholder feedback on updates in the Inventory through EPA's stakeholder process on oil and gas
in the Inventory. Stakeholder feedback is noted below in Recalculations Discussion and Planned Improvements.
More information on the stakeholder process can be found here: https://www.epa.gov/ghgemissions/natural-gas-
and-petroleum-systems.
Emission Factors. Key references for emission factors include Methane Emissions from the Natural Gas Industry by
the Gas Research Institute and EPA (GRI/EPA 1996), Estimates of Methane Emissions from the U.S. Oil Industry (EPA
1999), Compilation of Air Pollutant Emission Factors, AP-42 (EPA 1997), Global Emissions of Methane from
Petroleum Sources (API 1992), consensus of industry peer review panels, Bureau of Ocean Energy Management
(BOEM) reports, Nonpoint Oil and Gas Emission Estimation Tool (EPA 2017), and analysis of GHGRP data (EPA
2020).
Emission factors for hydraulically fractured (HF) oil well completions and workovers (in four control categories)
were developed using EPA's GHGRP data; year-specific data were used to calculate emission factors from 2016-
forward and the year 2016 emission factors were applied to all prior years in the time series. The emission factors
for all years for pneumatic controllers and chemical injection pumps were developed using GHGRP data for
reporting year 2014. The emission factors for tanks, well testing, and associated gas venting and flaring were
developed using year-specific GHGRP data for years 2015 forward; earlier years in the time series use 2015
3-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
emission factors. For miscellaneous production flaring, year-specific emission factors were developed for years
2015 forward from GHGRP data, an emission factor of 0 (assumption of no flaring) was assumed for 1990 through
1992, and linear interpolation was applied to develop emission factors for 1993 through 2014. For more
information, please see memoranda available online.76 For offshore oil production, emission factors were
calculated using BOEM data for offshore facilities in federal waters of the Gulf of Mexico (and these data were also
applied to facilities located in state waters of the Gulf of Mexico) and GHGRP data for offshore facilities off the
coasts of California and Alaska. For many other sources, emission factors were held constant for the period 1990
through 2019, and trends in emissions reflect changes in activity levels. Emission factors from EPA 1999 are used
for all other production and transportation activities.
For associated gas venting and flaring and miscellaneous production flaring, emission factors were developed on a
production basis (i.e., emissions per unit oil produced). Additionally, for these two sources, basin-specific activity
and emission factors were developed for each basin that in any year from 2011 forward contributed at least 10
percent of total source emissions (on a C02 Eq. basis) in the GHGRP. For associated gas venting and flaring, basin-
specific factors were developed for four basins: Williston, Permian, Gulf Coast, and Anadarko. For miscellaneous
production flaring, basin-specific factors were developed for three basins: Williston, Permian, and Gulf Coast. For
each source, data from all other basins were combined, and activity and emission factors were developed for the
other basins as a single group.
For the exploration and production segments, in general, C02 emissions for each source were estimated with
GHGRP data or by multiplying C02 content factors by the corresponding CH4 data, as the C02 content of gas relates
to its CH4 content. Sources with C02 emission estimates calculated using GHGRP data include HF completions and
workovers, associated gas venting and flaring, tanks, well testing, pneumatic controllers, chemical injection pumps,
miscellaneous production flaring, and certain offshore production facilities (those located off the coasts of
California and Alaska). For these sources, C02 was calculated using the same methods as used for CH4. Carbon
dioxide emission factors for offshore oil production in the Gulf of Mexico were derived using data from BOEM,
following the same methods as used for CH4 estimates. For other sources, the production field operations emission
factors for C02 are generally estimated by multiplying the CH4 emission factors by a conversion factor, which is the
ratio of C02 content and CH4 content in produced associated gas.
For the exploration and production segments, N20 emissions were estimated for flaring sources using GHGRP or
BOEM OGOR-B data and the same method used for C02. Sources with N20 emissions in the exploration segment
include well testing and HF completions with flaring. Sources with N20 emissions in the production segment
include associated gas flaring, tank flaring, miscellaneous production flaring, HF workovers with flaring, and flaring
from offshore production sources.
For crude oil transportation, emission factors for CH4 were largely developed using data from EPA (1997), API
(1992), and EPA (1999). Emission factors for C02 were estimated by multiplying the CH4 emission factors by a
conversion factor, which is the ratio of C02 content and CH4 content in whole crude post-separator.
For petroleum refining activities, year-specific emissions from 2010 forward were directly obtained from EPA's
GHGRP. All U.S. refineries have been required to report CH4, C02, and N20 emissions for all major activities starting
with emissions that occurred in 2010. The reported total CH4, C02, and N20 emissions for each activity was used
for the emissions in each year from 2010 forward. To estimate emissions for 1990 to 2009, the 2010 to 2013
emissions data from GHGRP along with the refinery feed data for 2010 to 2013 were used to derive CH4 and C02
emission factors (i.e., sum of activity emissions/sum of refinery feed) and 2010 to 2017 data were used to derive
N20 emission factors; these emission factors were then applied to the annual refinery feed in years 1990 to 2009.
GHGRP delayed coker CH4 emissions for 2010 through 2017 were increased using the ratio of certain reported
emissions for 2018 to 2017, to account for a more accurate GHGRP calculation methodology that was
implemented starting in reporting year 2018.
76 See .
Energy 3-75

-------
A complete list of references for emission factors and activity data by emission source is provided in Annex 3.5.
Activity Data. References for activity data include Enverus data (Enverus 2021), Energy Information Administration
(EIA) reports, Methane Emissions from the Natural Gas Industry by the Gas Research Institute and EPA (EPA/GRI
1996), Estimates of Methane Emissions from the U.S. Oil Industry (EPA 1999), consensus of industry peer review
panels, BOEM reports, the Oil & Gas Journal, the Interstate Oil and Gas Compact Commission, the United States
Army Corps of Engineers, and analysis of GHGRP data (EPA 2020).
For many sources, complete activity data were not available for all years of the time series. In such cases, one of
three approaches was employed to estimate values, consistent with IPCC good practice. Where appropriate, the
activity data were calculated from related statistics using ratios developed based on EPA/GRI 1996 and/or GHGRP
data. In some cases, activity data are developed by interpolating between recent data points (such as from GHGRP)
and earlier data points, such as from EPA/GRI 1996. Lastly, in limited instances the previous year's data were used
if current year data were not yet available.
A complete list of references for emission factors and activity data by emission source is provided in Annex 3.5. The
U.S. reports data to the UNFCCC using this Inventory report along with Common Reporting Format (CRF) tables.
This note is provided for those reviewing the CRF tables: The notation key "IE" is used for C02 and CH4 emissions
from venting and flaring in CRF table l.B.2. Disaggregating flaring and venting estimates across the Inventory
would involve the application of assumptions and could result in inconsistent reporting and, potentially, decreased
transparency. Data availability varies across segments within oil and gas activities systems, and emission factor
data available for activities that include flaring can include emissions from multiple sources (flaring, venting and
leaks).
Uncertainty and Time-Series Consistency
EPA conducted a quantitative uncertainty analysis using the IPCC Approach 2 methodology (Monte Carlo
Simulation technique) to characterize uncertainty for petroleum systems. For more information on the approach,
please see the memorandum Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2016: Natural Gas and
Petroleum Systems Uncertainty Estimates (2018 Uncertainty Memo).77
EPA used Microsoft Excels @RISK add-in tool to estimate the 95 percent confidence bound around CH4 and C02
emissions from petroleum systems for the current Inventory. Uncertainty estimates for N20 were not developed
given the minor contribution of N20 to emission totals. For the CH4 uncertainty analysis, EPA focused on the eight
highest methane-emitting sources for the year 2019, which together emitted 76 percent of methane from
petroleum systems in 2019, and extrapolated the estimated uncertainty for the remaining sources. Uncertainty
was not previously estimated specifically for C02 emissions, instead the uncertainty bounds calculated for CH4
were applied to C02 emissions estimates. As part of the stakeholder process for the current Inventory, EPA
developed an update to the uncertainty analysis for C02. The update is documented in the memorandum,
Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Natural Gas and Petroleum Systems
C02 Uncertainty Estimates.78 EPA ultimately applied the same approach as was developed for CH4. For the C02
uncertainty analysis, EPA focused on the 3 highest-emitting sources for the year 2018 (from the previous 1990-
2018 Inventory), which together emitted 80 percent of C02 from petroleum systems in 2018, and extrapolated the
estimated uncertainty for the remaining sources. The C02 uncertainty calculations were developed as part of the
stakeholder process and were based on the previous 1990-2018 Inventory; as a result, the uncertainty results from
last year's Inventory for year 2018 are applied for this year's uncertainty analysis. In future years, the C02
uncertainty bounds will be calculated using the most recent Inventory data. The @RISK add-in provides for the
specification of probability density functions (PDFs) for key variables within a computational structure that mirrors
the calculation of the inventory estimate. The IPCC guidance notes that in using this method, "some uncertainties
77	See .
78	Stakeholder materials, including draft and final memoranda for the current (i.e. 1990 to 2019) Inventory are available at
.
3-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
that are not addressed by statistical means may exist, including those arising from omissions or double counting,
or other conceptual errors, or from incomplete understanding of the processes that may lead to inaccuracies in
estimates developed from models." As a result, the understanding of the uncertainty of emission estimates for this
category evolves and improves as the underlying methodologies and datasets improve. The uncertainty bounds
reported below only reflect those uncertainties that EPA has been able to quantify and do not incorporate
considerations such as modeling uncertainty, data representativeness, measurement errors, misreporting or
misclassifi cation.
The results presented below provide the 95 percent confidence bound within which actual emissions from this
source category are likely to fall for the year 2019, using the recommended IPCC methodology. The results of the
Approach 2 uncertainty analysis are summarized in Table 3-44. Petroleum systems CH4 emissions in 2019 were
estimated to be between 29.7 and 50.3 MMT C02 Eq., while C02 emissions were estimated to be between 34.5
and 66.7 MMT C02 Eq. at a 95 percent confidence level. Uncertainty bounds for other years of the time series have
not been calculated, but uncertainty is expected to vary over the time series. For example, years where many
emission sources are calculated with interpolated data would likely have higher uncertainty than years with
predominantly year-specific data. In addition, the emission sources that contribute the most to CH4 and C02
emissions are different over the time series, particularly when comparing recent years to early years in the time
series. For example, associated gas venting emissions were higher and flaring emissions were lower in early years
of the time series, compared to recent years. Technologies also changed over the time series (e.g., reduced
emissions completions were not used early in the time series).
Table 3-44: Approach 2 Quantitative Uncertainty Estimates for ChU and CO2 Emissions from
Petroleum Systems (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTCOz Eq.)b
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Lower
Upper



Bound Bound
Bound
Bound
Petroleum Systems
ch4
39.1
29.7 50.3
-24%
+29%
Petroleum Systems
C02
47.3
34.5 66.7
-27%
+41%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for the year 2019 CH4 and year 2018 C02 emissions.
b All reported values are rounded after calculation. As a result, lower and upper bounds may not be duplicable from other
rounded values as shown in table.
EPA's GHGRP data, available starting in 2010 for refineries and in 2011 for other sources, have improved estimates
of emissions from petroleum systems. Many of the previously available datasets were collected in the 1990s. To
develop a consistent time series for sources with new data, EPA reviewed available information on factors that
may have resulted in changes over the time series (e.g., regulations, voluntary actions) and requested stakeholder
feedback on trends as well. For most sources, EPA developed annual data for 1993 through 2009 or 2014 by
interpolating activity data or emission factors or both between 1992 (when GRI/EPA data are available) and 2010
or 2015 data points. Information on time-series consistency for sources updated in this year's Inventory can be
found in the Recalculations Discussion below, with additional detail provided in supporting memos (relevant
memos are cited in the Recalculations Discussion). For information on other sources, please see the Methodology
Discussion above and Annex 3.5. Methodological recalculations were applied to the entire time series to ensure
time-series consistency from 1990 through 2019. Details on the emission trends through time are described in
more detail in the Methodology section, above.
QA/QC and Verification Discussion
The petroleum systems emission estimates in the Inventory are continually being reviewed and assessed to
determine whether emission factors and activity factors accurately reflect current industry practices. A QA/QC
Energy 3-77

-------
analysis was performed for data gathering and input, documentation, and calculation. QA/QC checks are
consistently conducted to minimize human error in the model calculations. EPA performs a thorough review of
information associated with new studies, GHGRP data, regulations, public webcasts, and the Natural Gas STAR
Program to assess whether the assumptions in the Inventory are consistent with current industry practices. EPA
has a multi-step data verification process for GHGRP data, including automatic checks during data-entry, statistical
analyses on completed reports, and staff review of the reported data. Based on the results of the verification
process, EPA follows up with facilities to resolve mistakes that may have occurred.79
As in previous years, EPA conducted early engagement and communication with stakeholders on updates prior to
public review. EPA held stakeholder webinars on greenhouse gas data for oil and gas in September and November
of 2020. EPA released memos detailing updates under consideration and requesting stakeholder feedback.
Stakeholder feedback received through these processes is discussed in the Recalculations Discussion and Planned
Improvements sections below.
In recent years, several studies have measured emissions at the source level and at the national or regional level
and calculated emission estimates that may differ from the Inventory. There are a variety of potential uses of data
from new studies, including replacing a previous estimate or factor, verifying or QA of an existing estimate or
factor, and identifying areas for updates. In general, there are two major types of studies related to oil and gas
greenhouse gas data: studies that focus on measurement or quantification of emissions from specific activities,
processes, and equipment, and studies that use tools such as inverse modeling to estimate the level of overall
emissions needed to account for measured atmospheric concentrations of greenhouse gases at various scales. The
first type of study can lead to direct improvements to or verification of Inventory estimates. In the past few years,
EPA has reviewed, and in many cases, incorporated data from these data sources. The second type of study can
provide general indications on potential over- and under-estimates.
One comment on the public review draft suggested that the inventory estimates be compared with an
observational analysis from a 2019 Lan et al. study.80 Lan et al. estimated an average increasing trend of U.S. oil
and gas methane emissions of 3.4 percent +/-1A percent per year between 2006 and 2015, based on three U.S.
measurement sites that were "substantially influenced by O&NG activities." This study did not address the
magnitude of emissions. Nationally, in the Inventory, methane emissions from oil and gas decreased by an average
of 1 percent per year from 2006 to 2015, largely driven by the natural gas distribution and transmission and
storage segments. A key challenge in using these types of studies to assess Inventory results is having a relevant
basis for comparison (e.g., the two data sets should have comparable time frames and geographic coverage, and
the independent study should assess data from the Inventory and not another data set, such as the Emissions
Database for Global Atmospheric Research, or "EDGAR"). In an effort to improve the ability to compare the
national-level Inventory with measurement results that may be at other scales, a team at Harvard University along
with EPA and other coauthors developed a gridded inventory of U.S. anthropogenic methane emissions with 0.1
degree x 0.1 degree spatial resolution, monthly temporal resolution, and detailed scale-dependent error
characterization.81 The gridded methane inventory is designed to be consistent with the U.S. EPA's Inventory of
U.S. Greenhouse Gas Emissions and Sinks: 1990-2014 estimates for the year 2012, which presents national totals.82
An updated version of the gridded inventory is being developed and will improve efforts to compare results of the
inventory with atmospheric studies.
As discussed above, refinery emissions are quantified by using the total emissions reported to GHGRP for the
refinery emission categories included in Petroleum Systems. Subpart Y has provisions that refineries are not
required to report under Subpart Y if their emissions fall below certain thresholds. Each year, a review is conducted
to determine whether an adjustment is needed to the Inventory emissions to include emissions from refineries
that stopped reporting to the GHGRP. The 2019 GHGRP data indicates that 3 refineries stopped reporting in 2019
79	See .
80	See .
81	See .
82	See .
3-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
(i.e., 2018 is the last reported year). Two of the refineries ceased refinery operations permanently and the other
was a refinery that discontinued reporting in 2019 without a valid reason. EPA did not adjust the 2019 refinery
emissions in the Inventory but will further consider if adjustments are warranted in the future for the refinery that
discontinued reporting without a valid reason.
Recalculations Discussion
EPA received information and data related to the emission estimates through GHGRP reporting, the annual
Inventory formal public notice periods, stakeholder feedback on updates under consideration, and new studies. In
September 2020, EPA released a draft memorandum that discussed changes under consideration to estimate
emissions from produced water and requested stakeholder feedback on those changes. EPA then created an
updated version of the memorandum to document the methodology implemented in the current Inventory.83 The
EPA memorandum Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Produced Water
Emissions (Produced Water memo) is cited below.
EPA thoroughly evaluated relevant information available and made an update to include an estimate for produced
water emissions, discussed in detail below. In addition, certain sources did not undergo methodological updates,
but CH4 and/or C02 emissions changed by greater than 0.05 MMT C02 Eq., comparing the previous estimate for
2018 to the current (recalculated) estimate for 2018. For the sources without methodological updates, the
emissions changes were mostly due to GHGRP data submission revisions and well count updates Enverus updates.
These sources are also discussed below and include hydraulically fractured oil well completions, associated gas
flaring, miscellaneous production flaring, production storage tanks, pneumatic controllers, chemical injection
pumps, oil wellheads (leaks), and gas engines.
The combined impact of revisions to 2018 petroleum systems CH4 emission estimates, compared to the previous
Inventory, is an increase from 36.2 to 39.4 MMT C02 Eq. (3.2 MMT C02 Eq., or 9 percent). The recalculations
resulted in an average increase in CH4 emission estimates across the 1990 through 2018 time series, compared to
the previous Inventory, of 1.1 MMT C02 Eq., or 3 percent, with the largest increase being in the estimate for 1990
(2.8 MMT C02 Eq. or 6 percent) primarily due to inclusion of produced water estimates.
The combined impact of revisions to 2018 petroleum systems C02 emission estimates, compared to the previous
Inventory, is an increase from 36.8 to 37.1 MMT C02 (0.3 MMT C02, or less than 1 percent). The recalculations
resulted in an average decrease in emission estimates across the 1990 through 2018 time series, compared to the
previous Inventory, of 0.1 MMT C02 Eq., or 0.4 percent with the largest changes being for 2016 (1.1 MMT C02 or 5
percent) primarily due to the recalculations for flaring from tanks.
The combined impact of revisions to 2018 petroleum systems N20 emission estimates, compared to the previous
Inventory, is a decrease of 0.03 MMT C02, Eq. or 41 percent. The emission changes were primarily driven by
reduction in flaring emissions from storage tanks and miscellaneous production flaring due to GHGRP data
submission revisions. The recalculations resulted in an average decrease in emission estimates across the 1990
through 2018 time series, compared to the previous Inventory, of 0.001 MMT C02 Eq., or 2 percent.
In Table 3-45 and Table 3-46 below are categories in Petroleum Systems with updated methodologies or with
recalculations resulting in a change of greater than 0.05 MMT C02 Eq., comparing the previous estimate for 2018
to the current (recalculated) estimate for 2018. For more information, please see the discussion below.
83 Stakeholder materials including draft and final memoranda for the current (i.e., 1990 to 2019) Inventory are available at
.
Energy 3-79

-------
Table 3-45: Recalculations of CO2 in Petroleum Systems (MMT CO2)
Segment/Source
Previous Estimate
Year 2018,
2020 Inventory
Current Estimate
Year 2018,
2021 Inventory
Current Estimate
Year 2019,
2021 Inventory
Exploration
2.8
2.9
2.1
HF Oil Well Completions
2.7
2.9
2.1
Production
30.3
30.5
40.2
Produced Water
NE
0.0
0.0
Tanks
6.4
6.3
6.1
Associated Gas Flaring
19.0
19.3
25.4
Miscellaneous Flaring
4.2
4.2
7.9
Transportation
+
+
+
Refining
3.7
3.7
5.0
Petroleum Systems Total
36.8
37.1
47.3
+ Does not exceed 0.05 MMT C02.
NE (Not Estimated)
Table 3-46: Recalculations of ChU in Petroleum Systems (MMT CO2 Eq.)
Segment/Source
Previous Estimate
Year 2018,
2020 Inventory
Current Estimate
Year 2018,
2021 Inventory
Current Estimate
Year 2019,
2021 Inventory
Exploration
0.4
0.4
0.3
HF Oil Well Completions
0.3
0.4
0.2
Production
34.9
35.9
37.7
Produced Water
NE
2.1
2.1
Pneumatic Controllers
18.4
17.3
17.5
Associated Gas Flaring
1.3
1.7
2.0
Miscellaneous Flaring
0.4
0.3
0.6
Chemical InjectionPumps
2.0
1.9
1.9
Oil Wellheads (Leaks)
1.5
1.4
1.4
Gas Engines
2.3
2.4
2.4
Transportation
0.2
0.2
0.2
Refining
0.8
0.8
0.9
Petroleum Systems Total
36.2
39.4
41.2
NE (Not Estimated)
Exploration
HF Oil Well Completions (Recalculation with Updated Data)
HF oil well completion C02 emissions decreased by an average of 1 percent across the time series and increased by
5 percent in 2018, compared the to the previous Inventory. The emissions changes were due to GHGRP data
submission revisions.
Table 3-47: HF Oil Well Completions National CO2 Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
HF Completions: Non-REC with Venting
2.5
4.0
1.4
0.2
0.2
+
+
HF Completions: Non-REC with Flaring
88
140
439
248
394
574
795
HF Completions: REC with Venting
0.0
0.0
0.2
0.1
0.1
0.1
0.1
HF Completions: REC with Flaring
0.0
0.0
1,494
926
1,270
2,300
1,283
Total Emissions
91
144
1,935
1,174
1,664
2,874
2,078
3-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Previous Estimate	92	143	1,966 1,192 1,529 2,730 NA
+ Does not exceed 0.05 kt C02.
NA (Not Applicable)
Production
Produced Water (Methodological Update)
EPA developed a new calculation methodology to estimate produced water emissions from oil wells. Previous
Inventories did not include emissions from produced water from oil wells. EPA's considerations for this source are
documented in the Produced Water Memo.84 Produced water quantities (i.e., bbl) were obtained for 36 oil-
producing states as described below:
•	Produced water quantities for 1990-2019 were obtained using Drillinglnfo and Prism datasets from
Enverus for 29 states (i.e., AK, AL, AR, AZ, CA, CO, FL, ID, KY, LA, MD, Ml, MN, MO, MS, MT, NC, ND, NE,
NM, NV, NY, OR, SD, TN, TX, UT, VA, and WY) (Enverus 2021). Linear interpolation was used to correct an
obviously inaccurate near-zero produced water quantity value in Colorado for 1998.
•	For four additional states, produced water quantities for 1990-2018 were obtained from state agency
websites - KS (Kansas Department of Health and Environment 2020), OH (Ohio Environmental Protection
Agency 2020), OK (Oklahoma Department of Environmental Quality 2020), and PA (Pennsylvania
Department of Environmental Protection 2020). Produced water quantities for 2018 were used as proxy
data for 2019 for these four states.
•	Produced water quantities for 1990-2018 were estimated for three states (IL, IN, and WV) using state-
level produced water production ratios for oil wells. Well-level produced water data for oil wells for 2011
were obtained from the Drillinglnfo dataset (Enverus 2021) and oil production data were obtained from
state agency websites - IL (Illinois Office of Oil and Gas Resource Management 2020), IN (Indiana Division
of Oil & Gas 2020), and WV (West Virginia Department of Environmental Protection 2020). Using these
well-level produced water data and the oil production data, production ratios were developed for oil wells
in each state. These production ratios were then applied to annual state-level oil production data (2000-
2018) from EIA (EIA 2020). Produced water quantities for 2018 were used as proxy data for 2019 for these
three states.
Based upon findings of the CenSARA emissions inventory (CenSARA 2012), EPA assumed that 73 percent of the
produced water was from low pressure oil wells (i.e., wells requiring artificial lifts) and that 27 percent of the
produced water was from regular pressure oil wells (i.e., wells not requiring artificial lifts).. EPA applied emission
factors unique to low pressure and regular pressure oil wells, obtained from the Production Module of the 2017 Oil
and Gas Tool. Produced water CH4 emissions average 73,000 mt CH4, over the 1990 to 2019 time series.
EPA received feedback on this update through its September 2020 memo and through the public review draft of
the inventory.
A stakeholder indicated that the typical practice is to route produced water to a tank battery, once it reaches the
surface and has been separated from the oil and gas. A stakeholder requested that data from the latest 2017
Ground Water Protection Council produced water management practices survey be used to determine the percent
of produced water that is stored in tanks. The stakeholder indicated that approximately 16 percent of produced
water has the potential of being stored in a tank battery that could potentially flash (based on the 2012 Ground
Water Protection Council produced water management practices survey). After further assessment of the 2012
and 2017 water management practice surveys, EPA has maintained the assumption that all produced water goes
through tanks and emissions are flashed, consistent with the approach used for the public review draft of the
Inventory.
84 See .
Energy 3-81

-------
A stakeholder commented that current regulations under 40 CFR 60 subpart OOOOa require that certain storage
vessels route emission vapors to a recovery device, flare, or other control device. EPA currently does not have
specific data to address the use of controls on produced water tanks but will continue to assess this issue in future
inventories should additional data become available.
Table 3-48: Produced Water National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Low Pressure Oil Wells
20,292
13,794
18,276
17,142
17,446
18,368
18,794
Regular Pressure Oil Wells
71,186
48,390
64,115
60,136
61,273
64,438
65,931
Total
91,478
62,184
82,392
77,278
78,739
82,806
84,726
Previous Estimate
NA
NA
NA
NA
NA
NA
NA
NA (Not Applicable)
Tanks (Recalculation with Updated Data)
Tank C02 emissions estimates increased by an average of 2 percent across the 1990 to 2018 time series and
decreased by 1 percent in 2018, compared to the previous inventory. The emission changes were due to GHGRP
data submission revisions.
Table 3-49: Tanks National CO2 Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
Large Tanks w/Flares
0
2,451
7,074
4,489
4,298
6,219
6,037
Large Tanks w/VRU
0
5
14
3
4
4
7
Large Tanks w/o Control
24
6
6
5
5
4
5
Small Tanks w/Flares
0
2
7
13
11
8
10
Small Tanks w/o Flares
6
3
5
4
4
4
4
Malfunctioning Separator Dump







Valves
86
50
104
32
43
39
34
Total Emissions
116
2,517
7,209
4,546
4,364
6,278
6,098
Previous Estimate
46
2,641
7,584
5,913
4,413
6,369
NA
NA (Not Applicable)
Pneumatic Controllers (Recalculation with Updated Data)
Pneumatic controller CH4 emission estimates decreased by an average of 3 percent across the 1990 to 2018 time
series and decreased by 6 percent in 2018, compared to the previous Inventory. The emission changes were due to
GHGRP data submission revisions and updated Enverus well counts.
Table 3-50: Pneumatic Controller National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
High Bleed
741,469
401,564
73,505
77,866
57,020
45,304
49,775
Low Bleed
50,606
42,080
23,868
16,358
18,283
28,496
34,092
Intermittent Bleed
0
229,126
652,946
690,799
724,193
620,175
615,621
Total Emissions
792,075
672,769
750,320
785,023
799,496
693,976
699,488
Previous Estimate
772,311
704,401
789,484
822,989
850,624
734,824
NA
NA (Not Applicable)
3-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Associated Gas Flaring (Recalculation with Updated Data)
Associated gas flaring C02 emission estimates increased by an average of 1 percent across the time series and
increased by 2 percent in 2018 in the current Inventory, compared to the previous Inventory. The emission
changes were due to GHGRP data submission revisions.
Table 3-51: Associated Gas Flaring National CO2 Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
220 - Gulf Coast Basin (LA, TX)
227
121
672
404
744
643
584
360 - Anadarko Basin
108
66
242
1
64
82
18
395 - Williston Basin
987
1,263
8,567
6,091
6,908
11,140
16,572
430 - Permian Basin
2,983
2,056
4,468
2,261
3,209
6,782
7,161
"Other" Basins
935
505
548
324
387
641
1,021
Total Emissions
5,241
4,011
14,498
9,081
11,313
19,287
25,356
220 - Gulf Coast Basin (LA, TX)
234
127
673
404
740
686
NA
360 - Anadarko Basin
108
65
238
2
57
37
NA
395 - Williston Basin
966
1,239
8,412
5,838
6,530
10,132
NA
430 - Permian Basin
2,983
2,046
4,443
2,246
3,148
7,249
NA
"Other" Basins
925
499
544
326
414
876
NA
Previous Estimate
5,217
3,977
14,311
8,815
10,889
18,980
NA
NA (Not Applicable)
Associated gas flaring CH4 emission estimates increased by an average of 2 percent across the time series in the
current Inventory, compared to the previous inventory. The CH4 estimates increased by 27 percent in 2018,
primarily due to Williston Basin data. The emission changes were due to GHGRP data submission revisions.
Table 3-52 Associated Gas Flaring National Cm Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
220 - Gulf Coast Basin (LA, TX)
896
479
2,654
1,572
2,936
2,448
2,626
360 - Anadarko Basin
472
288
1,056
4
277
358
87
395 - Williston Basin
2,931
3,750
25,437
16,948
20,707
37,754
48,453
430- Permian Basin
11,815
8,143
17,696
8,972
13,189
25,511
27,016
"Other" Basins
4,328
2,335
2,538
1,193
1,290
1,932
3,614
Total Emissions
20,441
14,995
49,380
28,689
38,399
68,004
81,797
220 - Gulf Coast Basin (LA, TX)
922
500
2,659
1,572
2,918
2,779
NA
360 - Anadarko Basin
471
285
1,038
7
252
190
NA
395 - Williston Basin
2,874
3,686
25,020
16,151
20,130
26,011
NA
430 - Permian Basin
11,816
8,104
17,596
8,913
12,974
22,597
NA
"Other" Basins
4,274
2,306
2,514
1,196
1,388
1,862
NA
Previous Estimate
20,357
14,881
48,826
27,839
37,662
53,438
NA
NA (Not Applicable)
Miscellaneous Production Flaring (Recalculation with Updated Data)
Miscellaneous production flaring CH4 emission estimates decreased by 27 percent in 2018 and decreased by an
average of 4 percent for other years of the time series, compared to the previous Inventory. The 2018 decrease
was primarily due to recalculations in the Permian and Gulf Coast basins, where GHGRP data showed lower CH4
flaring emissions, by 47 and 18 percent, respectively. The emission changes were due to GHGRP data submission
revisions.
Energy 3-83

-------
Table 3-53: Miscellaneous Production Flaring National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
220 - Gulf Coast Basin (LA, TX)
0
410
4,021
1,979
2,179
1,951
2,554
395 - Williston Basin
0
182
2,184
854
1,618
3,045
3,904
430 - Permian Basin
0
807
3,103
2,812
5,055
4,449
14,151
Other Basins
0
1,316
2,806
1,455
1,960
1,891
1,847
Total Emissions
0
2,715
12,114
7,101
10,812
11,336
22,457
220 - Gulf Coast Basin (LA, TX)
0
424
3,985
1,979
2,164
2,370
NA
395 - Williston Basin
0
191
2,293
888
1,603
2,947
NA
430 - Permian Basin
0
805
3,091
2,794
5,024
8,406
NA
Other Basins
0
1,440
3,074
1,452
2,018
1,812
NA
Previous Total Estimate
0
2,859
12,443
7,113
10,810
15,536
NA
NA (Not Applicable)
Miscellaneous production flaring C02 emission estimates decreased by 1 percent in 2018 and decreased by less
than 0.5 percent for other years of the time series, compared to the previous Inventory. The 2018 decrease was
primarily due to recalculations of C02 from flaring in the Gulf Coast Basin, where GHGRP data showed lower C02
emissions from flaring, by 16 percent. The emission changes were due to GHGRP data submission revisions.
Table 3-54: Miscellaneous Production Flaring National CO2 Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
220 - Gulf Coast Basin (LA, TX)
0
102
1,004
497
526
577
625
395 - Williston Basin
0
73
873
304
537
1,706
2,934
430 - Permian Basin
0
215
828
799
1,433
1,244
3,701
Other Basins
0
408
870
593
568
640
689
Total Emissions
0
799
3,575
2,192
3,063
4,167
7,949
220 - Gulf Coast Basin (LA, TX)
0
106
997
497
526
687
NA
395 - Williston Basin
0
73
882
315
531
1,653
NA
430 - Permian Basin
0
215
825
794
1,424
1,183
NA
Other Basins
0
407
870
592
585
703
NA
Previous Total Estimate
0
801
3,573
2,198
3,066
4,226
NA
NA (Not Applicable)
Chemical Injection Pumps (Recalculation with Updated Data)
Chemical injection pump CH4 emission estimates decreased by an average of 4 percent across the time series and
decreased by 6 percent in 2018, compared to the previous Inventory. The emission changes were due to updated
Enverus well counts.
Table 3-55: Chemical Injection Pump National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Chemical Injection Pump
Previous Estimate
50,806
49,368
64,259
68,097
81,103
86,529
78,351
83,705
77,061
82,180
76,014
81,294
75,182
NA
NA (Not Applicable)
Oil Wellheads (Recalculation with Updated Data)
3-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Oil wellhead CH4 emission estimates decreased by an average of 8 percent across the time series and decreased by
6 percent in 2018, compared to the previous Inventory. The emission changes were due to updated Enverus well
counts.
Table 3-56: Oil Wellhead National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Oil Wellheads (heavy crude)
32
28
35
34
34
33
33
Oil Wellheads (light crude)
55,064
48,648
61,163
59,088
58,115
57,326
56,699
Total Emissions
55,096
48,676
61,199
59,122
58,149
57,359
56,732
Previous Estimate
61,144
52,504
65,294
63,162
62,011
61,343
NA
NA (Not Applicable)
Gas Engines (Recalculation with Updated Data)
Gas engine (combustion slip) CH4 emission estimates increased by an average of 4 percent across the time series
and increased by 5 percent in 2018, compared to the previous Inventory. The emission changes were due to
updated Enverus well counts. Even though the well counts have decreased across the time-series, the 2018 gas
engine estimates are calculated using the ratio of 2018 to 1993 well counts. Since the 1993 well counts show a
larger decrease (-12%) compared to the 2018 well counts (-6%), the gas engine estimates increased.
Table 3-57: Gas Engine National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Total Gas Engine Emissions
Previous Estimate
87,854
85,744
73,659
69,999
98,896
93,414
94,771
89,565
94,311
89,063
96,338
91,459
97,828
NA
NA (Not Applicable)
Well Counts (Recalculation with Updated Data)
EPA uses annual producing oil well counts as an input for estimates of emissions from multiple sources in the
Inventory, including exploration well testing, pneumatic controllers, chemical injection pumps, well workovers, and
equipment leaks. Annual well count data are obtained from Enverus for the entire time series during each
Inventory cycle, and a new data processing methodology was implemented this year due to a restructuring of the
Enverus well count data (Enverus 2021). Due to of the data restructuring and the new processing methodology,
annual well counts decreased by an average of 8 percent across the 1990-2018 time series and decreased by 6
percent in 2018, compared to the previous Inventory.
Table 3-58: National Oil Well Counts
Source
1990
2005
2015
2016
2017
2018
2019
Oil Wells
Previous Estimate
506,730
562,356
447,683
482,887
562,857
600,519
543,759
580,917
534,806
570,331
527,544
564,186
521,771
NA
NA (Not Applicable)
In December 2020, EIA released an updated time series of national oil and gas well counts (covering 2000 through
2018). EIA estimates 969,136 total producing wells for year 2019. EPA's total well count for this year is 939,637.
EPA's well counts are generally lower than ElA's (e.g., around 3 percent lower in 2019). ElA's well counts include
side tracks (i.e., secondary wellbore away from original wellbore in order to bypass unusable formation, explore
nearby formations, or other reasons), completions, and recompletions, and therefore are expected to be higher
than EPA's which include only producing wells. EPA and EIA use a different threshold for distinguishing between oil
versus gas (EIA uses 6 mcf/bbl, while EPA uses 100 mcf/bbl), which results in EIA having a lower fraction of oil wells
Energy 3-85

-------
(e.g., 44 percent versus EPA's 56 percent in 2019) and a higher fraction of gas wells (e.g., 56 percent versus EPA's
44 percent in 2019) than EPA.
Transportation
Recalculations for the transportation segment have resulted in an average increase in calculated CH4 and C02
emissions over the time series from this segment of less than 0.02 percent, compared to the previous Inventory.
Refining
Recalculations due to resubmitted GHGRP data in the refining segment have resulted in an average increase in
calculated CH4 emissions over the time series from this segment of 0.1 percent and an average increase in
calculated C02 emissions over the time series of less than 0.01 percent, compared to the previous Inventory.
Planned Improvements
Mud Degassing
As part of the stakeholder process for the current (1990 to 2019) Inventory, EPA developed draft CH4 emission
estimates for mud degassing in the onshore exploration segment. To date, the Inventory has not included
emissions from onshore exploration mud degassing. EPA's considerations for this source are documented in the
EPA memorandum Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update Under Consideration
for Mud Degassing Emissions (Mud Degassing Memo).85 EPA estimated emissions using CH4 emission factors from
EPA (EPA 1977) and the count of wells drilled from Enverus Drillinglnfo data (Enverus Drillinglnfo 2019). To
calculate emissions per well drilled, EPA incorporated an estimate of 26 days as an average drilling duration and
61.2 percent (by weight) as the default CH4 content of associated gas. EPA developed national estimates for two
different scenarios: 1) EPA assumed 80 percent of drilling operations were performed using water-based muds and
the remaining 20 percent used oil-based muds; and 2) EPA assumed 100 percent of drilling operations were
performed using water-based muds. Mud degassing CH4 emissions averaged 107 kt over the time series for
scenario 1 and 126 kt for scenario 2, or around 3 MMT C02 Eq. This update would increase emissions from the
exploration segment but would have a small impact on overall CH4 emissions from petroleum systems.
EPA notes that estimates for mud degassing using similar assumptions are included in several other bottom-up
inventories for greenhouse gases and other gases, including New York state and the NEI.
EPA received feedback on this update through its September 2020 memo and through the public review draft of
the Inventory. A stakeholder indicated the 12 inch diameter borehole and 25 percent formation porosity
assumptions used in developing the CH4 emission factor for water-based muds are outdated and recommended
that an 8 inch borehole diameter and 10 percent porosity should be considered in developing the CH4 EF. A
stakeholder commented that current onshore practices are to drill with balanced or slightly over-balanced mud
systems that keep gas from being entrained in the drilling mud and that mud degassing systems are rarely needed
or used. A stakeholder also indicated that mainly oil-based muds are used for horizontal/lateral drilling and water-
based muds are more frequently used for vertical drilling.
Additionally, EPA also received comments on the average drilling duration used in developing the draft estimates
for onshore mud degassing. A stakeholder commentstated that EPA should only consider the duration the drill
spends in the producing formation. Another comment indicated that EPA's average drilling duration assumption of
26 days per well is high and presented 2 examples - a Marcellus well takes 10 days to drill with 2-3 days in the
producing formation; and the drilling duration in the Fayetteville shale dropped from 20 days in 2007 to 11 days in
2009. EPA's average drilling duration assumption of 26 days per well is comparable to average drilling duration
85 Stakeholder materials including draft memoranda for the current (i.e., 1990 to 2019) Inventory are available at
.
3-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
developed from other inventories (New York - 24 days/well and CenSARA - 22 days/well). Refer to the Mud
Degassing memo for further details.
EPA continues to seek feedback on average total drilling days and drilling days in the producing formation, CH4
content of the gas, and the effect of balanced and over-balanced mud degassing systems. EPA will further assess
the average drilling duration using updated Enverus data. Additionally, EPA is considering developing C02
estimates for onshore production mud degassing using the CH4 estimates and a ratio of C02-to-CH4.
Table 3-59: Draft Mud Degassing National ChU Emissions—Not Included in Totals (Metric
Tons CH4)
Source
1990

2005

2015
2016
2017
2018
2019
Scenario 1 (80/20)
105,862

98,024

93,555
56,256
101,179
101,179
101,179
Scenario 2 (100)
124,543

115,322

110,065
66,183
119,035
119,035
119,035
Previous Estimate
NA

NA

NA
NA
NA
NA
NA
NA (Not Applicable)
Scenario 1 (80/20) = 80% water-based mud usage and 20% oil-based mud usage
Scenario 2 (100) = 100% water-based mud usage
Upcoming Data, and Additional Data that Could Inform the Inventory
EPA will assess new data received by the Methane Challenge Program on an ongoing basis, which may be used to
confirm or improve existing estimates and assumptions.
EPA continues to track studies that contain data that may be used to update the Inventory. EPA will also continue
to assess studies that include and compare both top-down and bottom-up estimates, and which could lead to
improved understanding of unassigned high emitters (e.g., identification of emission sources and information on
frequency of high emitters) as recommended in stakeholder comments.
EPA also continues to seek new data that could be used to assess or update the estimates in the Inventory. For
example, in recent years, stakeholder comments have highlighted areas where additional data that could inform
the Inventory are currently limited or unavailable:
•	Tank and flaring malfunction and control efficiency data.
•	Improved equipment leak data
•	Activity data and emissions data for production facilities that do not report to GHGRP.
•	Associated gas venting and flaring data on practices from 1990 through 2010.
•	Refineries emissions data.
•	Anomalous leak events.
EPA will continue to seek available data on these and other sources as part of the process to update the Inventory.
Box 3-6: Carbon Dioxide Transport, Injection, and Geological Storage
Carbon dioxide is produced, captured, transported, and used for Enhanced Oil Recovery (EOR) as well as
commercial and non-EOR industrial applications, or is stored geologically. This C02 is produced from both
naturally-occurring C02 reservoirs and from industrial sources such as natural gas processing plants and
ammonia plants. In the Inventory, emissions of C02from naturally-occuring C02reservoirs are estimated based
on the specific application.
In the Inventory, C02 that is used in non-EOR industrial and commercial applications (e.g., food processing,
chemical production) is assumed to be emitted to the atmosphere during its industrial use. These emissions are
discussed in the Carbon Dioxide Consumption section, 4.15.
For EOR C02, as noted in the 2006IPCC Guidelines, "At the Tier 1 or 2 methodology levels [EOR C02 is]
indistinguishable from fugitive greenhouse gas emissions by the associated oil and gas activities." In the U.S.
estimates for oil and gas fugitive emissions, the Tier 2 emission factors for C02 include C02 that was originally
Energy 3-87

-------
injected and is emitted along with other gas from leak, venting, and flaring pathways, as measurement data
used to develop those factors would not be able to distinguish between C02 from EOR and C02 occurring in the
produced natural gas. Therefore, EOR C02 emitted through those pathways is included in C02 estimates in 1B2.
IPCC includes methodological guidance to estimate emissions from the capture, transport, injection, and
geological storage of C02. The methodology is based on the principle that the carbon capture and storage
system should be handled in a complete and consistent manner across the entire Energy sector. The approach
accounts for C02 captured at natural and industrial sites as well as emissions from capture, transport, and use.
For storage specifically, a Tier 3 methodology is outlined for estimating and reporting emissions based on site-
specific evaluations. However, IPCC (IPCC 2006) notes that if a national regulatory process exists, emissions
information available through that process may support development of C02 emission estimates for geologic
storage.
In the United States, facilities that produce C02 for various end-use applications (including capture facilities such
as acid gas removal plants and ammonia plants), importers of C02, exporters of C02, facilities that conduct
geologic sequestration of C02, and facilities that inject C02 underground, are required to report greenhouse gas
data annually to EPA through its GHGRP. Facilities reporting geologic sequestration of C02 to the GHGRP
develop and implement an EPA-approved site-specific monitoring, reporting and verification plan, and report
the amount of C02 sequestered using a mass balance approach.
GHGRP data relevant for this inventory estimate consists of national-level annual quantities of C02 captured and
extracted for EOR applications for 2010 to 2019 and data reported for geologic sequestration from 2016 to
2019.
The amount of C02 captured and extracted from natural and industrial sites for EOR applications in 2019 is
52,100 kt (52.1 MMT C02 Eq.) (see 6). The quantity of C02 captured and extracted is noted here for information
purposes only; C02 captured and extracted from industrial and commercial processes is generally assumed to be
emitted and included in emissions totals from those processes.
Table 3-60: Quantity of CO2 Captured and Extracted for EOR Operations (kt CO2)
Stage
2015
2016
2017
2018
2019
Total C02 Captured and Extracted Stage
54,000
46,700
49,600
48,400
52,100
Several facilities are reporting under GHGRP subpart RR (Geologic Sequestration of Carbon Dioxide). See Table
3-61 for the number of facilities reporting under subpart RR, the reported C02 sequestered in subsurface
geologic formations in each year, and of the quantity of C02 emitted from equipment leaks in each year. The
quantity of C02 sequestered and emitted is noted here for information purposes only; EPA is considering
updates to its approach in the Inventory for this source for future Inventories.
Table 3-61: Geologic Sequestration Information Reported Under GHGRP Subpart RR
Stage	2015 2016	2017	2018	2019
Number of Reporting Facilities NA	1	3	5	5
Reported Annual C02
Sequestered (kt) NA 3,091	5,958	7,662	8,332
Reported Annual C02 Emissions
from Equipment Leaks (kt) NA	10	10	11	16
3-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
3.7 Natural Gas Systems (CRF Source Cati—ry
lB2b)
The U.S. natural gas system encompasses hundreds of thousands of wells, hundreds of processing facilities, and
over a million miles of transmission and distribution pipelines. This IPCC category (lB2b) is for fugitive emissions,
which per IPCC include emissions from leaks, venting, and flaring. Total greenhouse gas emissions (CH4, C02, and
N20) from natural gas systems in 2019 were 194.9 MMT C02 Eq., a decrease of 11 percent from 1990, primarily
due to decreases in CH4 emissions, and an increase of 5 percent from 2018, primarily due to increases in CH4
emissions. From 2009, emissions increased by 6 percent, primarily due to increases in C02 emissions. National total
dry gas production in the United States increased by 91 percent from 1990 to 2019, by 10 percent from 2018 to
2019, and by 65 percent from 2009 to 2019. Of the overall greenhouse gas emissions (194.9 MMT C02 Eq.), 81
percent are CH4 emissions (157.6 MMT C02 Eq.), 19 percent are C02 emissions (37.2 MMT), and less than 0.01
percent are N20 emissions (0.01 MMT C02 Eq.).
Overall, natural gas systems emitted 157.6 MMT C02 Eq. (6,305 kt CH4) of CH4 in 2019, a 16 percent decrease
compared to 1990 emissions, and 3 percent increase compared to 2018 emissions (see Table 3-63 and Table 3-64).
For non-combustion C02, a total of 37.2 MMT C02 Eq. (37,234 kt) was emitted in 2019, a 16 percent increase
compared to 1990 emissions, and a 10 percent increase compared to 2018 levels. The 2019 N20 emissions were
estimated to be 0.01 MMT C02 Eq. (0.04 kt N20), a 123 percent increase compared to 1990 emissions, and a 1
percent increase compared to 2018 levels.
The 1990 to 2019 trend is not consistent across segments or gases. Overall, the 1990 to 2019 decrease in CH4
emissions is due primarily to the decrease in emissions from the following segments: distribution (69 percent
decrease), transmission and storage (35 percent decrease), processing (42 percent decrease), and exploration (87
percent decrease). Over the same time period, the production segment saw increased CH4 emissions of 59 percent
(with onshore production emissions increasing 44 percent, offshore production emissions decreasing 82 percent,
and gathering and boosting [G&B] emissions increasing 121 percent). The 1990 to 2019 increase in C02 emissions
is primarily due to an increase in C02 emissions in the production segment, where emissions from flaring have
increased over time.
Methane and C02 emissions from natural gas systems include those resulting from normal operations, routine
maintenance, and system upsets. Emissions from normal operations include natural gas engine and turbine
uncombusted exhaust, flaring, and leak emissions from system components. Routine maintenance emissions
originate from pipelines, equipment, and wells during repair and maintenance activities. Pressure surge relief
systems and accidents can lead to system upset emissions. Emissions of N20 from flaring activities are included in
the Inventory, with most of the emissions occurring in the processing and production segments. Note, C02
emissions exclude all combustion emissions (e.g., engine combustion) except for flaring C02 emissions. All
combustion C02 emissions (except for flaring) are accounted for in Section 3.1 - C02 from Fossil Fuel Combustion.
Each year, some estimates in the Inventory are recalculated with improved methods and/or data. These
improvements are implemented consistently across the previous Inventory's time series (i.e., 1990 to 2018) to
ensure that the trend is accurate. Recalculations in natural gas systems in this year's Inventory include:
•	Updated methodology for produced water (to expand included basins)
•	Updated methodology for customer meters to use data from GTI 2009 and GTI 2019
•	Updates to well counts using the most recent data from Enverus
•	Recalculations due to GHGRP submission revisions
The Recalculations Discussion section below provides more details on the updated methods.
Below is a characterization of the five major segments of the natural gas system: exploration, production (including
gathering and boosting), processing, transmission and storage, and distribution. Each of the segments is described
and the different factors affecting CH4, C02, and N20 emissions are discussed.
Energy 3-89

-------
Exploration. Exploration includes well drilling, testing, and completions. Emissions from exploration accounted for
less than 1 percent of CH4 emissions and of C02 emissions from natural gas systems in 2019. Well completions
accounted for approximately 95 percent of CH4 emissions from the exploration segment in 2019, with the rest
resulting from well testing and drilling. Flaring emissions account for most of the C02 emissions. Methane
emissions from exploration decreased by 87 percent from 1990 to 2019, with the largest decreases coming from
hydraulically fractured gas well completions without reduced emissions completions (RECs). Methane emissions
decreased 36 percent from 2018 to 2019 due to decreases in emissions from hydraulically fractured well
completions with RECs and venting. Methane emissions were highest from 2005 to 2008. Carbon dioxide emissions
from exploration decreased by 44 percent from 1990 to 2019 and decreased 34 percent from 2018 to 2019 due to
decreases in flaring. Carbon dioxide emissions were highest from 2006 to 2008. Nitrous oxide emissions decreased
73 percent from 1990 to 2019 and decreased 95 percent from 2018 to 2019.
Production (including gathering and boosting). In the production segment, wells are used to withdraw raw gas
from underground formations. Emissions arise from the wells themselves, and from well-site equipment and
activities such as pneumatic controllers, tanks and separators, and liquids unloading. Gathering and boosting
emission sources are included within the production sector. The gathering and boosting sources include gathering
and boosting stations (with multiple emission sources on site) and gathering pipelines. The gathering and boosting
stations receive natural gas from production sites and transfer it, via gathering pipelines, to transmission pipelines
or processing facilities (custody transfer points are typically used to segregate sources between each segment).
Boosting processes include compression, dehydration, and transport of gas to a processing facility or pipeline.
Emissions from production (including gathering and boosting) accounted for 59 percent of CH4 emissions and 29
percent of C02 emissions from natural gas systems in 2019. Emissions from gathering and boosting and pneumatic
controllers in onshore production accounted for most of the production segment CH4 emissions in 2019. Within
gathering and boosting, the largest sources of CH4 are compressor exhaust slip, compressor venting and leaks, and
tanks. Flaring emissions account for most of the C02 emissions from production, with the highest emissions coming
from flare stacks at gathering stations, miscellaneous onshore production flaring, and tank flaring. Methane
emissions from production increased by 59 percent from 1990 to 2019, due primarily to increases in emissions
from pneumatic controllers (due to an increase in the number of controllers, particularly in the number of
intermittent bleed controllers) and increases in emissions from compressor exhaust slip in gathering and boosting.
Methane emissions increased 3 percent from 2018 to 2019 due to increases in the number of intermittent bleed
controllers and increases in emissions from tanks in gathering and boosting. Carbon dioxide emissions from
production increased by approximately a factor of 3.6 from 1990 to 2019 due to increases in emissions at flare
stacks in gathering and boosting and miscellaneous onshore production flaring, and increased 11 percent from
2018 to 2019 due primarily to increases in emissions from flare stacks and acid gas removal in gathering and
boosting. Nitrous oxide emissions increased 28 percent from 1990 to 2019 and increased 10 percent from 2018 to
2019. The increase in N20 emissions from 1990 to 2019 and from 2018 to 2019 is primarily due to increase in
emissions from flare stacks at gathering and boosting stations.
Processing. In the processing segment, natural gas liquids and various other constituents from the raw gas are
removed, resulting in "pipeline quality" gas, which is injected into the transmission system. Methane emissions
from compressors, including compressor seals, are the primary emission source from this stage. Most of the C02
emissions come from acid gas removal (AGR) units, which are designed to remove C02 from natural gas. Processing
plants accounted for 8 percent of CH4 emissions and 67 percent of C02 emissions from natural gas systems.
Methane emissions from processing decreased by 42 percent from 1990 to 2019 as emissions from compressors
(leaks and venting) and equipment leaks decreased; and increased 3 percent from 2018 to 2019 due to increased
emissions from gas engines. Carbon dioxide emissions from processing decreased by 13 percent from 1990 to
2019, due to a decrease in AGR emissions, and increased 7 percent from 2018 to 2019 due to increased emissions
from flaring. Nitrous oxide emissions increased 39 percent from 2018 to 2019.
Transmission and Storage. Natural gas transmission involves high pressure, large diameter pipelines that transport
gas long distances from field production and processing areas to distribution systems or large volume customers
such as power plants or chemical plants. Compressor station facilities are used to move the gas throughout the
U.S. transmission system. Leak CH4 emissions from these compressor stations and venting from pneumatic
controllers account for most of the emissions from this stage. Uncombusted compressor engine exhaust and
3-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
pipeline venting are also sources of CH4 emissions from transmission. Natural gas is also injected and stored in
underground formations, or liquefied and stored in above ground tanks, during periods of low demand (e.g.,
summer), and withdrawn, processed, and distributed during periods of high demand (e.g., winter). Leak and
venting emissions from compressors are the primary contributors to CH4 emissions from storage. Emissions from
liquefied natural gas (LNG) stations and terminals are also calculated under the transmission and storage segment.
Methane emissions from the transmission and storage segment accounted for approximately 23 percent of
emissions from natural gas systems, while C02 emissions from transmission and storage accounted for 3 percent of
the C02 emissions from natural gas systems. CH4emissions from this source decreased by 35 percent from 1990 to
2019 due to reduced compressor station emissions (including emissions from compressors and leaks) and
increased 6 percent from 2018 to 2019 due to increased emissions from transmission compressors. C02 emissions
from transmission and storage were 6.9 times higher in 2019 than in 1990, due to increased emissions from LNG
export terminals, and increased by 128 percent from 2018 to 2019, also due to LNG export terminals. The quantity
of LNG exported from the U.S. increased by a factor of 35 from 1990 to 2019, and by 68 percent from 2018 to
2019. LNG emissions are about 1 percent of CH4 and 80 percent of C02 emissions from transmission and storage in
year 2019. Nitrous oxide emissions from transmission and storage increased by 145 percent from 1990 to 2019
and increased 169 percent from 2018 to 2019.
Distribution. Distribution pipelines take the high-pressure gas from the transmission system at "city gate" stations,
reduce the pressure and distribute the gas through primarily underground mains and service lines to individual end
users. There were 1,316,689 miles of distribution mains in 2019, an increase of 372,532 miles since 1990 (PHMSA
2020). Distribution system emissions, which accounted for 9 percent of CH4 emissions from natural gas systems
and less than 1 percent of C02 emissions, result mainly from leak emissions from pipelines and stations. An
increased use of plastic piping, which has lower emissions than other pipe materials, has reduced both CH4 and
C02 emissions from this stage, as have station upgrades at metering and regulating (M&R) stations. Distribution
system CH4 emissions in 2019 were 69 percent lower than 1990 levels and 1 percent lower than 2018 emissions.
Distribution system C02 emissions in 2019 were 69 percent lower than 1990 levels and 1 percent lower than 2018
emissions. Annual C02 emissions from this segment are less than 0.1 MMT C02 Eq. across the time series.
Total greenhouse gas emissions from the five major segments of natural gas systems are shown in MMT C02 Eq. in
Table 3-62. Total CH4 emissions for these same segments of natural gas systems are shown in MMT C02 Eq. (Table
3-63) and kt (Table 3-64). Most emission estimates are calculated using a net emission approach. However, a few
sources are still calculated with a potential emission approach. Reductions data are applied to those sources that
use a potential emissions approach. In recent years 6.3 MMT C02 Eq. CH4 are subtracted from production segment
emissions and 6.7 MMT C02 Eq. CH4 are subtracted from the transmission and storage segment to calculate net
emissions. More disaggregated information on potential emissions, net emissions, and reductions data is available
in Annex 3.6, Methodology for Estimating CH4 and C02 Emissions from Natural Gas Systems.
Table 3-62: Total Greenhouse Gas Emissions (CH4, CO2, and N2O) from Natural Gas Systems
(MMT COz Eq.)
Stage
1990
2005
2015
2016
2017
2018
2019
Exploration
4.6
12.0
1.3
0.9
1.7
1.2
0.8
Production
61.8
85.2
96.9
94.1
96.6
100.6
104.7
Processing
49.7
30.4
32.0
33.2
34.5
35.2
37.2
Transmission and Storage
57.4
36.2
34.4
34.8
32.9
35.3
38.2
Distribution
45.5
25.6
14.4
14.3
14.2
14.1
14.0
Total
219.0
189.4
179.0
177.4
179.9
186.4
194.9
Note: Totals may not sum due to independent rounding.
Table 3-63: ChU Emissions from Natural Gas Systems (MMT CO2 Eq.)a
Stage	1990	2005
Exploration*1	4.2	10.3
Production	58.8	80.4
2015 2016 2017 2018 2019
1.0	0.7	1.2	0.8	0.5
89.3	86.6 89.4 90.8 93.7
Energy 3-91

-------
Onshore Production
36.0
54.9
52.1
49.5
50.7
51.7
52.0
Gathering and Boosting0
18.5
23.9
36.6
36.3
38.0
38.3
40.9
Offshore Production
4.3
1.8
0.6
0.8
0.7
0.8
0.8
Processing
21.3
11.6
11.0
11.2
11.5
12.1
12.4
Transmission and Storage
57.2
36.1
34.1
34.5
32.4
34.8
37.0
Distribution
45.5
25.6
14.3
14.3
14.2
14.1
14.0
Total
186.9
164.2
149.8
147.3
148.7
152.5
157.6
Note: Totals may not sum due to independent rounding.
a These values represent CH4 emitted to the atmosphere. CH4 that is captured, flared, or otherwise controlled (and
not emitted to the atmosphere) has been calculated and removed from emission totals.
b Exploration includes well drilling, testing, and completions.
c Gathering and boosting includes gathering and boosting station routine vented and leak sources, gathering
pipeline leaks and blowdowns, and gathering and boosting station episodic events.
Table 3-64: ChU Emissions from Natural Gas Systems (kt)a
Stage
1990
2005
2015
2016
2017
2018
2019
Explorationb
167
412
42
28
49
33
21
Production
2,350
3,227
3,572
3,466
3,574
3,631
3,748
Onshore Production
1,441
2,197
2,085
1,981
2,026
2,068
2,081
Gathering and Boosting0
739
957
1,463
1,453
1,521
1,532
1,636
Offshore Production
170
73
24
32
26
31
31
Processing
853
463
440
448
460
483
497
Transmission and Storage
2,288
1,443
1,366
1,379
1,298
1,390
1,478
Distribution
1,819
1,023
574
573
569
565
560
Total
7,478
6,567
5,994
5,894
5,949
6,101
6,305
Note: Totals may not sum due to independent rounding.
a These values represent CH4 emitted to the atmosphere. CH4 that is captured, flared, or otherwise controlled (and
not emitted to the atmosphere) has been calculated and removed from emission totals.
b Exploration includes well drilling, testing, and completions.
c Gathering and boosting includes gathering and boosting station routine vented and leak sources, gathering
pipeline leaks and blowdowns, and gathering and boosting station episodic events.
Table 3-65: Non-combustion CO2 Emissions from Natural Gas Systems (MMT)
Stage
1990
2005
2015
2016
2017
2018
2019
Exploration
0.4
1.7
0.3
0.2
0.4
0.4
0.2
Production
3.0
4.5
7.6
7.5
7.3
9.8
11.0
Processing
28.3
18.8
21.0
22.0
23.0
23.1
24.8
Transmission and Storage
0.2
0.2
0.2
0.3
0.5
0.5
1.2
Distribution
0.1
+
+
+
+
+
+
Total
32.0
25.2
29.1
30.1
31.2
33.9
37.2
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 3-66: Non-combustion CO2 Emissions from Natural Gas Systems (kt)
Stage
1990
2005
2015
2016
2017
2018
2019
Exploration
421
1,651
282
193
445
355
236
Production
3,048
4,486
7,623
7,482
7,261
9,841
10,951
Processing
28,338
18,836
20,977
22,022
22,980
23,126
24,786
Transmission and Storage
181
176
228
339
498
546
1,244
Distribution
54
30
17
17
17
17
16
Total
32,042
25,179
29,127
30,054
31,200
33,885
37,234
Note: Totals may not sum due to independent rounding.
3-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 3-67: N2O Emissions from Natural Gas Systems (Metric Tons CO2 Eq.)
Stage
1990
2005
2015
2016
2017
2018
2019
Exploration
458
1,348
3,248
115
244
2,267
123
Production
4,359
5,804
9,835
8,892
4,453
5,094
5,591
Processing
NO
3,348
5,766
3,819
3,066
3,587
4,987
Transmission and Storage
257
309
346
382
462
234
630
Distribution
NO
NO
NO
NO
NO
NO
NO
Total
5,073
10,808
19,196
13,209
8,226
11,182
11,331
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
Table 3-68: N2O Emissions from Natural Gas Systems (Metric Tons N2O)
Stage
1990
2005
2015
2016
2017
2018
2019
Exploration
1.5
4.5
10.9
0.4
0.8
7.6
0.4
Production
14.6
19.5
33.0
29.8
14.9
17.1
18.8
Processing
NO
11.2
19.3
12.8
10.3
12.0
16.7
Transmission and Storage
0.9
1.0
1.2
1.3
1.6
0.8
2.1
Distribution
NO
NO
NO
NO
NO
NO
NO
Total
17.0
36.3
64.4
44.3
27.6
37.5
38.0
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
fviet had ©logy
See Annex 3.6 for the full time series of emissions data, activity data, and emission factors, and additional
information on methods and data sources—for example, the specific years of reporting data from EPA's
Greenhouse Gas Reporting Program (GHGRP) that are used to develop certain factors.
This section provides a general overview of the methodology for natural gas system emission estimates in the
Inventory, which involves the calculation of CH4, C02, and N20 emissions for over 100 emissions sources (i.e.,
equipment types or processes), and then the summation of emissions for each natural gas segment.
The approach for calculating emissions for natural gas systems generally involves the application of emission
factors to activity data. For most sources, the approach uses technology-specific emission factors or emission
factors that vary over time and take into account changes to technologies and practices, which are used to
calculate net emissions directly. For others, the approach uses what are considered "potential methane factors"
and emission reduction data to calculate net emissions. The estimates are developed with a Tier 2 approach. Tier 1
approaches are not used.
Emission Factors. Key references for emission factors for CH4 and C02 emissions from the U.S. natural gas industry
include a 1996 study published by the Gas Research Institute (GRI) and EPA (GRI/EPA 1996), the EPA's GHGRP (EPA
2020), and others.
The 1996 GRI/EPA study developed over 80 CH4 emission factors to characterize emissions from the various
components within the operating segments of the U.S. natural gas system. The GRI/EPA study was based on a
combination of process engineering studies, collection of activity data, and measurements at representative
natural gas facilities conducted in the early 1990s. Year-specific natural gas CH4 compositions are calculated using
U.S. Department of Energy's Energy Information Administration (EIA) annual gross production data for National
Energy Modeling System (NEMS) oil and gas supply module regions in conjunction with data from the Gas
Technology Institute (GTI, formerly GRI) Unconventional Natural Gas and Gas Composition Databases (GTI 2001).
These year-specific CH4 compositions are applied to emission factors, which therefore may vary from year to year
due to slight changes in the CH4 composition of natural gas for each NEMS region.
Energy 3-93

-------
GHGRP Subpart W data were used to develop CH4, C02, and N20 emission factors for many sources in the
Inventory. In the exploration and production segments, GHGRP data were used to develop emission factors used
for all years of the time series for well testing, gas well completions and workovers with and without hydraulic
fracturing, pneumatic controllers and chemical injection pumps, condensate tanks, liquids unloading,
miscellaneous flaring, gathering and boosting pipelines, and certain sources at gathering and boosting stations. In
the processing segment, for recent years of the times series, GHGRP data were used to develop emission factors
for leaks, compressors, flares, dehydrators, and blowdowns/venting. In the transmission and storage segment,
GHGRP data were used to develop factors for all years of the time series for LNG stations and terminals and
transmission pipeline blowdowns, and for pneumatic controllers for recent years of the times series.
Other data sources used for CH4 emission factors include Zimmerle et al. (2015) for transmission and storage
station leaks and compressors, GSI (2019) for underground storage well leaks, GTI (2009 and 2019) for commercial
and industrial meters, Lamb et al. (2015) for recent years for distribution pipelines and meter/regulator stations,
Zimmerle et al. (2019) for gathering and boosting stations, and Bureau of Ocean Energy Management (BOEM)
reports.
For C02 emissions from sources in the exploration, production and processing segments that use emission factors
not directly calculated from GHGRP data, data from the 1996 GRI/EPA study and a 2001 GTI publication were used
to adapt the CH4 emission factors into related C02 emission factors. For sources in the transmission and storage
segment that use emission factors not directly calculated from GHGRP data, and for sources in the distribution
segment, data from the 1996 GRI/EPA study and a 1993 GTI publication were used to adapt the CH4 emission
factors into non-combustion related C02 emission factors.
Flaring N20 emissions were estimated for flaring sources using GHGRP data.
See Annex 3.6 for more detailed information on the methodology and data used to calculate CH4, C02, and N20
emissions from natural gas systems.
Activity Data. Activity data were taken from various published data sets, as detailed in Annex 3.6. Key activity data
sources include data sets developed and maintained by EPA's GHGRP (EPA 2020); Enverus (Enverus 2020); BOEM;
Federal Energy Regulatory Commission (FERC); EIA; the Natural Gas STAR Program annual data; Oil and Gas
Journal; and PHMSA.
For a few sources, recent direct activity data are not available. For these sources, either 2018 data were used as a
proxy for 2019 data, or a set of industry activity data drivers was developed and used to calculate activity data over
the time series. Drivers include statistics on gas production, number of wells, system throughput, miles of various
kinds of pipe, and other statistics that characterize the changes in the U.S. natural gas system infrastructure and
operations. More information on activity data and drivers is available in Annex 3.6.
A complete list of references for emission factors and activity data by emission source is provided in Annex 3.6.
Calculating Net Emissions. For most sources, net emissions are calculated directly by applying emission factors to
activity data. Emission factors used in net emission approaches reflect technology-specific information, and take
into account regulatory and voluntary reductions. However, for production and transmission and storage, some
sources are calculated using potential emission factors, and the step of deducting CH4 that is not emitted from the
total CH4 potential estimates to develop net CH4 emissions is applied. To take into account use of such
technologies and practices that result in lower emissions but are not reflected in "potential" emission factors, data
are collected on both regulatory and voluntary reductions. Regulatory actions addressed using this method include
EPA National Emission Standards for Hazardous Air Pollutants (NESHAP) regulations for dehydrator vents.
Voluntary reductions included in the Inventory are those reported to Natural Gas STAR for certain sources.
Through EPA's stakeholder process on oil and gas in the Inventory, EPA received stakeholder feedback on updates
under consideration for the Inventory. Stakeholder feedback is noted below in Recalculations Discussion and
Planned Improvements.
The United States reports data to the UNFCCC using this Inventory report along with Common Reporting Format
(CRF) tables. This note is provided for those reviewing the CRF tables: The notation key "IE" is used for C02 and CH4
emissions from venting and flaring in CRF table l.B.2. Disaggregating flaring and venting estimates across the
3-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Inventory would involve the application of assumptions and could result in inconsistent reporting and, potentially,
decreased transparency. Data availability varies across segments within oil and gas activities systems, and emission
factor data available for activities that include flaring can include emissions from multiple sources (flaring, venting
and leaks).
Uncertainty and Time-Series Consistency
EPA has conducted a quantitative uncertainty analysis using the IPCC Approach 2 methodology (Monte Carlo
Simulation technique) to characterize the uncertainty for natural gas systems. For more information on the
approach, please see the memorandum Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2016: Natural
Gas and Petroleum Systems Uncertainty Estimates (2018 Uncertainty Memo).86
EPA used Microsoft Excels @RISK add-in tool to estimate the 95 percent confidence bound around CH4 and C02
emissions from natural gas systems for the current Inventory. Uncertainty estimates for N20 were not developed
given the minor contribution of N20 to emission totals. For the CH4 uncertainty analysis, EPA focused on the 14
highest-emitting sources for the year 2019, which together emitted 75 percent of methane from natural gas
systems in 2019, and extrapolated the estimated uncertainty for the remaining sources. Uncertainty was not
previously estimated specifically for C02 emissions, instead the uncertainty bounds calculated for CH4 were applied
to C02 emissions estimates. As part of the stakeholder process for the current Inventory, EPA developed an update
to the uncertainty analysis for C02. The update is documented in the memorandum, Inventory of U.S. Greenhouse
Gas Emissions and Sinks 1990-2019: Update for Natural Gas and Petroleum Systems C02 Uncertainty Estimates.87
EPA ultimately applied the same approach as was developed for CH4. For the C02 uncertainty analysis, EPA focused
on the 3 highest-emitting sources for the year 2018 (from the previous 1990-2018 Inventory), which together
emitted 82 percent of C02 from natural gas systems in 2018, and extrapolated the estimated uncertainty for the
remaining sources. The C02 uncertainty calculations were developed as part of the stakeholder process and were
based on the previous 1990-2018 Inventory; as a result, the uncertainty results from last year's Inventory for year
2018	are applied for this year's uncertainty analysis. In future years, the C02 uncertainty bounds will be calculated
using the most recent Inventory data. The @RISK add-in provides for the specification of probability density
functions (PDFs) for key variables within a computational structure that mirrors the calculation of the inventory
estimate. The IPCC guidance notes that in using this method, "some uncertainties that are not addressed by
statistical means may exist, including those arising from omissions or double counting, or other conceptual errors,
or from incomplete understanding of the processes that may lead to inaccuracies in estimates developed from
models." The uncertainty bounds reported below only reflect those uncertainties that EPA has been able to
quantify and do not incorporate considerations such as modeling uncertainty, data representativeness,
measurement errors, misreporting or misclassification. The understanding of the uncertainty of emission estimates
for this category evolves and improves as the underlying methodologies and datasets improve.
The results presented below provide the 95 percent confidence bound within which actual emissions from this
source category are likely to fall for the year 2019, using the IPCC methodology. The results of the Approach 2
uncertainty analysis are summarized in Table 3-69. Natural gas systems CH4 emissions in 2019 were estimated to
be between 133.4 and 180.1 MMT C02 Eq. at a 95 percent confidence level. Natural gas systems C02 emissions in
2019	were estimated to be between 31.3 and 44.3 MMT C02 Eq. at a 95 percent confidence level.
Uncertainty bounds for other years of the time series have not been calculated, but uncertainty is expected to vary
over the time series. For example, years where many emission sources are calculated with interpolated data would
likely have higher uncertainty than years with predominantly year-specific data. In addition, the emission sources
that contribute the most to CH4 and C02 emissions are different over the time series, particularly when comparing
recent years to early years in the time series. For example, venting emissions were higher and flaring emissions
were lower in early years of the time series, compared to recent years. Technologies also changed over the time
86	See .
87	Stakeholder materials, including draft and final memoranda for the current (i.e. 1990 to 2019) Inventory are available at
.
Energy 3-95

-------
series (e.g., liquids unloading with plunger lifts and reduced emissions completions were not used early in the time
series and cast iron distribution mains were more prevalent than plastic mains in early years). Transmission and
gas processing compressor leak and vent emissions were also higher in the early years of the time series.
Table 3-69: Approach 2 Quantitative Uncertainty Estimates for CH4 and Non-combustion CO2
Emissions from Natural Gas Systems (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTC02 Eq.)b
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Lower Upper



Boundb Bound'5
Boundb Bound'5
Natural Gas Systems
ch4
157.6
133.4 180.1
-15% +14%
Natural Gas Systems
C02
37.2
31.3 44.3
-16% 19%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for the year 2019 CH4 and year 2018 C02 emissions.
b All reported values are rounded after calculation. As a result, lower and upper bounds may not be duplicable from other
rounded values as shown in Table 3-63 and Table 3-64.
GHGRP data available (starting in 2011) and other recent data sources have improved estimates of emissions from
natural gas systems. To develop a consistent time series, for sources with new data, EPA reviewed available
information on factors that may have resulted in changes over the time series (e.g., regulations, voluntary actions)
and requested stakeholder feedback on trends as well. For most sources, EPA developed annual data for 1993
through 2010 by interpolating activity data or emission factors or both between 1992 and 2011 data points.
Information on time-series consistency for sources updated in this year's Inventory can be found in the
Recalculations Discussion below, with additional detail provided in supporting memos (relevant memos are cited in
the Recalculations Discussion). For detailed documentation of methodologies, please see Annex 3.5.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification Discussion
The natural gas systems emission estimates in the Inventory are continually being reviewed and assessed to
determine whether emission factors and activity factors accurately reflect current industry practices. A QA/QC
analysis was performed for data gathering and input, documentation, and calculation. QA/QC checks are
consistently conducted to minimize human error in the model calculations. EPA performs a thorough review of
information associated with new studies, GHGRP data, regulations, public webcasts, and the Natural Gas STAR
Program to assess whether the assumptions in the Inventory are consistent with current industry practices. The
EPA has a multi-step data verification process for GHGRP data, including automatic checks during data-entry,
statistical analyses on completed reports, and staff review of the reported data. Based on the results of the
verification process, the EPA follows up with facilities to resolve mistakes that may have occurred.88
As in previous years, EPA conducted early engagement and communication with stakeholders on updates prior to
public review. EPA held stakeholder webinars in September and November of 2020. EPA released memos detailing
updates under consideration and requesting stakeholder feedback.
In recent years, several studies have measured emissions at the source level and at the national or regional level
and calculated emission estimates that may differ from the Inventory. There are a variety of potential uses of data
from new studies, including replacing a previous estimate or factor, verifying or QA of an existing estimate or
88 See .
3-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
factor, and identifying areas for updates. In general, there are two major types of studies related to oil and gas
greenhouse gas data: studies that focus on measurement or quantification of emissions from specific activities,
processes and equipment, and studies that use tools such as inverse modeling to estimate the level of overall
emissions needed to account for measured atmospheric concentrations of greenhouse gases at various scales. The
first type of study can lead to direct improvements to or verification of Inventory estimates. In the past few years,
EPA has reviewed and in many cases, incorporated data from these data sources. The second type of study can
provide general indications of potential over- and under-estimates.
One comment on the public review draft suggested that the inventory estimates be compared with an
observational analysis from a 2019 Lan et al. study.89 Lan et al. estimated an average increasing trend of U.S. oil
and gas methane emissions of 3.4 percent +/-1A percent per year between 2006 and 2015, based on three U.S.
measurement sites that were "substantially influenced by O&NG activities." This study did not address the
magnitude of emissions. Nationally, in the Inventory, methane emissions from oil and gas decreased by an average
of 1 percent per year from 2006 to 2015, largely driven by the natural gas distribution and transmission and
storage segments. A key challenge in using these types of studies to assess Inventory results is having a relevant
basis for comparison (e.g., the two data sets should have comparable time frames and geographic coverage, and
the independent study should assess data from the Inventory and not another data set, such as the Emissions
Database for Global Atmospheric Research, or "EDGAR"). In an effort to improve the ability to compare the
national-level Inventory with measurement results that may be at other scales, a team at Harvard University along
with EPA and other coauthors developed a gridded inventory of U.S. anthropogenic methane emissions with 0.1
degree x 0.1 degree spatial resolution, monthly temporal resolution, and detailed scale-dependent error
characterization.90 The gridded methane inventory is designed to be consistent with the U.S. EPA's Inventory of
U.S. Greenhouse Gas Emissions and Sinks: 1990-2014 estimates for the year 2012, which presents national totals.91
An updated version of the gridded inventory is being developed and will improve efforts to compare results of the
Inventory with atmospheric studies.
Recalculations Discussion
EPA received information and data related to the emission estimates through GHGRP reporting, the annual
Inventory formal public notice periods, stakeholder feedback on updates under consideration, and new studies. In
September and November 2020, EPA released draft memoranda that discussed changes under consideration, and
requested stakeholder feedback on those changes. EPA then created updated versions of the memoranda to
document the methodology implemented in the current Inventory.92 Memoranda cited in the Recalculations
Discussion below are: Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Natural Gas
Customer Meter Emissions (Customer Meters memo) and Inventory of U.S. Greenhouse Gas Emissions and Sinks
1990-2019: Update for Produced Water Emissions (Produced Water memo).
EPA thoroughly evaluated relevant information available and made several updates to the Inventory, including
using revised emission factors and produced water volumes to calculate produced water emissions, and using GTI
2019 along with GTI 2009 study data to calculate customer meter emissions. These changes are discussed in detail
below. In addition, certain sources did not undergo methodological updates, but CH4 and/or C02 emissions
changed by greater than 0.05 MMT C02 Eq., comparing the previous estimate for 2018 to the current
(recalculated) estimate for 2018. For sources without methodological updates, the emissions changes were mostly
due to GHGRP data submission revisions and updates to well counts in the Enverus dataset.
89	See .
90	See .
91	See .
92	Stakeholder materials including draft and final memoranda for the current (i.e., 1990 to 2019) Inventory are available at
.
Energy 3-97

-------
The combined impact of revisions to 2018 natural gas sector CH4 emissions, compared to the previous Inventory, is
an increase from 140.0 to 152.5 MMT C02 Eq. (12.6 MMT C02 Eq., or 9 percent). The recalculations resulted in an
average increase in CH4 emission estimates across the 1990 through 2018 time series, compared to the previous
Inventory, of 6.6 MMT C02 Eq., or 4 percent.
The combined impact of revisions to 2018 natural gas sector C02 emissions, compared to the previous Inventory, is
a decrease from 35.0 MMT to 33.9 MMT, or 3 percent. The recalculations resulted in an average decrease in
emission estimates across the 1990 through 2018 time series, compared to the previous Inventory, of 0.1 MMT
C02 Eq., or 0.5 percent.
The combined impact of revisions to 2018 natural gas sector N20 emissions, compared to the previous Inventory, is
an increase from 10.4 kt C02 Eq. to 11.2 kt C02 Eq., or 8 percent. The recalculations resulted in an average increase
in emission estimates across the 1990 through 2018 time series, compared to the previous Inventory, of 6 percent.
In Table 3-70 and Table 3-71 below are categories in Natural Gas Systems with recalculations resulting in a change
of greater than 0.05 MMT C02 Eq., comparing the previous estimate for 2018 to the current (recalculated)
estimate for 2018. No changes made to N20 estimates resulted in a change greater than 0.05 MMT C02 Eq. For
more information, please see the Recalculations Discussion below.
Table 3-70: Recalculations of CO2 in Natural Gas Systems (MMT CO2)
Segment and Emission Source
Previous Estimate
Year 2018,
2020 Inventory
Current Estimate
Year 2018,
2021 Inventory
Current Estimate
Year 2019,
2021 Inventory
Exploration
0.4
0.4
0.2
HF Gas Well Completions
0.4
0.3
0.2
Production
9.6
9.8
11.0
Gathering Stations Flares Stacks
4.2
4.4
5.0
Processing
24.5
23.1
24.8
AGR Vents
17.5
16.7
16.5
Flares
7.0
6.4
8.3
Transmission and Storage
0.5
0.5
1.2
LNG Export Terminals
0.3
0.3
1.0
Distribution
+
+
+
Customer Meters
+
+
+
Total
35.0
33.9
37.2
+ Does not exceed 0.05 MMT C02.
Table 3-71: Recalculations of CH4 in Natural Gas Systems (MMT CO2 Eq.)
Previous
Estimate Year
2018,
Segment and Emission Source	2020 Inventory
Exploration
1.1
0.8
0.5
HF Gas Well Completions
1.0
0.8
0.5
Non-HF Gas Well Completions
0.1
+
+
Production
80.9
90.8
93.7
Produced Water (Onshore Production)
1.5
4.7
4.7
G&B Station Sources
31.4
35.0
37.3
Pneumatic Controllers (Onshore Production)
25.4
26.9
28.2
Liquids Unloading
4.4
5.1
4.4
HF Workovers
0.6
0.5
0.4
Chemical Injection Pumps
2.7
2.9
2.8
Kimray Pumps
1.8
1.9
1.8
Gas Engines
6.2
6.4
6.3
Compressors
1.6
1.7
1.7
Current Estimate
Year 2018, 2021
Inventory
Current Estimate
Year 2019, 2021
Inventory
3-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Processing	12.2	12.1	12.4
Reciprocating Compressors	1.6	1.5	1.2
Transmission and Storage	33.9	34.8	37.0
Reciprocating Compressors (Transmission)	9.2	9.3	10.2
Pneumatic Controllers (Storage)	+	0.6	0.6
Distribution	11.8	14.1	14.0
Customer Meters	1.4	3.7	3.7
Total	140.0	152.5	157.6
+ Does not exceed 0.05 MMT C02 Eq.
Exploration
HF Gas Well Completions (Recalculation with Updated Data)
HF gas well completions CH4 emissions estimates averaged no change across the time series. However, emissions
decreased by 18 percent in 2018, compared to the previous Inventory, with the largest change being in RECs with
Venting. These changes were due to GHGRP submission revisions.
Table 3-72: HF Gas Well Completions National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
HF Completions - Non-REC with







Venting
156,988
382,619
924
1,109
2,932
1,085
641
HF Completions - Non-REC with







Flaring
2,223
6,828
394
75
476
621
335
HF Completions - REC with







Venting
0
6,489
14,463
12,569
37,650
28,934
17,827
HF Completions - REC with







Flaring
0
1,855
8,760
4,581
5,656
1,344
1,139
Total Emissions
159,211
397,791
24,541
18,334
46,713
31,984
19,942
Previous Estimate
153,924
397,427
24,566
18,177
47,414
39,036
NA
NA (Not Applicable)
HF gas well completion C02 emissions estimates decreased by an average of approximately 1 percent across the
time series and decreased by 26 percent in 2018, compared to the previous Inventory, primarily due to decreases
in emissions from RECs with Flaring. These changes were due to GHGRP submission revisions.
Energy 3-99

-------
Table 3-73: HF Gas Well Completions National Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
HF Completions - Non-REC with







Venting
11
26
0.2
+
0.4
+
+
HF Completions - Non-REC with







Flaring
402
1,236
49
12
37
54
32
HF Completions - REC with







Venting
0
3
1
0
1
3
0
HF Completions - REC with Flaring
0
370
218
167
398
233
198
Total Emissions
413
1,634
268
179
436
290
230
Previous Estimate
399
1,633
268
177
449
392
NA
+ Does not exceed 0.05 kt C02
NA (Not Applicable)
Non-HF Gas Well Completions (Recalculation with Updated Data)
Non-HF gas well completion CH4 emissions estimates decreased by an average of 3 percent across the time series
and decreased by 82 percent in 2018, compared to the previous inventory. These changes were due to GHGRP
submission revisions and a correction to the Inventory calculations of the number of non-HF completions that
were vented versus flared in 2018.
Table 3-74: Non-HF Gas Well Completions National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Non-HF Completions - Vented
5,747
10,925
14,872
8,602
850
272
832
Non-HF Completions - Flared
20
38
40
82
714
481
0
Total Emissions
5,767
10,963
14,912
8,684
1,564
753
832
Previous Estimate
5,717
10,252
13,667
8,077
1,440
4,285
NA
NA (Not Applicable)
Production
Produced Water (Methodological Update)
EPA updated the calculation methodology for produced water to estimate emissions for all produced water from
natural gas wells. Previous inventories only estimated emissions for two CBM formations (i.e., Powder River in
Wyoming and Black Warrior in Alabama). The updated methodology includes updates to the produced water
quantities and the emission factor, each of which are discussed here. EPA's considerations for this source are
documented in the Produced Water Memo.93
Produced water quantities (i.e., bbl) from natural gas wells were obtained for 36 natural gas-producing states as
described below:
•	Produced water quantities for 1990-2018 were obtained using Drillinglnfo and Prism datasets from
Enverus for 29 states (i.e., AK, AL, AR, AZ, CA, CO, FL, ID, KY, LA, MD, Ml, MN, MO, MS, MT, NC, ND, NE,
NM, NV, NY, OR, SD, TN, TX, UT, VA, and WY) (Enverus 2021). Linear interpolation was used to correct an
obviously inaccurate new-zero produced water quantity value in Colorado for 1998.
•	For four additional states, produced water quantities for 1990-2018 were available on state agency
websites- KS (Kansas Department of Health and Environment 2020), OH (Ohio Environmental Protection
93 See .
3-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Agency 2020), OK (Oklahoma Department of Environmental Quality 2020), and PA (Pennsylvania
Department of Environmental Protection 2020). Produced water quantities for 2018 were used as proxy
data for 2019 for these four states.
• Produced water quantities for 1990-2018 were estimated for three states (IL, IN, and WV) using state-
level produced water production ratios for gas wells. Well-level produced water data for gas wells for
2011 were obtained from the Drillinglnfo dataset (Enverus 2021) and gas production data were obtained
from state agency websites - IL (Illinois Office of Oil and Gas Resource Management 2020), IN (Indiana
Division of Oil & Gas 2020), and WV (West Virginia Department of Environmental Protection 2020). Using
these well-level produced water data and the gas production data, production ratios were developed for
gas wells in each state. These production ratios were then applied to annual state-level gas production
data (2000-2019) from EIA (EIA 2020). Produced water quantities for 2018 were used as proxy data for
2019 for these three states.
EPA updated the produced water EF to use an EF consistent with the Production Module of the 2017 Oil and Gas
Tool,94 and applied this EF to all gas well produced water (EPA 2017). Overall, the update increases the emission
estimate for produced water (now including all gas production), by approximately three times in recent years,
compared to the previous Inventory.
EPA received feedback on this update through its September 2020 memo and through the public review draft of
the inventory.
A stakeholder indicated that the typical practice is to route produced water to a tank battery, once it reaches the
surface and has been separated from the oil and gas. A stakeholder requested that data from the latest 2017
Ground Water Protection Council produced water management practices survey be used to determine the percent
of produced water that is stored in tanks. The stakeholder indicated that approximately 16 percent of produced
water has the potential of being stored in a tank battery that could potentially flash (based on the 2012 Ground
Water Protection Council produced water management practices survey). After further assessment of the 2012
and 2017 water management practice surveys, EPA maintained the assumption that all produced water goes
through tanks and emissions are flashed, consistent with the approach used for the public review draft of the
Inventory.
A stakeholder commented that current regulations under 40 CFR 60 subpart OOOOa require that certain storage
vessels route emission vapors to a recovery device, flare, or other control device. EPA currently does not have
specific data to address the use of controls on produced water tanks but will continue to assess this issue in future
inventories should additional data become available.
Table 3-75: Produced Water National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Gas Well Produced Water
Previous Estimate
82,250
2,767
139,453
59,884
171,523
60,745
154,394
61,673
157,488
61,673
188,601
61,673
187,070
61,673
NA (Not Applicable)
Gathering and Boosting (G&B) Stations (Recalculation with Updated Data)
Methane emission estimates for sources at gathering and boosting stations increased in the current Inventory by
an average of 2 percent across the time series and increased by 11 percent in 2018, compared to the previous
Inventory. The G&B sources with the largest increase in CH4 emissions estimates for year 2018 are tanks (increase
of 70 kt, or 39 percent), gas engines (increase of 19 kt, or 5 percent), and station blowdowns (increase of 17 kt or
27 percent). These changes were due to GHGRP submission revisions.
94 Instructions for Using the 2017 EPA Nonpoint Oil and Gas Emissions Estimation Tool, Production Module. Produced by
Eastern Research Group, Inc. (ERG) for U.S. Environmental Protection Agency. October 2019.
Energy 3-101

-------
Flare stack C02 emissions at G&B stations increased in the current inventory by an average of 0.6 percent,
compared to the previous Inventory. These changes were due to GHGRP submission revisions.
Table 3-76: Gathering Stations Sources National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Compressors
130,744
165,028
261,792
260,325
278,692
296,279
309,465
Tanks
131,152
165,543
262,610
261,139
255,244
251,243
301,338
Station Blowdowns
20,715
26,148
41,479
41,247
63,833
78,550
68,773
Dehydrator Vents -
Large units
36,022
45,468
72,128
71,724
61,297
56,932
55,218
High-bleed Pneumatic
Devices
17,466
22,046
34,973
34,777
33,985
24,599
23,624
Intermittent Bleed
Pneumatic Devices
80,265
101,312
160,716
159,816
178,037
163,253
170,952
Low-Bleed Pneumatic
Devices
2,784
3,515
5,575
5,544
5,877
5,803
6,819
Gas Engines
173,040
218,415
346,483
344,542
369,192
392,459
410,376
Other Gathering
Sources
60,349
69,352
120,838
120,161
113,470
129,876
145,139
Total Emissions
652,538
823,648
1,306,595
1,299,276
1,359,628
1,398,994
1,491,704
Previous Estimate
641,624
815,454
1,293,262
1,281,711
1,281,484
1,257,799
NA
NA (Not Applicable)







ble 3-77: Gathering Stations Flare Stacks National CO2 Emissions (Metric Tons CO2)
Source
1990
2005
2015
2016
2017
2018
2019
Flare Stacks
1,367,178
1,725,682
2,737,537
2,722,202
2,317,495
4,386,761
5,005,631
Previous Estimate
1,354,751
1,721,783
2,730,646
2,706,255
2,300,171
4,205,760
NA
NA (Not Applicable)
Pneumatic Controllers (Recalculation with Updated Data)
Pneumatic controller CH4 emission estimates increased in the current Inventory by an average of 3.9 percent
across the time series, compared to the previous Inventory. This change was due to GHGRP submission revisions
which increased the number of intermittent bleed controllers and updates to well counts in the Enverus dataset.
Table 3-78: Production Segment Pneumatic Controller National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Low Bleed
0
24,049
33,107
32,610
35,856
35,671
33,538
High Bleed
291,948
481,227
106,286
107,714
112,764
92,886
74,565
Intermittent Bleed
190,386
557,410
979,719
923,468
954,461
947,089
1,018,428
Total Emissions
482,334
1,062,685
1,119,112
1,063,791
1,103,082
1,075,645
1,126,531
Previous Estimate
490,594
1,023,770
1,072,732
1,037,136
1,062,086
1,016,357
NA
NA (Not Applicable)
Liquids Unloading (Recalculation with Updated Data)
Liquids unloading CH4 emission estimates increased for 2018 by 15 percent in the current Inventory, compared to
the previous Inventory. Compared to the previous Inventory, on average across the time series, liquids unloading
3-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
CH4 emission estimates increased more than 2 percent. These changes were due to GHGRP submission revisions
and updates to well counts in the Enverus dataset.
Table 3-79: Liquids Unloading National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Unloading with Plunger Lifts
NO
127,621
65,375
61,359
60,724
83,841
76,937
Unloading without Plunger Lifts
355,840
250,973
102,166
84,096
90,767
120,146
98,892
Total Emissions
355,840
378,594
167,540
145,455
151,492
203,987
175,828
Previous Estimated Emissions
371,391
372,614
160,061
127,663
129,790
177,298
NA
NO (Not Occurring)
NA (Not Applicable)
HF Gas Well Workovers (Recalculation with Updated Data)
HF gas well workover CH4 emissions decreased an average of 1 percent across the time series and decreased by 16
percent in 2018, when comparing the current Inventory to the previous Inventory, mostly due to decreases in
emissions from RECs with Venting. These changes were due to GHGRP submission revisions.
Table 3-80: HF Gas Well Workovers National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
HF Workovers - Non-REC with
Venting
25,774
60,903
1,752
7,530
8,638
1,394
4,301
HF Workovers - Non-REC with
Flaring
365
953
80
72
521
1,094
606
HF Workovers - REC with Venting
NO
576
8,685
6,384
16,146
18,010
8,824
HF Workovers - REC with Flaring
NO
4
1,695
1,234
4,885
39
257
Total Emissions
26,139
62,437
12,212
15,220
30,190
20,537
13,988
Previous Estimate
26,139
62,437
12,175
15,155
31,485
24,422
NA
NO (Not Occurring)
NA (Not Applicable)
Gas Engines (Recalculation with Updated Data)
Gas engine (combustion slip) CH4 emissions increased an average of 5 percent across the time series, compared to
the previous Inventory. These changes were due to updates to well counts in the Enverus dataset.
Table 3-81: Gas Engine National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Gas Engines
Previous Estimate
116,684
116,558
129,715
123,713
132,093
125,843
124,835
119,100
120,272
114,599
116,437
110,432
111,886
NA
NA (Not Applicable)
Chemical Injection Pumps (Recalculation with Updated Data)
Chemical injection pump CH4 emissions estimates increased an average of 4 percent across the time series,
compared to the previous Inventory. These changes were due to updates to well counts in the Enverus dataset.
Table 3-82: Chemical Injection Pump National Emissions (Metric Tons ChU)
Source	1990	2005	2015 2016 2017 2018 2019
Energy 3-103

-------
Chemical Injection Pumps	26,060	87,007 117,857 116,038 115,322 114,636 112,843
Previous Estimate	26,323	83,687 113,336 113,243 111,421 109,376	NA
NA (Not Applicable)
Kimray Pumps (Recalculation with Updated Data)
CH4 emissions from Kimray pumps decreased by an average of 2 percent across time series, compared to the
previous Inventory. These changes were due to updates to well counts in the Enverus dataset.
Table 3-83: Kimray Pumps National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Kimray Pumps
Previous Estimate
141,304
148,064
109,923
114,936
76,818
73,850
75,485
73,660
75,042
72,475
74,596
71,125
73,426
NA
NA (Not Applicable)
Compressors (Recalculation with Updated Data)
Compressors CH4 emissions estimates increased an average of 1 percent across the time series, compared to the
previous Inventory. These changes were due to updates to well counts in the Enverus dataset.
Table 3-84: Compressors National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Compressors
Previous Estimate
28,189
28,645
56,528
55,520
70,121
67,412
68,904
67,239
68,500
66,157
68,093
64,925
67,025
NA
NA (Not Applicable)
Well Counts (Recalculation with Updated Data)
EPA uses annual producing gas well counts as an input for estimates of emissions from multiple sources in the
Inventory, including exploration well testing, pneumatic controllers, chemical injection pumps, well workovers, and
equipment leaks. Annual well count data are obtained from Enverus for the entire time series during each
Inventory cycle, and a new data processing methodology was implemented this year due to a restructuring of the
Enverus well count data (Enverus 2021). Due to the data restructuring and the new processing methodology,
annual gas well counts increased by an average of 1 percent across the 1990-2018 time series and increased by 5
percent in 2018, compared to the previous Inventory.
Table 3-85: National Gas Well Counts
Source
1990
2005
2015
2016
2017
2018
2019
Gas Wells
Previous Estimate
185,141
193,232
351,982
346,484
436,432
419,692
429,697
419,346
427,046
412,601
424,507
405,026
417,866
NA
NA (Not Applicable)
In December 2020, EIA released an updated time series of national oil and gas well counts (covering 2000 through
2018). EIA estimates 969,136 total producing wells for year 2019. EPA's total well count for this year is 939,637.
EPA's well counts are generally lower than ElA's (e.g., around 3 percent lower in 2019). ElA's well counts include
side tracks (i.e., secondary wellbore away from original wellbore in order to bypass unusable formation, explore
nearby formations, or other reasons) completions, and recompletions, and therefore are expected to be higher
than EPA's which include only producing wells. EPA and EIA use a different threshold for distinguishing between oil
versus gas (EIA uses 6 mcf/bbl, while EPA uses 100 mcf/bbl), which results in EIA having a lower fraction of oil wells
3-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
(e.g., 44 percent versus EPA's 56 percent in 2019) and a higher fraction of gas wells (e.g., 56 percent versus EPA's
44 percent in 2019) than EPA.
Processing
Acid Gas Removal (Recalculation with Updated Data)
Acid gas removal unit (AGR) C02 emission estimates decreased by less than 1 percent across the time series,
compared to the previous Inventory. The 2018 estimate decreased by 4 percent when compared to the previous
inventory. These changes are due to GHGRP submission revisions.
Table 3-86: AGR National CO2 Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
Acid Gas Removal
Previous Estimate
28,282
28,282
15,281
15,339
14,878
14,979
16,741
16,679
17,218
17,182
16,699
17,451
16,498
NA
NA (Not Applicable)
Flares (Recalculation with Updated Data)
Processing segment flare C02 emission estimates decreased by less than 1 percent across the 1993 to 2018 time
series in the current Inventory. Processing segment flare C02 emission estimates decreased by approximately 8
percent for 2018 in the current Inventory, compared to the previous Inventory. These changes are due to GHGRP
submission revisions.
Table 3-87: Processing Segment Flares National Emissions (kt CO2)
Source
1990
2005
2015
2016
2017
2018
2019
Flares
NO
3,517
6,057
5,246
5,726
6,394
8,257
Previous Estimate
NO
3,515
6,054
5,195
5,679
6,981
NA
NO (Not Occurring)
NA (Not Applicable)
Reciprocating Compressors (Recalculation with Updated Data)
Reciprocating compressor CH4 emission estimates decreased by less than 1 percent on average for 2011 to 2018 in
the current Inventory and decreased by 5 percent for 2018 in the current Inventory, compared to the previous
Inventory. This decrease in the CH4 emission estimates is due to GHGRP submission revisions.
Table 3-88: Processing Segment Reciprocating Compressors National Emissions (Metric
Tons CH4)
Source
1990
2005
2015
2016
2017
2018
2019
Reciprocating Compressors
Previous Estimate
324,939
324,939
NA
NA
67,988
67,982
63,565
63,682
64,789
64,955
59,373
62,574
46,652
NA
NA (Not Applicable)
Transmission and Storage
There were no methodological updates to the transmission and storage segment, but there were recalculations
due to updated data that resulted in an average increase in calculated emissions over the time series from this
segment of 0.19 MMT C02 Eq. of CH4 (or 0.5 percent) and less than 0.01 MMT C02 (or 2 percent).
Energy 3-105

-------
Transmission Station Reciprocating Compressors (Recalculation with Updated Data)
Methane emission estimates from reciprocating compressors at transmission compressor stations increased by an
average of 0.2 percent for 2011 to 2018, compared to the previous Inventory. This increase in the CH4 emission
estimates is due to GHGRP submission revisions.
Table 3-89: Transmission Station Reciprocating Compressors National Emissions (Metric
Tons CH4)
Source
1990
2005
2015
2016
2017
2018
2019
Transmission Station -







NO
NO
341,316
345,224
347,830
373,233
406,453
Reciprocating Compressors


Previous Estimate
NO
NO
341,316
345,224
346,527
369,976
NA
NO (Not Occurring)
NA (Not Applicable)
Storage Pneumatic Controllers (Recalculation with Updated Data)
Storage segment pneumatic controller CH4 emission estimates increased in the current Inventory for 2014-2018,
compared to the previous Inventory. Emission estimates for 2018 increased by 24,169 metric tons CH4, compared
to the previous Inventory. This increase in the CH4 emission estimates is due to GHGRP submission revisions.
Table 3-90: Storage Segment Pneumatic Controller National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Low Bleed
NE
NE
15,902
19,904
20,465
17,291
17,187
High Bleed
NE
NE
5,734
5,648
6,419
6,485
6,365
Intermittent Bleed
NE
NE
401
413
480
535
516
Total Emissions
44,441
35,263
22,038
25,965
27,364
24,310
24,067
Previous Estimate
44,441
35,263
22,094
1,402
27,364
141
NA
NE (Not Estimated)
NA (Not Applicable)
LNG Export Terminals (Recalculation with Updated Data)
LNG export terminal C02 emissions estimates for equipment leaks, compressors, and flares increased by 20
percent in 2018, compared to the previous Inventory. This increase in the C02 emission estimate for 2018 is due to
GHGRP submission revisions.
Table 3-91: LNG Export Terminal National Emissions (Metric Tons CO2)
Source 1990
2005
2015
2016
2017
2018
2019
LNG Export Terminals (eq. leaks, ^
compressors, flares)
Previous Estimate 23
23
23
23
23
97,935
97,935
277,979
277,979
327,535
273,956
979,142
NA
NA (Not Applicable)
Distribution
Customer Meters (Methodological Update)
3-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
EPA updated the commercial and industrial meters methodologies to use leak data from the GTI 2009 and GTI
2019 studies. The GTI 2019 study measured CH4 emissions rates at commercial and industrial meters in six regions
across the country and calculated population EFs for each meter type. The GTI 2009 study conducted similar
measurements and was used to calculate emissions for commercial and industrial meters in the previous (1990 to
2018) Inventory. EPA applied weighted average population EFs from the two studies across the time series for the
methodology implemented in the Inventory. The Customer Meters memo provides details on the methodology
implemented into the final inventory.
Commercial and industrial meter CH4 and C02 emissions increased by an average of 173 percent across the time
series, compared to the previous Inventory. The increase in both CH4 and C02 emissions is due to differences in the
EFs used in the current Inventory and the previous Inventory. The previous inventory used a lower EF (based on
commercial meter measurements only) and applied that EF to both commercial and industrial meter counts. The
updated methodology uses commercial meter data from both the 2009 and 2019 GTI studies to develop an EF that
is applied to commercial meter counts, and uses industrial meter data from both the 2009 (leak emissions only)
and GTI 2019 studies to develop an EF that is applied to industrial meter counts. No change was made to the
activity data approach.
EPA received comments on the September 2020 version of the Customer Meters Memo and through the public
review draft of the Inventory. These comments included a recommendation to delay updates until additional data
could be collected. The comments also recommended using separate EFs for commercial and industrial meters and
region-specific EFs. The largest source of emissions from customer meters in the 2009 study was vented emissions
from industrial meters, with an average emission factor per meter of 3,487 kg/year, compared with an average
emission factor per industrial meter from leaks of 105 kg/year. Venting emissions were observed and measured at
2 out of the 6 companies participating in the 2009 GTI study. This source of emissions was not studied in the 2019
GTI study. The final methodology for industrial meters uses an EF calculated only from leak emissions, which have
less variability, and does not include the more limited and highly variable vented emissions. EPA did not use
region-specific EFs due to the limited data available for each region, but did finalize separate EFs for commercial
and industrial meters that rely on the leak emissions from the 2009 and 2019 GTI studies. Using data from both
studies to calculate population EFs greatly increases the number of data points that serve as the basis of the EFs,
instead of only using the commercial meter EF from the 2009 GTI study. EPA seeks stakeholder feedback on
upcoming or ongoing research studies that measure vented emissions from industrial meters.
Table 3-92: Commercial and Industrial Meter National Emissions (Metric Tons ChU)
Source
1990
2005
2015
2016
2017
2018
2019
Commercial Meters
99,129
121,634
127,615
128,108
128,698
129,130
129,796
Industrial Meters
22,926
21,653
19,775
19,828
19,419
19,426
19,239
Total
122,055
143,287
147,390
147,936
148,118
148,555
149,036
Previous Estimate
43,362
52,605
54,919
55,129
55,324
56,140
NA
NA (Not Applicable)
Table 3-93: Commercial and Industrial Meter National Emissions (Metric Tons CO2)
Source
1990

2005
2015
2016
2017
2018
2019
Commercial Meters
2,919

3,581
3,757
3,772
3,789
3,802
3,822
Industrial Meters
675

638
582
584
572
572
566
Total
3,594

4,219
4,340
4,356
4,361
4,374
4,388
Previous Estimate
1,277

1,549
1,617
1,623
1,629
1,653
NA
NA (Not Applicable)
Energy 3-107

-------
Planned Improvements
EPA seeks stakeholder feedback on the improvements noted below.
Post-Meter Fugitive Emissions
The Inventory does not currently include estimates for post-meter fugitive (leakage) emissions. The 2019
Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2019) include methods and
default emission factors to estimate these emissions. In IPCC 2019, post-meter fugitives includes leak emissions
from appliances in commercial and residential sectors (leakage from house piping and appliances, including home
heating, water heating, stoves, and barbecues), leakage at industrial plants and power stations (leakage beyond
gas meters including internal piping), and leakage from natural gas-fueled vehicles (vehicles with fuels produced
from natural gas e.g., LNG, CNG, RNG).
For consistency with IPCC 2019, and to improve completeness of the Inventory estimates, EPA is considering
updating next year's Inventory to include this emission source. EPA will seek stakeholder feedback on emission
factors and activity data for this source.
Anomalous Emissions Events (Well Blowouts)
In recent years, a number of studies have assessed and, in some cases, quantified total emissions for gas well
blowout events.95,96 EPA is considering updating next year's inventory to include these events. EPA will seek
stakeholder feedback on estimates for this source.
Transmission and Storage
Storage Wells
As part of the stakeholder process for the current (1990 to 2019) Inventory, EPA developed draft CH4 emission
estimates for underground storage well leak emissions in the transmission and storage segment. EPA's
considerations for this source are documented in the EPA memorandum Inventory ofU. S. Greenhouse Gas
Emissions and Sinks 1990-2019: Updates Under Consideration to Natural Gas Underground Storage Well Emissions
(Underground Storage Wells memo).97 EPA presented multiple options to calculate storage well emissions in the
Underground Storage Wells memo, and in the public review draft of the Inventory, presented estimated emissions
using a 'per station' EF along with underground storage station counts. EPA received comment suggesting a recent
PHMSA data set for national storage well counts.98 EPA is considering the use of the national storage well data set
and the GSI emission factors in next year's Inventory.
Table 3-94: Draft Underground Storage Wells National Emissions (Metric Tons ChU) and Well
Counts - Not Included in Totals
Source	1990	2005	2015 2016 2017 2018 2019
Storage Wells CH4 (EPA
, ® 7,838 7,431	7,566 7,508 7,489 7,431 7,566
draft3 ' '	'
95	See .
96	See .
97	Stakeholder materials, including draft memoranda for the current (i.e., 1990 to 2019) Inventory are available at
.
98	See .
3-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Storage Wells CH4 (PHMSA
and GSI)b
7,523 7,483 7,426
Well Counts (PHMSA)
Current Estimate CH4C
Current Estimate Well
Counts
16,853 18,524
13,565 14,910
14,268 14,192 14,084
14,250 13,428 13,632 15,439 15,495
17,703 16,682 16,936 19,181 19,250
NA (Not Applicable)
a Estimate developed using a 'per station' weighted average EF (based on data from the GSI 2019 study,
GHGRP wellhead component counts, and field type distributions from EIA) and the number of
underground storage stations over the time series (already in the Inventory).
b Estimate developed using PHMSA storage well counts and GSI emission factors. EPA has not developed an
approach for the full time series for the activity data or emission factors, and is showing preliminary
estimates for 2017 through 2019 emissions because those are the years where PHMSA storage well counts
are available.
c The Current Estimate shows the routine storage well leak emissions only, to allow for a direct comparison,
and does not include the Aliso Canyon leak emissions that occurred in 2015 and 2016 and that are
included in the Inventory under the storage wells line item.
Transmission Station Counts
Stakeholder feedback suggested alternate approaches for calculating the annual number of transmission stations.
EPA will consider the update for next year's Inventory.
Mud Degassing
As part of the stakeholder process for the current (1990 to 2019) Inventory, EPA developed new CH4 emission
estimates for onshore mud degassing in the exploration segment. To date, the Inventory has not included
emissions from onshore exploration mud degassing. EPA's considerations for this source are documented in the
EPA memorandum Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update Under Consideration
for Mud Degassing Emissions (Mud Degassing memo). EPA estimated preliminary emissions using CH4 emission
factors from EPA (EPA 1977) and well counts data from Enverus Drillinglnfo data. In estimating national preliminary
CH4 estimates for mud degassing, EPA incorporated an estimate of 26 days as an average drilling duration and 61.2
percent (by weight) as the default CH4 content of natural gas. EPA developed preliminary national estimates for 2
different scenarios: 1) EPA assumed 80 percent of drilling operations were performed using water-based muds and
the remaining 20 percent used oil-based muds; and 2) EPA assumed 100 percent of drilling operations were
performed using water-based muds. Methane emissions from mud degassing averaged 73 kt over the time series
for scenario 1 and 86 kt for scenario 2 (100 percent water-based muds), or around 2 MMT C02 Eq. This update
would increase emissions from the exploration segment but would have a small impact on overall CH4 emissions
from natural gas systems.
EPA notes that estimates for mud degassing using similar assumptions are included in several other bottom-up
inventories for greenhouse gases and other gases, including New York state, and the NEI.
EPA received feedback on this update through its September 2020 memo and through the public review draft of
the Inventory. A stakeholder indicated the 12 inch diameter borehole and 25 percent formation porosity
assumptions used in developing the CH4 emission factor for water-based muds are outdated and recommended
that an 8 inch borehole diameter and 10 percent porosity should be considered in developing the CH4 EF. A
stakeholder commented that current onshore practices are to drill with balanced or slightly over-balanced mud
systems that keep gas from being entrained in the drilling mud and that mud degassing systems are rarely needed
or used. A stakeholder also indicated that mainly oil-based muds are used for horizontal/lateral drilling and water-
based muds are more frequently used for vertical drilling.
Additionally, EPA also received comments on the average drilling duration used in developing the draft estimates
for onshore mud degassing. A stakeholder comment stated that EPA should only consider the duration the drill
Energy 3-109

-------
spends in the producing formation. Another comment indicated that EPA's average drilling duration assumption of
26 days per well is high and presented 2 examples-a Marcellus well takes 10 days to drill with 2-3 days in the
producing formation; and the drilling duration in the Fayetteville shale dropped from 20 days in 2007 to 11 days in
2009. EPA's average drilling duration assumption of 26 days per well is comparable to average drilling duration
developed from other inventories (New York - 24 days/well and CenSARA - 22 days/well). Refer to the Mud
Degassing memo for further details.
EPA continues to seek feedback on average total drilling days and drilling days in the producing formation, CH4
content of the gas, and the effect of balanced and over-balanced mud degassing systems. EPA will further assess
the average drilling duration using updated Enverus data. Additionally, EPA is considering developing C02
estimates for onshore production mud degassing using the CH4 estimates and a ratio of C02-to-CH4.
Table 3-95: Draft Mud Degassing National ChU Emissions - Not Included in Totals (Metric
Tons CH4)
Source
1990
2005
2015
2016
2017
2018
2019
Scenario 1 (80/20)
95,133
146,766
18,211
11,736
19,301
19,301
19,301
Scenario 2 (100)
111,922
172,666
21,425
13,808
22,707
22,707
22,707
Previous Estimate
NA
NA
NA
NA
NA
NA
NA
NA (Not Applicable)
Scenario 1 (80/20) - 80% water-based mud usage and 20% oil-based mud usage
Scenario 2 (100) - 100% water-based mud usage
Upcoming Data, and Additional Data that Could Inform the Inventory
EPA will assess new data received by the EPA Methane Challenge Program on an ongoing basis, which may be used
to validate or improve existing estimates and assumptions.
EPA continues to track studies that contain data that may be used to update the Inventory. EPA will also continue
to assess studies that include and compare both top-down and bottom-up emission estimates, which could lead to
improved understanding of unassigned high emitters (e.g., identification of emission sources and information on
frequency of high emitters) as recommended in stakeholder comments.
EPA also continues to seek new data that could be used to assess or update the estimates in the Inventory. For
example, stakeholder comments have highlighted areas where additional data that could inform the Inventory are
currently limited or unavailable:
•	Tank and flaring malfunction and control efficiency data.
•	Improved equipment leak data
•	Activity data and emissions data for production facilities that do not report to GHGRP.
EPA will continue to seek available data on these and other sources as part of the process to update the Inventory.
3.8 Abandoned Oil and Gas Wells (CRF Source
Categories lB2a and lB2b)
The term "abandoned wells" encompasses various types of wells, including orphaned wells and other non-
producing wells:
•	Wells with no recent production, and not plugged. Common terms (such as those used in state databases)
might include: inactive, temporarily abandoned, shut-in, dormant, and idle.
3-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Wells with no recent production and no responsible operator. Common terms might include: orphaned,
deserted, long-term idle, and abandoned.
•	Wells that have been plugged to prevent migration of gas or fluids.
The U.S. population of abandoned wells (including orphaned wells and other non-producing wells) is around 3.4
million (with around 2.7 million abandoned oil wells and 0.6 million abandoned gas wells). The methods to
calculate emissions from abandoned wells involve calculating the total populations of plugged and unplugged
abandoned oil and gas wells in the U.S. An estimate of the number of orphaned wells within this population is not
developed as part of the methodology. Wells that are plugged have much lower average emissions than wells that
are unplugged (less than 1 kg CH4 per well per year, versus over 100 kg CH4 per well per year). Around 40 percent
of the abandoned well population in the United States is plugged. This fraction has increased over the time series
(from around 19 percent in 1990) as more wells fall under regulations and programs requiring or promoting
plugging of abandoned wells.
Abandoned oil wells. Abandoned oil wells emitted 209 kt CH4 and 4 kt C02 in 2019. Emissions of both gases
decreased by 10 percent from 1990, while the total population of abandoned oil wells increased 28 percent.
Abandoned gas wells. Abandoned gas wells emitted 55 kt CH4 and 2 kt C02 in 2019. Emissions of both gases
increased by 38 percent from 1990, as the total population of abandoned gas wells increased 84 percent.
Table 3-96: ChU Emissions from Abandoned Oil and Gas Wells (MMT CO2 Eq.)
Activity	1990	2005	2015 2016 2017 2018	2019
Abandoned Oil Wells Ti (To ^9 ^9 5J Ti	52
Abandoned Gas Wells	1.0 1.2	1.5 1.5 1.5 1.5	1.4
Total 6.8 7.2 7.4 7.4 7.2 7.3	6.6
Note: Totals may not sum due to independent rounding.
Table 3-97: ChU Emissions from Abandoned Oil and Gas Wells (kt)
Activity	1990	2005	2015	2016	2017	2018	2019
Abandoned Oil Wells 231 238 235 236 229 231	209
Abandoned Gas Wells	40	49	59	60	58	59	55
Total 271 287 294 296 288 290	263
Note: Totals may not sum due to independent rounding.
Table 3-98: CO2 Emissions from Abandoned Oil and Gas Wells (MMT CO2)
Activity 1990 2005 2015 2016 2017 2018	2019
Abandoned Oil Wells + + + + + +	+
Abandoned Gas Wells + + + + + +	+
Total + + + + + +	+
+ Does not exceed 0.05 MMT C02.
Energy 3-111

-------
Table 3-99: CO2 Emissions from Abandoned Oil and Gas Wells (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Abandoned Oil Wells
5
5
5
5
5
5
4
Abandoned Gas Wells
2
2
3
3
3
3
2
Total
6
7
7
7
7
7
7
Note: Totals may not sum due to independent rounding.
Methodology
EPA developed abandoned well CH4 emission factors using data from Kang et al. (2016) and Townsend-Small et al.
(2016). Plugged and unplugged abandoned well CH4 emission factors were developed at the national-level
(emission data from Townsend-Small et al.) and for the Appalachia region (using emission data from
measurements in Pennsylvania and Ohio conducted by Kang et al. and Townsend-Small et al., respectively). The
Appalachia region emissions factors were applied to abandoned wells in states in the Appalachian basin region,
and the national-level emission factors were applied to all other abandoned wells.
EPA developed abandoned well C02 emission factors using the CH4 emission factors and an assumed ratio of C02-
to-CH4 gas content, similar to the approach used to calculate C02 emissions for many sources in Petroleum
Systems and Natural Gas Systems. For abandoned oil wells, EPA used the Petroleum Systems default production
segment associated gas ratio of 0.020 MT C02/MT CH4, which was derived through API TankCalc modeling runs. For
abandoned gas wells, EPA used the Natural Gas Systems default production segment CH4 and C02 gas content
values (GRI/EPA 1996, GTI 2001) to develop a ratio of 0.044 MT COz/MT CH4.
The total population of abandoned wells over the time series was estimated using historical data and Enverus data.
The total abandoned well population was then split into plugged and unplugged wells by assuming that all
abandoned wells were unplugged in 1950, using year-specific Enverus data to calculate the fraction of plugged
abandoned wells (31 percent in 2016, 34 percent in 2017 and 2018, and 41 percent in 2019), and applying linear
interpolation between the 1950 value and 2016 value to calculate the plugged fraction for intermediate years. See
the memorandum Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2016: Abandoned Wells in Natural
Gas and Petroleum Systems (2018 Abandoned Wells Memo) for details." Due to changes in the structure of the
Enverus data, the fraction of abandoned wells that are plugged was calculated uniquely for 2019. See Planned
Improvements section below for more information.
Abandoned Oil Wells
Table 3-100: Abandoned Oil Wells Activity Data, ChU and CO2 Emissions (kt)
Source
1990
2005
2015
2016
2017
2018
2019
Plugged abandoned oil wells
394,907
624,930
789,418
809,774
890,458
905,866
1,109,167
Unplugged abandoned oil







wells
1,720,692
1,809,892
1,813,090
1,819,396
1,765,889
1,777,547
1,604,291
Total Abandoned Oil Wells
2,115,599
2,434,821
2,602,508
2,629,170
2,656,346
2,683,413
2,713,458
Abandoned oil wells in







Appalachia
26%
24%
23%
23%
23%
23%
23%
Abandoned oil wells outside







of Appalachia
74%
76%
77%
77%
77%
77%
77%
CH4 from plugged







abandoned oil wells (MT)
0
0
1
1
1
1
1
99 See .
3-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
CH4from unplugged
abandoned oil wells (MT)	231	238
Total CH4from Abandoned
oil wells (MT)	231	238
Total C02 from Abandoned
oil wells (MT)	5	5
235	236	229	230	208
235	236	229	231	209
Abandoned Gas Wells
Table 3-101: Abandoned Gas Wells Activity Data, ChU and CO2 Emissions (kt)
Source
1990
2005
2015
2016
2017
2018
2019
Plugged abandoned gas wells
65,559
119,655
179,428
186,116
206,735
212,338
264,277
Unplugged abandoned gas
wells
285,654
346,540
412,101
418,164
409,982
416,663
382,248
Total Abandoned Gas Wells
351,213
466,196
591,529
604,280
616,717
629,001
646,525
Abandoned gas wells in
Appalachia
28%
29%
30%
30%
30%
30%
30%
Abandoned gas wells outside
of Appalachia
72%
- /
71%
70%
70%
70%
70%
70%
CH4from plugged abandoned
gas wells (kt)
0
0
0
0
0
0
0
CH4from unplugged
abandoned gas wells (kt)
39
49
59
59
58
59
54
Total CH4 from abandoned
gas wells (kt)
40
49
59
60
58
59
55
Total C02 from abandoned
gas wells (kt)
2
2
3
3
3
3
2
Uncertainty and Time-Series Consistency
To characterize uncertainty surrounding estimates of abandoned well emissions, EPA conducted a quantitative
uncertainty analysis using the IPCC Approach 2 methodology (Monte Carlo simulation technique). See the 2018
Abandoned Wells Memo for details of the uncertainty analysis methods. EPA used Microsoft Excel's @ RISK add-in
tool to estimate the 95 percent confidence bound around total methane emissions from abandoned oil and gas
wells in year 2019, then applied the calculated bounds to both CH4 and C02 emissions estimates for each
population. The @RISK add-in provides for the specification of probability density functions (PDFs) for key variables
within a computational structure that mirrors the calculation of the inventory estimate. EPA used measurement
data from the Kang et al. (2016) and Townsend-Small et al. (2016) studies to characterize the CH4 emission factor
PDFs. For activity data inputs (e.g., total count of abandoned wells, split between plugged and unplugged), EPA
assigned default uncertainty bounds of ± 10 percent based on expert judgment.
The IPCC guidance notes that in using this method, "some uncertainties that are not addressed by statistical means
may exist, including those arising from omissions or double counting, or other conceptual errors, or from
incomplete understanding of the processes that may lead to inaccuracies in estimates developed from models." As
a result, the understanding of the uncertainty of emission estimates for this category evolves and improves as the
underlying methodologies and datasets improve. The uncertainty bounds reported below only reflect those
uncertainties that EPA has been able to quantify and do not incorporate considerations such as modeling
uncertainty, data representativeness, measurement errors, misreporting or misclassification.
The results presented below in Table 3-102 provide the 95 percent confidence bound within which actual
emissions from abandoned oil and gas wells are likely to fall for the year 2019, using the recommended IPCC
Energy 3-113

-------
methodology. Abandoned oil well CH4 emissions in 2019 were estimated to be between 0.9 and 16.5 MMT C02 Eq.,
while abandoned gas well CH4 emissions were estimated to be between 0.2 and 4.3 MMT C02 Eq. at a 95 percent
confidence level. Uncertainty bounds for other years of the time series have not been calculated, but uncertainty is
expected to vary over the time series.
Table 3-102: Approach 2 Quantitative Uncertainty Estimates for ChU and CO2 Emissions from
Petroleum and Natural Gas Systems (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTCOz Eq.)b
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Abandoned Oil Wells
ch4
5.2
0.9
16.5
-83%
+217%
Abandoned Gas Wells
ch4
1.4
0.2
4.3
-83%
+217%
Abandoned Oil Wells
co2
0.004
0.001
0.013
-83%
+217%
Abandoned Gas Wells
co2
0.002
0.0004
0.008
-83%
+217%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for total abandoned oil and gas well CH4 emissions in year 2019.
b All reported values are rounded after calculation. As a result, lower and upper bounds may not be duplicable from other
rounded values as shown in table.
To calculate a time series of emissions for abandoned wells, EPA developed annual activity data for 1990 through
2019 by summing an estimate of total abandoned wells not included in recent databases, to an annual estimate of
abandoned wells in the Enverus data set. As discussed above, the abandoned well population was split into
plugged and unplugged wells by assuming that all abandoned wells were unplugged in 1950, using year-specific
Enverus data to calculate the fraction of abandoned wells plugged in 2016 through 2019, and applying linear
interpolation between the 1950 value and 2016 value to calculate plugged fraction for intermediate years. The
same emission factors were applied to the corresponding categories for each year of the time series.
QA/QC and Verification Discussion
The emission estimates in the Inventory are continually being reviewed and assessed to determine whether
emission factors and activity factors accurately reflect current industry practices. A QA/QC analysis was performed
for data gathering and input, documentation, and calculation. QA/QC checks are consistently conducted to
minimize human error in the model calculations. EPA performs a thorough review of information associated with
new studies to assess whether the assumptions in the Inventory are consistent with industry practices and
whether new data is available that could be considered for updates to the estimates. As in previous years, EPA
conducted early engagement and communication with stakeholders on updates prior to public review. EPA held
stakeholder webinars on greenhouse gas data for oil and gas in September and November of 2020.
Recalculations Discussion
The counts of national abandoned wells were recalculated across the time series to use the latest Enverus data,
which resulted in changes to the total abandoned well population and the allocation between petroleum and
natural gas systems. The changes resulted from changes to the year-specific data for 1990 to 2019 available in the
restructured Enverus data, which led EPA to recalculate the 1975 estimate of historical wells not included in the
Enverus data set (which increased from 1,075,849 to 1,152,211 historical wells not completely included in
Enverus).
Compared with the previous Inventory, counts of abandoned oil and gas wells are on average 1 percent and 12
percent, respectively, higher over 1990 to 2018. Total methane emissions from abandoned wells are around 3
percent higher across the time series than the previous Inventory and CQ2 emissions are around 4 percent higher.
3-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Planned Improvements
EPA will continue to assess new data and stakeholder feedback on considerations (such as disaggregation of the
well population into regions other than Appalachia and non-Appalachia, and emission factor data from regions not
included in the measurement studies on which current emission factors are based) to improve the abandoned well
count estimates and emission factors.
As noted above in the Methodology section, Enverus, a key data source for the calculation of the number of
abandoned wells in the U.S., has restructured its information on the U.S. well population. EPA will seek stakeholder
feedback on how to consider the restructured Enverus dataset for future inventories.
In addition to the wells identified as abandoned through analysis of the Enverus population and included in the
Inventory estimates, for 2019, EPA identified approximately 900,000 wells in the Enverus dataset with only limited
data available for use in determining whether the wells should be included in the abandoned well population.
These wells may be included in some fraction of the estimate of historical wells estimated outside of the Enverus
data set, but the extent is unknown at this time. To develop the national count of abandoned wells in the
inventory for 2019, EPA did not include these approximately 900,000 wells along with the other abandoned wells
from the Enverus data set, due to the limited data, and still relied on the historical estimate to account for old
abandoned wells. Note, including these approximately 900,000 wells would have limited overall impact on the
total count of abandoned wells (including them would simply reduce the historical estimate for the number of
estimated abandoned wells missing from the Enverus data), but would impact the fraction of plugged and
unplugged abandoned wells, as discussed next.
Using the restructured Enverus dataset, for the year 2019, the well status for approximately 400,000 wells in Texas
changed from 'inactive' to 'P&A' (P&A = plugged and abandoned). Applying the same approach to calculating the
fraction of plugged wells as in prior Inventories would have resulted in a large change in plugging status from year
2018 (34 percent plugged) to year 2019 (62 percent plugged). For this year's Inventory, due to lack of clarity on the
900,000 wells noted in the paragraph above and on the 400,000 wells with changed plugging status, EPA
calculated that 41 percent of abandoned wells are plugged for year 2019 by incorporating an assumption that all
historical wells (those outside of the abandoned wells counts developed with Enverus) were not plugged. In other
years of the time series, EPA relies only on the plugging status data available in Enverus and applies the calculated
fraction to all abandoned wells.
EPA will seek stakeholder feedback on other approaches to estimate the total national abandoned well counts and
the plugged abandoned well population, including feedback on how the approximately 900,000 wells with limited
data should be considered.
3.9 Energy Sources of Precursor Greenhouse
Gas Emissions
In addition to the main greenhouse gases addressed above, energy-related activities are also sources of precursor
gases. The reporting requirements of the UNFCCC100 request that information be provided on precursor
greenhouse gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-CH4 volatile organic
compounds (NMVOCs), and sulfur dioxide (S02). These gases are not direct greenhouse gases, but indirectly affect
terrestrial radiation absorption by influencing the formation and destruction of tropospheric and stratospheric
ozone, or, in the case of S02, by affecting the absorptive characteristics of the atmosphere. Additionally, some of
these gases may react with other chemical compounds in the atmosphere to form compounds that are greenhouse
100 See .
Energy 3-115

-------
gases. Total emissions of NOx, CO, and NMVOCs from energy-related activities from 1990 to 2019 are reported in
Table 3-103. Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and Annex 6.3.
Table 3-103: NOx, CO, and NMVOC Emissions from Energy-Related Activities (kt)
Gas/Activity
1990
2005
2015
2016
2017
2018
2019
NOx
21,106
16,602
9,429
8,268
7,928
7,471
7,080
Mobile Fossil Fuel Combustion
10,862
10,295
5,634
4,739
4,563
4,123
3,862
Stationary Fossil Fuel Combustion
10,023
5,858
3,084
2,856
2,728
2,711
2,581
Oil and Gas Activities
139
321
622
594
565
565
565
Waste Combustion
82
128
88
80
71
71
71
International Bunker Fuelsa
1,956
1,704
1,363
1,470
1,481
1,462
1,296
CO
125,640
64,985
38,521
34,461
33,582
32,048
31,208
Mobile Fossil Fuel Combustion
119,360
58,615
32,635
28,789
28,124
26,590
25,749
Stationary Fossil Fuel Combustion
5,000
4,648
3,688
3,690
3,692
3,692
3,692
Waste Combustion
978
1,403
1,576
1,375
1,175
1,175
1,175
Oil and Gas Activities
302
318
622
607
592
592
592
International Bunker Fuelsa
103
133
144
150
156
160
157
NMVOCs
12,620
7,191
6,738
5,941
5,626
5,410
5,304
Mobile Fossil Fuel Combustion
10,932
5,724
3,458
2,873
2,758
2,543
2,437
Oil and Gas Activities
554
510
2,656
2,459
2,262
2,262
2,262
Stationary Fossil Fuel Combustion
912
716
493
489
496
496
496
Waste Combustion
222
241
132
121
109
109
109
International Bunker Fuels0
57
54
47
50
51
51
46
Note: Totals may not sum due to independent rounding.
a These values are presented for informational purposes only and are not included in totals.
Methodology
Emission estimates for 1990 through 2019 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2020), and disaggregated based on EPA (2003). Emissions were
calculated either for individual categories or for many categories combined, using basic activity data (e.g., the
amount of raw material processed) as an indicator of emissions. National activity data were collected for individual
applications from various agencies.
Activity data were used in conjunction with emission factors, which together relate the quantity of emissions to
the activity. Emission factors are generally available from the EPA's Compilation of Air Pollutant Emission Factors,
AP-42 (EPA 1997). The EPA currently derives the overall emission control efficiency of a source category from a
variety of information sources, including published reports, the 1985 National Acid Precipitation and Assessment
Program emissions inventory, and other EPA databases.
Uncertainty and Time-Series Consistency
Uncertainties in these estimates are partly due to the accuracy of the emission factors used and accurate estimates
of activity data. A quantitative uncertainty analysis was not performed.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
3-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
3.10 International Bunker Fuels (CRF Source
Category 1: Memo Items)
Emissions resulting from the combustion of fuels used for international transport activities, termed international
bunker fuels under the UNFCCC, are not included in national emission totals, but are reported separately based
upon location of fuel sales. The decision to report emissions from international bunker fuels separately, instead of
allocating them to a particular country, was made by the Intergovernmental Negotiating Committee in establishing
the Framework Convention on Climate Change.101 These decisions are reflected in the IPCC methodological
guidance, including IPCC (2006), in which countries are requested to report emissions from ships or aircraft that
depart from their ports with fuel purchased within national boundaries and are engaged in international transport
separately from national totals (IPCC 2006).102
Two transport modes are addressed under the IPCC definition of international bunker fuels: aviation and
marine.103 Greenhouse gases emitted from the combustion of international bunker fuels, like other fossil fuels,
include C02, CH4 and N20 for marine transport modes, and C02 and N20 for aviation transport modes. Emissions
from ground transport activities—by road vehicles and trains—even when crossing international borders are
allocated to the country where the fuel was loaded into the vehicle and, therefore, are not counted as bunker fuel
emissions.
The 2006 IPCC Guidelines distinguish between three different modes of air traffic: civil aviation, military aviation,
and general aviation. Civil aviation comprises aircraft used for the commercial transport of passengers and freight,
military aviation comprises aircraft under the control of national armed forces, and general aviation applies to
recreational and small corporate aircraft. The 2006 IPCC Guidelines further define international bunker fuel use
from civil aviation as the fuel combusted for civil (e.g., commercial) aviation purposes by aircraft arriving or
departing on international flight segments. However, as mentioned above, and in keeping with the 2006 IPCC
Guidelines, only the fuel purchased in the United States and used by aircraft taking-off (i.e., departing) from the
United States are reported here. The standard fuel used for civil and military aviation is kerosene-type jet fuel,
while the typical fuel used for general aviation is aviation gasoline.104
Emissions of C02 from aircraft are essentially a function of fuel consumption. Nitrous oxide emissions also depend
upon engine characteristics, flight conditions, and flight phase (i.e., take-off, climb, cruise, decent, and landing).
Recent data suggest that little or no CH4 is emitted by modern engines (Anderson et al. 2011), and as a result, CH4
emissions from this category are reported as zero. In jet engines, N20 is primarily produced by the oxidation of
atmospheric nitrogen, and the majority of emissions occur during the cruise phase.
International marine bunkers comprise emissions from fuels burned by ocean-going ships of all flags that are
engaged in international transport. Ocean-going ships are generally classified as cargo and passenger carrying,
military (i.e., U.S. Navy), fishing, and miscellaneous support ships (e.g., tugboats). For the purpose of estimating
greenhouse gas emissions, international bunker fuels are solely related to cargo and passenger carrying vessels,
which is the largest of the four categories, and military vessels. Two main types of fuels are used on sea-going
101	See report of the Intergovernmental Negotiating Committee for a Framework Convention on Climate Change on the work
of its ninth session, held at Geneva from 7 to 18 February 1994 (A/AC.237/55, annex I, para. lc).
102	Note that the definition of international bunker fuels used by the UNFCCC differs from that used by the International Civil
Aviation Organization.
103	Most emission related international aviation and marine regulations are under the rubric of the International Civil Aviation
Organization (ICAO) or the International Maritime Organization (IMO), which develop international codes, recommendations,
and conventions, such as the International Convention of the Prevention of Pollution from Ships (MARPOL).
104	Naphtha-type jet fuel was used in the past by the military in turbojet and turboprop aircraft engines.
Energy 3-117

-------
vessels: distillate diesel fuel and residual fuel oil. Carbon dioxide is the primary greenhouse gas emitted from
marine shipping.
Overall, aggregate greenhouse gas emissions in 2019 from the combustion of international bunker fuels from both
aviation and marine activities were 117.2 MMT C02 Eq., or 12.1 percent above emissions in 1990 (see Table 3-104
and Table 3-105). Emissions from international flights and international shipping voyages departing from the
United States have increased by 112.2 percent and decreased by 46.0 percent, respectively, since 1990. The
majority of these emissions were in the form of C02; however, small amounts of CH4 (from marine transport
modes) and N20 were also emitted.
Table 3-104: CO2, ChU, and N2O Emissions from International Bunker Fuels (MMT CO2 Eq.)
Gas/Mode
1990
2005
2015
2016
2017
2018
2019
CO?
103.5
113.2
110.9
116.6
120.1
122.1
116.1
Aviation
38.0
60.1
71.9
74.1
77.7
80.8
80.7
Commercial
30.0
55.6
68.6
70.8
74.5
77.7
77.6
Military
8.1
4.5
3.3
3.3
3.2
3.1
3.1
Marine
65.4
53.1
39.0
42.6
42.4
41.3
35.4
ch4
0.2
0.1
0.1
0.1
0.1
0.1
0.1
Aviation3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Marine
0.2
0.1
0.1
0.1
0.1
0.1
0.1
n2o
0.9
1.0
1.0
1.0
1.1
1.1
1.0
Aviation
0.4
0.6
0.7
0.7
0.7
0.8
0.8
Marine
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Total
104.5
114.3
112.0
117.7
121.3
123.3
117.2
Notes: Totals may not sum due to independent rounding. Includes aircraft cruise altitude emissions.

a CH4 emissions from aviation are estimated to be zero





ible 3-105: CO2, ChU, and N2O Emissions from International Bunker Fuels (kt)

Gas/Mode
1990
2005
2015
2016
2017
2018
2019
CO?
103,463
113,232
110,908
116,611
120,121
122,112
116,064
Aviation
38,034
60,125
71,942
74,059
77,696
80,788
80,714
Marine
65,429
53,107
38,967
42,552
42,425
41,324
35,350
ch4
7
5
4
4
4
4
4
Aviation3
0
0
0
0
0
0
0
Marine
7
5
4
4
4
4
4
n2o
3
3
3
3
4
4
3
Aviation
1
2
2
2
3
3
3
Marine
2
1
1
1
1
1
1
Notes: Totals may not sum due to independent rounding. Includes aircraft cruise altitude emissions.
a CH4 emissions from aviation are estimated to be zero.
Methodology
Emissions of C02 were estimated by applying C content and fraction oxidized factors to fuel consumption activity
data. This approach is analogous to that described under Section 3 - C02 from Fossil Fuel Combustion. Carbon
content and fraction oxidized factors for jet fuel, distillate fuel oil, and residual fuel oil are the same as used for C02
from Fossil Fuel Combustion and are presented in Annex 2.1, Annex 2.2, and Annex 3.8 of this Inventory. Density
conversions were taken from Chevron (2000), ASTM (1989), and USAF (1998). Heat content for distillate fuel oil
and residual fuel oil were taken from EIA (2020) and USAF (1998), and heat content for jet fuel was taken from EIA
(2020).
3-118 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
A complete description of the methodology and a listing of the various factors employed can be found in Annex
2.1. See Annex 3.8 for a specific discussion on the methodology used for estimating emissions from international
bunker fuel use by the U.S. military.
Emission estimates for CH4 and N20 were calculated by multiplying emission factors by measures of fuel
consumption by fuel type and mode. Emission factors used in the calculations of CH4 and N20 emissions were
obtained from the Revised 1996IPCC Guidelines (IPCC/UNEP/OECD/IEA 1997), which is also referenced in the 2006
IPCC Guidelines (IPCC 2006). For aircraft emissions, the following value, in units of grams of pollutant per kilogram
of fuel consumed (g/kg), was employed: 0.1 for N20 (IPCC 2006). For marine vessels consuming either distillate
diesel or residual fuel oil the following values (g/MJ), were employed: 0.315 for CH4 and 0.08 for N20. Activity data
for aviation included solely jet fuel consumption statistics, while the marine mode included both distillate diesel
and residual fuel oil.
Activity data on domestic and international aircraft fuel consumption were developed by the U.S. Federal Aviation
Administration (FAA) using radar-informed data from the FAA Enhanced Traffic Management System (ETMS) for
1990 and 2000 through 2019 as modeled with the Aviation Environmental Design Tool (AEDT). This bottom-up
approach is built from modeling dynamic aircraft performance for each flight occurring within an individual
calendar year. The analysis incorporates data on the aircraft type, date, flight identifier, departure time, arrival
time, departure airport, arrival airport, ground delay at each airport, and real-world flight trajectories. To generate
results for a given flight within AEDT, the radar-informed aircraft data is correlated with engine and aircraft
performance data to calculate fuel burn and exhaust emissions. Information on exhaust emissions for in-
production aircraft engines comes from the International Civil Aviation Organization (ICAO) Aircraft Engine
Emissions Databank (EDB). This bottom-up approach is in accordance with the Tier 3B method from the 2006 IPCC
Guidelines (IPCC 2006).
International aviation C02 estimates for 1990 and 2000 through 2019 were obtained directly from FAA's AEDT
model (FAA 2021). The radar-informed method that was used to estimate C02 emissions for commercial aircraft
for 1990 and 2000 through 2019 was not possible for 1991 through 1999 because the radar dataset was not
available for years prior to 2000. FAA developed Official Airline Guide (OAG) schedule-informed inventories
modeled with AEDT and great circle trajectories for 1990, 2000, and 2010. Because fuel consumption and C02
emission estimates for years 1991 through 1999 are unavailable, consumption estimates for these years were
calculated using fuel consumption estimates from the Bureau of Transportation Statistics (DOT 1991 through
2013), adjusted based on 2000 through 2005 data. See Annex 3.3 for more information on the methodology for
estimating emissions from commercial aircraft jet fuel consumption.
Data on U.S. Department of Defense (DoD) aviation bunker fuels and total jet fuel consumed by the U.S. military
was supplied by the Office of the Undersecretary of Defense (Installations and Environment), DoD. Estimates of
the percentage of each Service's total operations that were international operations were developed by DoD.
Military aviation bunkers included international operations, operations conducted from naval vessels at sea, and
operations conducted from U.S. installations principally over international water in direct support of military
operations at sea. Military aviation bunker fuel emissions were estimated using military fuel and operations data
synthesized from unpublished data from DoD's Defense Logistics Agency Energy (DLA Energy 2020). Together, the
data allow the quantity of fuel used in military international operations to be estimated. Densities for each jet fuel
type were obtained from a report from the U.S. Air Force (USAF 1998). Final jet fuel consumption estimates are
presented in Table 3-106. See Annex 3.8 for additional discussion of military data.
Energy 3-119

-------
Table 3-106: Aviation Jet Fuel Consumption for International Transport (Million Gallons)
Nationality
1990
2005
2015
2016
2017
2018
2019
U.S. and Foreign Carriers
3,222
5,983
7,383
7,610
8,011
8,352
8,344
U.S. Military
862
462
341
333
326
315
318
Total
4,084
6,445
7,725
7,943
8,338
8,667
8,662
Note: Totals may not sum due to independent rounding.
In order to quantify the civilian international component of marine bunker fuels, activity data on distillate diesel
and residual fuel oil consumption by cargo or passenger carrying marine vessels departing from U.S. ports were
collected for individual shipping agents on a monthly basis by the U.S. Customs and Border Protection. This
information was then reported in unpublished data collected by the Foreign Trade Division of the U.S. Department
of Commerce's Bureau of the Census (DOC 1991 through 2020) for 1990 through 2001, 2007 through 2019, and
the Department of Homeland Security's Bunker Report for 2003 through 2006 (DHS 2008). Fuel consumption data
for 2002 was interpolated due to inconsistencies in reported fuel consumption data. Activity data on distillate
diesel consumption by military vessels departing from U.S. ports were provided by DLA Energy (2020). The total
amount of fuel provided to naval vessels was reduced by 21 percent to account for fuel used while the vessels
were not-underway (i.e., in port). Data on the percentage of steaming hours underway versus not underway were
provided by the U.S. Navy. These fuel consumption estimates are presented in Table 3-107.
Table 3-107: Marine Fuel Consumption for International Transport (Million Gallons)
Fuel Type
1990
2005
2015
2016
2017
2018
2019
Residual Fuel Oil
4,781
3,881
2,718
3,011
2,975
2,790
2,246
Distillate Diesel Fuel & Other
617
444
492
534
568
684
702
U.S. Military Naval Fuels
522
471
326
314
307
285
281
Total
5,920
4,796
3,536
3,858
3,850
3,759
3,229
Note: Totals may not sum due to independent rounding.
Uncertainty and Time-Series Consistency
Emission estimates related to the consumption of international bunker fuels are subject to the same uncertainties
as those from domestic aviation and marine mobile combustion emissions; however, additional uncertainties
result from the difficulty in collecting accurate fuel consumption activity data for international transport activities
separate from domestic transport activities.105 For example, smaller aircraft on shorter routes often carry
sufficient fuel to complete several flight segments without refueling in order to minimize time spent at the airport
gate or take advantage of lower fuel prices at particular airports. This practice, called tankering, when done on
international flights, complicates the use of fuel sales data for estimating bunker fuel emissions. Tankering is less
common with the type of large, long-range aircraft that make many international flights from the United States,
however. Similar practices occur in the marine shipping industry where fuel costs represent a significant portion of
overall operating costs and fuel prices vary from port to port, leading to some tankering from ports with low fuel
costs.
Uncertainties exist with regard to the total fuel used by military aircraft and ships. Total aircraft and ship fuel use
estimates were developed from DoD records, which document fuel sold to the DoD Components (e.g., Army,
Department of Navy and Air Force) from the Defense Logistics Agency Energy. These data may not include fuel
used in aircraft and ships as a result of a Service procuring fuel from, selling fuel to, trading fuel with, or giving fuel
to other ships, aircraft, governments, or other entities.
Additionally, there are uncertainties in historical aircraft operations and training activity data. Estimates for the
quantity of fuel actually used in Navy and Air Force flying activities reported as bunker fuel emissions had to be
105 See uncertainty discussions under section 3.1 Carbon Dioxide Emissions from Fossil Fuel Combustion.
3-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
estimated based on a combination of available data and expert judgment. Estimates of marine bunker fuel
emissions were based on Navy vessel steaming hour data, which reports fuel used while underway and fuel used
while not underway. This approach does not capture some voyages that would be classified as domestic for a
commercial vessel. Conversely, emissions from fuel used while not underway preceding an international voyage
are reported as domestic rather than international as would be done for a commercial vessel. There is uncertainty
associated with ground fuel estimates for 1997 through 2019, including estimates for the quantity of jet fuel
allocated to ground transportation. Small fuel quantities may have been used in vehicles or equipment other than
that which was assumed for each fuel type.
There are also uncertainties in fuel end-uses by fuel type, emissions factors, fuel densities, diesel fuel sulfur
content, aircraft and vessel engine characteristics and fuel efficiencies, and the methodology used to back-
calculate the data set to 1990 using the original set from 1995. The data were adjusted for trends in fuel use based
on a closely correlating, but not matching, data set. All assumptions used to develop the estimate were based on
process knowledge, DoD data, and expert judgments. The magnitude of the potential errors related to the various
uncertainties has not been calculated but is believed to be small. The uncertainties associated with future military
bunker fuel emission estimates could be reduced through revalidation of assumptions based on data regarding
current equipment and operational tempo, however, it is doubtful data with more fidelity exist at this time.
Although aggregate fuel consumption data have been used to estimate emissions from aviation, the recommended
method for estimating emissions of gases other than C02 in the 2006IPCC Guidelines (IPCC 2006) is to use data by
specific aircraft type, number of individual flights and, ideally, movement data to better differentiate between
domestic and international aviation and to facilitate estimating the effects of changes in technologies. The IPCC
also recommends that cruise altitude emissions be estimated separately using fuel consumption data, while
landing and take-off (LTO) cycle data be used to estimate near-ground level emissions of gases other than C02.106
There is also concern regarding the reliability of the existing DOC (1991 through 2019) data on marine vessel fuel
consumption reported at U.S. customs stations due to the significant degree of inter-annual variation.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
In order to ensure the quality of the emission estimates from international bunker fuels, General (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved
checks specifically focusing on the activity data and emission factor sources and methodology used for estimating
C02, CH4, and N20 emissions from international bunker fuels in the United States. Emission totals for the different
sectors and fuels were compared and trends were investigated. No corrective actions were necessary.
106 U.S. aviation emission estimates for CO, NOx, and NMVOCs are reported by EPA's National Emission Inventory (NEI) Air
Pollutant Emission Trends website, and reported under the Mobile Combustion section. It should be noted that these estimates
are based solely upon LTO cycles and consequently only capture near ground-level emissions, which are more relevant for air
quality evaluations. These estimates also include both domestic and international flights. Therefore, estimates reported under
the Mobile Combustion section overestimate IPCC-defined domestic CO, NOx, and NMVOC emissions by including landing and
take-off (LTO) cycles by aircraft on international flights, but underestimate because they do not include emissions from aircraft
on domestic flight segments at cruising altitudes.
Energy 3-121

-------
Recalculations Discussion
EPA revised distillate fuel oil carbon contents, which affect marine distillate fuel oil consumption (EPA 2020).
Revisions resulted in an average annual increase of less than 0.05 MMT C02 Eq. in emissions from marine residual
and distillate fuel oil.
Planned Improvements
A longer-term effort is underway to consider the feasibility of including data from a broader range of domestic and
international sources for bunker fuels. Potential sources include the International Maritime Organization (IMO)
and their ongoing greenhouse gas analysis work, data from the U.S. Coast Guard on vehicle operation currently
used in criteria pollutant modeling, and data from the International Energy Agency.
3.11 Wood Biomass and Biofuels
Consumption (CRF Source Category 1A)
The combustion of biomass fuels—such as wood, charcoal, and wood waste and biomass-based fuels such as
ethanol, biogas, and biodiesel—generates C02 in addition to CH4 and N20 already covered in this chapter. In line
with the reporting requirements for inventories submitted under the UNFCCC, C02 emissions from biomass
combustion have been estimated separately from fossil fuel C02 emissions and are not directly included in the
energy sector contributions to U.S. totals. In accordance with IPCC methodological guidelines, any such emissions
are calculated by accounting for net carbon fluxes from changes in biogenic C reservoirs in wooded or crop lands.
For a more complete description of this methodological approach, see the Land Use, Land-Use Change, and
Forestry chapter (Chapter 6), which accounts for the contribution of any resulting C02 emissions to U.S. totals
within the Land Use, Land-Use Change, and Forestry sector's approach.
Therefore, C02 emissions from wood biomass and biofuel consumption are not included specifically in summing
energy sector totals. However, they are presented here for informational purposes and to provide detail on wood
biomass and biofuels consumption.
In 2019, total C02 emissions from the burning of woody biomass in the industrial, residential, commercial, and
electric power sectors were approximately 216.5 MMT C02 Eq. (216,533 kt) (see Table 3-108 and Table 3-109). As
the largest consumer of woody biomass, the industrial sector was responsible for 61.3 percent of the C02
emissions from this source. The residential sector was the second largest emitter, constituting 25.2 percent of the
total, while the electric power and commercial sectors accounted for the remainder.
Table 3-108: CO2 Emissions from Wood Consumption by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Industrial
135.3
136.3
138.5
138.3
135.4
135.0
132.6
Residential
59.8
44.3
52.9
45.6
43.8
53.3
54.5
Commercial
6.8
7.2
8.2
8.6
8.6
8.7
8.7
Electric Power
13.3
19.1
25.1
23.1
23.6
22.8
20.7
Total
215.2
206.9
224.7
215.7
211.5
219.8
216.5
ible 3-109: CO2 Emissions from Wood Consumption by End-Use Sector (kt)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Industrial
135,348
136,269
138,537
138,339
135,386
134,983
132,635
Residential
59,808
44,340
52,872
45,598
43,844
53,346
54,528
Commercial
6,779
7,218
8,176
8,635
8,634
8,669
8,693
3-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Electric Power	13,252	19,074	25,146 23,140 23,647 22,795 20,677
Total	215,186	206,901	224,730 215,712 211,511 219,794 216,533
Note: Totals may not sum due to independent rounding.
The transportation sector is responsible for most of the fuel ethanol consumption in the United States. Ethanol
used for fuel is currently produced primarily from corn grown in the Midwest, but it can be produced from a
variety of biomass feedstocks. Most ethanol for transportation use is blended with gasoline to create a 90 percent
gasoline, 10 percent by volume ethanol blend known as E-10 or gasohol.
In 2019, the United States transportation sector consumed an estimated 1,150.2 trillion Btu of ethanol (95 percent
of total), and as a result, produced approximately 78.7 MMT C02 Eq. (78,739 kt) (see Table 3-110 and Table 3-111)
of C02 emissions. Smaller quantities of ethanol were also used in the industrial and commercial sectors. Ethanol
fuel production and consumption has grown significantly since 1990 due to the favorable economics of blending
ethanol into gasoline and federal policies that have encouraged use of renewable fuels.
Table 3-110: CO2 Emissions from Ethanol Consumption (MMT CO2 Eq.)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation3
4.1
21.6
74.2
76.9
77.7
78.6
78.7
Industrial
0.1
1.2
1.9
1.8
1.9
1.4
1.6
Commercial
0.1
0.2
2.8
2.6
2.5
1.9
2.2
Total
4.2
22.9
78.9
81.2
82.1
81.9
82.6
Note: Totals may not sum due to independent rounding.
a See Annex 3.2, Table A-81 for additional information on transportation consumption of these fuels.
Table 3-111: CO2 Emissions from Ethanol Consumption (kt)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation3
4,059
21,616
74,187
76,903
77,671
78,603
78,739
Industrial
105
1,176
1,931
1,789
1,868
1,404
1,627
Commercial
63
151
2,816
2,558
2,550
1,910
2,212
Total
4,227
22,943
78,934
81,250
82,088
81,917
82,578
Note: Totals may not sum due to independent rounding.
a See Annex 3.2, Table A-81 for additional information on transportation consumption of these fuels.
The transportation sector is assumed to be responsible for all of the biodiesel consumption in the United States
(EIA 2020). Biodiesel is currently produced primarily from soybean oil, but it can be produced from a variety of
biomass feedstocks including waste oils, fats, and greases. Biodiesel for transportation use appears in low-level
blends (less than 5 percent) with diesel fuel, high-level blends (between 6 and 20 percent) with diesel fuel, and 100
percent biodiesel (EIA 2020b).
In 2019, the United States consumed an estimated 231.3 trillion Btu of biodiesel, and as a result, produced
approximately 17.1 MMT C02 Eq. (17,080 kt) (see Table 3-112 and Table 3-113) of C02 emissions. Biodiesel
production and consumption has grown significantly since 2001 due to the favorable economics of blending
biodiesel into diesel and federal policies that have encouraged use of renewable fuels (EIA 2020b). There was no
measured biodiesel consumption prior to 2001 EIA (2020).
Table 3-112: CO2 Emissions from Biodiesel Consumption (MMT CO2 Eq.)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation3
NO
0.9
14.1
19.6
18.7
17.9
17.1
Total
NO
0.9
14.1
19.6
18.7
17.9
17.1
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
a See Annex 3.2, Table A-81 for additional information on transportation consumption of these fuels.
Energy 3-123

-------
Table 3-113: CO2 Emissions from Biodiesel Consumption (kt)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation3
NO
856
14,077
19,648
18,705
17,936
17,080
Total
NO
856
14,077
19,648
18,705
17,936
17,080
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
a See Annex 3.2, Table A-81 for additional information on transportation consumption of these fuels.
Methodology
Woody biomass emissions were estimated by applying two gross heat contents from EIA (Lindstrom 2006) to U.S.
consumption data (EIA 2020) (see Table 3-114), provided in energy units for the industrial, residential, commercial,
and electric power sectors. One heat content (16.95 MMBtu/MT wood and wood waste) was applied to the
industrial sector's consumption, while the other heat content (15.43 MMBtu/MT wood and wood waste) was
applied to the consumption data for the other sectors. An EIA emission factor of 0.434 MT C/MT wood (Lindstrom
2006) was then applied to the resulting quantities of woody biomass to obtain C02 emission estimates. The woody
biomass is assumed to contain black liquor and other wood wastes, have a moisture content of 12 percent, and
undergo complete combustion to be converted into C02.
The amount of ethanol allocated across the transportation, industrial, and commercial sectors was based on the
sector allocations of ethanol-blended motor gasoline. The sector allocations of ethanol-blended motor gasoline
were determined using a bottom-up analysis conducted by EPA, as described in the Methodology section of Fossil
Fuel Combustion. Total U.S. ethanol consumption from EIA (2020) was allocated to individual sectors using the
same sector allocations as ethanol-blended motor gasoline. The emissions from ethanol consumption were
calculated by applying an emission factor of 18.67 MMT C/Qbtu (EPA 2010) to adjusted ethanol consumption
estimates (see Table 3-115). The emissions from biodiesel consumption were calculated by applying an emission
factor of 20.1 MMT C/Qbtu (EPA 2010) to U.S. biodiesel consumption estimates that were provided in energy units
(EIA 2020) (see Table 3-116).107
Table 3-114: Woody Biomass Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Industrial
1,441.9
1,451.7
1,475.9
1,473.8
1,442.3
1,438.0
1,413.0
Residential
580.0
430.0
512.7
442.2
425.2
517.3
528.8
Commercial
65.7
70.0
79.3
83.7
83.7
84.1
84.3
Electric Power
128.5
185.0
243.9
224.4
229.3
221.1
200.5
Total
2,216.2
2,136.7
2,311.8
2,224.1
2,180.6
2,260.5
2,226.6
Note: Totals may not sum due to independent rounding.





ble 3-115: Ethanol
Consumption by Sector
(Trillion
Btu)



End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation
59.3
315.8
1,083.7
1,123.4
1,134.6
1,148.2
1,150.2
Industrial
1.5
17.2
28.2
26.1
27.3
20.5
23.8
Commercial
0.9
2.2
41.1
37.4
37.2
27.9
32.3
Total
61.7
335.1
1,153.1
1,186.9
1,199.1
1,196.6
1,206.3
Note: Totals may not sum due to independent rounding.
107 C02 emissions from biodiesel do not include emissions associated with the C in the fuel that is from the methanol used in
the process. Emissions from methanol use and combustion are assumed to be accounted for under Non-Energy Use of Fuels.
See Annex 2.3 - Methodology for Estimating Carbon Emitted from Non-Energy Uses of Fossil Fuels.
3-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 3-116: Biodiesel Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2015
2016
2017
2018
2019
Transportation
NO
11.6
190.6
266.1
253.3
242.9
231.3
Total
NO
11.6
190.6
266.1
253.3
242.9
231.3
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
Uncertainty and Time-Series Consistency
It is assumed that the combustion efficiency for woody biomass is 100 percent, which is believed to be an
overestimate of the efficiency of wood combustion processes in the United States. Decreasing the combustion
efficiency would decrease emission estimates for C02. Additionally, the heat content applied to the consumption
of woody biomass in the residential, commercial, and electric power sectors is unlikely to be a completely accurate
representation of the heat content for all the different types of woody biomass consumed within these sectors.
Emission estimates from ethanol and biodiesel production are more certain than estimates from woody biomass
consumption due to better activity data collection methods and uniform combustion techniques.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
Recalculations Discussion
EIA (2020) revised approximate heat rates for electricity and the heat content of electricity for noncombustible
renewable energy, which impacted wood energy consumption by the industrial sector from 2016 through 2018.
Revisions to biomass consumption resulted in an average annual decrease of 0.7 MMT C02 Eq. (0.3 percent).
Planned Improvements
Future research will investigate the availability of data on woody biomass heat contents and carbon emission
factors the see if there are newer, improved data sources available for these factors.
The availability of facility-level combustion emissions through EPA's GHGRP will be examined to help better
characterize the industrial sector's energy consumption in the United States, and further classify woody biomass
consumption by business establishments according to industrial economic activity type. Most methodologies used
in EPA's GHGRP are consistent with IPCC, though for EPA's GHGRP, facilities collect detailed information specific to
their operations according to detailed measurement standards, which may differ with the more aggregated data
collected for the Inventory to estimate total, national U.S. emissions. In addition, and unlike the reporting
requirements for this chapter under the UNFCCC reporting guidelines, some facility-level fuel combustion
emissions reported under EPA's GHGRP may also include industrial process emissions.108
In line with UNFCCC reporting guidelines, fuel combustion emissions are included in this chapter, while process
emissions are included in the Industrial Processes and Product Use chapter of this report. In examining data from
EPA's GHGRP that would be useful to improve the emission estimates for the C02 from biomass combustion
category, particular attention will also be made to ensure time series consistency, as the facility-level reporting
data from EPA's GHGRP are not available for all inventory years as reported in this Inventory. Additionally, analyses
will focus on aligning reported facility-level fuel types and IPCC fuel types per the national energy statistics,
ensuring C02 emissions from biomass are separated in the facility-level reported data, and maintaining consistency
with national energy statistics provided by EIA. In implementing improvements and integration of data from EPA's
108 See .
Energy 3-125

-------
GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories will be relied
upon.109
Currently emission estimates from biomass and biomass-based fuels included in this Inventory are limited to
woody biomass, ethanol, and biodiesel. Additional forms of biomass-based fuel consumption include biogas, the
biogenic components of MSW, and other renewable diesel fuels. EPA will examine EIA data on biogas and other
renewable diesel fuels to see if it can be included in future inventories. EIA (2020) natural gas data already deducts
biogas used in the natural gas supply, so no adjustments are needed to the natural gas fuel consumption data to
account for biogas. Distillate fuel statistics are adjusted in this Inventory to remove other renewable diesel fuels as
well as biodiesel. Sources of estimates for the biogenic fraction of MSW will be examined, including the GHGRP,
EIA data, and EPA MSW characterization data.
Carbon dioxide emissions from biomass used in the electric power sector are calculated using woody biomass
consumption data from ElA's Monthly Energy Review (EIA 2020a), whereas non-C02 biomass emissions from the
electric power sector are estimated by applying technology and fuel use data from EPA's Clean Air Market Acid
Rain Program dataset (EPA 2021) to fuel consumption data from EIA (2020a). There were significant discrepancies
identified between the EIA woody biomass consumption data and the consumption data estimated using EPA's
Acid Rain Program dataset (see the Methodology section for CH4 and N20 from Stationary Combustion). EPA will
continue to investigate this discrepancy in order to apply a consistent approach to both C02 and non-C02 emission
calculations for woody biomass consumption in the electric power sector.
109 See .
3-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
4. Industrial Processes and Product Use
The Industrial Processes and Product Use (IPPU) chapter includes greenhouse gas emissions occurring from
industrial processes and from the use of greenhouse gases in products. The industrial processes and product use
categories included in this chapter are presented in Figure 4-1 and Figure 4-2. Greenhouse gas emissions from
industrial processes can occur in two different ways. First, they may be generated and emitted as the byproducts
of various non-energy-related industrial activities. Second, they may be emitted due to their use in manufacturing
processes or by end-consumers. Combustion-related energy use emissions from industry are reported in Chapter
3, Energy.
In the case of byproduct emissions, the emissions are generated by an industrial process itself and are not directly
a result of energy consumed during the process. For example, raw materials can be chemically or physically
transformed from one state to another. This transformation can result in the release of greenhouse gases such as
carbon dioxide (C02), methane (CH4), nitrous oxide (N20), and fluorinated greenhouse gases (e.g., HFC-23). The
greenhouse gas byproduct generating processes included in this chapter include iron and steel production and
metallurgical coke production, cement production, petrochemical production, lime production, ammonia
production, nitric acid production, other process uses of carbonates (e.g., flux stone, flue gas desulfurization, and
glass manufacturing), urea consumption for non-agricultural purposes, adipic acid production, HCFC-22
production, aluminum production, soda ash production and use, ferroalloy production, titanium dioxide
production, caprolactam production, glass production, zinc production, phosphoric acid production, lead
production, and silicon carbide production and consumption.
Greenhouse gases that are used in manufacturing processes or by end-consumers include man-made compounds
such as hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SF6), and nitrogen trifluoride
(NF3). The present contribution of HFCs, PFCs, SF6, and NF3 gases to the radiative forcing effect of all anthropogenic
greenhouse gases is small; however, because of their extremely long lifetimes, many of them will continue to
persist in the atmosphere long after they were first released. In addition, many of these gases have high global
warming potentials; SF6 is the most potent greenhouse gas the Intergovernmental Panel on Climate Change (IPCC)
has evaluated. Use of HFCs is growing rapidly since they are the primary substitutes for ozone depleting substances
(ODS), which are being phased-out under the Montreal Protocol on Substances that Deplete the Ozone Layer.
Hydrofluorocarbons, PFCs, SF6, and NF3 are employed and emitted by a number of other industrial sources in the
United States, such as semiconductor manufacture, electric power transmission and distribution, and magnesium
metal production and processing. Carbon dioxide is also consumed and emitted through various end-use
applications. In addition, nitrous oxide is used in and emitted by semiconductor manufacturing and anesthetic and
aerosol applications.
In 2019, IPPU generated emissions of 373.7 million metric tons of C02 equivalent (MMT C02 Eq.), or 5.7 percent of
total U.S. greenhouse gas emissions.1 Carbon dioxide emissions from all industrial processes were 166.6 MMT C02
1 Emissions reported in the IPPU chapter include those from all 50 states, including Hawaii and Alaska, as well as from U.S.
Territories to the extent of which industries are occurring.
Industrial Processes and Product Use 4-1

-------
Eq. (166,589 kt C02) in 2019, or 3.2 percent of total U.S. C02 emissions. Methane emissions from industrial
processes resulted in emissions of approximately 0.4 MMT C02 Eq. (15 kt CH4) in 2019, which was less than 1
percent of U.S. CH4 emissions. Nitrous oxide emissions from IPPU were 21.1 MMT C02 Eq. (71 kt N20) in 2019, or
4.6 percent of total U.S. N20 emissions. In 2019 combined emissions of HFCs, PFCs, SF6, and NF3 totaled 185.7
MMT C02 Eq. Total emissions from IPPU in 2019 were 8.1 percent more than 1990 emissions. Indirect greenhouse
gas emissions also result from IPPU and are presented in Table 4-112 in kilotons (kt).
Figure 4-1: 2019 Industrial Processes and Product Use Chapter Greenhouse Gas Sources
Substitution of Ozone Depleting Substances
Iron and Steel Production & Metallurgical Coke Production
Cement Production
Petrochemical Production
Ammonia Production
Lime Production
Nitric Acid Production
Other Process Uses of Carbonates
Urea Consumption for Non-Agricultural Purposes
Adipic Acid Production
Carbon Dioxide Consumption
Electronics Industry
Electrical Transmission and Distribution
N2O from Product Uses
HCFC-22 Production
Aluminum Production
Soda Ash Production
Ferroalloy Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Glass Production
Zinc Production
Magnesium Production and Processing
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
171
Industrial Processes and Product Use
as a Portion of All Emissions
Energy
I Agriculture
IPPU
Waste
< 0.5
10
20
30 40
MMT COz Eq.
50
60
70
4-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 4-2: Trends in Industrial Processes and Product Use Chapter Greenhouse Gas Sources
400
350
300
. 250
O

CTi CTi CTi (Ti  cn cri cn cn cr> cr» cri en cn
iNirMr\ir\itN«NirMr\|
The increase in overall IPPU emissions since 1990 reflects a range of emission trends among the emission sources,
as shown in Figure 4-2. Emissions resulting from most types of metal production have declined significantly since
1990, largely due to production shifting to other countries, but also due to transitions to less-emissive methods of
production (in the case of iron and steel) and to improved practices (in the case of PFC emissions from aluminum
production). Carbon dioxide and CH4 emissions from many chemical production sources have either decreased or
not changed significantly since 1990, with the exception of petrochemical production, Carbon Dioxide
Consumption, and Urea Consumption for Non-Agricultural Purposes which has steadily increased. Emissions from
mineral sources have either increased (e.g., Cement Production) or not changed significantly (e.g., Glass and Lime
Production) since 1990 but largely follow economic cycles. Hydrofluorocarbon emissions from the substitution of
ODS have increased drastically since 1990 and are the largest source of IPPU emissions (45.6 percent in 2019),
while the emissions of HFCs, PFCs, SF6, and NF3 from other sources have generally declined. Nitrous oxide
emissions from the production of adipic and nitric acid have decreased, while N20 emissions from product uses
have remained nearly constant over time. Some emission sources exhibit varied interannual trends. Trends are
explained further within each emission source category throughout the chapter. Table 4-1 summarizes emissions
for the IPPU chapter in MMT C02 Eq. using IPCC Fourth Assessment Report (AR4) GWP values, following the
requirements of the current United Nations Framework Convention on Climate Change (UNFCCC) reporting
guidelines for national inventories (IPCC 2007).2 Unweighted native gas emissions in kt are also provided in Table
4-2. The source descriptions that follow in the chapter are presented in the order as reported to the UNFCCC in the
Common Reporting Format (CRF) tables, corresponding generally to: mineral products, chemical production, metal
production, and emissions from the uses of HFCs, PFCs, SF6, and NF3.
Each year, some emission and sink estimates in the IPPU sector of the Inventory are recalculated and revised with
improved methods and/or data. In general, recalculations are made to the U.S. greenhouse gas emission estimates
either to incorporate new methodologies or, most commonly, to update recent historical data. These
improvements are implemented consistently across the previous Inventory's time series (i.e., 1990 to 2018) to
2 See .
Industrial Processes and Product Use 4-3

-------
ensure that the trend is accurate. This year's estimates of HFC emissions from use of Ozone Depleting Substances
Substitutes reflect updates to market size, substitute transitions, and charge size assumptions for Metered Dose
Inhalers (MDI) aerosols to align with stakeholder input and market research. Market transitions for the ice maker
end-use were updated based on manufacturer information on refrigerant use. In addition, several updates to the
foam sector were implemented. The commercial refrigeration foam end-use was replaced with ten discrete
commercial refrigeration application end-uses, in order to better define a market that was not adequately
encompassed by the current commercial refrigeration foam end-use. Within the domestic refrigerator foam end-
use, manufacturing emissions were adjusted to only include equipment manufactured within the United States,
including those that are produced for export, and excluding those that are imported with foam. Market size,
manufacturing loss rate, disposal lost rate, and post-life emission rate assumptions were also updated for PU and
PIR boardstock foams based on market research. Carbon content factors were also updated for the Iron and Steel
emissions calculations. Finally, the methods to estimate the C02 emission factors to recalculate emissions for
earlier parts of the time series (i.e., 1990 to 2009) for petrochemical subcategories ethylene, ethylene dichloride
and vinyl chloride monomer, and carbon black were updated to reflect GHGRP data updates. Together, these
updates decreased greenhouse gas emissions an average of 0.081 MMT C02 Eq. (0.04 percent) across the time
series.
Table 4-1: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
CO?
212.3
194.1
173.5
165.3
163.9
164.3
166.6
Iron and Steel Production &







Metallurgical Coke Production
104.7
70.1
47.9
43.6
40.6
42.6
41.3
Iron and Steel Production
99.1
66.2
43.5
41.0
38.6
41.3
39.9
Metallurgical Coke Production
5.6
3.9
4.4
2.6
2.0
1.3
1.4
Cement Production
33.5
46.2
39.9
39.4
40.3
39.0
40.9
Petrochemical Production
21.6
27.4
28.1
28.3
28.9
29.3
30.8
Ammonia Production
13.0
9.2
10.6
10.2
11.1
12.2
12.3
Lime Production
11.7
14.6
13.3
12.6
12.9
13.1
12.1
Other Process Uses of Carbonates
6.3
7.6
12.2
11.0
9.9
7.5
7.5
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
4.6
5.1
5.0
5.9
6.2
Carbon Dioxide Consumption
1.5
1.4
4.9
4.6
4.6
4.1
4.9
Aluminum Production
6.8
4.1
2.8
1.3
1.2
1.5
1.9
Soda Ash Production
1.4
1.7
1.7
1.7
1.8
1.7
1.8
Ferroalloy Production
2.2
1.4
2.0
1.8
2.0
2.1
1.6
Titanium Dioxide Production
1.2
1.8
1.6
1.7
1.7
1.5
1.5
Glass Production
1.5
1.9
1.3
1.2
1.3
1.3
1.3
Zinc Production
0.6
1.0
0.9
0.8
0.9
1.0
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
0.9
0.9
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Carbide Production and







Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Magnesium Production and







Processing
+
+
+
+
+
+
+
:h4
0.3
0.1
0.2
0.3
0.3
0.3
0.4
Petrochemical Production
0.2
0.1
0.2
0.2
0.3
0.3
0.3
Ferroalloy Production
+
+
+
+
+
+
+
Carbide Production and







Consumption
+
+
+
+
+
+
+
Iron and Steel Production &







Metallurgical Coke Production
+
+
+
+
+
+
+
m2o
33.3
24.9
22.2
23.3
22.7
25.8
21.1
Nitric Acid Production
12.1
11.3
11.6
10.1
9.3
9.6
10.0
4-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
AdipicAcid Production
15.2
7.1
4.3
7.0
7.4
10.3
5.3
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
1.9
1.7
1.5
1.4
1.4
Electronics Industry
+
0.1
0.2
0.2
0.3
0.3
0.2
HFCs
46.5
127.5
168.3
168.1
170.3
169.8
174.6
Substitution of Ozone Depleting







Substances3
0.2
107.3
163.6
164.9
164.7
166.0
170.5
HCFC-22 Production
46.1
20.0
4.3
2.8
5.2
3.3
3.7
Electronics Industry
0.2
0.2
0.3
0.3
0.4
0.4
0.3
Magnesium Production and







Processing
0.0
0.0
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.2
4.4
4.1
4.7
4.5
Electronics Industry
2.8
3.3
3.1
2.9
2.9
3.0
2.7
Aluminum Production
21.5
3.4
2.1
1.4
1.1
1.6
1.8
Substitution of Ozone Depleting







Substances
0.0
+
+
+
+
0.1
0.1
sf6
28.8
11.8
5.5
6.0
5.9
5.7
5.9
Electrical Transmission and







Distribution
23.2
8.4
3.8
4.1
4.2
3.9
4.2
Magnesium Production and







Processing
5.2
2.7
1.0
1.1
1.0
1.0
0.9
Electronics Industry
0.5
0.7
0.7
0.8
0.7
0.8
0.8
nf3
+
0.5
0.6
0.6
0.6
0.6
0.6
Electronics Industry
+
0.5
0.6
0.6
0.6
0.6
0.6
Unspecified Mix of HFCs, PFCs, SF6,







and NF3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total
345.6
365.7
375.4
368.0
367.7
371.3
373.7
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Small amounts of PFC emissions also result from this source.
Table 4-2: Emissions from Industrial Processes and Product Use (kt)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
co2
212,320
194,068
173,480
165,260
163,877
164,348
166,589
Iron and Steel Production &







Metallurgical Coke Production
104,732
70,076
47,941
43,621
40,566
42,627
41,310
Iron and Steel Production
99,124
66,155
43,525
40,979
38,587
41,345
39,944
Metallurgical Coke Production
5,608
3,921
4,417
2,643
1,978
1,282
1,366
Cement Production
33,484
46,194
39,907
39,439
40,324
38,971
40,896
Petrochemical Production
21,611
27,383
28,062
28,310
28,910
29,314
30,792
Ammonia Production
13,047
9,177
10,616
10,245
11,112
12,163
12,272
Lime Production
11,700
14,552
13,342
12,630
12,882
13,106
12,112
Other Process Uses of Carbonates
6,297
7,644
12,182
10,972
9,933
7,469
7,457
Urea Consumption for Non-







Agricultural Purposes
3,784
3,653
4,578
5,132
5,028
5,857
6,222
Carbon Dioxide Consumption
1,472
1,375
4,940
4,640
4,580
4,130
4,870
Aluminum Production
6,831
4,142
2,767
1,334
1,205
1,451
1,880
Soda Ash Production
1,431
1,655
1,714
1,723
1,753
1,714
1,792
Ferroalloy Production
2,152
1,392
1,960
1,796
1,975
2,063
1,598
Titanium Dioxide Production
1,195
1,755
1,635
1,662
1,688
1,541
1,474
Glass Production
1,535
1,928
1,299
1,249
1,296
1,305
1,280
Industrial Processes and Product Use 4-5

-------
Zinc Production
632
1,030
886
838
900
999
1,026
Phosphoric Acid Production
1,529
1,342
999
998
1,028
940
891
Lead Production
516
553
473
500
513
513
540
Carbide Production and







Consumption
370
213
176
170
181
184
175
Magnesium Production and







Processing
1
3
3
3
3
1
1
ch4
12
4
9
11
11
13
15
Petrochemical Production
9
3
7
10
10
12
13
Ferroalloy Production
1
+
1
1
1
1
+
Carbide Production and







Consumption
1
+
+
+
+
+
+
Iron and Steel Production &







Metallurgical Coke Production
1
1
+
+
+
+
+
n2o
112
84
74
78
76
87
71
Nitric Acid Production
41
38
39
34
31
32
34
AdipicAcid Production
51
24
14
23
25
35
18
N20 from Product Uses
14
14
14
14
14
14
14
Caprolactam, Glyoxal, and







Glyoxylic Acid Production
6
7
6
6
5
5
5
Electronics Industry
+
+
1
1
1
1
1
HFCs
M
M
M
M
M
M
M
Substitution of Ozone Depleting







Substances3
M
M
M
M
M
M
M
HCFC-22 Production
3
1
+
+
+
+
+
Electronics Industry
M
M
M
M
M
M
M
Magnesium Production and







Processing
0
0
+
+
+
+
+
PFCs
M
M
M
M
M
M
M
Electronics Industry
M
M
M
M
M
M
M
Aluminum Production
M
M
M
M
M
M
M
Substitution of Ozone Depleting







Substances
0
+
+
+
+
+
+
sf6
1
1
+
+
+
+
+
Electrical Transmission and







Distribution
1
+
+
+
+
+
+
Magnesium Production and







Processing
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
nf3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Unspecified Mix of HFCs, PFCs, SF6,







and NF3
M
M
M
M
M
M
M
Electronics Industry
M
M
M
M
M
M
M
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
M (Mixture of gases)
a Small amounts of PFC emissions also result from this source.
This chapter presents emission estimates calculated in accordance with the 2006IPCC Guidelines for National
Greenhouse Gas Inventories (2006 IPCC Guidelines) and its refinements. For additional detail on IPPU sources that
are not included in this Inventory report, please review Annex 5, Assessment of the Sources and Sinks of
Greenhouse Gas Emissions Not Included. These sources are not included due to various national circumstances,
such as that emissions from a source may not currently occur in the United States, data are not currently available
for those emission sources (e.g., ceramics, non-metallurgical magnesium production, glyoxal and glyoxylic acid
4-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
production, CH4 from direct reduced iron production), emissions are included elsewhere within the Inventory
report, or data suggest that emissions are not significant (e.g., various fluorinated gas emissions from the
electronics industry and other produce uses). In terms of geographic scope, emissions reported in the IPPU chapter
include those from all 50 states, including Hawaii and Alaska, as well as from District of Columbia and U.S.
Territories to the extent to which industries are occurring. While most IPPU sources do not occur in U.S. Territories
(e.g., electronics manufacturing does not occur in U.S. Territories), they are estimated and accounted for where
they are known to occur (e.g., cement production, lime production, and electrical transmission and distribution).
EPA will review this on an ongoing basis to ensure emission sources are included across all geographic areas if they
occur. Information on planned improvements for specific IPPU source categories can be found in the Planned
Improvements section of the individual source category.
In addition, as mentioned in the Energy chapter of this report (Box 3-5), fossil fuels consumed for non-energy uses
for primary purposes other than combustion for energy (including lubricants, paraffin waxes, bitumen asphalt, and
solvents) are reported in the Energy chapter. According to the 2006IPCC Guidelines, these non-energy uses of
fossil fuels are to be reported under the IPPU, rather than the Energy sector; however, due to national
circumstances regarding the allocation of energy statistics and carbon balance data, the United States reports
these non-energy uses in the Energy chapter of this Inventory. Although emissions from these non-energy uses are
reported in the Energy chapter, the methodologies used to determine emissions are compatible with the 2006
IPCC Guidelines and are well documented and scientifically based. The methodologies used are described in
Section 3.2, Carbon Emitted from Non-Energy Uses of Fossil Fuels and Annex 2.3, Methodology for Estimating
Carbon Emitted from Non-Energy Uses of Fossil Fuels. The emissions are reported under the Energy chapter to
improve transparency, report a more complete carbon balance, and avoid double counting. For example, only the
emissions from the first use of lubricants and waxes are to be reported under the IPPU sector, and emissions from
use of lubricants in 2-stroke engines and emissions from secondary use of lubricants and waxes in waste
incineration with energy recovery are to be reported under the Energy sector. Reporting non-energy use emissions
from only first use of lubricants and waxes under IPPU would involve making artificial adjustments to the non-
energy use carbon balance and could potentially result in double counting of emissions. These artificial
adjustments would also be required for asphalt and road oil and solvents (which are captured as part of
petrochemical feedstock emissions) and could also potentially result in double counting of emissions. For more
information, see the Methodology discussion in Section 3.1, C02 from Fossil Fuel Combustion, Section 3.2, Carbon
Emitted from Non-Energy Uses of Fossil Fuels and Annex 2.3, Methodology for Estimating Carbon Emitted from
Non-Energy Uses of Fossil Fuels.
Finally, as stated in the Energy chapter, portions of the fuel consumption data for seven fuel categories—coking
coal, distillate fuel, industrial other coal, petroleum coke, natural gas, residual fuel oil, and other oil—are
reallocated to the IPPU chapter, as they are consumed during non-energy related industrial process activity.
Emissions from uses of fossil fuels as feedstocks or reducing agents (e.g., petrochemical production, aluminum
production, titanium dioxide, zinc production) are reported in the IPPU chapter, unless otherwise noted due to
specific national circumstances. This approach is compatible with the 2006 IPCC Guidelines and is well documented
and scientifically based. The emissions from these feedstocks and reducing agents are reported under the IPPU
chapter to improve transparency and to avoid double counting of emissions under both the Energy and IPPU
sectors. More information on the methodology to adjust for these emissions within the Energy chapter is
described in the Methodology section of C02 from Fossil Fuel Combustion (3.1 Fossil Fuel Combustion [CRF Source
Category 1A]) and Annex 2.1, Methodology for Estimating Emissions of C02 from Fossil Fuel Combustion.
Additional information is listed within each IPPU emission source in which this approach applies.
Box 4-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals
In following the United Nations Framework Convention on Climate Change (UNFCCC) requirement under Article
4.1 to develop and submit national greenhouse gas emission inventories, the emissions and removals presented
in this report and this chapter are organized by source and sink categories and calculated using internationally
accepted methods provided by the Intergovernmental Panel on Climate Change (IPCC) in the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories (2006 IPCC Guidelines) and its supplements and
Industrial Processes and Product Use 4-7

-------
refinements. Additionally, the calculated emissions and removals in a given year for the United States are
presented in a common format in line with the UNFCCC reporting guidelines for the reporting of inventories
under this international agreement. The use of consistent methods to calculate emissions and removals by all
nations providing their inventories to the UNFCCC ensures that these reports are comparable. The presentation
of emissions and removals provided in the IPPU chapter do not preclude alternative examinations, but rather,
this chapter presents emissions and removals in a common format consistent with how countries are to report
Inventories under the UNFCCC. The report itself, and this chapter, follows this standardized format, and
provides an explanation of the application of methods used to calculate emissions and removals from industrial
processes and from the use of greenhouse gases in products.
QA/QC and Verification Procedures
For IPPU sources, a detailed QA/QC plan was developed and implemented for specific categories. This plan is
consistent with the U.S. Inventory QA/QC plan outlined in Annex 8 but tailored to include specific procedures
recommended for these sources. The IPPU QA/QC Plan does not replace the Inventory QA/QC Plan, but rather
provides more context for the IPPU sector. The IPPU QA/QC Plan provides the completed QA/QC forms for each
inventory reports, as well as, for certain source categories (e.g., key categories), more detailed documentation of
quality control checks and recalculations due to methodological changes.
Two types of checks were performed using this plan: (1) general (Tier 1) procedures consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines that focus on annual procedures and checks to be used when gathering,
maintaining, handling, documenting, checking, and archiving the data, supporting documents, and files; and (2)
source category-specific (Tier 2) procedures that focus on checks and comparisons of the emission factors, activity
data, and methodologies used for estimating emissions from the relevant industrial process and product use
sources. Examples of these procedures include: checks to ensure that activity data and emission estimates are
consistent with historical trends to identify significant changes; that, where possible, consistent and reputable data
sources are used and specified across sources; that interpolation or extrapolation techniques are consistent across
sources; and that common datasets, units, and conversion factors are used where applicable. The IPPU QA/QC
plan also checked for transcription errors in data inputs required for emission calculations, including activity data
and emission factors; and confirmed that estimates were calculated and reported for all applicable and able
portions of the source categories for all years.
For sources that use data from EPA's Greenhouse Gas Reporting Program (GHGRP), EPA verifies annual facility-
level reports through a multi-step process (e.g., including a combination of pre-and post-submittal electronic
checks and manual reviews by staff) to identify potential errors and ensure that data submitted to EPA are
accurate, complete, and consistent.3 Based on the results of the verification process, EPA follows up with facilities
to resolve mistakes that may have occurred. The post-submittals checks are consistent with a number of general
and category-specific QC procedures, including: range checks, statistical checks, algorithm checks, and year-to-year
checks of reported data and emissions. See Box 4-2 below for more information on use of GHGRP data in this
chapter.
General QA/QC procedures (Tier 1) and calculation-related QC (category-specific, Tier 2) have been performed for
all IPPU sources. Consistent with the 2006 IPCC Guidelines, additional category-specific QC procedures were
performed for more significant emission categories (such as the comparison of reported consumption with
modeled consumption using EPA's Greenhouse Gas Reporting Program (GHGRP) data within Substitution of Ozone
Depleting Substances) or sources where significant methodological and data updates have taken place. The QA/QC
implementation did not reveal any significant inaccuracies, and all errors identified were documented and
corrected. Application of these procedures, specifically category-specific QC procedures and
3 See .
4-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
updates/improvements as a result of QA processes (expert, public, and UNFCCC technical expert reviews), are
described further within respective source categories, in the Recalculations and Planned Improvement sections.
For most IPPU categories, activity data are obtained via aggregation of facility-level data from EPA's GHGRP (See
Box 4-2 below and Annex 9), national commodity surveys conducted by U.S. Geological Survey National Minerals
Information Center, U.S. Department of Energy (DOE), U.S. Census Bureau, industry associations such as Air-
Conditioning, Heating, and Refrigeration Institute (AHRI), American Chemistry Council (ACC), and American Iron
and Steel Institute (AISI) (specified within each source category). The emission factors used include those derived
from the EPA's GHGRP and application of IPCC default factors. Descriptions of uncertainties and assumptions for
activity data and emission factors are included within the uncertainty discussion sections for each IPPU source
category.
Box 4-2: Industrial Process and Product Use Data from EPA's Greenhouse Gas Reporting Program
EPA collects greenhouse gas emissions data from individual facilities and suppliers of certain fossil fuels and
industrial gases through its Greenhouse Gas Reporting Program (GHGRP). The GHGRP applies to direct
greenhouse gas emitters, fossil fuel suppliers, industrial gas suppliers, and facilities that inject C02 underground
for sequestration or other reasons and requires reporting by sources or suppliers in 41 industrial categories.
Annual reporting is at the facility level, except for certain suppliers of fossil fuels and industrial greenhouse
gases.
In general, the threshold for reporting is 25,000 metric tons or more of C02 Eq. per year, but reporting is
required for all facilities in some industries. Calendar year 2010 was the first year for which data were collected
for facilities subject to 40 CFR Part 98, though some source categories first collected data for calendar year
2011. For more information, see Annex 9, Use of EPA Greenhouse Gas Reporting Program in Inventory.
EPA uses annual GHGRP data in a number of categories to improve the national estimates presented in this
Inventory, consistent with IPCC guidelines (e.g., minerals, chemicals, product uses). Methodologies used in
EPA's GHGRP are consistent with IPCC guidelines, including higher tier methods; however, it should be noted
that the coverage and definitions for source categories (e.g., allocation of energy and IPPU emissions) in EPA's
GHGRP may differ from those used in this Inventory in meeting the UNFCCC reporting guidelines (IPCC 2011). In
line with the UNFCCC reporting guidelines, the Inventory is a comprehensive accounting of all emissions from
source categories identified in the 2006 IPCC Guidelines. EPA has paid particular attention to ensuring both
completeness and time-series consistency for major recalculations that have occurred from the incorporation of
GHGRP data into these categories, consistent with 2006 IPCC Guidelines and IPCC Technical Bulletin on Use of
Facility-Specific Data in National GHG Inventories,4
For certain source categories in this Inventory (e.g., nitric acid production, lime production, cement production,
petrochemical production, carbon dioxide consumption, ammonia production, and urea consumption for non-
agricultural purposes), EPA has integrated data values that have been calculated by aggregating GHGRP data
that are considered confidential business information (CBI) at the facility level. EPA, with industry engagement,
has put forth criteria to confirm that a given data aggregation shields underlying CBI from public disclosure. EPA
is only publishing data values that meet these aggregation criteria.5 Specific uses of aggregated facility-1 eve I
data are described in the respective methodological sections (e.g., including other sources using GHGRP data
that is not aggregated CBI, such as aluminum, electronics industry, electrical transmission and distribution,
HCFC-22 production, and magnesium production and processing.). For other source categories in this chapter,
as indicated in the respective planned improvements sections,6 EPA is continuing to analyze how facility-level
4	See .
5	U.S. EPA Greenhouse Gas Reporting Program. Developments on Publication of Aggregated Greenhouse Gas Data, November
25, 2014. See .
6	Ammonia Production, Glass Production, Lead Production, and Other Fluorinated Gas Production.
Industrial Processes and Product Use 4-9

-------
GHGRP data may be used to improve the national estimates presented in this Inventory, giving particular
consideration to ensuring time-series consistency and completeness.
Additionally, EPA's GHGRP has and will continue to enhance QA/QC procedures and assessment of uncertainties
within the IPPU categories (see those categories for specific QA/QC details regarding the use of GHGRP data).
4.1 Cement Production (CRF Source Category
2A1)	
Cement production is an energy- and raw material-intensive process that results in the generation of carbon
dioxide (C02) both from the energy consumed in making the clinker precursor to cement and from the chemical
process to make the clinker. Emissions from fuels consumed for energy purposes during the production of cement
are accounted for in the Energy chapter.
During the clinker production process, the key reaction occurs when calcium carbonate (CaC03), in the form of
limestone or similar rocks, is heated in a cement kiln at a temperature range of about 700 to 1,000 degrees Celsius
(1,300 to 1,800 degrees Fahrenheit) to form lime (i.e., calcium oxide, or CaO) and C02 in a process known as
calcination or calcining. The quantity of C02 emitted during clinker production is directly proportional to the lime
content of the clinker. During calcination, each mole of CaC03 heated in the clinker kiln forms one mole of CaO and
one mole of C02. The C02 is vented to the atmosphere as part of the kiln lime exhaust:
CaC03 + heat -» CaO + C02
Next, over a temperature range of 1000 to 1450 degrees Celsius, the CaO combines with alumina, iron oxide and
silica that are also present in the clinker raw material mix to form hydraulically reactive compounds within white-
hot semifused (sintered) nodules of clinker. Because these "sintering" reactions are highly exothermic, they
produce few C02 process emissions. The clinker is then rapidly cooled to maintain quality and then very finely
ground with a small amount of gypsum and potentially other materials (e.g., ground granulated blast furnace slag,
etc.) to make portland and similar cements. Masonry cement consists of plasticizers (e.g., ground limestone, lime,
etc.) and portland cement, and the amount of portland cement used accounts for approximately 3 percent of total
clinker production (USGS 2020). There are no additional emissions associated with the production of masonry
cement. Carbon dioxide emissions that result from the production of lime used to produce portland and masonry
cement are included in Section 4.2 Lime Production (CRF Source Category 2A2).
Carbon dioxide emitted from the chemical process of cement production is the second largest source of industrial
C02 emissions in the United States. Cement is produced in 34 states and Puerto Rico. Texas, California, Missouri,
Florida, Alabama, Michigan, and Pennsylvania were the leading cement-producing states in 2019 and accounted
for almost 60 percent of total U.S. production (USGS 2020). Clinker production in 2019 remained at relatively flat
levels, compared to 2018 (EPA 2020; USGS 2020). In 2019, cement sales increased slightly, and imports of clinker
for consumption increased by approximately 14 percent from 2018 (USGS 2020). In 2019, U.S. clinker production
totaled 78,600 kilotons (EPA 2020). The resulting C02 emissions were estimated to be 40.9 MMT C02 Eq. (40,896
kt) (see Table 4-3).
4-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-3: CO2 Emissions from Cement Production (MMT CO2 Eq. and kt)
Year
MMT CO? Eq.
kt
1990
33.5
33,484
2005
46.2
46,194
2015
2016
2017
2018
2019
39.9
39.4
40.3
39.0
40.9
39,907
39,439
40,324
38,971
40,896
Greenhouse gas emissions from cement production, which are primarily driven by production levels, increased
every year from 1991 through 2006 but decreased in the following years until 2009. Since 1990, emissions have
increased by 22 percent. Emissions from cement production were at their lowest levels in 2009 (2009 emissions
are approximately 28 percent lower than 2008 emissions and 12 percent lower than 1990) due to the economic
recession and the associated decrease in demand for construction materials. Since 2010, emissions have increased
by about 30 percent, due to increasing demand for cement. Cement continues to be a critical component of the
construction industry; therefore, the availability of public and private construction funding, as well as overall
economic conditions, have considerable impact on the level of cement production.
Carbon dioxide emissions from cement production were estimated using the Tier 2 methodology from the 2006
IPCC Guidelines as this is a key category. The Tier 2 methodology was used because detailed and complete data
(including weights and composition) for carbonate(s) consumed in clinker production are not available,7 and thus a
rigorous Tier 3 approach is impractical. Tier 2 specifies the use of aggregated plant or national clinker production
data and an emission factor, which is the product of the average lime fraction for clinker of 65 percent and a
constant reflecting the mass of C02 released per unit of lime. The U.S. Geological Survey (USGS) mineral
commodity expert for cement has confirmed that this is a reasonable assumption for the United States (Van Oss
2013a). This calculation yields an emission factor of 0.510 tons of C02 per ton of clinker produced, which was
determined as follows:
EFciinker = 0.650 CaO x [(44.01 g/mole CO2) -h (56.08 g/mole CaO)] = 0.510 tons CCh/ton clinker
During clinker production, some of the raw materials, partially reacted raw materials, and clinker enters the kiln
line's exhaust system as non-calcinated, partially calcinated, or fully calcinated cement kiln dust (CKD). To the
degree that the CKD contains carbonate raw materials which are then calcined, there are associated C02 emissions.
At some plants, essentially all CKD is directly returned to the kiln, becoming part of the raw material feed, or is
likewise returned to the kiln after first being removed from the exhaust. In either case, the returned CKD becomes
a raw material, thus forming clinker, and the associated C02 emissions are a component of those calculated for the
clinker overall. At some plants, however, the CKD cannot be returned to the kiln because it is chemically unsuitable
as a raw material or chemical issues limit the amount of CKD that can be so reused. Any clinker that cannot be
returned to the kiln is either used for other (non-clinker) purposes or is landfilled. The C02 emissions attributable
to the non-returned calcinated portion of the CKD are not accounted for by the clinker emission factor and thus a
CKD correction factor should be applied to account for those emissions. The USGS reports the amount of CKD used
7 As discussed further under "Planned Improvements," most cement-producing facilities that report their emissions to the
GHGRP use CEMS to monitor combined process and fuel combustion emissions for kilns, making it difficult to quantify the
process emissions on a facility-specific basis. In 2019, the percentage of facilities not using CEMS was 8 percent.
Methodology
Industrial Processes and Product Use 4-11

-------
to produce clinker, but no information is currently available on the total amount of CKD produced annually.8
Because data are not currently available to derive a country-specific CKD correction factor, a default correction
factor of 1.02 (2 percent) was used to account for CKD C02 emissions, as recommended by the IPCC (IPCC 2006).9
Total cement production emissions were calculated by adding the emissions from clinker production and the
emissions assigned to CKD.
Small amounts of impurities (i.e., not calcium carbonate) may exist in the raw limestone used to produce clinker.
The proportion of these impurities is generally minimal, although a small amount (1 to 2 percent) of magnesium
oxide (MgO) may be desirable as a flux. Per the IPCC Tier 2 methodology, a correction for MgO is not used, since
the amount of MgO from carbonate is likely very small and the assumption of a 100 percent carbonate source of
CaO already yields an overestimation of emissions (IPCC 2006).
The 1990 through 2012 activity data for clinker production (see Table 4-4) were obtained from USGS (Van Oss
2013a, Van Oss 2013b). Clinker production data for 2013 were also obtained from USGS (USGS 2014). USGS
compiled the data (to the nearest ton) through questionnaires sent to domestic clinker and cement manufacturing
plants, including facilities in Puerto Rico. Clinker production values in the current Inventory report utilize GHGRP
data for the years 2014 through 2019 (EPA 2020). Details on how this GHGRP data compares to USGS reported
data can be found in the section on QA/QC and Verification.
Table 4-4: Clinker Production (kt)
Year	Clinker
1990	64,355
2005	88,783
2015	76,700
2016	75,800
2017	77,500
2018	74,900
201	9	78,600	
Notes: Clinker production from 1990
through 2019 includes Puerto Rico
(relevant U.S. Territories).
Uncertainty and Time-Series Consistency
The uncertainties contained in these estimates are primarily due to uncertainties in the lime content of clinker and
in the percentage of CKD recycled inside the cement kiln. Uncertainty is also associated with the assumption that
all calcium-containing raw materials are CaC03, when a small percentage likely consists of other carbonate and
non-carbonate raw materials. The lime content of clinker varies from 60 to 67 percent; 65 percent is used as a
representative value (Van Oss 2013a). The amount of C02 from CKD loss can range from 1.5 to 8 percent
depending upon plant specifications. Additionally, some amount of C02 is reabsorbed when the cement is used for
construction. As cement reacts with water, alkaline substances such as calcium hydroxide are formed. During this
curing process, these compounds may react with C02 in the atmosphere to create calcium carbonate. This reaction
8	The USGS Minerals Yearbook: Cement notes that CKD values used for clinker production are likely underreported.
9	As stated on p. 2.12 of the 2006 IPCC Guidelines, Vol. 3, Chapter 2: "...As data on the amount of CKD produced may be scarce
(except possibly for plant-level reporting), estimating emissions from lost CKD based on a default value can be considered good
practice. The amount of C02 from lost CKD can vary, but ranges typically from about 1.5 percent (additional CO2 relative to that
calculated for clinker) for a modern plant to about 20 percent for a plant losing a lot of highly calcinated CKD (van Oss, 2005). In
the absence of data, the default CKD correction factor (CFckd) is 1.02 (i.e., add 2 percent to the CO2 calculated for clinker). If no
calcined CKD is believed to be lost to the system, the CKD correction factor will be 1.00 (van Oss, 2005)..."
4-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
only occurs in roughly the outer 0.2 inches of the total thickness. Because the amount of C02 reabsorbed is
thought to be minimal, it was not estimated.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-5. Based on the
uncertainties associated with total U.S. clinker production, the C02 emission factor for clinker production, and the
emission factor for additional C02 emissions from CKD, 2019 C02 emissions from cement production were
estimated to be between 38.5 and 43.4 MMT C02 Eq. at the 95 percent confidence level. This confidence level
indicates a range of approximately 6 percent below and 6 percent above the emission estimate of 40.9 MMT C02
Eq.
Table 4-5: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Cement
Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Cement Production
C02
40.9
38.5 43.4
-6% +6%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
EPA relied upon the latest guidance from the IPCC on the use of facility-level data in national inventories and
applied a category-specific QC process to compare activity data from EPA's GHGRP with existing data from USGS
surveys. This was to ensure time-series consistency of the emission estimates presented in the Inventory. Total
U.S. clinker production is assumed to have low uncertainty because facilities routinely measure this for economic
reasons and because both USGS and GHGRP take multiple steps to ensure that reported totals are accurate. EPA
verifies annual facility-level GHGRP reports through a multi-step process that is tailored to the reporting industry
(e.g., combination of electronic checks including range checks, statistical checks, algorithm checks, year-to-year
comparison checks, along with manual reviews involving outside data checks) to identify potential errors and
ensure that data submitted to EPA are accurate, complete, and consistent (EPA 2015). Based on the results of the
verification process, EPA follows up with facilities to resolve mistakes that may have occurred.10 Facilities are also
required to monitor and maintain records of monthly clinker production per section 98.84 of the GHGRP regulation
(40 CFR 98.84).
EPA's GHGRP requires all facilities producing Portland cement to report greenhouse gas emissions, including C02
process emissions from each kiln, C02 combustion emissions from each kiln, CH4 and N20 combustion emissions
from each kiln, and C02, CH4, and N20 emissions from each stationary combustion unit other than kilns (40 CFR
Part 98 Subpart H). Source-specific quality control measures for the Cement Production category are included in
section 98.84, Monitoring and QA/QC Requirements.
10 See GHGRP Verification Fact Sheet .
Industrial Processes and Product Use 4-13

-------
As mentioned above, EPA compares GHGRP clinker production data to the USGS clinker production data. For the
year 2014 and 2018, USGS and GHGRP clinker production data showed a difference of approximately 2 percent
and 3 percent, respectively. In 2015, 2016, 2017, and 2019, that difference was less than 1 percent between the
two sets of activity data. This difference resulted in an increase of emissions compared to USGS data by less than
0.1 MMT C02 Eq. in 2015, 2016, 2017, and 2019. The information collected by the USGS National Minerals
Information Center surveys continue to be an important data source.
Recalculations Discussion
Recalculations were performed for year 2018 based on updated clinker production data from EPA's GHGRP.
Compared to the previous Inventory, emissions for 2018 decreased by 3 percent (1,353 kt C02 Eq.).
Planned Improvements
EPA is continuing to evaluate and analyze data reported under EPA's GHGRP that would be useful to improve the
emission estimates for the Cement Production source category. Most cement production facilities reporting under
EPA's GHGRP use Continuous Emission Monitoring Systems (CEMS) to monitor and report C02 emissions, thus
reporting combined process and combustion emissions from kilns. In implementing further improvements and
integration of data from EPA's GHGRP, the latest guidance from the IPCC on the use of facility-level data in national
inventories will be relied upon, in addition to category-specific QC methods recommended by the 2006 IPCC
Guidelines.11 EPA's long-term improvement plan includes continued assessment of the feasibility of using
additional GHGRP information beyond aggregation of reported facility-level clinker data, in particular
disaggregating the combined process and combustion emissions reported using CEMS, to separately present
national process and combustion emissions streams consistent with IPCC and UNFCCC guidelines. This long-term
planned analysis is still in development and has not been applied for this current Inventory.
Finally, in response to feedback from Portland Cement Association (PCA) during the Public Review comment period
of a previous Inventory, EPA plans to work with PCA to discuss additional long-term improvements to review
methods and data used to estimate C02 emissions from cement production to account for both organic material
and magnesium carbonate in the raw material, and to discuss the carbonation that occurs across the duration of
the cement product. Priority work includes identifying data and studies on the average MgO content of clinker
produced in the United States, the average carbon content for organic materials in kiln feed in the United States,
and C02 reabsorption rates via carbonation for various cement products. This information is not reported by
facilities subject to report to GHGRP. EPA met with PCA in the fall of 2020 to discuss PCA's latest research on
carbonation.
4.2 Lime Production (CRF Source Category
2A2)	
Lime is an important manufactured product with many industrial, chemical, and environmental applications. Lime
production involves three main processes: stone preparation, calcination, and hydration. Carbon dioxide (C02) is
generated during the calcination stage, when limestone—mostly calcium carbonate (CaC03)—is roasted at high
temperatures in a kiln to produce calcium oxide (CaO) and C02. The C02 is given off as a gas and is normally
emitted to the atmosphere.
11 See IPCC Technical Bulletin on Use of Facility-Specific Data in National Greenhouse Gas Inventories .
4-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
CaCO3 —> C&0 + CO2
Some of the C02 generated during the production process, however, is recovered at some facilities for use in sugar
refining and precipitated calcium carbonate (PCC) production.12 Emissions from fuels consumed for energy
purposes during the production of lime are included in the Energy chapter.
For U.S. operations, the term "lime" actually refers to a variety of chemical compounds. These include CaO, or
high-calcium quicklime; calcium hydroxide (Ca(OH)2), or hydrated lime; dolomitic quicklime ([CaOMgO]); and
dolomitic hydrate ([Ca(OH)2*MgO] or [Ca(OH)2*Mg(OH)2]).
The current lime market is approximately distributed across five end-use categories, as follows: metallurgical uses,
34 percent; environmental uses, 30 percent; chemical and industrial uses, 21 percent; construction uses, 11
percent; and refractory dolomite, 1 percent (USGS 2020b). The major uses are in steel making, flue gas
desulfurization (FGD) systems at coal-fired electric power plants, construction, and water treatment, as well as
uses in mining, pulp and paper and precipitated calcium carbonate manufacturing. Lime is also used as a C02
scrubber, and there has been experimentation on the use of lime to capture C02 from electric power plants. Both
lime (CaO) and limestone (CaC03) can be used as a sorbent for FGD systems. Emissions from limestone
consumption for FGD systems are reported under Section 4.4 Other Process Uses of Carbonate Production (CRF
Source Category 2A4).
Emissions from lime production have increased and decreased over the time series depending on lime end-use
markets - primarily the steel making industry and FGD systems for utility and industrial plants - and also energy
costs. One significant change to lime end-use since 1990 has been the increase in demand for lime for FGD at coal-
fired electric power plants, which can be attributed to compliance with sulfur dioxide (S02) emission regulations of
the Clean Air Act Amendments of 1990. Phase I went into effect on January 1,1995, followed by Phase II on
January 1, 2000. To supply lime for the FGD market, the lime industry installed more than 1.8 million tons per year
of new capacity by the end of 1995 (USGS 1996). The need for air pollution controls continued to drive the FGD
lime market, which had doubled between 1990 and 2019 (USGS 1991 and 2020d).
The U.S. lime industry temporarily shut down individual gas-fired kilns and, in some case, entire lime plants during
2000	and 2001, due to significant increases in the price of natural gas. Lime production continued to decrease in
2001	and 2002, a result of lower demand from the steel making industry, lime's largest end-use market, when
domestic steel producers were affected by low priced imports and slowing demand (USGS 2002).
Emissions from lime production increased and then peaked in 2006 at approximately 30.3 percent above 1990
levels, due to strong demand from the steel and construction markets (road and highway construction projects),
before dropping to its lowest level in 2009 at approximately 2.5 percent below 1990 emissions, driven by the
economic recession and downturn in major markets including construction, mining, and steel (USGS 2007, 2008,
2010). In 2010, the lime industry began to recover as the steel, FGD, and construction markets also recovered
(USGS 2011 and 2012). Fluctuation in lime production since 2015 has been driven largely by demand from the steel
making industry (USGS 2018b, 2019, 2020b, 2020c).
Lime production in the United States—including Puerto Rico—was reported to be 16,897 kilotons in 2019 (USGS
2020a). Lime production in 2019 decreased by about 7 percent, compared to 2018 levels (USGS 2020a). Compared
to 1990, lime production increased by about 7 percent. At year-end 2019, there were 74 operating primary lime
plants in the United States, including Puerto Rico according to the USGS MCS (USGS 2020a).13 Principal lime
producing states are Missouri, Alabama, Ohio, Texas, and Kentucky (USGS 2020a).
U.S. lime production resulted in estimated net C02 emissions of 12.1 MMT C02 Eq. (12,112 kt) (see Table 4-6 and
Table 4-7). Carbon dioxide emissions from lime production decreased by about 8 percent compared to 2018 levels.
12	PCC is obtained from the reaction of C02 with calcium hydroxide. It is used as a filler and/or coating in the paper, food, and
plastic industries.
13	In 2019, 71 operating primary lime facilities in the United States reported to the EPA Greenhouse Gas Reporting Program due
to closures.
Industrial Processes and Product Use 4-15

-------
Compared to 1990, C02 emissions have increased 3.5 percent. The trends in C02 emissions from lime production
are directly proportional to trends in production, which are described above.
Table 4-6: CO2 Emissions from Lime Production (MMT CO2 Eq. and kt)
Year
MMTCOz Eq.
kt
1990
11.7
11,700
2005
14.6
14,552
2015
13.3
13,342
2016
12.6
12,630
2017
12.9
12,882
2018
13.1
13,106
2019
12.1
12,112
Table 4-7: Gross, Recovered, and Net CO2 Emissions from Lime Production (kt)
Year	Gross	Recovered3 Net Emissions
1990
11,959
259
11,700
2005
15,074
522
14,552
2015
13,764
422
13,342
2016
13,000
370
12,630
2017
13,283
401
12,882
2018
13,609
503
13,106
2019
12,676
564
12,112
Note: Totals may not sum due to independent rounding.
3 For sugar refining and PCC production.
Methodology
To calculate emissions, the amounts of high-calcium and dolomitic lime produced were multiplied by their
respective emission factors using the Tier 2 approach from the 2006IPCC Guidelines. The emission factor is the
product of the stoichiometric ratio between C02 and CaO, and the average CaO and MgO content for lime. The
CaO and MgO content for lime is assumed to be 95 percent for both high-calcium and dolomitic lime (IPCC 2006).
The emission factors were calculated as follows:
For high-calcium lime:
[(44.01 g/mole C02) 4- (56.08 g/mole CaO)] x (0.9500 CaO/lime) = 0.7455 g COz/g lime
For dolomitic lime:
[(88.02 g/mole C02) 4 (96.39 g/mole CaO)] x (0.9500 CaO/lime) = 0.8675 g COz/g lime
Production was adjusted to remove the mass of chemically combined water found in hydrated lime, determined
according to the molecular weight ratios of H20 to (Ca(OH)2 and [Ca(OH)2*Mg(OH)2]) (IPCC 2006). These factors set
the chemically combined water content to 27 percent for high-calcium hydrated lime, and 30 percent for dolomitic
hydrated lime.
The 2006 IPCC Guidelines (Tier 2 method) also recommends accounting for emissions from lime kiln dust (LKD)
through application of a correction factor. LKD is a byproduct of the lime manufacturing process typically not
recycled back to kilns. LKD is a very fine-grained material and is especially useful for applications requiring very
small particle size. Most common LKD applications include soil reclamation and agriculture. Emissions from the
4-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
application of lime for agricultural purposes are reported in the Agriculture chapter under 5.5 Liming (CRF Source
Category 3G). Currently, data on annual LKD production is not readily available to develop a country-specific
correction factor. Lime emission estimates were multiplied by a factor of 1.02 to account for emissions from LKD
(IPCC 2006). See the Planned Improvements section associated with efforts to improve uncertainty analysis and
emission estimates associated with LKD.
Lime emission estimates were further adjusted to account for the amount of C02 captured for use in on-site
processes. All the domestic lime facilities are required to report these data to EPA under its GHGRP. The total
national-level annual amount of C02 captured for on-site process use was obtained from EPA's GHGRP (EPA 2020)
based on reported facility-level data for years 2010 through 2019. The amount of C02 captured/recovered for on-
site process use is deducted from the total gross emissions (i.e., from lime production and LKD). The net lime
emissions are presented in Table 4-6 and Table 4-7. GHGRP data on C02 removals (i.e., C02 captured/recovered)
was available only for 2010 through 2019. Since GHGRP data are not available for 1990 through 2009, IPCC
"splicing" techniques were used as per the 2006 IPCC Guidelines on time-series consistency (IPCC 2006, Volume 1,
Chapter 5).
Lime production data by type (i.e., high-calcium and dolomitic quicklime, high-calcium and dolomitic hydrated
lime, and dead-burned dolomite) for 1990 through 2019 (see Table 4-8) were obtained from U.S. Geological Survey
(USGS) Minerals Yearbook (USGS 1992 through 2020d) and are compiled by USGS to the nearest ton. Dead-burned
dolomite data are additionally rounded by USGS to no more than one significant digit to avoid disclosing company
proprietary data. Natural hydraulic lime, which is produced from CaO and hydraulic calcium silicates, is not
manufactured in the United States (USGS 2018a). Total lime production was adjusted to account for the water
content of hydrated lime by converting hydrate to oxide equivalent based on recommendations from the IPCC and
using the water content values for high-calcium hydrated lime and dolomitic hydrated lime mentioned above, and
is presented in Table 4-9 (IPCC 2006). The CaO and CaOMgO contents of lime, both 95 percent, were obtained
from the IPCC (IPCC 2006). Since data for the individual lime types (high calcium and dolomitic) were not provided
prior to 1997, total lime production for 1990 through 1996 was calculated according to the three-year distribution
from 1997 to 1999.
Table 4-8: High-Calcium- and Dolomitic-Quicklime, High-Calcium- and Dolomitic-Hydrated,
and Dead-Burned-Dolomite Lime Production (kt)

High-Calcium
Dolomitic
High-Calcium
Dolomitic
Dead-Burned
Year
Quicklime
Quicklime
Hydrated
Hydrated
Dolomite
1990
11,166
2,234
1,781
319
342
2005
14,100
2,990
2,220
474
200
2015
13,100
2,550
2,150
279
200
2016
12,100
2,420
2,350
280
200
2017
12,200
2,650
2,360
276
200
2018
12,400
2,810
2,430
265
200
2019
11,300
2,700
2,430
267
200
Table 4-9:	Adjusted Lime Production (kt)
Year	High-Calcium	Dolomitic
1990	12,466	2,800
2005	15,721	3,522
2,945
2,816
3,043
3,196
2015	14,670
2016	13,816
2017	13,923
2018	14,174
Industrial Processes and Product Use 4-17

-------
2019	13,074
3,087
Note: Minus water content of hydrated
lime.
Uncertainty and Time-Series Consistency
The uncertainties contained in these estimates can be attributed to slight differences in the chemical composition
of lime products and C02 recovery rates for on-site process use over the time series. Although the methodology
accounts for various formulations of lime, it does not account for the trace impurities found in lime, such as iron
oxide, alumina, and silica. Due to differences in the limestone used as a raw material, a rigid specification of lime
material is impossible. As a result, few plants produce lime with exactly the same properties.
In addition, a portion of the C02 emitted during lime production will actually be reabsorbed when the lime is
consumed, especially at captive lime production facilities. As noted above, lime has many different chemical,
industrial, environmental, and construction applications. In many processes, C02 reacts with the lime to create
calcium carbonate (e.g., water softening). Carbon dioxide reabsorption rates vary, however, depending on the
application. For example, 100 percent of the lime used to produce precipitated calcium carbonate reacts with C02,
whereas most of the lime used in steel making reacts with impurities such as silica, sulfur, and aluminum
compounds. Quantifying the amount of C02 that is reabsorbed would require a detailed accounting of lime use in
the United States and additional information about the associated processes where both the lime and byproduct
C02 are "reused." Research conducted thus far has not yielded the necessary information to quantify C02
reabsorption rates.14 Some additional information on the amount of C02 consumed on site at lime facilities,
however, has been obtained from EPA's GHGRP.
In some cases, lime is generated from calcium carbonate byproducts at pulp mills and water treatment plants.15
The lime generated by these processes is included in the USGS data for commercial lime consumption. In the
pulping industry, mostly using the Kraft (sulfate) pulping process, lime is consumed in order to causticize a process
liquor (green liquor) composed of sodium carbonate and sodium sulfide. The green liquor results from the dilution
of the smelt created by combustion of the black liquor where biogenic carbon (C) is present from the wood. Kraft
mills recover the calcium carbonate "mud" after the causticizing operation and calcine it back into lime—thereby
generating C02—for reuse in the pulping process. Although this re-generation of lime could be considered a lime
manufacturing process, the C02 emitted during this process is mostly biogenic in origin and therefore is not
included in the industrial processes totals (Miner and Upton 2002). In accordance with IPCC methodological
guidelines, any such emissions are calculated by accounting for net C fluxes from changes in biogenic C reservoirs
in wooded or crop lands (see the Land Use, Land-Use Change, and Forestry chapter).
In the case of water treatment plants, lime is used in the softening process. Some large water treatment plants
may recover their waste calcium carbonate and calcine it into quicklime for reuse in the softening process. Further
research is necessary to determine the degree to which lime recycling is practiced by water treatment plants in the
United States.
Another uncertainty is the assumption that calcination emissions for LKD are around 2 percent. The National Lime
Association (NLA) has commented that the estimates of emissions from LKD in the United States could be closer to
6 percent. They also note that additional emissions (approximately 2 percent) may also be generated through
production of other byproducts/wastes (off-spec lime that is not recycled, scrubber sludge) at lime plants (Seeger
2013). Publicly available data on LKD generation rates, total quantities not used in cement production, and types of
14	Representatives of the National Lime Association estimate that C02 reabsorption that occurs from the use of lime may offset
as much as a quarter of the C02 emissions from calcination (Males 2003).
15	Some carbide producers may also regenerate lime from their calcium hydroxide byproducts, which does not result in
emissions of C02. In making calcium carbide, quicklime is mixed with coke and heated in electric furnaces. The regeneration of
lime in this process is done using a waste calcium hydroxide (hydrated lime) [CaC2 + 2H20 -> C2H2 + Ca(OH) 2], not calcium
carbonate [CaC03]. Thus, the calcium hydroxide is heated in the kiln to simply expel the water [Ca(OH)2 + heat -> CaO + H20],
and no C02 is released.
4-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
other byproducts/wastes produced at lime facilities are limited. NLA compiled and shared historical emissions
information and quantities for some waste products reported by member facilities associated with generation of
total calcined byproducts and LKD, as well as methodology and calculation worksheets that member facilities
complete when reporting. There is uncertainty regarding the availability of data across the time series needed to
generate a representative country-specific LKD factor. Uncertainty of the activity data is also a function of the
reliability and completeness of voluntarily reported plant-level production data. Further research, including
outreach and discussion with NLA, and data is needed to improve understanding of additional calcination
emissions to consider revising the current assumptions that are based on IPCC guidelines. More information can be
found in the Planned Improvements section below.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-10. Lime C02 emissions
for 2019 were estimated to be between 11.9 and 12.4 MMT C02 Eq. at the 95 percent confidence level. This
confidence level indicates a range of approximately 2 percent below and 2 percent above the emission estimate of
12.1 MMT C02 Eq.
Table 4-10: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lime
Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTC02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Lower
Upper



Bound Bound
Bound
Bound
Lime Production
C02
12.1
11.9 12.4
-2%
+2%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006 IPCC Guidelines as noted in the introduction
of the IPPU chapter (see Annex 8 for more details).
More details on the greenhouse gas calculation, monitoring and QA/QC methods associated with reporting on C02
captured for onsite use applicable to lime manufacturing facilities can be found under Subpart S (Lime
Manufacturing) of the GHGRP regulation (40 CFR Part 98).16 EPA verifies annual facility-level GHGRP reports
through a multi-step process (e.g., combination of electronic checks and manual reviews) to identify potential
errors and ensure that data submitted to EPA are accurate, complete, and consistent (EPA 2020).17 Based on the
results of the verification process, EPA follows up with facilities to resolve mistakes that may have occurred. The
post-submittals checks are consistent with a number of general and category-specific QC procedures, including:
range checks, statistical checks, algorithm checks, and year-to-year checks of reported data and emissions.
Recalculations Discussion
Recalculations were performed for year 2018 based on updated C02 captured for on-site process use data
obtained from EPA's GHGRP. Recalculations were performed for years 2015, 2016, 2017, and 2018 based on
updated high-calcium and dolomitic quicklime, high-calcium and dolomitic hydrated lime, and dead-burned
16	See .
17	See .
Industrial Processes and Product Use 4-19

-------
dolomite production data from revised USGS data (USGS 2020d). The updates resulted in less than a 1 percent
decrease in C02 emissions for 2018 and a less than 1 percent increase in C02 emissions for 2015, 2016, and 2017,
compared to the previous Inventory.
Planned Improvements
EPA plans to review GHGRP emissions and activity data reported to EPA under Subpart S of the GHGRP regulation
(40 CFR Part 98), and aggregated activity data on lime production by type in particular. In addition, initial review of
data has identified that several facilities use CEMS to report emissions. Under Subpart S, if a facility is using a
CEMS, they are required to report combined combustion emissions and process emissions. EPA continues to
review how best to incorporate GHGRP and notes that particular attention will be made to also ensuring time-
series consistency of the emissions estimates presented in future Inventory reports, consistent with IPCC and
UNFCCC guidelines. This is required because the facility-level reporting data from EPA's GHGRP, with the program's
initial requirements for reporting of emissions in calendar year 2010, are not available for all inventory years (i.e.,
1990 through 2009) as required for this Inventory. In implementing improvements and integration of data from
EPA's GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories will be
relied upon.18
Future improvements involve improving and/or confirming the representativeness of current assumptions
associated with emissions from production of LKD and other byproducts/wastes as discussed in the Uncertainty
and Time Series Consistency section, per comments from the NLA provided during a prior Public Review comment
period for this (i.e., 1990 through 2018) and previous Inventories. EPA met with NLA in summer of 2020 for
clarification on data needs and available data and to discuss planned research into GHGRP data. Previously, EPA
met with NLA in spring of 2015 to outline specific information required to apply IPCC methods to develop a
country-specific correction factor to more accurately estimate emissions from production of LKD. In 2016, NLA
compiled and shared historical emissions information reported by member facilities on an annual basis under
voluntary reporting initiatives from 2002 through 2011 associated with generation of total calcined byproducts and
LKD. Reporting of LKD was only differentiated for the years 2010 and 2011. This emissions information was
reported on a voluntary basis consistent with NLA's facility-level reporting protocol, which was also provided to
EPA. To reflect information provided by NLA, EPA updated the qualitative description of uncertainty. At the time of
this Inventory, this planned improvement is in process and has not been incorporated into this current Inventory
report.
4.3 Glass Production (CRF Source Category
2A3)	
Glass production is an energy and raw-material intensive process that results in the generation of carbon dioxide
(C02) from both the energy consumed in making glass and the glass production process itself. Emissions from fuels
consumed for energy purposes during the production of glass are included in the Energy sector.
Glass production employs a variety of raw materials in a glass-batch. These include formers, fluxes, stabilizers, and
sometimes colorants. The major raw materials (i.e., fluxes and stabilizers) that emit process-related C02 emissions
during the glass melting process are limestone, dolomite, and soda ash. The main former in all types of glass is
silica (Si02). Other major formers in glass include feldspar and boric acid (i.e., borax). Fluxes are added to lower the
temperature at which the batch melts. Most commonly used flux materials are soda ash (sodium carbonate,
Na2C03) and potash (potassium carbonate, K20). Stabilizers are used to make glass more chemically stable and to
keep the finished glass from dissolving and/or falling apart. Commonly used stabilizing agents in glass production
18 See .
4-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
are limestone (CaC03), dolomite (CaC03MgC03), alumina (Al203), magnesia (MgO), barium carbonate (BaC03),
strontium carbonate (SrC03), lithium carbonate (Li2C03), and zirconia (Zr02) (OIT 2002). Glass makers also use a
certain amount of recycled scrap glass (cullet), which comes from in-house return of glassware broken in the
process or other glass spillage or retention, such as recycling or from cullet broker services.
The raw materials (primarily soda ash, limestone, and dolomite) release C02 emissions in a complex high-
temperature chemical reaction during the glass melting process. This process is not directly comparable to the
calcination process used in lime manufacturing, cement manufacturing, and process uses of carbonates (i.e.,
limestone/dolomite use) but has the same net effect in terms of C02 emissions (IPCC 2006).
The U.S. glass industry can be divided into four main categories: containers, flat (window) glass, fiber glass, and
specialty glass. The majority of commercial glass produced is container and flat glass (EPA 2009). The United States
is one of the major global exporters of glass. Domestically, demand comes mainly from the construction, auto,
bottling, and container industries. There are more than 1,500 companies that manufacture glass in the United
States, with the largest being Corning, Guardian Industries, Owens-Illinois, and PPG Industries.19
In 2019, 2,220 kilotons of soda ash and 817 kilotons of limestone were consumed for glass production (USGS 2020;
USGS 2020a). Dolomite consumption data for glass manufacturing was reported to be zero for 2019. Use of
limestone and soda ash in glass production resulted in aggregate C02 emissions of 1.3 MMT C02 Eq. (1,280 kt) (see
Table 4-11). Overall, emissions have decreased 17 percent from 1990 through 2019.
Emissions in 2019 decreased approximately 2 percent from 2018 levels while, in general, emissions from glass
production have remained relatively constant over the time series with some fluctuations since 1990. In general,
these fluctuations were related to the behavior of the export market and the U.S. economy. Specifically, the
extended downturn in residential and commercial construction and automotive industries between 2008 and 2010
resulted in reduced consumption of glass products, causing a drop in global demand for limestone/dolomite and
soda ash and resulting in lower emissions. Furthermore, the glass container sector is one of the leading soda ash
consuming sectors in the United States. Some commercial food and beverage package manufacturers are shifting
from glass containers towards lighter and more cost-effective polyethylene terephthalate (PET) based containers,
putting downward pressure on domestic consumption of soda ash (USGS 1995 through 2015b).
Table 4-11: CO2 Emissions from Glass Production (MMT CO2 Eq. and kt)
Year
MMTCOz Eq.
kt
1990
1.5
1,535
2005
1.9
1,928
2015
1.3
1,299
2016
1.2
1,249
2017
1.3
1,296
2018
1.3
1,305
2019
1.3
1,280
Note: Totals may not sum due to
independent rounding.
19 Excerpt from Glass & Glass Product Manufacturing Industry Profile, First Research. Available online at:
.
Industrial Processes and Product Use 4-21

-------
Methodology
Carbon dioxide emissions were calculated based on the 2006IPCC Guidelines Tier 3 method by multiplying the
quantity of input carbonates (limestone, dolomite, and soda ash) by the carbonate-based emission factor (in
metric tons COVmetric ton carbonate): limestone, 0.43971; dolomite, 0.47732; and soda ash, 0.41492.
In 1991, the U.S. Bureau of Mines, now known as the U.S. Geological Survey (USGS), began compiling production
and end use information through surveys of crushed stone manufacturers. Each year, limestone and dolomite
make up approximately 70% of the total crushed stone manufactured in the United States (USGS 1995 through
2016a). Crushed stone manufacturers provided different levels of detail in the survey responses, so information
was divided into three categories: (1) production by end-use, as reported by manufacturers (i.e., "specified"
production); (2) production reported by manufacturers without end-uses specified (i.e., "unspecified-reported"
production); and (3) estimated additional production by manufacturers who did not respond to the survey (i.e.,
"unspecified-estimated" production).
The "specified" production portion of the report provides limestone and dolomite consumption for glass
manufacturing. Large quantities of limestone and dolomite consumption are reported under the categories
"unspecified-reported" and "unspecified-estimated" as well, and a portion of this consumption is believed to be
limestone or dolomite used for glass manufacturing. The quantities listed under both "unspecified" categories
were allocated to glass manufacturing according to the percentage of "specified" limestone or dolomite consumed
for glass manufacturing end-use for that year.20
During 1990 and 1992, the U.S. Bureau of Mines did not conduct a detailed survey of limestone and dolomite
consumption by end-use. Therefore, data on consumption by end use for 1990 was estimated by applying the 1991
ratios of total limestone and dolomite consumption by end use to total 1990 limestone and dolomite consumption
values. Similarly, the 1992 consumption figures were approximated by applying an average of the 1991 and 1993
ratios of total limestone and dolomite consumption by end use to the 1992 total limestone and dolomite
consumption values.
For 1990 through 1993, consumption data of limestone and dolomite used for glass manufacturing were obtained
from the U.S. Bureau of Mines (1991 and 1993a), For 1994 through 2018, consumption data of limestone and
dolomite used for glass manufacturing were obtained from the USGS Minerals Yearbook: Crushed Stone Annual
Report (1995 through 2016a), and 2018 preliminary data from the USGS Crushed Stone Commodity Expert (Willett
2020a). The total limestone and dolomite used for glass manufacturing was determined in the same manner as
described for 1991 above. For 2019, consumption data for limestone and dolomite used for glass manufacturing
were not available at the time of publication, so 2018 values were used as proxy.
Each year the USGS withholds data on certain limestone and dolomite end-uses due to confidentiality agreements
regarding company proprietary data. For the purposes of this analysis, emissive end-uses that contained withheld
data were estimated using one of the following techniques: (1) the value for all the withheld data points for
limestone or dolomite use was distributed evenly to all withheld end-uses; or (2) the average percent of total
limestone or dolomite for the withheld end-use in the preceding and succeeding years.
For 1990 through 2019, consumption data for soda ash used for glass manufacturing were obtained from the U.S.
Bureau of Mines (1991 and 1993a), the USGS Minerals Yearbook: Soda Ash Annual Report (1995 through 2015b)
(USGS 1995 through 2015b), and USGS Mineral Industry Surveys for Soda Ash in April 2020 (USGS 2020).
Based on the 2019 reported data, the estimated distribution of soda ash consumption for glass production
compared to total domestic soda ash consumption is 47 percent (USGS 2020). Emissions from soda ash production
are reported in 4.12 Soda Ash Production (CRF Source Category 2B7).
20 This approach was recommended by USGS.
4-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-12: Limestone, Dolomite, and Soda Ash Consumption Used in Glass Production (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Limestone
430
920
699
472
720
818
817
Dolomite
59
541
0
0
0
0
0
Soda Ash
3,177
3,050
2,390
2,510
2,360
2,280
2,220
Total
3,666
4,511
3,089
2,982
3,080
3,098
3,037
Uncertainty and Time-Series Consistency
The uncertainty levels presented in this section arise in part due to variations in the chemical composition of
limestone used in glass production. In addition to calcium carbonate, limestone may contain smaller amounts of
magnesia, silica, and sulfur, among other minerals (potassium carbonate, strontium carbonate and barium
carbonate, and dead burned dolomite). Similarly, the quality of the limestone (and mix of carbonates) used for
glass manufacturing will depend on the type of glass being manufactured.
The estimates below also account for uncertainty associated with activity data. Large fluctuations in reported
consumption exist, reflecting year-to-year changes in the number of survey responders. The uncertainty resulting
from a shifting survey population is exacerbated by the gaps in the time series of reports. The accuracy of
distribution by end use is also uncertain because this value is reported by the manufacturer of the input
carbonates (limestone, dolomite and soda ash) and not the end user. For 2019, there has been no reported
consumption of dolomite for glass manufacturing. These data have been reported to USGS by dolomite
manufacturers and not end-users (i.e., glass manufacturers). There is a high uncertainty associated with this
estimate, as dolomite is a major raw material consumed in glass production. Additionally, there is significant
inherent uncertainty associated with estimating withheld data points for specific end uses of limestone and
dolomite. The uncertainty of the estimates for limestone and dolomite used in glass making is especially high.
Lastly, much of the limestone consumed in the United States is reported as "other unspecified uses;" therefore, it
is difficult to accurately allocate this unspecified quantity to the correct end-uses. Further research is needed into
alternate and more complete sources of data on carbonate-based raw material consumption by the glass industry.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-13. In 2019, glass
production C02 emissions were estimated to be between 1.2 and 1.3 MMT C02 Eq. at the 95 percent confidence
level. This indicates a range of approximately 4 percent below and 4 percent above the emission estimate of 1.3
MMT C02 Eq.
Table 4-13: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Glass
Production (MMT CO2 Eq. and Percent)
Source
2019 Emission Estimate
Gas
Uncertainty Range Relative to Emission Estimate3
(MMTC02 Eq.)
(MMT C02 Eq.)
(%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
Glass Production
C02 1.3
1.2 1.3
-4% +4%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
Industrial Processes and Product Use 4-23

-------
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
Recalculations Discussion
For the current Inventory, updated USGS data on limestone and dolomite consumption became available for 2016
and 2018. The revised values resulted in increased emissions estimates for the years 2016 (increase of 0.6 percent)
and 2018 (increase of 1.7 percent), compared to the previous Inventory.
Planned Improvements
As noted in the prior annual publications of this report, current publicly available activity data shows consumption
of only limestone and soda ash for glass manufacturing. While limestone and soda ash are the predominant
carbonates used in glass manufacturing, other carbonates are also consumed for glass manufacturing but in
smaller quantities. EPA has initiated review of available activity data on carbonate consumption by type in the
glass industry, reported annually since 2010 from EPA's Greenhouse Gas Reporting Program (GHGRP) as well as in
USGS publications. This is a long-term planned improvement.
EPA has initiated review of GHGRP data to help understand the completeness of emission estimates and facilitate
category-specific QC per Volume 1 of the 2006 IPCC Guidelines for the Glass Production source category. GHGRP
has an emission threshold for reporting from this industry, so the assessment will also consider the completeness
of carbonate consumption data for glass production in the United States. Particular attention will be made to
ensure time-series consistency of the emissions estimates presented in future Inventory reports, consistent with
IPCC and UNFCCC guidelines. This is required as the facility-level reporting data from EPA's GHGRP, with the
program's initial requirements for reporting of emissions in calendar year 2010, are not available for all inventory
years (i.e., 1990 through 2009) as required for this Inventory. In implementing improvements and integration of
data from GHGRP, EPA will rely on the latest guidance from the IPCC on the use of facility-level data in national
inventories.21 These planned improvements are ongoing, and EPA may also initiate research into other sources of
activity data for carbonate consumption by the glass industry.
4.4 Other Process Uses of Carbonates (CRF
Source Category 2A4)
Limestone (CaC03), dolomite (CaC03MgC03),22 and other carbonates such as soda ash, magnesite, and siderite are
basic materials used by a wide variety of industries, including construction, agriculture, chemical, metallurgy, glass
production, and environmental pollution control. This section addresses only limestone, dolomite, and soda ash use.
For industrial applications, carbonates such as limestone and dolomite are heated sufficiently enough to calcine the
material and generate C02 as a byproduct.
CaCO3 —> CaO + C02
21	See .
22	Limestone and dolomite are collectively referred to as limestone by the industry, and intermediate varieties are seldom
distinguished.
4-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
MgC03 —> MgO + C02
Examples of such applications include limestone used as a flux or purifier in metallurgical furnaces, as a sorbent in
flue gas desulfurization (FGD) systems for utility and industrial plants, and as a raw material for the production of
glass, lime, and cement. Emissions from limestone and dolomite used in other process sectors, such as cement, lime,
glass production, and iron and steel, are excluded from the Other Process Uses of Carbonates category and reported
under their respective source categories (e.g., Section 4.3, Glass Production). Emissions from soda ash production
are reported under Section 4.12 Soda Ash Production (CRF Source Category 2B7). Emissions from soda ash
consumption associated with glass manufacturing are reported under Section 4.3 Glass Production (CRF Source
Category 2A3). Emissions from the use of limestone and dolomite in liming of agricultural soils are included in the
Agriculture chapter under Liming (CRF Source Category 3G). Emissions from fuels consumed for energy purposes
during these processes are accounted for in the Energy chapter under Section 3.1 Fossil Fuel Combustion (CRF
Source Category 1A). Both lime (CaO) and limestone (CaC03) can be used as a sorbent for FGD systems. Emissions
from lime consumption for FGD systems are reported under Section 4.3 Lime Production (CRF Source Category
2A2).
Limestone and dolomite are widely distributed throughout the world in deposits of varying sizes and degrees of
purity. Large deposits of limestone occur in nearly every state in the United States, and significant quantities are
extracted for industrial applications. In 2016, the leading limestone producing states were Texas, Florida, Missouri,
Ohio, and Pennsylvania, which contributed 44 percent of the total U.S. output (USGS 2020a). Dolomite deposits are
found in the United States, Canada, Mexico, Europe, Africa, and Brazil. In the United States, the leading dolomite
producing states are Pennsylvania, New York, and Utah which currently contribute more than a third of the total
U.S. output (USGS 2020a). Internationally, two types of soda ash are produced: natural and synthetic. In 2017, 93
percent of the global soda ash production came from China, the United States, Russia, Germany, India, Turkey,
Poland, and France. The United States only produces natural soda ash and only in two states: Wyoming and
California (USGS 2020b).
In 2019,13,779 kilotons of limestone, 2,066 kt of dolomite, and 2,497 kt of soda ash were consumed for these
emissive applications, excluding glass manufacturing (Willett 2020, USGS 2020b). Usage of limestone, dolomite and
soda ash resulted in aggregate C02 emissions of 7.5 MMT C02 Eq. (7,457 kt) (see Table 4-14 and Table 4-15).
Limestone and dolomite consumption data were not available for 2019, so 2018 data were used as a proxy. The
2018 and 2019 emissions decreased 25 percent compared to 2017, primarily as a result of decreased limestone
consumption attributed to sulfur oxide removal usage for FGD systems. Overall emissions have increased 18
percent from 1990 through 2019.
Table 4-14: CO2 Emissions from Other Process Uses of Carbonates (MMT CO2 Eq.)
Other
Flux	Magnesium	Soda Ash	Miscellaneous
Year Stone FGD Production	Consumption3	Uses'5	Total
1990 2Ł	1A	ol	1A	oi	63~
2005 2.6	3.0	0.0	1.3	0.7	7.6
2015
2.9
7.3
0.0
1.1
0.9
12.2
2016
2.6
6.2
0.0
1.1
1.1
11.0
2017
2.4
5.6
0.0
1.1
0.8
9.9
2018
2.8
2.2
0.0
1.1
1.4
7.5
2019
2.8
2.2
0.0
1.0
1.4
7.5
Note: Totals may not sum due to independent rounding.
a Soda ash consumption not associated with glass manufacturing.
b "Other miscellaneous uses" include chemical stone, mine dusting or acid water treatment, acid
neutralization, and sugar refining.
Industrial Processes and Product Use 4-25

-------
Table 4-15: CO2 Emissions from Other Process Uses of Carbonates (kt)
Other
Magnesium	Soda Ash Miscellaneous
Year Flux Stone FGD Production Consumption3	Uses'5	Total
1990	2,592	1,432	64	1,390	819	6,297
2005	2,649	2,973	0	1,305	718	7,644
2015
2,901
7,335
0
1,075
871
12,182
2016
2,585
6,164
0
1,082
1,140
10,972
2017
2,441
5,598
0
1,058
835
9,933
2018
2,800
2,233
0
1,069
1,367
7,469
2019
2,821
2,233
0
1,036
1,367
7,457
Note: Totals may not sum due to independent rounding
a Soda ash consumption not associated with glass manufacturing.
b "Other miscellaneous uses" include chemical stone, mine dusting or acid water treatment, acid
neutralization, and sugar refining.
IViethiMlGiGgf
Carbon dioxide emissions were calculated based on the 2006IPCC Guidelines Tier 2 method by multiplying the
quantity of limestone or dolomite consumed by the emission factor for limestone or dolomite calcination,
respectively: 0.43971 metric ton C02/metric ton carbonate for limestone and 0.47732 metric ton C02/metric ton
carbonate for dolomite.23 This methodology was used for flux stone, flue gas desulfurization systems, chemical
stone, mine dusting or acid water treatment, acid neutralization, and sugar refining. Flux stone used during the
production of iron and steel was deducted from the Other Process Uses of Carbonates source category estimate
and attributed to the Iron and Steel Production source category estimate. Similarly, limestone and dolomite
consumption for glass manufacturing, cement, and lime manufacturing are excluded from this category and
attributed to their respective categories.
Historically, the production of magnesium metal was the only other significant use of limestone and dolomite that
produced C02 emissions. At the end of 2001, the sole magnesium production plant operating in the United States
that produced magnesium metal using a dolomitic process that resulted in the release of C02 emissions ceased its
operations (USGS 1995b through 2020).
Consumption data for 1990 through 2018 of limestone and dolomite used for flux stone, flue gas desulfurization
systems, chemical stone, mine dusting or acid water treatment, acid neutralization, and sugar refining (see Table
4-16) were obtained from the U.S. Geological Survey (USGS) Minerals Yearbook: Crushed Stone Annual Report
(1995a through 2017, 2020a, 2020c), preliminary data for 2018 from USGS Crushed Stone Commodity Expert
(Willett 2020), American Iron and Steel Institute limestone and dolomite consumption data (AISI 2018 through
2020), and the U.S. Bureau of Mines (1991 and 1993a), which are reported to the nearest ton. For 2019, no data
on limestone and dolomite consumption were available at the time of publication, so 2018 values were used as a
proxy for these values. The production capacity data for 1990 through 2001 of dolomitic magnesium metal also
came from the USGS (1995b through 2002) and the U.S. Bureau of Mines (1990 through 1993b). During 1990 and
1992, the USGS did not conduct a detailed survey of limestone and dolomite consumption by end-use. Therefore,
data on consumption by end use for 1990 was estimated by applying the 1991 ratios of total limestone and
dolomite consumption by end use to total 1990 limestone and dolomite consumption values. Similarly, the 1992
consumption figures were approximated by applying an average of the 1991 and 1993 ratios of total limestone and
dolomite use by end uses to the 1992 total values.
23 2006 IPCC Guidelines, Volume 3: Chapter 2, Table 2.1.
4-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
In 1991, the U.S. Bureau of Mines, now known as the USGS, began compiling production and end use information
through surveys of crushed stone manufacturers. Manufacturers, provided different levels of detail in survey
responses, so information was divided into three categories: (1) production by end-use, as reported by
manufacturers (i.e., "specified" production); (2) production reported by manufacturers without end-uses specified
(i.e., "unspecified-reported" production); and (3) estimated additional production by manufacturers who did not
respond to the survey (i.e., "unspecified-estimated" production). Additionally, each year the USGS withholds data
on certain limestone and dolomite end-uses due to confidentiality agreements regarding company proprietary
data. For the purposes of this analysis, emissive end-uses that contained withheld data were estimated using one
of the following techniques: (1) the value for all the withheld data points for limestone or dolomite use was
distributed evenly to all withheld end-uses; (2) the average percent of total limestone or dolomite for the withheld
end-use in the preceding and succeeding years; or (3) the average fraction of total limestone or dolomite for the
end-use over the entire time period.
There is a large quantity of crushed stone reported to the USGS under the category "unspecified uses." A portion
of this consumption is believed to be limestone or dolomite used for emissive end uses. The quantity listed for
"unspecified uses" was, therefore, allocated to all other reported end-uses according to each end-use's fraction of
total consumption in that year.24
Table 4-16: Limestone and Dolomite Consumption (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Flux Stone
6,737
7,022
7,834
7,092
6,853
7,658
7,658
Limestone
5,804
3,165
4,590
4118
4,920
5,603
5,603
Dolomite
933
3,857
3,244
2,973
1,933
2,055
2,055
FGD
3,258
6,761
16,680
14,019
12,732
5,078
5,078
Other Miscellaneous Uses
1,835
1,632
1,982
2,592
1,900
3,108
3,108
Total
11,830
15,415
26,496
23,703
21,484
15,845
15,845
Once produced, most soda ash is consumed in chemical production, with minor amounts used in soap production,
pulp and paper, flue gas desulfurization, and water treatment (excluding soda ash consumption for glass
manufacturing). As soda ash is consumed for these purposes, additional C02 is usually emitted. In these
applications, it is assumed that one mole of carbon is released for every mole of soda ash used. Thus,
approximately 0.113 metric tons of carbon (or 0.415 metric tons of C02) are released for every metric ton of soda
ash consumed. The activity data for soda ash consumption for 1990 to 2019 (see Table 4-17) were obtained from
the U.S. Geological Survey (USGS) Minerals Yearbook for Soda Ash (1994 through 2015b) and USGS Mineral
Industry Surveys for Soda Ash (USGS 2017a, 2018, 2019, 2020). Soda ash consumption data were collected by the
USGS from voluntary surveys of the U.S. soda ash industry.
Table 4-17: Soda Ash Consumption Not Associated with Glass Manufacturing (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Soda Asha
3,351
3,144
2,592
2,608
2,550
2,576
2,497
Total
3,351
3,144
2,592
2,608
2,550
2,576
2,497
a Soda ash consumption is sales reported by producers which exclude imports. Historically, imported soda ash
is less than 1 percent of the total U.S. consumption (Kostick 2012).
Uncertainty and Time-Series Consistency
The uncertainty levels presented in this section account for uncertainty associated with activity data. Data on
limestone and dolomite consumption are collected by USGS through voluntary national surveys. USGS contacts the
24 This approach was recommended by USGS, the data collection agency.
Industrial Processes and Product Use 4-27

-------
mines (i.e., producers of various types of crushed stone) for annual sales data. Data on other carbonate
consumption are not readily available. The producers report the annual quantity sold to various end-users and
industry types. USGS estimates the historical response rate for the crushed stone survey to be approximately 70
percent, and the rest is estimated by USGS. Large fluctuations in reported consumption exist, reflecting year-to-
year changes in the number of survey responders. The uncertainty resulting from a shifting survey population is
exacerbated by the gaps in the time series of reports. The accuracy of distribution by end use is also uncertain
because this value is reported by the producer/mines and not the end user. Additionally, there is significant
inherent uncertainty associated with estimating withheld data points for specific end uses of limestone and
dolomite. Lastly, much of the limestone consumed in the United States is reported as "other unspecified uses;"
therefore, it is difficult to accurately allocate this unspecified quantity to the correct end-uses. EPA contacted the
USGS National Minerals Information Center Crushed Stone commodity expert to assess the current uncertainty
ranges associated with the limestone and dolomite consumption data compiled and published by USGS. During
this discussion, the expert confirmed that EPA's range of uncertainty was still reasonable (Willett 2017).
Uncertainty in the estimates also arises in part due to variations in the chemical composition of limestone. In
addition to calcium carbonate, limestone may contain smaller amounts of magnesia, silica, and sulfur, among
other minerals. The exact specifications for limestone or dolomite used as flux stone vary with the
pyrometallurgical process and the kind of ore processed.
For emissions from soda ash consumption, the primary source of uncertainty results from the fact that these
emissions are dependent upon the type of processing employed by each end-use. Specific emission factors for
each end-use are not available, so a Tier 1 default emission factor is used for all end-uses. Therefore, there is
uncertainty surrounding the emission factors from the consumption of soda ash. Additional uncertainty comes
from the reported consumption and allocation of consumption within sectors that is collected on a quarterly basis
by the USGS. Efforts have been made to categorize company sales within the correct end-use sector.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-18. Carbon dioxide
emissions from other process uses of carbonates in 2019 were estimated to be between 6.6 and 8.6 MMT C02 Eq.
at the 95 percent confidence level. This indicates a range of approximately 12 percent below and 15 percent above
the emission estimate of 7.5 MMT C02 Eq.
Table 4-18: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Other
Process Uses of Carbonates (MMT CO2 Eq. and Percent)


2019 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMTC02 Eq.)
(MMT C02
Eq.)
(%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Other Process Uses
of Carbonates
C02
7.5
6.6
8.6
-12% +15%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
4-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Recalculations Discussion
For the current Inventory, updated USGS data on limestone and dolomite consumption was available for 2016,
2017, and 2018, resulting in updated emissions estimates for those years. Compared to the previous Inventory,
emissions for 2016 increased by 4 percent (467 kt C02 Eq.), decreased by less than 1 percent (2 kt C02 Eq.) for
2017, and decreased by 25 percent (2,485 kt C02 Eq.) for 2018.
Planned Improvements
In response to comments received during previous Inventory reports from the UNFCCC, EPA has inquired to the
availability of ceramics and non-metallurgical magnesia data. The USGS notes that this data is not currently
reported by survey respondents. EPA continues to conduct outreach with other entities, but at this time, the
research has not yielded any alternative data on national levels of carbonates. This improvement remains ongoing,
and EPA plans to continue to update this Planned Improvements section in future reports as more information
becomes available.
EPA also plans to continue dialogue with USGS to assess uncertainty ranges for activity data used to estimate
emissions from other process use of carbonates. This planned improvement is currently planned as a medium-
term improvement.
4.5 Ammonia Production (CRF Source Category
2B1)	
Emissions of carbon dioxide (C02) occur during the production of synthetic ammonia (NH3), primarily through the
use of natural gas, petroleum coke, or naphtha as a feedstock. The natural gas-, naphtha-, and petroleum coke-
based processes produce C02 and hydrogen (H2), the latter of which is used in the production of ammonia. The
brine electrolysis process for production of ammonia does not lead to process-based C02 emissions. Due to
national circumstances, emissions from fuels consumed for energy purposes during the production of ammonia
are accounted for in the Energy chapter. More information on this approach can be found in the Methodology
section below.
Ammonia production requires a source of nitrogen (N) and hydrogen (H). Nitrogen is obtained from air through
liquid air distillation or an oxidative process where air is burnt and the residual nitrogen is recovered. In the United
States, the majority of ammonia is produced using a natural gas feedstock as the hydrogen source; however, one
synthetic ammonia production plant located in Kansas is producing ammonia from petroleum coke feedstock. In
some U.S. plants, some of the C02 produced by the process is captured and used to produce urea rather than
being emitted to the atmosphere. In 2019, there were 16 companies operating 35 ammonia producing facilities in
16 states. Approximately 60 percent of domestic ammonia production capacity is concentrated in the states of
Louisiana, Oklahoma, and Texas (USGS 2020).
There are five principal process steps in synthetic ammonia production from natural gas feedstock. The primary
reforming step converts methane (CH4) to C02, carbon monoxide (CO), and hydrogen (H2) in the presence of a
catalyst. Only 30 to 40 percent of the CH4 feedstock to the primary reformer is converted to CO and C02 in this
step of the process. The secondary reforming step converts the remaining CH4 feedstock to CO and C02. The CO in
the process gas from the secondary reforming step (representing approximately 15 percent of the process gas) is
converted to C02 in the presence of a catalyst, water, and air in the shift conversion step. Carbon dioxide is
removed from the process gas by the shift conversion process, and the H2 is combined with the nitrogen (N2) gas in
the process gas during the ammonia synthesis step to produce ammonia. The C02 is included in a waste gas stream
with other process impurities and is absorbed by a scrubber solution. In regenerating the scrubber solution, C02 is
released from the solution.
Industrial Processes and Product Use 4-29

-------
The conversion process for conventional steam reforming of CH4, including the primary and secondary reforming
and the shift conversion processes, is approximately as follows:
0.88C7/4 +1.26Air +1.24H20 0.88C02 + N2 +3H2
N2 +3H2 -> 2NH3
To produce synthetic ammonia from petroleum coke, the petroleum coke is gasified and converted to C02 and H2.
These gases are separated, and the H2 is used as a feedstock to the ammonia production process, where it is
reacted with N2 to form ammonia.
Not all of the C02 produced during the production of ammonia is emitted directly to the atmosphere. Some of the
ammonia and some of the C02 produced by the synthetic ammonia process are used as raw materials in the
production of urea [CO(NH2)2], which has a variety of agricultural and industrial applications.
The chemical reaction that produces urea is:
2nh3+ C02 -> NH2COONH4 -> CO(NH2)2 +h2o
Only the C02 emitted directly to the atmosphere from the synthetic ammonia production process is accounted for
in determining emissions from ammonia production. The C02 that is captured during the ammonia production
process and used to produce urea does not contribute to the C02 emission estimates for ammonia production
presented in this section. Instead, C02 emissions resulting from the consumption of urea are attributed to the urea
consumption or urea application source category (under the assumption that the carbon stored in the urea during
its manufacture is released into the environment during its consumption or application). Emissions of C02 resulting
from agricultural applications of urea are accounted for in Section 5.6 Urea Fertilization (CRF Source Category 3H)
of the Agriculture chapter. Emissions of C02 resulting from non-agricultural applications of urea (e.g., use as a
feedstock in chemical production processes) are accounted for in Section 4.6 Urea Consumption for Non-
Agricultural Purposes of this chapter.
Total emissions of C02 from ammonia production in 2019 were 12.3 MMT C02 Eq. (12,272 kt), and are summarized
in Table 4-19 and Table 4-20. Ammonia production relies on natural gas as both a feedstock and a fuel, and as
such, market fluctuations and volatility in natural gas prices affect the production of ammonia. Since 1990,
emissions from ammonia production have decreased by about 6 percent. Emissions in 2019 increased by about 1
percent from the 2018 levels. Emissions from ammonia production have increased steadily since 2016, due to the
addition of new ammonia production facilities and new production units at existing facilities in 2016, 2017, and
2018. Agriculture continues to drive demand for nitrogen fertilizers and the need for new ammonia production
capacity (USGS 2020).
Table 4-19: CO2 Emissions from Ammonia Production (MMT CO2 Eq.)
Source
1990
2005
2015
2016
2017
2018
2019
Ammonia Production
13.0
9.2
10.6
10.2
11.1
12.2
12.3
Total	13.0	9.2	10.6 10.2 11.1 12.2 12.3
Table 4-20: CO2 Emissions from Ammonia Production (kt)
Source
1990
2005
2015
2016
2017
2018
2019
Ammonia Production
13,047
9,177
10,616
10,245
11,112
12,163
12,272
Total	13,047 9,177	10,616 10,245 11,112 12,163 12,272
Methodology
For this Inventory, C02 emissions from the production of synthetic ammonia from natural gas feedstock are
estimated using a country-specific approach modified from the 2006IPCC Guidelines (IPCC 2006) Tier 1 and 2
methods. In the country-specific approach, emissions are not based on total fuel requirement per the 2006 IPCC
4-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Guidelines due to data disaggregation limitations of energy statistics provided by the Energy Information
Administration (EIA). Data on total fuel use (including fuel used for ammonia feedstock and fuel used for energy)
for ammonia production are not known in the U.S. EIA does not provide data broken out by industrial category,
only at the broad industry sector level. To estimate emissions, a country-specific emission factor is developed and
applied to national ammonia production to estimate ammonia-production emissions from feedstock fuel use.
Emissions from fuel used for energy at ammonia plants are included in the overall EIA Industrial sector energy use
and accounted for in the Energy Chapter.
The country-specific approach uses a C02 emission factor of 1.2 metric tons C02/metric ton NH3, which is
published by the European Fertilizer Manufacturers Association (EFMA) and is based on natural gas-based
ammonia production technologies that are similar to those employed in the United States (EFMA 2000a). The
EFMA reported an emission factor range of 1.15 to 1.30 metric tons C02 per metric ton NH3, with 1.2 metric tons
C02 per metric ton NH3 as a typical value (EFMA 2000a). Technologies (e.g., catalytic reforming process, etc.)
associated with this factor are found to closely resemble those employed in the United States for use of natural gas
as a feedstock. The EFMA reference also indicates that more than 99 percent of the CH4 feedstock to the catalytic
reforming process is ultimately converted to C02. This country-specific approach is compatible with the 2006IPCC
Guidelines as it is based on the same scientific approach that the carbon in the fuel used to produce ammonia is
released as C02. The C02 emission factor is applied to the percent of total annual domestic ammonia production
from natural gas feedstock.
Emissions of C02 from ammonia production are then adjusted to account for the use of some of the C02 produced
from ammonia production as a raw material in the production of urea. The C02 emissions reported for ammonia
production are reduced by a factor of 0.733 multiplied by total annual domestic urea production. This corresponds
to a stoichiometric CO^urea factor of 44/60, assuming complete conversion of ammonia (NH3) and C02 to urea
(IPCC 2006; EFMA 2000b).
All synthetic ammonia production and subsequent urea production are assumed to be from the same process-
conventional catalytic reforming of natural gas feedstock, with the exception of ammonia production from
petroleum coke feedstock at one plant located in Kansas. Annual ammonia and urea production are shown in
Table 4-21.
The implied C02 emission factor for total ammonia production is a combination of the emission factors for
ammonia production from natural gas and from petroleum coke. Changes in the relative production of ammonia
from natural gas and petroleum coke will impact overall emissions and emissions per ton of total ammonia
produced. For example, between 2000 and 2001 there were increases in the amount of ammonia produced from
petroleum coke which caused increases in the implied emission factor across those years.
In previous Inventories, the C02 emission factor of 3.57 metric tons C02 per metric ton NH3for the petroleum coke
feedstock process (Bark 2004) was applied to the percent of total annual domestic ammonia production from
petroleum coke feedstock. Beginning with this Inventory, the C02 emission factor for petroleum coke feedstock
was updated to 3.52 metric tons of C02 per metric ton of NH3. The updated emission factor is based on an average
of the ratio of ammonia production from petroleum coke for years 2010 through 2015 (ACC 2020) and the facility-
specific C02 emissions from the one ammonia production plant located in Kansas that is manufacturing ammonia
from petroleum coke feedstock for years 2010 through 2015 (GHGRP 2020). Ammonia and urea are assumed to be
manufactured in the same manufacturing complex, as both the raw materials needed for urea production are
produced by the ammonia production process.
As a result of further examining the large increase in the amount of ammonia produced from petroleum coke
between 2015 and 2016, another methodology change has been made for this Inventory. The amount of ammonia
produced from petroleum coke changed significantly in 2016 because the parent company, CVR Energy, acquired a
second plant that uses natural gas as a feedstock. Therefore, the amount of ammonia production reported by CVR
Energy is no longer specific to the use of petroleum coke as a feedstock.
To correct this, beginning in 2016, the amount of C02from the ammonia production plant located in Kansas that is
manufacturing ammonia from petroleum coke feedstock (as reported under GHGRP 2020) is now being used,
Industrial Processes and Product Use 4-31

-------
along with the emission factor of 3.52 metric tons of C02 per metric ton of NH3 to back-calculate the amount of
ammonia produced through the use of petroleum coke as feedstock.
The consumption of natural gas and petroleum coke as fossil fuel feedstocks for NH3 production are adjusted for
within the Energy chapter as these fuels were consumed during non-energy related activities. More information on
this methodology is described in Annex 2.1, Methodology for Estimating Emissions of C02from Fossil Fuel
Combustion. See the Planned Improvements section on improvements of reporting fuel and feedstock C02
emissions utilizing EPA's GHGRP data to improve consistency with 2006IPCC Guidelines.
The total ammonia production data for 2011 through 2019 were obtained from American Chemistry Council (ACC
2020). For years before 2011, ammonia production data (see Table 4-21) were obtained from Coffeyville Resources
(Coffeyville 2005, 2006, 2007a, 2007b, 2009, 2010, 2011, and 2012) and the Census Bureau of the U.S. Department
of Commerce (U.S. Census Bureau 1991 through 1994,1998 through 2011) as reported in Current Industrial
Reports Fertilizer Materials and Related Products annual and quarterly reports. Urea-ammonia nitrate production
from petroleum coke for years through 2011 was obtained from Coffeyville Resources (Coffeyville 2005, 2006,
2007a, 2007b, 2009, 2010, 2011, and 2012), and from CVR Energy, Inc. Annual Report (CVR 2012 through 2015) for
2012 through 2015. Urea production data for 1990 through 2008 were obtained from the Minerals Yearbook:
Nitrogen (USGS 1994 through 2009). Urea production data for 2009 through 2010 were obtained from the U.S.
Census Bureau (U.S. Census Bureau 2010 and 2011). The U.S. Census Bureau ceased collection of urea production
statistics in 2011. Urea production values for the years 2011 through 2019 utilize GHGRP data (EPA 2018; EPA
2020).
Table 4-21: Ammonia Production, Recovered CO2 Consumed for Urea Production, and Urea
Production (kt)


Total C02 Consumption

Year
Ammonia Production
for Urea Production
Urea Production
1990
15,425
5,463
7,450
2005
10,143
3,865
5,270
2015
11,765
4,312
5,880
2016
12,305
5,419
7,390
2017
14,070
6,622
9,030
2018
16,010
7,847
10,700
2019
16,410
8,360
11,400
Uncertainty and Time-Series Consistency
The uncertainties presented in this section are primarily due to how accurately the emission factor used represents
an average across all ammonia plants using natural gas feedstock. Uncertainties are also associated with ammonia
production estimates and the assumption that all ammonia production and subsequent urea production was from
the same process—conventional catalytic reforming of natural gas feedstock, with the exception of one ammonia
production plant located in Kansas that is manufacturing ammonia from petroleum coke feedstock. Uncertainty is
also associated with the representativeness of the emission factor used for the petroleum coke-based ammonia
process. It is also assumed that ammonia and urea are produced at co-located plants from the same natural gas
raw material. The uncertainty of the total urea production activity data, based on USGS Minerals Yearbook:
Nitrogen data, is a function of the reliability of reported production data and is influenced by the completeness of
the survey responses.
Recovery of C02 from ammonia production plants for purposes other than urea production (e.g., commercial sale,
etc.) has not been considered in estimating the C02 emissions from ammonia production, as data concerning the
disposition of recovered C02 are not available. Such recovery may or may not affect the overall estimate of C02
emissions depending upon the end use to which the recovered C02 is applied. Further research is required to
4-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
determine whether byproduct C02 is being recovered from other ammonia production plants for application to
end uses that are not accounted for elsewhere; however, for reporting purposes, C02 consumption for urea
production is provided in this chapter.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-22. Carbon dioxide
emissions from ammonia production in 2019 were estimated to be between 10.9 and 13.6 MMT C02 Eq. at the 95
percent confidence level. This indicates a range of approximately 11 percent below and 11 percent above the
emission estimate of 12.3 MMT C02 Eq.
Table 4-22: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Ammonia Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Ammonia Production
C02
12.3
10.9 13.6
-11% +11%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied to ammonia production emission
estimates consistent with the U.S. Inventory QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006
IPCC Guidelines as described in the introduction of the IPPU chapter (see Annex 8 for more details). More details
on the greenhouse gas calculation, monitoring and QA/QC methods applicable to ammonia facilities can be found
under Subpart G (Ammonia Production) of the regulation (40 CFR Part 98).25 EPA verifies annual facility-level
GHGRP reports through a multi-step process (e.g., combination of electronic checks and manual reviews) to
identify potential errors and ensure that data submitted to EPA are accurate, complete, and consistent.26 Based on
the results of the verification process, EPA follows up with facilities to resolve mistakes that may have occurred.
The post-submittals checks are consistent with a number of general and category-specific QC procedures, including
range checks, statistical checks, algorithm checks, and year-to-year checks of reported data and emissions.
More details on the greenhouse gas calculation, monitoring and QA/QC methods applicable to reporting of urea
produced at ammonia production facilities can be found under Section 4.6 Urea Consumption for Non-Agricultural
Purposes.
Recalculations Discussion
Recalculations of ammonia emissions were performed for the 1990 through 2018 portion of the time series as
described below.
For years 2000 through 2018, the C02 emission factor for petroleum coke feedstock was updated from 3.57 metric
tons of C02 per metric ton of NH3 (Bark 2004) to 3.52 metric tons of C02 per metric ton of NH3, as detailed in the
Methodology section.
25	See .
26	See .
Industrial Processes and Product Use 4-33

-------
For years 2016 through 2018, the methodologies for determining the amount of ammonia produced using
petroleum coke and for determining the amount of UAN produced using petroleum coke were changed, as
detailed in the Methodology section.
For the year 2018, the amount of ammonia production increased by 11 percent due to a correction to the value
used in the previous Inventory, based on the adjustments made in determining the amount of ammonia produced
using petroleum coke as mentioned in the Methodology section.
These changes resulted in recalculations of the estimated C02 emissions estimates shown in Table 4-19 and Table
4-20 for the 1990 through 2018 portion of the time series. For years 2000 through 2015, the values reported
decreased by less than 0.24 percent per year from the values reported in the previous Inventory report. For year
2016, the value decreased by 5 percent (593 kt C02); for year 2017, the value decreased by 16 percent (2,104 kt
C02), and for year 2018, the value decreased by 10 percent (1,369 kt C02).
Planned Improvements
Future improvements involve continuing to evaluate and analyze data reported under EPA's GHGRP to improve the
emission estimates for the Ammonia Production source category, in particular new data from updated reporting
requirements finalized in October of 2014 (79 FR 63750) and December 2016 (81 FR 89188),27 that include facility-
level ammonia production data and feedstock consumption. The data were first reported by facilities in 2018 and
available post-verification to assess in 2019 for use in future Inventories, if the data meet GHGRP CBI aggregation
criteria. The data are still being evaluated and will be incorporated in future Inventory reports, if possible.
Particular attention will be made to ensure time-series consistency of the emission estimates presented in future
Inventory reports, along with application of appropriate category-specific QC procedures consistent with IPCC and
UNFCCC guidelines. For example, data reported in 2018 will reflect activity in 2017 and may not be representative
of activity in prior years of the time series. This assessment is required as the new facility-1 eve I reporting data from
EPA's GHGRP associated with new requirements are only applicable starting with reporting of emissions in
calendar year 2017, and thus are not available for all inventory years (i.e., 1990 through 2016) as required for this
Inventory.
In implementing improvements and integration of data from EPA's GHGRP, the latest guidance from the IPCC on
the use of facility-level data in national inventories will be relied upon.28 Specifically, the planned improvements
include assessing the anticipated new data to update the emission factors to include both fuel and feedstock C02
emissions to improve consistency with 2006 IPCC Guidelines, in addition to reflecting C02 capture and storage
practices (beyond use of C02 for urea production). Methodologies will also be updated if additional ammonia
production plants are found to use hydrocarbons other than natural gas for ammonia production. Due to limited
resources and ongoing data collection efforts, this planned improvement is still in development and is not
incorporated into this Inventory. This is a long-term planned improvement.
4.6 Urea Consumption for Non-Agricultural
Purposes
Urea is produced using ammonia (NH3) and carbon dioxide (C02) as raw materials. All urea produced in the United
States is assumed to be produced at ammonia production facilities where both ammonia and C02 are generated.
There were 35 plants producing ammonia in the United States in 2019, with two additional plants sitting idle for
the entire year (USGS 2020).
27	See .
28	See .
4-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The chemical reaction that produces urea is:
2nh3+ C02 -> NH2COONH4 -> CO(NH2)2 +h2o
This section accounts for C02 emissions associated with urea consumed exclusively for non-agricultural purposes.
Emissions of C02 resulting from agricultural applications of urea are accounted for in Section 5.6 Urea Fertilization
(CRF Source Category 3H) of the Agriculture chapter.
The industrial applications of urea include its use in adhesives, binders, sealants, resins, fillers, analytical reagents,
catalysts, intermediates, solvents, dyestuffs, fragrances, deodorizers, flavoring agents, humectants and
dehydrating agents, formulation components, monomers, paint and coating additives, photosensitive agents, and
surface treatments agents. In addition, urea is used for abating nitrogen oxide (NOx) emissions from coal-fired
power plants and diesel transportation motors.
Emissions of C02 from urea consumed for non-agricultural purposes in 2019 were estimated to be 6.2 MMT C02
Eq. (6,222 kt), and are summarized in Table 4-23 and Table 4-24. Net C02 emissions from urea consumption for
non-agricultural purposes have increased by approximately 64 percent from 1990 to 2019. The increase in
emissions since 2018 can be attributed to an increase in production and consumption.
Table 4-23: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (MMT CO2
Eq.)
Source	1990	2005	2015 2016 2017 2018 2019
Urea Consumption	3.8	3.7	4.6 5.1 5.0 5.9 6.2
Total	3^8	3/7	4^6 SA eTo 5^9 6^2
Table 4-24: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (kt)
Source	1990	2005	2015 2016 2017 2018 2019
Urea Consumption	3,784	3,653	4,578 5,132 5,028 5,857 6,222
Total	3,784 3,653 4,578 5,132 5,028 5,857 6,222
Methodology
Emissions of C02 resulting from urea consumption for non-agricultural purposes are estimated by multiplying the
amount of urea consumed in the United States for non-agricultural purposes by a factor representing the amount
of C02 used as a raw material to produce the urea. This method is based on the assumption that all of the carbon
in urea is released into the environment as C02 during use, consistent with the Tier 1 method used to estimate
emissions from ammonia production in the 2006IPCC Guidelines (IPCC 2006) which states that the "C02 recovered
[from ammonia production] for downstream use can be estimated from the quantity of urea produced where C02
is estimated by multiplying urea production by 44/60, the stoichiometric ratio of C02 to urea."
The amount of urea consumed for non-agricultural purposes in the United States is estimated by deducting the
quantity of urea fertilizer applied to agricultural lands, which is obtained directly from the Agriculture chapter (see
Table 5-25), from the total domestic supply of urea as reported in Table 4-25. The domestic supply of urea is
estimated based on the amount of urea produced plus urea imports and minus urea exports. A factor of 0.733 tons
of C02 per ton of urea consumed is then applied to the resulting supply of urea for non-agricultural purposes to
estimate C02 emissions from the amount of urea consumed for non-agricultural purposes. The 0.733 tons of C02
per ton of urea emission factor is based on the stoichiometry of C in urea. This corresponds to a stoichiometric C02
to urea factor of 44/60, assuming complete conversion of C in urea to C02(IPCC 2006; EFMA 2000).
Urea production data for 1990 through 2008 were obtained from the U.S. Geological Survey (USGS) Minerals
Yearbook: Nitrogen (USGS 1994 through 2009a). Urea production data for 2009 through 2010 were obtained from
Industrial Processes and Product Use 4-35

-------
the U.S. Census Bureau (2011). The U.S. Census Bureau ceased collection of urea production statistics in 2011.
Starting with the Inventory report for the years 1990 through 2017, EPA began utilizing urea production data from
EPA's GHGRP to estimate emissions. Urea production values in the current Inventory report utilize GHGRP data for
the years 2011 through 2019 (EPA 2018, EPA 2020).
Urea import data for 2019 are not yet publicly available, so 2018 data have been used as a proxy. Urea import data
for 2013 to 2018 were obtained from the USGS Minerals Yearbook: Nitrogen (USGS 2019a). Urea import data for
2011 and 2012 were taken from U.S. Fertilizer Import/Exports from the United States Department of Agriculture
(USDA) Economic Research Service Data Sets (U.S. Department of Agriculture 2012). USDA suspended updates to
this data after 2012. Urea import data for the previous years were obtained from the U.S. Census Bureau Current
Industrial Reports Fertilizer Materials and Related Products annual and quarterly reports for 1997 through 2010
(U.S. Census Bureau 2001 through 2011), The Fertilizer Institute (TFI 2002) for 1993 through 1996, and the United
States International Trade Commission Interactive Tariff and Trade DataWeb (U.S. ITC 2002) for 1990 through 1992
(see Table 4-25).
Urea export data for 2019 are not yet publicly available, so 2018 data have been used as a proxy. Urea export data
for 2013 to 2018 were obtained from the USGS Minerals Yearbook: Nitrogen (USGS 2019a). Urea export data for
1990 through 2012 were taken from U.S. Fertilizer Import/Exports from USDA Economic Research Service Data Sets
(U.S. Department of Agriculture 2012). USDA suspended updates to this data after 2012.
Table 4-25: Urea Production, Urea Applied as Fertilizer, Urea Imports, and Urea Exports (kt)
Year
Urea
Production
Urea Applied
as Fertilizer
Urea
Imports
Urea
Exports
Urea Consumed for Non-
Agricultural Purposes
1990
7,450
3,296
1,860
854
5,160
2005
5,270
4,779
5,026
536
4,981
2015
5,880
6,447
7,190
380
6,243
2016
7,390
6,651
6,580
321
6,998
2017
9,030
6,888
5,510
795
6,857
2018
10,700
7,080
5,110
743
7,987
2019
11,400
7,283
5,110
743
8,484
Uncertainty and Time-Series Consistency
There is limited publicly available data on the quantities of urea produced and consumed for non-agricultural
purposes. Therefore, the amount of urea used for non-agricultural purposes is estimated based on a balance that
relies on estimates of urea production, urea imports, urea exports, and the amount of urea used as fertilizer. The
primary uncertainties associated with this source category are associated with the accuracy of these estimates as
well as the fact that each estimate is obtained from a different data source. Because urea production estimates are
no longer available from the USGS, there is additional uncertainty associated with urea produced beginning in
2011. There is also uncertainty associated with the assumption that all of the carbon in urea is released into the
environment as C02 during use.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-26. Carbon dioxide
emissions associated with urea consumption for non-agricultural purposes during 2019 were estimated to be
between 5.4 and 7.1 MMT C02 Eq. at the 95 percent confidence level. This indicates a range of approximately 13
percent below and 14 percent above the emission estimate of 6.2 MMT C02 Eq.
Table 4-26: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Urea
Consumption for Non-Agricultural Purposes (MMT CO2 Eq. and Percent)
2019 Emission Estimate Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.)	(MMT CP2 Eq.)	(%)
4-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Urea Consumption
for Non-Agricultural
Purposes
C02
6.2
5.4
7.1
-13%
+14%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
More details on the greenhouse gas calculation, monitoring and QA/QC methods applicable to reporting of urea
production occurring at ammonia facilities can be found under Subpart G (Ammonia Manufacturing) of the
regulation (40 CFR Part 98).29 EPA verifies annual facility-level GHGRP reports through a multi-step process (e.g.,
combination of electronic checks and manual reviews) to identify potential errors and ensure that data submitted
to EPA are accurate, complete, and consistent.30 Based on the results of the verification process, EPA follows up
with facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with a
number of general and category-specific QC procedures, including range checks, statistical checks, algorithm
checks, and year-to-year checks of reported data and emissions. EPA also conducts QA checks of GHGRP reported
urea production data against external datasets including the USGS Minerals Yearbook data. The comparison shows
consistent trends in urea production over time.
Recalculations Discussion
Based on updated urea production data from EPA's GHGRP for 2017 and 2018, updated urea imports from USGS
for 2018, and updated urea exports from USGS for 2017 and 2018, recalculations were performed for these two
years. Compared to the previous Inventory, C02 emissions from urea consumption for non-agricultural purposes
increased by 33 percent (1,717 kt C02) for 2017 and 61 percent (3,039 kt C02) for 2018, due to large increases in
urea production for both years.
4.7 Nitric Acid Production (CRF Source
Category 2B2)
Nitrous oxide (N20) is emitted during the production of nitric acid (HN03), an inorganic compound used primarily
to make synthetic commercial fertilizers. Nitric acid is also a major component in the production of adipic acid—a
feedstock for nylon—and explosives. Virtually all of the nitric acid produced in the United States is manufactured
by the high-temperature catalytic oxidation of ammonia (EPA 1998). There are two different nitric acid production
methods: weak nitric acid and high-strength nitric acid. The first method utilizes oxidation, condensation, and
absorption to produce nitric acid at concentrations between 30 and 70 percent nitric acid. High-strength acid (90
29	See .
30	See .
Industrial Processes and Product Use 4-37

-------
percent or greater nitric acid) can be produced from dehydrating, bleaching, condensing, and absorption of the
weak nitric acid. Most U.S. plants were built between 1960 and 2000. As of 2019, there were 31 active nitric acid
production plants, including one high-strength nitric acid production plant in the United States (EPA 2010; EPA
2020).
The basic process technology for producing nitric acid has not changed significantly over time. During this process,
N20 is formed as a byproduct and is released from reactor vents into the atmosphere. Emissions from fuels
consumed for energy purposes during the production of nitric acid are included in the Energy chapter.
Nitric acid is made from the reaction of ammonia (NH3) with oxygen (02) in two stages. The overall reaction is:
4NH3 + 802 -> 4HN03 + 4H2
Currently, the nitric acid industry controls emissions of NO and N02 (i.e., NOx). As such, the industry in the United
States uses a combination of non-selective catalytic reduction (NSCR) and selective catalytic reduction (SCR)
technologies. In the process of destroying NOx, NSCR systems are also very effective at destroying N20. NSCR units,
however, are generally not preferred in modern plants because of high energy costs and associated high gas
temperatures. NSCR systems were installed in nitric plants built between 1971 and 1977 and are used in
approximately one-third of the weak acid production plants. U.S. facilities are using both tertiary (i.e., NSCR) and
secondary controls (i.e., alternate catalysts).
Nitrous oxide emissions from this source were estimated to be 10.0 MMT C02 Eq. (34 kt of N20) in 2019 (see Table
4-27). Emissions from nitric acid production have decreased by 18 percent since 1990, while production has
increased by 12 percent over the same time period. Emissions have decreased by 31 percent since 1997, the
highest year of production in the time series.
Table 4-27: N2O Emissions from Nitric Acid Production (MMT CO2 Eq. and kt N2O)
Year
MMTCOz Eq.
kt N20
1990
12.1
41
2005
11.3
38
2015
11.6
39
2016
10.1
34
2017
9.3
31
2018
9.6
32
2019
10.0
34
Methodology
Emissions of N20 were calculated using the estimation methods provided by the 2006IPCC Guidelines and a
country-specific method utilizing EPA's GHGRP. The 2006 IPCCGuidelinesTier 2 method was used to estimate
emissions from nitric acid production for 1990 through 2009, and a country-specific approach similar to the IPCC
Tier 3 method was used to estimate N20 emissions for 2010 through 2019.
2010 through 2019
Process N20 emissions and nitric acid production data were obtained directly from EPA's GHGRP for 2010 through
2019 by aggregating reported facility-level data (EPA 2020).
Since 2010, in the United States, all nitric acid facilities that produce weak nitric acid (30 to 70 percent) have been
required to report annual greenhouse gas emissions data to EPA as per the requirements of the GHGRP (Subpart
V). Beginning with 2018, the rule was changed to include facilities that produce nitric acid of any strength. The only
facility that produces high-strength nitric acid also produces weak nitric acid. All greenhouse gas emissions from
nitric acid production originate from the production of weak nitric acid.
4-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Process emissions and nitric acid production reported to the GHGRP provide complete estimates of greenhouse
gas emissions for the United States because there are no reporting thresholds. While facilities are allowed to stop
reporting to the GHGRP if the total reported emissions from nitric acid production are less than 25,000 metric tons
C02 Eq. per year for five consecutive years or less than 15,000 metric tons C02 Eq. per year for three consecutive
years, no facilities have stopped reporting as a result of these provisions.31 All nitric acid facilities are required to
calculate process emissions using a site-specific emission factor that is the average of the emission factor
determined through annual performance tests for each nitric acid train under typical operating conditions or by
directly measuring N20 emissions using monitoring equipment.32
Emissions from facilities vary from year to year, depending on the amount of nitric acid produced with and without
abatement technologies and other conditions affecting the site-specific emission factor. To maintain consistency
across the time series and with the rounding approaches taken by other data sets, GHGRP nitric acid data are
rounded for consistency and are shown in Table 4-28.
1990 through 2009
Using GHGRP data for 2010,33 country-specific N20 emission factors were calculated for nitric acid production with
abatement and without abatement (i.e., controlled and uncontrolled emission factors). The following 2010
emission factors were derived for production with abatement and without abatement: 3.3 kg N20/metric ton
HN03 produced at plants using abatement technologies (e.g., tertiary systems such as NSCR systems) and 5.99 kg
N20/metric ton HN03 produced at plants not equipped with abatement technology. Country-specific weighted
emission factors were derived by weighting these emission factors by percent production with abatement and
without abatement over time periods 1990 through 2008 and 2009. These weighted emission factors were used to
estimate N20 emissions from nitric acid production for years prior to the availability of GHGRP data (i.e., 1990
through 2008 and 2009). A separate weighted emission factor is included for 2009 due to data availability for that
year. At that time, EPA had initiated compilation of a nitric acid database to improve estimation of emissions from
this industry and obtained updated information on application of controls via review of permits and outreach with
facilities and trade associations. The research indicated recent installation of abatement technologies at additional
facilities.
Based on the available data, it was assumed that emission factors for 2010 would be more representative of
operating conditions in 1990 through 2009 than more recent years. Initial review of historical data indicates that
percent production with and without abatement can change over time and from year to year due to changes in
application of facility-level abatement technologies, maintenance of abatement technologies, and also due to plant
closures and start-ups (EPA 2012, 2013; Desai 2012; CAR 2013). The installation dates of N20 abatement
technologies are not known at most facilities, but it is assumed that facilities reporting abatement technology use
have had this technology installed and operational for the duration of the time series considered in this report
(especially NSCRs).
The country-specific weighted N20 emission factors were used in conjunction with annual production to estimate
N20 emissions for 1990 through 2009, using the following equations:
El — Pi X EFWelgfr):eCl:l
EFweighted,i =	X EFc) + (%PUnc,i X EFunc)\
31	See 40 CFR 98.2(i)(l) and 40 CFR 98.2(i)(2) for more information about these provisions.
32	Facilities must use standard methods, either EPA Method 320 or ASTM D6348-03 for annual performance tests and must
follow associated QA/QC procedures consistent during these performance test consistent with category-specific QC of direct
emission measurements.
33	National N20 process emissions, national production, and national share of nitric acid production with abatement and
without abatement technology was aggregated from the GHGRP facility-level data for 2010 to 2017 (i.e., percent production
with and without abatement).
Industrial Processes and Product Use 4-39

-------
where,
E,	= Annual N20 Emissions for year i (kg/yr)
Pi	= Annual nitric acid production for year i (metric tons HN03)
EFweighted.i	= Weighted N20 emission factor for year i (kg N20/metric ton HN03)
%Pc,i	= Percent national production of HN03 with N20 abatement technology (%)
EFC	= N20 emission factor, with abatement technology (kg N20/metric ton HN03)
%PUnc,i	= Percent national production of HN03 without N20 abatement technology (%)
EFunc	= N20 emission factor, without abatement technology (kg N20/metric ton HN03)
i	= year from 1990 through 2009
•	For 2009: Weighted N20 emission factor = 5.46 kg N20/metric ton HN03.
•	For 1990 through 2008: Weighted N20 emission factor = 5.66 kg N20/metric ton HN03.
Nitric acid production data for the United States for 1990 through 2009 were obtained from the U.S. Census
Bureau (U.S. Census Bureau 2008, 2009, 2010a, 2010b) (see Table 4-28). Publicly available information on plant-
level abatement technologies was used to estimate the shares of nitric acid production with and without
abatement for 2008 and 2009 (EPA 2012, 2013; Desai 2012; CAR 2013). EPA has previously conducted a review of
operating permits to obtain more current information due to the lack of publicly-available data on use of
abatement technologies for 1990 through 2007, as stated previously; therefore, the share of national production
with and without abatement for 2008 was assumed to be constant for 1990 through 2007.
Table 4-28: Nitric Acid Production (kt)
Year	kt
1990 7,200
2005 6,710
2015	7,210
2016	7,810
2017	7,780
2018	8,210
2019	8,080
Uncertainty and Time-Series Consistency
Uncertainty associated with the parameters used to estimate N20 emissions includes the share of U.S. nitric acid
production attributable to each emission abatement technology over the time series (especially prior to 2010), and
the associated emission factors applied to each abatement technology type. While some information has been
obtained through outreach with industry associations, limited information is available over the time series
(especially prior to 2010) for a variety of facility level variables, including plant-specific production levels, plant
production technology (e.g., low, high pressure, etc.), and abatement technology type, installation date of
abatement technology, and accurate destruction and removal efficiency rates. Production data prior to 2010 were
obtained from National Census Bureau, which does not provide uncertainty estimates with their data. Facilities
reporting to EPA's GHGRP must measure production using equipment and practices used for accounting purposes.
While emissions are often directly proportional to production, the emission factor for individual facilities can vary
significantly from year to year due to site-specific fluctuations in ambient temperature and humidity, catalyst age
and condition, nitric acid production process changes, the addition or removal of abatement technologies, and the
number of nitric acid trains at the facility. At this time, EPA does not estimate uncertainty of the aggregated
facility-level information. As noted in the QA/QC and verification section below, EPA verifies annual facility-1 eve I
reports through a multi-step process (e.g., combination of electronic checks and manual reviews by staff) to
identify potential errors and ensure that data submitted to EPA are accurate, complete, and consistent. The annual
4-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
production reported by each nitric acid facility under EPA's GHGRP and then aggregated to estimate national N20
emissions is assumed to have low uncertainty.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-29. Nitrous oxide
emissions from nitric acid production were estimated to be between 9.5 and 10.5 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 5 percent below to 5 percent above the 2019 emissions
estimate of 10.0 MMT C02 Eq.
Table 4-29: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Nitric
Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTC02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Nitric Acid Production
N20
10.0
9.5 10.5
-5% +5%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of the 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details). More details on the greenhouse gas calculation,
monitoring and QA/QC methods applicable to nitric acid facilities can be found under Subpart V: Nitric Acid
Production of the GHGRP regulation (40 CFR Part 98).34
The main QA/QC activities are related to annual performance testing, which must follow either EPA Method 320 or
ASTM D6348-03. EPA verifies annual facility-level GHGRP reports through a multi-step process that is tailored to
the Subpart (e.g., combination of electronic checks including range checks, statistical checks, algorithm checks,
year-to-year comparison checks, along with manual reviews) to identify potential errors and ensure that data
submitted to EPA are accurate, complete, and consistent. Based on the results of the verification process, EPA
follows up with facilities to resolve mistakes that may have occurred (EPA 2015).35 EPA's review of observed trends
noted that while emissions have generally mirrored production, in 2015 and 2019 nitric acid production decreased
compared to the previous year and emissions increased. While review is ongoing, based on feedback from the
verification process to date, these changes are due to facility-specific changes (e.g. in the nitric production process
and management of abatement equipment).
Recalculations Discussion
Recalculations of emissions from nitric acid production were performed for the 1990 through 2018 time series
when the GHGRP data for 2018 were released in November 2020. Previously, the 2017 value was used as proxy for
2018. The change resulted in recalculations of the estimated C02 emissions estimates shown in Table 4-19 and
Table 4-20. Compared to the previous Inventory, the emissions value for 2018 increased by 3 percent (0.3 MMT
CQ2 Eq.), and the nitric acid production value for 2018 increased by 6 percent (430 kt).
34	See Subpart V monitoring and reporting regulation .
35	See GHGRP Verification Factsheet .
Industrial Processes and Product Use 4-41

-------
Planned Improvements
Pending resources, EPA is considering both near-term and long-term improvement to estimates and associated
characterization of uncertainty. In the short-term, with 8 years of EPA's GHGRP data, EPA anticipates completing
updates of category-specific QC procedures to potentially also improve both qualitative and quantitative
uncertainty estimates. In the next Inventory, EPA anticipates including information from GHGRP facilities on the
installation date of any N20 abatement equipment, per revisions finalized in December 2016 to EPA's GHGRP. This
information will enable more accurate estimation of N20 emissions from nitric acid production over the time
series.
4.8 Adipic Acid Production (CRF Source
Category 2B3)
Adipic acid is produced through a two-stage process during which nitrous oxide (N20) is generated in the second
stage. Emissions from fuels consumed for energy purposes during the production of adipic acid are accounted for
in the Energy chapter. The first stage of manufacturing usually involves the oxidation of cyclohexane to form a
cyclohexanone/cyclohexanol mixture. The second stage involves oxidizing this mixture with nitric acid to produce
adipic acid. Nitrous oxide is generated as a byproduct of the nitric acid oxidation stage and is emitted in the waste
gas stream (Thiemens and Trogler 1991). The second stage is represented by the following chemical reaction:
('CH2)5CO{cyclohexanone) + (CH2)5CHOH(cyclohexanol) + wHN03
-» HOOC(CH2)4COOH(adipic acid) + xN20 + yH20
Process emissions from the production of adipic acid vary with the types of technologies and level of emission
controls employed by a facility. In 1990, two major adipic acid-producing plants had N20 abatement technologies
in place and, as of 1998, three major adipic acid production facilities had control systems in place (Reimer et al.
1999). In 2018, catalytic reduction, non-selective catalytic reduction (NSCR), and thermal reduction abatement
technologies were applied as N20 abatement measures at adipic acid facilities (EPA 2019, 2020).
Worldwide, only a few adipic acid plants exist. The United States, Europe, and China are the major producers, with
the United States accounting for the largest share of global adipic acid production capacity in recent years. In 2019,
the United States had two companies with a total of two adipic acid production facilities (one in Texas and one in
Florida), following the ceased operations of a third major production facility at the end of 2015 (EPA 2019, 2020).
Adipic acid is a white crystalline solid used in the manufacture of synthetic fibers, plastics, coatings, urethane
foams, elastomers, and synthetic lubricants. Commercially, it is the most important of the aliphatic dicarboxylic
acids, which are used to manufacture polyesters. Eighty-four percent of all adipic acid produced in the United
States is used in the production of nylon 6,6; 9 percent is used in the production of polyester polyols; 4 percent is
used in the production of plasticizers; and the remaining 4 percent is accounted for by other uses, including
unsaturated polyester resins and food applications (ICIS 2007). Food grade adipic acid is used to provide some
foods with a "tangy" flavor (Thiemens and Trogler 1991).
National adipic acid production has increased by approximately 7 percent over the period of 1990 through 2019, to
approximately 810,000 metric tons (ACC 2020). Nitrous oxide emissions from adipic acid production were
estimated to be 5.3 MMT C02 Eq. (18 kt N20) in 2019 (see Table 4-30). Over the period 1990 through 2019,
emissions have been reduced by 65 percent due to both the widespread installation of pollution control measures
in the late 1990s and plant idling in the late 2000s. The total emissions from adipic acid production decreased by
approximately 49 percent from GHGRP Reporting Year (RY) 2018 to RY2019 due to a significant change in
emissions from one facility. The facility confirmed that there was a decrease in adipic acid production and an
increase in the use of the N20 abatement device in RY2019, resulting in a large decrease in greenhouse gas
emissions (EPA 2019, 2020). As noted above, changes in control measures and abatement technologies at adipic
4-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
acid production facilities, including maintenance of equipment, can result in annual emission fluctuations. Little
additional information is available on drivers of trends in adipic acid production as it is not reported under EPA's
GHGRP.
Table 4-30: N2O Emissions from Adipic Acid Production (MMT CO2 Eq. and kt N2O)
Year
MMT CO? Eq.
kt N20
1990
15.2
51
2005
7.1
24
2015
2016
2017
2018
2019
4.3
7.0
7.4
10.3
5.3
14
23
25
35
18
Methodology
Emissions are estimated using both Tier 2 and Tier 3 methods consistent with the 2006IPCC Guidelines. Due to
confidential business information (CBI), plant names are not provided in this section. Therefore, the four adipic
acid-producing facilities that have operated over the time series will be referred to as Plants 1 through 4. Overall,
as noted above, the two currently operating facilities use catalytic reduction, NSCR and thermal reduction
abatement technologies.
All emission estimates for 2010 through 2019 were obtained through analysis of GHGRP data (EPA 2010 through
2020), which is consistent with the 2006 IPCC Guidelines Tier 3 method. Facility-level greenhouse gas emissions
data were obtained from EPA's GHGRP for the years 2010 through 2019 (EPA 2010 through 2020) and aggregated
to national N20 emissions. Consistent with IPCC Tier 3 methods, all adipic acid production facilities are required to
calculate emissions using a facility-specific emission factor developed through annual performance testing under
typical operating conditions or by directly measuring N20 emissions using monitoring equipment.36
For years 1990 through 2009, which were prior to EPA's GHGRP reporting, for both Plants 1 and 2, emission
estimates were obtained directly from the plant engineers and account for reductions due to control systems in
place at these plants during the time series. These prior estimates are considered CBI and hence are not published
(Desai 2010, 2011). These estimates were based on continuous process monitoring equipment installed at the two
facilities.
For Plant 4,1990 through 2009 N20 emissions were estimated using the following Tier 2 equation from the 2006
IPCC Guidelines:
36 Facilities must use standard methods, either EPA Method 320 or ASTM D6348-03 for annual performance testing, and must
follow associated QA/QC procedures during these performance tests consistent with category-specific QC of direct emission
measurements.
2010 through 2019
1990 through 2009
Eaa = Qaa x EFaa X (1 ~ [DF X UF])
where,
Industrial Processes and Product Use 4-43

-------
Eaa	=	N20 emissions from adipic acid production, metric tons
Qaa	=	Quantity of adipic acid produced, metric tons
EFaa	=	Emission factor, metric ton N20/metric ton adipic acid produced
DF	=	N20 destruction factor
UF	=	Abatement system utility factor
The adipic acid production is multiplied by an emission factor (i.e., N20 emitted per unit of adipic acid produced),
which has been estimated to be approximately 0.3 metric tons of N20 per metric ton of product (IPCC 2006). The
"N20 destruction factor" in the equation represents the percentage of N20 emissions that are destroyed by the
installed abatement technology. The "abatement system utility factor" represents the percentage of time that the
abatement equipment operates during the annual production period. Plant-specific production data for Plant 4
were obtained across the time series through personal communications (Desai 2010, 2011). The plant-specific
production data were then used for calculating emissions as described above.
For Plant 3, 2005 through 2009 emissions were obtained directly from the plant (Desai 2010, 2011). For 1990
through 2004, emissions were estimated using plant-specific production data and the IPCC factors as described
above for Plant 4. Plant-level adipic acid production for 1990 through 2003 was estimated by allocating national
adipic acid production data to the plant level using the ratio of known plant capacity to total national capacity for
all U.S. plants (ACC 2020; CMR 2001,1998; CW 1999; C&EN 1992 through 1995). For 2004, actual plant production
data were obtained and used for emission calculations (CW 2005).
Plant capacities for 1990 through 1994 were obtained from Chemical & Engineering News, "Facts and Figures" and
"Production of Top 50 Chemicals" (C&EN 1992 through 1995). Plant capacities for 1995 and 1996 were kept the
same as 1994 data. The 1997 plant capacities were taken from Chemical Market Reporter, "Chemical Profile: Adipic
Acid" (CMR 1998). The 1998 plant capacities for all four plants and 1999 plant capacities for three of the plants
were obtained from Chemical Week, Product Focus: Adipic Acid/Adiponitrile (CW 1999). Plant capacities for the
year 2000 for three of the plants were updated using Chemical Market Reporter, "Chemical Profile: Adipic Acid"
(CMR 2001). For 2001 through 2003, the plant capacities for three plants were held constant at year 2000
capacities. Plant capacity for 1999 to 2003 for the one remaining plant was kept the same as 1998.
National adipic acid production data (see Table 4-31) from 1990 through 2019 were obtained from the American
Chemistry Council (ACC 2020).
Table 4-31: Adipic Acid Production (kt)
Year	kt
1990 755
2005 865
2015	1,055
2016	860
2017	830
2018	825
2019	810
Uncertainty and Time-Series Consistency
Uncertainty associated with N20 emission estimates includes the methods used by companies to monitor and
estimate emissions. While some information has been obtained through outreach with facilities, limited
information is available over the time series on these methods, abatement technology destruction and removal
efficiency rates, and plant-specific production levels.
The results of this Approach 2 quantitative uncertainty analysis are summarized in
4-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-32. Nitrous oxide emissions from adipic acid production for 2019 were estimated to be between 5.0 and
5.5 MMT C02 Eq. at the 95 percent confidence level. These values indicate a range of approximately 5 percent
below to 5 percent above the 2019 emission estimate of 5.3 MMT C02 Eq.
Table 4-32: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Adipic
Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Adipic Acid Production
N20
5.3
5.0 5.5
-5% +5%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of the 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
More details on the greenhouse gas calculation, monitoring and QA/QC methods applicable to adipic acid facilities
can be found under Subpart E (Adipic Acid Production) of the GHGRP regulation (40 CFR Part 98).37 The main
QA/QC activities are related to annual performance testing, which must follow either EPA Method 320 or ASTM
D6348-03. EPA verifies annual facility-level GHGRP reports through a multi-step process (e.g., combination of
electronic checks and manual reviews) to identify potential errors and ensure that data submitted to EPA are
accurate, complete, and consistent (EPA 2015).38 Based on the results of the verification process, EPA follows up
with facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with a
number of general and category-specific QC procedures, including range checks, statistical checks, algorithm
checks, and year-to-year comparisons of reported data.
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
EPA plans to review GHGRP facility reported information on the date of abatement technology installation in order
to better reflect trends and changes in emissions abatement within the industry across the time series. See more
detail on similar planned improvements within Section 4.7 on Nitric Acid Production presented above.
37	See .
38	See .
Industrial Processes and Product Use 4-45

-------
4.9 Caprolactam, Glyoxal and Glyoxylic Acid
Production (CRF Source Category 2B4)
Caprolactam
Caprolactam (C6HnNO) is a colorless monomer produced for nylon-6 fibers and plastics. A substantial proportion
of the fiber is used in carpet manufacturing. Most commercial processes used for the manufacture of caprolactam
begin with benzene, but toluene can also be used. The production of caprolactam can give rise to significant
emissions of nitrous oxide (N20).
During the production of caprolactam, emissions of N20 can occur from the ammonia oxidation step, emissions of
carbon dioxide (C02) from the ammonium carbonate step, emissions of sulfur dioxide (S02) from the ammonium
bisulfite step, and emissions of non-methane volatile organic compounds (NMVOCs). Emissions of C02, S02 and
NMVOCs from the conventional process are unlikely to be significant in well-managed plants. Modified
caprolactam production processes are primarily concerned with elimination of the high volumes of ammonium
sulfate that are produced as a byproduct of the conventional process (IPCC 2006).
Where caprolactam is produced from benzene, the main process, the benzene is hydrogenated to cyclohexane
which is then oxidized to produce cyclohexanone (C6Hi0O). The classical route (Raschig process) and basic reaction
equations for production of caprolactam from cyclohexanone are (IPCC 2006):
Oxidation of NH3 to N0/N02
I
NH3 reacted with C02/H20 to yield ammonium carbonate (NH4)2C03
I
(NH4)2C03 reacted with N0/N02 (from NH3 oxidation) to yield ammonium nitrite (NH4N02)
I
NH3 reacted with S02/H20 to yield ammonium bisulphite (NH4HS03)
I
NH4N02 and (NH4HS03) reacted to yield hydroxylamine disulphonate (N0H(S03NH4)2)
I
(N0H(S03NH4)2) hydrolised to yield hydroxylamine sulphate ((NH20H)2.H2S04) and
ammonium sulphate ((NH4)2S04)
I
Cylohexanone reaction-.
1
C6H10O + — (NH20H)2.H2S04(+NH3 and H2S04) -> C6H10NOH + (NH4)2S04 + H20
I
Beckmann rearrangement:
C6H10NOH (+H2S04 and S02) -> C^H^NO. H2S04 (+4NH3 and H20) -> C^H^NO + 2(NH4)2S04
In 1999, there were four caprolactam production facilities in the United States. As of 2019, the United States had
two companies that produce caprolactam with a total of two caprolactam production facilities: AdvanSix in Virginia
(AdvanSix 2020) and BASF in Texas (BASF 2020). Caprolactam production at Fibrant LLC in Georgia ceased in 2018
(Cline 2019).
4-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Nitrous oxide emissions from caprolactam production in the United States were estimated to be 1.4 MMT C02 Eq.
(5 kt N20) in 2019 (see Table 4-33). National emissions from caprolactam production decreased by approximately
18 percent over the period of 1990 through 2019. Emissions in 2019 decreased by approximately 3 percent from
the 2018 levels.
Table 4-33: N2O Emissions from Caprolactam Production (MMT CO2 Eq. and kt N2O)
Year MMT CP2 Eq. kt N2Q
1990	1.7	6
2005	2.1	7
2015	1.9	6
2016	1.7	6
2017	1.5	5
2018	1.4	5
2019	1.4	5
Glyoxal
Glyoxal is mainly used as a crosslinking agent for vinyl acetate/acrylic resins, disinfectant, gelatin hardening agent,
textile finishing agent (permanent-press cotton, rayon fabrics), and wet-resistance additive (paper coatings) (IPCC
2006). It is also used for enhanced oil-recovery. It is produced from oxidation of acetaldehyde with concentrated
nitric acid, or from the catalytic oxidation of ethylene glycol, and N20 is emitted in the process of oxidation of
acetaldehyde.
Glyoxal (ethanedial) (C2H202) is produced from oxidation of acetaldehyde (ethanal) (C2H40) with concentrated
nitric acid (HN03). Glyoxal can also be produced from catalytic oxidation of ethylene glycol (ethanediol)
(CH2OHCH2OH).
Glyoxylic Acid
Glyoxylic acid is produced by nitric acid oxidation of glyoxal. Glyoxylic acid is used for the production of synthetic
aromas, agrochemicals, and pharmaceutical intermediates (IPCC 2006).
EPA does not currently estimate the emissions associated with the production of Glyoxal and Glyoxylic Acid due to
data availability and a lack of publicly available information on the industry in the United States. See Annex 5 for
additional information.
Methodology
Emissions of N20 from the production of caprolactam were calculated using the estimation methods provided by
the 2006 IPCC Guidelines. The 2006 IPCC Guidelines Tier 1 method was used to estimate emissions from
caprolactam production for 1990 through 2019, as shown in this formula:
en2o = EF x CP
where,
EN20	= Annual N20 Emissions (kg)
EF	= N20 emission factor (default) (kg N20/metric ton caprolactam produced)
CP	= Caprolactam production (metric tons)
During the caprolactam production process, N20 is generated as a byproduct of the high temperature catalytic
oxidation of ammonia (NH3), which is the first reaction in the series of reactions to produce caprolactam. The
Industrial Processes and Product Use 4-47

-------
amount of N20 emissions can be estimated based on the chemical reaction shown above. Based on this formula,
which is consistent with an IPCC Tier 1 approach, approximately 111.1 metric tons of caprolactam are required to
generate one metric ton of N20, resulting in an emission factor of 9.0 kg N20 per metric ton of caprolactam (IPCC
2006). When applying the Tier 1 method, the 2006 IPCC Guidelines state that it is good practice to assume that
there is no abatement of N20 emissions and to use the highest default emission factor available in the guidelines.
In addition, EPA did not find support for the use of secondary catalysts to reduce N20 emissions, such as those
employed at nitric acid plants. Thus, the 515 thousand metric tons (kt) of caprolactam produced in 2019 (ACC
2020) resulted in N20 emissions of approximately 1.4 MMT C02 Eq. (5 kt).
The activity data for caprolactam production (see Table 4-34) from 1990 to 2019 were obtained from the American
Chemistry Council's Guide to the Business of Chemistry (ACC 2020). EPA will continue to analyze and assess
alternative sources of production data as a quality control measure.
Table 4-34: Caprolactam Production (kt)
Year	kt_
1990 626
2005 795
2015	700
2016	640
2017	545
2018	530
2019	515
Carbon dioxide and methane (CH4) emissions may also occur from the production of caprolactam, but currently the
IPCC does not have methodologies for calculating these emissions associated with caprolactam production.
Uncertainty and Time-Series Consistency
Estimation of emissions of N20 from caprolactam production can be treated as analogous to estimation of
emissions of N20 from nitric acid production. Both production processes involve an initial step of NH3 oxidation,
which is the source of N20 formation and emissions (IPCC 2006). Therefore, uncertainties for the default emission
factor values in the 2006 IPCC Guidelines are an estimate based on default values for nitric acid plants. In general,
default emission factors for gaseous substances have higher uncertainties because mass values for gaseous
substances are influenced by temperature and pressure variations and gases are more easily lost through process
leaks. The default values for caprolactam production have a relatively high level of uncertainty due to the limited
information available (IPCC 2006).
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-35. Nitrous oxide
emissions from Caprolactam, Glyoxal and Glyoxylic Acid Production for 2019 were estimated to be between 1.0
and 1.8 MMT C02 Eq. at the 95 percent confidence level. These values indicate a range of approximately 31
percent below to 32 percent above the 2019 emission estimate of 1.4 MMT C02 Eq.
Table 4-35: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from
Caprolactam, Glyoxal and Glyoxylic Acid Production (MMT CO2 Eq. and Percent)
2019 Emission Estimate Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.)	(MMTCOz Eq.)	(%)
4-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Caprolactam Production
N20
1.4
1.0
1.8
-31%
+32%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of the 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
Pending resources, EPA will research other available datasets for caprolactam production and industry trends,
including facility-level data. EPA will also research the production process and emissions associated with the
production of glyoxal and glyoxylic acid. During the Expert Review period for the current Inventory report, EPA
continues to seek expert solicitation on data available for these emission source categories. This planned
improvement is subject to data availability and will be implemented in the medium- to long-term.
4.10 Carbide Production and Consumption
(CRF Source Category 2B5)
Carbon dioxide (C02) and methane (CH4) are emitted from the production of silicon carbide (SiC), a material used
for industrial abrasive, metallurgical and other non-abrasive applications in the United States. Emissions from fuels
consumed for energy purposes during the production of silicon carbide are accounted for in the Energy chapter.
To produce SiC, silica sand or quartz (Si02) is reacted with carbon (C) in the form of petroleum coke. A portion
(about 35 percent of the carbon contained in the petroleum coke is retained in the SiC. The remaining C is emitted
as C02, CH4, or carbon monoxide (CO). The overall reaction is shown below, but in practice, it does not proceed
according to stoichiometry:
Si02 + 3C —> SiC + 2CO (+ 02 —> 2COz)
Carbon dioxide and CH4 are also emitted during the production of calcium carbide, a chemical used to produce
acetylene. Carbon dioxide is implicitly accounted for in the storage factor calculation for the non-energy use of
petroleum coke in the Energy chapter. As noted in Annex 5 to this report, CH4 emissions from calcium carbide
production are not estimated because data are not available. EPA is continuing to investigate the inclusion of these
emissions in future Inventory reports.
Markets for manufactured abrasives, including SiC, are heavily influenced by activity in the U.S. manufacturing
sector, especially in the aerospace, automotive, furniture, housing, and steel manufacturing sectors. The U.S.
Geological Survey (USGS) reports that a portion (approximately 50 percent) of SiC is used in metallurgical and
Industrial Processes and Product Use 4-49

-------
other non-abrasive applications, primarily in iron and steel production (USGS 1991a through 2017). As a result of
the economic downturn in 2008 and 2009, demand for SiC decreased in those years. Low-cost imports, particularly
from China, combined with high relative operating costs for domestic producers, continue to put downward
pressure on the production of SiC in the United States. Demand for SiC consumption in the United States has
recovered somewhat from its low in 2009 (USGS 1991a through 2015). Abrasive-grade silicon carbide was
manufactured at one facility in 2017 in the United States (USGS 2020).
Carbon dioxide emissions from SiC production and consumption in 2019 were 0.2 MMT C02 Eq. (175 kt C02) (see
Table 4-36 and Table 4-37). Approximately 52 percent of these emissions resulted from SiC production, while the
remainder resulted from SiC consumption. Methane emissions from SiC production in 2019 were 0.01 MMT C02
Eq. (0.4 kt CH4) (see Table 4-36 and Table 4-37). Emissions have not fluctuated greatly in recent years, but 2019
emissions are about 50 percent lower than emissions in 1990.
Table 4-36: CO2 and ChU Emissions from Silicon Carbide Production and Consumption (MMT
COz Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
C02
0.4
0.2
0.2
0.2
0.2
0.2
0.2
ch4
+
+
+
+
+
+
+
Total
0.4
0.2
0.2
0.2
0.2
0.2
0.2
+ Does not exceed 0.05 MMT C02 Eq.
Table 4-37: CO2 and ChU Emissions from Silicon Carbide Production and Consumption (kt)
Year
C02
CH4
1990
370
1
2005
213
+
2015
176
+
2016
170
+
2017
181
+
2018
184
+
2019
175
+
+ Does not exceed 0.5 kt
Methodology
Emissions of C02 and CH4 from the production of SiC were calculated39 using the Tier 1 method provided by the
2006IPCC Guidelines. Annual estimates of SiC production were multiplied by the appropriate emission factor, as
shown below:
Esc,C02 = EFsc,C02 X Qsc
1 metric ton\
where,
Esc,C02
EFsc,C02
Qsc
Esc,CH4
EFSC CH4
Esc,CH4 ~ EFsc,CH4 * Qsc *
1000 kg
C02 emissions from production of SiC, metric tons
Emission factor for production of SiC, metric ton C02/metric ton SiC
Quantity of SiC produced, metric tons
CH4 emissions from production of SiC, metric tons
Emission factor for production of SiC, kilogram CH4/metric ton SiC
Emission factors were taken from the 2006 IPCC Guidelines:
39 EPA has not integrated aggregated facility-level GHGRP information to inform these estimates. The aggregated information
(e.g., activity data and emissions) associated with silicon carbide did not meet criteria to shield underlying confidential business
information (CBI) from public disclosure.
4-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	2.62 metric tons C02/metric ton SiC
•	11.6 kg CHVmetric ton SiC
Production data for metallurgical and other non-abrasive applications of SiC are not available; therefore, both C02
and CH4 estimates for SiC are based solely upon production data for SiC for industrial abrasive applications.
Silicon carbide industrial abrasives production data for 1990 through 2013 were obtained from the U.S. Geological
Survey (USGS) Minerals Yearbook: Manufactured Abrasives (USGS 1991a through 2015). Production data for 2014
through 2017 were obtained from the Mineral Commodity Summaries: Abrasives (Manufactured) (USGS 2019).
Production data for 2018 and 2019 were obtained from the Mineral Industry Surveys, Manufactured Abrasives
(USGS 2019a, USGS 2020a). Silicon carbide production data obtained through the USGS National Minerals
Information Center has been rounded to the nearest 5,000 metric tons to avoid disclosing company proprietary
data. SiC consumption for the entire time series is estimated using USGS consumption data (USGS 1991b through
2015, USGS 2017c) and data from the U.S. International Trade Commission (USITC) database on net imports and
exports of SiC provided by the U.S. Census Bureau (2005 through 2020) (see Table 4-38). Total annual SiC
consumption (utilization) was estimated by subtracting annual exports of SiC by the annual total of national SiC
production and net imports.
Emissions of C02 from SiC consumption for metallurgical uses were calculated by multiplying the annual utilization
of SiC for metallurgical uses (reported annually in the USGS Minerals Yearbook: Silicon) by the carbon content of
SiC (30.0 percent), which was determined according to the molecular weight ratio of SiC. USGS has not published
consumption data for metallurgical uses since 2016 due to concerns of disclosing company-specific sensitive
information, and there is uncertainty about the future availability of these data from the USGS. Other options are
being explored that would allow the estimation of SiC consumption for metallurgical uses. The 2016 consumption
data will be used as a proxy until a suitable approach is developed.
Emissions of C02from SiC consumption for other non-abrasive uses were calculated by multiplying the annual SiC
consumption for non-abrasive uses by the carbon content of SiC (30 percent). The annual SiC consumption for non-
abrasive uses was calculated by multiplying the annual SiC consumption (production plus net imports) by the
percentage used in metallurgical and other non-abrasive uses (50 percent) (USGS 1991a through 2017) and then
subtracting the SiC consumption for metallurgical use.
The petroleum coke portion of the total C02 process emissions from silicon carbide production is adjusted for
within the Energy chapter, as these fuels were consumed during non-energy related activities. Additional
information on the adjustments made within the Energy sector for Non-Energy Use of Fuels is described in both
the Methodology section of C02 from Fossil Fuel Combustion (Section 3.1) and Annex 2.1, Methodology for
Estimating Emissions of C02 from Fossil Fuel Combustion.
Table 4-38: Production and Consumption of Silicon Carbide (Metric Tons)
Year
Production
Consumption
1990
105,000
172,465
2005
35,000
220,149
2015
35,000
153,474
2016
35,000
142,104
2017
35,000
163,492
2018
35,000
168,526
2019
35,000
152,410
Uncertainty and Time-Series Consistency
There is uncertainty associated with the emission factors used because they are based on stoichiometry as
opposed to monitoring of actual SiC production plants. An alternative is to calculate emissions based on the
Industrial Processes and Product Use 4-51

-------
quantity of petroleum coke used during the production process rather than on the amount of silicon carbide
produced. However, these data were not available. For CH4, there is also uncertainty associated with the
hydrogen-containing volatile compounds in the petroleum coke (IPCC 2006). There is also uncertainty associated
with the use or destruction of CH4 generated from the process, in addition to uncertainty associated with levels of
production, net imports, consumption levels, and the percent of total consumption that is attributed to
metallurgical and other non-abrasive uses.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-39. Silicon carbide
production and consumption C02 emissions from 2019 were estimated to be between 9 percent below and 9
percent above the emission estimate of 0.18 MMT C02 Eq. at the 95 percent confidence level. Silicon carbide
production CH4 emissions were estimated to be between 9 percent below and 9 percent above the emission
estimate of 0.01 MMT C02 Eq. at the 95 percent confidence level.
Table 4-39: Approach 2 Quantitative Uncertainty Estimates for ChU and CO2 Emissions from
Silicon Carbide Production and Consumption (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTCOz Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Silicon Carbide Production
and Consumption
C02
0.18
0.16
0.19
-9% +9%
Silicon Carbide Production
ch4
+
+
+
-9% +9%
+ Does not exceed 0.05 MMT C02 Eq.
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006 IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
Recalculations Discussion
During annual QC, a transcription error in the molecular weight value for silicon was identified that impacts the
percentage of carbon in SiC. This corrected percentage (30 percent) was applied to the entire time series for SiC
consumption. This change resulted in annual emissions decreases ranging from 3 to 9 kt C02 between 1990 and
2018.
4.11 Titanium Dioxide Production (CRF
Source Category 2B6)
Titanium dioxide (Ti02) is manufactured using one of two processes: the chloride process and the sulfate process.
The chloride process uses petroleum coke and chlorine as raw materials and emits process-related carbon dioxide
4-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
(C02). Emissions from fuels consumed for energy purposes during the production of titanium dioxide are
accounted for in the Energy chapter. The chloride process is based on the following chemical reactions:
2FeTi03 + 7Cl2 + 3C —> 2TiCl^ + 2FeCl^ + 3C02
2TiCl4 + 202 ~~* 2Ti02 ~l~
The sulfate process does not use petroleum coke or other forms of carbon as a raw material and does not emit
C02.
The carbon in the first chemical reaction is provided by petroleum coke, which is oxidized in the presence of the
chlorine and FeTi03 (rutile ore) to form C02. Since 2004, all Ti02 produced in the United States has been produced
using the chloride process, and a special grade of "calcined" petroleum coke is manufactured specifically for this
purpose.
The principal use of Ti02 is as a pigment in white paint, lacquers, and varnishes. It is also used as a pigment in the
manufacture of plastics, paper, and other products. In 2019, U.S. Ti02 production totaled 1,100,000 metric tons
(USGS 2020). There were five plants producing Ti02 in the United States in 2019.
Emissions of C02 from titanium dioxide production in 2019 were estimated to be 1.5 MMT C02 Eq. (1,474 kt C02),
which represents an increase of 23 percent since 1990 (see Table 4-40). Compared to 2018, emissions from
titanium dioxide production decreased by 4 percent in 2019, likely due to a 4 percent decrease in production.
Table 4-40: CO2 Emissions from Titanium Dioxide (MMT CO2 Eq. and kt)
Year MMT C02 Eq.	Kt
1990
1.2
1,195
2005
1.8
1,755
2015
1.6
1,635
2016
1.7
1,662
2017
1.7
1,688
2018
1.5
1,541
2019
1.5
1,474
Methodology
Jff
Emissions of C02 from Ti02 production were calculated by multiplying annual national Ti02 production by chloride
process-specific emission factors using a Tier 1 approach provided in 2006IPCC Guidelines. The Tier 1 equation is
as follows:
Etd = EFtd X Qtd
where,
Etd =	C02 emissions from Ti02 production, metric tons
EFtd =	Emission factor (chloride process), metric ton C02/metric ton Ti02
Qtd =	Quantity of Ti02 produced
The petroleum coke portion of the total C02 process emissions from Ti02 production is adjusted for within the
Energy chapter as these fuels were consumed during non-energy related activities. Additional information on the
adjustments made within the Energy sector for Non-Energy Use of Fuels is described in both the Methodology
section of C02 from Fossil Fuel Combustion (Section 3.1 Fossil Fuel Combustion) and Annex 2.1, Methodology for
Estimating Emissions of C02 from Fossil Fuel Combustion.
Industrial Processes and Product Use 4-53

-------
Data were obtained for the total amount of Ti02 produced each year. For years prior to 2004, it was assumed that
Ti02 was produced using the chloride process and the sulfate process in the same ratio as the ratio of the total U.S.
production capacity for each process. As of 2004, the last remaining sulfate process plant in the United States
closed; therefore, 100 percent of production since 2004 used the chloride process (USGS 2005). An emission factor
of 1.34 metric tons C02/metric ton Ti02 was applied to the estimated chloride-process production (IPCC 2006). It
was assumed that all Ti02 produced using the chloride process was produced using petroleum coke, although
some Ti02 may have been produced with graphite or other carbon inputs.
The emission factor for the Ti02 chloride process was taken from the 2006 IPCC Guidelines. Titanium dioxide
production data and the percentage of total Ti02 production capacity that is chloride process for 1990 through
2013 (see Table 4-41) were obtained through the U.S. Geological Survey (USGS) Minerals Yearbook: Titanium
Annual Report (USGS 1991 through 2015). Production data for 2014 through 2019 were obtained from the
Minerals Commodity Summaries: Titanium and Titanium Dioxide (USGS 2020).40 Data on the percentage of total
Ti02 production capacity that is chloride process were not available for 1990 through 1993, so data from the 1994
USGS Minerals Yearbook were used for these years. Because a sulfate process plant closed in September 2001, the
chloride process percentage for 2001 was estimated based on a discussion with Joseph Gambogi (2002). By 2002,
only one sulfate process plant remained online in the United States, and this plant closed in 2004 (USGS 2005).
Table 4-41: Titanium Dioxide Production (kt)
Year
kt
1990
979
2005
1,310
2015
1,220
2016
1,240
2017
1,260
2018
1,150
2019
1,100
Uncertainty and Time-Series Consistency
Each year, the USGS collects titanium industry data for titanium mineral and pigment production operations. If
Ti02 pigment plants do not respond, production from the operations is estimated based on prior year production
levels and industry trends. Variability in response rates varies from 67 to 100 percent of Ti02 pigment plants over
the time series.
Although some Ti02 may be produced using graphite or other carbon inputs, information and data regarding these
practices were not available. Titanium dioxide produced using graphite inputs, for example, may generate differing
amounts of C02per unit of Ti02 produced as compared to that generated using petroleum coke in production.
While the most accurate method to estimate emissions would be to base calculations on the amount of reducing
agent used in each process rather than on the amount of Ti02 produced, sufficient data were not available to do
so.
As of 2004, the last remaining sulfate-process plant in the United States closed. Since annual Ti02 production was
not reported by USGS by the type of production process used (chloride or sulfate) prior to 2004 and only the
percentage of total production capacity by process was reported, the percent of total Ti02 production capacity that
was attributed to the chloride process was multiplied by total Ti02 production to estimate the amount of Ti02
40 EPA has not integrated aggregated facility-level GHGRP information for Titanium Dioxide production facilities (40 CFR Part 98
Subpart EE). The relevant aggregated information (activity data, emission factor) from these facilities did not meet criteria to
shield underlying CBI from public disclosure.
4-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
produced using the chloride process. Finally, the emission factor was applied uniformly to all chloride-process
production, and no data were available to account for differences in production efficiency among chloride-process
plants. In calculating the amount of petroleum coke consumed in chloride-process Ti02 production, literature data
were used for petroleum coke composition. Certain grades of petroleum coke are manufactured specifically for
use in the Ti02 chloride process; however, this composition information was not available.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-42. Titanium dioxide
consumption C02 emissions from 2019 were estimated to be between 1.3 and 1.7 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 12 percent below and 13 percent above the emission
estimate of 1.5 MMT C02 Eq.
Table 4-42: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Titanium
Dioxide Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.)
(MMT C02 Eq.)
(%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Titanium Dioxide Production
C02
1.5
1.3 1.7
-12% +13%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of the 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
EPA plans to examine the use of GHGRP titanium dioxide emissions data for possible use in emission estimates
consistent with both Volume 1, Chapter 6 of the 2006 IPCC Guidelines and the latest IPCC guidance on the use of
facility-level data in national inventories.41 This planned improvement is ongoing and has not been incorporated
into this Inventory report. This is a long-term planned improvement.
41 See .
Industrial Processes and Product Use 4-55

-------
4.12 Soda Ash Production (CRF Source
Category 2B7)
Carbon dioxide (C02) is generated as a byproduct of calcining trona ore to produce soda ash and is eventually
emitted into the atmosphere. In addition, C02 may also be released when soda ash is consumed. Emissions from
soda ash consumption in chemical production processes are reported under Section 4.4 Other Process Uses of
Carbonates (CRF Category 2A4), and emissions from fuels consumed for energy purposes during the production
and consumption of soda ash are accounted for in the Energy chapter.
Calcining involves placing crushed trona ore into a kiln to convert sodium bicarbonate into crude sodium carbonate
that will later be filtered into pure soda ash. The emission of C02 during trona-based production is based on the
following reaction:
2Na2C03 ¦ NaHC03 ¦ 2H20(Trona) -» 3Na2C03(Soda Ash) + 5H20 +C02
Soda ash (sodium carbonate, Na2C03) is a white crystalline solid that is readily soluble in water and strongly
alkaline. Commercial soda ash is used as a raw material in a variety of industrial processes and in many familiar
consumer products such as glass, soap and detergents, paper, textiles, and food. The largest use of soda ash is for
glass manufacturing. Emissions from soda ash used in glass production are reported under Section 4.3, Glass
Production (CRF Source Category 2A3). In addition, soda ash is used primarily to manufacture many sodium-based
inorganic chemicals, including sodium bicarbonate, sodium chromates, sodium phosphates, and sodium silicates
(USGS 2018b). Internationally, two types of soda ash are produced: natural and synthetic. The United States
produces only natural soda ash and is second only to China in total soda ash production. Trona is the principal ore
from which natural soda ash is made.
The United States represents about one-fifth of total world soda ash output (USGS 2020a). Only two states
produce natural soda ash: Wyoming and California. Of these two states, net emissions of C02 from soda ash
production were only calculated for Wyoming, due to specifics regarding the production processes employed in
the state.42 Based on 2019 reported data, the estimated distribution of soda ash by end-use in 2019 (excluding
glass production) was chemical production, 55 percent; other uses, 14 percent; soap and detergent manufacturing,
11 percent; wholesale distributors (e.g., for use in agriculture, water treatment, and grocery wholesale), 10
percent; flue gas desulfurization, 6 percent; water treatment, 2 percent, and pulp and paper production, 2 percent
(USGS 2020b).43
U.S. natural soda ash is competitive in world markets because it is generally considered a better-quality raw
material than synthetically produced soda ash, and the majority of the world output of soda ash is made
synthetically. Although the United States continues to be a major supplier of soda ash, China surpassed the United
States in soda ash production in 2003, becoming the world's leading producer.
42	In California, soda ash is manufactured using sodium carbonate-bearing brines instead of trona ore. To extract the sodium
carbonate, the complex brines are first treated with C02 in carbonation towers to convert the sodium carbonate into sodium
bicarbonate, which then precipitates from the brine solution. The precipitated sodium bicarbonate is then calcined back into
sodium carbonate. Although C02 is generated as a byproduct, the C02 is recovered and recycled for use in the carbonation stage
and is not emitted. A third state, Colorado, produced soda ash until the plant was idled in 2004. The lone producer of sodium
bicarbonate no longer mines trona ore in the state. For a brief time, sodium bicarbonate was produced using soda ash
feedstocks mined in Wyoming and shipped to Colorado. Prior to 2004, because the trona ore was mined in Wyoming, the
production numbers given by the USGS included the feedstocks mined in Wyoming and shipped to Colorado. In this way, the
sodium bicarbonate production that took place in Colorado was accounted for in the Wyoming numbers.
43	Percentages may not add up to 100 percent due to independent rounding.
4-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
In 2019, C02 emissions from the production of soda ash from trona ore were 1.8 MMT C02 Eq. (1,792 kt C02) (see
Table 4-43). Total emissions from soda ash production in 2019 increased by approximately 5 percent from
emissions in 2018 and have increased by approximately 25 percent from 1990 levels.
Emissions have remained relatively constant over the time series with some fluctuations since 1990. In general,
these fluctuations were related to the behavior of the export market and the U.S. economy. The U.S. soda ash
industry continued a trend of increased production and value in 2019 since experiencing a decline in domestic and
export sales caused by adverse global economic conditions in 2009.
Table 4-43: CO2 Emissions from Soda Ash Production (MMT CO2 Eq. and kt CO2)
Year	MMT CP2 Eq. ktCP2
1990	1.4	1,431
2005	1.7	1,655
2015	1.7	1,714
2016	1.7	1,723
2017	1.8	1,753
2018	1.7	1,714
2019	1.8	1,792
Methodology
During the soda ash production process, trona ore is calcined in a rotary kiln and chemically transformed into a
crude soda ash that requires further processing. Carbon dioxide and water are generated as byproducts of the
calcination process. Carbon dioxide emissions from the calcination of trona ore can be estimated based on the
chemical reaction shown above. Based on this formula, which is consistent with an IPCC Tier 1 approach,
approximately 10.27 metric tons of trona ore are required to generate one metric ton of C02, or an emission factor
of 0.0974 metric tons C02 per metric ton of trona ore (IPCC 2006). Thus, the 18.4 million metric tons of trona ore
mined in 2019 for soda ash production (USGS 2020b) resulted in C02 emissions of approximately 1.8 MMT C02 Eq.
(1,792 kt).
Once produced, most soda ash is consumed in chemical production, with minor amounts used in soap production,
pulp and paper, flue gas desulfurization, and water treatment (excluding soda ash consumption for glass
manufacturing). As soda ash is consumed for these purposes, additional C02 is usually emitted. Consistent with the
2006 IPCC Guidelines for National Greenhouse Gas Inventories, emissions from soda ash consumption in chemical
production processes are reported under Section 4.4 Other Process Uses of Carbonates (CRF Category 2A4).
Data is not currently available for the quantity of trona used in soda ash production. Trona ore produced is used
primarily for soda ash production, and for the current Inventory report, EPA assumes that all trona produced was
used in soda ash production. The activity data for trona ore production (see Table 4-44) for 1990 through 2019
were obtained from the U.S. Geological Survey (USGS) Minerals Yearbook for Soda Ash (1994 through 2015b) and
USGS Mineral Industry Surveys for Soda Ash (USGS 2016 through 2017, 2018a, 2019, 2020b). Soda ash
production44 data were collected by the USGS from voluntary surveys of the U.S. soda ash industry. EPA will
continue to analyze and assess opportunities to use facility-level data from EPA's GHGRP to improve the emission
estimates for the Soda Ash Production source category consistent with IPCC45 and UNFCCC guidelines.
44	EPA has assessed the feasibility of using emissions information (including activity data) from EPA's GHGRP program. At this
time, the aggregated information associated with production of soda ash did not meet criteria to shield underlying confidential
business information (CBI) from public disclosure.
45	See .
Industrial Processes and Product Use 4-57

-------
Table 4-44: Trona Ore Use (kt)
Year
Use3
1990
14,700
2005
17,000
2015
17,600
2016
17,700
2017
18,000
2018
17,600
2019
18,400
a Trona ore use is
assumed to be equal
to trona ore
production.
Uncertainty and Time-Series Consistency
Emission estimates from soda ash production have relatively low associated uncertainty levels because reliable
and accurate data sources are available for the emission factor and activity data for trona-based soda ash
production. One source of uncertainty is the purity of the trona ore used for manufacturing soda ash. The emission
factor used for this estimate assumes the ore is 100 percent pure and likely overestimates the emissions from soda
ash manufacture. The average water-soluble sodium carbonate-bicarbonate content for ore mined in Wyoming
ranges from 85.5 to 93.8 percent (USGS 1995c).
EPA is aware of one facility producing soda ash from a liquid alkaline feedstock process, based on EPA's GHGRP.
Soda ash production data was collected by the USGS from voluntary surveys. A survey request was sent to each of
the five soda ash producers, all of which responded, representing 100 percent of the total production data (USGS
2020b).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-45. Soda ash production
C02 emissions for 2019 were estimated to be between 1.5 and 1.8 MMT C02 Eq. at the 95 percent confidence
level. This indicates a range of approximately 9 percent below and 8 percent above the emission estimate of 1.8
MMT C02 Eq.
Table 4-45: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Soda Ash
Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMTC02 Eq.)
(MMTC02 Eq.)
(%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Soda Ash Production
C02
1.8
1.5 1.8
-9% +8%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions estimates
from 1990 through 2019.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
4-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
EPA plans to review USGS data to improve use of activity data to estimate emissions, consistent with the
methodological decision trees in 2006IPCC Guidelines. EPA also plans to use GHGRP data for conducting category-
specific QC of emission estimates, consistent with both Volume 1, Chapter 6 of the 2006 IPCC Guidelines and the
latest IPCC guidance on the use of facility-level data in national inventories.46 This planned improvement is
ongoing and has not been incorporated into this Inventory report. This is a medium-term planned improvement
and expected to be completed by the next (i.e., 2021) Inventory submission.
4.13 Petrochemical Production (CRF Source
Category 2B8)
The production of some petrochemicals results in the release of carbon dioxide (C02) and methane (CH4)
emissions. Petrochemicals are chemicals isolated or derived from petroleum or natural gas. Carbon dioxide
emissions from the production of acrylonitrile, carbon black, ethylene, ethylene dichloride, ethylene oxide, and
methanol, and CH4 emissions from the production of methanol and acrylonitrile are presented here and reported
under IPCC Source Category 2B8. The petrochemical industry uses primary fossil fuels (i.e., natural gas, coal,
petroleum, etc.) for non-fuel purposes in the production of carbon black and other petrochemicals. Emissions from
fuels and feedstocks transferred out of the system for use in energy purposes (e.g., indirect or direct process heat
or steam production) are currently accounted for in the Energy sector. The allocation and reporting of emissions
from feedstocks transferred out of the system for use in energy purposes to the Energy chapter is consistent with
the 2006 IPCC Guidelines.
Worldwide, more than 90 percent of acrylonitrile (vinyl cyanide, C3H3N) is made by way of direct ammoxidation of
propylene with ammonia (NH3) and oxygen over a catalyst. This process is referred to as the SOHIO process,
named after the Standard Oil Company of Ohio (SOHIO) (IPCC 2006). The primary use of acrylonitrile is as the raw
material for the manufacture of acrylic and modacrylic fibers. Other major uses include the production of plastics
(acrylonitrile-butadiene-styrene [ABS] and styrene-acrylonitrile [SAN]), nitrile rubbers, nitrile barrier resins,
adiponitrile, and acrylamide. All U.S. acrylonitrile facilities use the SOHIO process (AN 2014). The SOHIO process
involves a fluidized bed reaction of chemical-grade propylene, ammonia, and oxygen over a catalyst. The process
produces acrylonitrile as its primary product, and the process yield depends on the type of catalyst used and the
process configuration. The ammoxidation process produces byproduct C02, carbon monoxide (CO), and water
from the direct oxidation of the propylene feedstock and produces other hydrocarbons from side reactions.
Carbon black is a black powder generated by the incomplete combustion of an aromatic petroleum- or coal-based
feedstock at a high temperature. Most carbon black produced in the United States is added to rubber to impart
strength and abrasion resistance, and the tire industry is by far the largest consumer. The other major use of
carbon black is as a pigment. The predominant process used in the United States to produce carbon black is the
furnace black (or oil furnace) process. In the furnace black process, carbon black oil (a heavy aromatic liquid) is
continuously injected into the combustion zone of a natural gas-fired furnace. Furnace heat is provided by the
natural gas and a portion of the carbon black feedstock; the remaining portion of the carbon black feedstock is
pyrolyzed to carbon black. The resultant C02 and uncombusted CH4 emissions are released from thermal
incinerators used as control devices, process dryers, and equipment leaks. Three facilities in the United States use
46 See .
Industrial Processes and Product Use 4-59

-------
other types of carbon black processes. Specifically, one facility produces carbon black by the thermal cracking of
acetylene-containing feedstocks (i.e., acetylene black process), a second facility produces carbon black by the
thermal cracking of other hydrocarbons (i.e., thermal black process), and a third facility produces carbon black by
the open burning of carbon black feedstock (i.e., lamp black process) (EPA 2000).
Ethylene (C2H4) is consumed in the production processes of the plastics industry including polymers such as high,
low, and linear low density polyethylene (HDPE, LDPE, LLDPE); polyvinyl chloride (PVC); ethylene dichloride;
ethylene oxide; and ethylbenzene. Virtually all ethylene is produced from steam cracking of ethane, propane,
butane, naphtha, gas oil, and other feedstocks. The representative chemical equation for steam cracking of ethane
to ethylene is shown below:
^2^6	+ H2
Small amounts of CH4 are also generated from the steam cracking process. In addition, C02 and CH4 emissions are
also generated from combustion units.
Ethylene dichloride (C2H4CI2) is used to produce vinyl chloride monomer, which is the precursor to polyvinyl
chloride (PVC). Ethylene dichloride was also used as a fuel additive until 1996 when leaded gasoline was phased
out. Ethylene dichloride is produced from ethylene by either direct chlorination, oxychlorination, or a combination
of the two processes (i.e., the "balanced process"); most U.S. facilities use the balanced process. The direct
chlorination and oxychlorination reactions are shown below:
C2H4 + Cl2 -> C2H4Cl2 (direct chlorination)
C2H4 + |02 + 2HCI -» C2H4Cl2 + 2H20 (oxychlorination)
C2H4 + 302 -» 2C02 + 2H20 (direct oxidation of ethylene during oxychlorination)
In addition to the byproduct C02 produced from the direction oxidation of the ethylene feedstock, C02 and CH4
emissions are also generated from combustion units.
Ethylene oxide (C2H40) is used in the manufacture of glycols, glycol ethers, alcohols, and amines. Approximately 70
percent of ethylene oxide produced worldwide is used in the manufacture of glycols, including monoethylene
glycol. Ethylene oxide is produced by reacting ethylene with oxygen over a catalyst. The oxygen may be supplied to
the process through either an air (air process) or a pure oxygen stream (oxygen process). The byproduct C02 from
the direct oxidation of the ethylene feedstock is removed from the process vent stream using a recycled carbonate
solution, and the recovered C02 may be vented to the atmosphere or recovered for further utilization in other
sectors, such as food production (IPCC 2006). The combined ethylene oxide reaction and byproduct C02 reaction is
exothermic and generates heat, which is recovered to produce steam for the process. The ethylene oxide process
also produces other liquid and off-gas byproducts (e.g., ethane, etc.) that may be burned for energy recovery
within the process. Almost all facilities, except one in Texas, use the oxygen process to manufacture ethylene oxide
(EPA 2008).
Methanol (CH3OH) is a chemical feedstock most often converted into formaldehyde, acetic acid and olefins. It is
also an alternative transportation fuel, as well as an additive used by municipal wastewater treatment facilities in
the denitrification of wastewater. Methanol is most commonly synthesized from a synthesis gas (i.e., "syngas" - a
mixture containing H2, CO, and C02) using a heterogeneous catalyst. There are a number of process techniques
that can be used to produce syngas. Worldwide, steam reforming of natural gas is the most common method;
most methanol producers in the United States also use steam reforming of natural gas to produce syngas. Other
syngas production processes in the United States include partial oxidation of natural gas and coal gasification.
Emissions of C02 and CH4 from petrochemical production in 2019 were 30.8 MMT C02 Eq. (30,792 kt C02) and 0.3
MMT C02 Eq. (13 kt CH4), respectively (see Table 4-46 and Table 4-47). Since 1990, total C02 emissions from
petrochemical production increased by 42 percent. Carbon dioxide emissions from petrochemical production are
driven primarily from ethylene, while CH4 emissions are mainly from methanol production. Overall emissions from
methanol production reached a low in 2011, given declining methanol production; however, emissions have been
increasing every year since 2011 and are now 53 percent greater than in 1990 (though still 4 percent less than the
peak of 4.0 MMT C02 Eq. in 1997) due to a rebound in methanol production.
4-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-46: CO2 and Cm Emissions from Petrochemical Production (MMT CO2 Eg.)
Year
1990
2005
2015
2016
2017
2018
2019
Total C02
21.6
27.4
28.1
28.3
28.9
29.3
30.8
Carbon Black
3.4
4.3
3.3
3.2
3.3
3.4
3.3
Ethylene
13.1
19.0
20.1
19.8
20.0
19.4
20.7
Ethylene Dichloride
0.3
0.5
0.4
0.4
0.4
0.4
0.5
Methanol
2.5
0.8
2.1
2.8
2.9
3.5
3.8
Ethylene Oxide
1.1
1.5
1.2
1.1
1.3
1.3
1.4
Acrylonitrile
1.2
1.3
1.1
1.0
1.0
1.3
1.1
Total CH4
0.2
0.1
0.2
0.2
0.3
0.3
0.3
Methanol
0.2
0.1
0.2
0.2
0.2
0.3
0.3
Acrylonitrile
+
+
+
+
+
+
+
Total
21.8
27.5
28.2
28.6
29.2
29.6
31.1
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 4-47: CO2 and Cm Emissions from Petrochemical Production (kt)
Year
1990
2005
2015
2016
2017
2018
2019
Total C02
21,611
27,383
28,062
28,310
28,910
29,314
30,792
Carbon Black
3,381
4,269
3,260
3,160
3,330
3,440
3,300
Ethylene
13,126
19,024
20,100
19,800
20,000
19,400
20,700
Ethylene Dichloride
254
455
398
447
412
440
503
Methanol
2,513
821
2,054
2,848
2,878
3,484
3,839
Ethylene Oxide
1,123
1,489
1,200
1,100
1,250
1,300
1,370
Acrylonitrile
1,214
1,325
1,050
955
1,040
1,250
1,080
Total CH4
9
3
7
10
10
12
13
Methanol
9
3
7
10
10
12
13
Acrylonitrile	+ y,	+ /	+	+	+	+	+
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt CH4.
Methodology
Emissions of C02 and CH4 were calculated using the estimation methods provided by the 2006IPCC Guidelines and
country-specific methods from EPA's GHGRP. The 2006 IPCC Guidelines Tier 1 method was used to estimate C02
and CH4 emissions from production of acrylonitrile and methanol,47 and a country-specific approach similar to the
IPCC Tier 2 method was used to estimate C02 emissions from production of carbon black, ethylene oxide, ethylene,
and ethylene dichloride. The Tier 2 method for petrochemicals is a total feedstock carbon (C) mass balance
method used to estimate total C02 emissions, but it is not applicable for estimating CH4 emissions.
As noted in the 2006 IPCC Guidelines, the total feedstock C mass balance method (Tier 2) is based on the
assumption that all of the C input to the process is converted either into primary and secondary products or into
C02. Further, the guideline states that while the total C mass balance method estimates total C emissions from the
process, it does not directly provide an estimate of the amount of the total C emissions emitted as C02, CH4, or
non-CH4 volatile organic compounds (NMVOCs). This method accounts for all the C as C02, including CH4.
Note, a small subset of facilities reporting under EPA's GHGRP use Continuous Emission Monitoring Systems
(CEMS) to monitor C02 emissions from process vents and/or stacks from stationary combustion units, these
47 EPA has not integrated aggregated facility-level GHGRP information for acrylonitrile and methanol production. The
aggregated information associated with production of these petrochemicals did not meet criteria to shield underlying CBI from
public disclosure.
Industrial Processes and Product Use 4-61

-------
facilities are required to also report C02, CH4 and N20 emissions from combustion of process off-gas in flares. The
C02 from flares are included in aggregated C02 results. Preliminary analysis of aggregated annual reports shows
that flared CH4 and N20 emissions are less than 500 kt C02 Eq./year. EPA's GHGRP is still reviewing these data
across reported years, and EPA plans to address this more completely in future reports.
Carbon Black, Ethylene, Ethylene Dichloride, and Ethylene Oxide
2010 through 2019
Carbon dioxide emissions and national production were aggregated directly from EPA's GHGRP dataset for 2010
through 2019 (EPA 2019, 2020). In 2019, data reported to the GHGRP included C02 emissions of 3,300,000 metric
tons from carbon black production; 20,700,000 metric tons of C02from ethylene production; 503,000 metric tons
of C02 from ethylene dichloride production; and 1,370,000 metric tons of C02 from ethylene oxide production.
These emissions reflect application of a country-specific approach similar to the IPCC Tier 2 method and were used
to estimate C02 emissions from the production of carbon black, ethylene, ethylene dichloride, and ethylene oxide.
Since 2010, EPA's GHGRP, under Subpart X, requires all domestic producers of petrochemicals to report annual
emissions and supplemental emissions information (e.g., production data, etc.) to facilitate verification of reported
emissions. Under EPA's GHGRP, most petrochemical production facilities are required to use either a mass balance
approach or CEMS to measure and report emissions for each petrochemical process unit to estimate facility-level
process C02 emissions; ethylene production facilities also have a third option. The mass balance method is used by
most facilities48 and assumes that all the carbon input is converted into primary and secondary products,
byproducts, or is emitted to the atmosphere as C02. To apply the mass balance, facilities must measure the volume
or mass of each gaseous and liquid feedstock and product, mass rate of each solid feedstock and product, and
carbon content of each feedstock and product for each process unit and sum for their facility. To apply the
optional combustion methodology, ethylene production facilities must measure the quantity, carbon content, and
molecular weight of the fuel to a stationary combustion unit when that fuel includes any ethylene process off-gas.
These data are used to calculate the total C02 emissions from the combustion unit. The facility must also estimate
the fraction of the emissions that is attributable to burning the ethylene process off-gas portion of the fuel. This
fraction is multiplied by the total emissions to estimate the emissions from ethylene production. The QA/QC and
Verification section below has a discussion of non-C02 emissions from ethylene production facilities.
All non-energy uses of residual fuel and some non-energy uses of "other oil" are assumed to be used in the
production of carbon black; therefore, consumption of these fuels is adjusted for within the Energy chapter to
avoid double-counting of emissions from fuel used in the carbon black production presented here within IPPU
sector. Additional information on the adjustments made within the Energy sector for Non-Energy Use of Fuels is
described in both the Methodology section of C02from Fossil Fuel Combustion (3.1 Fossil Fuel Combustion (IPCC
Source Category 1A)) and Annex 2.1, Methodology for Estimating Emissions of C02 from Fossil Fuel Combustion.
1990 through 2009
Prior to 2010, for each of these 4 types of petrochemical processes, an average national C02 emission factor was
calculated based on the GHGRP data and applied to production for earlier years in the time series (i.e., 1990
through 2009) to estimate C02 emissions from carbon black, ethylene, ethylene dichloride, and ethylene oxide
production. For carbon black, ethylene, ethylene dichloride, and ethylene oxide carbon dioxide emission factors
were derived from EPA's GHGRP data by dividing annual C02 emissions for petrochemical type "\" with annual
production for petrochemical type "i" and then averaging the derived emission factors obtained for each calendar
year 2010 through 2013 (EPA 2019). The years 2010 through 2013 were used in the development of carbon dioxide
emission factors as these years are more representative of operations in 1990 through 2009 for these facilities.
The average emission factors for each petrochemical type were applied across all prior years because
48 A few facilities producing ethylene dichloride, ethylene, and methanol used C02 CEMS; those C02 emissions have been
included in the aggregated GHGRP emissions presented here.
4-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
petrochemical production processes in the United States have not changed significantly since 1990, though some
operational efficiencies have been implemented at facilities over the time series.
The average country-specific C02 emission factors that were calculated from the GHGRP data are as follows:
•	2.59 metric tons C02/metric ton carbon black produced
•	0.79 metric tons C02/metric ton ethylene produced
•	0.040 metric tons C02/metric ton ethylene dichloride produced
•	0.46 metric tons C02/metric ton ethylene oxide produced
Annual production data for carbon black for 1990 through 2009 were obtained from the International Carbon
Black Association (Johnson 2003 and 2005 through 2010). Annual production data for ethylene, ethylene
dichloride, and ethylene oxide for 1990 through 2009 were obtained from the American Chemistry Council's
(ACC's) Business of Chemistry (ACC 2020).
Acrylonitrile
Carbon dioxide and methane emissions from acrylonitrile production were estimated using the Tier 1 method in
the 2006IPCC Guidelines. Annual acrylonitrile production data were used with IPCC default Tier 1C02 and CH4
emission factors to estimate emissions for 1990 through 2019. Emission factors used to estimate acrylonitrile
production emissions are as follows:
•	0.18 kg CH4/metric ton acrylonitrile produced
•	1.00 metric tons C02/metric ton acrylonitrile produced
Annual acrylonitrile production data for 1990 through 2019 were obtained from ACC's Business of Chemistry (ACC
2020). EPA is not able to apply the integrated aggregated facility-level GHGRP information for acrylonitrile
production needed for a Tier 2 approach. The aggregated information associated with production of these
petrochemicals did not meet criteria to shield underlying CBI from public disclosure.
Methanol
Carbon dioxide and methane emissions from methanol production were estimated using the Tier 1 method in the
2006 IPCC Guidelines. Annual methanol production data were used with IPCC default Tier 1C02 and CH4 emission
factors to estimate emissions for 1990 through 2019. Emission factors used to estimate methanol production
emissions are as follows:
•	2.3 kg CH4/metric ton methanol produced
•	0.67 metric tons C02/metric ton methanol produced
Annual methanol production data for 1990 through 2019 were obtained from the ACC's Business of Chemistry (ACC
2020). EPA is not able to apply the integrated aggregated facility-level GHGRP information for methanol
production needed for a Tier 2 approach. The aggregated information associated with production of these
petrochemicals did not meet criteria to shield underlying CBI from public disclosure.
Table 4-48: Production of Selected Petrochemicals (kt)
Chemical
1990
2005
2015
2016
2017
2018
2019
Carbon Black
1,310
1,650
1,220
1,190
1,240
1,280
1,210
Ethylene
16,500
24,000
26,900
26,600
27,800
30,500
32,400
Ethylene Dichloride
6,280
11,300
11,300
11,700
12,400
12,500
12,600
Ethylene Oxide
2,430
3,220
3,240
3,270
3,350
3,310
3,800
Acrylonitrile
1,210
1,330
1,050
960
1,040
1,250
1,080
Methanol
3,750
1,230
3,070
4,250
4,300
5,200
5,730
Industrial Processes and Product Use 4-63

-------
As noted earlier in the introduction section of the Petrochemical Production section, the allocation and reporting
of emissions from both fuels and feedstocks transferred out of the system for use in energy purposes to the Energy
chapter differs slightly from the 2006IPCC Guidelines. According to the 2006IPCC Guidelines, emissions from fuel
combustion from petrochemical production should be allocated to this source category within the IPPU chapter.
Due to national circumstances, EIA data on primary fuel for feedstock use within the energy balance are presented
by commodity only, with no resolution on data by industry sector (i.e., petrochemical production). In addition,
under EPA's GHGRP, reporting facilities began reporting in 2014 on annual feedstock quantities for mass balance
and CEMS methodologies (79 FR 63794), as well as the annual average carbon content of each feedstock (and
molecular weight for gaseous feedstocks) for the mass balance methodology beginning in reporting year 2017 (81
FR 89260).49 The United States is currently unable to report non-energy fuel use from petrochemical production
under the IPPU chapter due to CBI issues. Therefore, consistent with 2006 IPCC Guidelines, fuel consumption data
reported by EIA are modified to account for these overlaps to avoid double-counting. More information on the
non-energy use of fossil fuel feedstocks for petrochemical production can be found in Annex 2.3.
Uncertainty and Time-Series Consistency
The CH4 and C02 emission factors used for acrylonitrile and methanol production are based on a limited number of
studies. Using plant-specific factors instead of default or average factors could increase the accuracy of the
emission estimates; however, such data were not available for the current Inventory report.
The results of the quantitative uncertainty analysis for the C02 emissions from carbon black production, ethylene,
ethylene dichloride, and ethylene oxide are based on reported GHGRP data. Refer to the Methodology section for
more details on how these emissions were calculated and reported to EPA's GHGRP. There is some uncertainty in
the applicability of the average emission factors for each petrochemical type across all prior years. While
petrochemical production processes in the United States have not changed significantly since 1990, some
operational efficiencies have been implemented at facilities over the time series.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-49. Petrochemical
production C02 emissions from 2019 were estimated to be between 29.1 and 32.5 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 6 percent below to 6 percent above the emission
estimate of 30.8 MMT C02 Eq. Petrochemical production CH4 emissions from 2019 were estimated to be between
0.12 and 0.41 MMT C02 Eq. at the 95 percent confidence level. This indicates a range of approximately 57 percent
below to 47 percent above the emission estimate of 0.3 MMT C02 Eq.
Table 4-49: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from
Petrochemical Production and CO2 Emissions from Petrochemical Production (MMT CO2 Eq.
and Percent)
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3


(MMT CO? Eq.)
(MMTCOz Eq.)

(%)



Lower Upper
Lower
Upper



Bound Bound
Bound
Bound
Petrochemical
Production
C02
30.8
29.1 32.5
-6%
+6%
Petrochemical
Production
ch4
0.3
0.12 0.41
-57%
+47%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019.
49 See .
4-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
QA/QC and Verification
For Petrochemical Production, QA/QC activities were conducted consistent with the U.S. Inventory QA/QC plan, as
described in the QA/QC and Verification Procedures section of the IPPU chapter and Annex 8. Source-specific
quality control measures for this category included the QA/QC requirements and verification procedures of EPA's
GHGRP. More details on the greenhouse gas calculation, monitoring and QA/QC methods applicable to
petrochemical facilities can be found under Subpart X (Petrochemical Production) of the regulation (40 CFR Part
98).50 EPA verifies annual facility-level GHGRP reports through a multi-step process (e.g., combination of electronic
checks and manual reviews) to identify potential errors and ensure that data submitted to EPA are accurate,
complete, and consistent (EPA 2015).51 Based on the results of the verification process, EPA follows up with
facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with a number of
general and category-specific QC procedures, including range checks, statistical checks, algorithm checks, and year-
to-year checks of reported data and emissions. EPA also conducts QA checks of GHGRP reported production data
by petrochemical type against external datasets.
For ethylene, ethylene dichloride, and ethylene oxide it is possible to compare C02 emissions calculated using the
GHGRP data to the C02 emissions that would have been calculated using the Tier 1 approach if GHGRP data were
not available. For ethylene, the GHGRP emissions typically are within 5 percent of the emissions calculated using
the Tier 1 approach (except for 2018 and 2019 when the differences were 18 percent and 17 percent,
respectively). For ethylene dichloride, the GHGRP emissions are typically within 25 percent of the Tier 1 emissions.
For ethylene oxide, GHGRP emissions vary from 17 percent less than the Tier 1 emissions to 20 percent more than
the Tier 1 emissions, depending on the year.
EPA's GHGRP mandates that all petrochemical production facilities report their annual emissions of C02, CH4, and
N20 from each of their petrochemical production processes. Source-specific quality control measures for the
Petrochemical Production category included the QA/QC requirements and verification procedures of EPA's GHGRP.
The QA/QC requirements differ depending on the calculation methodology used.
As part of a planned improvement effort, EPA has assessed the potential of using GHGRP data to estimate CH4
emissions from ethylene production. As discussed in the Methodology section above, C02 emissions from ethylene
production in this chapter are based on data reported under the GHGRP, and these emissions are calculated using
a Tier 2 approach that assumes all of the carbon in the fuel (i.e., ethylene process off-gas) is converted to C02.
Ethylene production facilities also calculate and report CH4 emissions under the GHGRP when they use the optional
combustion methodology. The facilities calculate CH4 emissions from each combustion unit that burns off-gas from
an ethylene production process unit using a Tier 1 approach based on the total quantity of fuel burned, a default
higher heating value, and a default emission factor. Because multiple other types of fuel in addition to the ethylene
process unit off-gas may be burned in these combustion units, the facilities also report an estimate of the fraction
of emissions that is due to burning the ethylene process off-gas component of the total fuel. Multiplying the total
emissions by the estimated fraction provides an estimate of the CH4 emissions from the ethylene production
process unit. These ethylene production facilities also calculate CH4 emissions from flares that burn process vent
emissions from ethylene processes. The emissions are calculated using either a Tier 2 approach based on
measured gas volumes and measured carbon content or higher heating value, or a Tier 1 approach based on the
measured gas flow and a default emission factor. Nearly all ethylene production facilities use the optional
combustion methodology under the GHGRP, and the sum of reported CH4 emissions from combustion in stationary
combustion units and flares at all of these facilities is on the same order of magnitude as the combined CH4
emissions presented in this chapter from methanol and acrylonitrile production. The CH4 emissions from ethylene
production under the GHGRP have not been included in this chapter because this approach double counts carbon
(i.e., all of the carbon in the CH4 emissions is also included in the C02 emissions from the ethylene process units).
50	See .
51	See .
Industrial Processes and Product Use 4-65

-------
EPA continues to assess the GHGRP data for ways to better disaggregate the data and incorporate it into the
inventory.
These facilities are also required to report emissions of N20 from combustion of ethylene process off-gas in both
stationary combustion units and flares. Facilities using CEMS (consistent with a Tier 3 approach) are also required
to report emissions of CH4 and N20 from combustion of petrochemical process-off gases in flares. Preliminary
analysis of the aggregated reported CH4 and N20 emissions from facilities using CEMS and N20 emissions from
facilities using the optional combustion methodology suggests that these annual emissions are less than 500 kt/yr,
which is not significant enough to prioritize for inclusion in the report at this time. Pending resources and
significance, EPA may include these N20 emissions in future reports to enhance completeness.
Future QC efforts to validate the use of Tier 1 default emission factors and report on the comparison of Tier 1
emission estimates and GHGRP data are described below in the Planned Improvements section.
Recalculations Discussion
The 2018 data for production and emissions from carbon black, ethylene, ethylene dichloride, and ethylene oxide
have been updated with updated GHGRP data for 2018 for this Inventory (EPA 2020). These changes resulted in a
0.4 percent decrease in total petrochemical emissions for 2018, compared to the previous Inventory.
Planned Improvements
Improvements include completing category-specific QC of activity data and emission factors, along with further
assessment of CH4 and N20 emissions to enhance completeness in reporting of emissions from U.S. petrochemical
production, pending resources, significance and time-series consistency considerations. For example, EPA is
planning additional assessment of ways to use CH4 data from the GHGRP in the Inventory. One possible approach
EPA is assessing would be to adjust the C02 emissions from the GHGRP downward by subtracting the carbon that is
also included in the reported CH4 emissions, per the discussion in the Petrochemical Production QA/QC and
Verification section, above. As of this current report, timing and resources have not allowed EPA to complete this
analysis of activity data, emissions, and emission factors and remains a priority improvement within the IPPU
chapter.
Pending resources, a secondary potential improvement for this source category would focus on continuing to
analyze the fuel and feedstock data from EPA's GHGRP to better disaggregate energy-related emissions and
allocate them more accurately between the Energy and IPPU sectors of the Inventory. Some degree of double
counting may occur between C02 estimates of non-energy use of fuels in the energy sector and C02 process
emissions from petrochemical production in this sector. This is not considered to be a significant issue since the
non-energy use industrial release data includes different categories of sources than those included in this sector.
As noted previously in the methodology section, data integration is not feasible at this time as feedstock data from
the EIA used to estimate non-energy uses of fuels are aggregated by fuel type, rather than disaggregated by both
fuel type and particular industries. As described in the methodology section of this source category, EPA is
currently unable to use GHGRP-reported data on quantities of fuel consumed as feedstocks by petrochemical
producers, only feedstock type, due to the data failing GHGRP CBI aggregation criteria. Incorporating this data into
future Inventories will allow for easier data integration between the non-energy uses of fuels category and the
petrochemicals category presented in this chapter. This planned improvement is still under development and has
not been completed to report on progress in this current Inventory.
4-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
4.14 HCFC-22 Production (CRF Source
Category 2B9a)
Trifluoromethane (HFC-23 or CHF3) is generated as a byproduct during the manufacture of chlorodifluoromethane
(HCFC-22), which is primarily employed in refrigeration and air conditioning systems and as a chemical feedstock
for manufacturing synthetic polymers. Between 1990 and 2000, U.S. production of HCFC-22 increased significantly
as HCFC-22 replaced chlorofluorocarbons (CFCs) in many applications. Between 2000 and 2007, U.S. production
fluctuated but generally remained above 1990 levels. In 2008 and 2009, U.S. production declined markedly and has
remained near 2009 levels since. Because HCFC-22 depletes stratospheric ozone, its production for non-feedstock
uses is scheduled to be phased out by 2020 under the U.S. Clean Air Act.52 Feedstock production, however, is
permitted to continue indefinitely.
HCFC-22 is produced by the reaction of chloroform (CHCI3) and hydrogen fluoride (HF) in the presence of a catalyst,
SbCI5. The reaction of the catalyst and HF produces SbClxFy, (where x + y = 5), which reacts with chlorinated
hydrocarbons to replace chlorine atoms with fluorine. The HF and chloroform are introduced by submerged piping
into a continuous-flow reactor that contains the catalyst in a hydrocarbon mixture of chloroform and partially
fluorinated intermediates. The vapors leaving the reactor contain HCFC-21 (CHCI2F), HCFC-22 (CHCIF2), HFC-23
(CHF3), HCI, chloroform, and HF. The under-fluorinated intermediates (HCFC-21) and chloroform are then
condensed and returned to the reactor, along with residual catalyst, to undergo further fluorination. The final
vapors leaving the condenser are primarily HCFC-22, HFC-23, HCI and residual HF. The HCI is recovered as a useful
byproduct, and the HF is removed. Once separated from HCFC-22, the HFC-23 may be released to the atmosphere,
recaptured for use in a limited number of applications, or destroyed.
Two facilities produced HCFC-22 in the United States in 2019. Emissions of HFC-23 from this activity in 2019 were
estimated to be 3.7 MMT C02 Eq. (0.3 kt) (see Table 4-50). This quantity represents a 13 percent increase from
2018 emissions and a 92 percent decrease from 1990 emissions. The decrease from 1990 emissions was caused
primarily by changes in the HFC-23 emission rate (kg HFC-23 emitted/kg HCFC-22 produced). The increase from
2018 emissions was caused primarily by an increase in the HFC-23 emission rate at one plant. The long-term
decrease in the emission rate is primarily attributable to six factors: (a) five plants that did not capture and destroy
the HFC-23 generated have ceased production of HCFC-22 since 1990; (b) one plant that captures and destroys the
HFC-23 generated began to produce HCFC-22; (c) one plant implemented and documented a process change that
reduced the amount of HFC-23 generated; (d) the same plant began recovering HFC-23, primarily for destruction
and secondarily for sale; (e) another plant began destroying HFC-23; and (f) the same plant, whose emission rate
was higher than that of the other two plants, ceased production of HCFC-22 in 2013.
52 As construed, interpreted, and applied in the terms and conditions of the Montreal Protocol on Substances that Deplete the
Ozone Layer [42 U.S.C. §7671m(b), CAA §614],
Industrial Processes and Product Use 4-67

-------
Table 4-50: HFC-23 Emissions from HCFC-22 Production (MMT CCh Eq. and kt HFC-23)
Year
MMT CO? Eq.
kt HFC-23
1990
46.1
3
2005
20.0
1
2015
2016
2017
2018
2019
4.3
5.2
3.3
3.7
2.8
0.3
0.2
0.3
0.2
0.3
Methodology
To estimate HFC-23 emissions for five of the eight HCFC-22 plants that have operated in the United States since
1990, methods comparable to the Tier 3 methods in the 2006IPCC Guidelines (IPCC 2006) were used. Emissions for
2010 through 2019 were obtained through reports submitted by U.S. HCFC-22 production facilities to EPA's
Greenhouse Gas Reporting Program (GHGRP). EPA's GHGRP mandates that all HCFC-22 production facilities report
their annual emissions of HFC-23 from HCFC-22 production processes and HFC-23 destruction processes.
Previously, data were obtained by EPA through collaboration with an industry association that received voluntarily
reported HCFC-22 production and HFC-23 emissions annually from all U.S. HCFC-22 producers from 1990 through
2009. These emissions were aggregated and reported to EPA on an annual basis.
For the other three plants, the last of which closed in 1993, methods comparable to the Tier 1 method in the 2006
IPCC Guidelines were used. Emissions from these three plants have been calculated using the recommended
emission factor for unoptimized plants operating before 1995 (0.04 kg HCFC-23/kg HCFC-22 produced).
The five plants that have operated since 1994 measure (or, for the plants that have since closed, measured)
concentrations of HFC-23 as well as mass flow rates of process streams to estimate their generation of HFC-23.
Plants using thermal oxidation to abate their HFC-23 emissions monitor the performance of their oxidizers to verify
that the HFC-23 is almost completely destroyed. One plant that releases a small fraction of its byproduct HFC-23
periodically measures HFC-23 concentrations at process vents using gas chromatography. This information is
combined with information on quantities of products (e.g., HCFC-22) to estimate HFC-23 emissions.
To estimate 1990 through 2009 emissions, reports from an industry association were used that aggregated HCFC-
22 production and HFC-23 emissions from all U.S. HCFC-22 producers and reported them to EPA (ARAP 1997,1999,
2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, and 2010). To estimate 2010 through 2019
emissions, facility-level data (including both HCFC-22 production and HFC-23 emissions) reported through EPA's
GHGRP were analyzed. In 1997 and 2008, comprehensive reviews of plant-level estimates of HFC-23 emissions and
HCFC-22 production were performed (RTI1997; RTI 2008). The 1997 and 2008 reviews enabled U.S. totals to be
reviewed, updated, and where necessary, corrected, and also for plant-level uncertainty analyses (Monte-Carlo
simulations) to be performed for 1990,1995, 2000, 2005, and 2006. Estimates of annual U.S. HCFC-22 production
are presented in Table 4-51.
4-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-51: HCFC-22 Production (kt)
Year	kt_
1990	139
2005	156
2012	96
2013-2019	C_
C(CBI)
Note: HCFC-22 production in
2013 through 2019 is
considered Confidential
Business Information (CBI) as
there were only two producers
of HCFC-22 in those years.
Uncertainty and Time-Series Consistency
The uncertainty analysis presented in this section was based on a plant-level Monte Carlo Stochastic Simulation for
2006. The Monte Carlo analysis used estimates of the uncertainties in the individual variables in each plant's
estimating procedure. This analysis was based on the generation of 10,000 random samples of model inputs from
the probability density functions for each input. A normal probability density function was assumed for all
measurements and biases except the equipment leak estimates for one plant; a log-normal probability density
function was used for this plant's equipment leak estimates. The simulation for 2006 yielded a 95-percent
confidence interval for U.S. emissions of 6.8 percent below to 9.6 percent above the reported total.
The relative errors yielded by the Monte Carlo Stochastic Simulation for 2006 were applied to the U.S. emission
estimate for 2019. The resulting estimates of absolute uncertainty are likely to be reasonably accurate because (1)
the methods used by the two remaining plants to estimate their emissions are not believed to have changed
significantly since 2006, and (2) although the distribution of emissions among the plants has changed between
2006 and 2019 (because one plant has closed), the plant that currently accounts for most emissions had a relative
uncertainty in its 2006 (as well as 2005) emissions estimate that was similar to the relative uncertainty for total
U.S. emissions. Thus, the closure of one plant is not likely to have a large impact on the uncertainty of the national
emission estimate.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-52. HFC-23 emissions
from HCFC-22 production were estimated to be between 3.5 and 4.1 MMT C02 Eq. at the 95 percent confidence
level. This indicates a range of approximately 7 percent below and 10 percent above the emission estimate of 3.7
MMT C02 Eq.
Table 4-52: Approach 2 Quantitative Uncertainty Estimates for HFC-23 Emissions from
HCFC-22 Production (MMT CO2 Eq. and Percent)
Source
2019 Emission Estimate
Gas
Uncertainty Range Relative to Emission Estimate3
(MMTC02 Eq.)
(MMTC02 Eq.)
(%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
HCFC-22 Production
HFC-23 3.7
3.5 4.1
-7% +10%
a Range of emissions reflects a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions estimates
from 1990 through 2019. See Methods discussion of this section above.
Industrial Processes and Product Use 4-69

-------
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details). Under the GHGRP, EPA verifies annual facility-level
reports through a multi-step process (e.g., including a combination of pre-and post-submittal electronic checks and
manual reviews by staff) to identify potential errors and ensure that data submitted to EPA are accurate,
complete, and consistent (EPA 2015).53 Based on the results of the verification process, EPA follows up with
facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with a number of
general and category-specific QC procedures, including: range checks, statistical checks, algorithm checks, and
year-to-year checks of reported data and emissions.
The GHGRP also requires source-specific quality control measures for the HCFC-22 Production category. Under
EPA's GHGRP, HCFC-22 producers are required to (1) measure concentrations of HFC-23 and HCFC-22 in the
product stream at least weekly using equipment and methods (e.g., gas chromatography) with an accuracy and
precision of 5 percent or better at the concentrations of the process samples, (2) measure mass flows of HFC-23
and HCFC-22 at least weekly using measurement devices (e.g., flowmeters) with an accuracy and precision of 1
percent of full scale or better, (3) calibrate mass measurement devices at the frequency recommended by the
manufacturer using traceable standards and suitable methods published by a consensus standards organization,
(4) calibrate gas chromatographs at least monthly through analysis of certified standards, and (5) document these
calibrations.
4.15 Carbon Dioxide Consumption (CRF
Source Category 2B10)
Carbon dioxide (C02) is used for a variety of commercial applications, including food processing, chemical
production, carbonated beverage production, and refrigeration, and is also used in petroleum production for
enhanced oil recovery (EOR). C02 used for EOR is injected underground to enable additional petroleum to be
produced. For the purposes of this analysis, C02 used in commercial applications other than EOR is assumed to be
emitted to the atmosphere. A further discussion of C02 used in EOR is described in the Energy chapter in Box 3-6
titled "Carbon Dioxide Transport, Injection, and Geological Storage" and is not included in this section.
Carbon dioxide is produced from naturally-occurring C02 reservoirs, as a byproduct from the energy and industrial
production processes (e.g., ammonia production, fossil fuel combustion, ethanol production), and as a byproduct
from the production of crude oil and natural gas, which contain naturally occurring C02 as a component.
In 2019, the amount of C02 produced and captured for commercial applications and subsequently emitted to the
atmosphere was 4.9 MMT C02 Eq. (4,870 kt) (see Table 4-53). This is an 18 percent increase (740 kt) from 2018
levels and is an increase of approximately 231 percent since 1990.
Table 4-53: CO2 Emissions from CO2 Consumption (MMT CO2 Eq. and kt)
Year MMT CP2 Eq.	kt_
1990	1.5	1,472
2005	1.4	1,375
53 EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
4-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Year
MMTCOz Eq.
kt
2015
2016
2017
2018
2019
4.9
4.6
4.6
4.1
4.9
4,940
4,640
4,580
4,130
4,870
Methodology
Carbon dioxide emission estimates for 1990 through 2019 were based on the quantity of C02 extracted and
transferred for industrial applications (i.e., non-EOR end-uses). Some of the C02 produced by these facilities is used
for EOR, and some is used in other commercial applications (e.g., chemical manufacturing, food production). It is
assumed that 100 percent of the C02 production used in commercial applications other than EOR is eventually
released into the atmosphere.
For 2010 through 2019, data from EPA's GHGRP (Subpart PP) were aggregated from facility-level reports to
develop a national-level estimate for use in the Inventory (EPA 2020). Facilities report C02 extracted or produced
from natural reservoirs and industrial sites, and C02 captured from energy and industrial processes and transferred
to various end-use applications to EPA's GHGRP. This analysis includes only reported C02 transferred to food and
beverage end-uses. EPA is continuing to analyze and assess integration of C02 transferred to other end-uses to
enhance the completeness of estimates under this source category. Other end-uses include industrial applications,
such as metal fabrication. EPA is analyzing the information reported to ensure that other end-use data excludes
non-emissive applications and publication will not reveal CBI. Reporters subject to EPA's GHGRP Subpart PP are
also required to report the quantity of C02 that is imported and/or exported. Currently, these data are not publicly
available through the GHGRP due to data confidentiality reasons and hence are excluded from this analysis.
Facilities subject to Subpart PP of EPA's GHGRP are required to measure C02 extracted or produced. More details
on the calculation and monitoring methods applicable to extraction and production facilities can be found under
Subpart PP: Suppliers of Carbon Dioxide of the regulation, Part 98.54 The number of facilities that reported data to
EPA's GHGRP Subpart PP (Suppliers of Carbon Dioxide) for 2010 through 2019 is much higher (ranging from 44 to
53) than the number of facilities included in the Inventory for the 1990 to 2009 time period prior to the availability
of GHGRP data (4 facilities). The difference is largely due to the fact the 1990 to 2009 data includes only C02
transferred to end-use applications from naturally occurring C02 reservoirs and excludes industrial sites.
For 1990 through 2009, data from EPA's GHGRP are not available. For this time period, C02 production data from
four naturally-occurring C02 reservoirs were used to estimate annual C02 emissions. These facilities were Jackson
Dome in Mississippi, Bravo and West Bravo Domes in New Mexico, and McCallum Dome in Colorado. The facilities
in Mississippi and New Mexico produced C02 for use in both EOR and in other commercial applications (e.g.,
chemical manufacturing, food production). The fourth facility in Colorado (McCallum Dome) produced C02 for
commercial applications only (New Mexico Bureau of Geology and Mineral Resources 2006).
Carbon dioxide production data and the percentage of production that was used for non-EOR applications for the
Jackson Dome, Mississippi facility were obtained from Advanced Resources International (ARI 2006, 2007) for 1990
to 2000, and from the Annual Reports of Denbury Resources (Denbury Resources 2002 through 2010) for 2001 to
2009 (see Table 4-54). Denbury Resources reported the average C02 production in units of MMCF C02 per day for
54 See .
2010 through 2019
1990 through 2009
Industrial Processes and Product Use 4-71

-------
2001 through 2009 and reported the percentage of the total average annual production that was used for EOR.
Production from 1990 to 1999 was set equal to 2000 production, due to lack of publicly available production data
for 1990 through 1999. Carbon dioxide production data for the Bravo Dome and West Bravo Dome were obtained
from ARI for 1990 through 2009 (ARI1990 to 2010). Data for the West Bravo Dome facility were only available for
2009. The percentage of total production that was used for non-EOR applications for the Bravo Dome and West
Bravo Dome facilities for 1990 through 2009 were obtained from New Mexico Bureau of Geology and Mineral
Resources (Broadhead 2003; New Mexico Bureau of Geology and Mineral Resources 2006). Production data for the
McCallum Dome (Jackson County), Colorado facility were obtained from the Colorado Oil and Gas Conservation
Commission (COGCC) for 1999 through 2009 (COGCC 2014). Production data for 1990 to 1998 and percentage of
production used for EOR were assumed to be the same as for 1999, due to lack of publicly available data.
Table 4-54: CO2 Production (kt CO2) and the Percent Used for Non-EOR Applications
Year
Jackson Dome,
Bravo Dome,
West Bravo Dome,
McCallum Dome,
Total C02
%

MS
NM
NM
CO
Production
Non-

C02 Production
CO2 Production
C02 Production
C02 Production
from Extraction
EOR3

(kt) (% Non-EOR)
(kt) (% Non-EOR)
(kt) (% Non-EOR)
(kt) (% Non-EOR)
and Capture






Facilities (kt)

1990
1,344 (100%)
63 (1%)
+
65 (100%)
NA
NA
2005
1,254 (27%)
58(1%)
+
63 (100%)
NA
NA
2015
NA
NA
NA
NA
64,800b
8%
2016
NA
NA
NA
NA
55,900b
8%
2017
NA
NA
NA
NA
59,900b
8%
2018
NA
NA
NA
NA
58,400b
7%
2019
NA
NA
NA
NA
61,300b
8%
+ Does not exceed 0.5 percent.
NA (Not Available)
a Includes only food & beverage applications.
b For 2010 through 2019, the publicly available GHGRP data were aggregated at the national level based on GHGRP CBI
criteria.
Uncertainty and Time-Series Consistency
There is uncertainty associated with the data reported through EPA's GHGRP. Specifically, there is uncertainty
associated with the amount of C02 consumed for food and beverage applications, given the GHGRP does have
provisions that Subpart PP reporters are not required to report to the GHGRP if their emissions fall below certain
thresholds, in addition to the exclusion of the amount of C02 transferred to all other end-use categories. This latter
category might include C02 quantities that are being used for non-EOR industrial applications such as firefighting.
Second, uncertainty is associated with the exclusion of imports/exports data for C02 suppliers. Currently these
data are not publicly available through EPA's GHGRP and hence are excluded from this analysis. EPA verifies annual
facility-level reports through a multi-step process (e.g., combination of electronic checks and manual reviews by
staff) to identify potential errors and ensure that data submitted to EPA are accurate, complete, and consistent.
Based on the results of the verification process, EPA follows up with facilities to resolve mistakes that may have
occurred.55
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-55. Carbon dioxide
consumption C02 emissions for 2019 were estimated to be between 4.6 and 5.1 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 5 percent below to 5 percent above the emission
estimate of 4.9 MMT C02 Eq.
55 See .
4-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-55: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from CO2
Consumption (MMT CO2 Eq. and Percent)
Source Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.)
(MMTC02 Eq.)
(%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
C02 Consumption C02
4.9
4.6 5.1
-5% +5%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details). More details on the greenhouse gas calculation,
monitoring and QA/QC methods applicable to C02 Consumption can be found under Subpart PP (Suppliers of
Carbon Dioxide) of the regulation (40 CFR Part 98).56 EPA verifies annual facility-level GHGRP reports through a
multi-step process (e.g., combination of electronic checks and manual reviews) to identify potential errors and
ensure that data submitted to EPA are accurate, complete, and consistent (EPA 2015).57 Based on the results of the
verification process, EPA follows up with facilities to resolve mistakes that may have occurred. The post-submittals
checks are consistent with a number of general and category-specific QC procedures, including range checks,
statistical checks, algorithm checks, and year-to-year checks of reported data and emissions.
Recalculations Discussion
For the current Inventory, updated GHGRP Subpart PP data were available for the 2015 through 2018 portion of
the time series, resulting in recalculations for each of these years. Data from EPA's GHGRP (Subpart PP) were
previously unavailable for use for the years 2015 through 2018, so the emissions estimates for 2015 through 2018
had been held constant from 2014. Compared to the previous Inventory, emissions increased by 10 percent for
2015 (470 kt C02 Eq.), increased by 4 percent for 2016 (170 kt C02 Eq.), increased by 2 percent for 2017 (110 kt
C02 Eq.) and decreased by 8 percent for 2018 (340 kt C02 Eq.).
Planned Improvements
EPA will continue to evaluate the potential to include additional GHGRP data on other emissive end-uses to
improve the accuracy and completeness of estimates for this source category. Particular attention will be made to
ensuring time-series consistency of the emissions estimates presented in future Inventory reports, consistent with
IPCC and UNFCCC guidelines. This is required as the facility-level reporting data from EPA's GHGRP, with the
program's initial requirements for reporting of emissions in calendar year 2010, are not available for all inventory
years (i.e., 1990 through 2009) as required for this Inventory. In implementing improvements and integration of
data from EPA's GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories
will be relied upon.58
56	See .
57	See .
58	See .
Industrial Processes and Product Use 4-73

-------
These improvements are still in process and will be incorporated into future Inventory reports. These are near- to
medium-term improvements.
4.16 Phosphoric Acid Production (CRF
Source Category 2B10)
Phosphoric acid (H3PO4) is a basic raw material used in the production of phosphate-based fertilizers. Phosphoric
acid production from natural phosphate rock is a source of carbon dioxide (C02) emissions, due to the chemical
reaction of the inorganic carbon (calcium carbonate) component of the phosphate rock.
Phosphate rock is mined in Florida and North Carolina, which account for more than 75 percent of total domestic
output, and in Idaho and Utah (USGS 2020). It is used primarily as a raw material for wet-process phosphoric acid
production. The composition of natural phosphate rock varies, depending on the location where it is mined.
Natural phosphate rock mined in the United States generally contains inorganic carbon in the form of calcium
carbonate (limestone) and may also contain organic carbon.
The phosphoric acid production process involves chemical reaction of the calcium phosphate (Ca3(P04)2)
component of the phosphate rock with sulfuric acid (H2S04) and recirculated phosphoric acid (H3PO4) (EFMA 2000).
The generation of C02, however, is due to the associated limestone-sulfuric acid reaction, as shown below:
CaCO3 + //2S04 + H20 —* CaS04 ¦ 2H20 + C02
Total U.S. phosphate rock production used in 2019 was an estimated 23.0 million metric tons (USGS 2020). Total
imports of phosphate rock to the United States in 2019 were estimated to be approximately 2.0 million metric tons
(USGS 2020). Between 2015 and 2018, most of the imported phosphate rock (79 percent) came from Peru, with 20
percent from Morocco and 1 percent from other sources (USGS 2020). All phosphate rock mining companies in the
U.S. are vertically integrated with fertilizer plants that produce phosphoric acid located near the mines. The
phosphoric acid production facilities that use imported phosphate rock are located in Louisiana.
Over the 1990 to 2019 period, domestic phosphoric acid production has decreased by nearly 54 percent. Total C02
emissions from phosphoric acid production were 0.9 MMT C02 Eq. (891 kt C02) in 2019 (see Table 4-56). Domestic
consumption of phosphate rock in 2019 was estimated to have decreased 4 percent relative to 2018 levels (USGS
2020).
Table 4-56: CO2 Emissions from Phosphoric Acid Production (MMT CO2 Eq. and kt)
Year
MMT C02 Eq.
kt
1990
1.5
1,529

2005
1.3
1,342

2015
1.0
999
2016
1.0
998
2017
1.0
1,028
2018
0.9
940
2019
0.9
891
Methodology
Carbon dioxide emissions from production of phosphoric acid from phosphate rock are estimated by multiplying
the average amount of inorganic carbon (expressed as C02) contained in the natural phosphate rock as calcium
4-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
carbonate by the amount of phosphate rock that is used annually to produce phosphoric acid, accounting for
domestic production and net imports for consumption. The estimation methodology is as follows:
where,
Qpr
C02 emissions from phosphoric acid production, metric tons
Average amount of carbon (expressed as C02) in natural phosphate rock, metric ton
CO2/ metric ton phosphate rock
Quantity of phosphate rock used to produce phosphoric acid
The C02 emissions calculation methodology assumes that all of the inorganic C (calcium carbonate) content of the
phosphate rock reacts to produce C02 in the phosphoric acid production process and is emitted with the stack gas.
The methodology also assumes that none of the organic C content of the phosphate rock is converted to C02 and
that all of the organic C content remains in the phosphoric acid product. The United States uses a country-specific
methodology consistent with an IPCC Tier 1 approach to calculate emissions from production of phosphoric acid
from phosphate rock.59
From 1993 to 2004, the U.S. Geological Survey (USGS) Mineral Yearbook: Phosphate Rock disaggregated phosphate
rock mined annually in Florida and North Carolina from phosphate rock mined annually in Idaho and Utah, and
reported the annual amounts of phosphate rock exported and imported for consumption (see Table 4-57). For the
years 1990 through 1992, and 2005 through 2019, only nationally aggregated mining data was reported by USGS.
For the years 1990,1991, and 1992, the breakdown of phosphate rock mined in Florida and North Carolina, and
the amount mined in Idaho and Utah, are approximated using data reported by USGS for the average share of U.S.
production in those states from 1993 to 2004. For the years 2005 through 2019, the same approximation method
is used, but data for the share of U.S. production in those states were obtained from the USGS commodity
specialist for phosphate rock (USGS 2012). Data for domestic sales or consumption of phosphate rock, exports of
phosphate rock (primarily from Florida and North Carolina), and imports of phosphate rock for consumption for
1990 through 2010 were obtained from USGS Minerals Yearbook: Phosphate Rock (USGS 1994 through 2015b),
and from USGS Minerals Commodity Summaries: Phosphate Rock (USGS 2016 through 2020). From 2004 through
2019, the USGS reported no exports of phosphate rock from U.S. producers (USGS 2020).
The carbonate content of phosphate rock varies depending upon where the material is mined. Composition data
for domestically mined and imported phosphate rock were provided by the Florida Institute of Phosphate Research
(FIPR 2003a). Phosphate rock mined in Florida contains approximately 1 percent inorganic C, and phosphate rock
imported from Morocco contains approximately 1.46 percent inorganic C. Calcined phosphate rock mined in North
Carolina and Idaho contains approximately 0.41 percent and 0.27 percent inorganic C, respectively (see Table
4-58). Similar to the phosphate rock mined in Morocco, phosphate rock mined in Peru contains approximately 5
percent C02 (Golder Associates and M3 Engineering 2016).
Carbonate content data for phosphate rock mined in Florida are used to calculate the C02 emissions from
consumption of phosphate rock mined in Florida and North Carolina (more than 75 percent of domestic
production), and carbonate content data for phosphate rock mined in Morocco and Peru are used to calculate C02
emissions from consumption of imported phosphate rock. The C02 emissions calculation assumes that all of the
domestic production of phosphate rock is used in uncalcined form. As of 2006, the USGS noted that one phosphate
rock producer in Idaho produces calcined phosphate rock; however, no production data were available for this
single producer (USGS 2006). The USGS confirmed that no significant quantity of domestic production of
phosphate rock is in the calcined form (USGS 2012).
59 The 2006 IPCC Guidelines do not provide a method for estimating process emissions (C02) from Phosphoric Acid Production.
Industrial Processes and Product Use 4-75

-------
Table 4-57: Phosphate Rock Domestic Consumption, Exports, and Imports (kt)
Location/Year
1990
2005
2015
2016
2017
2018 2019
U.S. Domestic Consumption
49,800
35,200
26,200
26,700
26,300
23,300 23,000
FL and NC
42,494
28,160
20,960
21,360
21,040
18,640 18,400
ID and UT
7,306
7,040
5,240
5,340
5,260
4,660 4,600
Exports—FL and NC
6,240
0
0
0
0
0 0
Imports
451
2,630
1,960
1,590
2,470
2,770 2,000
Total U.S. Consumption
44,011
37,830
28,160
28,290
28,770
26,070 25,000
rable 4-58: Chemical Composition of Phosphate Rock (Percent by Weight)

Central
North
North Carolina
Idaho

Composition
Florida
Florida
(calcined)
(calcined) Morocco
Total Carbon (as C)
1.60
1.76
0.76

0.60
1.56
Inorganic Carbon (as C)
1.00
0.93
0.41

0.27
1.46
Organic Carbon (as C)
0.60
0.83
0.35

0.00
0.10
Inorganic Carbon (as C02)
3.67
3.43
1.50

1.00
5.00
Source: FIPR (2003a).
Uncertainty and Time-Series Consistency
Phosphate rock production data used in the emission calculations were developed by the USGS through monthly
and semiannual voluntary surveys of the active phosphate rock mines during 2019. Prior to 2006, USGS provided
the data disaggregated regionally; however, beginning in 2006, only total U.S. phosphate rock production was
reported. Regional production for 2019 was estimated based on regional production data from 2005 to 2011 and
multiplied by regionally-specific emission factors. There is uncertainty associated with the degree to which the
estimated 2019 regional production data represents actual production in those regions. Total U.S. phosphate rock
production data are not considered to be a significant source of uncertainty because all the domestic phosphate
rock producers report their annual production to the USGS. Data for exports of phosphate rock used in the
emission calculations are reported to the USGS by phosphate rock producers and are not considered to be a
significant source of uncertainty. Data for imports for consumption are based on international trade data collected
by the U.S. Census Bureau. These U.S. government economic data are not considered to be a significant source of
uncertainty.
An additional source of uncertainty in the calculation of C02 emissions from phosphoric acid production is the
carbonate composition of phosphate rock, as the composition of phosphate rock varies depending upon where the
material is mined and may also vary over time. The Inventory relies on one study (FIPR 2003a) of chemical
composition of the phosphate rock; limited data are available beyond this study. Another source of uncertainty is
the disposition of the organic carbon content of the phosphate rock. A representative of FIPR indicated that in the
phosphoric acid production process the organic C content of the mined phosphate rock generally remains in the
phosphoric acid product, which is what produces the color of the phosphoric acid product (FIPR 2003b). Organic
carbon is therefore not included in the calculation of C02 emissions from phosphoric acid production.
A third source of uncertainty is the assumption that all domestically-produced phosphate rock is used in
phosphoric acid production and used without first being calcined. Calcination of the phosphate rock would result
in conversion of some of the organic C in the phosphate rock into C02. However, according to air permit
information available to the public, at least one facility has calcining units permitted for operation (NCDENR 2013).
Finally, USGS indicated that in 2017 less than 5 percent of domestically-produced phosphate rock was used to
manufacture elemental phosphorus and other phosphorus-based chemicals, rather than phosphoric acid (USGS
2019b). According to USGS, there is only one domestic producer of elemental phosphorus, in Idaho, and no data
were available concerning the annual production of this single producer. Elemental phosphorus is produced by
reducing phosphate rock with coal coke, and it is therefore assumed that 100 percent of the carbonate content of
the phosphate rock will be converted to C02 in the elemental phosphorus production process. The calculation for
4-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
C02 emissions assumes that phosphate rock consumption, for purposes other than phosphoric acid production,
results in C02emissions from 100 percent of the inorganic carbon content in phosphate rock, but none from the
organic carbon content.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-59. 2019 phosphoric acid
production C02 emissions were estimated to be between 0.8 and 1.1 MMT C02 Eq. at the 95 percent confidence
level. This indicates a range of approximately 19 percent below and 21 percent above the emission estimate of 0.9
MMT C02 Eq.
Table 4-59: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Phosphoric Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Phosphoric Acid Production
C02
0.9
1
1
00
O
-19% +21%
3 Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions estimates
from 1990 through 2019. Details on the emission trends through time are described in more detail in the
Methodology section, above.
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with the U.S.
Inventory QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details).
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
EPA continues to evaluate potential improvements to the Inventory estimates for this source category, which
include direct integration of EPA's GHGRP data for 2010 through 2019 along with assessing applicability of
reported GHGRP data to update the inorganic C content of phosphate rock for prior years to ensure time series
consistency. Specifically, EPA would need to assess that averaged inorganic C content data (by region or other
approaches) meets GHGRP confidential business information (CBI) screening criteria. EPA would then need to
assess the applicability of GHGRP data for the averaged inorganic C content (by region or other approaches) from
2010 through 2019, along with other information to inform estimates in prior years in the required time series
(1990 through 2009) based on the sources of phosphate rock used in production of phosphoric acid over time. In
implementing improvements and integration of data from EPA's GHGRP, the latest guidance from the IPCC on the
use of facility-level data in national inventories will be relied upon.60 These long-term planned improvements are
still in development by EPA and have not been implemented into the current Inventory report.
60 See .
Industrial Processes and Product Use 4-77

-------
4.17 Iron and Steel Production (CRF Source
Category 2C1) and Metallurgical Coke
Production
Iron and steel production is a multi-step process that generates process-related emissions of carbon dioxide (C02)
and methane (CH4) as raw materials are refined into iron and then transformed into crude steel. Emissions from
conventional fuels (e.g., natural gas, fuel oil) consumed for energy purposes during the production of iron and steel
are accounted for in the Energy chapter.
Iron and steel production includes six distinct production processes: coke production, sinter production, direct
reduced iron (DRI) production, pig iron61 production, electric arc furnace (EAF) steel production, and basic oxygen
furnace (BOF) steel production. The number of production processes at a particular plant is dependent upon the
specific plant configuration. Most process C02 generated from the iron and steel industry is a result of the
production of crude iron.
In addition to the production processes mentioned above, C02 is also generated at iron and steel mills through the
consumption of process byproducts (e.g., blast furnace gas, coke oven gas) used for various purposes including
heating, annealing, and electricity generation. Process byproducts sold for use as synthetic natural gas are also
included in these calculations. In general, C02 emissions are generated in these production processes through the
reduction and consumption of various carbon-containing inputs (e.g., ore, scrap, flux, coke byproducts). Fugitive
CH4 emissions can also be generated from these processes, as well as from sinter, direct iron, and pellet
production.
Currently, there are approximately nine integrated iron and steel steelmaking facilities that utilize BOFs to refine
and produce steel from iron. As of 2018, these facilities have 21 active blast furnaces between them. Almost 100
steelmaking facilities utilize EAFs to produce steel primarily from recycled ferrous scrap (USGS 2019). The trend in
the United States for integrated facilities has been a shift towards fewer BOFs and more EAFs. EAFs use scrap steel
as their main input and use significantly less energy than BOFs. There are also 14 cokemaking facilities, of which 3
facilities are co-located with integrated iron and steel facilities (ACCCI 2020). In the United States, four states
account for roughly 51 percent of total raw steel production: Indiana, Ohio, Michigan, and Pennsylvania (USGS
2019).
Total annual production of crude steel in the United States was fairly constant between 2000 and 2008 and ranged
from a low of 99,320,000 tons to a high of 109,880,000 tons (2001 and 2004, respectively). Due to the decrease in
demand caused by the global economic downturn (particularly from the automotive industry), crude steel
production in the United States sharply decreased to 65,459,000 tons in 2009. Crude steel production was fairly
constant from 2011 through 2014, and after a dip in production from 2014 to 2015, crude steel production has
slowly and steadily increased for the past few years. The United States was the fourth largest producer of raw steel
in the world, behind China, India and Japan, accounting for approximately 4.6 percent of world production in 2019
(AISI 2004 through 2020).
The majority of C02 emissions from the iron and steel production process come from the use of metallurgical coke
in the production of pig iron and from the consumption of other process byproducts, with lesser amounts emitted
from the use of flux and from the removal of carbon from pig iron used to produce steel.
61 Pig iron is the common industry term to describe what should technically be called crude iron. Pig iron is a subset of crude
iron that has lost popularity over time as industry trends have shifted. Throughout this report, pig iron will be used
interchangeably with crude iron, but it should be noted that in other data sets or reports pig iron and crude iron may not be
used interchangeably and may provide different values.
4-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
According to the 2006IPCC Guidelines, the production of metallurgical coke from coking coal is considered to be an
energy use of fossil fuel, and the use of coke in iron and steel production is considered to be an industrial process
source. Therefore, the 2006 IPCC Guidelines suggest that emissions from the production of metallurgical coke
should be reported separately in the Energy sector, while emissions from coke consumption in iron and steel
production should be reported in the Industrial Processes and Product Use sector. The approaches and emission
estimates for both metallurgical coke production and iron and steel production, however, are presented here
because much of the relevant activity data is used to estimate emissions from both metallurgical coke production
and iron and steel production. For example, some byproducts (e.g., coke oven gas) of the metallurgical coke
production process are consumed during iron and steel production, and some byproducts of the iron and steel
production process (e.g., blast furnace gas) are consumed during metallurgical coke production. Emissions
associated with the consumption of these byproducts are attributed at the point of consumption. Emissions
associated with the use of conventional fuels (e.g., natural gas, fuel oil) for electricity generation, heating and
annealing, or other miscellaneous purposes downstream of the iron and steelmaking furnaces are reported in the
Energy chapter.
Metallurgical Coke Production
Emissions of C02 from metallurgical coke production in 2019 were 1.4 MMT C02 Eq. (1,366 kt C02) (see Table 4-60
and Table 4-61). Emissions increased slightly in 2019 by 7 percent from 2018 levels and have decreased by 76
percent (4.2 MMT C02 Eq.) since 1990. Coke production in 2019 was about 2 percent lower than in 2018 and 51
percent below 1990.
Table 4-60: CO2 Emissions from Metallurgical Coke Production (MMT CO2 Eq.)
Gas	1990	2005	2015 2016 2017 2018 2019
C02	5J5	3J)	4.4 2.6 2.0 1.3 1.4
Total	516	3^9	4.4 2.6 2.0 1.3 1.4
Table 4-61: CO2 Emissions from Metallurgical Coke Production (kt)
Gas 1990 2005	2015 2016 2017 2018 2019
C02 5,608 A 3,921 A 4,417 2,643 1,978 1,282 1,366
Total 5,608 3,921	4,417 2,643 1,978 1,282 1,366
Iron and Steel Production
Emissions of C02 and CH4 from iron and steel production in 2019 were 39.9 MMT C02 Eq. (39,944 kt) and 0.0077
MMT C02 Eq. (0.3 kt CH4), respectively (see Table 4-62 through Table 4-65), totaling approximately 39.9 MMT C02
Eq. Emissions slightly decreased in 2019 from 2018 and have decreased overall since 1990 due to restructuring of
the industry, technological improvements, and increased scrap steel utilization. Carbon dioxide emission estimates
include emissions from the consumption of carbonaceous materials in the blast furnace, EAF, and BOF, as well as
blast furnace gas and coke oven gas consumption for other activities at the steel mill.
In 2019, domestic production of pig iron decreased by 7 percent from 2018 levels. Overall, domestic pig iron
production has declined since the 1990s. Pig iron production in 2019 was 53 percent lower than in 2000 and 55
percent below 1990. Carbon dioxide emissions from iron production have decreased by 80 percent since 1990.
Carbon dioxide emissions from steel production have decreased by 28 percent (2.2 MMT C02 Eq.) since 1990,
while overall C02 emissions from iron and steel production have declined by 60 percent (59.2 MMT C02 Eq.) from
1990 to 2019.
Table 4-62: CO2 Emissions from Iron and Steel Production (MMT CO2 Eq.)
Source/Activity Data	1990	2005	2015 2016 2017 2018 2019
Sinter Production	2.4	1.7	1.0	0.9	0.9	0.9	0.9
Industrial Processes and Product Use 4-79

-------
Iron Production
45.7
17.7
10.3
9.9
8.2
9.6
9.3
Pellet Production
1.8
1.5
1.0
0.9
0.9
0.9
0.9
Steel Production
8.0
9.4
6.9
6.9
6.2
5.8
5.8
Other Activities3
41.2
35.9
24.3
22.5
22.4
24.1
23.2
Total
99.1
66.2
43.5
41.0
38.6
41.3
39.9
Note: Totals may not sum due to independent rounding.
a Includes emissions from blast furnace gas and coke oven gas combustion for activities at the steel mill
other than consumption in blast furnace, EAFs, or BOFs.
Table 4-63: CO2 Emissions from Iron and Steel Production (kt)
Source/Activity Data
1990
2005
2015
2016
2017
2018
2019
Sinter Production
2,448
1,663
1,016
877
869
937
876
Iron Production
45,701
17,660
10,330
9,928
8,236
9,580
9,273
Pellet Production
1,817
1,503
964
869
867
924
867
Steel Production
7,964
9,395
6,935
6,854
6,218
5,754
5,770
Other Activities3
41,194
35,934
24,280
22,451
22,396
24,149
23,158
Total
99,124
66,155
43,525
40,979
38,587
41,345
39,944
Note: Totals may not sum due to independent rounding.
a Includes emissions from blast furnace gas and coke oven gas combustion for activities at the
steel mill other than consumption in blast furnace, EAFs, or BOFs.
Table 4-64: ChU Emissions from Iron and Steel Production (MMT CO2 Eq.)
Source/Activity Data
1990
2005
2015
2016
2017
2018
2019
Sinter Production
+
+
+
+
+
+
+
Total
+
+
+
+
+
+
+
+ Does not exceed 0.05 MMT C02 Eq.
Table 4-65: ChU Emissions from Iron and Steel Production (kt)
Source/Activity Data
1990
2005
2015
2016
2017
2018
2019
Sinter Production
0.9
0.6
0.3
0.3
0.3
0.3
0.3
Methodology
Emission estimates presented in this chapter utilize a country-specific approach based on Tier 2 methodologies
provided by the 2006IPCC Guidelines. These Tier 2 methodologies call for a mass balance accounting of the
carbonaceous inputs and outputs during the iron and steel production process and the metallurgical coke
production process. Tier 1 methods are used for certain iron and steel production processes (i.e., sinter
production, pellet production and DRI production) for which available data are insufficient to apply a Tier 2
method (e.g., country-specific carbon contents of inputs and outputs are not known). However, sinter production,
pellet production and DRI production only account for roughly 8 percent of total iron and steel production
emissions and therefore the majority of category emissions are captured with higher tier estimates.
The Tier 2 methodology equation is as follows:
Em- —
^(<2a X Q) — X Q)
44
12
where,
ECo2 =	Emissions from coke, pig iron, EAF steel, or BOF steel production, metric tons
a	=	Input material a
4-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
b	=	Output material b
Qa	=	Quantity of input material a, metric tons
Ca	=	Carbon content of input material a, metric tons C/metric ton material
Qb	=	Quantity of output material b, metric tons
Cb	=	Carbon content of output material b, metric tons C/metric ton material
44/12	=	Stoichiometric ratio of C02 to C
The Tier 1 methodology equations are as follows:
ES,P = QSX EFs,p
Ed,C02 = Qd X EFd,c02
Ep,C02 = Qp X EFp,co2
Emissions from sinter production process for pollutant p (C02 or CH4), metric ton
Quantity of sinter produced, metric tons
Emission factor for pollutant p (C02 or CH4), metric ton p/metric ton sinter
Emissions from DRI production process for C02, metric ton
Quantity of DRI produced, metric tons
Emission factor for C02, metric ton C02/metric ton DRI
Quantity of pellets produced, metric tons
Emission factor for C02, metric ton C02/metric ton pellets produced
where,
Es,p =
a
EFs,P
Ed,C02 =
Qd
EFd,co2 =
QP
EFPico2 =
Metallurgical Coke Production
Coking coal is used to manufacture metallurgical coke which is used primarily as a reducing agent in the production
of iron and steel, but is also used in the production of other metals including zinc and lead (see Zinc Production and
Lead Production sections of this chapter). Emissions associated with producing metallurgical coke from coking coal
are estimated and reported separately from emissions that result from the iron and steel production process. To
estimate emissions from metallurgical coke production, a Tier 2 method provided by the 2006IPCC Guidelines was
utilized. The amount of carbon contained in materials produced during the metallurgical coke production process
(i.e., coke, coke breeze and coke oven gas) is deducted from the amount of carbon contained in materials
consumed during the metallurgical coke production process (i.e., natural gas, blast furnace gas, and coking coal).
Light oil, which is produced during the metallurgical coke production process, is excluded from the deductions due
to data limitations. The amount of carbon contained in these materials is calculated by multiplying the material-
specific carbon content by the amount of material consumed or produced (see Table 4-66). The amount of coal tar
produced was approximated using a production factor of 0.03 tons of coal tar per ton of coking coal consumed.
The amount of coke breeze produced was approximated using a production factor of 0.075 tons of coke breeze per
ton of coking coal consumed (AISI 2008; DOE 2000). Data on the consumption of carbonaceous materials (other
than coking coal) as well as coke oven gas production were available for integrated steel mills only (i.e., steel mills
with co-located coke plants). Therefore, carbonaceous material (other than coking coal) consumption and coke
oven gas production were excluded from emission estimates for merchant coke plants. Carbon contained in coke
oven gas used for coke-oven underfiring was not included in the deductions to avoid double-counting.
Industrial Processes and Product Use 4-81

-------
Table 4-66: Material Carbon Contents for Metallurgical Coke Production
Material
kg C/kg
Coal Tara
0.62
Cokea
0.83
Coke Breeze3
0.83
Coking Coalb
0.75
Material
kg C/GJ
Coke Oven Gasc
12.1
Blast Furnace Gasc
70.8
a Source: IPCC (2006), Vol. 3 Chapter 4, Table 4.3
b Source: EIA (2017c)
c Source: IPCC (2006), Vol. 2 Chapter 1, Table 1.3
Although the 2006 IPCC Guidelines provide a Tier 1 CH4 emission factor for metallurgical coke production (i.e., 0.1 g
CH4 per metric ton of coke production), it is not appropriate to use because C02 emissions were estimated using
the Tier 2 mass balance methodology. The mass balance methodology makes a basic assumption that all carbon
that enters the metallurgical coke production process either exits the process as part of a carbon-containing
output or as C02 emissions. This is consistent with a preliminary assessment of aggregated facility-level
greenhouse gas CH4 emissions reported by coke production facilities under EPA's GHGRP. The assessment indicates
that CH4 emissions from coke production are insignificant and below 500 kt or 0.05 percent of total national
emissions. Pending resources and significance, EPA continues to assess the possibility of including these emissions
in future Inventories to enhance completeness but has not incorporated these emissions into this report.
Data relating to the mass of coking coal consumed at metallurgical coke plants and the mass of metallurgical coke
produced at coke plants were taken from the Energy Information Administration (EIA) Quarterly Coal Report:
October through December (EIA 1998 through 2019) (see Table 4-67). Data on the volume of natural gas
consumption, blast furnace gas consumption, and coke oven gas production for metallurgical coke production at
integrated steel mills were obtained from the American Iron and Steel Institute (AISI) Annual Statistical Report
(AISI 2004 through 2020) and through personal communications with AISI (AISI 2008) (see Table 4-68). The factor
for the quantity of coal tar produced per ton of coking coal consumed was provided by AISI (AISI 2008). The factor
for the quantity of coke breeze produced per ton of coking coal consumed was obtained through Table 2-1 of the
report Energy and Environmental Profile of the U.S. Iron and Steel Industry (DOE 2000). Currently, data on natural
gas consumption and coke oven gas production at merchant coke plants were not available and were excluded
from the emission estimate. Carbon contents for metallurgical coke, coal tar, coke oven gas, and blast furnace gas
were provided by the 2006 IPCC Guidelines. The C content for coke breeze was assumed to equal the C content of
coke. Carbon contents for coking coal was from EIA.
Table 4-67: Production and Consumption Data for the Calculation of CO2 Emissions from
Metallurgical Coke Production (Thousand Metric Tons)
Source/Activity Data
1990
2005
2015
2016
2017
2018
2019
Metallurgical Coke Production







Coking Coal Consumption at Coke Plants
35,269
21,259
17,879
14,955
15,910
16,635
16,261
Coke Production at Coke Plants
25,054
15,167
12,479
10,755
11,746
12,525
12,215
Coal Breeze Production
2,645
1,594
1,341
1,122
1,193
1,248
1,220
Coal Tar Production
1,058
638
536
449
477
499
488
Table 4-68: Production and Consumption Data for the Calculation of CO2 Emissions from
Metallurgical Coke Production (Million ft3)
Source/Activity Data	1990 2005	2015	2016	2017 2018	2019
Metallurgical Coke Production
Coke Oven Gas Production 250,767 114,213 84,336	74,807	74,997	80,750	77,692
Natural Gas Consumption 599 , ,. 2,996 2,338	2,077	2,103	2,275	2,189
4-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Blast Furnace Gas Consumption
24,602
4,460
4,185 3,741 3,683 4,022 3,914
Iron and Steel Production
To estimate emissions from pig iron production in the blast furnace, the amount of carbon contained in the
produced pig iron and blast furnace gas were deducted from the amount of carbon contained in inputs (i.e.,
metallurgical coke, sinter, natural ore, pellets, natural gas, fuel oil, coke oven gas, carbonate fluxes or slagging
materials, and direct coal injection). The carbon contained in the pig iron, blast furnace gas, and blast furnace
inputs was estimated by multiplying the material-specific C content by each material type (see Table 4-69). Carbon
in blast furnace gas used to pre-heat the blast furnace air is combusted to form C02 during this process. Carbon
contained in blast furnace gas used as a blast furnace input was not included in the deductions to avoid double-
counting.
Emissions from steel production in EAFs were estimated by deducting the carbon contained in the steel produced
from the carbon contained in the EAF anode, charge carbon, and scrap steel added to the EAF. Small amounts of
carbon from DRI and pig iron to the EAFs were also included in the EAF calculation. For BOFs, estimates of carbon
contained in BOF steel were deducted from C contained in inputs such as natural gas, coke oven gas, fluxes (e.g.
burnt lime or dolomite), and pig iron. In each case, the carbon was calculated by multiplying material-specific
carbon contents by each material type (see Table 4-69). For EAFs, the amount of EAF anode consumed was
approximated by multiplying total EAF steel production by the amount of EAF anode consumed per metric ton of
steel produced (0.002 metric tons EAF anode per metric ton steel produced [AISI 2008]). The amount of flux (e.g.,
burnt lime or dolomite) used in pig iron production was deducted from the "Other Process Uses of Carbonates"
source category (CRF Source Category 2A4) to avoid double-counting.
Carbon dioxide emissions from the consumption of blast furnace gas and coke oven gas for other activities
occurring at the steel mill were estimated by multiplying the amount of these materials consumed for these
purposes by the material-specific carbon content (see Table 4-69).
Carbon dioxide emissions associated with the sinter production, direct reduced iron production, pig iron
production, steel production, and other steel mill activities were summed to calculate the total C02 emissions from
iron and steel production (see Table 4-62 and Table 4-63).
Table 4-69: Material Carbon Contents for Iron and Steel Production
Material
kg C/kg
Coke
0.83
Direct Reduced Iron
0.02
Dolomite
0.13
EAF Carbon Electrodes
0.82
EAF Charge Carbon
0.83
Limestone
0.12
Pig Iron
0.04
Steel
0.01
Material
kg C/GJ
Coke Oven Gas
12.1
Blast Furnace Gas
70.8
Source: IPCC (2006), Table 4.3. Coke Oven Gas and
Blast Furnace Gas, Table 1.3.
The production process for sinter results in fugitive emissions of CH4, which are emitted via leaks in the production
equipment, rather than through the emission stacks or vents of the production plants. The fugitive emissions were
calculated by applying Tier 1 emission factors taken from the 2006 IPCC Guidelines for sinter production (see Table
4-70). Although the 1995 IPCC Guidelines (IPCC/UNEP/OECD/IEA 1995) provide a Tier 1CH4 emission factor for pig
iron production, it is not appropriate to use because C02 emissions were estimated using the Tier 2 mass balance
methodology. The mass balance methodology makes a basic assumption that all carbon that enters the pig iron
production process either exits the process as part of a carbon-containing output or as C02 emissions; the
Industrial Processes and Product Use 4-83

-------
estimation of CH4 emissions is precluded. Annual analysis of facility-level emissions reported during iron
production further supports this assumption and indicates that CH4 emissions are below 500 kt C02 Eq. and well
below 0.05 percent of total national emissions. The production of direct reduced iron also results in emissions of
CH4 through the consumption of fossil fuels (e.g., natural gas, etc.); however, these emission estimates are
excluded due to data limitations. Pending further analysis and resources, EPA may include these emissions in
future reports to enhance completeness. EPA is still assessing the possibility of including these emissions in future
reports and have not included this data in the current report.
Table 4-70: ChU Emission Factors for Sinter and Pig Iron Production
Material Produced
Factor
Unit
Sinter
0.07
kg CH4/metric ton
Source: IPCC (2006), Table 4.2.
Emissions of C02 from sinter production, direct reduced iron production and pellet production were estimated by
multiplying total national sinter production and the total national direct reduced iron production by Tier 1 C02
emission factors (see Table 4-71). Because estimates of sinter production, direct reduced iron production, and
pellet production were not available, production was assumed to equal consumption.
Table 4-71: CO2 Emission Factors for Sinter Production, Direct Reduced Iron Production, and
Pellet Production

Metric Ton C02/Metric
Material Produced
Ton
Sinter
0.2
Direct Reduced Iron
0.7
Pellet Production
0.03
Source: IPCC (2006), Table 4.1.
The consumption of coking coal, natural gas, distillate fuel, and coal used in iron and steel production are adjusted
for within the Energy chapter to avoid double-counting of emissions reported within the IPPU chapter as these
fuels were consumed during non-energy related activities. More information on this methodology and examples of
adjustments made between the IPPU and Energy chapters are described in Annex 2.1, Methodology for Estimating
Emissions of C02 from Fossil Fuel Combustion.
Sinter consumption and pellet consumption data for 1990 through 2019 were obtained from AISI's Annual
Statistical Report (AISI 2004 through 2020) and through personal communications with AISI (AISI 2008) (see Table
4-72). In general, direct reduced iron (DRI) consumption data were obtained from the U.S. Geological Survey
(USGS) Minerals Yearbook- Iron and Steel Scrap (USGS 1991 through 2017) and personal communication with the
USGS Iron and Steel Commodity Specialist (Tuck 2020); however, data for DRI consumed in EAFs were not available
for the years 1990 and 1991. EAF DRI consumption in 1990 and 1991 was calculated by multiplying the total DRI
consumption for all furnaces by the EAF share of total DRI consumption in 1992. Also, data for DRI consumed in
BOFs were not available for the years 1990 through 1993. BOF DRI consumption in 1990 through 1993 was
calculated by multiplying the total DRI consumption for all furnaces (excluding EAFs and cupola) by the BOF share
of total DRI consumption (excluding EAFs and cupola) in 1994.
The Tier 1 C02 emission factors for sinter production, direct reduced iron production and pellet production were
obtained through the 2006 IPCC Guidelines (IPCC 2006). Time-series data for pig iron production, coke, natural gas,
fuel oil, sinter, and pellets consumed in the blast furnace; pig iron production; and blast furnace gas produced at
the iron and steel mill and used in the metallurgical coke ovens and other steel mill activities were obtained from
AISI's Annual Statistical Report (AISI 2004 through 2020) and through personal communications with AISI (AISI
2008) (see Table 4-72 and Table 4-73).
Data for EAF steel production, flux, EAF charge carbon, and natural gas consumption were obtained from AISI's
Annual Statistical Report (AISI 2004 through 2020) and through personal communications with AISI (AISI 2006
through 2016, AISI 2008). The factor for the quantity of EAF anode consumed per ton of EAF steel produced was
4-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
provided by AISI (AISI 2008). Data for BOF steel production, flux, natural gas, natural ore, pellet, sinter
consumption as well as BOF steel production were obtained from AISI's Annual Statistical Report (AISI 2004
through 2020) and through personal communications with AISI (AISI 2008). Data for EAF and BOF scrap steel, pig
iron, and DRI consumption were obtained from the USGS Minerals Yearbook- Iron and Steel Scrap (USGS 1991
through 2017). Data on coke oven gas and blast furnace gas consumed at the iron and steel mill (other than in the
EAF, BOF, or blast furnace) were obtained from AISI's Annual Statistical Report (AISI 2004 through 2020) and
through personal communications with AISI (AISI 2008).
Data on blast furnace gas and coke oven gas sold for use as synthetic natural gas were obtained from ElA's Natural
Gas Annual (EIA 2020). Carbon contents for direct reduced iron, EAF carbon electrodes, EAF charge carbon,
limestone, dolomite, pig iron, and steel were provided by the 2006IPCC Guidelines. The carbon contents for
natural gas, fuel oil, and direct injection coal were obtained from EIA (EIA 2017c) and EPA (EPA 2010). Heat
contents for fuel oil and direct injection coal were obtained from EIA (EIA 1992, 2011); natural gas heat content
was obtained from Table 37 of AISI's Annual Statistical Report (AISI 2004 through 2020). Heat contents for coke
oven gas and blast furnace gas were provided in Table 37 of AISI's Annual Statistical Report (AISI 2004 through
2020) and confirmed by AISI staff (Carroll 2016).
Table 4-72: Production and Consumption Data for the Calculation of CO2 and ChU Emissions
from Iron and Steel Production (Thousand Metric Tons)
Source/Activity Data
1990
2005
2015
2016
2017
2018
2019
Sinter Production
Sinter Production
Direct Reduced Iron Production
Direct Reduced Iron Production
Pellet Production
Pellet Production
Pig Iron Production
Coke Consumption
Pig Iron Production
Direct Injection Coal
Consumption
EAF Steel Production
EAF Anode and Charge Carbon
Consumption
Scrap Steel Consumption
Flux Consumption
EAF Steel Production
BOF Steel Production
Pig Iron Consumption
Scrap Steel Consumption
Flux Consumption
BOF Steel Production
12,239
517
60,563
24,946
49,669
1,485
67
42,691
319
33,511
47,307
14,713
576
43,973
8,315
1,303
50,096
13,832
37,222
2,573
1,127
46,600
695
52,194
34,400
11,400
582
42,705
4,347
C
32,146 28,967 28,916
5,079 4,385
2,722	C
7,969 7,124 7,101
25,436 22,293 22,395
2,275
1,072
44,000
998
1,935
1,120
C
998
2,125
1,127
C
998
4,687	4,378
C	C
30,793	28,916
7,618	7,291
24,058	22,302
2,569	2,465
1,133
C
998
20,300
4,530
454
29,396 25,8
C
C
408
C
C
408
25,788
C
C
408
1,137
C
998
49,451 52,589 55,825 58,904 61,172
C
C
363
27,704 26,591
C-Confidential
Table 4-73: Production and Consumption Data for the Calculation of CO2 Emissions from
Iron and Steel Production (Million ft3 unless otherwise specified)
Source/Activity Data
1990
2005
2015
2016
2017
2018
2019
Pig Iron Production
Natural Gas Consumption	56,273
Fuel Oil Consumption
(thousand gallons)	163,397
Coke Oven Gas Consumption	22,033
Blast Furnace Gas Production	1,439,380
59,844
16,170
16,557
1,299,980
43,294 38,396 38,142 40,204 37,934
9,326 6,124 4,352
13,921 12,404 12,459
874,670 811,005 808,499
3,365 2,321
13,337 12,926
871,860 836,033
Industrial Processes and Product Use 4-85

-------
EAF Steel Production
Natural Gas Consumption
15,905
19,985
8,751
3,915
8,105
8,556
9,115
BOF Steel Production







Coke Oven Gas Consumption
3,851
524
386
367
374
405
389
Other Activities







Coke Oven Gas Consumption
224,883
97,132
70,029
62,036
62,164
67,008
64,377
Blast Furnace Gas







Consumption
1,414,778
1,295,520
870,485
807,264
804,816
867,838
832,119
Uncertainty and Time-Series Consistency
The estimates of C02 emissions from metallurgical coke production are based on assessing uncertainties in
material production and consumption data and average carbon contents. Uncertainty is associated with the total
U.S. coking coal consumption, total U.S. coke production, and materials consumed during this process. Data for
coking coal consumption and metallurgical coke production are from different data sources (EIA) than data for
other carbonaceous materials consumed at coke plants (AISI), which does not include data for merchant coke
plants. There is uncertainty associated with the fact that coal tar and coke breeze production were estimated
based on coke production because coal tar and coke breeze production data were not available. Since merchant
coke plant data is not included in the estimate of other carbonaceous materials consumed at coke plants, the mass
balance equation for C02 from metallurgical coke production cannot be reasonably completed. Therefore, for the
purpose of this analysis, uncertainty parameters are applied to primary data inputs to the calculation (i.e., coking
coal consumption and metallurgical coke production) only.
The estimates of C02 emissions from iron and steel production are based on material production and consumption
data and average C contents. There is uncertainty associated with the assumption that pellet production, direct
reduced iron and sinter consumption are equal to production. There is uncertainty with the representativeness of
the associated IPCC default emission factors. There is uncertainty associated with the assumption that all coal used
for purposes other than coking coal is for direct injection coal. There is also uncertainty associated with the C
contents for pellets, sinter, and natural ore, which are assumed to equal the C contents of direct reduced iron,
when consumed in the blast furnace. There is uncertainty associated with the consumption of natural ore under
current industry practices. For EAF steel production, there is uncertainty associated with the amount of EAF anode
and charge carbon consumed due to inconsistent data throughout the time series. Also for EAF steel production,
there is uncertainty associated with the assumption that 100 percent of the natural gas attributed to "steelmaking
furnaces" by AISI is process-related and nothing is combusted for energy purposes. Uncertainty is also associated
with the use of process gases such as blast furnace gas and coke oven gas. Data are not available to differentiate
between the use of these gases for processes at the steel mill versus for energy generation (i.e., electricity and
steam generation); therefore, all consumption is attributed to iron and steel production. These data and carbon
contents produce a relatively accurate estimate of C02 emissions. However, there are uncertainties associated
with each.
For calculating the emissions estimates from iron and steel and metallurgical coke production, EPA utilizes a
number of data points taken from the AISI Annual Statistical Report (ASR). This report serves as a benchmark for
information on steel companies in United States, regardless if they are a member of AISI, which represents
integrated producers (i.e., blast furnace and EAF). During the compilation of the 1990 through 2016 Inventory
report EPA initiated conversation with AISI to better understand and update the qualitative and quantitative
uncertainty metrics associated with AISI data elements. AISI estimates their data collection response rate to range
from 75 to 90 percent, with certain sectors of the iron and steel industry not being covered by the ASR. Therefore,
there is some inherent uncertainty in the values provided in the AISI ASR, including material production and
consumption data. There is also some uncertainty to which materials produced are exported to Canada. As
indicated in the introduction to this section, the trend for integrated facilities has moved to more use of EAFs and
fewer BOFs. This trend may not be completely captured in the current data which also increases uncertainty. EPA
currently uses an uncertainty range of ±10 percent for the primary data inputs to calculate overall uncertainty
from iron and steel production, consistent with 2006 IPCC Guidelines. During EPA's discussion with AISI, AISI noted
that an uncertainty range of ±5 percent would be a more appropriate approximation to reflect their coverage of
4-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
integrated steel producers in the United States. EPA will continue to assess the best range of uncertainty for these
values.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-74 for metallurgical coke
production and iron and steel production. Total C02 emissions from metallurgical coke production and iron and
steel production for 2019 were estimated to be between 33.3 and 49.1 MMT C02 Eq. at the 95 percent confidence
level. This indicates a range of approximately 19 percent below and 19 percent above the emission estimate of
41.3 MMT C02 Eq. Total CH4 emissions from metallurgical coke production and iron and steel production for 2019
were estimated to be between 0.006 and 0.009 MMT C02 Eq. at the 95 percent confidence level. This indicates a
range of approximately 19 percent below and 19 percent above the emission estimate of 0.008 MMT C02 Eq.
Table 4-74: Approach 2 Quantitative Uncertainty Estimates for CO2 and ChU Emissions from
Iron and Steel Production and Metallurgical Coke Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTCOz Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Metallurgical Coke & Iron
and Steel Production
C02
41.3
33.3
49.1
-19%
+19%
Metallurgical Coke & Iron
ch4



-19%
+19%
and Steel Production



+ Does not exceed 0.05 MMT C02 Eq.
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019.
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter.
Recalculations Discussion
No recalculations or changes to calculation methodology were made to the iron and steel production and
metallurgical coke production inventory in 2019.
Planned Improvements
Future improvements involve improving activity data and emission factor sources for C02 and CH4 emissions
estimations from pellet production. EPA will also evaluate and analyze data reported under EPA's GHGRP to
improve the emission estimates for this and other Iron and Steel Production process categories. Particular
attention will be made to ensure time-series consistency of the emissions estimates presented in future Inventory
reports, consistent with IPCC and UNFCCC guidelines. This is required as the facility-level reporting data from EPA's
GHGRP, with the program's initial requirements for reporting of emissions in calendar year 2010, are not available
for all inventory years (i.e., 1990 through 2009) as required for this Inventory. In implementing improvements and
integration of data from EPA's GHGRP, the latest guidance from the IPCC on the use of facility-level data in national
Industrial Processes and Product Use 4-87

-------
inventories will be relied upon.62 This is a medium-term improvement and EPA estimates that earliest this
improvement could be incorporated is the 2022 Inventory submission.
Additional improvements include accounting for emission estimates for the production of metallurgical coke in the
Energy chapter as well as identifying the amount of carbonaceous materials, other than coking coal, consumed at
merchant coke plants. Other potential improvements include identifying the amount of coal used for direct
injection and the amount of coke breeze, coal tar, and light oil produced during coke production. Efforts will also
be made to identify information to better characterize emissions from the use of process gases and fuels within
the Energy and IPPU chapters. Additional efforts will be made to improve the reporting between the IPPU and
Energy chapters, particularly the inclusion of a quantitative summary of the carbon balance in the United States.
This planned improvement is a long-term improvement and is still in development; therefore, it is not included in
this current Inventory report and is not expected until a future (i.e., 2022) Inventory submission.
EPA also received comments during the Expert Review cycle of a previous (i.e., 1990 through 2016) Inventory on
recommendations to improve the description of the iron and steel industry and emissive processes. EPA began
incorporating some of these recommendations into a previous Inventory (i.e., 1990 through 2016) and will require
some additional time to implement other substantive changes.
4.18 Ferroalloy Production (CRF Source
Category 2C2)
Carbon dioxide (C02) and methane (CH4) are emitted from the production of several ferroalloys. Ferroalloys are
composites of iron (Fe) and other elements such as silicon (Si), manganese (Mn), and chromium (Cr). Emissions
from fuels consumed for energy purposes during the production of ferroalloys are accounted for in the Energy
chapter. Emissions from the production of two types of ferrosilicon (25 to 55 percent and 56 to 95 percent silicon),
silicon metal (96 to 99 percent silicon), and miscellaneous alloys (32 to 65 percent silicon) have been calculated.
Emissions from the production of ferrochromium and ferromanganese are not included here because of the small
number of manufacturers of these materials in the United States, and therefore, government information
disclosure rules prevent the publication of production data for these production facilities.
Similar to emissions from the production of iron and steel, C02 is emitted when metallurgical coke is oxidized
during a high-temperature reaction with iron and the selected alloying element. Due to the strong reducing
environment, CO is initially produced and eventually oxidized to C02. A representative reaction equation for the
production of 50 percent ferrosilicon (FeSi) is given below:
Fe203 + 2Si02 + 7C —> 2FeSi + 7C0
While most of the carbon contained in the process materials is released to the atmosphere as C02, a percentage is
also released as CH4 and other volatiles. The amount of CH4 that is released is dependent on furnace efficiency,
operation technique, and control technology.
When incorporated in alloy steels, ferroalloys are used to alter the material properties of the steel. Ferroalloys are
used primarily by the iron and steel industry, and production trends closely follow that of the iron and steel
industry. As of 2016,11 facilities in the United States produce ferroalloys (USGS 2020).
Emissions of C02 from ferroalloy production in 2019 were 1.6 MMT C02 Eq. (1,598 kt C02) (see Table 4-75 and
Table 4-76), which is a 26 percent reduction since 1990. Emissions of CH4 from ferroalloy production in 2019 were
0.01 MMT C02 Eq. (0.4 kt CH4), which is a 34 percent decrease since 1990. These decreases in emissions were
mostly caused by the shutdown of two ferroalloy facilities during 2018.
62 See .
4-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-75: CO2 and ChU Emissions from Ferroalloy Production (MMT CO2 Eq.)
Gas
1990
2005
2015
2016
2017
2018
2019
C02
2.2
1.4
2.0
1.8
2.0
2.1
1.6
ch4
+
+
+
+
+
+
+
Total
2.2
1.4
2.0
1.8
2.0
2.1
1.6
+ Does not exceed 0.05 MMT C02 Eq.
Table 4-76: CO2 and CH4 Emissions from Ferroalloy Production (kt)
Gas
1990
2005
2015
2016
2017
2018
2019
C02
2,152
1,392
1,960
1,796
1,975
2,063
1,598
CH4
1
+
1
1
1
1
+
+ Does not exceed 0.5 kt
Methodology
Emissions of C02 and CH4 from ferroalloy production were calculated63 using a Tier 1 method from the 2006IPCC
Guidelines by multiplying annual ferroalloy production by material-specific default emission factors provided by
IPCC (IPCC 2006). The Tier 1 equations for C02 and CH4 emissions are as follows:
Eco2 = Y^MPi X EFi)
i
where,
ECo2 =	C02 emissions, metric tons
MP, =	Production of ferroalloy type/', metric tons
EF, =	Generic emission factor for ferroalloy type /', metric tons C02/metric ton specific
ferroalloy product
ECHi =	X EF^
i
where,
ECh4 =	CH4 emissions, kg
MP, =	Production of ferroalloy type/', metric tons
EF, =	Generic emission factor for ferroalloy type /', kg CH4/metric ton specific ferroalloy
product
Default emission factors were used because country-specific emission factors are not currently available. The
following emission factors were used to develop annual C02 and CH4 estimates:
•	Ferrosilicon, 25 to 55 percent Si and Miscellaneous Alloys, 32 to 65 percent Si: 2.5 metric tons CO^metric
ton of alloy produced, 1.0 kg CH4/metric ton of alloy produced.
•	Ferrosilicon, 56 to 95 percent Si: 4.0 metric tons C02/metric ton alloy produced, 1.0 kg CH4/metric ton of
alloy produced.
•	Silicon Metal: 5.0 metric tons C02/metric ton metal produced, 1.2 kg CH4/metric ton metal produced.
63 EPA has not integrated aggregated facility-level GHGRP information to inform these estimates. The aggregated information
(e.g., activity data and emissions) associated with production of ferroalloys did not meet criteria to shield underlying
confidential business information (CBI) from public disclosure.
Industrial Processes and Product Use 4-89

-------
It was assumed that 100 percent of the ferroalloy production was produced using petroleum coke in an electric arc
furnace process (IPCC 2006), although some ferroalloys may have been produced with coking coal, wood, other
biomass, or graphite carbon inputs. The amount of petroleum coke consumed in ferroalloy production was
calculated assuming that the petroleum coke used is 90 percent carbon (C) and 10 percent inert material (Onder
and Bagdoyan 1993).
The use of petroleum coke for ferroalloy production is adjusted for within the Energy chapter as this fuel was
consumed during non-energy related activities. Additional information on the adjustments made within the Energy
sector for Non-Energy Use of Fuels is described in both the Methodology section of C02 from Fossil Fuel
Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating
Emissions of C02 from Fossil Fuel Combustion.
Ferroalloy production data for 1990 through 2019 (see Table 4-77) were obtained from the U.S. Geological Survey
(USGS) through the Minerals Yearbook: Silicon (USGS 1996 through 2015) and the Mineral Industry Surveys: Silicon
(USGS 2014, 2015b, 2016b, 2017, 2018b, 2019, 2020). The following data were available from the USGS
publications for the time series:
•	Ferrosilicon, 25 to 55 percent Si: Annual production data were available from 1990 through 2010.
•	Ferrosilicon, 56 to 95 percent Si: Annual production data were available from 1990 through 2010.
•	Silicon Metal: Annual production data were available from 1990 through 2005. The production data for
2005 were used as proxy for 2006 through 2010.
•	Miscellaneous Alloys, 32 to 65 percent Si: Annual production data were available from 1990 through
1998. Starting 1999, USGS reported miscellaneous alloys and ferrosilicon containing 25 to 55 percent
silicon as a single category.
Starting with the 2011 publication, USGS ceased publication of production quantity by ferroalloy product and
began reporting all the ferroalloy production data as a single category (i.e., Total Silicon Materials Production). This
is due to the small number of ferroalloy manufacturers in the United States and government information
disclosure rules. Ferroalloy product shares developed from the 2010 production data (i.e., ferroalloy product
production/total ferroalloy production) were used with the total silicon materials production quantity to estimate
the production quantity by ferroalloy product type for 2011 through 2019 (USGS 2013, 2014, 2015b, 2016b, 2017,
2018b, 2019, 2020).
Table 4-77: Production of Ferroalloys (Metric Tons)
Year Ferrosilicon Ferrosilicon Silicon Metal Misc. Alloys
	25%-55%	56%-95%	32-65%
1990	321,385	109,566	145,744	72,442
2005	123,000	86,100	148,000	NA
2015	180,372	159,151	174,477	NA
2016	165,282	145,837	159,881	NA
2017	181,775	160,390	175,835	NA
2018	189,846	167,511	183,642	NA
2019	147,034	129,736	142,229	NA
NA - Not Available for product type, aggregated along with ferrosilicon (25-55% Si)
Uncertainty and Time-Series Consistency
Annual ferroalloy production was reported by the USGS in three broad categories until the 2010 publication:
ferroalloys containing 25 to 55 percent silicon (including miscellaneous alloys), ferroalloys containing 56 to 95
percent silicon, and silicon metal (through 2005 only, 2005 value used as proxy for 2005 through 2010). Starting
with the 2011 Minerals Yearbook, USGS started reporting all the ferroalloy production under a single category:
total silicon materials production. The total silicon materials quantity was allocated across the three categories,
4-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
based on the 2010 production shares for the three categories. Refer to the Methodology section for further
details. Additionally, production data for silvery pig iron (alloys containing less than 25 percent silicon) are not
reported by the USGS to avoid disclosing proprietary company data. Emissions from this production category,
therefore, were not estimated.
Some ferroalloys may be produced using wood or other biomass as a primary or secondary carbon source
(carbonaceous reductants); however, information and data regarding these practices were not available. Emissions
from ferroalloys produced with wood or other biomass would not be counted under this source because wood-
based carbon is of biogenic origin.64 Even though emissions from ferroalloys produced with coking coal or graphite
inputs would be counted in national trends, they may be generated with varying amounts of C02 per unit of
ferroalloy produced. The most accurate method for these estimates would be to base calculations on the amount
of reducing agent used in the process, rather than the amount of ferroalloys produced. These data, however, were
not available, and are also often considered confidential business information.
Emissions of CH4 from ferroalloy production will vary depending on furnace specifics, such as type, operation
technique, and control technology. Higher heating temperatures and techniques such as sprinkle charging would
reduce CH4 emissions; however, specific furnace information was not available or included in the CH4 emission
estimates.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-78. Ferroalloy
production C02 emissions from 2019 were estimated to be between 1.4 and 1.8 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 12 percent below and 12 percent above the emission
estimate of 2.1 MMT C02 Eq. Ferroalloy production CH4 emissions were estimated to be between a range of
approximately 12 percent below and 12 percent above the emission estimate of 0.01 MMT C02 Eq.
Table 4-78: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Ferroalloy Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Ferroalloy Production
C02
1.6
1.4
1.8
-12%
+12%
Ferroalloy Production
ch4
+
+
+
-12%
+12%
+ Does not exceed 0.05 MMT C02 Eq.
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter and Annex 8.
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
64 Emissions and sinks of biogenic carbon are accounted for in the Land Use, Land-Use Change, and Forestry chapter.
Industrial Processes and Product Use 4-91

-------
Planned Improvements
Pending available resources and prioritization of improvements for more significant sources, EPA will continue to
evaluate and analyze data reported under EPA's GHGRP that would be useful to improve the emission estimates
and category-specific QC procedures for the Ferroalloy Production source category. Given the small number of
facilities and reporting thresholds, particular attention will be made to ensure completeness and time-series
consistency of the emissions estimates presented in future Inventory reports, consistent with IPCC and UNFCCC
guidelines. This is required as the facility-level reporting data from EPA's GHGRP, with the program's initial
requirements for reporting of emissions in calendar year 2010, are not available for all inventory years (i.e., 1990
through 2009) as required for this Inventory. In implementing improvements and integration of data from EPA's
GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories will be relied
upon.65 This is a long-term planned improvement and EPA is still assessing the possibility of incorporating this
improvement into the Inventory. This improvement has not been included in the current Inventory report.
4.19 Aluminum Production (CRF Source
Category 2C3)
Aluminum is a light-weight, malleable, and corrosion-resistant metal that is used in many manufactured products,
including aircraft, automobiles, bicycles, and kitchen utensils. As of recent reporting, the United States was the
twelfth largest producer of primary aluminum, with approximately 1 percent of the world total production (USGS
2019). The United States was also a major importer of primary aluminum. The production of primary aluminum—
in addition to consuming large quantities of electricity—results in process-related emissions of carbon dioxide
(C02) and two perfluorocarbons (PFCs): perfluoromethane (CF4) and perfluoroethane (C2F6).
Carbon dioxide is emitted during the aluminum smelting process when alumina (aluminum oxide, Al203) is reduced
to aluminum using the Hall-Heroult reduction process. The reduction of the alumina occurs through electrolysis in
a molten bath of natural or synthetic cryolite (Na3AIF6). The reduction cells contain a carbon (C) lining that serves
as the cathode. Carbon is also contained in the anode, which can be a C mass of paste, coke briquettes, or
prebaked C blocks from petroleum coke. During reduction, most of this C is oxidized and released to the
atmosphere as C02.
Process emissions of C02from aluminum production were estimated to be 1.9 MMT C02 Eq. (1,880 kt) in 2019 (see
Table 4-79). The C anodes consumed during aluminum production consist of petroleum coke and, to a minor
extent, coal tar pitch. The petroleum coke portion of the total C02 process emissions from aluminum production is
considered to be a non-energy use of petroleum coke, and is accounted for here and not under the C02 from Fossil
Fuel Combustion source category of the Energy sector. Similarly, the coal tar pitch portion of these C02 process
emissions is accounted for here.
Table 4-79: CO2 Emissions from Aluminum Production (MMT CO2 Eq. and kt)
Year MMT CP2 Eq. kt
1990	6.8	6,831
2005
4.1
4,142
2015	2.8	2,767
2016	1.3	1,334
2017	1.2	1,205
65 See .
4-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2018	1.5	1,451
2019	1.9	1,880
In addition to C02 emissions, the aluminum production industry is also a source of PFC emissions. During the
smelting process, when the alumina ore content of the electrolytic bath falls below critical levels required for
electrolysis, rapid voltage increases occur, which are termed High Voltage Anode Effects (HVAEs) HVAEs cause C
from the anode and fluorine from the dissociated molten cryolite bath to combine, thereby producing fugitive
emissions of CF4 and C2F6. In general, the magnitude of emissions for a given smelter and level of production
depends on the frequency and duration of these anode effects. As the frequency and duration of the anode effects
increase, emissions increase. Another type of anode effect, Low Voltage Anode Effects (LVAEs), became a concern
in the early 2010s as the aluminum industry increasingly began to use cell technologies with higher amperage and
additional anodes (IPCC 2019). LVAEs emit CF4, and are included in PFC emission totals from 2006 forward.
Since 1990, emissions of CF4 and C2F6 have declined by 92 percent and 90 percent, respectively, to 1.4 MMT C02
Eq. of CF4 (0.19 kt) and 0.4 MMT C02 Eq. of C2F6 (0.03 kt) in 2019 as shown in
Table 4-80 and Table 4-81. This decline is due both to reductions in domestic aluminum production and to actions
taken by aluminum smelting companies to reduce the frequency and duration of anode effects. These actions
include technology and operational changes such as employee training, use of computer monitoring, and changes
in alumina feeding techniques. Since 1990, aluminum production has declined by 72 percent, while the combined
CF4 and C2F6 emission rate (per metric ton of aluminum produced) has been reduced by 70 percent. PFC emissions
increased by approximately 8 percent between 2018 and 2019 due to increases in aluminum production. The
decrease in the ratio of C2F6toCF4 emissions between 2018 and 2019 may be due to a combination of a decrease
in the measured C2F6 to CF4 weight fraction at some facilities and changes in how production is distributed among
facilities with different C2F6 to CF4 weight fractions.
Table 4-80: PFC Emissions from Aluminum Production (MMT CO2 Eq.)
Year CF4 C2F6 Total
1990 17.9 3.5 21.5
2005 2.9 0.6	3.4
2015
1.6
0.5
2.1
2016
1.0
0.4
1.4
2017
0.7
0.4
1.1
2018
1.2
0.4
1.6
2019
1.4
0.4
1.8
Note: Totals may not sum due to
independent rounding.
Industrial Processes and Product Use 4-93

-------
Table 4-81: PFC Emissions from Aluminum Production (kt)
Year CF4 C2F6
1990 2.4	0.3
2005 0.4	+
2015	0.2	+
2016	0.1	+
2017	0.1	+
2018	0.2	+
2019	0.2	+_
+ Does not exceed 0.05 kt.
In 2019, U.S. primary aluminum production totaled approximately 1.1 million metric tons, a 26 percent increase
from 2018 production levels (USAA 2020). In 2019, three companies managed production at seven operational
primary aluminum smelters. Two smelters operated at full capacity during 2019, while the other five operated at
reduced capacity (USGS 2020). During 2019, monthly U.S. primary aluminum production was higher for every
month when compared to the corresponding months in 2018, with the exception of December (USAA 2020).
For 2020, total production for the January to August period was approximately 0.7 million metric tons compared to
0.8 million metric tons for the same period in 2018, a 4.5 percent decrease (USAA 2020). Based on the decrease in
production, process C02 and PFC emissions may be lower in 2020 compared to 2019 if there are no significant
changes in process controls at operational facilities.
Methodology
Process C02 and PFC (i.e., CF4 and C2F6) emission estimates from primary aluminum production for 2010 through
2019 are available from EPA's GHGRP—Subpart F (Aluminum Production) (EPA 2020). Under EPA's GHGRP,
facilities began reporting primary aluminum production process emissions (for 2010) in 2011; as a result, GHGRP
data (for 2010 through 2019) are available to be incorporated into the Inventory. EPA's GHGRP mandates that all
facilities that contain an aluminum production process must report: CF4 and C2F6 emissions from anode effects in
all prebake and S0derberg electrolysis cells, C02 emissions from anode consumption during electrolysis in all
prebake and S0derberg cells, and all C02 emissions from onsite anode baking. To estimate the process emissions,
EPA's GHGRP uses the process-specific equations detailed in Subpart F (aluminum production).66 These equations
are based on the Tier 2/Tier 3 IPCC (2006) methods for primary aluminum production, and Tier 1 methods when
estimating missing data elements. It should be noted that the same methods (i.e., 2006 IPCC Guidelines) were used
for estimating the emissions prior to the availability of the reported GHGRP data in the Inventory. Prior to 2010,
aluminum production data were provided through EPA's Voluntary Aluminum Industrial Partnership (VAIP).
As previously noted, the use of petroleum coke for aluminum production is adjusted for within the Energy chapter
as this fuel was consumed during non-energy related activities. Additional information on the adjustments made
within the Energy sector for Non-Energy Use of Fuels is described in both the Methodology section of C02 from
Fossil Fuel Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for
Estimating Emissions of C02 from Fossil Fuel Combustion.
66 Code of Federal Regulations, Title 40: Protection of Environment, Part 98: Mandatory Greenhouse Gas Reporting, Subpart
F—Aluminum Production. See .
4-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Process C02 Emissions from Anode Consumption and Anode Baking
Carbon dioxide emission estimates for the years prior to the introduction of EPA's GHGRP in 2010 were estimated
using 2006IPCC Guidelines methods, but individual facility reported data were combined with process-specific
emissions modeling. These estimates were based on information previously gathered from EPA's Voluntary
Aluminum Industrial Partnership (VAIP) program, U.S. Geological Survey (USGS) Mineral Commodity reviews, and
The Aluminum Association (USAA) statistics, among other sources. Since pre- and post-GHGRP estimates use the
same methodology, emission estimates are comparable across the time series.
Most of the C02 emissions released during aluminum production occur during the electrolysis reaction of the C
anode, as described by the following reaction:
2AI2O3 + 3C -> 4A1 + 3C02
For prebake smelter technologies, C02 is also emitted during the anode baking process. These emissions can
account for approximately 10 percent of total process C02 emissions from prebake smelters.
Depending on the availability of smelter-specific data, the C02 emitted from electrolysis at each smelter was
estimated from: (1) the smelter's annual anode consumption, (2) the smelter's annual aluminum production and
rate of anode consumption (per ton of aluminum produced) for previous and/or following years, or (3) the
smelter's annual aluminum production and IPCC default C02 emission factors. The first approach tracks the
consumption and carbon content of the anode, assuming that all C in the anode is converted to C02. Sulfur, ash,
and other impurities in the anode are subtracted from the anode consumption to arrive at a C consumption figure.
This approach corresponds to either the IPCC Tier 2 or Tier 3 method, depending on whether smelter-specific data
on anode impurities are used. The second approach interpolates smelter-specific anode consumption rates to
estimate emissions during years for which anode consumption data are not available. This approach avoids
substantial errors and discontinuities that could be introduced by reverting to Tier 1 methods for those years. The
last approach corresponds to the IPCC Tier 1 method (IPCC 2006) and is used in the absence of present or historic
anode consumption data.
The equations used to estimate C02 emissions in the Tier 2 and 3 methods vary depending on smelter type (IPCC
2006). For Prebake cells, the process formula accounts for various parameters, including net anode consumption,
and the sulfur, ash, and impurity content of the baked anode. For anode baking emissions, the formula accounts
for packing coke consumption, the sulfur and ash content of the packing coke, as well as the pitch content and
weight of baked anodes produced. For S0derberg cells, the process formula accounts for the weight of paste
consumed per metric ton of aluminum produced, and pitch properties, including sulfur, hydrogen, and ash
content.
Through the VAIP, anode consumption (and some anode impurity) data have been reported for 1990, 2000, 2003,
2004, 2005, 2006, 2007, 2008, and 2009. Where available, smelter-specific process data reported under the VAIP
were used; however, if the data were incomplete or unavailable, information was supplemented using industry
average values recommended by IPCC (2006). Smelter-specific C02 process data were provided by 18 of the 23
operating smelters in 1990 and 2000, by 14 out of 16 operating smelters in 2003 and 2004,14 out of 15 operating
smelters in 2005,13 out of 14 operating smelters in 2006, 5 out of 14 operating smelters in 2007 and 2008, and 3
out of 13 operating smelters in 2009. For years where C02 emissions data or C02 process data were not reported
by these companies, estimates were developed through linear interpolation, and/or assuming representative (e.g.,
previously reported or industry default) values.
In the absence of any previous historical smelter-specific process data (i.e., 1 out of 13 smelters in 2009; 1 out of
14 smelters in 2006, 2007, and 2008; 1 out of 15 smelters in 2005; and 5 out of 23 smelters between 1990 and
2003), C02 emission estimates were estimated using Tier 1 S0derberg and/or Prebake emission factors (metric ton
of C02 per metric ton of aluminum produced) from IPCC (2006).
Industrial Processes and Product Use 4-95

-------
Process PFC Emissions from Anode Effects
High Voltage Anode Effects
Smelter-specific PFC emissions from aluminum production for 2010 through 2019 were reported to EPA under its
GHGRP. To estimate their PFC emissions from HVAEs and report them under EPA's GHGRP, smelters use an
approach identical to the Tier 3 approach in the 2006IPCC Guidelines (IPCC 2006). Specifically, they use a smelter-
specific slope coefficient as well as smelter-specific operating data to estimate an emission factor using the
following equation:
PFC = S xAE
AE = F xD
where,
PFC =
CF4 or C2F6, kg/MT aluminum
S
Slope coefficient, PFC/AE
AE
Anode effect, minutes/cell-day
F
Anode effect frequency per cell-day
D
Anode effect duration, minutes
They then multiply this emission factor by aluminum production to estimate PFC emissions from HVAEs. All U.S.
aluminum smelters are required to report their emissions under EPA's GHGRP.
Perfluorocarbon emissions for the years prior to 2010 were estimated using the same equation, but the slope-
factor used for some smelters was technology-specific rather than smelter-specific, making the method a Tier 2
rather than a Tier 3 approach for those smelters. Emissions and background data were reported to EPA under the
VAIP. For 1990 through 2009, smelter-specific slope coefficients were available and were used for smelters
representing between 30 and 94 percent of U.S. primary aluminum production. The percentage changed from year
to year as some smelters closed or changed hands and as the production at remaining smelters fluctuated. For
smelters that did not report smelter-specific slope coefficients, IPCC technology-specific slope coefficients were
applied (IPCC 2006). The slope coefficients were combined with smelter-specific anode effect data collected by
aluminum companies and reported under the VAIP to estimate emission factors over time. For 1990 through 2009,
smelter-specific anode effect data were available for smelters representing between 80 and 100 percent of U.S.
primary aluminum production. Where smelter-specific anode effect data were not available, representative values
(e.g., previously reported or industry averages) were used.
For all smelters, emission factors were multiplied by annual production to estimate annual emissions at the
smelter level. For 1990 through 2009, smelter-specific production data were available for smelters representing
between 30 and 100 percent of U.S. primary aluminum production. (For the years after 2000, this percentage was
near the high end of the range.) Production at non-reporting smelters was estimated by calculating the difference
between the production reported under VAIP and the total U.S. production supplied by USGS or USAA, and then
allocating this difference to non-reporting smelters in proportion to their production capacity. Emissions were then
aggregated across smelters to estimate national emissions.
Low Voltage Anode Effects
LVAE emissions of CF4 were estimated for 2006 through 2019 based on the Tier 1 (technology-specific, production-
based) method in the 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC
4-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2019). Prior to 2006, LVAE emissions are believed to have been negligible.67 The following equation was used to
estimate LVAE PFC emissions:
LVAE ECF4 = LVAEEFCF4 X MP
where,
LVAE ECf4 =	LVAE emissions of CF4 from aluminium production, kg CF4
LVAE EFCf4 =	LVAE emission factor for CF4 (default by cell technology type)
MP	=	metal production by cell technology type, tons Al.
LVAE emissions estimates were then combined with HVAE emissions estimates to calculate total PFC emissions
from aluminum production.
Production Data
Between 1990 and 2009, production data were provided under the VAIP by 21 of the 23 U.S. smelters that
operated during at least part of that period. For the non-reporting smelters, production was estimated based on
the difference between reporting smelters and national aluminum production levels (USGS and USAA 1990
through 2009), with allocation to specific smelters based on reported production capacities (USGS 1990 through
2009).
National primary aluminum production data for 2019 were obtained via USAA (USAA 2020). For 1990 through
2001, and 2006 (see Table 4-82) data were obtained from USGS Mineral Industry Surveys: Aluminum Annual Report
(USGS 1995,1998, 2000, 2001, 2002, 2007). For 2002 through 2005, and 2007 through 2019, national aluminum
production data were obtained from the USAA's Primary Aluminum Statistics (USAA 2004 through 2006, 2008
through 2020).
Table 4-82: Production of Primary Aluminum (kt)
Year
kt
1990
4,048
2005
2,478
2015
1,587
2016
818
2017
741
2018
897
2019
1,126
Uncertainty and Time-Series Consistency
Uncertainty was estimated for the C02, CF4, and C2F6 emission values reported by each individual facility to EPA's
GHGRP, taking into consideration the uncertainties associated with aluminum production, anode effect minutes,
and slope factors. The uncertainty bounds used for these parameters were established based on information
collected under the VAIP and held constant through 2019. Uncertainty surrounding the reported C02, CF4, and C2F6
67 The 2019 Refinement states, "Since 2006, the global aluminium industry has undergone changes in technology and operating
conditions that make LVAE emissions much more prevalent12; these changes have occurred not only through uptake of newer
technologies (e.g., PFPBL to PFPBM) but also during upgrades within the same technology in order to maximise productivity and
reduce energy use" (IPCC 2019). Footnote #12 uses the example of PFPBL, which is prevalent in the U.S., as an older technology
that has been upgraded.
Industrial Processes and Product Use 4-97

-------
emission values were determined to have a normal distribution with uncertainty ranges of ±6, ±16, and ±20
percent, respectively.
For the LVAE emission values not reported through EPA's GHGRP but estimated instead through a Tier 1
methodology, the analysis examined uncertainty associated with primary capacity data as well as technology-
specific emission factors. Uncertainty for each facility's primary capacity, reported in the USGS Yearbook, was
estimated to have a Pert Beta distribution with an uncertainty range of-10/+7 percent based on the uncertainty of
reported capacity data, the number of years since the facility reported new capacity data, and uncertainty in
capacity utilization. Uncertainty was applied to LVAE emission factors according to technology using the
uncertainty ranges provided in the 2019 Refinement to the 2006IPCC Guidelines. An uncertainty range for
Horizontal Stud S0derberg (HSS) technology was not provided in the 2019 Refinement to the 2006 IPCC Guidelines
due to insufficient data, so a normal distribution and uncertainty range of ±99 percent was applied for that
technology based on expert judgment. A Monte Carlo analysis was applied to estimate the overall uncertainty of
the C02, CF4, and C2F6 emission estimates for the U.S. aluminum industry as a whole, and the results are provided
below.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-83. Aluminum
production-related C02 emissions were estimated to be between 1.84 and 1.92 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 2 percent below to 2 percent above the emission
estimate of 1.88 MMT C02 Eq. Also, production-related CF4 emissions were estimated to be between 1.22 and 1.43
MMT C02 Eq. at the 95 percent confidence level. This indicates a range of approximately 8 percent below to 8
percent above the emission estimate of 1.33 MMT C02 Eq. Finally, aluminum production-related C2F6 emissions
were estimated to be between 0.31 and 0.41 MMT C02 Eq. at the 95 percent confidence level. This indicates a
range of approximately 13 percent below to 14 percent above the emission estimate of 0.36 MMT C02 Eq.
Table 4-83: Approach 2 Quantitative Uncertainty Estimates for CO2 and PFC Emissions from
Aluminum Production (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Aluminum Production
C02
1.8
1.8
1.9
-2%
2%
Aluminum Production
cf4
1.4
1.3
1.5
-8%
8%
Aluminum Production
c2f6
0.4
0.3
0.4
-14%
14%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time-series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006 IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details). For the GHGRP data, EPA verifies annual facility-
level reports through a multi-step process (e.g., including a combination of pre-and post-submittal electronic
checks and manual reviews by staff) to identify potential errors and ensure that data submitted to EPA are
accurate, complete, and consistent (EPA 20 15).68 Based on the results of the verification process, EPA follows up
68 GHGRP Report Verification Factsheet. .
4-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
with facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with a
number of general and category-specific QC procedures, including: range checks, statistical checks, algorithm
checks, and year-to-year checks of reported data and emissions.
Recalculations Discussion
In a few instances, GHGRP-reporting facilities revised their GHGRP reports due to previously identified reporting
errors in 2015 and 2018, resulting in a slight decrease in total emissions of PFCs.
Based on the 2019 Refinement, Low Voltage Anode Effect (LVAE) emissions were estimated for the years 2006
to2019 using the Tier 1 production-based methodology (IPCC 2019). The Tier 1 LVAE method uses technology-
based default emissions factors, multiplied by metal production. Metal production was estimated based on
primary aluminum production capacity. The addition of LVAE emissions estimates resulted in slightly higher CF4
estimates from 2006 through 2019, with LVAE emissions accounting for 5 percent of total CF4 emissions in 2019.
4.20 Magnesium Production and Processing
(CRF Source Category 2C4)
The magnesium metal production and casting industry uses sulfur hexafluoride (SF6) as a cover gas to prevent the
rapid oxidation of molten magnesium in the presence of air. Sulfur hexafluoride has been used in this application
around the world for more than thirty years. A dilute gaseous mixture of SF6 with dry air and/or carbon dioxide
(C02) is blown over molten magnesium metal to induce and stabilize the formation of a protective crust. A small
portion of the SF6 reacts with the magnesium to form a thin molecular film of mostly magnesium oxide and
magnesium fluoride. The amount of SF6 reacting in magnesium production and processing is considered to be
negligible and thus all SF6 used is assumed to be emitted into the atmosphere. Alternative cover gases, such as
AM-cover™ (containing HFC-134a), Novec™ 612 (FK-5-1-12) and dilute sulfur dioxide (S02) systems can and are
being used by some facilities in the United States. However, many facilities in the United States are still using
traditional SF6 cover gas systems.
The magnesium industry emitted 0.9 MMT C02 Eq. (0.04 kt) of SF6, 0.1 MMT C02 Eq. (0.05 kt) of HFC-134a, and
0.001 MMT C02 Eq. (1.4 kt) of C02 in 2019. This represents a decrease of approximately 12 percent from total 2018
emissions (see Table 4-84) and a decrease in SF6 emissions by 11 percent. In 2019, total HFC-134a emissions
decreased from 0.079 MMT C02 Eq. to 0.066 MMT C02 Eq., or a 16 percent decrease as compared to 2018
emissions. FK 5-1-12 emissions were held constant from 2018. The emissions of the carrier gas, C02, decreased
from 1.43 kt in 2018 to 1.35 kt in 2019, or 6 percent.
Table 4-84: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and
Processing (MMT CO2 Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
sf6
5.2
2.7
1.0
1.1
1.0
1.0
0.9
HFC-134a
0.0
0.0
0.1
0.1
0.1
0.1
0.1
C02
+
+
+
+
+
+
+
FK 5-1-12°
0.0
0.0
+
+
+
+
+
Total
5.2
2.7
1.1
1.2
1.1
1.1
1.0
+ Does not exceed 0.05 MMT C02 Eq.
a Emissions of FK 5-1-12 are not included in totals.
Industrial Processes and Product Use 4-99

-------
Table 4-85: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and
Processing (kt)
Year
1990
2005
2015
2016
2017
2018
2019
sf6
0.2
0.1
+
+
+
+
+
HFC-134a
0.0
0.0
0.1
0.1
0.1
0.1
+
C02
1.4
2.9
2.6
2.7
3.1
1.4
1.4
FK 5-1-12°
0.0
0.0
+
+
+
+
+
+ Does not exceed 0.05 kt
a Emissions of FK 5-1-12 are not included in totals.
Methodology
Emission estimates for the magnesium industry incorporate information provided by industry participants in EPA's
SF6 Emission Reduction Partnership for the Magnesium Industry as well as emissions data reported through
Subpart T (Magnesium Production and Processing) of EPA's GHGRP. The Partnership started in 1999 and, in 2010,
participating companies represented 100 percent of U.S. primary and secondary production and 16 percent of the
casting sector production (i.e., die, sand, permanent mold, wrought, and anode casting). SF6 emissions for 1999
through 2010 from primary production, secondary production (i.e., recycling), and die casting were generally
reported by Partnership participants. Partners reported their SF6 consumption, which is assumed to be equivalent
to emissions. Along with SF6, some Partners reported their HFC-134a and FK 5-1-12 consumed, which is also
assumed to be equal to emissions. The last reporting year under the Partnership was 2010. Emissions data for
2011 through 2019 are obtained through EPA's GHGRP. Under the program, owners or operators of facilities that
have a magnesium production or casting process must report emissions from use of cover or carrier gases, which
include SF6, HFC-134a, FK 5-1-12 and C02. Consequently, cover and carrier gas emissions from magnesium
production and processing were estimated for three time periods, depending on the source of the emissions data:
1990 through 1998 (pre-EPA Partnership), 1999 through 2010 (EPA Partnership), and 2011 through 2019 (EPA
GHGRP). The methodologies described below also make use of magnesium production data published by the U.S.
Geological Survey (USGS) as available.
1990 through 1998
To estimate emissions for 1990 through 1998, industry SF6 emission factors were multiplied by the corresponding
metal production and consumption (casting) statistics from USGS. For this period, it was assumed that there was
no use of HFC-134a or FK 5-1-12 cover gases, and hence emissions were not estimated for these alternatives.
Sulfur hexafluoride emission factors from 1990 through 1998 were based on a number of sources and
assumptions. Emission factors for primary production were available from U.S. primary producers for 1994 and
1995. The primary production emission factors were 1.2 kg SF6 per metric ton for 1990 through 1993, and 1.1 kg
SF6 per metric ton for 1994 through 1997. The emission factor for secondary production from 1990 through 1998
was assumed to be constant at the 1999 average Partner value. An emission factor for die casting of 4.1 kg SF6 per
metric ton, which was available for the mid-1990s from an international survey (Gjestland and Magers 1996), was
used for years 1990 through 1996. For 1996 through 1998, the emission factor for die casting was assumed to
decline linearly to the level estimated based on Partner reports in 1999. This assumption is consistent with the
trend in SF6 sales to the magnesium sector that was reported in the RAND survey of major SF6 manufacturers,
which showed a decline of 70 percent from 1996 to 1999 (RAND 2002). Sand casting emission factors for 1990
through 2001 were assumed to be the same as the 2002 emission factor. The emission factors for the other
processes (i.e., permanent mold, wrought, and anode casting), about which less is known, were assumed to remain
constant at levels defined in Table 4-84. The emission factors for the other processes (i.e., permanent mold,
wrought, and anode casting) were based on discussions with industry representatives.
The quantities of C02 carrier gas used for each production type have been estimated using the 1999 estimated C02
emissions data and the annual calculated rate of change of SF6 use in the 1990 through 1999 time period. For each
year and production type, the rate of change of SF6 use between the current year and the subsequent year was
4-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
first estimated. This rate of change was then applied to the C02 emissions of the subsequent year to determine the
C02 emission of the current year. The emissions of carrier gases for permanent mold, wrought, and anode
processes are not estimated in this Inventory.
1999 through 2010
The 1999 through 2010 emissions from primary and secondary production are based on information provided by
EPA's industry Partners. In some instances, there were years of missing Partner data, including SF6 consumption
and metal processed. For these situations, emissions were estimated through interpolation where possible, or by
holding company-reported emissions (as well as production) constant from the previous year. For alternative cover
gases, including HFC-134a and FK 5-1-12, mainly reported data was relied upon. That is, unless a Partner reported
using an alternative cover gas, it was not assumed it was used. Emissions of alternate gases were also estimated
through linear interpolation where possible.
The die casting emission estimates for 1999 through 2010 were also based on information supplied by industry
Partners. When a Partner was determined to be no longer in production, its metal production and usage rates
were set to zero. Missing data on emissions or metal input was either interpolated or held constant at the last
available reported value. In 1999 through 2010, Partners were assumed to account for all die casting tracked by
USGS. For 1999, die casters who were not Partners were assumed to be similar to Partners who cast small parts.
Due to process requirements, these casters consume larger quantities of SF6 per metric ton of processed
magnesium than casters that process large parts. Consequently, emission estimates from this group of die casters
were developed using an average emission factor of 5.2 kg SF6 per metric ton of magnesium. This emission factor
was developed using magnesium production and SF6 usage data for the year 1999. In 2008, the derived emission
factor for die casting began to increase after many years of largely decreasing emission factors. As determined
through an analysis of activity data reported from the USGS, this increase is due to a temporary decrease in
production at many facilities between 2008 and 2010, which reflects the change in production that occurred
during the recession.
The emissions from other casting operations were estimated by multiplying emission factors (kg SF6 per metric ton
of metal produced or processed) by the amount of metal produced or consumed from USGS, with the exception of
some years for which Partner sand casting emissions data are available. The emission factors for sand casting
activities were acquired through the data reported by the Partnership for 2002 to 2006. For 1999-2001, the sand
casting emission factor was held constant at the 2002 Partner-reported level. For 2007 through 2010, the sand
casting Partner did not report and the reported emission factor from 2005 was applied to the Partner and to all
other sand casters. Activity data for 2005 was obtained from USGS (USGS 2005b).
The emission factors for primary production, secondary production and sand casting for the 1999 to 2010 are not
published to protect company-specific production information. However, the emission factor for primary
production has not risen above the average 1995 Partner value of 1.1 kg SF6 per metric ton. The emission factors
for the other industry sectors (i.e., permanent mold, wrought, and anode casting) were based on discussions with
industry representatives. The emission factors for casting activities are provided below in Table 4-86.
The emissions of HFC-134a and FK-5-1-12 were included in the estimates for only instances where Partners
reported that information to the Partnership. Emissions of these alternative cover gases were not estimated for
instances where emissions were not reported.
Carbon dioxide carrier gas emissions were estimated using the emission factors developed based on GHGRP-
reported carrier gas and cover gas data, by production type. It was assumed that the use of carrier gas, by
production type, is proportional to the use of cover gases. Therefore, an emission factor, in kg C02 per kg cover gas
and weighted by the cover gases used, was developed for each of the production types. GHGRP data on which
these emissions factors are based was available for primary, secondary, die casting and sand casting. The emission
factors were applied to the total quantity of all cover gases used (SF6, HFC-134a, and FK-5-1-12) by production type
in this time period. Carrier gas emissions for the 1999 through 2010 time period were only estimated for those
Partner companies that reported using C02 as a carrier gas through the GHGRP. Using this approach helped ensure
Industrial Processes and Product Use 4-101

-------
time-series consistency. The emissions of carrier gases for permanent mold, wrought, and anode processes are not
estimated in this Inventory.
Table 4-86: SF6 Emission Factors (kg SF6 per metric ton of magnesium)
Year
Die Casting3
Permanent Mold
Wrought
Anodes
1999
1.75b
2
1
1
2000
0.72
2
1
1
2001
0.72
2
1
1
2002
0.71
2
1
1
2003
0.81
2
1
1
2004
0.79
2
1
1
2005
0.77
2
1
1
2006
0.88
2
1
1
2007
0.64
2
1
1
2008
0.97
2
1
1
2009
1.41
2
1
1
2010
1.43
2
1
1
a Weighted average includes all die casters, Partners and non-Partners, hor
the majority of the time series (2000 through 2010), Partners made up 100
percent of die casters in the United States.
b Weighted average that includes an estimated emission factor of 5.2 kg
SF6 per metric ton of magnesium for die casters that do not participate in
the Partnership.
2011 through 2019
For 2011 through 2019, for the primary and secondary producers, GHGRP-reported cover and carrier gases
emissions data were used. For sand and die casting, some emissions data was obtained through EPA's GHGRP.
Additionally, in 2018 a new GHGRP reporter began reporting permanent mold emissions. The balance of the
emissions for this industry segment was estimated based on previous Partner reporting (i.e., for Partners that did
not report emissions through EPA's GHGRP) or were estimated by multiplying emission factors by the amount of
metal produced or consumed. Partners who did not report through EPA's GHGRP were assumed to have continued
to emit SF6 at the last reported level, which was from 2010 in most cases, unless publicly available sources
indicated that these facilities have closed or otherwise eliminated SF6 emissions from magnesium production (ARB
2015). Many Partners that did report through the GHGRP showed increases in SF6 emissions driven by increased
production related to a continued economic recovery after the 2008 recession. One Partner in particular reported
an anonymously large increase in SF6 emissions from 2010-2011, further driving increases in emissions between
the two time periods of inventory estimates. All Partners were assumed to have continued to consume magnesium
at the last reported level. Where the total metal consumption estimated for the Partners fell below the U.S. total
reported by USGS, the difference was multiplied by the emission factors discussed in the section above, i.e., non-
partner emission factors. For the other types of production and processing (i.e., permanent mold, wrought, and
anode casting), emissions were estimated by multiplying the industry emission factors with the metal production
or consumption statistics obtained from USGS (USGS 2020). USGS data for 2019 was not yet available at the time
of the analysis, so the 2017 values were held constant through 2019 as a proxy. Where data was submitted late or
with errors or not available for 2019 through the GHGRP, EPA held values constant at previous year's levels for
emissions.
Uncertainty and Time-Series Consistency
Uncertainty surrounding the total estimated emissions in 2019 is attributed to the uncertainties around SF6, HFC-
134a, and C02 emission estimates. To estimate the uncertainty surrounding the estimated 2019 SF6 emissions from
magnesium production and processing, the uncertainties associated with three variables were estimated: (1)
emissions reported by magnesium producers and processors for 2019 through EPA's GHGRP, (2) emissions
4-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
estimated for magnesium producers and processors that reported via the Partnership in prior years but did not
report 2019 emissions through EPA's GHGRP, and (3) emissions estimated for magnesium producers and
processors that did not participate in the Partnership or report through EPA's GHGRP. An uncertainty of 5 percent
was assigned to the emissions (usage) data reported by each GHGRP reporter for all the cover and carrier gases
(per the 2006IPCC Guidelines). If facilities did not report emissions data during the current reporting year through
EPA's GHGRP, SF6 emissions data were held constant at the most recent available value reported through the
Partnership. The uncertainty associated with these values was estimated to be 30 percent for each year of
extrapolation. The uncertainty of the total inventory estimate remained relatively constant between 2018 and
2019.
Alternate cover gas and carrier gases data was set equal to zero if the facilities did not report via the GHGRP. For
those industry processes that are not represented in the Partnership, such as permanent mold and wrought
casting, SF6 emissions were estimated using production and consumption statistics reported by USGS and
estimated process-specific emission factors (see Table 4-87). The uncertainties associated with the emission
factors and USGS-reported statistics were assumed to be 75 percent and 25 percent, respectively. Emissions
associated with die casting and sand casting activities utilized emission factors based on Partner reported data
with an uncertainty of 75 percent. In general, where precise quantitative information was not available on the
uncertainty of a parameter, a conservative (upper-bound) value was used.
Additional uncertainties exist in these estimates that are not addressed in this methodology, such as the basic
assumption that SF6 neither reacts nor decomposes during use. The melt surface reactions and high temperatures
associated with molten magnesium could potentially cause some gas degradation. Previous measurement studies
have identified SF6 cover gas degradation in die casting applications on the order of 20 percent (Bartos et al. 2007).
Sulfur hexafluoride may also be used as a cover gas for the casting of molten aluminum with high magnesium
content; however, the extent to which this technique is used in the United States is unknown.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-87. Total emissions
associated with magnesium production and processing were estimated to be between 0.90 and 1.05 MMT C02 Eq.
at the 95 percent confidence level. This indicates a range of approximately 8 percent below to 8 percent above the
2019 emission estimate of 0.97 MMT C02 Eq. The uncertainty estimates for 2019 are similar to the uncertainty
reported for 2018 in the previous Inventory.
Table 4-87: Approach 2 Quantitative Uncertainty Estimates for SFe, HFC-134a and CO2
Emissions from Magnesium Production and Processing (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTC02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Magnesium
Production
SF6, HFC-
134a, C02
0.97
0.90
1.05
SP
0s-
00
SP
0s-
0?
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above and QA/QC and Verification section below.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Volume 1, Chapter 6 of 2006 IPCC Guidelines as described in the
introduction of the IPPU chapter (see Annex 8 for more details). For the GHGRP data, EPA verifies annual facility-
level reports through a multi-step process (e.g., including a combination of pre-and post-submittal electronic
Industrial Processes and Product Use 4-103

-------
checks and manual reviews by staff) to identify potential errors and ensure that data submitted to EPA are
accurate, complete, and consistent (EPA 20 15).69 Based on the results of the verification process, EPA follows up
with facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with a
number of general and category-specific QC procedures, including: range checks, statistical checks, algorithm
checks, and year-to-year checks of reported data and emissions. Note, corrections to the total U.S. activity data for
die casting completed for the previous Inventory (see Recalculations Discussion) resolved an apparent increase in
the implied emission factor for magnesium production that was caused by recalculations to the diecasting
emissions in the 1990 to 2016 Inventory without a concurrent adjustment to the activity data. With both
corrections in the Inventory, the implied emission factor for magnesium has a roughly flat trend for 2009 to 2011
and an overall downward trend for 1999 to 2014. There remains a spike in emissions in 2011 partially due to
unusually high emissions from two facilities in 2011.
Recalculations Discussion
Die casting and sand volumes were updated based on the release of the 2017 USGS Minerals Yearbook (USGS,
2017). Additionally, one facility's emissions were revised for 2017 and 2018 to reflect its confirmed closure. These
revisions resulted in slightly higher emissions in 2017 and reduced emissions in 2018.
Planned Improvements
Cover gas research conducted over the last decade has found that SF6 used for magnesium melt protection can
have degradation rates on the order of 20 percent in die casting applications (Bartos et al. 2007). Current emission
estimates assume (per the 2006IPCC Guidelines) that all SF6 utilized is emitted to the atmosphere. Additional
research may lead to a revision of the 2006 IPCC Guidelines to reflect this phenomenon and until such time,
developments in this sector will be monitored for possible application to the Inventory methodology. Usage and
emission details of carrier gases in permanent mold, wrought, and anode processes will be researched as part of a
future Inventory. Based on this research and data from a permanent mold facility newly reporting the GHGRP, it
will be determined if C02 carrier gas emissions are to be estimated.
Additional emissions are generated as byproducts from the use of alternate cover gases, which are not currently
accounted for. Research on this topic is developing, and as reliable emission factors become available, these
emissions will be incorporated into the Inventory.
4.21 Lead Production (CRF Source Category
2C5)	
In 2019, lead was produced in the United States only using secondary production processes. Until 2014, lead
production in the United States involved both primary and secondary processes—both of which emit carbon
dioxide (C02) (Sjardin 2003). Emissions from fuels consumed for energy purposes during the production of lead are
accounted for in the Energy chapter.
Primary production of lead through the direct smelting of lead concentrate produces C02 emissions as the lead
concentrates are reduced in a furnace using metallurgical coke (Sjardin 2003). Primary lead production, in the form
of direct smelting, previously occurred at a single smelter in Missouri. This primary lead smelter was closed at the
69 GHGRP Report Verification Factsheet. .
4-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
end of 2013. In 2014, the smelter processed a small amount of residual lead during demolition of the site (USGS
2015), and in 2018, the smelter processed no lead (USGS 2016, 2019).
Similar to primary lead production, C02 emissions from secondary lead production result when a reducing agent,
usually metallurgical coke, is added to the smelter to aid in the reduction process. Carbon dioxide emissions from
secondary production also occur through the treatment of secondary raw materials (Sjardin 2003). Secondary
production primarily involves the recycling of lead acid batteries and post-consumer scrap at secondary smelters.
Secondary lead production has increased in the United States over the past decade, while primary lead production
has decreased to production levels of zero. In 2019, secondary lead production accounted for 100 percent of total
lead production. The lead-acid battery industry accounted for more than 93 percent of the reported U.S. lead
consumption in 2019 (USGS 2020).
In 2019, total secondary lead production in the United States increased from 2018. The United States has become
more reliant on imported refined lead in recent years, owing to the closure of the last primary lead smelter in
2013. Exports of spent SLI batteries have been generally decreasing since 2014, however they have increased back
up to the 2014 level for 2019 (USGS 2015 through 2020). In the first 10 months of 2019, 22.9 million spent SLI lead-
acid batteries were exported, essentially unchanged compared with that in the same time period in 2018 (USGS
2020).
As in 2018, U.S. primary lead production remained at production levels of zero for 2019. This is due to the closure
of the only domestic primary lead smelter in 2013 (year-end), as stated previously. In 2019, U.S. secondary lead
production was greater than 2018 levels and has increased by 30 percent since 1990 (USGS 1995 through 2020).
In 2019, U.S. lead production totaled 1,200,000 metric tons (USGS 2020). The resulting emissions of C02 from 2019
lead production were estimated to be 0.5 MMT C02 Eq. (540 kt) (see Table 4-88).
The United States was the fourth largest mine producer of lead in the world, behind China, Australia, and Peru, and
accounted for approximately 6 percent of world production in 2019 (USGS 2020).
Table 4-88: CO2 Emissions from Lead Production (MMT CO2 Eq. and kt)
Year
MMTCOz Eq.
Kt
1990
0.5
516
2005
0.6
553
2015
0.5
473
2016
0.5
500
2017
0.5
513
2018
0.5
513
2019
0.5
540
After a steady increase in total emissions from 1995 to 2000, total emissions have gradually decreased since 2000,
slightly increased since 2015, and are currently 5 percent higher than 1990 levels.
Methodology
The methods used to estimate emissions for lead production70 are based on Sjardin's work (Sjardin 2003) for lead
production emissions and Tier 1 methods from the 2006IPCC Guidelines. The Tier 1 equation is as follows:
70 EPA has not integrated aggregated facility-level Greenhouse Gas Reporting Program (GHGRP) information to inform these
estimates. The aggregated information (e.g., activity data and emissions) associated with Lead Production did not meet criteria
to shield underlying confidential business information (CBI) from public disclosure.
Industrial Processes and Product Use 4-105

-------
where,
C02 Emissions = (DS x EFDS) + (5 x EFS)
DS	=	Lead produced by direct smelting, metric ton
S	=	Lead produced from secondary materials
EFds	=	Emission factor for direct Smelting, metric tons C02/metric ton lead product
EFS	=	Emission factor for secondary materials, metric tons C02/metric ton lead product
For primary lead production using direct smelting, Sjardin (2003) and the IPCC (2006) provide an emission factor of
0.25 metric tons C02/metric ton lead. For secondary lead production, Sjardin (2003) and IPCC (2006) provide an
emission factor of 0.25 metric tons C02/metric ton lead for direct smelting, as well as an emission factor of 0.2
metric tons CO^metric ton lead produced for the treatment of secondary raw materials (i.e., pretreatment of lead
acid batteries). Since the secondary production of lead involves both the use of the direct smelting process and the
treatment of secondary raw materials, Sjardin recommends an additive emission factor to be used in conjunction
with the secondary lead production quantity. The direct smelting factor (0.25) and the sum of the direct smelting
and pretreatment emission factors (0.45) are multiplied by total U.S. primary and secondary lead production,
respectively, to estimate C02 emissions.
The production and use of coking coal for lead production is adjusted for within the Energy chapter as this fuel was
consumed during non-energy related activities. Additional information on the adjustments made within the Energy
sector for Non-Energy Use of Fuels is described in both the Methodology section of C02 from Fossil Fuel
Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating
Emissions of C02 from Fossil Fuel Combustion.
The 1990 through 2019 activity data for primary and secondary lead production (see Table 4-89) were obtained
from the U.S. Geological Survey (USGS 1995 through 2020).
Table 4-89: Lead Production (Metric Tons)
Year
Primary
Secondary
1990
404,000
922,000
2005
143,000
1,150,000
2015
0
1,050,000
2016
0
1,110,000
2017
0
1,140,000
2018
0
1,140,000
2019
0
1,200,000
Uncertainty and Time-Series Consistency
Uncertainty associated with lead production relates to the emission factors and activity data used. The direct
smelting emission factor used in primary production is taken from Sjardin (2003) who averaged the values
provided by three other studies (Dutrizac et al. 2000; Morris et al. 1983; Ullman 1997). For secondary production,
Sjardin (2003) added a C02 emission factor associated with battery treatment. The applicability of these emission
factors to plants in the United States is uncertain. There is also a smaller level of uncertainty associated with the
accuracy of primary and secondary production data provided by the USGS which is collected via voluntary surveys;
the uncertainty of the activity data is a function of the reliability of reported plant-level production data and the
completeness of the survey response.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-90. Lead production C02
emissions in 2019 were estimated to be between 0.5 and 0.6 MMT C02 Eq. at the 95 percent confidence level. This
indicates a range of approximately 14 percent below and 16 percent above the emission estimate of 0.5 MMT C02
Eq.
4-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-90: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lead
Production (MMT CO2 Eq. and Percent)
Source Gas
2019 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTCOzEq.) (%)


Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Lead Production C02
0.5
0.5 0.6
-14%
+16%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches discussed below were applied to applicable years to ensure time-series consistency in
emissions from 1990 through 2019. Details on the emission trends through time are described in more detail in the
Methodology section, above.
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter.
Initial review of activity data show that EPA's GHGRP Subpart R lead production data and resulting emissions are
fairly consistent with those reported by USGS. EPA is still reviewing available GHGRP data, reviewing QC analysis to
understand differences in data reporting (i.e., threshold implications), and assessing the possibility of including this
planned improvement in future Inventory reports (see Planned Improvements section below). Currently, GHGRP
data are used for QA purposes only.
Recalculations Discussion
No emissions recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
Pending resources and prioritization of improvements for more significant sources, EPA will continue to evaluate
and analyze data reported under EPA's GHGRP that would be useful to improve the emission estimates and
category-specific QC for the Lead Production source category, in particular considering completeness of reported
lead production given the reporting threshold. Particular attention will be made to ensuring time-series
consistency of the emissions estimates presented in future Inventory reports, consistent with IPCC and UNFCCC
guidelines. This is required as the facility-level reporting data from EPA's GHGRP, with the program's initial
requirements for reporting of emissions in calendar year 2010, are not available for all inventory years (i.e., 1990
through 2009) as required for this Inventory. In implementing improvements and integration of data from EPA's
GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories will be relied
upon.71
71 See .
Industrial Processes and Product Use 4-107

-------
4.22 Zinc Production (CRF Source Category
2C6)
Zinc production in the United States consists of both primary and secondary processes. Of the primary and
secondary processes used in the United States, only the electrothermic and Waelz kiln secondary processes result
in non-energy carbon dioxide (C02) emissions (Viklund-White 2000). Emissions from fuels consumed for energy
purposes during the production of zinc are accounted for in the Energy chapter.
The majority of zinc produced in the United States is used for galvanizing. Galvanizing is a process where zinc
coating is applied to steel in order to prevent corrosion. Zinc is used extensively for galvanizing operations in the
automotive and construction industry. Zinc is also used in the production of zinc alloys and brass and bronze alloys
(e.g., brass mills, copper foundries, and copper ingot manufacturing). Zinc compounds and dust are also used, to a
lesser extent, by the agriculture, chemicals, paint, and rubber industries.
Primary production in the United States is conducted through the electrolytic process, while secondary techniques
include the electrothermic and Waelz kiln processes, as well as a range of other metallurgical, hydrometallurgical,
and pyrometallurgical processes. Worldwide primary zinc production also employs a pyrometallurgical process
using the Imperial Smelting Furnace process; however, this process is not used in the United States (Sjardin 2003).
In the electrothermic process, roasted zinc concentrate and secondary zinc products enter a sinter feed where
they are burned to remove impurities before entering an electric retort furnace. Metallurgical coke is added to the
electric retort furnace as a carbon-containing reductant. This concentration step, using metallurgical coke and high
temperatures, reduces the zinc oxides and produces vaporized zinc, which is then captured in a vacuum
condenser. This reduction process also generates non-energy C02 emissions.
ZnO + C -» Zn(gas) + C02 (Reaction 1)
ZnO + CO -» Zn(gas) + C02 (Reaction 2)
In the Waelz kiln process, electric arc furnace (EAF) dust, which is captured during the recycling of galvanized steel,
enters a kiln along with a reducing agent (typically carbon-containing metallurgical coke). When kiln temperatures
reach approximately 1,100 to 1,200 degrees Celsius, zinc fumes are produced, which are combusted with air
entering the kiln. This combustion forms zinc oxide, which is collected in a baghouse or electrostatic precipitator,
and is then leached to remove chloride and fluoride. The use of carbon-containing metallurgical coke in a high-
temperature fuming process results in non-energy C02 emissions. Through this process, approximately 0.33 metric
tons of zinc is produced for every metric ton of EAF dust treated (Viklund-White 2000).
The only companies in the United States that use emissive technology to produce secondary zinc products are
American Zinc Recycling (AZR) (formerly "Horsehead Corporation"), Steel Dust Recycling (SDR), and PIZO. For AZR,
EAF dust is recycled in Waelz kilns at their Calumet, IL; Palmerton, PA; Rockwood, TN; and Barnwell, SC facilities.
The AZR facility in Beaumont, TX also processed EAF dust via flame reactor from 1993 through 2009 (AZR 2021,
Horsehead 2014). These Waelz kiln and flame reactor facilities produce intermediate zinc products (crude zinc
oxide or calcine), most of which was transported to their Monaca, PA facility where the products were smelted
into refined zinc using electrothermic technology. In April 2014, AZR permanently shut down their Monaca
smelter. This was replaced by their new facility in Mooresboro, NC in 2014.
The new Mooresboro facility uses a hydrometallurgical process (i.e., solvent extraction with electrowinning
technology) to produce zinc products. Hydrometallurgical production processes are assumed to be non-emissive
since no carbon is used in these processes (Sjardin 2003). The current capacity of the new facility is 155,000 short
tons. Production at the Mooresboro facility was idled in April 2016 and re-started in February 2020, with plans to
be at full capacity by 2021 (Recycling Today 2020). Direct consumption of coal, coke, and natural gas have been
replaced with electricity consumption at the new Mooresboro facility. The new facility is reported to have a
significantly lower greenhouse gas and other air emissions than the Monaca smelter (Horsehead 2012b).
4-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The Mooresboro facility uses leaching and solvent extraction (SX) technology combined with electrowinning,
melting, and casting technology. In this process, Waelz Oxide (WOX) is first washed in water to remove soluble
elements such as chlorine, potassium, and sodium, and then is leached in a sulfuric acid solution to dissolve the
contained zinc creating a pregnant liquor solution (PLS). The PLS is then processed in a solvent extraction step in
which zinc is selectively extracted from the PLS using an organic solvent creating a purified zinc-loaded electrolyte
solution. The loaded electrolyte solution is then fed into the electrowinning process in which electrical energy is
applied across a series of anodes and cathodes submerged in the electrolyte solution causing the zinc to deposit on
the surfaces of the cathodes. As the zinc metal builds up on these surfaces, the cathodes are periodically harvested
in order to strip the zinc from their surfaces (Horsehead 2015).
SDR and PIZO recycle EAF dust into intermediate zinc products using Waelz kilns and sell the intermediate products
to companies who smelt it into refined products.
Emissions of C02 from zinc production in 2019 were estimated to be 1.0 MMT C02 Eq. (1,026 kt C02) (see Table
4-91). All 2019 C02 emissions resulted from secondary zinc production processes. Emissions from zinc production
in the United States have increased overall since 1990 due to a gradual shift from non-emissive primary production
to emissive secondary production. In 2019, emissions were estimated to be 62 percent higher than they were in
1990.
Table 4-91: CO2 Emissions from Zinc Production (MMT CO2 Eq. and kt)
Year MMT CP2 Eq.	kt_
1990	0.6	632
2005	1.0	1,030
2015	0.9	886
2016	0.8	838
2017	0.9	900
2018	1.0	999
2019	1.0	1,026
In 2019, United States primary and secondary refined zinc production were estimated to total 120,000 metric tons
(USGS 2020) (see Table 4-92). Domestic zinc mine production decreased in 2019, owing partially to the closure of
the Pend Oreille Mine in Washington State in July after current reserves were exhausted. The mine was reopened
in December 2014 after being closed since 2009 (USGS 2020). Primary zinc production (primary slab zinc) in 2018 is
used as a proxy for 2019, while secondary zinc production in 2019 increased by 27 percent compared to 2018.
Table 4-92: Zinc Production (Metric Tons)
Year
Primary
Secondary
Total
1990
262,704
95,708
358,412
2005
191,120
156,000
347,120
2015
122,857
49,143
172,000
2016
111,000
15,000
126,000
2017
117,000
15,000
132,000
2018
101,000
15,000
116,000
2019
101,000
19,000
120,000
Industrial Processes and Product Use 4-109

-------
Methodology
The methods used to estimate non-energy C02 emissions from zinc production72 using the electrothermic primary
production and Waelz kiln secondary production processes are based on Tier 1 methods from the 2006IPCC
Guidelines (IPCC 2006). The Tier 1 equation used to estimate emissions from zinc production is as follows:
Eco2 ~ Zn x EF[iej:auit
where,
ECo2 =	C02 emissions from zinc production, metric tons
Zn =	Quantity of zinc produced, metric tons
EFdefauit =	Default emission factor, metric tons C02/metric ton zinc produced
The Tier 1 emission factors provided by IPCC for Waelz kiln-based secondary production were derived from
metallurgical coke consumption factors and other data presented in Vikland-White (2000). These coke
consumption factors as well as other inputs used to develop the Waelz kiln emission factors are shown below. IPCC
does not provide an emission factor for electrothermic processes due to limited information; therefore, the Waelz
kiln-specific emission factors were also applied to zinc produced from electrothermic processes. Starting in 2014,
refined zinc produced in the United States used hydrometallurgical processes and is assumed to be non-emissive.
For Waelz kiln-based production, IPCC recommends the use of emission factors based on EAF dust consumption, if
possible, rather than the amount of zinc produced since the amount of reduction materials used is more directly
dependent on the amount of EAF dust consumed. Since only a portion of emissive zinc production facilities
consume EAF dust, the emission factor based on zinc production is applied to the non-EAF dust consuming
facilities, while the emission factor based on EAF dust consumption is applied to EAF dust consuming facilities.
The Waelz kiln emission factor based on the amount of zinc produced was developed based on the amount of
metallurgical coke consumed for non-energy purposes per ton of zinc produced (i.e., 1.19 metric tons coke/metric
ton zinc produced) (Viklund-White 2000), and the following equation:
1.19 metric tons coke 0.85 metric tons C 3.67 metric tons C02 3.70 metric tons C02
Ktill	*	*	^	^	,	,	,
metric tons zinc metric tons coke	metric tons C	metric tons zinc
Refined zinc production levels for AZR's Monaca, PA facility (utilizing electrothermic technology) were available
from the company for years 2005 through 2013 (Horsehead 2008, 2011, 2012, 2013, and 2014). The Monaca
facility was permanently shut down in April 2014 and replaced by AZR's new facility in Mooresboro, NC. The new
facility uses hydrometallurgical process to produce refined zinc products. Hydrometallurgical production processes
are assumed to be non-emissive since no carbon is used in these processes (Sjardin 2003).
Metallurgical coke consumption for non-EAF dust consuming facilities for 1990 through 2004 were extrapolated
using the percentage change in annual refined zinc production at secondary smelters in the United States, as
provided by the U.S. Geological Survey (USGS) Minerals Yearbook: Zinc (USGS 1995 through 2006). Metallurgical
coke consumption for 2005 through 2013 were based on the secondary zinc production values obtained from the
Horsehead Corporation Annual Report Form 10-k: 2005 through 2008 from the 2008 10-k (Horsehead Corp 2009);
2009 and 2010 from the 2010 10-k (Horsehead Corp. 2011); 2011 from the 201110-k (Horsehead Corp. 2012a);
2012 from the 2012 10-k (Horsehead Corp. 2013); and 2013 from the 2013 10-k (Horsehead Corp. 2014).
Metallurgical coke consumption levels for 2014 and later were zero due to the closure of the AZR (formerly
"Horsehead Corporation") Monaca, PA electrothermic furnace facility. The secondary zinc produced values for
72 EPA has not integrated aggregated facility-level Greenhouse Gas Reporting Program (GHGRP) information to inform these
estimates. The aggregated information (e.g., activity data and emissions) associated with Zinc Production did not meet criteria
to shield underlying confidential business information (CBI) from public disclosure.
4-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
each year were then multiplied by the 3.70 metric tons CO^metric ton zinc produced emission factor to develop
C02 emission estimates for the AZR electrothermic furnace facility.
The Waelz kiln emission factor based on the amount of EAF dust consumed was developed based on the amount
of metallurgical coke consumed per ton of EAF dust consumed (i.e., 0.4 metric tons coke/metric ton EAF dust
consumed) (Viklund-White 2000), and the following equation:
OA metric tons coke 0.85 metric tons C 3.67 metric tons C02 1.24 metric tons C02
EFgAp Dust —	x	x	—
ivci nt tons EAF Dust metric tons coke	metric tons C	metric tons EAF Dust
Metallurgical coke consumption for EAF dust consuming facilities for 1990 through 2019 were calculated based on
the values of EAF dust consumed. The values of EAF dust consumed for AZR, SDR, and PIZO are explained below.
The total amount of EAF dust consumed by AZR at their Waelz kilns was available from AZR (formerly "Horsehead
Corporation") financial reports for years 2006 through 2015 (Horsehead 2007, 2008, 2010a, 2011, 2012a, 2013,
2014, 2015, and 2016) and from AZR for 2016, 2017, 2018, and 2019 (AZR 2020). The EAF dust consumption values
for each year were then multiplied by the 1.24 metric tons C02/metric ton EAF dust consumed emission factor to
develop C02 emission estimates for AZR's Waelz kiln facilities.
The amount of EAF dust consumed by SDR and their total production capacity were obtained from SDR's facility in
Alabama for the years 2011 through 2019 (SDR 2012, 2014, 2015, 2017, 2018, 2021). The SDR facility has been
operational since 2008, underwent expansion in 2011 to include a second unit (operational since early- to mid-
2012), and expanded its capacity again in 2017 (SDR 2018). Annual consumption data for SDR was not publicly
available for the years 2008, 2009, and 2010. These data were estimated using data for AZR's Waelz kilns for 2008
through 2010 (Horsehead 2007, 2008, 2010a, 2010b, and 2011). Annual capacity utilization ratios were calculated
using AZR's annual consumption and total capacity for the years 2008 through 2010. AZR's annual capacity
utilization ratios were multiplied with SDR's total capacity to estimate SDR's consumption for each of the years,
2008 through 2010 (SDR 2013). The 1.24 metric tons COVmetric ton EAF dust consumed emission factor was then
applied to SDR's estimated EAF dust consumption to develop C02 emission estimates for those Waelz kiln facilities.
PIZO Technologies Worldwide LLC's facility in Arkansas was operational from 2009 to 2012 (PIZO 2021). The
amount of EAF dust consumed by PIZO's facility for 2009 through 2012 was not publicly available. EAF dust
consumption for PIZO's facility for 2009 and 2010 were estimated by calculating annual capacity utilization of
AZR's Waelz kilns and multiplying this utilization ratio by PIZO's total capacity (PIZO 2012). EAF dust consumption
for PIZO's facility for 2011 through 2012 were estimated by applying the average annual capacity utilization rates
for AZR and SDR (Grupo PROMAX) to PIZO's annual capacity (Horsehead 2012; SDR 2012; PIZO 2012). The 1.24
metric tons COz/metric ton EAF dust consumed emission factor was then applied to PIZO's estimated EAF dust
consumption to develop C02 emission estimates for those Waelz kiln facilities.
The production and use of coking coal for zinc production is adjusted for within the Energy chapter as this fuel was
consumed during non-energy related activities. Additional information on the adjustments made within the Energy
sector for Non-Energy Use of Fuels is described in both the Methodology section of C02 from Fossil Fuel
Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating
Emissions of C02 from Fossil Fuel Combustion.
Beginning with the 2017 USGS Minerals Commodity Summary: Zinc, United States primary and secondary refined
zinc production were reported as one value, total refined zinc production. Prior to this publication, primary and
secondary refined zinc production statistics were reported separately. For the current Inventory report, EPA
sought expert judgment from the USGS mineral commodity expert to assess approaches for splitting total
production into primary and secondary values. For years 2016 through 2019, only one facility produced primary
zinc. Primary zinc produced from this facility was subtracted from the USGS 2016 to 2019 total zinc production
statistic to estimate secondary zinc production for these years.
Uncertainty and Time-Series Consistency
The uncertainty associated with these estimates is two-fold, relating to activity data and emission factors used.
Industrial Processes and Product Use 4-111

-------
First, there is uncertainty associated with the amount of EAF dust consumed in the United States to produce
secondary zinc using emission-intensive Waelz kilns. The estimate for the total amount of EAF dust consumed in
Waelz kilns is based on (1) an EAF dust consumption value reported annually by AZR/Horsehead Corporation as
part of its financial reporting to the Securities and Exchange Commission (SEC) and provided by AZR, and (2) an EAF
dust consumption value obtained from the Waelz kiln facility operated in Alabama by Steel Dust Recycling LLC.
Since actual EAF dust consumption information is not available for PIZO's facility (2009 through 2010) and SDR's
facility (2008 through 2010), the amount is estimated by multiplying the EAF dust recycling capacity of the facility
(available from the company's website) by the capacity utilization factor for AZR (which is available from
Horsehead Corporation financial reports).The EAF dust consumption for PIZO's facility for 2011 through 2012 was
estimated by multiplying the average capacity utilization factor developed from AZR and SDR's annual capacity
utilization rates by PIZO's EAF dust recycling capacity. Therefore, there is uncertainty associated with the
assumption used to estimate PIZO's annual EAF dust consumption values for 2009 through 2012 and SDR's annual
EAF dust consumption values for 2008 through 2010.
Second, there is uncertainty associated with the emission factors used to estimate C02 emissions from secondary
zinc production processes. The Waelz kiln emission factors are based on materials balances for metallurgical coke
and EAF dust consumed as provided by Viklund-White (2000). Therefore, the accuracy of these emission factors
depend upon the accuracy of these materials balances. Data limitations prevented the development of emission
factors for the electrothermic process. Therefore, emission factors for the Waelz kiln process were applied to both
electrothermic and Waelz kiln production processes.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-93. Zinc production C02
emissions from 2019 were estimated to be between 0.8 and 1.2 MMT C02 Eq. at the 95 percent confidence level.
This indicates a range of approximately 19 percent below and 21 percent above the emission estimate of 1.0 MMT
C02 Eq.
Table 4-93: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Zinc
Production (MMT CO2 Eq. and Percent)
Source
Gas 2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3

(MMT CO? Eq.)
(MMTCOz Eq.)
(%)



Lower Upper
Lower
Upper


Bound Bound
Bound
Bound
Zinc Production
C02 1.0

-------
Planned Improvements
Pending resources and prioritization of improvements for more significant sources, EPA will continue to evaluate
and analyze data reported under EPA's GHGRP that would be useful to improve the emission estimates and
category-specific QC for the Zinc Production source category, in particular considering completeness of reported
zinc production given the reporting threshold. Given the small number of facilities in the United States, particular
attention will be made to risks for disclosing CBI and ensuring time-series consistency of the emissions estimates
presented in future Inventory reports, consistent with IPCC and UNFCCC guidelines. This is required as the facility-
level reporting data from EPA's GHGRP, with the program's initial requirements for reporting of emissions in
calendar year 2010, are not available for all inventory years (i.e., 1990 through 2009) as required for this Inventory.
In implementing improvements and integration of data from EPA's GHGRP, the latest guidance from the IPCC on
the use of facility-level data in national inventories will be relied upon.73 This is a long-term planned improvement,
and EPA is still assessing the possibility of including this improvement in future Inventory reports.
4.23 Electronics Industry (CRF Source
Category 2E)
The electronics industry uses multiple greenhouse gases in its manufacturing processes. In semiconductor
manufacturing, these include long-lived fluorinated greenhouse gases used for plasma etching and chamber
cleaning (CRF Source Category 2E1), fluorinated heat transfer fluids (CRF Source Category 2E4) used for
temperature control and other applications, and nitrous oxide (N20) used to produce thin films through chemical
vapor deposition (reported under CRF Source Category 2H3). Similar to semiconductor manufacturing, the
manufacturing of micro-electro-mechanical systems (MEMS) devices (reported under CRF Source Category 2E5
Other) and photovoltaic cells (CRF Source Category 2E3) requires the use of multiple long-lived fluorinated
greenhouse gases for various processes.
The gases most commonly employed in plasma etching and chamber cleaning are trifluoromethane
(hydrofluorocarbon (HFC)-23 or CHF3), perfluoromethane (CF4), perfluoroethane (C2F6), nitrogen trifluoride (NF3),
and sulfur hexafluoride (SF6), although other fluorinated compounds such as perfluoropropane (C3F8) and
perfluorocyclobutane (c-C4F8) are also used. The exact combination of compounds is specific to the process
employed.
In addition to emission estimates for these seven commonly used fluorinated gases, this Inventory contains
emissions estimates for N20 and a combination of other HFCs and unsaturated, low-GWP PFCs such as C5F8,C4F6,
HFC-32, and HFC-134a. These additional HFCs and PFCs are emitted from etching and chamber cleaning processes
in much smaller amounts, accounting for less than 0.02 percent of emissions (in C02 Eq.) from these processes.
These gases have been grouped as "other fluorinated gases" for the purpose of this analysis.
For semiconductors, a single 300 mm silicon wafer that yields between 400 to 600 semiconductor products
(devices or chips) may require more than 100 distinct fluorinated-gas-using process steps, principally to deposit
and pattern dielectric films. Plasma etching (or patterning) of dielectric films, such as silicon dioxide and silicon
nitride, is performed to provide pathways for conducting material to connect individual circuit components in each
device. The patterning process uses plasma-generated fluorine atoms, which chemically react with exposed
dielectric film to selectively remove the desired portions of the film. The material removed as well as undissociated
fluorinated gases flow into waste streams and, unless emission abatement systems are employed, into the
73 See .
Industrial Processes and Product Use 4-113

-------
atmosphere. Plasma enhanced chemical vapor deposition (PECVD) chambers, used for depositing dielectric films,
are cleaned periodically using fluorinated and other gases. During the cleaning cycle the gas is converted to
fluorine atoms in plasma, which etches away residual material from chamber walls, electrodes, and chamber
hardware. Undissociated fluorinated gases and other products pass from the chamber to waste streams and,
unless abatement systems are employed, into the atmosphere.
In addition to emissions of unreacted gases, some fluorinated compounds can also be transformed in the plasma
processes into different fluorinated compounds which are then exhausted, unless abated, into the atmosphere.
For example, when C2F6 is used in cleaning or etching, CF4 is typically generated and emitted as a process
byproduct. In some cases, emissions of the byproduct gas can rival or even exceed emissions of the input gas, as is
the case for NF3 used in remote plasma chamber cleaning, which often generates CF4 as a byproduct.
Besides dielectric film etching and PECVD chamber cleaning, much smaller quantities of fluorinated gases are used
to etch polysilicon films and refractory metal films like tungsten.
Nitrous oxide is used in manufacturing semiconductor devices to produce thin films by CVD and nitridation
processes as well as for N-doping of compound semiconductors and reaction chamber conditioning (Doering
2000).
Liquid perfluorinated compounds are also used as heat transfer fluids (F-HTFs) for temperature control, device
testing, cleaning substrate surfaces and other parts, and soldering in certain types of semiconductor
manufacturing production processes. Leakage and evaporation of these fluids during use is a source of fluorinated
gas emissions (EPA 2006). Unweighted F-HTF emissions consist primarily of perfluorinated amines,
hydrofluoroethers, perfluoropolyethers (specifically, PFPMIEs), and perfluoroalkylmorpholines. One percent or less
consist of HFCs, PFCs, and SF6 (where PFCs are defined as compounds including only carbon and fluorine). With the
exceptions of the hydrofluoroethers and most of the HFCs, all of these compounds are very long-lived in the
atmosphere and have global warming potentials (GWPs) near 10,000.74
For 2019, total GWP-weighted emissions of all fluorinated greenhouse gases and N20 from deposition, etching,
and chamber cleaning processes in the U.S. semiconductor industry were estimated to be 4.6 MMT C02 Eq. Less
than 0.02 percent of total emissions from semiconductor manufacturing consist of a combination of HFCs other
than HFC-23 and unsaturated, low-GWP PFCs including C4F6, C4F80, C5F8, HFC-32, HFC-41, and HFC-134a. These
gases have been grouped as "Other F-GHGs". Emissions from all fluorinated greenhouse gases and N20 are
presented in Table 4-94 and Table 4-95 below for the years 1990, 2005, and the period 2015 to 2019. Emissions of
F-HTFs that are HFCs, PFCs or SF6 are presented in Table 4-94. Table 4-96 shows F-HTF emissions in tons by
compound group based on reporting to EPA's Greenhouse Gas Reporting Program (GHGRP) during years 2013
through 2019. Emissions of F-HTFs that are not HFCs, PFCs or SF6 are not included in inventory totals and are
included for informational purposes only.
The rapid growth of this industry and the increasing complexity (growing number of layers)75 of semiconductor
products led to an increase in emissions of 153 percent between 1990 and 1999, when emissions peaked at 9.1
MMT C02 Eq. Emissions began to decline after 1999, reaching a low point in 2009 before rebounding slightly and
more or less plateauing at the current level, which represents a 49 percent decline from 1999 levels. Together,
industrial growth, adoption of emissions reduction technologies (including but not limited to abatement
74	The GWP of PFPMIE, a perfluoropolyether used as an F-HTF, is included in the IPCC Fourth Assessment Report with a value of
10,300. The GWPs of the perfluorinated amines and perfluoroalkylmorpholines that are used as F-HTFs have not been
evaluated in the peer-reviewed literature. However, evaluations by the manufacturer indicate that their GWPs are near 10,000
(78 FR 20632), which is expected given that these compounds are both saturated and fully fluorinated. EPA assigns a default
GWP of 10,000 to compounds that are both saturated and fully fluorinated and that do not have chemical-specific GWPs in
either the Fourth or the Fifth Assessment Reports.
75	Complexity is a term denoting the circuit required to connect the active circuit elements (transistors) on a chip. Increasing
miniaturization, for the same chip size, leads to increasing transistor density, which, in turn, requires more complex
interconnections between those transistors. This increasing complexity is manifested by increasing the levels (i.e., layers) of
wiring, with each wiring layer requiring fluorinated gas usage for its manufacture.
4-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
technologies) and shifts in gas usages resulted in a net increase in emissions of approximately 29 percent between
1990 and 2019. Total emissions from semiconductor manufacture in 2019 were slightly lower than 2018 emissions,
decreasing by 9 percent. This is likely due in part to reduced production in 2019 as compared to 2018. Increased
abatement of F-GHGs also contributed to the decrease in emissions.
The emissions reported by facilities manufacturing MEMS included emissions of C2F6, C3F8, c-C4F8, CF4, HFC-23, NF3,
and SF6,76 and were equivalent to only 0.09 percent to 0.34 percent of the total reported emissions from
electronics manufacturing in 2011 to 2019. Emissions ranged from 0.0009 to 0.0185 MMT C02 Eq. from 1991 to
2019. Based upon information in the World Fab Forecast (WFF), it appears that some GHGRP reporters that
manufacture both semiconductors and MEMS are reporting their emissions as only from semiconductor
manufacturing (GHGRP reporters must choose a single classification per fab). Emissions from non-reporters have
not been estimated.
Total GWP-weighted emissions from manufacturing of photovoltaic cells were estimated to range from 0.0003
MMT C02 Eq. to 0.0326 MMT C02 Eq. from 1998 to 2019 and were equivalent to between 0.003 percent to 0.67
percent of the total reported emissions from electronics manufacturing. Emissions from manufacturing of
photovoltaic cells were estimated based on reported data from a single manufacturer between 2015 and 2017.
Reported emissions from photovoltaic cell manufacturing consisted of CF4, C2F6, c-C4F8, and CHF3.77
Emissions from all fluorinated greenhouse gases from photovoltaic and MEMS manufacturing are in Table 4-94.
While EPA has developed an elementary methodology to estimate emissions from non-reporters and to back-cast
emissions from these sources for the entire time series, there is very high uncertainty associated with these
emissions.
Only F-HTF emissions that consist of HFC, PFC and SF6 are included in the Inventory totals; emissions of other F-
HTFs, which account for the vast majority of F-HTF emissions, are provided for informational purposes and are not
included in the Inventory totals. Since reporting of F-HTF emissions began under EPA's GHGRP in 2011, total F-HTF
emissions (reported and estimated non-reported) have fluctuated between 0.6 MMT C02 Eq. and 0.9 MMT C02
Eq., with an overall declining trend. An analysis of the data reported to EPA's GHGRP indicates that F-HTF
emissions account for anywhere between 13 percent and 19 percent of total annual emissions (F-GHG, N20 and F-
HTFs) from semiconductor manufacturing.78 Table 4-96 shows F-HTF emissions in tons by compound group based
on reporting to EPA's GHGRP during years 2012 through 2019.79
Table 4-94: PFC, HFC, SFe, NF3, and N2O Emissions from Electronics Manufacture80 (MMT
COz Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
cf4
0.8
1.1
1.5
1.5
1.6
1.7
1.6
c2f6
2.0
2.0
1.3
1.2
1.2
1.1
0.9
C3Fs
+
0.1
0.1
0.1
0.1
0.1
0.1
76	Gases not reported by MEMS manufacturers to the GHGRP are currently listed as "NE" in the CRF. Since no facilities report
using these gases, emissions of these gases are not estimated for this sub-sector. However, there is insufficient data to
definitively conclude that they are not used by non-reporting facilities.
77	Gases not reported by PV manufacturers to the GHGRP are currently listed as "NE" in the CRF. Since no facilities report using
these gases, emissions of these gases are not estimated for this sub-sector. However, there is insufficient data to definitively
conclude that they are not used by non-reporting facilities.
78	Emissions data for HTFs (in tons of gas) from the semiconductor industry from 2011 through 2019 were obtained from the
EPA GHGRP annual facility emissions reports.
79	Many fluorinated heat transfer fluids consist of perfluoropolymethylisopropyl ethers (PFPMIEs) of different molecular
weights and boiling points that are distilled from a mixture. "BP 200 °C" (and similar terms below) indicate the boiling point of
the fluid in degrees Celsius. For more information, see .
80	An extremely small portion of emissions included in the totals for Electronics Manufacture are from the manufacturing of
MEMS and photovoltaic cells.
Industrial Processes and Product Use 4-115

-------
c-C4Fs
0.0
0.1
0.1
0.1
0.1
0.1
0.1
HFC-23
0.2
0.2
0.3
0.3
0.4
0.4
0.3
sf6
0.5
0.7
0.7
0.8
0.7
0.8
0.8
nf3
+
0.5
0.6
0.6
0.6
0.6
0.6
Other F-GHGs
+
+
+
+
+
+
+
Total F-GHGs
3.6
4.6
4.7
4.7
4.6
4.8
4.3
N2081
+
0.1
0.2
0.2
0.3
0.3
0.2
HFC, PFC and SF6 F-HTFs
0.000
0.012
0.019
0.018
0.021
0.022
0.029
MEMS
0.000
0.013
0.006
0.005
0.006
0.008
0.011
PV
0.000
0.005
0.033
0.029
0.029
0.029
0.029
Total
3.6
4.8
5.0
5.0
4.9
5.1
4.6
+ Does not exceed 0.05 MMT C02 Eq.
Table 4-95: PFC, HFC, SFe, NF3, and N2O Emissions from Electronics Manufacture (Metric
Tons)
Year
1990
2005
2015
2016
2017
2018
2019
cf4
115
146
206
208
218
232
214
c2f6
160
162
110
99
96
92
77
CsFs
0
9
16
14
12
12
10
c-C4Fs
0
11
6
5
6
6
6
HFC-23
15
14
22
23
24
25
21
sf6
22
31
32
36
31
33
33
nf3
3
29
33
33
35
36
35
n2o
120
412
789
792
920
857
755
Total
435
813
1,213
1,212
1,342
1,293
1,151
Table 4-96: F-HTF Emissions from Electronics Manufacture by Compound Group (Metric
Tons)
Year
2013
2014
2015
2016
2017
2018
2019
HFCs
4.6
2.0
1.6
2.7
1.7
1.7
2.4
PFCs
0.4
0.2
0.3
0.3
0.2
0.4
0.3
sf6
0.4
0.9
0.6
0.5
0.7
0.6
0.3
HFEs
25.4
25.2
19.0
13.5
16.5
23.3
7.2
PFPMIEs
18.8
18.2
20.8
17.3
14.3
17.7
16.2
Perfluoalkylromorpholines
10.7
10.8
8.1
7.6
5.2
5.9
5.7
Perfluorotrialkylamines
29.5
49.3
43.8
38.6
37.6
40.0
34.4
Total F-HTFs
89.9
106.7
94.4
80.6
76.2
89.7
66.4
Table 4-97: F-GHGa Emissions from PV and MEMS manufacturing (MMT CO2 Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
MEMS
0.0
0.013
0.006
0.005
0.006
0.008
0.011
PV
0.0
0.005
0.033
0.029
0.029
0.029
0.029
a F-GHGs from PV manufacturing include an unspecified mix of HFCs and PFCs, F-GHGs from MEMS
manufacturing includes those gases but also NF3 and SF6.
81 Emissions of N20 from semiconductor manufacturing are reported in the CRF under 2H3.
4-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
Emissions are based on data reported through Subpart I, Electronics Manufacture, of EPA's GHGRP, Partner-
reported emissions data received through EPA's PFC82 Reduction/Climate Partnership, EPA's PFC Emissions Vintage
Model (PEVM)—a model that estimates industry emissions from etching and chamber cleaning processes in the
absence of emission control strategies (Burton and Beizaie 2001)83—and estimates of industry activity (i.e., total
manufactured layer area). The availability and applicability of reported emissions data from the EPA Partnership
and EPA's GHGRP and activity data differ across the 1990 through 2019 time series. Consequently, fluorinated
greenhouse gas (F-GHG) emissions from etching and chamber cleaning processes for semiconductors were
estimated using seven distinct methods, one each for the periods 1990 through 1994,1995 through 1999, 2000
through 2006, 2007 through 2010, 2011 and 2012, 2013 and 2014, and 2015 through 2019. Nitrous oxide
emissions were estimated using five distinct methods, one each for the period 1990 through 1994,1995 through
2010, 2011 and 2012, 2013 and 2014, and 2015 through 2019. The methodology discussion below for these time
periods focuses on semiconductor emissions from etching, chamber cleaning, and uses of N20. Other emissions for
MEMS, PV, and HTFs were estimated using the approaches described immediately below.
GHGRP-reported emissions from the manufacturing of MEMS are available for the years 2011 to 2019. Emissions
from fabs that reported to the GHGRP as manufacturing MEMS are not included in the semiconductor
manufacturing totals reported above. Emissions from manufacturing of MEMS for years prior to 2011 were
calculated by linearly interpolating emissions between 1990 (at zero MMT C02 Eq.) and 2011, the first year where
emissions from manufacturing of MEMS was reported to the GHGRP. Based upon information in the World Fab
Forecast (WFF), it appears that some GHGRP reporters that manufacture both semiconductors and MEMS are
reporting their emissions as only from semiconductor manufacturing; however, emissions from MEMS
manufacturing are likely being included in semiconductor totals. Emissions were not estimated for non-reporters.
GHGRP-reported emissions from the manufacturing of photovoltaic cells are only available between 2015 and
2017 and are from a single manufacturer. These reported emissions are scaled by the ratio of reporters to non-
reporters to estimate the total U.S. emissions from PV. EPA estimates the emissions from manufacturing of PVs
from non-reporting facilities by multiplying the estimated capacity of non-reporters by a calculated emissions
factor based on GHGRP reported emissions per megawatt from 2015 and 2016. Manufacturing capacities in
megawatts were drawn from DisplaySearch, a 2015 Congressional Research Service Report on U.S. Solar
Photovoltaic Manufacturing,84 and self-reported capacity by the GHGRP reporter.85 EPA estimated that during the
2015 to 2017 period, 28 percent of emissions were reported through the GHGRP. These capacities are estimated
for the full time series by linearly scaling the total U.S. capacity between zero in 1997 to the total capacity reported
of crystalline silicon (c-Si) PV manufacturing in 2000 in DisplaySearch and then linearly scaling between the total
capacity of c-Si PV manufacturing in DisplaySearch in 2009 to the total capacity of c-Si PV manufacturing reported
in the Congressional Research Service report in 2012. Capacities were held constant for non-reporters for 2012 to
2019. Average emissions per MW from the GHGRP reporter in 2015 and 2016 were then applied to the total
capacity prior to 2015. Emissions for 2014 from the GHGRP reporter were scaled to the number of months open in
2014. For 2016 and 2017, emissions per MW (capacity) from the GHGRP reporter were applied to the non-
reporters. For 2018 and 2019, emissions were held constant to 2017 estimates, since there is no evidence that
much growth has occurred in the U.S. PV cell manufacturing industry in the last two years.
Facility emissions of F-HTFs from semiconductor manufacturing are reported to EPA under its GHGRP and are
available for the years 2011 through 2019. EPA estimates the emissions of F-HTFs from non-reporting facilities by
82	In the context of the EPA Partnership and PEVM, PFC refers to perfluorocompounds, not perfluorocarbons.
83	A Partner refers to a participant in the U.S. EPA PFC Reduction/Climate Partnership for the Semiconductor Industry. Through
a Memorandum of Understanding (MoU) with the EPA, Partners voluntarily reported their PFC emissions to the EPA by way of a
third party, which aggregated the emissions through 2010.
84	Platzer, Michaela D. (2015) U.S. Solar Photovoltaic Manufacturing: Industry Trends, Global Competition, Federal Support.
Congressional Research Service. January 27, 2015. .
85	DisplaySearch. 2010. DisplaySearch Q4'09 Quarterly FPD Supply/Demand and Capital Spending Report. DisplaySearch, LLC.
Industrial Processes and Product Use 4-117

-------
calculating the ratio of GHGRP-reported fluorinated HTF emissions to GHGRP reported F-GHG emissions from
etching and chamber cleaning processes, and then multiplying this ratio by the F-GHG emissions from etching and
chamber cleaning processes estimated for non-reporting facilities. Fluorinated HTF use in semiconductor
manufacturing is assumed to have begun in the early 2000s and to have gradually displaced other HTFs (e.g., de-
ionized water and glycol) in electronics manufacturing (EPA 2006). For time-series consistency, EPA interpolated
the share of F-HTF emissions to F-GHG emissions between 2000 (at 0 percent) and 2011 (at 22 percent) and
applied these shares to the unadjusted F-GHG emissions during those years to estimate the fluorinated HTF
emissions.
1990 through 1994
From 1990 through 1994, Partnership data were unavailable and emissions were modeled using PEVM (Burton and
Beizaie 2001).86 The 1990 to 1994 emissions are assumed to be uncontrolled, since reduction strategies such as
chemical substitution and abatement were yet to be developed.
PEVM is based on the recognition that fluorinated greenhouse gas emissions from semiconductor manufacturing
vary with: (1) the number of layers that comprise different kinds of semiconductor devices, including both silicon
wafer and metal interconnect layers, and (2) silicon consumption (i.e., the area of semiconductors produced) for
each kind of device. The product of these two quantities, Total Manufactured Layer Area (TMLA), constitutes the
activity data for semiconductor manufacturing. PEVM also incorporates an emission factor that expresses
emissions per unit of manufactured layer-area. Emissions are estimated by multiplying TMLA by this emission
factor.
PEVM incorporates information on the two attributes of semiconductor devices that affect the number of layers:
(1) linewidth technology (the smallest manufactured feature size),87 and (2) product type (discrete, memory or
logic).88 For each linewidth technology, a weighted average number of layers is estimated using VLSI product-
specific worldwide silicon demand data in conjunction with complexity factors (i.e., the number of layers per
Integrated Circuit (IC) specific to product type (Burton and Beizaie 2001; ITRS 2007). PEVM derives historical
consumption of silicon (i.e., square inches) by linewidth technology from published data on annual wafer starts
and average wafer size (VLSI Research, Inc. 2012).
The emission factor in PEVM is the average of four historical emission factors, each derived by dividing the total
annual emissions reported by the Partners for each of the four years between 1996 and 1999 by the total TMLA
estimated for the Partners in each of those years. Over this period, the emission factors varied relatively little (i.e.,
the relative standard deviation for the average was 5 percent). Since Partners are believed not to have applied
significant emission reduction measures before 2000, the resulting average emission factor reflects uncontrolled
emissions and hence may be use here to estimate 1990 through 1994 emissions. The emission factor is used to
estimate U.S. uncontrolled emissions using publicly-available data on world (including U.S.) silicon consumption.
86	Various versions of the PEVM exist to reflect changing industrial practices. From 1990 to 1994 emissions estimates are from
PEVM vl.0, completed in September 1998. The emission factor used to estimate 1990 to 1994 emissions is an average of the
1995 and 1996 emissions factors, which were derived from Partner reported data for those years.
87	By decreasing features of Integrated Circuit components, more components can be manufactured per device, which
increases its functionality. However, as those individual components shrink it requires more layers to interconnect them to
achieve the functionality. For example, a microprocessor manufactured with 65 nm feature sizes might contain as many as 1
billion transistors and require as many as 11 layers of component interconnects to achieve functionality, while a device
manufactured with 130 nm feature size might contain a few hundred million transistors and require 8 layers of component
interconnects (ITRS 2007).
88	Memory devices manufactured with the same feature sizes as microprocessors (a logic device) require approximately one-
half the number of interconnect layers, whereas discrete devices require only a silicon base layer and no interconnect layers
(ITRS 2007). Since discrete devices did not start using PFCs appreciably until 2004, they are only accounted for in the PEVM
emissions estimates from 2004 onwards.
4-118 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
As it was assumed for this time period that there was no consequential adoption of fluorinated-gas-reducing
measures, a fixed distribution of fluorinated-gas use was assumed to apply to the entire U.S. industry to estimate
gas-specific emissions. This distribution was based upon the average fluorinated-gas purchases made by
semiconductor manufacturers during this period and the application of IPCC default emission factors for each gas
(Burton and Beizaie 2001).
PEVM only addressed the seven main F-GHGs (CF4, C2F6, C3F8, c-C4F8, HFC-23, SF6, and NF3) used in semiconductor
manufacturing. Through reporting under Subpart I of EPA's GHGRP, data on other F-GHGs (C4F6, C5F8, HFC-32, HFC-
41, HFC-134a) used in semiconductor manufacturing became available and EPA was therefore able to extrapolate
this data across the entire 1990 to 2018 timeseries. To estimate emissions for these "other F-GHGs", emissions
data from Subpart I were used to estimate the average share or percentage contribution of these gases as
compared to total F-GHG emissions and then these shares were applied to all years prior to reported data from
Subpart I (1990 through 2010) and to the emissions from non-reporters from 2011 to 2018.
To estimate N20 emissions, it is assumed the proportion of N20 emissions estimated for 1995 (discussed below)
remained constant for the period of 1990 through 1994.
1995 through 1999
For 1995 through 1999, total U.S. emissions were extrapolated from the total annual emissions reported by the
Partners (1995 through 1999). Partner-reported emissions are considered more representative (e.g., in terms of
capacity utilization in a given year) than PEVM-estimated emissions, and are used to generate total U.S. emissions
when applicable. The emissions reported by the Partners were divided by the ratio of the total capacity of the
plants operated by the Partners and the total capacity of all of the semiconductor plants in the United States; this
ratio represents the share of capacity attributable to the Partnership. This method assumes that Partners and non-
Partners have identical capacity utilizations and distributions of manufacturing technologies. Plant capacity data is
contained in the World Fab Forecast (WFF) database and its predecessors, which is updated quarterly. Gas-specific
emissions were estimated using the same method as for 1990 through 1994.
For this time period emissions of other F-GHGs (C4F6, C5F8, HFC-32, HFC-41, HFC-134a) were estimated using the
method described above for 1990 to 1994.
For this time period, the N20 emissions were estimated using an emission factor that was applied to the annual,
total U.S. TMLA manufactured. The emission factor was developed using a regression-through-the-origin (RTO)
model: GHGRP reported N20 emissions were regressed against the corresponding TMLA of facilities that reported
no use of abatement systems. Details on EPA's GHGRP reported emissions and development of emission factor
using the RTO model are presented in the 2011 through 2012 section. The total U.S. TMLA for 1995 through 1999
was estimated using PEVM.
2000 through 2006
Emissions for the years 2000 through 2006—the period during which Partners began the consequential application
of fluorinated greenhouse gas-reduction measures—were estimated using a combination of Partner-reported
emissions and adjusted PEVM modeled emissions. The emissions reported by Partners for each year were
accepted as the quantity emitted from the share of the industry represented by those Partners. Remaining
emissions, those from non-Partners, were estimated using PEVM, with one change. To ensure time-series
consistency and to reflect the increasing use of remote clean technology (which increases the efficiency of the
production process while lowering emissions of fluorinated greenhouse gases), the average non-Partner emission
factor (PEVM emission factor) was assumed to begin declining gradually during this period. Specifically, the non-
Partner emission factor for each year was determined by linear interpolation, using the end points of 1999 (the
original PEVM emission factor) and 2011 (a new emission factor determined for the non-Partner population based
on GHGRP-reported data, described below).
The portion of the U.S. total emissions attributed to non-Partners is obtained by multiplying PEVM's total U.S.
Industrial Processes and Product Use 4-119

-------
emissions figure by the non-Partner share of U.S. total silicon capacity for each year as described above.89 Gas-
specific emissions from non-Partners were estimated using linear interpolation between the gas-specific emissions
distributions of 1999 (assumed to be the same as that of the total U.S. Industry in 1994) and 2011 (calculated from
a subset of non-Partners that reported through the GHGRP as a result of emitting more than 25,000 MT C02 Eq.
per year). Annual updates to PEVM reflect published figures for actual silicon consumption from VLSI Research,
Inc., revisions and additions to the world population of semiconductor manufacturing plants, and changes in IC
fabrication practices within the semiconductor industry (see ITRS 2008 and Semiconductor Equipment and
Materials Industry 2011).9a 91,92
For this time period emissions of other F-GHGs (C4F6, C5F8, HFC-32, HFC-41, HFC-134a) were estimated using the
method described above for 1990 to 1994.
Nitrous oxide emissions were estimated using the same methodology as the 1995 through 1999 methodology.
2007 through 2010
For the years 2007 through 2010, emissions were also estimated using a combination of Partner reported
emissions and adjusted PEVM modeled emissions to provide estimates for non-Partners; however, two
improvements were made to the estimation method employed for the previous years in the time series. First, the
2007 through 2010 emission estimates account for the fact that Partners and non-Partners employ different
distributions of manufacturing technologies, with the Partners using manufacturing technologies with greater
transistor densities and therefore greater numbers of layers.93 Second, the scope of the 2007 through 2010
estimates was expanded relative to the estimates for the years 2000 through 2006 to include emissions from
research and development (R&D) fabs. This additional enhancement was feasible through the use of more detailed
data published in the WFF. PEVM databases were updated annually as described above. The published world
average capacity utilization for 2007 through 2010 was used for production fabs, while for R&D fabs a 20 percent
89	This approach assumes that the distribution of linewidth technologies is the same between Partners and non-Partners. As
discussed in the description of the method used to estimate 2007 emissions, this is not always the case.
90	Special attention was given to the manufacturing capacity of plants that use wafers with 300 mm diameters because the
actual capacity of these plants is ramped up to design capacity, typically over a 2 to 3 year period. To prevent overstating
estimates of partner-capacity shares from plants using 300 mm wafers, design capacities contained in WFF were replaced with
estimates of actual installed capacities for 2004 published by Citigroup Smith Barney (2005). Without this correction, the
partner share of capacity would be overstated, by approximately 5 percent. For perspective, approximately 95 percent of all
new capacity additions in 2004 used 300 mm wafers, and by year-end those plants, on average, could operate at approximately
70 percent of the design capacity. For 2005, actual installed capacities were estimated using an entry in the World Fab Watch
database (April 2006 Edition) called "wafers/month, 8-inch equivalent," which denoted the actual installed capacity instead of
the fully-ramped capacity. For 2006, actual installed capacities of new fabs were estimated using an average monthly ramp rate
of 1100 wafer starts per month (wspm) derived from various sources such as semiconductor fabtech, industry analysts, and
articles in the trade press. The monthly ramp rate was applied from the first-quarter of silicon volume (FQSV) to determine the
average design capacity over the 2006 period.
91	In 2006, the industry trend in co-ownership of manufacturing facilities continued. Several manufacturers, who are Partners,
now operate fabs with other manufacturers, who in some cases are also Partners and in other cases are not Partners. Special
attention was given to this occurrence when estimating the Partner and non-Partner shares of U.S. manufacturing capacity.
92	Two versions of PEVM are used to model non-Partner emissions during this period. For the years 2000 to 2003 PEVM
v3.2.0506.0507 was used to estimate non-Partner emissions. During this time, discrete devices did not use PFCs during
manufacturing and therefore only memory and logic devices were modeled in the PEVM v3.2.0506.0507. From 2004 onwards,
discrete device fabrication started to use PFCs, hence PEVM v4.0.0701.0701, the first version of PEVM to account for PFC
emissions from discrete devices, was used to estimate non-Partner emissions for this time period.
93	EPA considered applying this change to years before 2007, but found that it would be difficult due to the large amount of
data (i.e., technology-specific global and non-Partner TMLA) that would have to be examined and manipulated for each year.
This effort did not appear to be justified given the relatively small impact of the improvement on the total estimate for 2007
and the fact that the impact of the improvement would likely be lower for earlier years because the estimated share of
emissions accounted for by non-Partners is growing as Partners continue to implement emission-reduction efforts.
4-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
figure was assumed (SIA 2009).
In addition, publicly-available utilization data was used to account for differences in fab utilization for
manufacturers of discrete and IC products for 2010 emissions for non-Partners. The Semiconductor Capacity
Utilization (SICAS) Reports from SIA provides the global semiconductor industry capacity and utilization,
differentiated by discrete and IC products (SIA 2009 through 2011). PEVM estimates were adjusted using
technology-weighted capacity shares that reflect the relative influence of different utilization. Gas-specific
emissions for non-Partners were estimated using the same method as for 2000 through 2006.
For this time period emissions of other F-GHGs (C5F8, CH2F2, CH3F, CH2FCF3, C2H2F4) were estimated using the
method described above for 1990 to 1994.
Nitrous oxide emissions were estimated using the same methodology as the 1995 through 1999 methodology.
2011 through 2012
The fifth method for estimating emissions from semiconductor manufacturing covers the period 2011 through
2012. This methodology differs from previous years because the EPA's Partnership with the semiconductor
industry ended (in 2010) and reporting under EPA's GHGRP began. Manufacturers whose estimated uncontrolled
emissions equal or exceed 25,000 MT C02 Eq. per year (based on default F-GHG-specific emission factors and total
capacity in terms of substrate area) are required to report their emissions to EPA. This population of reporters to
EPA's GHGRP included both historical Partners of EPA's PFC Reduction/Climate Partnership as well as non-Partners
some of which use gallium arsenide (GaAs) technology in addition to Si technology.94 Emissions from the
population of manufacturers that were below the reporting threshold were also estimated for this time period
using EPA-developed emission factors and estimates of facility-specific production obtained from WFF. Inventory
totals reflect the emissions from both reporting and non-reporting populations.
Under EPA's GHGRP, semiconductor manufacturing facilities report emissions of F-GHGs (for all types of F-GHGs)
used in etch and clean processes as well as emissions of fluorinated heat transfer fluids. (Fluorinated heat transfer
fluids are used to control process temperatures, thermally test devices, and clean substrate surfaces, among other
applications.) They also report N20 emissions from CVD and other processes. The F-GHGs and N20 were
aggregated, by gas, across all semiconductor manufacturing GHGRP reporters to calculate gas-specific emissions
for the GHGRP-reporting segment of the U.S. industry. At this time, emissions that result from heat transfer fluid
use that are HFC, PFC and SF6 are included in the total emission estimates from semiconductor manufacturing, and
these GHGRP-reported emissions have been compiled and presented in Table 4-94. F-HTF emissions resulting from
other types of gases (e.g., HFEs) are not presented in semiconductor manufacturing totals in Table 4-94 and Table
4-95 but are shown in Table 4-96 for informational purposes.
Changes to the default emission factors and default destruction or removal efficiencies (DREs) used for GHGRP
reporting affected the emissions trend between 2013 and 2014. These changes did not reflect actual emission rate
changes but data improvements. Therefore, for the current Inventory, EPA adjusted the time series of GHGRP-
reported data for 2011 through 2013 to ensure time-series consistency using a series of calculations that took into
account the characteristics of a facility (e.g., wafer size and abatement use). To adjust emissions for facilities that
did not report abatement in 2011 through 2013, EPA simply applied the revised emission factors to each facility's
estimated gas consumption by gas, process type and wafer size. In 2014, EPA also started collecting information on
fab-wide DREs and the gases abated by process type, which were used in calculations for adjusting emissions from
facilities that abated F-GHGs in 2011 through 2013.
• To adjust emissions for facilities that abated emissions in 2011 through 2013, EPA first calculated the
quantity of gas abated in 2014 using reported F-GHG emissions, the revised default DREs (or the
94 GaAs and Si technologies refer to the wafer on which devices are manufactured, which use the same PFCs but in different
ways.
Industrial Processes and Product Use 4-121

-------
estimated site-specific DRE,95 if a site-specific DRE was indicated), and the fab-wide DREs reported in
2014.96 To adjust emissions for facilities that abated emissions in 2011 through 2013, EPA first estimated
the percentage of gas passing through abatement systems for remote plasma clean in 2014 using the ratio
of emissions reported for CF4 and NF3.
•	EPA then estimated the quantity of NF3 abated for remote plasma clean in 2014 using the ratio of
emissions reported for CF4 (which is not abated) and NF3. This abated quantity was then subtracted from
the total abated quantity calculated as described in the bullet above.
•	To account for the resulting remaining abated quantity, EPA assumed that the percentage of gas passing
through abatement systems was the same across all remaining gas and process type combinations where
abatement was reported for 2014.
•	The percentage of gas abated was then assumed to be the same in 2011 through 2013 (if the facility
claimed abatement that year) as in 2014 for each gas abated in 2014.
The revised emission factors and DREs were then applied to the estimated gas consumption for each facility by gas,
process type and wafer size.97
For the segment of the semiconductor industry that is below EPA's GHGRP reporting threshold, and for R&D
facilities, which are not covered by EPA's GHGRP, emission estimates are based on EPA-developed emission factors
for the F-GHGs and N20 and estimates of manufacturing activity. The new emission factors (in units of mass of C02
Eq./TMLA [million square inches (MSI)]) are based on the emissions reported under EPA's GHGRP by facilities
without abatement and on the TMLA estimates for these facilities based on the WFF (SEMI 2012, 2013).98 In a
refinement of the method used to estimate emissions for the non-Partner population for prior years, different
emission factors were developed for different subpopulations of fabs, disaggregated by wafer size (200 mm and
300 mm). For each of these groups, a subpopulation-specific emission factor was obtained using a regression-
through-the-origin (RTO) model: facility-reported aggregate emissions of seven F-GHGs (CF4, C2F6, C3F8, c-C4F8,
CHF3, SF6 and NF3)99 were regressed against the corresponding TMLA to estimate an aggregate F-GHG emissions
factor (C02 Eq./MSI TMLA), and facility-reported N20 emissions were regressed against the corresponding TMLA to
estimate a N20 emissions factor (C02 Eq./MSI TMLA). For each subpopulation, the slope of the RTO model is the
emission factor for that subpopulation. Information on the use of point-of-use abatement by non-reporting fabs
was not available; thus, EPA conservatively assumed that non-reporting facilities did not use point-of-use
abatement.
For 2011 and 2012, estimates of TMLA relied on the capacity utilization of the fabs published by the U.S. Census
Bureau's Historical Data Quarterly Survey of Plant Capacity Utilization (USCB 2011, 2012). Similar to the
assumption for 2007 through 2010, facilities with only R&D activities were assumed to utilize only 20 percent of
their manufacturing capacity. All other facilities in the United States are assumed to utilize the average percent of
the manufacturing capacity without distinguishing whether fabs produce discrete products or logic products.
95	EPA generally assumed site-specific DREs were as follows: CF4, Etch (90 percent); all other gases, Etch (98 percent); NF3,
Clean (95 percent); CF4, Clean (80 percent), and all other gases, Clean (80 percent). There were a few exceptions where a higher
DRE was assumed to ensure the calculations operated correctly when there was 100 percent abatement.
96	If abatement information was not available for 2014 or the reported incorrectly in 2014, data from 2015 or 2016 was
substituted.
97	Since facilities did not report by fab before 2014, fab-wide DREs were averaged if a facility had more than one fab. For
facilities that reported more than one wafer size per facility, the percentages of a facility's emissions per wafer size were
estimated in 2014 and applied to earlier years, if possible. If the percentage of emissions per wafer size were unknown, a 50/50
split was used.
98	EPA does not have information on fab-wide DREs for this time period, so it is not possible to estimate uncontrolled emissions
from fabs that reported point-of-use abatement. These fabs were therefore excluded from the regression analysis. (They are
still included in the national totals.)
99	Only seven gases were aggregated because inclusion of F-GHGs that are not reported in the Inventory results in
overestimation of emission factor that is applied to the various non-reporting subpopulations.
4-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Non-reporting fabs were then broken out into subpopulations by wafer size (200 mm and 300 mm), using
information available through the WFF. The appropriate emission factor was applied to the total TMLA of each
subpopulation of non-reporting facilities to estimate the GWP-weighted emissions of that subpopulation.
Gas-specific, GWP-weighted emissions for each subpopulation of non-reporting facilities were estimated using the
corresponding reported distribution of gas-specific, GWP-weighted emissions from which the aggregate emission
factors, based on GHGRP-reported data, were developed. Estimated in this manner, the non-reporting population
accounted for 4.9 and 5.0 percent of U.S. emissions in 2011 and 2012, respectively. The GHGRP-reported emissions
and the calculated non-reporting population emissions are summed to estimate the total emissions from
semiconductor manufacturing.
2013 and 2014
For 2013 and 2014, as for 2011 and 2012, F-GHG and N20 emissions data received through EPA's GHGRP were
aggregated, by gas, across all semiconductor-manufacturing GHGRP reporters to calculate gas-specific emissions
for the GHGRP-reporting segment of the U.S. industry. However, for these years WFF data was not available.
Therefore, an updated methodology that does not depend on the WFF derived activity data was used to estimate
emissions for the segment of the industry that are not covered by EPA's GHGRP. For the facilities that did not
report to the GHGRP (i.e., which are below EPA's GHGRP reporting threshold or are R&D facilities), emissions were
estimated based on the proportion of total U.S. emissions attributed to non-reporters for 2011 and 2012. EPA used
a simple averaging method by first estimating this proportion for both F-GHGs and N20 for 2011, 2012, and 2015
through 2019, resulting in one set of proportions for F-GHGs and one set for N20, and then applied the average of
each set to the 2013 and 2014 GHGRP reported emissions to estimate the non-reporters' emissions. Fluorinated
gas-specific, GWP-weighted emissions for non-reporters were estimated using the corresponding reported
distribution of gas-specific, GWP-weighted emissions reported through EPA's GHGRP for 2013 and 2014.
GHGRP-reported emissions in 2013 were adjusted to capture changes to the default emission factors and default
destruction or removal efficiencies used for GHGRP reporting, affecting the emissions trend between 2013 and
2014. EPA used the same method to make these adjustments as described above for 2011 and 2012 GHGRP data.
2015 through 2019
Similar to the methods described above for 2011 and 2012, and 2013 and 2014, EPA relied upon emissions data
reported directly through the GHGRP. For 2015 through 2019, EPA took an approach similar to the one used for
2011 and 2012 to estimate emissions for the segment of the semiconductor industry that is below EPA's GHGRP
reporting threshold, and for R&D facilities, which are not covered by EPA's GHGRP. However, in a change from
previous years, EPA was able to develop new annual emission factors for 2015 through 2019 using TMLA from WFF
and a more comprehensive set of emissions, i.e., fabs with as well as without abatement control, as new
information about the use of abatement in GHGRP fabs and fab-wide were available. Fab-wide DREs represent
total fab C02 Eq.-weighted controlled F-GHG and N20 emissions (emissions after the use of abatement) divided by
total fab C02 Eq.-weighted uncontrolled F-GHG and N20 emissions (emission prior to the use of abatement).
Using information about reported emissions and the use of abatement and fab-wide DREs, EPA was able to
calculate uncontrolled emissions (each total F-GHG and N20) for every GHGRP reporting fab. Using this, coupled
with TMLA estimated using methods described above (see 2011 through 2012), EPA derived emission factors by
year, gas type (F-GHG or N20), and wafer size (200 mm and less or 300 mm) by dividing the total annual emissions
reported by GHGRP reporters by the total TMLA estimated for those reporters. These emission factors were
multiplied by estimates of non-reporter TMLA to arrive at estimates of total F-GHG and N20 emissions for non-
reporters for each year. For each wafer size, the total F-GHG emissions were disaggregated into individual gases
using the shares of total emissions represented by those gases in the emissions reported to the GHGRP by
unabated fabs producing that wafer size.
Industrial Processes and Product Use 4-123

-------
Data Sources
GHGRP reporters, which consist of former EPA Partners and non-Partners, estimated their emissions using a
default emission factor method established by EPA. Like the Tier 2c Method in the 2019 Refinement to the 2006
IPCC Guidelines, this method uses different emission and byproduct generation factors for different F-GHGs and
process types and uses factors for different wafer sizes (i.e., 300mm vs. 150 and 200mm) and CVD clean subtypes
(in situ thermal, in situ plasma, and remote plasma). Starting with 2014 reported emissions, EPA's GHGRP required
semiconductor manufacturers to apply updated emission factors to estimate their F-GHG emissions. For the years
2011	through 2013 reported emissions, semiconductor manufacturers used older emission factors to estimate
their F-GHG emissions (Federal Register / Vol. 75, No. 230 /December 1, 2010, 74829). Subpart I emission factors
were updated for 2014 by EPA as a result of a larger set of emission factor data becoming available as part of the
Subpart I petition process, which took place from 2011 through 2013.
Historically, partners estimated and reported their emissions using a range of methods and uneven
documentation. It is assumed that most Partners used a method at least as accurate as the IPCC's Tier 2a
Methodology, recommended in the 2006 IPCC Guidelines. Partners are estimated to have accounted for between
56 and 79 percent of F-GHG emissions from U.S. semiconductor manufacturing between 1995 and 2010, with the
percentage declining in recent years as Partners increasingly implemented abatement measures.
Estimates of operating plant capacities and characteristics for Partners and non-Partners were derived from the
Semiconductor Equipment and Materials Industry (SEMI) WFF (formerly World Fab Watch) database (1996 through
2012	and 2015) (e.g., Semiconductor Materials and Equipment Industry 2017). Actual worldwide capacity
utilizations for 2008 through 2010 were obtained from Semiconductor International Capacity Statistics (SICAS) (SIA
2009 through 2011). Estimates of the number of layers for each linewidth was obtained from International
Technology Roadmap for Semiconductors: 2013 Edition (Burton and Beizaie 2001; ITRS 2007; ITRS 2008; ITRS 2011;
ITRS 2013). PEVM utilized the WFF, SICAS, and ITRS, as well as historical silicon consumption estimates published
by VLSI. Actual quarterly U.S. capacity utilizations for 2011, 2012, 2015 and 2016 were obtained from the U.S.
Census Bureau's Historical Data Quarterly Survey of Plant Capacity Utilization (USCB 2011, 2012, 2015, and 2016).
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis of this source category was performed using the IPCC-recommended Approach
2 uncertainty estimation methodology, the Monte Carlo Stochastic Simulation technique. The equation used to
estimate uncertainty is:
Total Emissions (Et) = GHGRP Reported F-GHG Emissions (Er.f-ghg) + Non-Reporters' Estimated F-GHG
Emissions (Enr,f-ghg) + GHGRP Reported N2O Emissions (Er,n2o) + Non-Reporters' Estimated N2O Emissions
(Enr,N2o)
where ER and ENR denote totals for the indicated subcategories of emissions for F-GHG and N20, respectively.
The uncertainty in ET presented in Table 4-98 below results from the convolution of four distributions of emissions,
each reflecting separate estimates of possible values of Er,F-ghg, Er,N2o, Enr,f-ghg, and Enr,n2o- The approach and
methods for estimating each distribution and combining them to arrive at the reported 95 percent confidence
interval (CI) are described in the remainder of this section.
The uncertainty estimate of Er,F-ghg, or GHGRP-reported F-GHG emissions, is developed based on gas-specific
uncertainty estimates of emissions for two industry segments, one processing 200 mm wafers and one processing
300 mm wafers. Uncertainties in emissions for each gas and industry segment were developed during the
assessment of emission estimation methods for the Subpart I GHGRP rulemaking in 2012 (see Technical Support
for Modifications to the Fluorinated Greenhouse Gas Emission Estimation Method Option for Semiconductor
4-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Facilities under Subpart I, docket EPA-HQ-OAR-2011-0028).100 The assessment relied on facility-specific gas
information by gas and wafer size, and incorporated uncertainty associated with both emission factors and gas
consumption quantities. The 2012 analysis did not take into account the use of abatement. For the industry
segment that processed 200 mm wafers, estimates of uncertainties at a 95 percent CI ranged from ±29 percent for
C3F8 to ±10 percent for CF4. For the corresponding 300 mm industry segment, estimates of the 95 percent CI
ranged from ±36 percent for C4F8 to ±16 percent for CF4. These gas and wafer-specific uncertainty estimates are
applied to the total emissions across all the facilities that did not abate emissions as reported under EPA's GHGRP.
For those facilities reporting abatement of emissions under EPA's GHGRP, estimates of uncertainties for the no
abatement industry segments are modified to reflect the use of full abatement (abatement of all gases from all
cleaning and etching equipment) and partial abatement. These assumptions used to develop uncertainties for the
partial and full abatement facilities are identical for 200 mm and 300 mm wafer processing facilities. For all
facilities reporting gas abatement, a triangular distribution of destruction or removal efficiency is assumed for each
gas. The triangular distributions range from an asymmetric and highly uncertain distribution of zero percent
minimum to 90 percent maximum with 70 percent most likely value for CF4 to a symmetric and less uncertain
distribution of 85 percent minimum to 95 percent maximum with 90 percent most likely value for C4F8, NF3, and
SF6. For facilities reporting partial abatement, the distribution of fraction of the gas fed through the abatement
device, for each gas, is assumed to be triangularly distributed as well. It is assumed that no more than 50 percent
of the gases are abated (i.e., the maximum value) and that 50 percent is the most likely value and the minimum is
zero percent. Consideration of abatement then resulted in four additional industry segments, two 200-mm wafer-
processing segments (one fully and one partially abating each gas) and two 300-mm wafer-processing segment
(one fully and the other partially abating each gas). Gas-specific emission uncertainties were estimated by
convolving the distributions of unabated emissions with the appropriate distribution of abatement efficiency for
fully and partially abated facilities using a Monte Carlo simulation.
The uncertainty in ERiF.Ghg is obtained by allocating the estimates of uncertainties to the total GHGRP-reported
emissions from each of the six industry segments, and then running a Monte Carlo simulation which results in the
95 percent CI for emissions from GHGRP-reporting facilities (ER;F-ghg)-
The uncertainty in ER,N2o is obtained by assuming that the uncertainty in the emissions reported by each of the
GHGRP reporting facilities results from the uncertainty in quantity of N20 consumed and the N20 emission factor
(or utilization). Similar to analyses completed for Subpart I (see Technical Support for Modifications to the
Fluorinated Greenhouse Gas Emission Estimation Method Option for Semiconductor Facilities under Subpart I,
docket EPA-HQ-OAR-2011-0028), the uncertainty of N20 consumed was assumed to be 20 percent. Consumption
of N20 for GHGRP reporting facilities was estimated by back-calculating from emissions reported and assuming no
abatement. The quantity of N20 utilized (the complement of the emission factor) was assumed to have a triangular
distribution with a minimum value of zero percent, mode of 20 percent and maximum value of 84 percent. The
minimum was selected based on physical limitations, the mode was set equivalent to the Subpart I default N20
utilization rate for chemical vapor deposition, and the maximum was set equal to the maximum utilization rate
found in ISMI Analysis of Nitrous Oxide Survey Data (ISMI 2009). The inputs were used to simulate emissions for
100 On November 13, 2013, EPA published a final rule revising Subpart I (Electronics Manufacturing) of the GHGRP (78 FR
68162). The revised rule includes updated default emission factors and updated default destruction and removal efficiencies
that are slightly different from those that semiconductor manufacturers were required to use to report their 2012 emissions.
The uncertainty analyses that were performed during the development of the revised rule focused on these updated defaults,
but are expected to be reasonably representative of the uncertainties associated with the older defaults, particularly for
estimates at the country level. (They may somewhat underestimate the uncertainties associated with the older defaults at the
facility level.) For simplicity, the 2012 estimates are assumed to be unbiased although in some cases, the updated (and
therefore more representative) defaults are higher or lower than the older defaults. Multiple models and sensitivity scenarios
were run for the Subpart I analysis. The uncertainty analysis presented here made use of the Input gas and wafer size model
(Model 1) under the following conditions: Year = 2010, f = 20, n = SIA3.
Industrial Processes and Product Use 4-125

-------
each of the GHGRP reporting, N20-emitting facilities. The uncertainty for the total reported N20 emissions was
then estimated by combining the uncertainties of each facilities' reported emissions using Monte Carlo simulation.
The estimate of uncertainty in ENr, f-ghg and ENr, N20 entailed developing estimates of uncertainties for the emissions
factors and the corresponding estimates of TMLA.
The uncertainty in TMLA depends on the uncertainty of two variables—an estimate of the uncertainty in the
average annual capacity utilization for each level of production of fabs (e.g., full scale or R&D production) and a
corresponding estimate of the uncertainty in the number of layers manufactured. For both variables, the
distributions of capacity utilizations and number of manufactured layers are assumed triangular for all categories
of non-reporting fabs. The most probable utilization is assumed to be 82 percent, with the highest and lowest
utilization assumed to be 89 percent, and 70 percent, respectively. For the triangular distributions that govern the
number of possible layers manufactured, it is assumed the most probable value is one layer less than reported in
the ITRS; the smallest number varied by technology generation between one and two layers less than given in the
ITRS and largest number of layers corresponded to the figure given in the ITRS.
The uncertainty bounds for the average capacity utilization and the number of layers manufactured are used as
inputs in a separate Monte Carlo simulation to estimate the uncertainty around the TMLA of both individual
facilities as well as the total non-reporting TMLA of each sub-population.
The uncertainty around the emission factors for non-reporting facilities is dependent on the uncertainty of the
total emissions (MMT C02 Eq. units) and the TMLA of each reporting facility in that category. For each wafer size
for reporting facilities, total emissions were regressed on TMLA (with an intercept forced to zero) for 10,000
emission and 10,000 TMLA values in a Monte Carlo simulation, which results in 10,000 total regression coefficients
(emission factors). The 2.5th and the 97.5th percentile of these emission factors are determined, and the bounds
are assigned as the percent difference from the estimated emission factor.
The final step in estimating the uncertainty in emissions of reporting and non-reporting facilities is convolving the
distribution of reported emissions, emission factors, and TMLA using Monte Carlo simulation. For this final Monte
Carlo simulation, the distributions of the reported F-GHG gas- and wafer size-specific emissions are assumed to be
normally distributed and the uncertainty bounds are assigned at 1.96 standard deviations around the estimated
mean. The were some instances, though, where departures from normality were observed for variables, including
for the distributions of the gas- and wafer size-specific N20 emissions, TMLA, and non-reporter emission factors,
both for F-GHGs and N20. As a result, the distributions for these parameters were assumed to follow a pert beta
distribution.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-98, which is also
obtained by convolving—using Monte Carlo simulation—the distributions of emissions for each reporting and non-
reporting facility. The emissions estimate for total U.S. F-GHG and N20 emissions from semiconductor
manufacturing were estimated to be between 4.3 and 4.8 MMT C02 Eq. at a 95 percent confidence level. This
range represents 6 percent below to 6 percent above the 2019 emission estimate of 4.6 MMT C02 Eq. for
semiconductor emissions for the main seven gases. This range and the associated percentages apply to the
estimate of total emissions rather than those of individual gases. Uncertainties associated with individual gases will
be somewhat higher than the aggregate, but were not explicitly modeled.
4-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-98: Approach 2 Quantitative Uncertainty Estimates for HFC, PFC, SFe, NF3 and N2O
Emissions from Semiconductor Manufacture (MMT CO2 Eq. and Percent)3
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimateb


(MMT CO? Eq.)
(MMT CO? Eq.)
(%)




Lower
Upper
Lower
Upper



Boundc
Boundc
Bound
Bound
Semiconductor
Manufacture
HFC, PFC, SF6,
NF3, and N20
4.6
4.3
4.8
-6%
6%
a This uncertainty analysis does not include quantification of the uncertainty of emissions from other F-GHGs for
semiconductors, heat transfer fluids, PV, and MEMS.
b Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
c Absolute lower and upper bounds were calculated using the corresponding lower and upper bounds in percentages.
It should be noted that the uncertainty analysis for this source category does not quantify the uncertainty of HFC,
PFC, and SF6 emissions from the use of heat transfer fluids or the other F-GHGs. While these emissions are
included in the semiconductor manufacturing F-GHG total emissions, they make up a small portion of total
emissions from the source category (less than 1 percent). Any uncertainty of these emissions would have minimal
impact on the overall uncertainty estimates, and therefore the uncertainties associated for HTF HFC, PFC, and SF6
emissions was not included in this analysis for this Inventory year.
Similarly, the uncertainty was not quantified for emissions from the manufacturing of photovoltaics and micro-
electro-mechanical devices. These emissions make up a small portion of total emissions from the source category.
Any uncertainty of these emissions would have minimal impact on the overall uncertainty estimates, and therefore
associated uncertainties were not included.
In an effort to improve the uncertainty analysis for this source category other F-GHGs from semiconductor
manufacturing, HFC, PFC, and SF6 emissions from the use of heat transfer fluids and manufacturing of PVs and
MEMS may be added in future inventory years (see Planned Improvements section below). The emissions reported
under EPA's GHGRP for 2014, 2015, 2016, 2017, 2018, and 2019, which are included in the overall emissions
estimates, were based on an updated set of default emission factors.
QA/QC and Verification
For its GHGRP data, EPA verifies annual facility-level reports through a multi-step process (e.g., including a
combination of pre-and post-submittal electronic checks and manual reviews by staff) to identify potential errors
and ensure that data submitted to EPA are accurate, complete, and consistent (EPA 2015).101 Based on the results
of the verification process, EPA follows up with facilities to resolve mistakes that may have occurred. The post-
submittals checks are consistent with a number of general and category-specific QC procedures including range
checks, statistical checks, algorithm checks, and year-to-year checks of reported data and emissions.
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter and Annex 8 for more details.
Recalculations Discussion
Emissions from 2011 through 2019 were updated to reflect updated emissions reporting in EPA's GHGRP, relative
to the previous Inventory. Additionally, EPA made the following changes:
101 GHGRP Report Verification Factsheet. .
Industrial Processes and Product Use 4-127

-------
•	Previously, capacity utilization numbers used to develop TMLA estimates for 2015 through 2018 were
based on utilizations for quarter four of a given year reported to the U.S. Census Bureau. EPA adjusted
this approach to take an average utilization as reported across all four quarters of a reporting year. This
minimally affected the emission estimates for non-reporters for F-GHGs and N20.
•	To estimate non-reporter F-GHG and N20 emissions, EPA relies on data reported through Subpart I and
the World Fab Forecast. This process requires EPA to map facilities that report through Subpart I and
which are also represented in the World Fab Forecast. For this inventory update, EPA identified and made
corrections to a few instances of this mapping based on new information and additional reviews of the
data. This had minimal effects on emission estimates.
•	As discussed in the Methodology section, emission estimates for 2011 and 2012 were recently updated to
take into account the revised emission factors used in 2014 and later years. For this inventory update,
EPA identified and made corrections to a few instances where the adjusted emissions used in the
regression for developing the non-reporter emission factors were recorded incorrectly.
•	EPA updated its approach for adjusting the non-Partner emission factors applied to the 1999-2010 time
series to ensure time series consistency. This adjustment is based on a linear interpolation between a
PEVM-calculated emission factor from 1999 and an emission factor calculated using 2011 GHGRP-
reported data and World Fab Forecast data. EPA updated the 2011 emission factor used to make this
adjustment in order to reflect the adjustments made to the 2011 GHGRP reported data in the previous
inventory cycle and to ensure time series consistency for data reported through Subpart I. EPA is
continuing to evaluate this approach and may make additional improvements per the discussion below.
•	Previously, the capacity used to develop emissions estimates from photovoltaics prior to 2012 was
estimated using a linear extrapolation between 0 in 1997 and the known U.S. production capacity in
2012. EPA identified a new source of capacity data for 2000 through 2009 and used this data to adjust the
estimated emissions between 1997 and 2012. A linear interpolation is now used only for the years
between 1997 and 2000, and again between 2009 and 2012.
•	For the Monte Carlo simulations the bounds of TMLA, non-reporter emission factors, and reported N20
emissions are assumed to follow a pert beta distribution given their asymmetry. Previously, these bounds
were assumed to be normally distributed.
Planned Improvements
The Inventory methodology uses data reported through the EPA Partnership (for earlier years) and EPA's GHGRP
(for later years) to extrapolate the emissions of the non-reporting population. While these techniques are well
developed, the understanding of the relationship between the reporting and non-reporting populations is limited.
Further analysis of the reporting and non-reporting populations could aid in the accuracy of the non-reporting
population extrapolation in future years. In addition, the accuracy of the emissions estimates for the non-reporting
population could be further increased through EPA's further investigation of and improvement upon the accuracy
of estimated activity in the form of TMLA.
The Inventory uses utilization from two different sources for various time periods-SEMI to develop PEVM and to
estimate non-Partner emissions for the period 1995 to 2010 and U.S. Census Bureau for 2011 through 2014. SEMI
reported global capacity utilization for manufacturers through 2011. U.S. Census Bureau capacity utilization
include U.S. semiconductor manufacturers as well as assemblers. Further analysis on the impacts of using a new
and different source of utilization data could prove to be useful in better understanding of industry trends and
impacts of utilization data sources on historical emission estimates.
The current Inventory now includes HFC, PFC, and SF6 emissions resulting from the use of heat transfer fluids in the
total estimates of F-GHG emissions from semiconductor manufacturing. A point of consideration for future
Inventory reports is the inclusion of the uncertainty surrounding these emissions in the source category
uncertainty analysis (see also Uncertainty and Time-Series Consistency section).
Estimates of non-reporter and non-Partner emissions are based on EPA-developed emission factors for the time
periods pre-2010, 2011 through 2012, and 2015 through 2019. Based on the data available for these time periods,
the methods used to develop emission factors for non-reporters and non-Partners are slightly inconsistent (e.g.,
4-128 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
how data representing emissions and TMLAfrom the manufacture of various wafer sizes are aggregated or
disaggregated for purposes of calculating emission factors). Further analyses to support potentially adjusting the
methods for developing these emission factors could be done to better ensure consistency across the time series.
4.24 Substitution of Ozone Depleting
Substances (CRF Source Category 2F)
Hydrofluorocarbons (HFCs) and perfluorocarbons (PFCs) are used as alternatives to several classes of ozone-
depleting substances (ODSs) that are being phased out under the terms of the Montreal Protocol and the Clean Air
Act Amendments of 1990.102 Ozone depleting substances—chlorofluorocarbons (CFCs), halons, carbon
tetrachloride, methyl chloroform, and hydrochlorofluorocarbons (HCFCs)—are used in a variety of industrial
applications including refrigeration and air conditioning equipment, solvent cleaning, foam production,
sterilization, fire extinguishing, and aerosols. Although HFCs and PFCs are not harmful to the stratospheric ozone
layer, they are potent greenhouse gases. Emission estimates for HFCs and PFCs used as substitutes for ODSs are
provided in Table 4-99 and Table 4-100.103
Table 4-99: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.)
Gas
1990
2005
2015
2016
2017
2018
2019
HFC-23
0.0
+
+
+
+
+
+
HFC-32
0.0
0.3
3.9
4.6
5.3
6.0
6.8
HFC-125
+
9.0
43.4
46.9
50.1
53.8
58.5
HFC-134a
+
80.2
73.2
68.8
64.1
61.1
59.8
HFC-143a
+
9.4
27.6
28.2
28.0
27.7
27.8
HFC-236fa
0.0
1.2
1.3
1.3
1.2
1.2
1.1
cf4
0.0
+
+
+
+
0.1
0.1
Others3
0.2
7.3
14.0
15.0
15.9
16.3
16.4
Total
0.2
107.3
163.6
164.9
164.7
166.1
170.6
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Others represent an unspecified mix of HFCs and PFCs, which includes HFC-152a, HFC-227ea, HFC-
245fa, HFC-365mfc, HFC-43-10mee, HCFO-1233zd(E), HFO-1234yf, HFO-1234ze(E), HFO-1336mzz(Z),
C4F10, and PFC/PFPEs, the latter being a proxy for a diverse collection of PFCs and perfluoropolyethers
(PFPEs) employed for solvent applications. For estimating purposes, the GWP value used for PFC/PFPEs
was based upon C6Fi4.
102	[42 U.S.C § 7671, CAA Title VI],
103	Emissions of ODS are not included here consistent with UNFCCC reporting guidelines for national inventories noted in Box
4-1. See Annex 6.2 for more details on emissions of ODS.
Industrial Processes and Product Use 4-129

-------
Table 4-100: Emissions of HFCs and PFCs from ODS Substitution (Metric Tons)
Gas
1990
2005
2015
2016
2017
2018
2019
HFC-23
0
1
2
2
2
2
2
HFC-32
0
397
5,843
6,801
7,842
8,948
10,094
HFC-125
+
2,580
12,401
13,413
14,327
15,362
16,720
HFC-134a
+
56,052
51,213
48,126
44,797
42,694
41,814
HFC-143a
+
2,093
6,178
6,320
6,264
6,188
6,230
HFC-236fa
0
118
134
129
124
118
112
cf4
0
2
6
6
6
7
7
Others3
M
M
M
M
M
M
M
+ Does not exceed 0.5 MT.
M (Mixture of Gases).
a Others represent an unspecified mix of HFCs and PFCs, which includes HFC-152a, HFC-227ea, HFC-245fa,
HFC-365mfc, HFC-43-10mee, HCFO-1233zd(E), HFO-1234yf, HFO-1234ze(E), HFO-1336mzz(Z), C4Fi0, and
PFC/PFPEs, the latter being a proxy for a diverse collection of PFCs and perfluoropolyethers (PFPEs)
employed for solvent applications.
In 1990 and 1991, the only significant emissions of HFCs and PFCs as substitutes to ODSs were relatively small
amounts of HFC-152a—used as an aerosol propellant and also a component of the refrigerant blend R-500 used in
chillers. Beginning in 1992, HFC-134a was used in growing amounts as a refrigerant in motor vehicle air-
conditioners and in refrigerant blends such as R-404A.104 In 1993, the use of HFCs in foam production began, and
in 1994 ODS substitutes for halons entered widespread use in the United States as halon production was phased
out. In 1995, these compounds also found applications as solvents.
The use and subsequent emissions of HFCs and PFCs as ODS substitutes has been increasing from small amounts in
1990 to 170.6 MMT C02 Eq. emitted in 2019. This increase was in large part the result of efforts to phase out CFCs
and other ODSs in the United States. In the short term, this trend is expected to continue, and will likely continue
over the next decade as HCFCs, which are interim substitutes in many applications, are themselves phased-out
under the provisions of the Copenhagen Amendments to the Montreal Protocol. Improvements in the technologies
associated with the use of these gases and the introduction of alternative gases and technologies, however, may
help to offset this anticipated increase in emissions.
Table 4-101 presents emissions of HFCs and PFCs as ODS substitutes by end-use sector for 1990 through 2019. The
end-use sectors that contributed the most toward emissions of HFCs and PFCs as ODS substitutes in 2019 include
refrigeration and air-conditioning (133.4 MMT C02 Eq., or approximately 78 percent), aerosols (16.3 MMT C02 Eq.,
or approximately 10 percent), and foams (16.1 MMT C02 Eq., or approximately 9 percent). Within the refrigeration
and air-conditioning end-use sector, large retail food was the highest emitting end-use (33.6 MMT C02 Eq.),
followed by residential unitary air conditioning. Each of the end-use sectors is described in more detail below.
Table 4-101: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.) by Sector
Sector
1990
2005
2015
2016
2017
2018
2019
Refrigeration/Air Conditioning
+
89.7
124.8
126.5
126.9
129.4
133.4
Aerosols
0.2
10.7
20.8
19.2
17.6
16.0
16.3
Foams
+
4.1
13.9
15.0
15.8
16.2
16.1
Solvents
+
1.7
1.8
1.9
1.9
2.0
2.0
Fire Protection
+
1.1
2.3
2.4
2.5
2.6
2.8
Total
0.2
107.3
163.6
164.9
164.7
166.1
170.6
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
104 R-4Q4A contains HFC-125, HFC-143a, and HFC-134a.
4-130 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Refrigeration/Air Conditioning
The refrigeration and air-conditioning sector includes a wide variety of equipment types that have historically used
CFCs or HCFCs. End-uses within this sector include motor vehicle air-conditioning, retail food refrigeration,
refrigerated transport (e.g., ship holds, truck trailers, railway freight cars), household refrigeration, residential and
small commercial air-conditioning and heat pumps, chillers (large comfort cooling), cold storage facilities, and
industrial process refrigeration (e.g., systems used in food processing, chemical, petrochemical, pharmaceutical, oil
and gas, and metallurgical industries). As the ODS phaseout has taken effect, most equipment has been retrofitted
or replaced to use HFC-based substitutes. Common HFCs in use today in refrigeration/air-conditioning equipment
are HFC-134a, R-410A,105 R-404A, and R-507A.106 Lower-GWP options such as hydrofluoroolefin (HFO)-1234yf in
motor vehicle air-conditioning, R-717 (ammonia) in cold storage and industrial applications, and R-744 (carbon
dioxide) and HFC/HFO blends in retail food refrigeration, are also being used. These refrigerants are emitted to the
atmosphere during equipment operation (as a result of component failure, leaks, and purges), as well as at
manufacturing (if charged at the factory), installation, servicing, and disposal events.
Aerosols
Aerosol propellants are used in metered dose inhalers (MDIs) and a variety of personal care products and
technical/specialty products (e.g., duster sprays and safety horns). Pharmaceutical companies that produce MDIs—
a type of inhaled therapy used to treat asthma and chronic obstructive pulmonary disease—have replaced the use
of CFCs with HFC-propellant alternatives. The earliest ozone-friendly MDIs were produced with HFC-134a, but the
industry is using HFC-227ea as well. Conversely, since the use of CFC propellants was banned in 1978, most non-
medical consumer aerosol products have not transitioned to HFCs, but to "not-in-kind" technologies, such as solid
or roll-on deodorants and finger-pump sprays. The transition away from ODS in specialty aerosol products has also
led to the introduction of non-fluorocarbon alternatives (e.g., hydrocarbon propellants) in certain applications, in
addition to HFC-134a or HFC-152a. Other low-GWP options such as HFO-1234ze(E) are being used as well. These
propellants are released into the atmosphere as the aerosol products are used.
Foams
Chlorofluorocarbons and HCFCs have traditionally been used as foam blowing agents to produce polyurethane
(PU), polystyrene, polyolefin, and phenolic foams, which are used in a wide variety of products and applications.
Since the Montreal Protocol, flexible PU foams as well as other types of foam, such as polystyrene sheet,
polyolefin, and phenolic foam, have transitioned almost completely away from fluorocompounds, into alternatives
such as C02 and hydrocarbons. The majority of rigid PU foams have transitioned to HFCs—primarily HFC-134a and
HFC-245fa. Today, these HFCs are used to produce PU appliance, PU commercial refrigeration, PU spray, and PU
panel foams—used in refrigerators, vending machines, roofing, wall insulation, garage doors, and cold storage
applications. In addition, HFC-152a, HFC-134a, and C02 are used to produce polystyrene sheet/board foam, which
is used in food packaging and building insulation. Low-GWP fluorinated foam blowing agents in use include HFO-
1234ze(E) and HCFO-1233zd(E). Emissions of blowing agents occur when the foam is manufactured as well as
during the foam lifetime and at foam disposal, depending on the particular foam type.
Solvents
Chlorofluorocarbons, methyl chloroform (1,1,1-trichloroethane or TCA), and to a lesser extent carbon tetrachloride
(CCI4) were historically used as solvents in a wide range of cleaning applications, including precision, electronics,
and metal cleaning. Since their phaseout, metal cleaning end-use applications have primarily transitioned to non-
fluorocarbon solvents and not-in-kind processes. The precision and electronics cleaning end-uses have transitioned
in part to high-GWP gases, due to their high reliability, excellent compatibility, good stability, low toxicity, and
105	R-410A contains HFC-32 and HFC-125.
106	R-507A, also called R-507, contains HFC-125 and HFC-143a.
Industrial Processes and Product Use 4-131

-------
selective solvency. These applications rely on HFC-43-10mee, HFC-365mfc, HFC-245fa, and to a lesser extent, PFCs.
Electronics cleaning involves removing flux residue that remains after a soldering operation for printed circuit
boards and other contamination-sensitive electronics applications. Precision cleaning may apply to either
electronic components or to metal surfaces, and is characterized by products, such as disk drives, gyroscopes, and
optical components, that require a high level of cleanliness and generally have complex shapes, small clearances,
and other cleaning challenges. The use of solvents yields fugitive emissions of these HFCs and PFCs.
Fire Protection
Fire protection applications include portable fire extinguishers ("streaming" applications) that originally used halon
1211, and total flooding applications that originally used halon 1301, as well as some halon 2402. Since the
production and import of virgin halons were banned in the United States in 1994, the halon replacement agent of
choice in the streaming sector has been dry chemical, although HFC-236fa is also used to a limited extent. In the
total flooding sector, HFC-227ea has emerged as the primary replacement for halon 1301 in applications that
require clean agents. Other HFCs, such as HFC-23 and HFC-125, are used in smaller amounts. The majority of HFC-
227ea in total flooding systems is used to protect essential electronics, as well as in civil aviation, military mobile
weapons systems, oil/gas/other process industries, and merchant shipping. Fluoroketone FK-5-1-12 is also used as
a low-GWP option and 2-BTP is being considered. As fire protection equipment is tested or deployed, emissions of
HFCs occur.
Methodology
A detailed Vintaging Model of ODS-containing equipment and products was used to estimate the actual—versus
potential—emissions of various ODS substitutes, including HFCs and PFCs. The name of the model refers to the fact
that it tracks the use and emissions of various compounds for the annual "vintages" of new equipment that enter
service in each end-use. The Vintaging Model predicts ODS and ODS substitute use in the United States based on
modeled estimates of the quantity of equipment or products sold each year containing these chemicals and the
amount of the chemical required to manufacture and/or maintain equipment and products over time. Emissions
for each end-use were estimated by applying annual leak rates and release profiles, which account for the lag in
emissions from equipment as they leak over time. By aggregating the data for 78 different end-uses, the model
produces estimates of annual use and emissions of each compound. Further information on the Vintaging Model is
contained in Annex 3.9.
Uncertainty and Time-Series Consistency
Given that emissions of ODS substitutes occur from thousands of different kinds of equipment and from millions of
point and mobile sources throughout the United States, emission estimates must be made using analytical tools
such as the Vintaging Model or the methods outlined in IPCC (2006). Though the model is more comprehensive
than the IPCC default methodology, significant uncertainties still exist with regard to the levels of equipment sales,
equipment characteristics, and end-use emissions profiles that were used to estimate annual emissions for the
various compounds.
The uncertainty analysis quantifies the level of uncertainty associated with the aggregate emissions across the 78
end-uses in the Vintaging Model. In order to calculate uncertainty, functional forms were developed to simplify
some of the complex "vintaging" aspects of some end-use sectors, especially with respect to refrigeration and air-
conditioning, and to a lesser degree, fire extinguishing. These sectors calculate emissions based on the entire
lifetime of equipment, not just equipment put into commission in the current year, thereby necessitating
simplifying equations. The functional forms used variables that included growth rates, emission factors, transition
from ODSs, change in charge size as a result of the transition, disposal quantities, disposal emission rates, and
either stock for the current year or original ODS consumption. Uncertainty was estimated around each variable
within the functional forms based on expert judgment, and a Monte Carlo analysis was performed.
4-132 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The most significant sources of uncertainty for this source category include the charge size for technical aerosols
using HFC-134a, as well as total stock of refrigerant installed in industrial process refrigeration and cold storage
equipment. For this Inventory year, uncertainty was defined for charge sizes of consumer and technical aerosols,
which were assumed to be constant in the uncertainty analysis for previous Inventory reports. The updates to the
uncertainty analysis for the aerosols sector resulted in increased overall uncertainty for this source category;
however, the results reflect a more robust uncertainty analysis for the consumer and technical aerosol end-uses.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-102. Substitution of
ozone depleting substances HFC and PFC emissions were estimated to be between 164.3 and 192.9 MMT C02 Eq.
at the 95 percent confidence level. This indicates a range of approximately 3.7 percent below to 13.1 percent
above the emission estimate of 170.6 MMT C02 Eq.
Table 4-102: Approach 2 Quantitative Uncertainty Estimates for HFC and PFC Emissions
from ODS Substitutes (MMT CO2 Eq. and Percent)
Source
Gases
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3


(MMTCOz Eq.)
(MMTCOz Eq.)

(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Substitution of Ozone
HFCs and





Depleting Substances

170.6
164.3
192.9
-3.7%
+13.1%
PFCs
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter. Category specific QC findings are described below.
Comparison of Reported Consumption to Modeled Consumption of HFCs
Data from EPA's Greenhouse Gas Reporting Program (GHGRP)107 was also used to perform quality control as a
reference scenario check on the modeled emissions from this source category as specified in 2006 IPCC Guidelines
for National Greenhouse Gas Inventories. To do so, consumption patterns demonstrated through data reported
under GHGRP Subpart OO—Suppliers of Industrial Greenhouse Gases and Subpart QQ—Importers and Exporters of
Fluorinated Greenhouse Gases Contained in Pre-Charged Equipment or Closed-Cell Foams were compared to the
modeled demand for new saturated HFCs (excluding HFC-23) used as ODS substitutes from the Vintaging Model.
The collection of data from suppliers of HFCs enables EPA to calculate the reporters' aggregated net supply-the
sum of the quantities of chemical produced or imported into the United States less the sum of the quantities of
chemical transformed (used as a feedstock in the production of other chemicals), destroyed, or exported from the
107 For the GHGRP data, EPA verifies annual facility-level and company-level reports through a multi-step process (e.g.,
including a combination of pre-and post-submittal electronic checks and manual reviews by staff) to identify potential errors
and ensure that data submitted to EPA are accurate, complete, and consistent (EPA (2015). Based on the results of the
verification process, EPA follows up with facilities to resolve mistakes that may have occurred. The post-submittals checks are
consistent with a number of general and category-specific QC procedures, including range checks, statistical checks, algorithm
checks, and year-to-year checks of reported data.
Industrial Processes and Product Use 4-133

-------
United States.108 This allows for a quality control check on emissions from this source because the Vintaging Model
uses modeled demand for new chemical as a proxy for total amount supplied, which is similar to net supply, as an
input to the emission calculations in the model.
The QA/QC and verification process for individual gases and sources in the Vintaging Model includes regular review
against up-to-date market information, including equipment stock estimates, leak rates, and sector transitions. In
addition, comparisons against published emission and consumption sources by gas and by source are performed
when available as described further below. Independent peer reviews of the Vintaging Model are periodically
performed, including one conducted in 2017 (EPA 2018), to confirm Vintaging Model estimates and identify
updates. The HFCs and PFCs within the unspecified mix of HFCs and PFCs are modelled and verified individually in
the same process as all other gases and sources in the Vintaging Model. The HFCs and PFCs are grouped in the
unspecified mix of HFCs and PFCs category only for the purposes of reporting emissions to protect Confidential
Business Information (CBI).
Reported Net Supply (GHGRP Top-Down Estimate)
Under EPA's GHGRP, suppliers (i.e., producers, importers, and exporters) of HFCs under Subpart 00 began
annually reporting their production, transformation, destruction, imports, and exports to EPA in 2011 (for supply
that occurred in 2010) and suppliers of HFCs under Subpart QQ began annually reporting their imports and exports
to EPA in 2012 (for supply that occurred in 2011). Beginning in 2015, bulk consumption data for aggregated HFCs
reported under Subpart 00 were made publicly available under EPA's GHGRP. Data include all saturated HFCs
(except HFC-23) reported to EPA across the GHGRP-reporting time series. The data include all 26 such saturated
HFCs listed in Table A-l of 40 CFR Part 98, where regulations for EPA's GHGRP are promulgated, though not all
species were reported in each reporting year. For the first time in 2016, net imports of HFCs contained in pre-
charged equipment or closed-cell foams reported under Subpart QQ were made publicly available under EPA's
GHGRP.
Modeled Consumption (Vintaging Model Bottom-Up Estimate)
The Vintaging Model, used to estimate emissions from this source category, calculates chemical demand based on
the quantity of equipment and products sold, serviced and retired each year, and the amount of the chemical
required to manufacture and/or maintain the equipment and products.109 It is assumed that the total demand
equals the amount supplied by either new production, chemical import, or quantities recovered (often reclaimed)
and placed back on the market. In the Vintaging Model, demand for new chemical, as a proxy for consumption, is
calculated as any chemical demand (either for new equipment or for servicing existing equipment) that cannot be
met through recycled or recovered material. No distinction is made in the Vintaging Model between whether that
need is met through domestic production or imports. To calculate emissions, the Vintaging Model estimates the
quantity released from equipment over time. Thus, verifying the Vintaging Model's calculated consumption against
GHGRP reported data is one way to check the Vintaging Model's emission estimates.
There are eleven saturated HFC species modeled in the Vintaging Model: HFC-23, HFC-32, HFC-125, HFC-134a,
HFC-143a, HFC-152a, HFC-227ea, HFC-236fa, HFC-245fa, HFC-365mfc, and HFC-43-10mee. For the purposes of this
comparison, only ten HFC species are included (HFC-23 is excluded), to more closely align with the aggregated total
reported under EPA's GHGRP. While some amounts of less-used saturated HFCs, including isomers of those
included in the Vintaging Model, are reportable under EPA's GHGRP, the data are believed to represent an amount
comparable to the modeled estimates as a quality control check.
108	Chemical that is exported, transformed, or destroyed—unless otherwise imported back to the United States—will never be
emitted in the United States.
109	The model builds an inventory of the in-use stock of equipment and products and ODSs and HFCs in each of the sub-
applications. Emissions are subsequently estimated by applying annual and disposal emission rates to each population of
equipment and products.
4-134 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Comparison Results and Discussion
Comparing the estimates of consumption from these two approaches (i.e., reported and modeled) ultimately
supports and improves estimates of emissions, as noted in the 2006IPCC Guidelines (which refer to fluorinated
greenhouse gas consumption based on supplies as "potential emissions"):
[W]hen considered along with estimates of actual emissions, the potential emissions approach can assist
in validation of completeness of sources covered and as a QC check by comparing total domestic
consumption as calculated in this 'potential emissions approach' per compound with the sum of all
activity data of the various uses (IPCC 2006).
Table 4-103 and Figure 4-3 compare the published net supply of saturated HFCs (excluding HFC-23) in MMT C02
Eq. as determined from Subpart 00 (supply of HFCs in bulk) and Subpart QQ (supply of HFCs in products and
foams) of EPA's GHGRP for the years 2010 through 2019 (U.S. EPA 2021a) and the chemical demand as calculated
by the Vintaging Model for the same time series.
Table 4-103: U.S. HFC Supply (MMT COz Eq.)

2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Reported Net Supply (GHGRP)
235
248
245
295
279
290
268
317
338
324
Industrial GHG Suppliers
235
241
227
278
254
264
240
285
304
294
HFCs in Products and Foams3
NA
7
18
17
25
26
28
32
34
30
Modeled Supply (Vintaging Model)
265
271
275
281
287
285
288
278
283
278
Percent Difference
13%
9%
12%
-5%
3%
-2%
7%
-12%
-16%
-14%
NA (Not Available)
a Importers and exporters of fluorinated gases in products were not required to report 2010 data.
Figure 4-3: U.S. HFC Consumption (MMT CO2 Eq.)
¦	Reported Importsin Productsand Foams
¦	Modeled Consumption
¦	Reported Bulk Supply
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
400
350 -
As shown, the estimates from the Vintaging Model are lower than the GHGRP estimates by an average of 0.5
percent across the time series (i.e., 2010 through 2019). Potential reasons for the differences between the
reported and modeled data, include:
Industrial Processes and Product Use 4-135

-------
•	The Vintaging Model includes fewer saturated HFCs than are reported to EPA's GHGRP. However, the
additional reported HFCs represent a small fraction of total HFC use for this source category, both in
GWP-weighted and unweighted terms, and as such, it is not expected that the additional HFCs reported to
EPA are a major driver for the difference between the two sets of estimates. To the extent lower-GWP
isomers were used in lieu of the modeled chemicals (e.g., HFC-134 instead of HFC-134a), lower C02 Eq.
amounts in the GHGRP data compared to the modeled estimates would be expected.
•	Because the top-down data are reported at the time of actual production or import, and the bottom-up
data are calculated at the time of actual placement on the market, there could be a temporal discrepancy
when comparing data. Because the GHGRP data and the Vintaging Model estimates generally increase
over time (although some year-to-year variations exist), EPA would expect the modeled estimates to be
slightly lower than the corresponding GHGRP data due to this temporal effect.
•	An additional temporal effect can result from the stockpiling of chemicals by suppliers and distributors.
Suppliers might decide to produce or import additional quantities of HFCs for various reasons such as
expectations that prices may increase, or supplies may decrease, in the future. Such stockpiling behavior
was seen during ODS phasedowns, but it is unclear if such behavior exists amongst HFC suppliers in
anticipation of potential future controls on HFCs. Any such activity would increase the GHGRP data as
compared to the modeled data. This effect may be a major reason why the GHGRP data in 2017, 2018,
and 2019 are significantly higher than the modeled data.
•	Under EPA's GHGRP, all facilities that produce HFCs are required to report their quantities, whereas
importers or exporters of HFCs or pre-charged equipment and closed-cell foams that contain HFCs are
only required to report if either their total imports or their total exports of greenhouse gases are greater
than or equal to 25,000 metric tons of C02 Eq. per year. Thus, some imports may not be accounted for in
the GHGRP data. On the other hand, some exports might also not be accounted for in this data.
•	There could be noncompliance with the GHGRP. EPA routinely reviews import data provided by U.S.
Customs and Border Protection (CBP) to verify reported supply data and identify facilities that may be
subject to the GHGRP. Based on this review and other information, there appear to be companies that
imported or exported more than 25,000 metric tons C02 Eq. of HFCs annually that have not reported
imports or exports to the GHGRP.
•	In some years, imports and exports may be greater than consumption because the excess is being used to
increase chemical or equipment stockpiles as discussed above; in other years, the opposite may hold true.
Similarly, relocation of manufacturing facilities or recovery from the recession could contribute to
variability in imports or exports. Averaging imports and exports net supplies over multiple years can
minimize the impact of such fluctuations. For example, when the 2012 and 2013 net additions to the
supply are averaged, as shown in Table 4-104, the percent difference between the consumption estimates
decreases compared to the 2012-only and 2013-only estimates.
Table 4-104: Averaged U.S. HFC Demand (MMTCOz Eq.)

2010-
2011-
2012-
2013-
2014-
2015-
2016-
2017-
2018-

2011
2012
2013
2014
2015
2016
2017
2018
2019

Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Avg
Reported Net
Supply (GHGRP)
242
247
270
287
285
279
293
328
331
Modeled Demand
(Vintaging Model)
268
273
278
284
286
287
283
280
280
Percent Difference
11%
11%
3%
-1%
1%
3%
-3%
-14%
-15%
• The Vintaging Model does not reflect the dynamic nature of reported HFC consumption, with significant
differences seen in each year. Whereas the Vintaging Model projects a slowly increasing overall demand,
with some annual fluctuations, actual consumption for specific chemicals or equipment may vary over
4-136 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
time and could even switch from positive to negative (indicating more chemical exported, transformed,
and destroyed than produced and imported in a given year). Furthermore, consumption as calculated in
the Vintaging Model is a function of demand not met by disposal recovery. If, in any given year, a
significant number of units are disposed, there will be a large amount of additional recovery in that year
that can cause an unexpected and not modeled decrease in demand and thus a decrease in consumption.
On the other hand, if market, economic, or other factors cause less than expected disposal and recovery,
actual supply would decrease, and hence consumption would increase to meet that demand not satisfied
by recovered quantities, increasing the GHGRP amounts.
•	The Vintaging Model is used to estimate the emissions that occur in the United States. As such, all
equipment or products that contain ODS or alternatives, including saturated HFCs, are assumed to
consume and emit chemicals equally as like equipment or products originally produced in the United
States. The GHGRP data from Subpart 00 (industrial greenhouse gas suppliers) includes HFCs produced or
imported and used to fill or manufacture products that are then exported from the United States. The
Vintaging Model estimates of demand and supply are not meant to incorporate such chemical. Likewise,
chemicals may be used outside the United States to create products or charge equipment that is then
imported to and used in the United States. The Vintaging Model estimates of demand and supply are
meant to capture this chemical, as it will lead to emissions inside the United States. The GHGRP data from
Subpart QQ (supply of HFCs in products) accounts for some of these differences; however, the scope of
Subpart QQ does not cover all such equipment or products and the chemical contained therein.
Depending on whether the United States is a net importer or net exporter of such chemical, this factor
may account for some of the difference shown above or might lead to a further discrepancy.
One factor, however, would only lead to modeled estimates to be even higher than the estimates shown and
hence for some years possibly higher than GHGRP data:
•	Saturated HFCs are also known to be used as a cover gas in the production of magnesium. The Vintaging
Model estimates here do not include the amount of HFCs for this use, but rather only the amount for uses
that traditionally were served by ODS. Nonetheless, EPA expects this supply not included in the Vintaging
Model estimates to be very small compared to the ODS substitute use for the years analyzed. An
indication of the different magnitudes of these categories is seen in the fact that the 2019 emissions from
that non-modeled source (0.1 MMT C02 Eq.) are much smaller than those for the ODS substitute sector
(170.6 MMT C02 Eq.).
Using a Tier 2 bottom-up modeling methodology to estimate emissions requires assumptions and expert
judgment. Comparing the Vintaging Model's estimates to GHGRP-reported estimates, particularly for more widely
used chemicals, can help validate the model but it is expected that the model will have limitations. This
comparison shows that Vintaging Model consumption estimates are well within the same order of magnitude as
the actual consumption data as reported to EPA's GHGRP although the differences in reported net supply and
modeled demand are still significant. Although it can be difficult to capture the observed market variability, the
Vintaging Model is periodically reviewed and updated to ensure that the model reflects the current and future
trajectory of ODS and ODS substitutes across all end-uses and the Vintaging Model will continue to be compared to
available top-down estimates in order to ensure the model accurately estimates HFC consumption and emissions.
Recalculations Discussion
For the current Inventory, updates to the Vintaging Model included:
•	Updating market size, substitute transitions, and charge size assumptions for Metered Dose Inhaler (MDI)
aerosols to align with stakeholder input and market research (EPA 2020i);
•	Replacing the commercial refrigeration foam end-use with ten discrete commercial refrigeration
application end-uses:
o Vending machine foam (EPA 2020h),
o Stand-alone equipment foam (EPA 2020g),
Industrial Processes and Product Use 4-137

-------
o Ice machine foam (EPA 2020f),
o Refrigerated food processing and dispensing equipment foam (EPA 2020e),
o Small walk-in cooler foam,
o Large walk-in cooler foam (EPA 2020a),
o Display case foam (CFC-11) and display case foam (CFC-12) (EPA 2020b),
o Road transport foam, and
o Intermodal container foam (EPA 2020c);
•	Updating market transitions for the ice maker end-use based on manufacturer information on refrigerant
use (EPA 2020d);
•	Adjusting manufacturing emissions for domestic refrigerator foam to only include equipment
manufactured within the United States, including those that are produced for export, and excluding those
that are imported with foam;
•	Updating market size, manufacturing loss rate, disposal lost rate, and post-life emission rate assumptions
for PU and PIR boardstock foams (EPA 2020j); and
•	Updating market size of residential unitary AC, small commercial unitary AC, and large commercial unitary
AC to align with AHRI and EIA data (EPA 2021b).
Together, these updates decreased greenhouse gas emissions on average by 0.7 percent between 1990 and 2018.
Planned Improvements
Future improvements to the Vintaging Model are planned for the Fire Suppression sector. Specifically, streaming
agent fire suppression lifetimes, market size, and growth rates are under review to align more closely with real
world activities. We expect this revision to be prepared for the 2022 or 2023 report. In future reports, EPA also
plans to compare atmospheric emissions of select HFCs to the modeled results as an additional QA/QC and
Verification analysis.
4.25 Electrical Transmission and Distribution
(CRF Source Category 2G1)
The largest use of sulfur hexafluoride (SF6), both in the United States and internationally, is as an electrical
insulator and interrupter in equipment that transmits and distributes electricity (RAND 2004). The gas has been
employed by the electric power industry in the United States since the 1950s because of its dielectric strength and
arc-quenching characteristics. It is used in gas-insulated substations, circuit breakers, and other switchgear. SF6 has
replaced flammable insulating oils in many applications and allows for more compact substations in dense urban
areas.
Fugitive emissions of SF6 can escape from gas-insulated substations and switchgear through seals, especially from
older equipment. The gas can also be released during equipment manufacturing, installation, servicing, and
disposal. Emissions of SF6 from equipment manufacturing and from electrical transmission and distribution
systems were estimated to be 4.2 MMT C02 Eq. (0.2 kt) in 2019. This quantity represents an 82 percent decrease
from the estimate for 1990 (see Table 4-105 and Table 4-106). There are a few potential causes for this decrease: a
sharp increase in the price of SF6 during the 1990s and a growing awareness of the environmental impact of SF6
emissions through programs such as EPA's voluntary SF6 Emission Reduction Partnership for Electric Power
Systems (Partnership) and EPA's GHGRP, regulatory drivers at the state and local levels, and research and
development of alternative gases to SF6 that can be used in gas-insulated substations. Utilities participating in the
Partnership have lowered their emission factor from 13 percent in 1999 (kg SF6 emitted per kg of nameplate
capacity) to less than 2 percent in 2019. A recent examination of the SF6 emissions reported by electric power
systems to EPA's GHGRP revealed that SF6 emissions from reporters have decreased by 35 percent from 2011 to
4-138 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2019,110 with much of the reduction seen from utilities that are not participants in the Partnership. These utilities
may be making relatively large reductions in emissions as they take advantage of relatively large and/or
inexpensive emission reduction opportunities (i.e., "low hanging fruit," such as replacing major leaking circuit
breakers) that Partners have already taken advantage of under the voluntary program (Ottinger et al. 2014). Total
emissions from electrical transmission and distribution in 2019 were higher than 2018 emissions, increasing by 8.5
percent. The increase in emissions may be attributed to a combination of increasing nameplate capacity and
transmission miles in the United States, and an increase in the average emission rate reported to the GHGRP in
2019.
Table 4-105: SF6 Emissions from Electric Power Systems and Electrical Equipment
Manufacturers (MMT CO2 Eq.)


Electrical


Electric Power
Equipment

Year
Systems
Manufacturers
Total
1990
22.8
0.3
23.2
2005
7.7
0.7
8.4
2015
3.5
0.3
3.8
2016
3.8
0.3
4.1
2017
3.8
0.3
4.2
2018
3.6
0.3
3.9
2019
3.9
0.3
4.2
Note: Totals may not sum due to independent rounding.
Table 4-106: SF6 Emissions from Electric Power Systems and Electrical Equipment
Manufacturers (kt)
Year Emissions
1990	To
2005	0.4
2015	0.2
2016	0.2
2017	0.2
2018	0.2
2019	0.2
110 Analysis of emission trends from facilities reporting to EPA's GHGRP is imperfect due to an inconsistent group of reporters
year to year. A facility that has reported total non-biogenic greenhouse gas emissions below 15,000 metric tons of carbon
dioxide equivalent (MT C02 Eq.) for three consecutive years or below 25,000 MT C02 Eq. for five consecutive years to EPA's
GHGRP can discontinue reporting for all direct emitter subparts. For this sector, most of the variability in the group of reporters
is due to facilities exiting the GHGRP due to being below one of these thresholds; however, facilities must re-enter the program
if their emissions at a later date are above 25,000 MT C02 Eq., which may occur for a variety of reasons, including changes in
facility size and changes in emission rates.
Industrial Processes and Product Use 4-139

-------
Methodology
The estimates of emissions from Electrical Transmission and Distribution are comprised of emissions from electric
power systems and emissions from the manufacture of electrical equipment. The methodologies for estimating
both sets of emissions are described below.
1990 through 1998 Emissions from Electric Power Systems
Emissions from electric power systems from 1990 through 1998 were estimated based on (1) the emissions
estimated for this source category in 1999, which, as discussed in the next section, were based on the emissions
reported during the first year of EPA's SF6 Emission Reduction Partnership for Electric Power Systems (Partnership),
and (2) the RAND survey of global SF6 emissions. Because most utilities participating in the Partnership reported
emissions only for 1999 through 2011, modeling was used to estimate SF6 emissions from electric power systems
for the years 1990 through 1998. To perform this modeling, U.S. emissions were assumed to follow the same
trajectory as global emissions from this source during the 1990 to 1999 period. To estimate global emissions, the
RAND survey of global SF6 sales was used, together with the following equation for estimating emissions, which is
derived from the mass-balance equation for chemical emissions (Volume 3, Equation 7.3) in the 2006IPCC
Guidelines.m (Although Equation 7.3 of the 2006 IPCC Guidelines appears in the discussion of substitutes for
ozone-depleting substances, it is applicable to emissions from any long-lived pressurized equipment that is
periodically serviced during its lifetime.)
Emissions (kilograms SF6) = SF6 purchased to refill existing equipment (kilograms) + nameplate capacity of retiring
equipment (kilograms)112
Note that the above equation holds whether the gas from retiring equipment is released or recaptured; if the gas
is recaptured, it is used to refill existing equipment, thereby lowering the amount of SF6 purchased by utilities for
this purpose.
Gas purchases by utilities and equipment manufacturers from 1961 through 2003 are available from the RAND
(2004) survey. To estimate the quantity of SF6 released or recovered from retiring equipment, the nameplate
capacity of retiring equipment in a given year was assumed to equal 81.2 percent of the amount of gas purchased
by electrical equipment manufacturers 40 years previous (e.g., in 2000, the nameplate capacity of retiring
equipment was assumed to equal 81.2 percent of the gas purchased in 1960). The remaining 18.8 percent was
assumed to have been emitted at the time of manufacture. The 18.8 percent emission factor is an average of IPCC
default SF6 emission rates for Europe and Japan for 1995 (IPCC 2006). The 40-year lifetime for electrical equipment
is also based on IPCC (2006). The results of the two components of the above equation were then summed to yield
estimates of global SF6 emissions from 1990 through 1999.
U.S. emissions between 1990 and 1999 are assumed to follow the same trajectory as global emissions during this
period. To estimate U.S. emissions, global emissions for each year from 1990 through 1998 were divided by the
estimated global emissions from 1999. The result was a time series of factors that express each year's global
emissions as a multiple of 1999 global emissions. Historical U.S. emissions were estimated by multiplying the factor
for each respective year by the estimated U.S. emissions of SF6 from electric power systems in 1999 (estimated to
be 13.6 MMTCOz Eq.).
Two factors may affect the relationship between the RAND sales trends and actual global emission trends. One is
utilities' inventories of SF6 in storage containers. When SF6 prices rise, utilities are likely to deplete internal
inventories before purchasing new SF6 at the higher price, in which case SF6 sales will fall more quickly than
emissions. On the other hand, when SF6 prices fall, utilities are likely to purchase more SF6 to rebuild inventories, in
111	Ideally, sales to utilities in the United States between 1990 and 1999 would be used as a model. However, this information
was not available. There were only two U.S. manufacturers of SF6 during this time period, so it would not have been possible to
conceal sensitive sales information by aggregation.
112	Nameplate capacity is defined as the amount of SF6 within fully charged electrical equipment.
4-140 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
which case sales will rise more quickly than emissions. This effect was accounted for by applying 3-year smoothing
to utility SF6 sales data. The other factor that may affect the relationship between the RAND sales trends and
actual global emissions is the level of imports from and exports to Russia and China. SF6 production in these
countries is not included in the RAND survey and is not accounted for in any another manner by RAND. However,
atmospheric studies confirm that the downward trend in estimated global emissions between 1995 and 1998 was
real (see the Uncertainty discussion below).
1999 through 2019 Emissions from Electric Power Systems
Emissions from electric power systems from 1999 to 2019 were estimated based on: (1) reporting from utilities
participating in EPA's SF6 Emission Reduction Partnership for Electric Power Systems (Partners), which began in
1999; (2) reporting from utilities covered by EPA's GHGRP, which began in 2012 for emissions occurring in 2011
(GHGRP-Only Reporters); and (3) the relationship between utilities' reported emissions and their transmission
miles as reported in the 2001, 2004, 2007, 2010, 2013, and 2016 Utility Data Institute (UDI) Directories of Electric
Power Producers and Distributors (UDI 2001, 2004, 2007, 2010, 2013, and 2017), which was applied to the electric
power systems that do not report to EPA (Non-Reporters). (Transmission miles are defined as the miles of lines
carrying voltages above 34.5 kV).
Partners
Over the period from 1999 to 2019, Partner utilities, which for inventory purposes are defined as utilities that
either currently are or previously have been part of the Partnership,113 represented 49 percent, on average, of
total U.S. transmission miles. Partner utilities estimated their emissions using a Tier 3 utility-level mass balance
approach (IPCC 2006). If a Partner utility did not provide data for a particular year, emissions were interpolated
between years for which data were available or extrapolated based on Partner-specific transmission mile growth
rates. In 2012, many Partners began reporting their emissions (for 2011 and later years) through EPA's GHGRP
(discussed further below) rather than through the Partnership. In 2019, approximately 1 percent of the total
emissions attributed to Partner utilities were reported through Partnership reports. Approximately 96 percent of
the total emissions attributed to Partner utilities were reported and verified through EPA's GHGRP. Partners
without verified 2019 data accounted for approximately 3 percent of the total emissions attributed to Partner
utilities.114
The GHGRP program has an "offramp" provision (40 CFR Part 98.2(i)) that exempts facilities from reporting under
certain conditions. If reported total greenhouse gas emissions are below 15,000 metric tons of carbon dioxide
equivalent (MT C02 Eq.) for three consecutive years or below 25,000 MT C02 Eq. for five consecutive years, the
facility may elect to discontinue reporting. GHGRP reporters that have off-ramped are extrapolated for three years
of non-reporting using a utility-specific transmission mile growth rate. After three consecutive years of non-
reporting, they are treated as non-reporters, as described in the section below on non-reporters. Partners that
have years of non-reporting between reporting years are gap filled by interpolating between reported values.
GHGRP-Only Reporters
EPA's GHGRP requires users of SF6 in electric power systems to report emissions if the facility has a total SF6
nameplate capacity that exceeds 17,820 pounds. (This quantity is the nameplate capacity that would result in
annual SF6 emissions equal to 25,000 metric tons of C02 equivalent at the historical emission rate reported under
113	Starting in the 1990 to 2015 Inventory, partners who had reported three years or less of data prior to 2006 were removed.
Most of these Partners had been removed from the list of current Partners but remained in the Inventory due to the
extrapolation methodology for non-reporting partners.
114	Only data reported as of September 28, 2020 are used in the emission estimates for the prior year of reporting. Emissions
for Partners that did not report to the Partnership or GHGRP are extrapolated for three years using a utility-specific
transmission mile growth rate. After four consecutive years of non-reporting they are included in the 'non-reporting Partners'
category. It should be noted that data reported through EPA's GHGRP must go through a verification process. For electric
power systems, verification involved a series of electronic range, completeness, and algorithm checks for each report
submitted.
Industrial Processes and Product Use 4-141

-------
the Partnership.) As under the Partnership, electric power systems that report their SF6 emissions under EPA's
GHGRP are required to use the Tier 3 utility-level mass-balance approach. Many Partners began reporting their
emissions through EPA's GHGRP in 2012 (reporting emissions for 2011 and later years) because their nameplate
capacity exceeded the reporting threshold. Some Partners who did not report through EPA's GHGRP continued to
report through the Partnership.
In addition, many non-Partners began reporting to EPA for the first time through its GHGRP in 2012. Non-Partner
emissions reported and verified under EPA's GHGRP were compiled to form a new category of reported data
(GHGRP-Only Reporters). GHGRP-Only Reporters accounted for 21 percent of U.S. transmission miles and 24
percent of estimated U.S. emissions from electric power system in 2019.115
Emissions for GHGRP-only reporters that off-ramp are extrapolated for three years of non-reporting using a utility-
specific transmission mile growth rate. After three consecutive years of non-reporting, they are treated as non-
reporters, and emissions are subsequently estimated based on the methodology described below.
Non-Reporters
Emissions from Non-Reporters (i.e., utilities other than Partners and GHGRP-Only Reporters) in every year since
1999 were estimated using the results of a regression analysis that correlated emissions from reporting utilities
(using verified data from both Partners and GHGRP-Only Reporters) with their transmission miles.116 As noted
above, non-Partner emissions were reported to the EPA for the first time through its GHGRP in 2012 (representing
2011 emissions). This set of reported data was of particular interest because it provided insight into the emission
rate of non-Partners, which previously was assumed to be equal to the historical (1999) emission rate of Partners.
Specifically, emissions were estimated for Non-Reporters as follows:
•	Non-Reporters, 1999 to 2011: First, the 2011 emission rates (per kg nameplate capacity and per
transmission mile) reported by Partners and GHGRP-Only Reporters were reviewed to determine whether
there was a statistically significant difference between these two groups. Transmission mileage data for
2011 was reported through GHGRP, with the exception of transmission mileage data for Partners that did
not report through GHGRP, which was obtained from UDI. It was determined that there is no statistically
significant difference between the emission rates of Partners and GHGRP-Only reporters; therefore,
Partner and GHGRP-Only reported data for 2011 were combined to develop regression equations to
estimate the emissions of Non-Reporters. Historical emissions from Non-Reporters were estimated by
linearly interpolating between the 1999 regression coefficient (based on 1999 Partner data) and the 2011
regression coefficient.
•	Non-Reporters, 2012 to Present: It was determined that there continued to be no statistically significant
difference between the emission rates reported by Partners and by GHGRP-Only Reporters. Therefore,
the emissions data from both groups were combined to develop regression equations for 2012. This was
repeated for 2013 through 2019 using Partner and GHGRP-Only Reporter data for each year.
o The 2019 regression equation for reporters was developed based on the emissions reported by a
subset of Partner utilities and GHGRP-Only utilities who reported non-zero emissions and non-zero
transmission miles (representing approximately 65 percent of total U.S. transmission miles). The
regression equation for 2019 is:
115	GHGRP-reported and Partner transmission miles from a number of facilities were equal to zero with non-zero emissions.
These facilities emissions were added to the emissions totals for their respective parent companies when identifiable and not
included in the regression equation when not identifiable or applicable. Other facilities reported non-zero transmission miles
with zero emissions, or zero transmission miles and zero emissions. These facilities were not included in the development of the
regression equations (discussed further below). These emissions are already implicitly accounted for in the relationship
between transmission miles and emissions.
116	In the United States, SF6 is contained primarily in transmission equipment rated above 34.5 kV.
4-142 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Emissions (kg) = 0.226 x Transmission Miles
Table 4-107 below shows the percentage of transmission miles covered by reporters (i.e., associated with reported
data) and the regression coefficient for 1999 (the first year data was reported), and for 2011 through present (the
years with GHGRP reported data). The coefficient increased between 2015 and 2019.
Table 4-107: Transmission Mile Coverage (Percent) and Regression Coefficients (kg per

1999
2005
2015
2016
2017
2018
2019
Percentage of Miles Covered by Reporters
50%
50%
73%
73%
74%
70%
65%
Regression Coefficient3
0.71
0.35
0.19
0.21
0.24
0.21
0.23
a Regression coefficient for emissions is calculated utilizing transmission miles as the explanatory variable and emissions
as the response variable. The equation utilizes a constant intercept of zero. When calculating the regression coefficient,
outliers are also removed from the analysis when the standard residual for that reporter exceeds the value 3.0.
Data on transmission miles for each Non-Reporter for the years 2000, 2003, 2006, and 2009, 2012, and 2016 were
obtained from the 2001, 2004, 2007, 2010, 2013, and 2017 UDI Directories of Electric Power Producers and
Distributors, respectively (UDI 2001, 2004, 2007, 2010, 2013, and 2017). The following trends in transmission miles
have been observed over the time series:
•	The U.S. transmission system grew by over 22,000 miles between 2000 and 2003 yet declined by almost
4,000 miles between 2003 and 2006. Given these fluctuations, periodic increases are assumed to occur
gradually. Therefore, transmission mileage was assumed to increase at an annual rate of 1.2 percent
between 2000 and 2003 and decrease by 0.20 percent between 2003 and 2006.
•	The U.S. transmission system's annual growth rate grew to 1.7 percent from 2006 to 2009 as transmission
miles increased by more than 33,000 miles.
•	The annual growth rate for 2009 through 2012 was calculated to be 1.5 percent as transmission miles
grew yet again by over 30,000 miles during this time period.
•	The annual transmission mile growth rate for 2012 through 2019 was calculated to be 0.6 percent, as
transmission miles increased by approximately 43,000 miles.
Transmission miles for each year for non-reporters were calculated by interpolating between UDI reported values
obtained from the 2001, 2004, 2007, 2010, 2013 and 2017 UDI directories. In cases where a non-reporter
previously reported the GHGRP or the Partnership, transmission miles were interpolated between the most
recently reported value and the next available UDI value.
Total Industry Emissions
As a final step, total electric power system emissions from 1999 through 2019 were determined for each year by
summing the Partner reported and estimated emissions (reported data was available through the EPA's SF6
Emission Reduction Partnership for Electric Power Systems), the GHGRP-only reported emissions, and the non-
reporting utilities' emissions (determined using the regression equations).
1990 through 2019 Emissions from Manufacture of Electrical Equipment
Three different methods were used to estimate 1990 to 2019 emissions from original electrical equipment
manufacturers (OEMs).
•	OEM emissions from 1990 through 2000 were derived by assuming that manufacturing emissions equaled
10 percent of the quantity of SF6 provided with new equipment. The 10 percent emission rate is the
average of the "ideal" and "realistic" manufacturing emission rates (4 percent and 17 percent,
respectively) identified in a paper prepared under the auspices of the International Council on Large
Electric Systems (CIGRE) in February 2002 (O'Connell et al. 2002). The quantity of SF6 provided with new
equipment was estimated based on statistics compiled by the National Electrical Manufacturers
Association (NEMA). These statistics were provided for 1990 to 2000.
•	OEM emissions from 2000 through 2010 were estimated by (1) interpolating between the emission rate
estimated for 2000 (10 percent) and an emission rate estimated for 2011 based on reporting by OEMs
Industrial Processes and Product Use 4-143

-------
through the GHGRP (5.7 percent), and (2) estimating the quantities of SF6 provided with new equipment
for 2001 to 2010. The quantities of SF6 provided with new equipment were estimated using Partner
reported data and the total industry SF6 nameplate capacity estimate (156.5 MMT C02 Eq. in 2010).
Specifically, the ratio of new nameplate capacity to total nameplate capacity of a subset of Partners for
which new nameplate capacity data was available from 1999 to 2010 was calculated. These ratios were
then multiplied by the total industry nameplate capacity estimate for each year to derive the amount of
SF6 provided with new equipment for the entire industry. Additionally, to obtain the 2011 emission rate
(necessary for estimating 2001 through 2010 emissions), the estimated 2011 emissions (estimated using
the third methodology listed below) were divided by the estimated total quantity of SF6 provided with
new equipment in 2011. The 2011 quantity of SF6 provided with new equipment was estimated in the
same way as the 2001 through 2010 quantities.
• OEM emissions from 2011 through 2019 were estimated using the SF6 emissions from OEMs reporting to
the GHGRP, and an assumption that these reported emissions account for a conservatively low estimate
of 50 percent of the total emissions from all U.S. OEMs.
Uncertainty and Time-Series Consistency
To estimate the uncertainty associated with emissions of SF6 from Electrical Transmission and Distribution,
uncertainties associated with four quantities were estimated: (1) emissions from Partners, (2) emissions from
GHGRP-Only Reporters, (3) emissions from Non-Reporters, and (4) emissions from manufacturers of electrical
equipment. A Monte Carlo analysis was then applied to estimate the overall uncertainty of the emissions estimate.
Total emissions from the SF6 Emission Reduction Partnership include emissions from both reporting (through the
Partnership or EPA's GHGRP) and non-reporting Partners. For reporting Partners, individual Partner-reported SF6
data was assumed to have an uncertainty of 10 percent. Based on a Monte Carlo analysis, the cumulative
uncertainty of all Partner-reported data was estimated to be 6.0 percent. The uncertainty associated with
extrapolated or interpolated emissions from non-reporting Partners was assumed to be 20 percent.
For GHGRP-Only Reporters, reported SF6 data was assumed to have an uncertainty of 20 percent.117 Based on a
Monte Carlo analysis, the cumulative uncertainty of all GHGRP-Only reported data was estimated to be 8.5
percent.
There are two sources of uncertainty associated with the regression equations used to estimate emissions in 2019
from Non-Reporters: (1) uncertainty in the coefficients (as defined by the regression standard error estimate), and
(2) the uncertainty in total transmission miles for Non-Reporters. Uncertainties were also estimated regarding (1)
estimates of SF6 emissions from OEMs reporting to EPA's GHGRP, and (2) the assumption on the percent share of
OEM emissions from OEMs reporting to EPA's GHGRP.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-108. Electrical
Transmission and Distribution SF6 emissions were estimated to be between 3.6 and 5.0 MMT C02 Eq. at the 95
percent confidence level. This indicates a range of approximately 16 percent below and 18 percent above the
emission estimate of 4.2 MMT C02 Eq.
117 Uncertainty is assumed to be higher for the GHGRP-Only category, because 2011 is the first year that those utilities have
reported to EPA.
4-144 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 4-108: Approach 2 Quantitative Uncertainty Estimates for SF6 Emissions from
Electrical Transmission and Distribution (MMT CO2 Eq. and Percent)
Source
Gas 2019 Emission Estimate
Uncertainty Range Relative to 2018 Emission Estimate3

(MMTCOz Eq.)
(MMT C02
: Eq.)
(%)



Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Electrical Transmission
and Distribution
SF6 4.2
3.6
5.0
-16%
+18%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
In addition to the uncertainty quantified above, there is uncertainty associated with using global SF6 sales data to
estimate U.S. emission trends from 1990 through 1999. However, the trend in global emissions implied by sales of
SF6 appears to reflect the trend in global emissions implied by changing SF6 concentrations in the atmosphere. That
is, emissions based on global sales declined by 29 percent between 1995 and 1998 (RAND 2004), and emissions
based on atmospheric measurements declined by 17 percent over the same period (Levin et al. 2010).
Several pieces of evidence indicate that U.S. SF6 emissions were reduced as global emissions were reduced. First,
the decreases in sales and emissions coincided with a sharp increase in the price of SF6 that occurred in the mid-
1990s and that affected the United States as well as the rest of the world. A representative from DILO, a major
manufacturer of SF6 recycling equipment, stated that most U.S. utilities began recycling rather than venting SF6
within two years of the price rise. Finally, the emissions reported by the one U.S. utility that reported its emissions
for all the years from 1990 through 1999 under the Partnership showed a downward trend beginning in the mid-
1990s.
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
For the GHGRP data, EPA verifies annual facility-level reports through a multi-step process (e.g., including a
combination of pre-and post-submittal electronic checks and manual reviews by staff) to identify potential errors
and ensure that data submitted to EPA are accurate, complete, and consistent (EPA 2015).118 Based on the results
of the verification process, EPA follows up with facilities to resolve mistakes that may have occurred. The post-
submittals checks are consistent with a number of general and category-specific QC procedures including: range
checks, statistical checks, algorithm checks, and year-to-year checks of reported data and emissions.
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter and Annex 8 for more details.
Recalculations Discussion
The historical emissions estimated for this source category have undergone the following revisions for the period
1990 through 2018.
• GHGRP report resubmissions: Historical estimates for the period 2011 through 2018 were updated
relative to the previous report based on revisions to reported historical data in EPA's GHGRP.
118 GHGRP Report Verification Factsheet. .
Industrial Processes and Product Use 4-145

-------
As a result of the recalculations, SF6 emissions from electrical transmission and distribution decreased by 3.9
percent for 2018 relative to the previous report, and SF6 nameplate capacity decreased by 0.3 percent for 2018
relative to the previous report. On average, SF6 emission estimates for the entire time series decreased by
approximately 0.1 percent per year.
Planned Improvements
EPA plans to more closely examine transmission miles data by company provided by the UDI data sets, which has
been historically purchased every three years, to identify inconsistencies in the companies included in the data
sets and improve the transmission mile estimates to address data gaps, as necessary. In future inventory years,
EPA plans to identify additional sources for transmission miles data by company due to a discontinuation of this
specific data set in 2017.
Additionally, as the information on the type of new and retiring equipment is collected through GHGRP reporting,
EPA expects this data to provide insight into the relative importance of the two types of equipment as potential
emission sources. Historically, hermetically sealed pressure equipment has been considered to be a relatively small
source of SF6 in the United States; however, better estimating its potential source of emissions upon end-of-life
(i.e., disposal emissions) is an area for further analysis.
EPA also plans to investigate ways in which the electric transmission and distribution sector estimates can be
disaggregated to state level emissions estimates in order to provide greater clarity and specificity in emissions
rates by region.
4.26 Nitrous Oxide from Product Uses (CRF
Source Category 2G3)
Nitrous oxide (N20) is a clear, colorless, oxidizing liquefied gas with a slightly sweet odor which is used in a wide
variety of specialized product uses and applications. The amount of N20 that is actually emitted depends upon the
specific product use or application.
There are a total of three N20 production facilities currently operating in the United States (Ottinger 2021). Nitrous
oxide is primarily used in carrier gases with oxygen to administer more potent inhalation anesthetics for general
anesthesia, and as an anesthetic in various dental and veterinary applications. The second main use of N20 is as a
propellant in pressure and aerosol products, the largest application being pressure-packaged whipped cream.
Small quantities of N20 also are used in the following applications:
•	Oxidizing agent and etchant used in semiconductor manufacturing;
•	Oxidizing agent used, with acetylene, in atomic absorption spectrometry;
•	Production of sodium azide, which is used to inflate airbags;
•	Fuel oxidant in auto racing; and
•	Oxidizing agent in blowtorches used by jewelers and others (Heydorn 1997).
Production of N20 in 2019 was approximately 15 kt (see Table 4-109).
Table 4-109: N2O Production (kt)
Year kt
1990 16
2005 15
2015 15
4-146 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2016	15
2017	15
2018	15
2019	15
Nitrous oxide emissions were 4.2 MMT C02 Eq. (14 kt N20) in 2019 (see Table 4-110). Production of N20 stabilized
during the 1990s because medical markets had found other substitutes for anesthetics, and more medical
procedures were being performed on an outpatient basis using local anesthetics that do not require N20. The use
of N20 as a propellant for whipped cream has also stabilized due to the increased popularity of cream products
packaged in reusable plastic tubs (Heydorn 1997).
Table 4-110: N2O Emissions from N2O Product Usage (MMT CO2 Eq. and kt)
Year
MMTCOz Eq.
kt
1990
4.2
14
2005
4.2
14
2015
4.2
14
2016
4.2
14
2017
4.2
14
2018
4.2
14
2019
4.2
14
Methodology
Emissions from N20 product uses were estimated using the following equation:
Epu = X X ERa)
a
where,
Epu —
N20 emissions from product uses, metric tons
P
Total U.S. production of N20, metric tons
a =
specific application
Sa
Share of N20 usage by application a
ERa
Emission rate for application a, percent
The share of total quantity of N20 usage by end-use represents the share of national N20 produced that is used by
the specific subcategory (e.g., anesthesia, food processing). In 2019, the medical/dental industry used an
estimated 89.5 percent of total N20 produced, followed by food processing propellants at 6.5 percent. All other
subcategories, including semiconductor manufacturing, atomic absorption spectrometry, sodium azide production,
auto racing, and blowtorches, used the remainder of the N20 produced. This subcategory breakdown has changed
only slightly over the past decade. For instance, the small share of N20 usage in the production of sodium azide
declined significantly during the 1990s. Due to the lack of information on the specific time period of the phase-out
in this market subcategory, most of the N20 usage for sodium azide production is assumed to have ceased after
1996, with the majority of its small share of the market assigned to the larger medical/dental consumption
subcategory (Heydorn 1997). For 1990 through 1996, N20 usage was allocated across the following subcategories:
medical applications, food processing propellant, and sodium azide production. A usage emissions rate was then
applied for each subcategory to estimate the amount of N20 emitted.
Only the medical/dental and food propellant subcategories were estimated to release emissions into the
atmosphere that are not captured under another source category, and therefore these subcategories were the
only usage subcategories with emission rates. Emissions of N20 from semiconductor manufacturing are described
in Section 4.23 Electronics Industry (CRF Source Category 2E) and reported under CRF Source Category 2H3. For
Industrial Processes and Product Use 4-147

-------
the medical/dental subcategory, due to the poor solubility of N20 in blood and other tissues, none of the N20 is
assumed to be metabolized during anesthesia and quickly leaves the body in exhaled breath. Therefore, an
emission factor of 100 percent was used for this subcategory (IPCC 2006). For N20 used as a propellant in
pressurized and aerosol food products, none of the N20 is reacted during the process and all of the N20 is emitted
to the atmosphere, resulting in an emission factor of 100 percent for this subcategory (IPCC 2006). For the
remaining subcategories, all of the N20 is consumed or reacted during the process, and therefore the emission rate
was considered to be zero percent (Tupman 2003).
The 1990 through 1992 N20 production data were obtained from SRI Consulting's Nitrous Oxide, North America
(Heydorn 1997). Nitrous oxide production data for 1993 through 1995 were not available. Production data for
1996 was specified as a range in two data sources (Heydorn 1997; Tupman 2003). In particular, for 1996, Heydorn
(1997) estimates N20 production to range between 13.6 and 18.1 thousand metric tons. Tupman (2003) provided a
narrower range (15.9 to 18.1 thousand metric tons) for 1996 that falls within the production bounds described by
Heydorn (1997). Tupman (2003) data are considered more industry-specific and current; therefore, the midpoint
of the narrower production range was used to estimate N20 emissions for years 1993 through 2001 (Tupman
2003). The 2002 and 2003 N20 production data were obtained from the Compressed Gas Association Nitrous
Oxide Fact Sheet and Nitrous Oxide Abuse Hotline (CGA 2002, 2003). These data were also provided as a range. For
example, in 2003, CGA (2003) estimates N20 production to range between 13.6 and 15.9 thousand metric tons.
Due to the unavailability of data, production estimates for years 2004 through 2019 were held constant at the
2003 value.
The 1996 share of the total quantity of N20 used by each subcategory was obtained from SRI Consulting's Nitrous
Oxide, North America (Heydorn 1997). The 1990 through 1995 share of total quantity of N20 used by each
subcategory was kept the same as the 1996 number provided by SRI Consulting. The 1997 through 2001 share of
total quantity of N20 usage by sector was obtained from communication with a N20 industry expert (Tupman
2003). The 2002 and 2003 share of total quantity of N20 usage by sector was obtained from CGA (2002, 2003). Due
to the unavailability of data, the share of total quantity of N20 usage data for years 2004 through 2019 was
assumed to equal the 2003 value. The emissions rate for the food processing propellant industry was obtained
from SRI Consulting's Nitrous Oxide, North America (Heydorn 1997) and confirmed by a N20 industry expert
(Tupman 2003). The emissions rate for all other subcategories was obtained from communication with a N20
industry expert (Tupman 2003). The emissions rate for the medical/dental subcategory was obtained from the
2006 IPCC Guidelines.
Uncertainty and Time-Series Consistency
The overall uncertainty associated with the 2019 N20 emission estimate from N20 product usage was calculated
using the 2006 IPCC Guidelines (2006) Approach 2 methodology. Uncertainty associated with the parameters used
to estimate N20 emissions include production data, total market share of each end use, and the emission factors
applied to each end use, respectively.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-111. Nitrous oxide
emissions from N20 product usage were estimated to be between 3.2 and 5.2 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 24 percent below to 24 percent above the emission
estimate of 4.2 MMT C02 Eq.
Table 4-111: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from N2O
Product Usage (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate3


(MMT CO? Eq.)
(MMT CO? Eq.)
(%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
N20 from Product Uses
N20
4.2
3.2 5.2
-24% +24%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
4-148 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter.
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
Planned Improvements
EPA recently initiated an evaluation of alternative production statistics for cross-verification and updating time-
series activity data, emission factors, assumptions, etc., and a reassessment of N20 product use subcategories that
accurately represent trends. This evaluation includes conducting a literature review of publications and research
that may provide additional details on the industry. This work remains ongoing and thus far no additional sources
of data have been found to update this category.
Pending additional resources and planned improvement prioritization, EPA may also evaluate production and use
cycles, and the potential need to incorporate a time lag between production and ultimate product use and
resulting release of N20. Additionally, planned improvements include considering imports and exports of N20 for
product uses.
Finally, for future Inventories, EPA will examine data from EPA's GHGRP to improve the emission estimates for the
N20 product use subcategory. Particular attention will be made to ensure aggregated information can be published
without disclosing CBI and time-series consistency, as the facility-level reporting data from EPA's GHGRP are not
available for all inventory years as required in this Inventory. This is a lower priority improvement, and EPA is still
assessing the possibility of incorporating aggregated GHGRP CBI data to estimate emissions; therefore, this
planned improvement is still in development and not incorporated in the current Inventory report.
4.27 Industrial Processes and Product Use
Sources of Precursor Gases
In addition to the main greenhouse gases addressed above, many industrial processes can result in emissions of
various ozone precursors. The reporting requirements of the UNFCCC119 request that information be provided on
precursor greenhouse gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-CH4 volatile organic
compounds (NMVOCs), and sulfur dioxide (S02). These gases are not direct greenhouse gases, but indirectly affect
terrestrial radiation absorption by influencing the formation and destruction of tropospheric and stratospheric
ozone, or, in the case of S02, by affecting the absorptive characteristics of the atmosphere. Additionally, some of
these gases may react with other chemical compounds in the atmosphere to form compounds that are greenhouse
gases. As some of industrial applications also employ thermal incineration as a control technology, combustion
byproducts, such as CO and NOx, are also reported with this source category. NMVOCs, commonly referred to as
119 See .
Industrial Processes and Product Use 4-149

-------
"hydrocarbons," are the primary gases emitted from most processes employing organic or petroleum-based
products, and can also result from the product storage and handling.
Accidental releases of greenhouse gases associated with product use and handling can constitute major emissions
in this category. In the United States, emissions from product use are primarily the result of solvent evaporation,
whereby the lighter hydrocarbon molecules in the solvents escape into the atmosphere. The major categories of
product uses include: degreasing, graphic arts, surface coating, other industrial uses of solvents (e.g., electronics),
dry cleaning, and non-industrial uses (e.g., uses of paint thinner). Product usage in the United States also results in
the emission of small amounts of hydrofluorocarbons (HFCs) and hydrofluoroethers (HFEs), which are included
under Substitution of Ozone Depleting Substances in this chapter.
Total emissions of NOx, CO, and NMVOCs from non-energy industrial processes and product use from 1990 to 2019
are reported in Table 4-112. Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and Annex
6.3.
Table 4-112: NOx, CO, and NMVOC Emissions from Industrial Processes and Product Use (kt)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
NOx
592
572
408
402
397
397
397
Industrial Processes







Other Industrial Processes3
343
437
297
294
291
291
291
Metals Processing
88
60
62
61
60
60
60
Chemical and Allied Product







Manufacturing
152
55
41
39
37
37
37
Storage and Transport
3
15
5
5
5
5
5
Miscellaneous15
5
2
2
2
3
3
3
Product Uses







Surface Coating
1
3
1
1
1
1
1
Graphic Arts
+
0
0
0
0
0
0
Degreasing
+
0
0
0
0
0
0
Dry Cleaning
+
0
0
0
0
0
0
Other Industrial Processes3
+
0
0
0
0
0
0
Non-Industrial Processes0
+
0
0
0
0
0
0
Other
NA
0
0
0
0
0
0
CO
4,129
1,557
1,163
1,075
1,006
1,006
1,006
Industrial Processes







Metals Processing
2,395
752
510
468
425
425
425
Other Industrial Processes3
487
484
488
447
406
406
406
Chemical and Allied Product







Manufacturing
1,073
189
114
110
107
107
107
Miscellaneous15
101
32
42
42
61
61
61
Storage and Transport
69
97
7
7
7
7
7
Product Uses







Surface Coating
+
2
1
1
1
1
1
Other Industrial Processes3
4
0
0
0
0
0
0
Dry Cleaning
+
0
0
0
0
0
0
Degreasing
+
0
0
0
0
0
0
Graphic Arts
+
0
0
0
0
0
0
Non-Industrial Processes0
+
0
0
0
0
0
0
Other
NA
0
0
0
0
0
0
NMVOCs
7,638
5,849
3,796
3,776
3,767
3,767
3,767
Industrial Processes







Storage and Transport
1,352
1,308
619
626
633
633
633
Other Industrial Processes3
364
414
314
314
314
314
314
Chemical and Allied Product







Manufacturing
575
213
69
69
68
68
68
Miscellaneous15
20
17
24
24
35
35
35
Metals Processing
111
45
24
22
20
20
20
4-150 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Product Uses
Surface Coating
2,289
1,578
1,124
1,114
1,105
1,105
1,105
Non-Industrial Processes0
1,724
1,446
1,030
1,021
1,012
1,012
1,012
Degreasing
675
280
200
198
196
196
196
Dry Cleaning
195
230
164
163
161
161
161
Graphic Arts
249
194
138
137
136
136
136
Other Industrial Processes3
85
88
62
62
61
61
61
Other
+
36
26
25
25
25
25
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
NA (Not Available)
a Includes rubber and plastics manufacturing, and other miscellaneous applications.
b Miscellaneous includes the following categories: catastrophic/accidental release, other combustion, health
services, cooling towers, and fugitive dust. It does not include agricultural fires or slash/prescribed burning,
which are accounted for under the Field Burning of Agricultural Residues source.
c Includes cutback asphalt, pesticide application adhesives, consumer solvents, and other miscellaneous
applications.
Methodology
Emission estimates for 1990 through 2019 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2020) and disaggregated based on EPA (2003). Data were
collected for emissions of CO, NOx, volatile organic compounds (VOCs), and S02 from metals processing, chemical
manufacturing, other industrial processes, transport and storage, and miscellaneous sources. Emissions were
calculated either for individual source categories or for many categories combined, using basic activity data (e.g.,
the amount of raw material processed or the amount of solvent purchased) as an indicator of emissions. National
activity data were collected for individual categories from various agencies. Depending on the category, these
basic activity data may include data on production, fuel deliveries, raw material processed, etc.
Emissions for product use were calculated by aggregating product use data based on information relating to
product uses from different applications such as degreasing, graphic arts, etc. Emission factors for each
consumption category were then applied to the data to estimate emissions. For example, emissions from surface
coatings were mostly due to solvent evaporation as the coatings solidify. By applying the appropriate product-
specific emission factors to the amount of products used for surface coatings, an estimate of NMVOC emissions
was obtained. Emissions of CO and NOx under product use result primarily from thermal and catalytic incineration
of solvent-laden gas streams from painting booths, printing operations, and oven exhaust.
Activity data were used in conjunction with emission factors, which together relate the quantity of emissions to
the activity. Emission factors are generally available from the EPA's Compilation of Air Pollutant Emission Factors,
AP-42 (EPA 1997). The EPA currently derives the overall emission control efficiency of a source category from a
variety of information sources, including published reports, the 1985 National Acid Precipitation and Assessment
Program emissions inventory, and other EPA databases.
Uncertainty and Time-Series Consistency
Uncertainties in these estimates are partly due to the accuracy of the emission factors and activity data used. A
quantitative uncertainty analysis was not performed.
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
Industrial Processes and Product Use 4-151

-------
QA/QC and Verification
For more information on the general QA/QC process applied to this source category, consistent with Volume 1,
Chapter 6 of the 2006IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter.
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series.
4-152 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
5. Agriculture
Agricultural activities contribute directly to emissions of greenhouse gases through a variety of processes. This
chapter provides an assessment of methane (CH4) and nitrous oxide (N20) emissions from enteric fermentation in
domestic livestock, livestock manure management, rice cultivation, agricultural soil management, and field burning
of agricultural residues; as well as carbon dioxide (C02) emissions from liming and urea fertilization (see Figure
5-1). Additional C02, CH4 and N20 fluxes from agriculture-related land-use and land-use conversion activities, such
as cultivation of cropland, grassland fires, aquaculture, and conversion of forest land to cropland, are presented in
the Land Use, Land-Use Change, and Forestry (LULUCF) chapter. Carbon dioxide emissions from stationary and
mobile on-farm energy use and CH4 and N20 emissions from stationary on-farm energy use are reported in the
Energy chapter under the Industrial sector emissions. Methane and N20 emissions from mobile on-farm energy
use are reported in the Energy chapter under mobile fossil fuel combustion emissions.
Figure 5-1: 2019 Agriculture Chapter Greenhouse Gas Emission Sources
Agricultural Soil Management
Enteric Fermentation
Manure Management
Rice Cultivation
Urea Fertilization
Liming
Field Burning of Agricultural Residues
0 20 40 60 80 100 120 140 160 180 200
MMT COz Eq.
In 2019, the Agriculture sector was responsible for emissions of 628.6 MMT C02 Eq.,1 or 9.6 percent of total U.S.
greenhouse gas emissions. Methane emissions from enteric fermentation and manure management represent
27.1 percent and 9.5 percent of total CH4 emissions from anthropogenic activities, respectively. Of all domestic
animal types, beef and dairy cattle were the largest emitters of CH4. Rice cultivation and field burning of
agricultural residues were minor sources of CH4. Emissions of N20 by agricultural soil management through
activities such as fertilizer application and other agricultural practices that increased nitrogen availability in the soil
was the largest source of U.S. N20 emissions, accounting for 75.4 percent. Manure management and field burning
Agriculture as a Portion of
All Emissions
9.6% 	
Energy
¦ Agriculture
IPPU
Waste
1 Following the current reporting requirements under the United Nations Framework Convention on Climate Change (UNFCCC),
this Inventory report presents C02 equivalent values based on the IPCC Fourth Assessment Report (AR4) GWP values. See the
Introduction chapter for more information.
Agriculture 5-1

-------
of agricultural residues were also small sources of N20 emissions. Urea fertilization and liming accounted for 0.10
percent and 0.05 percent of total C02 emissions from anthropogenic activities, respectively.
Table 5- and Table 5-2 present emission estimates for the Agriculture sector. Between 1990 and 2019, C02and CH4
emissions from agricultural activities increased by 9.9 percent and 17.5 percent, respectively, while N20 emissions
from agricultural activities fluctuated from year to year, but increased by 10.4 percent overall. Trends in sources of
agricultural emissions over the 1990 to 2019 time series are shown in Figure 5-2.
Figure 5-2: Trends in Agriculture Chapter Greenhouse Gas Emission Sources
650
600
550
500
450
400
S
8 350
I—
S 300
250
200
100
50
0
Field Burning of Agricultural Residues
I Urea Fertilization
I Liming
Rice Cultivation
I Manure Management
I Enteric Fermentation
] Agricultural Soil Management
rvj ro
¦J")	r-N.
- ;
i
o-^Hrsjn-^-LOkor^cocno-^Hrsjro^rLnvD
oooooooooo»-iT-i.r-iT-i*-iT-i.rH
ooooooooooooooooo
in in r-sj in oj m oj in r>j in m im m in r\ in m
IN CM
Each year, some emission estimates in the Agriculture sector of the Inventory are recalculated and revised with
improved methods and/or data. In general, recalculations are made to the U.S. greenhouse gas emission estimates
either to incorporate new methodologies or, most commonly, to update recent historical data. These
improvements are implemented consistently across the previous Inventory's time series (i.e., 1990 through 2018)
to ensure that the trend is accurate. This year's notable updates include (1) Enteric Fermentation: default national
emission factors were updated for sheep and goats; (2) Field Burning of Agricultural Residues: updated parameters
within the methodology for combustion efficiency; (3) Urea Fertilization: updated methodology based on the
analytical solution from the Monte Carlo analysis; (4) Rice Cultivation: correction in splicing method; (5) Liming:
updated activity data from USGS; and (6) Agricultural Soil Management: using surrogate date method to update
the time series of PRP and manure N available for application to soils. In total, the improvements made to the
Agriculture sector in this Inventory increased greenhouse gas emissions by 2.5 MMT C02 Eq. (0.4 percent) in 2018.
For more information on specific methodological updates, please see the Recalculations discussions within the
respective source category sections of this chapter.
Emissions reported in the Agriculture chapter include those from all states; however, for Hawaii and Alaska some
agricultural practices that can increase nitrogen availability in the soil, and thus cause N20 emissions, are not
included (see chapter sections on "Uncertainty and Time-Series Consistency" and "Planned Improvements" for
more details). In addition, U.S. Territories and the District of Columbia are not estimated with the exception of
Urea Fertilization in Puerto Rico due to incomplete data. EPA continues to review available data on an ongoing
basis to include emissions from territories in future inventories to the extent they are occurring. Many U.S.
5-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
territories in the Pacific Islands have no permanent populations and therefore EPA assumes no agriculture
activities are occuring. See Annex 5 for more information on EPA's assessment of the sources not included in this
inventory.
Table 5-1: Emissions from Agriculture (MMT CO2 Eq.)
Gas/Source
1990

2005

2015
2016
2017
2018
2019
C02
7.1

7.9

8.5
8.0
8.1
7.4
7.8
Urea Fertilization
2.4

3.5

4.7
4.9
5.1
5.2
5.3
Liming
4.7

4.3

3.7
3.1
3.1
2.2
2.4
ch4
218.2

239.3

241.4
248.1
251.0
255.7
256.4
Enteric Fermentation
164.7

169.3

166.9
172.2
175.8
178.0
178.6
Manure Management
37.1

51.6

57.9
59.6
59.9
61.7
62.4
Rice Cultivation
16.0

18.0

16.2
15.8
14.9
15.6
15.1
Field Burning of Agricultural Residues
0.4

0.4

0.4
0.4
0.4
0.4
0.4
n2o
330.1

329.9

366.2
348.4
346.4
357.9
364.4
Agricultural Soil Management
315.9

313.4

348.5
330.1
327.6
338.2
344.6
Manure Management
14.0

16.4

17.5
18.1
18.7
19.4
19.6
Field Burning of Agricultural Residues
0.2

0.2

0.2
0.2
0.2
0.2
0.2
Total
555.3

577.1

616.1
604.4
605.5
621.0
628.6
Note: Totals may not sum due to independent rounding.






ible 5-2: Emissions from Agriculture (kt)






Gas/Source
1990

2005

2015
2016
2017
2018
2019
C02
7,084

7,854

8,464
7,959
8,131
7,440
7,782
Urea Fertilization
2,417

3,504

4,728
4,877
5,051
5,192
5,341
Liming
4,667

4,349

3,737
3,081
3,080
2,248
2,442
ch4
8,728

9,572

9,656
9,923
10,040
10,226
10,256
Enteric Fermentation
6,588

6,772

6,675
6,890
7,032
7,119
7,142
Manure Management
1,485

2,062

2,316
2,385
2,395
2,467
2,495
Rice Cultivation
640

720

648
631
596
623
602
Field Burning of Agricultural Residues
15

17

18
17
17
17
17
n2o
1,108

1,107

1,229
1,169
1,162
1,201
1,223
Agricultural Soil Management
1,060

1,052

1,169
1,108
1,099
1,135
1,156
Manure Management
47

55

59
61
63
65
66
Field Burning of Agricultural Residues
1

1

1
1
1
1
1
Note: Totals may not sum due to independent rounding.
Box 5-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals
In following the United Nations Framework Convention on Climate Change (UNFCCC) requirement under Article
4.1 to develop and submit national greenhouse gas emission inventories, the emissions and removals presented
in this report and this chapter, are organized by source and sink categories and calculated using internationally-
accepted methods provided by the Intergovernmental Panel on Climate Change (IPCC) in the 2006IPCC
Guidelines for National Greenhouse Gas Inventories (2006 IPCC Guidelines). Additionally, the calculated
emissions and removals in a given year for the United States are presented in a common format in line with the
UNFCCC reporting guidelines for the reporting of inventories under this international agreement. The use of
consistent methods to calculate emissions and removals by all nations providing their inventories to the
UNFCCC ensures that these reports are comparable. The presentation of emissions provided in the Agriculture
chapter do not preclude alternative examinations, but rather, this chapter presents emissions in a common
format consistent with how countries are to report Inventories under the UNFCCC. The report itself, and this
chapter, follows this standardized format, and provides an explanation of the application of methods used to
Agriculture 5-3

-------
calculate emissions from agricultural activities.
5.1 Enteric Fermentation (CRF Source Category
3A)	
Methane is produced as part of normal digestive processes in animals. During digestion, microbes resident in an
animal's digestive system ferment food consumed by the animal. This microbial fermentation process, referred to
as enteric fermentation, produces CH4 as a byproduct, which can be exhaled or eructated by the animal. The
amount of CH4 produced and emitted by an individual animal depends primarily upon the animal's digestive
system, and the amount and type of feed it consumes.2
Ruminant animals (e.g., cattle, buffalo, sheep, goats, and camels) are the major emitters of CH4 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 CH4 emissions per unit of body mass among all animal
types.
Non-ruminant animals (e.g., swine, horses, and mules and asses) also produce CH4 emissions through enteric
fermentation, although this microbial fermentation occurs in the large intestine. These non-ruminants emit
significantly less CH4 on a per-animal-mass basis than ruminants because the capacity of the large intestine to
produce CH4 is lower.
In addition to the type of digestive system, an animal's feed quality and feed intake also affect CH4 emissions. In
general, lower feed quality and/or higher feed intake leads to higher CH4 emissions. Feed intake is positively
correlated to animal size, growth rate, level of activity and production (e.g., milk production, wool growth,
pregnancy, or work). Therefore, feed intake varies among animal types as well as among different management
practices for individual animal types (e.g., animals in feedlots or grazing on pasture).
Methane emission estimates from enteric fermentation are provided in Table 5-3 and Table 5-4. Total livestock CH4
emissions in 2019 were 178.6 MMT C02 Eq. (7,142 kt). Beef cattle remain the largest contributor of CH4 emissions
from enteric fermentation, accounting for 72 percent in 2019. Emissions from dairy cattle in 2019 accounted for 24
percent, and the remaining emissions were from horses, sheep, swine, goats, American bison, mules and asses.3
Table 5-3: ChU Emissions from Enteric Fermentation (MMT CO2 Eq.)
Livestock Type
1990
2005
2015
2016
2017
2018
2019
Beef Cattle
119.1
125.2
118.0
123.0
126.3
128.1
129.1
Dairy Cattle
39.4
37.6
42.6
43.0
43.3
43.6
43.2
2	C02 emissions from livestock are not estimated because annual net C02 emissions are assumed to be zero - the C02
photosynthesized by plants is returned to the atmosphere as respired C02 (IPCC 2006).
3	Enteric fermentation emissions from poultry are not estimated because no IPCC method has been developed for determining
enteric fermentation CH4 emissions from poultry; at this time, developing of a country-specific method would require a
disproportionate amount of resources given the small magnitude of this source category. Enteric fermentation emissions from
camels are not estimated because there is no significant population of camels in the United States. Given the insignificance of
estimated camel emissions in terms of the overall level and trend in national emissions, there are no immediate improvement
plans to include this emissions category in the Inventory. See Annex 5 for more information on significance of estimated camel
emissions.
5-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Swine
2.0
2.3
2.6
2.6
2.7
2.8
2.9
Horses
1.0
1.7
1.4
1.4
1.3
1.2
1.1
Sheep
2.6
1.4
1.2
1.2
1.2
1.2
1.2
Goats
0.6
0.7
0.6
0.6
0.6
0.6
0.6
American Bison
0.1
0.4
0.4
0.4
0.4
0.4
0.4
Mules and Asses
+
0.1
0.1
0.1
0.1
0.1
0.1
Total
164.7
169.3
166.9
172.2
175.8
178.0
178.6
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 5-4: ChU Emissions from Enteric Fermentation (kt)
Livestock Type
1990
2005
2015
2016
2017
2018
2019
Beef Cattle
4,763
5,007
4,722
4,919
5,052
5,125
5,162
Dairy Cattle
1,574
1,503
1,706
1,722
1,730
1,744
1,729
Swine
81
92
102
105
108
111
115
Horses
40
70
57
54
51
48
46
Sheep
102
55
47
48
47
47
47
Goats
23
26
24
24
24
24
24
American Bison
4
17
14
15
15
15
16
Mules and Asses
1
2
3
3
3
3
3
Total	6,588	6,772	6,675 6,890 7,032 7,119 7,142
Note: Totals may not sum due to independent rounding.
From 1990 to 2019, emissions from enteric fermentation have increased by 8.4 percent. From 2018 to 2019,
emissions increased by 0.3 percent, largely driven by an increase in beef cattle populations. While emissions
generally follow trends in cattle populations, over the long term there are exceptions. For example, while dairy
cattle emissions increased 9.8 percent over the entire time series, the population has declined by 3.1 percent, and
milk production increased 58 percent (USDA 2019). These trends indicate that while emissions per head are
increasing, emissions per unit of product (i.e., meat, milk) are decreasing.
Generally, from 1990 to 1995 emissions from beef cattle increased and then decreased from 1996 to 2004. These
trends were mainly due to fluctuations in beef cattle populations and increased digestibility of feed for feedlot
cattle. Beef cattle emissions generally increased from 2004 to 2007, as beef cattle populations increased, and an
extensive literature review indicated a trend toward a decrease in feed digestibility for those years. Beef cattle
emissions decreased again from 2007 to 2014, as populations again decreased, but increased from 2015 to 2019,
consistent with another increase in population over those same years. Emissions from dairy cattle generally
trended downward from 1990 to 2004, along with an overall dairy cattle population decline during the same
period. Similar to beef cattle, dairy cattle emissions rose from 2004 to 2007 due to population increases and a
decrease in feed digestibility (based on an analysis of more than 350 dairy cow diets used by producers across the
United States). Dairy cattle emissions have continued to trend upward since 2007, in line with dairy cattle
population increases. Regarding trends in other animals, populations of sheep have steadily declined, with an
overall decrease of 54 percent since 1990. Horse populations are 15 percent greater than they were in 1990, but
their numbers have been declining by an average of 4 percent annually since 2007. Goat populations increased by
about 20 percent through 2007, steadily decreased through 2012, then increased again, by about 1 percent
annually, through 2019. Swine populations have trended upward through most of the time series, increasing 43
percent from 1990 to 2019. The population of American bison more than tripled over the 1990 to 2019 time
period, while the population of mules and asses increased by a factor of four.
Methodology
Livestock enteric fermentation emission estimate methodologies fall into two categories: cattle and other
domesticated animals. Cattle, due to their large population, large size, and particular digestive characteristics,
Agriculture 5-5

-------
account for the majority of enteric fermentation CH4 emissions from livestock in the United States. A more detailed
methodology (i.e., IPCC Tier 2) was therefore applied to estimate emissions for all cattle. Emission estimates for
other domesticated animals (horses, sheep, swine, goats, American bison, and mules and asses) were estimated
using the IPCC Tier 1 approach, as suggested by the 2006 IPCC Guidelines (see the Planned Improvements section).
While the large diversity of animal management practices cannot be precisely characterized and evaluated,
significant scientific literature exists that provides the necessary data to estimate cattle emissions using the IPCC
Tier 2 approach. The Cattle Enteric Fermentation Model (CEFM), developed by EPA and used to estimate cattle CH4
emissions from enteric fermentation, incorporates this information and other analyses of livestock population,
feeding practices, and production characteristics. For the current Inventory, CEFM results for 1990 through 2017
were carried over from the 1990 to 2017 Inventory (i.e., 2019 Inventory submission) to focus resources on CEFM
improvements, and a simplified approach was used to estimate 2018 and 2019 enteric emissions from cattle.
See Annex 3.10 for more detailed information on the methodology and data used to calculate CH4 emissions from
enteric fermentation. In addition, variables and the resulting emissions are also available at the state level in Annex
3.10.
1990 to 2017 Inventory Methodology for Cattle
National cattle population statistics were disaggregated into the following cattle sub-populations:
•	Dairy Cattle
o Calves
o Heifer Replacements
o Cows
•	Beef Cattle
o Calves
o Heifer Replacements
o Heifer and Steer Stockers
o Animals in Feedlots (Heifers and Steer)
o Cows
o Bulls
Calf birth rates, end-of-year population statistics, detailed feedlot placement information, and slaughter weight
data were used to create a transition matrix that models cohorts of individual animal types and their specific
emission profiles. The key variables tracked for each of the cattle population categories are described in Annex
3.10. These variables include performance factors such as pregnancy and lactation as well as average weights and
weight gain. Annual cattle population data were obtained from the U.S. Department of Agriculture's (USDA)
National Agricultural Statistics Service (NASS) QuickStats database (USDA 2016).
Diet characteristics were estimated by region for dairy, grazing beef, and feedlot beef cattle. These diet
characteristics were used to calculate digestible energy (DE) values (expressed as the percent of gross energy
intake digested by the animal) and CH4 conversion rates (Ym) (expressed as the fraction of gross energy converted
to CH4) for each regional population category. The IPCC recommends Ym ranges of 3.0±1.0 percent for feedlot
cattle and 6.5±1.0 percent for other well-fed cattle consuming temperate-climate feed types (IPCC 2006). Given
the availability of detailed diet information for different regions and animal types in the United States, DE and Ym
values unique to the United States were developed. The diet characterizations and estimation of DE and Ym values
were based on information from state agricultural extension specialists, a review of published forage quality
studies and scientific literature, expert opinion, and modeling of animal physiology.
The diet characteristics for dairy cattle were based on Donovan (1999) and an extensive review of nearly 20 years
of literature from 1990 through 2009. Estimates of DE were national averages based on the feed components of
the diets observed in the literature for the following year groupings: 1990 through 1993,1994 through 1998,1999
5-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
through 2003, 2004 through 2006, 2007, and 2008 onward.4 Base year Ym values by region were estimated using
Donovan (1999). As described in ERG (2016), a ruminant digestion model (COWPOLL, as selected in Kebreab et al.
2008) was used to evaluate Ym for each diet evaluated from the literature, and a function was developed to adjust
regional values over time based on the national trend. Dairy replacement heifer diet assumptions were based on
the observed relationship in the literature between dairy cow and dairy heifer diet characteristics.
For feedlot animals, the DE and Ym values used for 1990 were recommended by Johnson (1999). Values for DE and
Ym for 1991 through 1999 were linearly extrapolated based on the 1990 and 2000 data. DE and Ym values for 2000
onwards were based on survey data in Galyean and Gleghorn (2001) and Vasconcelos and Galyean (2007).
For grazing beef cattle, Ym values were based on Johnson (2002), DE values for 1990 through 2006 were based on
specific diet components estimated from Donovan (1999), and DE values from 2007 onwards were developed from
an analysis by Archibeque (2011), based on diet information in Preston (2010) and USDA-APHIS:VS (2010). Weight
and weight gains for cattle were estimated from Holstein (2010), Doren et al. (1989), Enns (2008), Lippke et al.
(2000), Pinchack et al. (2004), Platter et al. (2003), Skogerboe et al. (2000), and expert opinion. See Annex 3.10 for
more details on the method used to characterize cattle diets and weights in the United States.
Calves younger than 4 months are not included in emission estimates because calves consume mainly milk and the
IPCC recommends the use of a Ym of zero for all juveniles consuming only milk. Diets for calves aged 4 to 6 months
are assumed to go through a gradual weaning from milk decreasing to 75 percent at 4 months, 50 percent at age 5
months, and 25 percent at age 6 months. The portion of the diet made up with milk still results in zero emissions.
For the remainder of the diet, beef calf DE and Ym are set equivalent to those of beef replacement heifers, while
dairy calf DE is set equal to that of dairy replacement heifers and dairy calf Ym is provided at 4 and 7 months of age
by Soliva (2006). Estimates of Ym for 5 and 6 month old dairy calves are linearly interpolated from the values
provided for 4 and 7 months.
To estimate CH4 emissions, the population was divided into state, age, sub-type (i.e., dairy cows and replacements,
beef cows and replacements, heifer and steer stockers, heifers and steers in feedlots, bulls, beef calves 4 to 6
months, and dairy calves 4 to 6 months), and production (i.e., pregnant, lactating) groupings to more fully capture
differences in CH4 emissions from these animal types. The transition matrix was used to simulate the age and
weight structure of each sub-type on a monthly basis in order to more accurately reflect the fluctuations that
occur throughout the year. Cattle diet characteristics were then used in conjunction with Tier 2 equations from
IPCC (2006) to produce CH4 emission factors for the following cattle types: dairy cows, beef cows, dairy
replacements, beef replacements, steer stockers, heifer stockers, steer feedlot animals, heifer feedlot animals,
bulls, and calves. To estimate emissions from cattle, monthly population data from the transition matrix were
multiplied by the calculated emission factor for each cattle type. More details are provided in Annex 3.10.
2018 and 2019 Inventory Methodology for Cattle
As noted above, a simplified approach for cattle enteric emissions was used in lieu of the CEFM for 2018 and 2019
to focus resources on CEFM improvements. First, 2018 and 2019 populations for each of the CEFM cattle sub-
populations were estimated, then these populations were multiplied by the corresponding implied emission
factors developed from the CEFM for the 1990 to 2017 Inventory year. Dairy cow, beef cow, and bull populations
for 2019 were based on data directly from the USDA-NASS QuickStats database (USDA 2020, USDA 2019). Because
the remaining CEFM cattle sub-population categories do not correspond exactly to the remaining QuickStats cattle
categories, 2018 and 2019 populations for these categories were estimated by extrapolating the 2017 populations
based on percent changes from 2017 to 2018 and 2018 to 2019 in similar QuickStats categories, consistent with
Volume 1, Chapter 5 of the 2006 IPCC Guidelines on time-series consistency. Table 5-5 lists the QuickStats
categories used to estimate the percent change in population for each of the CEFM categories.
4 Due to inconsistencies in the 2003 literature values, the 2002 values were used for 2003 as well.
Agriculture 5-7

-------
Table 5-5: Cattle Sub-Population Categories for 2018 Population Estimates
CEFM Cattle Category
USDA-NASS QuickStats Cattle Category
Dairy Calves
Dairy Cows
Dairy Replacements 7-11 months
Dairy Replacements 12-23 months
Bulls
Beef Calves
Beef Cows
Beef Replacements 7-11 months
Beef Replacements 12-23 months
Steer Stockers
Heifer Stockers
Steer Feedlot
Heifer Feedlot
Cattle, Calves
Cattle, Cows, Milk
Cattle, Heifers, GE 500 lbs, Milk Replacement
Cattle, Heifers, GE 500 lbs, Milk Replacement
Cattle, Bulls, GE 500 lbs
Cattle, Calves
Cattle, Cows, Beef
Cattle, Heifers, GE 500 lbs, Beef Replacement
Cattle, Heifers, GE 500 lbs, Beef Replacement
Cattle, Steers, GE 500 lbs
Cattle, Heifers, GE 500 lbs, (Excl. Replacement)
Cattle, On Feed
Cattle, On Feed
Non-Cattle Livestock
Emission estimates for other animal types were based on average emission factors (Tier 1 default IPCC emission
factors) representative of entire populations of each animal type. Methane emissions from these animals
accounted for a minor portion of total CH4 emissions from livestock in the United States from 1990 through 2019.
Additionally, the variability in emission factors for each of these other animal types (e.g., variability by age,
production system, and feeding practice within each animal type) is less than that for cattle.
Annual livestock population data for 1990 to 2019 for sheep; swine; goats; horses; mules and asses; and American
bison were obtained for available years from USDA-NASS (USDA 2016). Horse, goat and mule and ass population
data were available for 1987,1992,1997, 2002, 2007, and 2012 (USDA 1992,1997, 2016); the remaining years
between 1990 and 2019 were interpolated and extrapolated from the available estimates (with the exception of
goat populations being held constant between 1990 and 1992). American bison population estimates were
available from USDA for 2002, 2007, and 2012 (USDA 2016) and from the National Bison Association (1999) for
1990 through 1999. Additional years were based on observed trends from the National Bison Association (1999),
interpolation between known data points, and extrapolation beyond 2012, as described in more detail in Annex
3.10.
Methane emissions from sheep, goats, swine, horses, American bison, and mules and asses were estimated by
using emission factors utilized in Crutzen et al. (1986, cited in IPCC 2006). These emission factors are
representative of typical animal sizes, feed intakes, and feed characteristics in developed countries. For American
bison, the emission factor for buffalo was used and adjusted based on the ratio of live weights to the 0.75 power.
The methodology is the same as that recommended by IPCC (2006).
A quantitative uncertainty analysis for this source category was performed using the IPCC-recommended Approach
2 uncertainty estimation methodology based on a Monte Carlo Stochastic Simulation technique as described in ICF
(2003). These uncertainty estimates were developed for the 1990 through 2001 Inventory (i.e., 2003 submission to
the UNFCCC). While there are plans to update the uncertainty to reflect recent methodological updates and
forthcoming changes (see Planned Improvements, below), at this time the uncertainty estimates were directly
applied to the 2019 emission estimates in this Inventory.
A total of 185 primary input variables (177 for cattle and 8 for non-cattle) were identified as key input variables for
the uncertainty analysis. A normal distribution was assumed for almost all activity- and emission factor-related
input variables. Triangular distributions were assigned to three input variables (specifically, cow-birth ratios for the
5-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019
Uncertainty and Time-Series Consistency

-------
three most recent years included in the 2001 model run) to ensure only positive values would be simulated. For
some key input variables, the uncertainty ranges around their estimates (used for inventory estimation) were
collected from published documents and other public sources; others were based on expert opinion and best
estimates. In addition, both endogenous and exogenous correlations between selected primary input variables
were modeled. The exogenous correlation coefficients between the probability distributions of selected activity-
related variables were developed through expert judgment.
Among the individual cattle sub-source categories, beef cattle account for the largest amount of CH4 emissions, as
well as the largest degree of uncertainty in the emission estimates—due mainly to the difficulty in estimating the
diet characteristics for grazing members of this animal group. Among non-cattle, horses represent the largest
percent of uncertainty in the previous uncertainty analysis because the Food and Agricultural Organization of the
United Nations (FAO) population estimates used for horses at that time had a higher degree of uncertainty than for
the USDA population estimates used for swine, goats, and sheep. The horse populations are now from the same
USDA source as the other animal types, and therefore the uncertainty range around horses is likely overestimated.
Cattle calves, American bison, mules and asses were excluded from the initial uncertainty estimate because they
were not included in emission estimates at that time.
The uncertainty ranges associated with the activity data-related input variables were plus or minus 10 percent or
lower. However, for many emission factor-related input variables, the lower- and/or the upper-bound uncertainty
estimates were over 20 percent. The results of the quantitative uncertainty analysis are summarized in Table 5-6.
Based on this analysis, enteric fermentation CH4 emissions in 2019 were estimated to be between 158.9 and 210.7
MMT C02 Eq. at a 95 percent confidence level, which indicates a range of 11 percent below to 18 percent above
the 2019 emission estimate of 178.6 MMT C02 Eq.
Table 5-6: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Enteric
Fermentation (MMT CO2 Eq. and Percent)


2019 Emission


Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3' b'c


(MMTCOz Eq.)
(MMT CO? Eq.)
(%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Enteric Fermentation
ch4
178.6
158.9 210.7
-11% +18%
a Range of emissions estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
b Note that the relative uncertainty range was estimated with respect to the 2001 emission estimates from the
2003 submission and applied to the 2019 estimates.
c The overall uncertainty calculated in 2003, and applied to the 2019 emission estimate, did not include uncertainty
estimates for calves, American bison, and mules and asses. Additionally, for bulls the emissions estimate was based
on the Tier 1 methodology. Since bull emissions are now estimated using the Tier 2 method, the uncertainty
surrounding their estimates is likely lower than indicated by the previous uncertainty analysis.
Methodological approaches, changes to historic data, and other parameters were applied to the entire time series
to ensure consistency in emissions estimates from 1990 through 2019. Details on the emission trends and
methodologies through time are described in more detail in the Introduction and Methodology sections.
QA/QC and Verification
In order to ensure the quality of the emission estimates from enteric fermentation, the General (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. Category-specific or Tier 2 QA procedures included
independent review of emission estimate methodologies from previous inventories.
Agriculture 5-9

-------
Over the past few years, particular importance has been placed on harmonizing the data exchange between the
enteric fermentation and manure management source categories. The current Inventory now utilizes the transition
matrix from the CEFM for estimating cattle populations and weights for both source categories, and the CEFM is
used to output volatile solids and nitrogen excretion estimates using the diet assumptions in the model in
conjunction with the energy balance equations from the IPCC (2006). This approach facilitates the QA/QC process
for both of these source categories. As noted in the Methodology discussion above, a simplified approach for cattle
enteric emissions was used in lieu of the CEFM for 2018 and 2019.
Recalculations Discussion
For sheep and goats, default national emission factors were updated to reflect revisions made in the 2019 IPCC
Refinement to the 2006 IPCC Guidelines and improve the accuracy of emissions. These revised emission factors
were applied to the entire time series and resulted in between 5 to 12 kt and 9 to 12 kt higher emissions for sheep
and goat livestock categories, respectively.
Planned Improvements
Regular annual data reviews and updates are necessary to maintain an emissions inventory that reflects the
current base of knowledge. EPA conducts the following list of regular annual assessments of data availability when
updating the estimates to extend time series each year:
•	Further research to improve the estimation of dry matter intake (as gross energy intake) using data from
appropriate production systems;
•	Updating input variables that are from older data sources, such as beef births by month, beef and dairy
annual calving rates, and beef cow lactation rates;
•	Investigating the availability of data for dairy births by month, to replace the current assumption that
births are evenly distributed throughout the year;
•	Updating the diet data to incorporate monthly or annual milk fat data in place of the fixed IPCC default
value of 4 percent milk fat. EPA has investigated the availability of data across the time series and plans to
incorporate annual U.S. milk fat values into the CEFM calculations in the next (i.e., 1990 to 2020)
Inventory, as opposed to using a default 4 percent milk fat across the entire time series;
•	Investigating the availability of annual data for the DE, Ym, and crude protein values of specific diet and
feed components for grazing and feedlot animals;
•	Further investigation on additional sources or methodologies for estimating DE for dairy cattle, given the
many challenges in characterizing dairy cattle diets;
•	Further evaluation of the assumptions about weights and weight gains for beef cows, such that trends
beyond 2007 are updated, rather than held constant;
•	Further evaluation of the estimated weight for dairy cows (i.e., 1,500 lbs) that is based solely on Holstein
cows as mature dairy cow weight is likely slightly overestimated, based on knowledge of the breeds of
dairy cows in the United States.
Depending upon the outcome of ongoing investigations, future improvement efforts for enteric fermentation
could include some of the following options which are additional to the regular updates, and may or may have
implications for regular updates once addressed:
•	Potentially updating to a Tier 2 methodology for other animal types (i.e., sheep, swine, goats, horses);
efforts to move to Tier 2 will consider the emissions significance of livestock types;
•	Investigation of methodologies and emission factors for including enteric fermentation emission
estimates from poultry;
5-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Comparison of the current CEFM processing of animal population data to estimates developed using
annual average populations to determine if the model could be simplified to use annual population data;
•	Comparison of the current CEFM with other models that estimate enteric fermentation emissions for
quality assurance and verification;
•	Investigation of recent research implications suggesting that certain parameters in enteric models may be
simplified without significantly diminishing model accuracy;
•	Recent changes that have been implemented to the CEFM warrant an assessment of the current
uncertainty analysis; therefore, a revision of the quantitative uncertainty surrounding emission estimates
from this source category will be initiated. EPA plans to perform this uncertainty analysis following the
completed updates to the CEFM; and
•	Analysis and integration of a more representative spatial distribution of animal populations by state,
particularly for poultry animal populations.
EPA received comments during recent Public Review periods of the Inventory regarding the CEFM model and data
and assumptions used to calculate enteric fermentation cattle emissions. Many of the comments received are
consistent with potential planned improvement options listed above. EPA is continuously investigating these
recommendations and potential improvements and working with USDA and other experts to utilize the best
available data and methods for estimating emissions. Many of these improvements are major updates and may
take multiple years to implement in full. In addition, EPA received comments during the Public Review period of
the current (1990 through 2019) and previous (1990 through 2018) Inventory regarding the use of alternate
metrics for weighting non-C02 emissions such as methane that differ from those required in reporting under the
UNFCCC to facilitate comparability as described in Box 5-1.
5.2 Manure Management (CRF Source
Category 3B)
The treatment, storage, and transportation of livestock manure can produce anthropogenic CH4 and N20
emissions.5 Methane is produced by the anaerobic decomposition of manure and nitrous oxide is produced from
direct and indirect pathways through the processes of nitrification and denitrification; in addition, there are many
underlying factors that can affect these resulting emissions from manure management, as described below.
When livestock manure is stored or treated in systems that promote anaerobic conditions (e.g., as a liquid/slurry in
lagoons, ponds, tanks, or pits), the decomposition of the volatile solids component in the manure tends to produce
CH4. When manure is handled as a solid (e.g., in stacks or drylots) or deposited on pasture, range, or paddock
lands, it tends to decompose aerobically and produce C02 and little or no CH4. Ambient temperature, moisture,
and manure storage or residency time affect the amount of CH4 produced because they influence the growth of
the bacteria responsible for CH4 formation. For non-liquid-based manure systems, moist conditions (which are a
function of rainfall and humidity) can promote CH4 production. Manure composition, which varies by animal diet,
growth rate, and animal type (particularly the different animal digestive systems), also affects the amount of CH4
produced. In general, the greater the energy content of the feed, the greater the potential for CH4 emissions.
However, some higher-energy feeds also are more digestible than lower quality forages, which can result in less
overall waste excreted from the animal.
As previously stated, N20 emissions are produced through both direct and indirect pathways. Direct N20 emissions
are produced as part of the nitrogen (N) cycle through the nitrification and denitrification of the N in livestock dung
5 C02 emissions from livestock are not estimated because annual net C02 emissions are assumed to be zero - the C02
photosynthesized by plants is returned to the atmosphere as respired C02 (IPCC 2006).
Agriculture 5-11

-------
and urine.6 There are two pathways for indirect N20 emissions. The first is the result of the volatilization of N in
manure (as NH3 and NOx) and the subsequent deposition of these gases and their products (NH4+ and N03") onto
soils and the surface of lakes and other waters. The second pathway is the runoff and leaching of N from manure
into the groundwater below, into riparian zones receiving drain or runoff water, or into the ditches, streams,
rivers, and estuaries into which the land drainage water eventually flows.
The production of direct N20 emissions from livestock manure depends on the composition of the manure
(manure includes both feces and urine), the type of bacteria involved in the process, and the amount of oxygen
and liquid in the manure system. For direct N20 emissions to occur, the manure must first be handled aerobically
where organic N is mineralized or decomposed to NH4 which is then nitrified to N03 (producing some N20 as a
byproduct) (nitrification). Next, the manure must be handled anaerobically where the nitrate is then denitrified to
N20 and N2 (denitrification). NOx can also be produced during denitrification (Groffman et al. 2000; Robertson and
Groffman 2015). These emissions are most likely to occur in dry manure handling systems that have aerobic
conditions, but that also contain pockets of anaerobic conditions due to saturation. A very small portion of the
total N excreted is expected to convert to N20 in the waste management system (WMS).
Indirect N20 emissions are produced when nitrogen is lost from the system through volatilization (as NH3 or NOx)
or through runoff and leaching. The vast majority of volatilization losses from these operations are NH3. Although
there are also some small losses of NOx, there are no quantified estimates available for use, so losses due to
volatilization are only based on NH3 loss factors. Runoff losses would be expected from operations that house
animals or store manure in a manner that is exposed to weather. Runoff losses are also specific to the type of
animal housed on the operation due to differences in manure characteristics. Little information is known about
leaching from manure management systems as most research focuses on leaching from land application systems.
Since leaching losses are expected to be minimal, leaching losses are coupled with runoff losses and the
runoff/leaching estimate provided in this chapter does not account for any leaching losses.
Estimates of CH4 emissions from manure management in 2019 were 62.4 MMT C02 Eq. (2,495 kt); in 1990,
emissions were 37.1 MMT C02 Eq. (1,485 kt). This represents a 68 percent increase in emissions from 1990.
Emissions increased on average by 0.8 MMT C02 Eq. (2 percent) annually over this period. The majority of this
increase is due to swine and dairy cow manure, where emissions increased 49 and 117 percent, respectively. From
2018 to 2019, there was a 1 percent increase in total CH4 emissions from manure management, due to an increase
in animal populations.
Although a large quantity of managed manure in the United States is handled as a solid, producing little CH4, the
general trend in manure management, particularly for dairy cattle and swine (which are both shifting towards
larger facilities), is one of increasing use of liquid systems. Also, new regulations controlling the application of
manure nutrients to land have shifted manure management practices at smaller dairies from daily spread systems
to storage and management of the manure on site. In many cases, manure management systems with the most
substantial methane emissions are those associated with confined animal management operations where manure
is handled in liquid-based systems. Nitrous oxide emissions from manure management vary significantly between
the types of management system used and can also result in indirect emissions due to other forms of nitrogen loss
from the system (IPCC 2006).
While national dairy animal populations have decreased since 1990, some states have seen increases in their dairy
cattle populations as the industry becomes more concentrated in certain areas of the country and the number of
animals contained on each facility increases. These areas of concentration, such as California, New Mexico, and
Idaho, tend to utilize more liquid-based systems to manage (flush or scrape) and store manure. Thus, the shift
toward larger dairy cattle and swine facilities since 1990 has translated into an increasing use of liquid manure
management systems, which have higher potential CH4 emissions than dry systems. This significant shift in both
6 Direct and indirect N20 emissions from dung and urine spread onto fields either directly as daily spread or after it is removed
from manure management systems (i.e., lagoon, pit, etc.) and from livestock dung and urine deposited on pasture, range, or
paddock lands are accounted for and discussed in the Agricultural Soil Management source category within the Agriculture
sector.
5-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
the dairy cattle and swine industries was accounted for by incorporating state and WMS-specific CH4 conversion
factor (MCF) values in combination with the 1992,1997, 2002, 2007, 2012, and 2017 farm-size distribution data
reported in the U.S. Department of Agriculture (USDA) Census of Agriculture (USDA 2019d).
In 2019, total N20 emissions from manure management were estimated to be 19.6 MMT C02 Eq. (66 kt); in 1990,
emissions were 14.0 MMT C02 Eq. (47 kt). These values include both direct and indirect N20 emissions from
manure management. Nitrous oxide emissions have increased since 1990. Small changes in N20 emissions from
individual animal groups exhibit the same trends as the animal group populations, with the overall net effect that
N20 emissions showed a 40 percent increase from 1990 to 2019 and a 0.9 percent increase from 2018 through
2019. Overall shifts toward liquid systems have driven down the emissions per unit of nitrogen excreted as dry
manure handling systems have greater aerobic conditions that promote N20 emissions.
Table 5-7 and Table 5-8 provide estimates of CH4 and N20 emissions from manure management by animal
category.7
Table 5-7: ChU and N2O Emissions from Manure Management (MMT CO2 Eq.)
Gas/Animal Type
1990
2005
2015
2016
2017
2018
2019
CH4a
37.1
51.6
57.9
59.6
59.9
61.7
62.4
Dairy Cattle
14.7
24.3
30.8
31.5
31.8
32.3
32.0
Swine
15.5
20.3
20.2
21.1
21.0
22.2
23.1
Poultry
3.3
3.2
3.4
3.4
3.4
3.5
3.6
Beef Cattle
3.1
3.3
3.1
3.3
3.4
3.4
3.4
Horses
0.2
0.3
0.2
0.2
0.2
0.2
0.2
Sheep
0.2
0.1
0.1
0.1
0.1
0.1
0.1
Goats
+
+
+
+
+
+
+
American Bison
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
N2Ob
14.0
16.4
17.5
18.1
18.7
19.4
19.6
Beef Cattle
5.9
7.2
7.7
8.1
8.6
9.2
9.4
Dairy Cattle
5.3
5.5
6.0
6.1
6.1
6.1
6.1
Swine
1.2
1.6
1.8
1.9
2.0
2.0
2.1
Poultry
1.4
1.6
1.6
1.6
1.6
1.7
1.7
Sheep
0.1
0.3
0.3
0.3
0.3
0.3
0.3
Horses
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Goats
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
American Bisonc
NA
NA
NA
NA
NA
NA
NA
Total
51.1
67.9
75.4
77.7
78.5
81.1
82.0
Notes: N20 emissions from manure deposited on pasture, range and paddock are included in the
Agricultural Soils Management sector. Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
NA (Not Available)
a Accounts for CH4 reductions due to capture and destruction of CH4 at facilities using anaerobic
digesters.
b Includes both direct and indirect N20 emissions.
c There are no American bison N20 emissions from managed systems; American bison are
maintained entirely on pasture, range, and paddock.
7 Manure management emissions from camels are not estimated because there is no significant population of camels in the
United States. Given the insignificance of estimated camel emissions in terms of the overall level and trend in national
emissions, there are no immediate improvement plans to include this emissions category in the Inventory. See Annex 5 for
more information on significance of estimated camel emissions.
Agriculture 5-13

-------
Table 5-8: ChU and N2O Emissions from Manure Management (kt)
Gas/Animal Type
1990
2005
2015
2016
2017
2018
2019
CH4a
1,485
2,062
2,316
2,385
2,395
2,467
2,495
Dairy Cattle
589
970
1,233
1,259
1,270
1,292
1,281
Swine
622
812
808
846
840
888
924
Poultry
131
129
136
136
137
141
142
Beef Cattle
126
133
126
132
136
135
136
Horses
9
12
8
8
7
7
7
Sheep
7
3
3
3
3
3
3
Goats
1
1
1
1
1
1
1
American Bison
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
N2Ob
47
55
59
61
63
65
66
Beef Cattle
20
24
26
27
29
31
31
Dairy Cattle
18
18
20
20
20
21
20
Swine
4
5
6
6
7
7
7
Poultry
5
5
5
5
5
6
6
Sheep
+
1
1
1
1
1
1
Horses
+
+
+
+
+
+
+
Goats
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
American Bisonc
NA
NA
NA
NA
NA
NA
NA
Notes: N20 emissions from manure deposited on pasture, range and paddock are included in the
Agricultural Soils Management sector. Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
NA (Not Available)
a Accounts for CH4 reductions due to capture and destruction of CH4 at facilities using anaerobic
digesters.
b Includes both direct and indirect N20 emissions.
cThere are no American bison N20 emissions from managed systems; American bison are
maintained entirely on pasture, range, and paddock.
Methodology
The methodologies presented in IPCC (2006) form the basis of the CH4 and N20 emission estimates for each animal
type, including Tier 1, Tier 2, and use of the CEFM previously described for Enteric Fermentation. This combination
of Tier 1 and Tier 2 methods was applied to all livestock animal types. This section presents a summary of the
methodologies used to estimate CH4 and N20 emissions from manure management. For the current Inventory,
time-series results were carried over from the 1990 to 2018 Inventory (i.e., 2020 submission) and a simplified
approach was used to estimate manure management emissions for 2019.
See Annex 3.11 for more detailed information on the methodology (including detailed formulas and emission
factors), data used to calculate CH4 and N20 emissions, and emission results (including input variables and results
at the state-level) from manure management.
Methane Calculation Methods
The following inputs were used in the calculation of manure management CH4 emissions for 1990 through 2018:
•	Animal population data (by animal type and state);
•	Typical animal mass (TAM) data (by animal type);
•	Portion of manure managed in each WMS, by state and animal type;
•	Volatile solids (VS) production rate (by animal type and state or United States);
•	Methane producing potential (B0) of the volatile solids (by animal type); and
5-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Methane conversion factors (MCF), the extent to which the CH4 producing potential is realized for each
type of WMS (by state and manure management system, including the impacts of any biogas collection
efforts).
Methane emissions were estimated by first determining activity data, including animal population, TAM, WMS
usage, and waste characteristics. The activity data sources are described below:
•	Annual animal population data for 1990 through 2018 for all livestock types, except goats, horses, mules
and asses, and American bison were obtained from the USDA-NASS. For cattle, the USDA populations
were utilized in conjunction with birth rates, detailed feedlot placement information, and slaughter
weight data to create the transition matrix in the Cattle Enteric Fermentation Model (CEFM) that models
cohorts of individual animal types and their specific emission profiles. The key variables tracked for each
of the cattle population categories are described in Section 5.1 and in more detail in Annex 3.10. Goat
population data for 1992,1997, 2002, 2007, 2012, and 2017; horse and mule and ass population data for
1987,1992,1997, 2002, 2007, 2012, and 2017; and American bison population for 2002, 2007, 2012, and
2017 were obtained from the Census of Agriculture (USDA 2019d). American bison population data for
1990 through 1999 were obtained from the National Bison Association (1999).
•	The TAM is an annual average weight that was obtained for animal types other than cattle from
information in USDA's Agricultural Waste Management Field Handbook (USDA 1996), the American
Society of Agricultural Engineers, Standard D384.1 (ASAE 1998) and others (Meagher 1986; EPA 1992;
Safley 2000; ERG 2003b; IPCC 2006; ERG 2010a). For a description of the TAM used for cattle, see Annex
3.10.
•	WMS usage was estimated for swine and dairy cattle for different farm size categories using state and
regional data from USDA (USDA APHIS 1996; Bush 1998; Ott 2000; USDA 2016c) and EPA (ERG 2000a; EPA
2002a and 2002b; ERG 2018, ERG 2019). For beef cattle and poultry, manure management system usage
data were not tied to farm size but were based on other data sources (ERG 2000a; USDA APHIS 2000; UEP
1999). For other animal types, manure management system usage was based on previous estimates (EPA
1992). American bison WMS usage was assumed to be the same as not on feed (NOF) cattle, while mules
and asses were assumed to be the same as horses.
•	VS production rates for all cattle except for calves were calculated by head for each state and animal type
in the CEFM. VS production rates by animal mass for all other animals were determined using data from
USDA's Agricultural Waste Management Field Handbook (USDA 1996 and 2008; ERG 2010b and 2010c)
and data that was not available in the most recent Handbook were obtained from the American Society of
Agricultural Engineers, Standard D384.1 (ASAE 1998) or the 2006 IPCC Guidelines (IPCC 2006). American
bison VS production was assumed to be the same as NOF bulls.
•	Bo was determined for each animal type based on literature values (Morris 1976; Bryant et al. 1976;
Hashimoto 1981; Hashimoto 1984; EPA 1992; Hill 1982; Hill 1984).
•	MCFs for dry systems were set equal to default IPCC factors based on state climate for each year (IPCC
2006). MCFs for liquid/slurry, anaerobic lagoon, and deep pit systems were calculated based on the
forecast performance of biological systems relative to temperature changes as predicted in the van't Hoff-
Arrhenius equation which is consistent with IPCC (2006) Tier 2 methodology.
•	Data from anaerobic digestion systems with CH4 capture and combustion were obtained from the EPA
AgSTAR Program, including information available in the AgSTAR project database (EPA 2019). Anaerobic
digester emissions were calculated based on estimated methane production and collection and
destruction efficiency assumptions (ERG 2008).
•	For all cattle except for calves, the estimated amount of VS (kg per animal-year) managed in each WMS
for each animal type, state, and year were taken from the CEFM, assuming American bison VS production
to be the same as NOF bulls. For animals other than cattle, the annual amount of VS (kg per year) from
manure excreted in each WMS was calculated for each animal type, state, and year. This calculation
multiplied the animal population (head) by the VS excretion rate (kg VS per 1,000 kg animal mass per
Agriculture 5-15

-------
day), the TAM (kg animal mass per head) divided by 1,000, the WMS distribution (percent), and the
number of days per year (365.25).
The estimated amount of VS managed in each WMS was used to estimate the CH4 emissions (kg CH4 per year) from
each WMS. The amount of VS (kg per year) were multiplied by the B0 (m3 CH4 per kg VS), the MCF for that WMS
(percent), and the density of CH4 (kg CH4per m3 CH4). The CH4 emissions for each WMS, state, and animal type
were summed to determine the total U.S. CH4 emissions. See details in Step 5 of Annex 3.11.
The following approach was used in the calculation of manure management CH4 emissions for 2019:
•	EPA obtained 2019 national-level animal population data: Sheep, poultry, and swine data were
downloaded from USDA-NASS Quickstats (USDA 2020). Cattle populations were obtained from the CEFM
(see NIR Section 5.1 and Annex 3.10). Data for goats, horses, bison, mules, and asses were extrapolated
based on the 2009 through 2018 population values to reflect recent trends in animal populations.
•	EPA multiplied the national populations by the animal-specific 2018 implied emission factors8 for CH4 to
calculate national-level 2019 CH4 emissions estimates by animal type. These methods were utilized in
order to maintain time-series consistency as referenced in Volume 1, Chapter 5 of the 2006IPCC
Guidelines.
Nitrous Oxide Calculation Methods
The following inputs were used in the calculation of direct and indirect manure management N20 emissions for
1990 through 2018:
•	Animal population data (by animal type and state);
•	TAM data (by animal type);
•	Portion of manure managed in each WMS (by state and animal type);
•	Total Kjeldahl N excretion rate (Nex);
•	Direct N20 emission factor (EFwms);
•	Indirect N20 emission factor for volatilization (EFVOiatiiization);
•	Indirect N20 emission factor for runoff and leaching (EFrun0ff/ieach);
•	Fraction of N loss from volatilization of NH3 and NOx (Fracgas); and
•	Fraction of N loss from runoff and leaching (Fracrunoff/ieach).
Nitrous oxide emissions were estimated by first determining activity data, including animal population, TAM, WMS
usage, and waste characteristics. The activity data sources (except for population, TAM, and WMS, which were
described above) are described below:
•	Nex for all cattle except for calves were calculated by head for each state and animal type in the CEFM.
Nex rates by animal mass for all other animals were determined using data from USDA's Agricultural
Waste Management Field Handbook (USDA 1996 and 2008; ERG 2010b and 2010c) and data from the
American Society of Agricultural Engineers, Standard D384.1 (ASAE 1998) and IPCC (2006). American bison
Nex were assumed to be the same as NOF bulls.9
•	All N20 emission factors (direct and indirect) were taken from IPCC (2006).
8	An implied emission factor is defined as emissions divided by the relevant measure of activity; the implied emission factor is
equal to emissions per activity data unit. For source/sink categories that are composed of several subcategories, the emissions
and activity data are summed up across all subcategories. Hence, the implied emission factors are generally not equivalent to
the emission factors used to calculate emission estimates, but are average values that could be used, with caution, in data
comparisons (UNFCCC 2017).
9	Nex of American bison on grazing lands are accounted for and discussed in the Agricultural Soil Management source category
and included under pasture, range and paddock (PRP) emissions. Because American bison are maintained entirely on
unmanaged WMS and N20 emissions from unmanaged WMS are not included in the Manure Management source category,
there are no N20 emissions from American bison included in the Manure Management source category.
5-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Country-specific estimates for the fraction of N loss from volatilization (Fracgas) and runoff and leaching
(Fracrunoff/ieach) were developed. Fracgas values were based on WMS-specific volatilization values as
estimated from EPA's National Emission Inventory - Ammonia Emissions from Animal Agriculture
Operations (EPA 2005). Fracrunoff/ieaching values were based on regional cattle runoff data from EPA's Office
of Water (EPA 2002b; see Annex 3.11).
To estimate N20 emissions for cattle (except for calves), the estimated amount of N excreted (kg per animal-year)
that is managed in each WMS for each animal type, state, and year were taken from the CEFM. For calves and
other animals, the amount of N excreted (kg per year) in manure in each WMS for each animal type, state, and
year was calculated. The population (head) for each state and animal was multiplied by TAM (kg animal mass per
head) divided by 1,000, the nitrogen excretion rate (Nex, in kg N per 1,000 kg animal mass per day), WMS
distribution (percent), and the number of days per year.
Direct N20 emissions were calculated by multiplying the amount of N excreted (kg per year) in each WMS by the
N20 direct emission factor for that WMS (EFwms, in kg N20-N per kg N) and the conversion factor of N20-N to N20.
These emissions were summed over state, animal, and WMS to determine the total direct N20 emissions (kg of
N20 per year). See details in Step 6 of Annex 3.11.
Indirect N20 emissions from volatilization (kg N20 per year) were then calculated by multiplying the amount of N
excreted (kg per year) in each WMS by the fraction of N lost through volatilization (Fracgas) divided by 100, the
emission factor for volatilization (EFVO|atiiization, in kg N20 per kg N), and the conversion factor of N20-N to N20.
Indirect N20 emissions from runoff and leaching (kg N20 per year) were then calculated by multiplying the amount
of N excreted (kg per year) in each WMS by the fraction of N lost through runoff and leaching (Fracrunoff/ieach)
divided by 100, and the emission factor for runoff and leaching (EFrunoff/ieach, in kg N20 per kg N), and the conversion
factor of N20-N to N20. The indirect N20 emissions from volatilization and runoff and leaching were summed to
determine the total indirect N20 emissions. See details in Step 6 of Annex 3.11.
Following these steps, direct and indirect N20 emissions were summed to determine total N20 emissions (kg N20
per year) for the years 1990 to 2018.
The following approach was used in the calculation of manure management N20 emissions for 2019:
•	EPA obtained 2019 national-level animal population data: Sheep, poultry, and swine data were
downloaded from USDA-NASS Quickstats (USDA 2020). Cattle populations were obtained from the CEFM,
see Section 5.1 and Annex 3.10 (Enteric Fermentation). Data for goats, horses, bison, mules, and asses
were extrapolated based on the 2009 through 2018 population values to reflect recent trends in animal
populations.
•	The national populations were multiplied by the animal-specific 2018 implied emission factors for N20
(which combines both direct and indirect N20) to calculate national-level 2019 N20 emissions estimates
by animal type. These methods were utilized in order to maintain time-series consistency as referenced in
Volume 1, Chapter 5 of the 2006IPCC Guidelines.
Uncertainty and Time-Series Consistency
An analysis (ERG 2003a) was conducted for the manure management emission estimates presented in the 1990
through 2001 Inventory (i.e., 2003 submission to the UNFCCC) to determine the uncertainty associated with
estimating CH4 and N20 emissions from livestock manure management. The quantitative uncertainty analysis for
this source category was performed in 2002 through the IPCC-recommended Approach 2 uncertainty estimation
methodology, the Monte Carlo Stochastic Simulation technique. The uncertainty analysis was developed based on
the methods used to estimate CH4 and N20 emissions from manure management systems. A normal probability
distribution was assumed for each source data category. The series of equations used were condensed into a single
equation for each animal type and state. The equations for each animal group contained four to five variables
around which the uncertainty analysis was performed for each state. While there are plans to update the
uncertainty to reflect recent manure management updates and forthcoming changes (see Planned Improvements,
below), at this time the uncertainty estimates were directly applied to the 2019 emission estimates.
Agriculture 5-17

-------
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 5-9. Manure management
CH4 emissions in 2019 were estimated to be between 51.1 and 74.8 MMT C02 Eq. at a 95 percent confidence level,
which indicates a range of 18 percent below to 20 percent above the actual 2019 emission estimate of 62.4 MMT
C02 Eq. At the 95 percent confidence level, N20 emissions were estimated to be between 16.5 and 24.3 MMT C02
Eq. (or approximately 16 percent below and 24 percent above the actual 2019 emission estimate of 19.6 MMT C02
Eq.).
Table 5-9: Approach 2 Quantitative Uncertainty Estimates for CH4 and N2O (Direct and
Indirect) Emissions from Manure Management (MMT CO2 Eq. and Percent)


2019 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMTCOz Eq.)
(MMT CO?
Eq.)
(%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Manure Management
ch4
62.4
51.1
74.8
-18% +20%
Manure Management
n2o
19.6
16.5
24.3
-16% +24%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Methodological approaches, changes to historic data, and other parameters were applied to the entire time series
to ensure consistency in emissions estimates from 1990 through 2019. Details on the emission trends and
methodologies through time are described in more detail in the Introduction and Methodology sections.
QA/QC and Verification
General (Tier 1) and category-specific (Tier 2) QA/QC activities were conducted consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. Tier 2 activities focused on comparing estimates for the previous and current
Inventories for N20 emissions from managed systems and CH4 emissions from livestock manure. All errors
identified were corrected. Order of magnitude checks were also conducted, and corrections made where needed.
In addition, manure N data were checked by comparing state-level data with bottom-up estimates derived at the
county level and summed to the state level. Similarly, a comparison was made by animal and WMS type for the full
time series, between national level estimates for N excreted and the sum of county estimates for the full time
series.
Time-series data, including population, are validated by experts to ensure they are representative of the best
available U.S.-specific data. The U.S.-specific values for TAM, Nex, VS, B0, and MCF were also compared to the IPCC
default values and validated by experts. Although significant differences exist in some instances, these differences
are due to the use of U.S.-specific data and the differences in U.S. agriculture as compared to other countries. The
U.S. manure management emission estimates use the most reliable country-specific data, which are more
representative of U.S. animals and systems than the IPCC (2006) default values.
For additional verification of the 1990 to 2018 estimates, the implied CH4 emission factors for manure
management (kg of CH4 per head per year) were compared against the default IPCC (2006) values.10 Table 5-10
presents the implied emission factors of kg of CH4 per head per year used for the manure management emission
estimates as well as the IPCC (2006) default emission factors. The U.S. implied emission factors fall within the
range of the IPCC (2006) default values, except in the case of sheep, goats, and some years for horses and dairy
cattle. The U.S. implied emission factors are greater than the IPCC (2006) default value for those animals due to
the use of U.S.-specific data for typical animal mass and VS excretion. There is an increase in implied emission
factors for dairy cattle and swine across the time series. This increase reflects the dairy cattle and swine industry
10 CH4 implied emission factors were not calculated for 2019 due to the simplified emissions estimation approach used to
estimate emissions for that year; therefore, those values are consistent with 2018.
5-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
trend towards larger farm sizes; large farms are more likely to manage manure as a liquid and therefore produce
more CH4 emissions.
Table 5-10: IPCC (2006) Implied Emission Factor Default Values Compared with Calculated
Values for ChU from Manure Management (kg/head/year)

IPCC Default







Animal Type
CH4 Emission
Factors

Implied CH4 Emission Factors (kg/head/year)


fke/head/vear)a
1990
2005
2015
2016
2017
2018
2019
Dairy Cattle
48-112
30.2
54.5
65.6
66.8
67.2
67.9
67.9
Beef Cattle
1-2
1.5
1.6
1.7
1.7
1.7
1.6
1.6
Swine
10-45
11.5
13.3
11.8
12.1
11.7
12.0
12.0
Sheep
0.19-0.37
0.6
0.6
0.5
0.5
0.5
0.5
0.5
Goats
0.13-0.26
0.4
0.3
0.3
0.3
0.3
0.3
0.3
Poultry
0.02-1.4
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Horses
1.56-3.13
4.3
3.1
2.6
2.6
2.6
2.6
2.6
American Bison
NA
1.8
2.0
2.1
2.1
2.1
2.1
2.1
Mules and Asses
0.76-1.14
0.9
1.0
1.0
1.0
1.0
1.0
1.0
Note: CH4 implied emission factors were not calculated for 2019 due to the simplified emissions estimation
approach used to estimate emissions for that year. 2018 values were used for 2019.
NA (Not Applicable)
a Ranges reflect 2006 IPCC Guidelines (Volume 4, Table 10.14) default emission factors for North America across
different climate zones.
In addition, default IPCC (2006) emission factors for N20 were compared to the U.S. Inventory implied N20
emission factors. Default N20 emission factors from the 2006 IPCC Guidelines were used to estimate N20 emission
from each WMS in conjunction with U.S.-specific Nex values. The implied emission factors differed from the U.S.
Inventory values due to the use of U.S.-specific Nex values and differences in populations present in each WMS
throughout the time series.
Recalculations Discussion
No recalculations were performed for the 1990 to 2018 estimates. The 2019 estimates were developed using a
simplified approach, as discussed in the Methodology section.
Planned Improvements
Regular annual data reviews and updates are necessary to maintain an emissions inventory that reflects the
current base of knowledge. EPA conducts the following list of regular annual assessments of data availability when
updating the estimates to extend time series each year. EPA is actively pursuing the following updates but notes
that implementation may be based on available resources and data availability:
•	Continuing to investigate new sources of WMS data. EPA is working with the USDA Natural Resources
Conservation Service to collect data for potential improvements to the Inventory. EPA expects the next
WMS systems to be updated for the next (i.e., 1990 to 2020) Inventory submission include poultry and
beef cattle.
•	Updating the B0 data used in the Inventory, as data become available. EPA is conducting outreach with
counterparts from USDA as to available data and research on B0.
•	Revising the methodology for population distribution to states where USDA population data are withheld
due to disclosure concerns. These updates will be made in collaboration with the EPA National Emissions
Inventory staff to improve consistency across U.S. inventories. EPA plans to incorporate these updates
into the next (i.e., 1990 to 2020) Inventory submission.
Agriculture 5-19

-------
IPCC's 2019 Refinement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories provides updated
emission factors that EPA plans to review and implement for manure management (IPCC 2019). EPA maintains
from previous reports that many of the improvements identified below are major updates and may take multiple
years to fully implement. Potential improvements (long-term improvements) for future Inventory years include:
•	Revising the anaerobic digestion estimates to estimate CH4 emissions reductions due to the use of
anaerobic digesters (the Inventory currently estimates only emissions from anaerobic digestion systems).
•	Investigating improved emissions estimate methodologies for swine pit systems with less than one month
of storage (the recently updated swine WMS data included this WMS category).
•	Comparing CH4 and N20 emission estimates with estimates from other models and more recent studies
and compare the results to the Inventory.
•	Comparing manure management emission estimates with on-farm measurement data to identify
opportunities for improved estimates.
•	Comparing VS and Nex data to literature data to identify opportunities for improved estimates.
•	Improving collaboration with the Enteric Fermentation source category estimates. For future inventories,
it may be beneficial to have the CEFM and Manure Management calculations in the same model, as they
rely on much of the same activity data and they depend on each other's outputs to properly calculate
emissions.
•	Revising the uncertainty analysis to address changes that have been implemented to the CH4 and N20
estimates. EPA plans to line up the timing of performing the updated Manure Management uncertainty
analysis with the uncertainty analysis for Enteric Fermentation.
5.3 Rice Cultivation (CRF Source Category 3C)
Most of the world's rice is grown on flooded fields (Baicich 2013) that create anaerobic conditions leading to CH4
production through a process known as methanogenesis. Approximately 60 to 90 percent of the CH4 produced by
methanogenic bacteria in flooded rice fields is oxidized in the soil and converted to C02 by methanotrophic
bacteria. The remainder is emitted to the atmosphere (Holzapfel-Pschorn et al. 1985; Sass et al. 1990) or
transported as dissolved CH4 into groundwater and waterways (Neue et al. 1997). Methane is transported to the
atmosphere primarily through the rice plants, but some CH4 also escapes via ebullition (i.e., bubbling through the
water) and to a much lesser extent by diffusion through the water (van Bodegom et al. 2001).
Water management is arguably the most important factor affecting CH4 emissions in rice cultivation, and improved
water management has the largest potential to mitigate emissions (Yan et al. 2009). Upland rice fields are not
flooded, and therefore do not produce CH4, but large amounts of CH4can be emitted in continuously irrigated
fields, which is the most common practice in the United States (USDA 2012). Single or multiple aeration events
with drainage of a field during the growing season can significantly reduce these emissions (Wassmann et al.
2000a), but drainage may also increase N20 emissions. Deepwater rice fields (i.e., fields with flooding depths
greater than one meter, such as natural wetlands) tend to have fewer living stems reaching the soil, thus reducing
the amount of CH4 transport to the atmosphere through the plant compared to shallow-flooded systems (Sass
2001).
Other management practices also influence CH4 emissions from flooded rice fields including rice residue straw
management and application of organic amendments, in addition to cultivar selection due to differences in the
5-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
amount of root exudates11 among rice varieties (Neue et al. 1997). These practices influence the amount of
organic matter available for methanogenesis, and some practices, such as mulching rice straw or composting
organic amendments, can reduce the amount of labile carbon and limit CH4 emissions (Wassmann et al. 2000b).
Fertilization practices also influence CH4 emissions, particularly the use of fertilizers with sulfate (Wassmann et al.
2000b; Linquist et al. 2012), which can reduce CH4 emissions. Other environmental variables also impact the
methanogenesis process such as soil temperature and soil type. Soil temperature regulates the activity of
methanogenic bacteria, which in turn affects the rate of CH4 production. Soil texture influences decomposition of
soil organic matter, but is also thought to have an impact on oxidation of CH4 in the soil (Sass et al. 1994).
Rice is currently cultivated in thirteen states, including Arkansas, California, Florida, Illinois, Kentucky, Louisiana,
Minnesota, Mississippi, Missouri, New York, South Carolina, Tennessee and Texas. Soil types, rice varieties, and
cultivation practices vary across the United States, but most farmers apply fertilizers and do not harvest crop
residues. In addition, a second, ratoon rice crop is grown in the Southeastern region of the country. Ratoon crops
are produced from regrowth of the stubble remaining after the harvest of the first rice crop. Methane emissions
from ratoon crops are higher than those from the primary crops due to the increased amount of labile organic
matter available for anaerobic decomposition in the form of relatively fresh crop residue straw. Emissions tend to
be higher in rice fields if the residues have been in the field for less than 30 days before planting the next rice crop
(Lindau and Bollich 1993; IPCC 2006; Wang et al. 2013).
A combination of Tier 1 and 3 methods are used to estimate CH4 emissions from rice cultivation across most of the
time series, while a surrogate data method has been applied to estimate national emissions for 2016 to 2019 in
this Inventory due to lack of data in the later years of the time series. National emission estimates based on
surrogate data will be recalculated in a future Inventory with the Tier 1 and 3 methods as data becomes available.
Overall, rice cultivation is a minor source of CH4 emissions in the United States relative to other source categories
(see Table 5-11, Table 5-12, and Figure 5-3). Most emissions occur in Arkansas, California, Louisiana, Mississippi,
Missouri and Texas. In 2019, CH4 emissions from rice cultivation were 15.1 MMT C02 Eq. (602 kt). Annual emissions
fluctuate between 1990 and 2019, which is largely due to differences in the amount of rice harvested areas over
time, which has been decreasing over the past two decades. Consequently, emissions in 2019 are six percent lower
than emissions in 1990.
Table 5-11: ChU Emissions from Rice Cultivation (MMT CO2 Eq.)
State
1990
2005
2015
2016
2017
2018
2019
Arkansas
5.4
7.9
6.4
NE
NE
NE
NE
California
3.3
3.4
4.1
NE
NE
NE
NE
Florida
+
+
+
NE
NE
NE
NE
Illinois
+
+
+
NE
NE
NE
NE
Kentucky
+
+
+
NE
NE
NE
NE
Louisiana
2.6
2.8
2.6
NE
NE
NE
NE
Minnesota
+
0.1
+
NE
NE
NE
NE
Mississippi
1.1
1.4
1.0
NE
NE
NE
NE
Missouri
0.6
1.1
0.7
NE
NE
NE
NE
New York
+
+
+
NE
NE
NE
NE
South Carolina
+
+
+
NE
NE
NE
NE
Tennessee
+
+
+
NE
NE
NE
NE
Texas
3.0
1.3
1.4
NE
NE
NE
NE
Total
16.0
18.0
16.2
15.8
14.9
15.6
15.1
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
11 The roots of rice plants add organic material to the soil through a process called "root exudation." Root exudation is thought
to enhance decomposition of the soil organic matter and release nutrients that the plant can absorb and use to stimulate more
production. The amount of root exudate produced by a rice plant over a growing season varies among rice varieties.
Agriculture 5-21

-------
NE (Not Estimated). State-level emissions are not estimated for 2016 through 2019 in this Inventory
because data are unavailable. A surrogate data method is used to estimate emissions for these years and
are produced only at the national scale.
Table 5-12: CHU Emissions from Rice Cultivation (kt)
State
1990

2005

2015
2016
2017
2018
2019
Arkansas
216

315

256
NE
NE
NE
NE
California
131

134

166
NE
NE
NE
NE
Florida
+

1

+
NE
NE
NE
NE
Illinois
+

+

+
NE
NE
NE
NE
Kentucky
+

+

+
NE
NE
NE
NE
Louisiana
103

113

103
NE
NE
NE
NE
Minnesota
1

2

+
NE
NE
NE
NE
Mississippi
45

55

40
NE
NE
NE
NE
Missouri
22

45

26
NE
NE
NE
NE
New York
+

+

+
NE
NE
NE
NE
South Carolina
+

+

+
NE
NE
NE
NE
Tennessee
+

+

+
NE
NE
NE
NE
Texas
122

54

57
NE
NE
NE
NE
Total
640

720

648
631
596
623
602
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
NE (Not Estimated). State-level emissions are not estimated for 2016 through 2019 in this Inventory
because data are unavailable. A surrogate data method is used to estimate emissions for these years and
are produced only at the national scale.
Figure 5-3: Annual CHU Emissions from Rice Cultivation, 2015
MT C02 Eq. ha1 yr
~ < 5
1 5 to 10
J 10 to 15
¦ 15 to 20
¦ >20
Note: Only national-scale emissions are estimated for 2016 through 2019 in this Inventory using the surrogate data method
5-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
described in the Methodology section; therefore, the fine-scale emission patterns in this map are based on the estimates for
2015.
Methodology
The methodology used to estimate CH4 emissions from rice cultivation is based on a combination of IPCC Tier 1 and
3 approaches. The Tier 3 method utilizes the DayCent process-based model to estimate CH4 emissions from rice
cultivation (Cheng et al. 2013), and has been tested in the United States (see Annex 3.12) and Asia (Cheng et al.
2013, 2014). The model simulates hydrological conditions and thermal regimes, organic matter decomposition,
root exudation, rice plant growth and its influence on oxidation of CH4, as well as CH4 transport through the plant
and via ebullition (Cheng et al. 2013). The method captures the influence of organic amendments and rice straw
management on methanogenesis in the flooded soils, and ratooning of rice crops with a second harvest during the
growing season. In addition to CH4 emissions, DayCent simulates soil C stock changes and N20 emissions (Parton et
al. 1987 and 1998; Del Grosso et al. 2010), and allows for a seamless set of simulations for crop rotations that
include both rice and non-rice crops.
The Tier 1 method is applied to estimate CH4 emissions from rice when grown in rotation with crops that are not
simulated by DayCent, such as vegetable crops. The Tier 1 method is also used for areas converted between
agriculture (i.e., cropland and grassland) and other land uses, such as forest land, wetland, and settlements. In
addition, the Tier 1 method is used to estimate CH4 emissions from organic soils (i.e., Histosols) and from areas
with very gravelly, cobbly, or shaley soils (greater than 35 percent by volume). The Tier 3 method using DayCent
has not been fully tested for estimating emissions associated with these crops and rotations, land uses, as well as
organic soils or cobbly, gravelly, and shaley mineral soils.
The Tier 1 method for estimating CH4 emissions from rice production utilizes a default base emission rate and
scaling factors (IPCC 2006). The base emission rate represents emissions for continuously flooded fields with no
organic amendments. Scaling factors are used to adjust the base emission rate for water management and organic
amendments that differ from continuous flooding with no organic amendments. The method accounts for pre-
season and growing season flooding; types and amounts of organic amendments; and the number of rice
production seasons within a single year (i.e., single cropping, ratooning, etc.). The Tier 1 analysis is implemented in
the Agriculture and Land Use National Greenhouse Gas Inventory (ALU) software (Ogle et al. 2016).12
Rice cultivation areas are based on cropping and land use histories recorded in the USDA National Resources
Inventory (NRI) survey (USDA-NRCS 2018). The NRI is a statistically-based sample of all non-federal land, and
includes 489,178 survey locations in agricultural land for the conterminous United States and Hawaii of which
1,960 include one or more years of rice cultivation. The Tier 3 method is used to estimate CH4 emissions from
1,655 of the NRI survey locations, and the remaining 305 survey locations are estimated with the Tier 1 method.
Each NRI survey location is associated with an "expansion factor" that allows scaling of CH4 emission to the entire
land base with rice cultivation (i.e., each expansion factor represents the amount of area with the same land-
use/management history as the survey location). Land-use and some management information in the NRI (e.g.,
crop type, soil attributes, and irrigation) were collected on a 5-year cycle beginning in 1982, along with cropping
rotation data in 4 out of 5 years for each 5-year time period (i.e., 1979 to 1982,1984 to 1987,1989 to 1992, and
1994 to 1997). The NRI program began collecting annual data in 1998, with data currently available through 2015
(USDA-NRCS 2018). The current Inventory only uses NRI data through 2015 because newer data are not available,
but will be incorporated when additional years of data are released by USDA-NRCS. The harvested rice areas in
each state are presented in Table 5-13.
Table 5-13: Rice Area Harvested (1,000 Hectares)	
State/Crop	1990	2005	2015 2016 2017 2018 2019
Arkansas	600	784	679 NE NE NE NE
California	249	236	280 NE NE NE NE
12 See .
Agriculture 5-23

-------
Florida
0
4
0
NE
NE
NE
NE
Illinois
0
0
0
NE
NE
NE
NE
Kentucky
0
0
0
NE
NE
NE
NE
Louisiana
381
402
368
NE
NE
NE
NE
Minnesota
4
9
1
NE
NE
NE
NE
Mississippi
123
138
98
NE
NE
NE
NE
Missouri
48
94
62
NE
NE
NE
NE
New York
1
0
0
NE
NE
NE
NE
South Carolina
0
0
0
NE
NE
NE
NE
Tennessee
0
1
0
NE
NE
NE
NE
Texas
302
118
131
NE
NE
NE
NE
Total
1,707
1,788
1,619
NE
NE
NE
NE
Note: Totals may not sum due to independent rounding.
NE (Not Estimated). State-level area data are not available for 2016 through 2019 but will be
added in a future Inventory with release of new NRI survey data.
The Southeastern states have sufficient growing periods for a ratoon crop in some years (Table 5-14). For example,
the growing season length is occasionally sufficient for ratoon crops to be grown on about 1 percent of the rice
fields in Arkansas. No data are available about ratoon crops in Missouri or Mississippi, and the average amount of
ratooning in Arkansas was assigned to these states. Ratoon cropping occurs much more frequently in Louisiana
(LSU 2015 for years 2000 through 2013, 2015) and Texas (TAMU 2015 for years 1993 through 2015), averaging 32
percent and 45 percent of rice acres planted, respectively. Florida also has a large fraction of area with a ratoon
crop (49 percent). Ratoon rice crops are not grown in California.
Table 5-14: Average Ratooned Area as Percent of Primary Growth Area (Percent)
State
1990-2015
Arkansas3
1%
California
0%
Florida15
49%
Louisiana0
32%
Mississippi3
1%
Missouri3
1%
Texasd
45%
a Arkansas: 1990-2000 (Slaton 1999 through 2001); 2001-2011 (Wilson 2002 through 2007, 2009 through 2012); 2012-2013
(Hardke 2013, 2014). Estimates of ratooning for Missouri and Mississippi are based on the data from Arkansas.
b Florida - Ratoon: 1990-2000 (Schueneman 1997,1999 through 2001); 2001 (Deren 2002); 2002-2003 (Kirstein 2003
through 2004, 2006); 2004 (Cantens 2004 through 2005); 2005-2013 (Gonzalez 2007 through 2014).
c Louisiana: 1990-2013 (Linscombe 1999, 2001 through 2014).
d Texas: 1990-2002 (Klosterboer 1997,1999 through 2003); 2003-2004 (Stansel 2004 through 2005); 2005 (Texas Agricultural
Experiment Station 2006); 2006-2013 (Texas Agricultural Experiment Station 2007 through 2014).
While rice crop production in the United States includes a minor amount of land with mid-season drainage or
alternate wet-dry periods, the majority of rice growers use continuously flooded water management systems
(Hardke 2015; UCCE 2015; Hollier 1999; Way et al. 2014). Therefore, continuous flooding was assumed in the
DayCent simulations and the Tier 1 method. Variation in flooding can be incorporated in future Inventories if water
management data are collected.
Winter flooding is another key practice associated with water management in rice fields, and the impact of winter
flooding on CH4 emissions is addressed in the Tier 3 and Tier 1 analyses. Flooding is used to prepare fields for the
next growing season, and to create waterfowl habitat (Young 2013; Miller et al. 2010; Fleskes et al. 2005).
Fitzgerald et al. (2000) suggests that as much as 50 percent of the annual emissions may occur during winter
flooding. Winter flooding is a common practice with an average of 34 percent of fields managed with winter
flooding in California (Miller et al. 2010; Fleskes et al. 2005), and approximately 21 percent of the fields managed
with winter flooding in Arkansas (Wilson and Branson 2005 and 2006; Wilson and Runsick 2007 and 2008; Wilson
5-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
et al. 2009 and 2010; Hardke and Wilson 2013 and 2014; Hardke 2015). No data are available on winter flooding
for Texas, Louisiana, Florida, Missouri, or Mississippi. For these states, the average amount of flooding is assumed
to be similar to Arkansas. In addition, the amount of flooding is assumed to be relatively constant over the
Inventory time series.
A surrogate data method is used to estimate emissions from 2016 to 2019 associated with the rice CH4 emissions
for Tier 1 and 3 methods. Specifically, a linear regression model with autoregressive moving-average (ARMA)
errors was used to estimate the relationship between the surrogate data and emissions data from 1990 through
2015, which were derived using the Tier 1 and 3 methods (Brockwell and Davis 2016). Surrogate data are based on
rice commodity statistics from USDA-NASS.13 See Box 5-2 for more information about the surrogate data method.
Box 5-2: Surrogate Data Method
An approach to extend the time series is needed to estimate emissions from Rice Cultivation because there are
gaps in activity data at the end of the time series. This is mainly due to the fact that the National Resources
Inventory (NRI) does not release data every year, and the NRI is a key data source for estimating greenhouse gas
emissions.
A surrogate data method has been selected to impute missing emissions at the end of the time series. A linear
regression model with autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) is used to
estimate the relationship between the surrogate data and the observed 1990 to 2015 emissions data that has
been compiled using the inventory methods described in this section. The model to extend the time series is
given by
Y=xp+ Ł,
where Y is the response variable (e.g., CH4 emissions), xp is the surrogate data that is used to predict the
missing emissions data, and Ł is the remaining unexplained error. Models with a variety of surrogate data were
tested, including commodity statistics, weather data, or other relevant information. Parameters are estimated
from the observed data for 1990 to 2015 using standard statistical techniques, and these estimates are used to
predict the missing emissions data for 2016 to 2019.
A critical issue in using splicing methods is to adequately account for the additional uncertainty introduced by
predicting emissions with related information without compiling the full inventory. For example, predicting CH4
emissions will increase the total variation in the emission estimates for these specific years, compared to those
years in which the full inventory is compiled. This added uncertainty is quantified within the model framework
using a Monte Carlo approach. The approach requires estimating parameters for results in each Monte Carlo
simulation for the full inventory (i.e., the surrogate data model is refit with the emissions estimated in each
Monte Carlo iteration from the full inventory analysis with data from 1990 to 2015).
Uncertainty and Time-Series Consistency
Sources of uncertainty in the Tier 3 method include management practices, uncertainties in model structure (i.e.,
algorithms and parameterization), and variance associated with the NRI sample. Sources of uncertainty in the IPCC
(2006) Tier 1 method include the emission factors, management practices, and variance associated with the NRI
sample. A Monte Carlo analysis was used to propagate uncertainties in the Tier 1 and 3 methods. For 2016 to 2019,
there is additional uncertainty propagated through the Monte Carlo analysis associated with the surrogate data
method (See Box 5-2 for information about propagating uncertainty with the surrogate data method). The
uncertainties from the Tier 1 and 3 approaches are combined to produce the final CH4 emissions estimate using
simple error propagation (IPCC 2006). Additional details on the uncertainty methods are provided in Annex 3.12.
13 See .
Agriculture 5-25

-------
Rice cultivation CH4 emissions in 2019 were estimated to be between 3.8 and 37.5 MMT C02 Eq. at a 95 percent
confidence level, which indicates a range of 75 percent below to 149 percent above the 2019 emission estimate of
15.1 MMT C02 Eq. (see Table 5-15).
Table 5-15: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Rice
Cultivation (MMT CO2 Eq. and Percent)
Source
Inventory
Method
Gas
2019 Emission
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMTC02 Eq.)
(mmtco2
Eq.)
(%)





Lower
Upper
Lower
Upper




Bound
Bound
Bound
Bound
Rice Cultivation
Tier 3
ch4
12.5
1.4
23.7
-89%
+89%
Rice Cultivation
Tier 1
ch4
2.5
1.3
3.7
-48%
+48%
Rice Cultivation
Total
ch4
15.1
3.8
37.5
-75%
+149%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches, changes to historic data, and other parameters were applied to the entire time series
to ensure consistency in emissions estimates from 1990 through 2019. Details on the emission trends and
methodologies through time are described in more detail in the Introduction and Methodology sections.
QA/QC and Verification
General (Tier 1) and category-specific (Tier 2) QA/QC activities were conducted consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. Quality control measures include checking input data, model scripts, and results
to ensure data are properly handled throughout the inventory process. Inventory reporting forms and text are
reviewed and revised as needed to correct transcription errors. One error was found in the Tier-3 linear regression
with ARMA surrogate data method and corrected. For each Monte Carlo iteration, total CH4 emissions data were
transformed using a constant scaler to meet the model requirement, however during the back-transformation only
one constant was used for all Monte Carlo iteration. This results in a bias model prediction and lower uncertainty
in the previous year's inventory. The estimates were corrected by updating the code and emissions were re-
estimated for the years 2016 to 2019.
Model results are compared to field measurements to verify if results adequately represent CH4 emissions. The
comparisons included over 17 long-term experiments, representing about 238 combinations of management
treatments across all the sites. A statistical relationship was developed to assess uncertainties in the model
structure, adjusting the estimates for model bias and assessing precision in the resulting estimates (methods are
described in Ogle et al. 2007). See Annex 3.12 for more information.
Recalculations Discussion
Emissions data from 2016 to 2018 were corrected based on an error in the data splicing method (see QA/QC and
Verification section). This change resulted in an average increase in CH4 emissions of 2.2 MMT C02 Eq., or 2.3
percent, from 2016 to 2018 relative to the previous Inventory.
Planned Improvements
A key planned improvement for rice cultivation is to fill several gaps in the management activity including
compiling new data on water management, organic amendments and ratooning practices in rice cultivation
systems. This improvement is expected to be completed for the next Inventory, but may be prioritized considering
overall improvements to make best use of available resources.
5-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
5.4 Agricultural Soil Management (CRF Source
Category 3D)
Nitrous oxide is naturally produced in soils through the microbial processes of nitrification and denitrification that
is driven by the availability of mineral nitrogen (N) (Firestone and Davidson 1989).14 Mineral N is made available in
soils through decomposition of soil organic matter and plant litter, as well as asymbiotic fixation of N from the
atmosphere.15 Several agricultural activities increase mineral N availability in soils that lead to direct N20
emissions at the site of a management activity (see Figure 5-4) (Mosier et al. 1998). These activities include
synthetic N fertilization; application of managed livestock manure; application of other organic materials such as
biosolids (i.e., treated sewage sludge); deposition of manure on soils by domesticated animals in pastures, range,
and paddocks (PRP) (i.e., unmanaged manure); retention of crop residues (N-fixing legumes and non-legume crops
and forages); and drainage of organic soils16 (i.e., Histosols) (IPCC 2006). Additionally, agricultural soil management
activities, including irrigation, drainage, tillage practices, cover crops, and fallowing of land, can influence N
mineralization from soil organic matter and levels of asymbiotic N fixation. Indirect emissions of N20 occur when N
is transported from a site and is subsequently converted to N20; there are two pathways for indirect emissions: (1)
volatilization and subsequent atmospheric deposition of applied/mineralized N, and (2) surface runoff and leaching
of applied/mineralized N into groundwater and surface water.17 Direct and indirect emissions from agricultural
lands are included in this section (i.e., cropland and grassland as defined in Section 6.1 Representation of the U.S.
Land Base). Nitrous oxide emissions from Forest Land and Settlements soils are found in Sections 6.2 and 6.10,
respectively.
14	Nitrification and denitrification are driven by the activity of microorganisms in soils. Nitrification is the aerobic microbial
oxidation of ammonium (NH4+) to nitrate (NOs ), and denitrification is the anaerobic microbial reduction of nitrate to N2. Nitrous
oxide is a gaseous intermediate product in the reaction sequence of nitrification and denitrification.
15	Asymbiotic N fixation is the fixation of atmospheric N2 by bacteria living in soils that do not have a direct relationship with
plants.
16	Drainage of organic soils in former wetlands enhances mineralization of N-rich organic matter, thereby increasing N20
emissions from these soils.
17	These processes entail volatilization of applied or mineralized N as NH3 and NOx, transformation of these gases in the
atmosphere (or upon deposition), and deposition of the N primarily in the form of particulate NH4+, nitric acid (HN03), and NOx.
In addition, hydrological processes lead to leaching and runoff of N03" that is converted to N20 in aquatic systems, e.g.,
wetlands, rivers, streams and lakes. Note: N20 emissions are not estimated for aquatic systems associated with N inputs from
terrestrial systems in order to avoid double-counting.
Agriculture 5-27

-------
Figure 5-4: Sources and Pathways of N that Result in N2O Emissions from Agricultural Soil
Management
Sources and Pathways of N that Result in N2O Emissions from Agricultural Soil Management
N Volatilization
Synthetic N Fertilizers
fEffniTZER
Synthetic IM fertilizer applied to soil
N Inputs to
Managed Soils
Organic
Amendments
Direct N20
Emissions
Includes both commercial and
non-co,rn mercisl fertilizers (i.e.,
animal manure, com post
sewage sludge, tankage etc.)
N Volatilization
and Deposition
Urine and Dung from
Grazing Animals
Indirect N20
Emissions
Manure deposited on pasture rang^
and paddock
Crop Residues
Ind udes above- and belowground
residues for all crops (non-N and In-
fixing (and from perennial forage
crops and pastures following renewal
Mineralization of
Soil Organic Matter
Includes N converted to mineral form
upon decomposition of soil organic
Asymbiotic Fixation
Fixation of atm ospheric by bacteria
living in soilsthat do not have a direct
relationshipwith plants
Histosol
Cultivation
This graphic illustrates the sources and pathways of nitrogen that result
in direct and indirect N,0 emissions from soils using the methodologies
described in this Inventory. Emission pathways are shown with arrows.
On the lower right-hand side is a cut-away view of a representative
section of a managed soil; histosol cultivation is represented here.
Groundwater
5-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Agricultural soils produce the majority of N20 emissions in the United States. Estimated emissions in 2019 are
344.6 MMT C02 Eq. (1,156 kt) (see Table 5-16 and Table 5-17). Annual N20 emissions from agricultural soils are 9
percent greater in the 2019 compared to 1990, but emissions fluctuated between 1990 and 2019 due to inter-
annual variability largely associated with weather patterns, synthetic fertilizer use, and crop production. From
1990 to 2019, cropland accounted for 68 percent of total direct emissions on average, while grassland accounted
for 32 percent. On average, 79 percent of indirect emissions are from croplands and 21 percent from grasslands.
Estimated direct and indirect N20 emissions by sub-source category are shown in Table 5-18 and Table 5-19.
Table 5-16: N2O Emissions from Agricultural Soils (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Direct
272.5
272.6
295.0
281.4
280.3
285.9
290.4
Cropland
185.9
183.7
199.5
190.8
190.4
195.1
196.4
Grassland
86.6
88.8
95.4
90.6
89.9
90.9
94.0
Indirect
43.4
40.8
53.5
48.7
47.3
52.3
54.2
Cropland
34.2
31.6
42.7
38.8
37.4
42.3
43.8
Grassland
9.2
9.2
10.8
9.9
9.8
10.0
10.4
Total
315.9
313.4
348.5
330.1
327.6
338.2
344.6
Notes: Estimates after 2015 are based on a data splicing method (See Methodology section). Totals
may not sum due to independent rounding.
Table 5-17: N2O Emissions from Agricultural Soils (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Direct
914.5
914.7
989.9
944.3
940.6
959.5
974.5
Cropland
623.8
616.6
669.6
640.3
639.0
654.5
659.1
Grassland
290.7
298.1
320.2
304.1
301.6
305.0
315.5
Indirect
145.6
137.0
179.6
163.4
158.6
175.5
181.9
Cropland
114.8
106.1
143.2
130.3
125.5
142.0
147.1
Grassland
30.7
30.9
36.4
33.1
33.0
33.4
34.8
Total
1,060.1
1,051.6
1,169.4
1,107.7
1,099.2
1,135.0
1,156.4
Notes: Estimates after 2015 are based on a data splicing method (See Methodology section). Totals may
not sum due to independent rounding.
Table 5-18: Direct N2O Emissions from Agricultural Soils by Land Use Type and N Input Type
(MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Cropland
185.8
183.7
199.5
190.8
190.4
195.0
196.4
Mineral Soils
182.1
180.0
196.1
187.4
187.0
191.6
193.0
Synthetic Fertilizer
63.1
64.0
64.8
68.8
68.5
70.1
70.4
Organic Amendment3
12.6
13.0
13.4
14.5
14.3
14.3
14.2
Residue Nb
39.3
39.6
39.0
40.1
40.1
41.2
41.6
Mineralization and







Asymbiotic Fixation
67.1
63.3
78.9
64.0
64.1
66.1
66.8
Drained Organic Soils
3.8
3.7
3.4
3.4
3.4
3.4
3.4
Grassland
86.7
88.9
95.5
90.6
89.9
90.9
94.0
Mineral Soils
84.2
86.5
93.0
88.2
87.4
88.4
91.6
Synthetic Fertilizer
+
+
+
+
+
+
+
PRP Manure
14.6
13.4
12.8
12.8
12.8
12.9
13.2
Managed Manurec
+
+
+
+
+
+
+
Biosolids (i.e., treated







Sewage Sludge)
0.2
0.5
0.6
0.6
0.6
0.6
0.7
Residue Nd
29.7
30.8
30.4
31.5
31.2
31.6
32.8
Mineralization and







Asymbiotic Fixation
39.5
41.7
49.2
43.2
42.8
43.3
44.9
Agriculture 5-29

-------
Drained Organic Soils	23	2.4	2.5 2.5 2.5 2.5 2.5
Total	272.5	272.6 295.0 281.4 280.3 285.9 290.4
Notes: Estimates after 2015 are based on a data splicing method (See Methodology section). Totals may not
sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Organic amendment inputs include managed manure, daily spread manure, and commercial organic
fertilizers (i.e., dried blood, dried manure, tankage, compost, and other).
b Cropland residue N inputs include N in unharvested legumes as well as crop residue N.
c Managed manure inputs include managed manure and daily spread manure amendments that are applied
to grassland soils.
d Grassland residue N inputs include N in ungrazed legumes as well as ungrazed grass residue N.
Table 5-19: Indirect N2O Emissions from Agricultural Soils (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
Cropland
34.2
31.6
42.7
38.8
37.4
42.3
43.8
Volatilization & Atm.







Deposition
6.5
7.3
8.5
8.1
7.9
8.0
7.9
Surface Leaching & Run-Off
27.7
24.4
34.2
30.7
29.5
34.4
35.9
Grassland
9.2
9.2
10.8
9.9
9.8
10.0
10.4
Volatilization & Atm.







Deposition
3.6
3.6
3.7
3.5
3.5
3.5
3.6
Surface Leaching & Run-Off
5.6
5.6
7.2
6.4
6.3
6.4
6.8
Total
43.4
40.8
53.5
48.7
47.3
52.3
54.2
Notes: Estimates after 2015 are based on a data splicing method (See Methodology section). Totals may not
sum due to independent rounding.
Figure 5-5 and Figure 5-6 show regional patterns for direct N20 emissions. Figure 5-7 and Figure 5-8 show indirect
N20 emissions from volatilization, and Figure 5-9 and Figure 5-10 show the indirect N20 emissions from leaching
and runoff in croplands and grasslands, respectively.
5-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 5-5: Croplands, 2015 Annual Direct N2O Emissions Estimated Using the Tier 3
DayCent Model
Note: Only national-scale emissions are estimated for 2016 to 2019 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Direct N20 emissions from croplands occur throughout all of the cropland regions but tend to be high in the
Midwestern Corn Belt Region (Illinois, Iowa, Indiana, Ohio, southern Minnesota and Wisconsin, and eastern
Nebraska), where a large portion of the land is used for growing highly fertilized corn and N-fixing soybean crops
(see Figure 5-5). Kansas, South Dakota and North Dakota have relatively high emissions from large areas of crop
production that are found in the Great Plains region. Emissions are also high in the Lower Mississippi River Basin
from Missouri to Louisiana, and highly productive irrigated areas, such as Platte River, which flows from Colorado
through Nebraska, Snake River Valley in Idaho and the Central Valley in California. Direct emissions are low in
many parts of the eastern United States because only a small portion of land is cultivated, and in many western
states where rainfall and access to irrigation water are limited.
Direct emissions from grasslands are more evenly distributed throughout the United States (see Figure 5-6), but
total emissions tend be highest in the Great Plains and western United States where a large proportion of the land
is dominated by grasslands with cattle and sheep grazing. However, there are relatively large emissions from local
areas in the Eastern United States, particularly Kentucky and Tennessee, in addition to areas in Missouri and Iowa,
where there can be higher rates of Pasture/Range/Paddock (PRP) manure N additions on a relatively small amount
of pasture. These areas have greater stocking rates of livestock per unit of area, compared to other regions of the
United States.
Agriculture 5-31

-------
Figure 5-6: Grasslands, 2015 Annual Direct N2O Emissions Estimated Using the Tier 3
DayCent Model
Note: Only national-scale emissions are estimated for 2016 to 2019 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Indirect N20 emissions from volatilization in croplands have a similar pattern as the direct N20 emissions with
higher emissions in the Midwestern Corn Belt, Lower Mississippi River Basin and Great Plains. Indirect N20
emissions from volatilization in grasslands are higher in the Southeastern United States, along with portions of the
Mid-Atlantic and southern Iowa. The higher emissions in this region are mainly due to large additions of PRP
manure N on relatively small but productive pastures that support intensive grazing, which in turn, stimulates NH3
volatilization.
Indirect l\l20 emissions from surface runoff and leaching of applied/mineralized N in croplands is highest in the
Midwestern Corn Belt. There are also relatively high emissions associated with N management in the Lower
Mississippi River Basin, Piedmont region of the Southeastern United States and the Mid-Atlantic states. In addition,
areas of high emissions occur in portions of the Great Plains that have relatively large areas of irrigated croplands
with high leaching rates of applied/mineralized N. Indirect N20 emissions from surface runoff and leaching of
applied/mineralized N in grasslands are higher in the eastern United States and coastal Northwest region. These
regions have greater precipitation and higher levels of leaching and runoff compared to arid to semi-arid regions in
the Western United States.
5-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 5-7: Croplands, 2015 Annual Indirect N2O Emissions from Volatilization Using the
Tier 3 DayCent Model
Note: Only national-scale emissions are estimated for 2016 to 2019 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Figure 5-8: Grasslands, 2015 Annual Indirect N2O Emissions from Volatilization Using the
Tier 3 DayCent Model
Note: Only national-scale emissions are estimated for 2016 to 2019 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Agriculture 5-33

-------
Figure 5-9: Croplands, 2015 Annual Indirect N2O Emissions from Leaching and Runoff Using
the Tier 3 DayCent Model
Note: Only national-scale emissions are estimated for 2016 to 2019 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Figure 5-10: Grasslands, 2015 Annual Indirect N2O Emissions from Leaching and Runoff
Using the Tier 3 DayCent Model
Note: Only national-scale emissions are estimated for 2016 to 2019 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
5-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
The 2006IPCC Guidelines (IPCC 2006) divide emissions from the agricultural soil management source category into
five components, including (1) direct emissions from N additions to cropland and grassland mineral soils from
synthetic fertilizers, biosolids (i.e., treated sewage sludge), crop residues (legume N-fixing and non-legume crops),
and organic amendments; (2) direct emissions from soil organic matter mineralization due to land use and
management change; (3) direct emissions from drainage of organic soils in croplands and grasslands; (4) direct
emissions from soils due to manure deposited by livestock on PRP grasslands; and (5) indirect emissions from soils
and water from N additions and manure deposition to soils that lead to volatilization, leaching, or runoff of N and
subsequent conversion to N20.
In this source category, the United States reports on all croplands, as well as all managed grasslands, whereby
anthropogenic greenhouse gas emissions are estimated consistent with the managed land concept (IPCC 2006),
including direct and indirect N20 emissions from asymbiotic fixation18 and mineralization of N associated with
decomposition of soil organic matter and residues. One recommendation from IPCC (2006) that has not been
completely adopted is the estimation of emissions from grassland pasture renewal, which involves occasional
plowing to improve forage production in pastures. Currently no data are available to address pasture renewal.
Direct N20 Emissions
The methodology used to estimate direct N20 emissions from agricultural soil management in the United States is
based on a combination of IPCC Tier 1 and 3 approaches, along with application of a splicing method for latter
years in the Inventory time series (IPCC 2006; Del Grosso et al. 2010) where data are not yet available. A Tier 3
process-based model (DayCent) is used to estimate direct emissions from a variety of crops that are grown on
mineral (i.e., non-organic) soils, as well as the direct emissions from non-federal grasslands except for applications
of biosolids (i.e., treated sewage sludge) (Del Grosso et al. 2010). The Tier 3 approach has been specifically
designed and tested to estimate N20 emissions in the United States, accounting for more of the environmental and
management influences on soil N20 emissions than the IPCC Tier 1 method (see Box 5-3 for further elaboration).
Moreover, the Tier 3 approach addresses direct N20 emissions and soil C stock changes from mineral cropland soils
in a single analysis. Carbon and N dynamics are linked in plant-soil systems through biogeochemical processes of
microbial decomposition and plant production (McGill and Cole 1981). Coupling the two source categories (i.e.,
agricultural soil C and N20) in a single inventory analysis ensures that there is consistent activity data and
treatment of the processes, and interactions are considered between C and N cycling in soils.
The Tier 3 approach is based on the crop and land use histories recorded in the USDA National Resources Inventory
(NRI) (USDA-NRCS 2018a). The NRI is a statistically-based sample of all non-federal land,19 and includes 349,464
points on agricultural land for the conterminous United States that are included in the Tier 3 method. The Tier 1
approach is used to estimate the emissions from 175,527 locations in the NRI survey across the time series, which
are designated as cropland or grassland (discussed later in this section). Each survey location is associated with an
"expansion factor" that allows scaling of N20 emissions from NRI points to the entire country (i.e., each expansion
factor represents the amount of area with the same land-use/management history as the survey location). Each
NRI survey location was sampled on a 5-year cycle from 1982 until 1997. For cropland, data were collected in 4 out
of 5 years in the cycle (i.e., 1979 through 1982,1984 through 1987,1989 through 1992, and 1994 through 1997).
In 1998, the NRI program began collecting annual data, which are currently available through 2015 (USDA-NRCS
2018a).
18	N inputs from asymbiotic N fixation are not directly addressed in 2006 IPCC Guidelines, but are a component of the N inputs
and total emissions from managed lands and are included in the Tier 3 approach developed for this source.
19	The NRI survey does include sample points on federal lands, but the program does not collect data from those sample
locations.
Agriculture 5-35

-------
Box 5-3: Tier 1 vs. Tier 3 Approach for Estimating N20 Emissions
The IPCC (2006) Tier 1 approach is based on multiplying activity data on different N inputs (i.e., synthetic
fertilizer, manure, N fixation, etc.) by the appropriate default IPCC emission factors to estimate N20 emissions
on an input-by-input basis. The Tier 1 approach requires a minimal amount of activity data, readily available in
most countries (e.g., total N applied to crops); calculations are simple; and the methodology is highly
transparent. In contrast, the Tier 3 approach developed for this Inventory is based on application of a process-
based model (i.e., DayCent) that represents the interaction of N inputs, land use and management, as well as
environmental conditions at specific locations, such as freeze-thaw effects that generate hot moments of N20
emissions (Wagner-Riddle et al. 2017). Consequently, the Tier 3 approach accounts for land-use and
management impacts and their interaction with environmental factors, such as weather patterns and soil
characteristics, in a more comprehensive manner, which will enhance or dampen anthropogenic influences.
However, the Tier 3 approach requires more detailed activity data (e.g., crop-specific N fertilization rates),
additional data inputs (e.g., daily weather, soil types), and considerable computational resources and
programming expertise. The Tier 3 methodology is less transparent, and thus it is critical to evaluate the output
of Tier 3 methods against measured data in order to demonstrate that the method is an improvement over
lower tier methods for estimating emissions (IPCC 2006). Another important difference between the Tier 1 and
Tier 3 approaches relates to assumptions regarding N cycling. Tier 1 assumes that N added to a system is subject
to N20 emissions only during that year and cannot be stored in soils and contribute to N20 emissions in
subsequent years. This is a simplifying assumption that may create bias in estimated N20 emissions for a specific
year. In contrast, the process-based model in the Tier 3 approach includes the legacy effect of N added to soils
in previous years that is re-mineralized from soil organic matter and emitted as N20 during subsequent years.
DayCent is used to estimate N20 emissions associated with production of alfalfa hay, barley, corn, cotton, grass
hay, grass-clover hay, oats, peanuts, potatoes, rice, sorghum, soybeans, sugar beets, sunflowers, tobacco and
wheat, but is not applied to estimate N20 emissions from other crops or rotations with other crops,20 such as
sugarcane, some vegetables, and perennial/horticultural crops. Areas that are converted between agriculture (i.e.,
cropland and grassland) and other land uses, such as forest land, wetland and settlements, are not simulated with
DayCent. DayCent is also not used to estimate emissions from land areas with very gravelly, cobbly, or shaley soils
in the topsoil (greater than 35 percent by volume in the top 30 cm of the soil profile), or to estimate emissions
from drained organic soils (Histosols). The Tier 3 method has not been fully tested for estimating N20 emissions
associated with these crops and rotations, land uses, as well as organic soils or cobbly, gravelly, and shaley mineral
soils. In addition, federal grassland areas are not simulated with DayCent due to limited activity data on land use
histories. For areas that are not included in the DayCent simulations, Tier 1 methods are used to estimate
emissions, including (1) direct emissions from N inputs for crops on mineral soils that are not simulated by
DayCent; (2) direct emissions from PRP N additions on federal grasslands; (3) direct emissions for land application
of biosolids (i.e., treated sewage sludge) to soils; and (4) direct emissions from drained organic soils in croplands
and grasslands.
A splicing method is used to estimate soil N20 emissions from 2016 to 2019 at the national scale because new NRI
activity data are not available for those years. Specifically, linear regression models with autoregressive moving-
average (ARMA) errors (Brockwell and Davis 2016) are used to estimate the relationship between surrogate data
and the 1990 to 2015 emissions that are derived using the Tier 3 method. Surrogate data for these regression
models includes corn and soybean yields from USDA-NASS statistics,21 and weather data from the PRISM Climate
Group (PRISM 2018). For the Tier 1 method, a linear-time series model is used to estimate emissions from 2016 to
2019 without surrogate data for most of the N sources (exceptions include biosolids, drainage of organic soils, and
20	A small proportion of the major commodity crop production, such as corn and wheat, is included in the Tier 1 analysis
because these crops are rotated with other crops or land uses (e.g., forest lands) that are not simulated by DayCent.
21	See .
5-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
crop residue N). See Box 5-4 for more information about the splicing method. Emission estimates for 2016 to 2019
will be recalculated in future Inventory reports when new NRI data are available.
Box 5-4: Surrogate Data Method
An approach to extend the time series is needed for Agricultural Soil Management because there are typically
activity data gaps at the end of the time series. This is mainly because the NRI survey program, which provides
critical information for estimating greenhouse gas emissions and removals, does not release data every year.
Splicing methods have been used to impute missing data at the end of the emission time series for both the Tier
1 and 3 methods. Specifically, a linear regression model with autoregressive moving-average (ARMA) errors
(Brockwell and Davis 2016) is used to estimate emissions based on the modeled 1990 to 2015 emissions data,
which has been compiled using the inventory methods described in this section. The model to extend the time
series is given by
Y = xp + Ł,
where Y is the response variable (e.g., soil nitrous oxide), xp for the Tier 3 method contains specific surrogate
data depending on the response variable, and Ł is the remaining unexplained error. Models with a variety of
surrogate data were tested, including commodity statistics, weather data, or other relevant information. The
term xp for the Tier 1 method only contains year as a predictor of emission patterns over the time series
(change in emissions per year), and therefore, is a linear time series model with no surrogate data. Parameters
are estimated from the emissions data for 1990 to 2015 using standard statistical techniques, and these
estimates are used in the model described above to predict the missing emissions data for 2016 to 2019.
A critical issue with splicing methods is to account for the additional uncertainty introduced by predicting
emissions without compiling the full inventory. Specifically, uncertainty will increase for years with imputed
estimates based on the splicing methods, compared to those years in which the full inventory is compiled. This
additional uncertainty is quantified within the model framework using a Monte Carlo approach. Consequently,
the uncertainty from the original inventory data is combined with the uncertainty in the data splicing model.
The approach requires estimating parameters in the data splicing models in each Monte Carlo simulation for the
full inventory (i.e., the surrogate data model is refit with the draws of parameters values that are selected in
each Monte Carlo iteration, and used to produce estimates with inventory data from 1990 to 2015). Therefore,
the data splicing method generates emissions estimates from each surrogate data model in the Monte Carlo
analysis, which are used to derive confidence intervals in the estimates for the missing emissions data from
2016 to 2019. Furthermore, the 95 percent confidence intervals are estimated using the 3 sigma rules assuming
a unimodal density (Pukelsheim 1994).
Tier 3 Approach for Mineral Cropland Soils
The DayCent biogeochemical model (Parton et al. 1998; Del Grosso et al. 2001 and 2011) is used to estimate direct
N20 emissions from mineral cropland soils that are managed for production of a wide variety of crops (see list in
previous section) based on the crop histories in the 2015 NRI (USDA-NRCS 2018a). Crops simulated by DayCent are
grown on approximately 85 percent of total cropland area in the United States. The model simulates net primary
productivity (NPP) using the NASA-CASA production algorithm MODIS Enhanced Vegetation Index (EVI) products,
MOD13Q1 and MYD13Q122 (Potter et al. 1993, 2007). The model simulates soil temperature and water dynamics,
using daily weather data from a 4-kilometer gridded product developed by the PRISM Climate Group (2018), and
22 NPP is estimated with the NASA-CASA algorithm for most of the cropland that is used to produce major commodity crops in
the central United States from 2000 to 2015. Other regions and years prior to 2000 are simulated with a method that
incorporates water, temperature, and moisture stress on crop production (see Metherell et al. 1993), but does not incorporate
the additional information about crop condition provided with remote sensing data.
Agriculture 5-37

-------
soil attributes from the Soil Survey Geographic Database (SSURGO) (Soil Survey Staff 2019). DayCent is used to
estimate direct N20 emissions due to mineral N available from the following sources: (1) application of synthetic
fertilizers; (2) application of livestock manure; (3) retention of crop residues in the field for N-fixing legumes and
non-legume crops and subsequent mineralization of N during microbial decomposition (i.e., leaving residues in the
field after harvest instead of burning or collecting residues); (4) mineralization of N from decomposition of soil
organic matter; and (5) asymbiotic fixation.
Management activity data from several sources supplement the activity data from the NRI. The USDA-NRCS
Conservation Effects and Assessment Project (CEAP) provides data on a variety of cropland management activities,
and is used to inform the inventory analysis about tillage practices, mineral fertilization, manure amendments,
cover crop management, as well as planting and harvest dates (USDA-NRCS 2018b; USDA-NRCS 2012). CEAP data
are collected at a subset of NRI survey locations, and currently provide management information from
approximately 2002 to 2006. These data are combined with other datasets in an imputation analysis that extend
the time series from 1990 to 2015. This imputation analysis is comprised of three steps: a) determine the trends in
management activity across the time series by combining information from several datasets (discussed below), b)
use an artificial neural network to determine the likely management practice at a given NRI survey location (Cheng
and Titterington 1994), and c) assign management practices from the CEAP survey to specific NRI locations using
predictive mean matching methods that are adapted to reflect the trending information (Little 1988, van Buuren
2012). The artificial neural network is a machine learning method that approximates nonlinear functions of inputs
and searches through a very large class of models to impute an initial value for management practices at specific
NRI survey locations. The predictive mean matching method identifies the most similar management activity
recorded in the CEAP survey that matches the prediction from the artificial neural network. The matching ensures
that imputed management activities are realistic for each NRI survey location, and not odd or physically
unrealizable results that could be generated by the artificial neural network. There are six complete imputations of
the management activity data using these methods.
To determine trends in mineral fertilization and manure amendments from 1979 to 2015, CEAP data are combined
with information on fertilizer use and rates by crop type for different regions of the United States from the USDA
Economic Research Service. The data collection program was known as the Cropping Practices Surveys through
1995 (USDA-ERS 1997), and is now part of data collection known as the Agricultural Resource Management
Surveys (ARMS) (USDA-ERS 2018). Additional data on fertilization practices are compiled through other sources
particularly the National Agricultural Statistics Service (USDA-NASS 1992,1999, 2004). The donor survey data from
CEAP contain both mineral fertilizer rates and manure amendment rates, so that the selection of a donor via
predictive mean matching yields the joint imputation of both rates. This approach captures the relationship
between mineral fertilization and manure amendment practices for U.S. croplands based directly on the observed
patterns in the CEAP survey data.
To determine the trends in tillage management from 1979 to 2015, CEAP data are combined with Conservation
Technology Information Center data between 1989 and 2004 (CTIC 2004) and USDA-ERS Agriculture Resource
Management Surveys (ARMS) data from 2002 to 2015 (Claasen et al. 2018). The CTIC data are adjusted for long-
term adoption of no-till agriculture (Towery 2001). It is assumed that the majority of agricultural lands are
managed with full tillage prior to 1985.
For cover crops, CEAP data are combined with information from 2011 to 2016 in the USDA Census of Agriculture
(USDA-NASS 2012, 2017). It is assumed that cover crop management was minimal prior to 1990 and the rates
increased linearly over the decade to the levels of cover crop management in the CEAP survey.
The IPCC method considers crop residue N and N mineralized from soil organic matter as activity data. However,
they are not treated as activity data in DayCent simulations because residue production, symbiotic N fixation (e.g.,
legumes), mineralization of N from soil organic matter, and asymbiotic N fixation are internally generated by the
model as part of the simulation. In other words, DayCent accounts for the influence of symbiotic N fixation,
mineralization of N from soil organic matter and crop residue retained in the field, and asymbiotic N fixation on
N20 emissions, but these are not model inputs.
The N20 emissions from crop residues are reduced by approximately 3 percent (the assumed average burned
portion for crop residues in the United States) to avoid double counting associated with non-C02 greenhouse gas
5-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
emissions from agricultural residue burning. Estimated levels of residue burning are based on state inventory data
(ILENR 1993; Oregon Department of Energy 1995; Noller 1996; Wisconsin Department of Natural Resources 1993;
Cibrowski 1996).
Uncertainty in the emission estimates from DayCent is associated with input uncertainty due to missing
management data in the NRI survey that is imputed from other sources; model uncertainty due to incomplete
specification of C and N dynamics in the DayCent model parameters and algorithms; and sampling uncertainty
associated with the statistical design of the NRI survey. To assess input uncertainty, C and N dynamics at each NRI
survey location are simulated six times using the imputation product and other model driver data. Uncertainty in
parameterization and model algorithms are determined using a structural uncertainty estimator derived from
fitting a linear mixed-effect model (Ogle et al. 2007; Del Grosso et al. 2010). Sampling uncertainty is assessed using
NRI replicate sampling weights. These data are combined in a Monte Carlo stochastic simulation with 1,000
iterations for 1990 through 2015. For each iteration, there is a random selection of management data from the
imputation product (select one of the six imputations), random selection of parameter values and random effects
for the linear mixed-effect model (i.e., structural uncertainty estimator), and random selection of a set of survey
weights from the replicates associated with the NRI survey design.
Nitrous oxide emissions and 95 percent confidence intervals are estimated for each year between 1990 and 2015
using the DayCent model. However, note that the areas have been modified in the original NRI survey through a
process in which the Forest Inventory and Analysis (FIA) survey data and the National Land Cover Dataset (Yang et
al. 2018) are harmonized with the NRI data. This process ensures that the land use areas are consistent across all
land use categories (See Section 6.1, Representation of the U.S. Land Base for more information). Further
elaboration on the methodology and data used to estimate N20 emissions from mineral soils are described in
Annex 3.12.
For the Tier 3 method, soil N20 emissions from 2016 to 2019 associated with mineral soils in croplands are
estimated using a splicing method that accounts for uncertainty in the original inventory data and the splicing
method (See Box 5-4). Annual data are currently available through 2015 (USDA-NRCS 2018a), and the Inventory
time series will be updated in the future when new NRI data are released.
Nitrous oxide emissions from managed agricultural lands are the result of interactions among anthropogenic
activities (e.g., N fertilization, manure application, tillage) and other driving variables, such as weather and soil
characteristics. These factors influence key processes associated with N dynamics in the soil profile, including
immobilization of N by soil microbial organisms, decomposition of organic matter, plant uptake, leaching, runoff,
and volatilization, as well as the processes leading to N20 production (nitrification and denitrification). It is not
possible to partition N20 emissions into each anthropogenic activity directly from model outputs due to the
complexity of the interactions (e.g., N20 emissions from synthetic fertilizer applications cannot be distinguished
from those resulting from manure applications). To approximate emissions by activity, the amount of mineral N
added to the soil, or made available through decomposition of soil organic matter and plant litter, as well as
asymbiotic fixation of N from the atmosphere, is determined for each N source and then divided by the total
amount of mineral N in the soil according to the DayCent model simulation. The percentages are then multiplied
by the total of direct N20 emissions in order to approximate the portion attributed to N management practices.
This approach is only an approximation because it assumes that all N made available in soil has an equal
probability of being released as N20, regardless of its source, which is unlikely to be the case (Delgado et al. 2009).
However, this approach allows for further disaggregation of emissions by source of N, which is valuable for
reporting purposes and is analogous to the reporting associated with the IPCC (2006) Tier 1 method, in that it
associates portions of the total soil N20 emissions with individual sources of N.
Tier 1 Approach for Mineral Cropland Soils
The IPCC (2006) Tier 1 methodology is used to estimate direct N20 emissions for mineral cropland soils that are not
simulated by DayCent (e.g., DayCent has not been parametrized to simulate all crop types and some soil types such
as Histosols). For the Tier 1 method, estimates of direct N20 emissions from N applications are based on mineral
soil N that is made available from the following practices: (1) the application of synthetic commercial fertilizers; (2)
application of managed manure and non-manure commercial organic fertilizers; and (3) decomposition and
Agriculture 5-39

-------
mineralization of nitrogen from above- and below-ground crop residues in agricultural fields (i.e., crop biomass
that is not harvested). Non-manure commercial organic amendments are only included in the Tier 1 analysis
because these data are not available at the county-level, which is necessary for the DayCent simulations.
Consequently, all commercial organic fertilizer, as well as manure that is not added to crops in the DayCent
simulations, are included in the Tier 1 analysis. The following sources are used to derive activity data:
•	A process-of-elimination approach is used to estimate synthetic N fertilizer additions for crop areas that
are not simulated by DayCent. The total amount of fertilizer used on farms has been estimated at the
county-level by the USGS using sales records from 1990 to 2012 (Brakebill and Gronberg 2017). For 2013
through 2015, county-level fertilizer used on-farms is adjusted based on annual fluctuations in total U.S.
fertilizer sales (AAPFCO 2013 through 20 17).23 After subtracting the portion of fertilizer applied to crops
and grasslands simulated by DayCent (see Tier 3 Approach for Mineral Cropland Soils and Direct N20
Emissions from Grassland Soils sections for information on data sources), the remainder of the total
fertilizer used on farms is assumed to be applied to crops that are not simulated by DayCent.
•	Similarly, a process-of-elimination approach is used to estimate manure N additions for crops that are not
simulated by DayCent. The total amount of manure available for land application to soils has been
estimated with methods described in the Manure Management section (Section 5.2) and annex (Annex
3.11). The amount of manure N applied in the Tier 3 approach to crops and grasslands is subtracted from
total annual manure N available for land application (see Tier 3 Approach for Mineral Cropland Soils and
Direct N20 Emissions from Grassland Soils sections for information on data sources). This difference is
assumed to be applied to crops that are not simulated by DayCent.
•	Commercial organic fertilizer additions are based on organic fertilizer consumption statistics, which are
converted from mass of fertilizer to units of N using average organic fertilizer N content, which range
between 2.3 to 4.2 percent across the time series (TVA 1991 through 1994; AAPFCO 1995 through 2017).
Commercial fertilizers do include dried manure and biosolids (i.e., treated sewage sludge), but the
amounts are removed from the commercial fertilizer data to avoid double counting24 with the manure N
dataset described above and the biosolids (i.e., treated sewage sludge) amendment data discussed later
in this section.
•	Crop residue N is derived by combining amounts of above- and below-ground biomass, which are
determined based on NRI crop area data (USDA-NRCS 2018a), crop production yield statistics (USDA-NASS
2019), dry matter fractions (IPCC 2006), linear equations to estimate above-ground biomass given dry
matter crop yields from harvest (IPCC 2006), ratios of below-to-above-ground biomass (IPCC 2006), and N
contents of the residues (IPCC 2006). N inputs from residue were reduced by 3 percent to account for
average residue burning portions in the United States.
The total increase in soil mineral N from applied fertilizers and crop residues is multiplied by the IPCC (2006)
default emission factor to derive an estimate of direct N20 emissions using the Tier 1 method. Further elaboration
on the methodology and data used to estimate N20 emissions from mineral soils are described in Annex 3.12.
Soil N20 emissions from 2016 to 2019 for Tier 1 mineral soil emissions are estimated using a splicing method that is
described in Box 5-4, with the exception of the crop residue N, which is only estimated with the data splicing
method for 2019. As with the Tier 3 method, the time series that is based on the splicing methods will be
recalculated in a future Inventory report when updated activity data are available.
23	The fertilizer consumption data in AAPFCO are recorded in "fertilizer year" totals, (i.e., July to June), but are converted to
calendar year totals. This is done by assuming that approximately 35 percent of fertilizer usage occurred from July to December
and 65 percent from January to June (TVA 1992b).
24	Commercial organic fertilizers include dried blood, tankage, compost, and other, but the dried manure and biosolids (i.e.,
treated sewage sludge) are also included in other datasets in this Inventory. Consequently, the proportions of dried manure and
biosolids, which are provided in the reports (TVA 1991 through 1994; AAPFCO 1995 through 2017), are used to estimate the N
amounts in dried manure and biosolids. To avoid double counting, the resulting N amounts for dried manure and biosolids are
subtracted from the total N in commercial organic fertilizers before estimating emissions using the Tier 1 method.
5-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Tier 1 and 3 Approaches for Direct N2O Emissions from Mineral Grassland Soils
As with N20 emissions from croplands, the Tier 3 process-based DayCent model and Tier 1 method described in
IPCC (2006) are combined to estimate emissions from non-federal grasslands and PRP manure N additions for
federal grasslands, respectively. Grassland includes pasture and rangeland that produce grass or mixed
grass/legume forage primarily for livestock grazing. Rangelands are extensive areas of native grassland that are not
intensively managed, while pastures are seeded grassland (possibly following tree removal) that may also have
additional management, such as irrigation, fertilization, or inter-seeding legumes. DayCent is used to simulate N20
emissions from NRI survey locations (USDA-NRCS 2018a) on non-federal grasslands resulting from manure
deposited by livestock directly onto pastures and rangelands (i.e., PRP manure), N fixation from legume seeding,
managed manure amendments (i.e., manure other than PRP manure such as Daily Spread or manure collected
from other animal waste management systems such as lagoons and digesters), and synthetic fertilizer application.
Other N inputs are simulated within the DayCent framework, including N input from mineralization due to
decomposition of soil organic matter and N inputs from senesced grass litter, as well as asymbiotic fixation of N
from the atmosphere. The simulations used the same weather, soil, and synthetic N fertilizer data as discussed
under the Tier 3 Approach in the Mineral Cropland Soils section. Mineral N fertilization rates are based on data
from the Carbon Sequestration Rural Appraisals (CSRA) conducted by the USDA-NRCS (USDA-NRCS, unpublished
data). The CSRA was a solicitation of expert knowledge from USDA-NRCS staff throughout the United States to
support the Inventory. Biological N fixation is simulated within DayCent, and therefore is not an input to the
model.
Manure N deposition from grazing animals in PRP systems (i.e., PRP manure N) is a key input of N to grasslands.
The amounts of PRP manure N applied on non-federal grasslands for each NRI survey location are based on the
amount of N excreted by livestock in PRP systems that is estimated in the Manure Management section (See
Section 5.2 and Annex 3.10). The total amount of N excreted in each county is divided by the grassland area to
estimate the N input rate associated with PRP manure. The resulting rates are a direct input into the DayCent
simulations. The N input is subdivided between urine and dung based on a 50:50 split. DayCent simulations of non-
federal grasslands accounted for approximately 61 percent of total PRP manure N in aggregate across the
country.25 The remainder of the PRP manure N in each state is assumed to be excreted on federal grasslands, and
the N20 emissions are estimated using the IPCC (2006) Tier 1 method.
Biosolids (i.e., treated sewage sludge) are assumed to be applied on grasslands. Application of biosolids is
estimated from data compiled by EPA (1993,1999, 2003), McFarland (2001), and NEBRA (2007) (see Section 7.2
Wastewater Treatment for a detailed discussion of the methodology for estimating treated sewage sludge
available for land application application). Biosolids data are only available at the national scale, and it is not
possible to associate application with specific soil conditions and weather at NRI survey locations. Therefore,
DayCent could not be used to simulate the influence of biosolids on N20 emissions from grassland soils, and
consequently, emissions from biosolids are estimated using the IPCC (2006) Tier 1 method.
Soil N20 emission estimates from DayCent are adjusted using a structural uncertainty estimator accounting for
uncertainty in model algorithms and parameter values (Del Grosso et al. 2010). There is also sampling uncertainty
for the NRI survey that is propagated through the estimate with replicate sampling weights associated with the
survey. N20 emissions for the PRP manure N deposited on federal grasslands and applied biosolids N are estimated
using the Tier 1 method by multiplying the N input by the default emission factor. Emissions from manure N are
estimated at the state level and aggregated to the entire country, but emissions from biosolids N are calculated
exclusively at the national scale. Further elaboration on the methodology and data used to estimate N20 emissions
from mineral soils are described in Annex 3.12.
Soil N20 emissions and 95 percent confidence intervals are estimated for each year between 1990 and 2015 based
on the Tier 1 and 3 methods, with the exception of biosolids (discussed below). Emissions from 2016 to 2019 are
estimated using a splicing method as described in Box 5-4. As with croplands, estimates for 2016 to 2019 will be
25 A small amount of PRP N (less than 1 percent) is deposited in grazed pasture that is in rotation with annual crops, and is
reported in the grassland N20 emissions.
Agriculture 5-41

-------
recalculated in a future Inventory when new NRI data are released by USDA. Biosolids application data are
compiled through 2019 in this Inventory, and therefore soil N20 emissions and confidence intervals are estimated
using the Tier 1 method for all years in the time series without application of the splicing method.
Tier 1 Approach for Drainage of Organic Soils in Croplands and Grasslands
The IPCC (2006) Tier 1 method is used to estimate direct N20 emissions due to drainage of organic soils in
croplands and grasslands at a state scale. State-scale estimates of the total area of drained organic soils are
obtained from the 2015 NRI (USDA-NRCS 2018a) using soils data from the Soil Survey Geographic Database
(SSURGO) (Soil Survey Staff 2019). Temperature data from the PRISM Climate Group (PRISM 2018) are used to
subdivide areas into temperate and tropical climates according to the climate classification from IPCC (2006). To
estimate annual emissions, the total temperate area is multiplied by the IPCC default emission factor for
temperate regions, and the total tropical area is multiplied by the IPCC default emission factor for tropical regions
(IPCC 2006). Annual NRI data are only available between 1990 and 2015, but the time series was adjusted using
data from the Forest Inventory and Analysis Program (USFS 2019) in order to estimate emissions from 2016 to
2018. The land representation data have not been updated for this Inventory so the amount of drained organic
soils is assumed to be the same in 2019 as the estimated areas in 2018, and consequently the emissions in 2019
are also assumed to the same as 2018. Further elaboration on the methodology and data used to estimate N20
emissions from organic soils are described in Annex 3.12.
Total Direct N20 Emissions from Cropland and Grassland Soils
Annual direct emissions from the Tier 1 and 3 approaches for mineral and drained organic soils occurring in both
croplands and grasslands are summed to obtain the total direct N20 emissions from agricultural soil management
(see Table 5-16 and Table 5-17).
Indirect N20 Emissions Associated with Nitrogen Management in Cropland and
Grasslands
Indirect N20 emissions occur when mineral N applied or made available through anthropogenic activity is
transported from the soil either in gaseous or aqueous forms and later converted into N20. There are two
pathways leading to indirect emissions. The first pathway results from volatilization of N as NOx and NH3 following
application of synthetic fertilizer, organic amendments (e.g., manure, biosolids), and deposition of PRP manure.
Nitrogen made available from mineralization of soil organic matter and residue, including N incorporated into
crops and forage from symbiotic N fixation, and input of N from asymbiotic fixation also contributes to volatilized
N emissions. Volatilized N can be returned to soils through atmospheric deposition, and a portion of the deposited
N is emitted to the atmosphere as N20. The second pathway occurs via leaching and runoff of soil N (primarily in
the form of N03") that is made available through anthropogenic activity on managed lands, mineralization of soil
organic matter and residue, including N incorporated into crops and forage from symbiotic N fixation, and inputs of
N into the soil from asymbiotic fixation. The N03" is subject to denitrification in water bodies, which leads to N20
emissions. Regardless of the eventual location of the indirect N20 emissions, the emissions are assigned to the
original source of the N for reporting purposes, which here includes croplands and grasslands.
Tier 1 and 3 Approaches for Indirect N2O Emissions from Atmospheric Deposition of Volatilized N
The Tier 3 DayCent model and IPCC (2006) Tier 1 methods are combined to estimate the amount of N that is
volatilized and eventually emitted as N20. DayCent is used to estimate N volatilization for land areas whose direct
emissions are simulated with DayCent (i.e., most commodity and some specialty crops and most grasslands). The N
inputs included are the same as described for direct N20 emissions in the Tier 3 Approach for Mineral Cropland
Soils and Direct N20 Emissions from Grassland Soils sections. Nitrogen volatilization from all other areas is
estimated using the Tier 1 method with default IPCC fractions for N subject to volatilization (i.e., N inputs on
5-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
croplands not simulated by DayCent, PRP manure N excreted on federal grasslands, and biosolids [i.e., treated
sewage sludge] application on grasslands).
The IPCC (2006) default emission factor is multiplied by the amount of volatilized N generated from both DayCent
and Tier 1 methods to estimate indirect N20 emissions occurring following re-deposition of the volatilized N (see
Table 5-19). Further elaboration on the methodology and data used to estimate indirect N20 emissions are
described in Annex 3.12.
Tier 1 and 3 Approaches for Indirect N2O Emissions from Leaching/Runoff
As with the calculations of indirect emissions from volatilized N, the Tier 3 DayCent model and IPCC (2006) Tier 1
method are combined to estimate the amount of N that is subject to leaching and surface runoff into water bodies,
and eventually emitted as N20. DayCent is used to simulate the amount of N transported from lands in the Tier 3
Approach. Nitrogen transport from all other areas is estimated using the Tier 1 method and the IPCC (2006) default
factor for the proportion of N subject to leaching and runoff associated with N applications on croplands that are
not simulated by DayCent, applications of biosolids on grasslands, and PRP manure N excreted on federal
grasslands.
For both the DayCent Tier 3 and IPCC (2006) Tier 1 methods, nitrate leaching is assumed to be an insignificant
source of indirect N20 in cropland and grassland systems in arid regions, as discussed in IPCC (2006). In the United
States, the threshold for significant nitrate leaching is based on the potential evapotranspiration (PET) and rainfall
amount, similar to IPCC (2006), and is assumed to be negligible in regions where the amount of precipitation does
not exceed 80 percent of PET (Note: All irrigated systems are assumed to have significant amounts of leaching of N
even in drier climates).
For leaching and runoff data estimated by the Tier 3 and Tier 1 approaches, the IPCC (2006) default emission factor
is used to estimate indirect N20 emissions that occur in groundwater and waterways (see Table 5-19). Further
elaboration on the methodology and data used to estimate indirect N20 emissions are described in Annex 3.12.
Indirect soil N20 emissions from 2016 to 2019 are estimated using the splicing method that is described in Box 5-4.
As with the direct N20 emissions, the time series will be recalculated in a future Inventory report when new
activity data are compiled.
Uncertainty and Time-Series Consistency
Uncertainty is estimated for each of the following five components of N20 emissions from agricultural soil
management: (1) direct emissions simulated by DayCent; (2) the components of indirect emissions (N volatilized
and leached or runoff) simulated by DayCent; (3) direct emissions estimated with the IPCC (2006) Tier 1 method;
(4) the components of indirect emissions (N volatilized and leached or runoff) estimated with the IPCC (2006) Tier
1 method; and (5) indirect emissions estimated with the IPCC (2006) Tier 1 method. Uncertainty in direct emissions
as well as the components of indirect emissions that are estimated from DayCent are derived from a Monte Carlo
Analysis (consistent with IPCC Approach 2), addressing uncertainties in model inputs and structure (i.e., algorithms
and parameterization) (Del Grosso et al. 2010). For 2016 to 2019, there is additional uncertainty propagated
through the Monte Carlo Analysis associated with the splicing method (See Box 5-4).
Simple error propagation methods (IPCC 2006) are used to derive confidence intervals for direct emissions
estimated with the IPCC (2006) Tier 1 method, the proportion of volatilization and leaching or runoff estimated
with the IPCC (2006) Tier 1 method, and indirect N20 emissions. Uncertainty in the splicing method is also included
in the error propagation for 2016 to 2019 (see Box 5-4). Additional details on the uncertainty methods are
provided in Annex 3.12.
Table 5-20 shows the combined uncertainty for direct soil N20 emissions. The estimated emissions ranges from 31
percent below to 31 percent above the 2019 emission estimate of 290.4 MMT C02 Eq. The combined uncertainty
for indirect soil N20 emissions ranges from 71 percent below to 154 percent above the 2019 estimate of 54.2 MMT
C02 Eq.
Agriculture 5-43

-------
Table 5-20: Quantitative Uncertainty Estimates of N2O Emissions from Agricultural Soil
Management in 2019 (MMT CO2 Eq. and Percent)


2019 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate


(MMT C02 Eq.)
(MMT CO:
1 Eq-)
(%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Direct Soil N20 Emissions
N20
290.4
200.7
380.1
-31% 31%
Indirect Soil N20 Emissions
n2o
54.2
16.0
137.5
-71% 154%
Note: Due to lack of data, uncertainties in PRP manure N production, other organic fertilizer amendments, and
biosolids (i.e., treated sewage sludge) amendments to soils are currently treated as certain; these sources of
uncertainty will be included in future Inventory reports.
Additional uncertainty is associated with an incomplete estimation of N20 emissions from managed croplands and
grasslands in Hawaii and Alaska. The Inventory currently includes the N20 emissions from mineral fertilizer and PRP
N additions in Alaska and Hawaii, and drained organic soils in Hawaii. Land areas used for agriculture in Alaska and
Hawaii are small relative to major crop commodity states in the conterminous United States, so the emissions are
likely to be small for the other sources of N (e.g., crop residue inputs), which are not currently included in the
Inventory.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
QA/QC and Verification
General (Tier 1) and category-specific (Tier 2) QA/QC activities were conducted consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. DayCent results for N20 emissions and N03" leaching are compared with field data
representing various cropland and grassland systems, soil types, and climate patterns (Del Grosso et al. 2005; Del
Grosso et al. 2008), and further evaluated by comparing the model results to emission estimates produced using
the IPCC (2006) Tier 1 method for the same sites. Nitrous oxide measurement data for cropland are available for
64 sites representing 796 different combinations of fertilizer treatments and cultivation practices, and
measurement data for grassland are available for 13 sites representing 36 different management treatments.
Nitrate leaching data are available for 12 sites, representing 279 different combinations of fertilizer treatments and
tillage practices. In general, DayCent predicted N20 emission and nitrate leaching for these sites reasonably well.
See Annex 3.12 for more detailed information about the comparisons.
Spreadsheets containing input data and probability distribution functions required for DayCent simulations of
croplands and grasslands and unit conversion factors have been checked, in addition to the program scripts that
are used to run the Monte Carlo uncertainty analysis. Links between spreadsheets have also been checked,
updated, and corrected when necessary. Spreadsheets containing input data, emission factors, and calculations
required for the Tier 1 method have been checked and updated as needed.
Recalculations Discussion
One improvement has been implemented in this Inventory leading to the need for recalculations. This
improvement was an update to the time series of PRP and manure N available for application to soils, in order to
be consistent with the data generated for the Manure Management section of this Inventory. The surrogate data
method was also applied to re-estimate N20 emissions from 2016 to 2018. These changes resulted in an average
increase in emissions of 0.1 percent from 1990 to 2018 relative to the previous Inventory.
5-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Planned Improvements
A key improvement for a future Inventory will be to incorporate additional management activity data from the
USDA-NRCS Conservation Effects Assessment Project survey. This survey has compiled new data in recent years
that will be available for the Inventory analysis by next year. The latest land use data will also be incorporated from
the USDA National Resources Inventory and related management data from USDA-ERS ARMS surveys.
Several planned improvements are underway associated with improving the DayCent biogeochemical model.
These improvements include a better representation of plant phenology, particularly senescence events following
grain filling in crops. In addition, crop parameters associated with temperature and water stress effects on plant
production will be further improved in DayCent with additional model calibration. Model development is
underway to represent the influence of nitrification inhibitors and slow-release fertilizers (e.g., polymer-coated
fertilizers) on N20 emissions. Experimental study sites will continue to be added for quantifying model structural
uncertainty. Studies that have continuous (daily) measurements of N20 (e.g., Scheer et al. 2013) will be given
priority.
Improvements are underway to simulate crop residue burning in the DayCent model based on the amount of crop
residues burned according to the data that is used in the Field Burning of Agricultural Residues source category
(see Section 5.7). Alaska and Hawaii are not included for all sources in the current Inventory for agricultural soil
management, with the exception of N20 emissions from drained organic soils in croplands and grasslands for
Hawaii, synthetic fertilizer and PRP N amendments for grasslands in Alaska and Hawaii. There is also an
improvement based on updating the Tier 1 emission factor for N20 emissions from drained organic soils by using
the revised factor in the 2013 Supplement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories:
Wetlands (IPCC 2013).
In addition, there is a planned improvement associated with implementation of the Tier 1 method. Specifically, soil
N20 emissions will be estimated and reported for N mineralization from soil organic matter decomposition that is
accelerated with Forest Land Converted to Cropland and Grassland Converted to Cropland. A review of available
data on biosolids (i.e., treated sewage sludge) application will also be undertaken to improve the distribution of
biosolids application on croplands, grasslands and settlements.
These improvements are expected to be completed for the next full Inventory analysis (i.e., 2022 submission to the
UNFCCC, 1990 through 2020 Inventory). However, the timeline may be extended if there are insufficient resources
to fund all or part of these planned improvements.
5.5 Liming (CRF Source Category 3G)
Crushed limestone (CaC03) and dolomite (CaMg(C03)2) are added to soils by land managers to increase soil pH
(i.e., to reduce acidification). Carbon dioxide emissions occur as these compounds react with hydrogen ions in
soils. The rate of degradation of applied limestone and dolomite depends on the soil conditions, soil type, climate
regime, and whether limestone or dolomite is applied. Emissions from limestone and dolomite that are used in
industrial processes (e.g., cement production, glass production, etc.) are reported in the IPPU chapter. Emissions
from liming of soils have fluctuated between 1990 and 2019 in the United States, ranging from 2.2 MMT C02 Eq. to
6.0 MMT C02 Eq. across the entire time series. In 2019, liming of soils in the United States resulted in emissions of
2.4 MMT C02 Eq. (0.7 MMT C), representing a 52 percent decrease in emissions since 1990 (see Table 5-21 and
Table 5-22). The trend is driven by variation in the amount of limestone and dolomite applied to soils over the time
period.
Table 5-21: Emissions from Liming (MMT CO2 Eq.)
Source
1990
2005
2015
2016
2017
2018
2019
Limestone
4.1
3.9
3.5
2.8
2.9
2.0
2.2
Dolomite
0.6
0.4
0.3
0.3
0.2
0.2
0.2
Agriculture 5-45

-------
Total	<17	43	3.7 3.1 3.1	2.2	2.4
Note: Totals may not sum due to independent rounding.
Table 5-22: Emissions from Liming (MMT C)
Source 1990

2005

2015 2016 2017 2018 2019
Limestone 1.1
Dolomite 0.2

1.1
0.1

0.9 0.8 0.8 0.6 0.6
0.1 0.1 0.1 0.1 0.1
Total 1.3

1.2

o
to
o
CO
o
CO
o
o
rH
Note: Totals may not sum due to independent rounding.
Methodology
Carbon dioxide emissions from application of limestone and dolomite to soils were estimated using a Tier 2
methodology consistent with IPCC (2006). The annual amounts of limestone and dolomite, which are applied to
soils (see Table 5-23), were multiplied by C02 emission factors from West and McBride (2005). These country-
specific emission factors (0.059 metric ton C/metric ton limestone, 0.064 metric ton C/metric ton dolomite) are
lower than the IPCC default emission factors because they account for the portion of carbonates that are
transported from soils through hydrological processes and eventually deposited in ocean basins (West and
McBride 2005). This analysis of lime dissolution is based on studies in the Mississippi River basin, where the vast
majority of lime application occurs in the United States (West 2008). Moreover, much of the remaining lime
application is occurring under similar precipitation regimes, and so the emission factors are considered a
reasonable approximation for all lime application in the United States (West 2008) (See Box 5-5).
The annual application rates of limestone and dolomite were derived from estimates and industry statistics
provided in the Minerals Yearbook (Tepordei 1993 through 2006; Willett 2007a, 2007b, 2009, 2010, 2011a, 2011b,
2013a, 2014, 2015, 2016, 2017, 2020a), as well as preliminary data that will eventually be published in the
Minerals Yearbook for the latter part of the time series (Willett 2019, 2020b). Data for the final year of the
inventory is based on the Mineral Industry Surveys, as discussed below (USGS 2020). The U.S. Geological Survey
(USGS; U.S. Bureau of Mines prior to 1997) compiled production and use information through surveys of crushed
stone manufacturers. However, manufacturers provided different levels of detail in survey responses so the
estimates of total crushed limestone and dolomite production and use were divided into three components: (1)
production by end-use, as reported by manufacturers (i.e., "specified" production); (2) production reported by
manufacturers without end-uses specified (i.e., "unspecified" production); and (3) estimated additional production
by manufacturers who did not respond to the survey (i.e., "estimated" production).
Box 5-5: Comparison of the Tier 2 U.S. Inventory Approach and IPCC (2006) Default Approach
Emissions from liming of soils were estimated using a Tier 2 methodology based on emission factors specific to
the United States that are lower than the IPCC (2006) emission default factors. Most lime application in the
United States occurs in the Mississippi River basin, or in areas that have similar soil and rainfall regimes as the
Mississippi River basin. Under these conditions, a significant portion of dissolved agricultural lime leaches
through the soil into groundwater. Groundwater moves into channels and is transported to larger rives and
eventually the ocean where CaC03 precipitates to the ocean floor (West and McBride 2005). The U.S.-specific
emission factors (0.059 metric ton C/metric ton limestone and 0.064 metric ton C/metric ton dolomite) are
about half of the IPCC (2006) emission factors (0.12 metric ton C/metric ton limestone and 0.13 metric ton
C/metric ton dolomite). For comparison, the 2019 U.S. emission estimate from liming of soils is 2.4 MMT C02
Eq. using the country-specific factors. In contrast, emissions would be estimated at 5.0 MMT C02 Eq. using the
IPCC (2006) default emission factors.
5-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Data on "specified" limestone and dolomite amounts were used directly in the emission calculation because the
end use is provided by the manufacturers and can be used to directly determine the amount applied to soils.
However, it is not possible to determine directly how much of the limestone and dolomite is applied to soils for
manufacturer surveys in the "unspecified" and "estimated" categories. For these categories, the amounts of
crushed limestone and dolomite applied to soils were determined by multiplying the percentage of total
"specified" limestone and dolomite production that is applied to soils, by the total amounts of "unspecified" and
"estimated" limestone and dolomite production. In other words, the proportion of total "unspecified" and
"estimated" crushed limestone and dolomite that was applied to soils is proportional to the amount of total
"specified" crushed limestone and dolomite that was applied to soils.
In addition, data were not available for 1990,1992, and 2019 on the fractions of total crushed stone production
that were limestone and dolomite, and on the fractions of limestone and dolomite production that were applied to
soils. To estimate the 1990 and 1992 data, a set of average fractions were calculated using the 1991 and 1993
data. These average fractions were applied to the quantity of "total crushed stone produced or used" reported for
1990 and 1992 in the 1994 Minerals Yearbook (Tepordei 1996). To estimate 2019 data, 2018 fractions were applied
to a 2019 estimate of total crushed stone presented in the USGS Mineral Industry Surveys: Crushed Stone and Sand
and Gravel in the First Quarter of 2020 (USGS 2020).
The primary source for limestone and dolomite activity data is the Minerals Yearbook, published by the Bureau of
Mines through 1996 and by the USGS from 1997 to the present. In 1994, the "Crushed Stone" chapter in the
Minerals Yearbook began rounding (to the nearest thousand metric tons) quantities for total crushed stone
produced or used. It then reported revised (rounded) quantities for each of the years from 1990 to 1993. In order
to minimize the inconsistencies in the activity data, these revised production numbers have been used in all of the
subsequent calculations.
Table 5-23: Applied Minerals (MMT)
Mineral
1990
2005
2015
2016
2017
2018
2019
Limestone
19.0
18.1
16.0
13.0
13.4
9.4
10.2
Dolomite
2.4
1.9
1.2
1.1
0.8
0.9
1.0
Uncertainty and Time-Series Consistency
Uncertainty regarding the amount of limestone and dolomite applied to soils was estimated at ±15 percent with
normal densities (Tepordei 2003; Willett 2013b). Analysis of the uncertainty associated with the emission factors
included the fraction of lime dissolved by nitric acid versus the fraction that reacts with carbonic acid, and the
portion of bicarbonate that leaches through the soil and is transported to the ocean. Uncertainty regarding the
time associated with leaching and transport was not addressed in this analysis, but is assumed to be a relatively
small contributor to the overall uncertainty (West 2005). The probability distribution functions for the fraction of
lime dissolved by nitric acid and the portion of bicarbonate that leaches through the soil were represented as
triangular distributions between ranges of zero and 100 percent of the estimates. The uncertainty surrounding
these two components largely drives the overall uncertainty.
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the uncertainty in C02 emissions from
liming. The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 5-24. Carbon
dioxide emissions from carbonate lime application to soils in 2019 were estimated to be between -0.27 and 4.61
MMT C02 Eq. at the 95 percent confidence level. This confidence interval represents a range of 111 percent below
to 88 percent above the 2019 emission estimate of 2.4 MMT C02 Eq. Note that there is a small probability of a
negative emissions value leading to a net uptake of C02 from the atmosphere. Net uptake occurs due to the
dominance of the carbonate lime dissolving in carbonic acid rather than nitric acid (West and McBride 2005).
Agriculture 5-47

-------
Table 5-24: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Liming
(MMT CO2 Eq. and Percent)
Source
„ 2019 Emission Estimate
Gas
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)


Lower Upper
Bound Bound
Lower Upper
Bound Bound
Liming
C02 2.4
(0.27) 4.61
-111% 88%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
QA/QC and Verification
A source-specific QA/QC plan for liming has been developed and implemented, consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. The quality control effort focused on the Tier 1 procedures for this Inventory. No
errors were found.
Recalculations Discussion
An adjustment was made in the current Inventory to improve the results; limestone and dolomite application data
for 2018 were updated with the recently published data from USGS (2020), rather than approximated by a ratio
method, which was used in the previous Inventory. With this revision in the activity data, the emissions decreased
by 28.6 percent for 2018 relative to the previous Inventory.
5.6 Urea Fertilization (CRF Source Category 3H)
The use of urea (CO(NH2)2) as a fertilizer leads to greenhouse gas emissions through the release of C02that was
fixed during the production of urea. In the presence of water and urease enzymes, urea that is applied to soils as
fertilizer is converted into ammonium (NH4+), hydroxyl ion (OH), and bicarbonate (HC03 ). The bicarbonate then
evolves into C02 and water. Emissions from urea fertilization in the United States is 5.3 MMT C02 Eq. (1.5 MMT C)
in 2019 (Table 5-25 and Table 5-26). Carbon dioxide emissions have increased by 121 percent between 1990 and
2019 due to an increasing amount of urea that is applied to soils. The variation in emissions across the time series
is driven by differences in the amounts of fertilizer applied to soils each year. Carbon dioxide emissions associated
with urea that is used for non-agricultural purposes are reported in the IPPU chapter (Section 4.6).
Table 5-25: CO2 Emissions from Urea Fertilization (MMT CO2 Eq.)
Source	1990	2005	2015 2016 2017 2018 2019
Urea Fertilization	2.4 J 3.5	4.7 4.9 5.1 5.2 5.3
Table 5-26: CO2 Emissions from Urea Fertilization (MMT C)
Source	1990	2005 2015 2016 2017 2018 2019
Urea Fertilization	0.7	1.0	1.3 1.3 1.4 1.4 1.5
5-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
Carbon dioxide emissions from the application of urea to agricultural soils were estimated using the IPCC (2006)
Tier 1 methodology. The method assumes that C in the urea is released after application to soils and converted to
C02. The annual amounts of urea applied to croplands (see Table 5-27) were derived from the state-level fertilizer
sales data provided in Commercial Fertilizer reports (TVA 1991,1992,1993,1994; AAPFCO 1995 through 2018).26
These amounts were multiplied by the default IPCC (2006) emission factor (0.20 metric tons of C per metric ton of
urea), which is equal to the C content of urea on an atomic weight basis. The calculations were made using a
Monte Carlo analysis as described in the Uncertainty section below.
Fertilizer sales data are reported in fertilizer years (July previous year through June current year) so a calculation
was performed to convert the data to calendar years (January through December). According to monthly fertilizer
use data (TVA 1992b), 35 percent of total fertilizer used in any fertilizer year is applied between July and December
of the previous calendar year, and 65 percent is applied between January and June of the current calendar year.
Fertilizer sales data for the 2016 through 2019 fertilizer years were not available for this Inventory. Therefore, urea
application in the 2016 through 2019 fertilizer years were estimated using a linear, least squares trend of
consumption over the data from the previous five years (2011 through 2015) at the state scale. A trend of five
years was chosen as opposed to a longer trend as it best captures the current inter-state and inter-annual
variability in consumption. State-level estimates of C02 emissions from the application of urea to agricultural soils
were summed to estimate total emissions for the entire United States. The fertilizer year data is then converted
into calendar year (Table 5-27) data using the method described above.
Table 5-27: Applied Urea (MMT)

1990
2005
2015
2016
2017
2018
2019
Urea Fertilizer3
3.3
4.8
6.4
6.7
6.9
7.1
7.3
a These numbers represent amounts applied to all agricultural land, including Cropland Remaining
Cropland, Land Converted to Cropland, Grassland Remaining Grassland, Land Converted to Grassland,
Settlements Remaining Settlements, Land Converted to Settlements, Forest Land Remaining Forest Land
and Land Converted to Forest Land, as it is not currently possible to apportion the data by land-use
category.
Uncertainty and Time-Series Consistency
An Approach 2 Monte Carlo analysis is conducted as described by the IPCC (2006). The largest source of
uncertainty is the default emission factor, which assumes that 100 percent of the C in CO(NH2)2 applied to soils is
emitted as C02. The uncertainty surrounding this factor incorporates the possibility that some of the C may not be
emitted to the atmosphere, and therefore the uncertainty range is set from 50 percent emissions to the maximum
emission value of 100 percent using a triangular distribution. In addition, urea consumption data have uncertainty
that is represented as a normal density. Due to the highly skewed distribution of the resulting emissions from the
Monte Carlo uncertainty analysis, the estimated emissions are based on the analytical solution to the equation,
and the confidence interval is approximated based on the values at 2.5 and 97.5 percentiles.
Carbon dioxide emissions from urea fertilization of agricultural soils in 2019 are estimated to be between 3.06 and
5.51 MMT C02 Eq. at the 95 percent confidence level. This indicates a range of 43 percent below to 3 percent
above the 2019 emission estimate of 5.3 MMT C02 Eq. (Table 5-28).
26 The amount of urea consumed for non-agricultural purposes in the United States is reported in the Industrial Processes and
Product Use chapter, Section 4.6 Urea Consumption for Non-Agricultural Purposes.
Agriculture 5-49

-------
Table 5-28: Quantitative Uncertainty Estimates for CO2 Emissions from Urea Fertilization
(MMT CO2 Eq. and Percent)


Uncertainty Range Relative to Emission
Source Gas
2019 Emission Estimate
Estimate3

(MMT CO? Eq.)
(MMT CO? Eq.)
(%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
Urea Fertilization C02
5.3
3.06 5.51
-43% +3%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
There are additional uncertainties that are not quantified in this analysis. There is uncertainty surrounding the
assumptions underlying conversion of fertilizer years to calendar years. These uncertainties are negligible over
multiple years because an over- or under-estimated value in one calendar year is addressed with a corresponding
increase or decrease in the value for the subsequent year. In addition, there is uncertainty regarding the fate of C
in urea that is incorporated into solutions of urea ammonium nitrate (UAN) fertilizer. Emissions of C02 from UAN
applications to soils are not estimated in the current Inventory (see Planned Improvements).
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies are described in the Introduction and
Methodology sections.
QA/QC and Verification
A source-specific QA/QC plan for Urea Fertilization has been developed and implemented, consistent with the U.S.
Inventory QA/QC plan. One quality control issue was raised by the expert review team (ERT) from the UNFCCC for
this emission source. In the previous (i.e., 1990 through 2018) Inventory, estimates of C02 emissions were based
on the results from the Monte Carlo uncertainty analysis. Specifically, the mode from the Monte Carlo uncertainty
analysis was used as the most probable estimate of emissions. The mode differs from the analytical solution to the
equation due to the pattern in the probability distribution for C02 emissions from the Monte Carlo uncertainty
analysis, which combined a normal density for the urea application data with the right triangle distribution for the
emission factor. For this Inventory, the analytical solution has been adopted as the estimate of C02 emissions for
urea fertilization to be consistent with recommendations from ERT. The ERT considered the analytical solution to
be more representative of the emissions than the mode from the Monte Carlo uncertainty analysis.
Recalculations Discussion
Emissions estimates were derived directly from the Monte Carlo uncertainty analysis in the previous Inventory as
discussed in the QA/QC and Verification section. For this Inventory, the entire time series was recalculated using
the analytical solution rather than the mode from the Monte Carlo uncertainty analysis. This change in emission
estimates averaged about 15 percent higher across the time series compared to the previous Inventory.
Planned Improvements
A key planned improvement is to incorporate Urea Ammonium Nitrate (UAN) in the estimation of Urea C02
emissions. Activity data for UAN have been identified, but additional information is needed to fully incorporate this
type of fertilizer into the analysis, which will be completed in a future Inventory.
5-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
5.7 Field Burning of Agricultural Residues (CRF
Source Category 3F)
Crop production creates large quantities of agricultural crop residues, which farmers manage in a variety of ways.
For example, crop residues can be left in the field and possibly incorporated into the soil with tillage; collected and
used as fuel, animal bedding material, supplemental animal feed, or construction material; composted and applied
to soils; transported to landfills; or burned in the field. Field burning of crop residues is not considered a net source
of C02 emissions because the C released to the atmosphere as C02 during burning is reabsorbed during the next
growing season by the crop. However, crop residue burning is a net source of CH4, N20, CO, and NOx, which are
released during combustion.
In the United States, field burning of agricultural residues commonly occurs in southeastern states, the Great
Plains, and the Pacific Northwest (McCarty 2011). The primary crops that are managed with residue burning
include corn, cotton, lentils, rice, soybeans, sugarcane and wheat (McCarty 2009). In 2019, CH4 and N20 emissions
from field burning of agricultural residues were 0.4 MMT C02 Eq. (17 kt) and 0.2 MMT C02 Eq. (1 kt), respectively
(Table 5-29 and Table 5-30). Annual emissions of CH4 and N20 have increased from 1990 to 2019 by 14 percent and
16 percent, respectively. The increase in emissions over time is partly due to higher yielding crop varieties with
larger amounts of residue production and fuel loads, but also linked with an increase in the area burned for some
of the crop types.
Table 5-29: ChU and N2O Emissions from Field Burning of Agricultural Residues (MMT CO2
Eq.)
Gas/Crop Type
1990
2005
2015
2016
2017
2018
2019
ch4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
Maize
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Rice
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wheat
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
+
+
+
+
+
+
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
+
+
+
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
n2o
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Maize
+
+
0.1
0.1
0.1
0.1
0.1
Rice
+
+
+
+
+
+
+
Wheat
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Agriculture 5-51

-------
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
+
+
+
+
+
+
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
+
+
+
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
Total
0.5
0.6
0.6
0.6
0.6
0.6
0.6
+ Does not exceed 0.05 MMT C02 Eq.
Note: Totals may not sum due to independent rounding.
Table 5-30: ChU, N2O, CO, and NOx Emissions from Field Burning of Agricultural Residues
(kt)
Gas/Crop Type
1990
2005
2015
2016
2017
2018
2019
ch4
15
17
18
17
17
17
17
Maize
2
4
5
5
5
5
5
Rice
3
3
3
2
3
2
3
Wheat
6
6
5
5
5
5
5
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
1
2
1
1
1
1
1
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
1
2
2
2
2
2
2
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
n2o
1
1
1
1
1
1
1
5-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Maize
+
+
+
+
+
+
+
Rice
+
+
+
+
+
+
+
Wheat
+
+
+
+
+
+
+
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
+
+
+
+
+
+
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
+
+
+
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
CO
315
363
342
340
339
338
337
NOx
13
15
14
14
14
14
14
+ Does not exceed 0.5 kt.
Note: Totals may not sum due to independent rounding.
Methodology
A country-specific Tier 2 method is used to estimate greenhouse gas emissions from field burning of agricultural
residues from 1990 to 2014 (for more details comparing the country-specific approach to the IPCC (2006) default
approach, see Box 5-6), and a data splicing method with a linear extrapolation is applied to complete the emissions
time series from 2015 to 2019. The following equation is used to estimate the amounts of C and N released
(Ri, where i is C or N) from burning.
Ri = CPx RCR x DMF x Ftx FB x CE
where,
Crop Production (CP)
Residue: Crop Ratio (RCR)
Dry Matter Fraction (DMF)
Fraction C or N (Ft)
Fraction Burned (FB)
FB =
AB
CAH
Annual production of crop, by state, kt crop production
Amount of residue produced per unit of crop production, kt residue/kt crop
production
Amount of dry matter per unit of residue biomass for a crop, kt residue dry
matter/ kt residue biomass
Fraction of C or N per unit of dry matter for a crop, kt C or N /kt residue dry
matter
Proportion of residue biomass consumed, unitless
Agriculture 5-53

-------
Combustion Efficiency (CE) = Proportion of C or N released with respect to the total amount of C or N
available in the burned material, respectively, unitless
Area Burned (AB)	= Total area of crop burned, by state, ha
Crop Area Harvested (CAH) = Total area of crop harvested, by state, ha
Crop production data are available by state and year from USDA (2019) for twenty-one crops that are burned in
the conterminous United States, including maize, rice, wheat, barley, oats, other small grains, sorghum, cotton,
grass hay, legume hay, peas, sunflower, tobacco, vegetables, chickpeas, dry beans, lentils, peanuts, soybeans,
potatoes, and sugarbeets.27 Crop area data are based on the 2015 National Resources Inventory (NRI) (USDA-NRCS
2018). In order to estimate total crop production, the crop yield data from USDA Quick Stats crop yields is
multiplied by the NRI crop areas. The production data for the crop types are presented in Table 5-31. Alaska and
Hawaii are not included in the current analysis, but there is a planned improvement to estimate residue burning
emissions for these two states in a future Inventory.
The amount of elemental C or N released through oxidation of the crop residues is used in the following equation
to estimate the amount of CH4, CO, N20, and NOx emissions (Eg, where g is the specific gas, i.e., CH4, CO, N20, and
NOx) from the Field Burning of Agricultural Residues:
Eg = X EFg x CF
where,
Emission ratio [EFg)	= emission ratio by gas, g CH4-C or CO-C/g C released, or g N20-N or NOx-
N/g N released
Conversion Factor (CF)	= conversion by molecular weight ratio of CH4-C to C (16/12), CO-C to C
(28/12), N20-N to N (44/28), or NOx-N to N (30/14)
Box 5-6: Comparison of Tier 2 U.S. Inventory Approach and IPCC (2006) Default Approach
Emissions from Field Burning of Agricultural Residues are calculated using a Tier 2 methodology that is based on
the method developed by the IPCC/UNEP/OECD/IEA (1997). The rationale for using the IPCC/UNEP/OECD/IEA
(1997) approach rather than the method provided in the 2006 IPCC Guidelines is as follows: (1) the equations
from both guidelines rely on the same underlying variables (though the formats differ); (2) the IPCC (2006)
equation was developed to be broadly applicable to all types of biomass burning, and, thus, is not specific to
agricultural residues; (3) the IPCC (2006) method provides emission factors based on the dry matter content
rather emission rates related to the amount of C and N in the residues; and (4) the IPCC (2006) default factors
are provided only for four crops (corn, rice, sugarcane, and wheat) while this Inventory includes emissions from
twenty-one crops.
A comparison of the methods in the current Inventory and the default IPCC (2006) approach was undertaken for
2014 to determine the difference in estimates between the two approaches. To estimate greenhouse gas
emissions from field burning of agricultural residues using the IPCC (2006) methodology, the following
equation—cf. IPCC (2006) Equation 2.27—was used with default factors and country-specific values for mass of
fuel.
Emissions (kt) =AB x (MbX Cf) x Get x 10~6
where,
Area Burned (AB)	= Total area of crop burned (ha)
Mass of Fuel (MB x Cf) = IPCC (2006) default carbon fractions with fuel biomass consumption U.S.-
27 Sugarcane and Kentucky bluegrass (produced on farms for turf grass installations) may have small areas of burning that are
not captured in the sample of locations that were used in the remote sensing analysis (see Planned Improvements).
5-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Specific Values using NASS Statistics28 (metric tons dry matter burnt ha 1)
Emission Factor (Gef) = IPCC (2006) emission factor (g kg1 dry matter burnt)
The IPCC (2006) Tier 1 method approach resulted in 33 percent lower emissions of CH4 and 53 percent lower
emissions of N20 compared to this Inventory. In summary, the IPCC/UNEP/OECD/IEA (1997) method is
considered more appropriate for U.S. conditions because it is more flexible for incorporating country-specific
data. Emissions are estimated based on specific C and N content of the fuel, which is converted into CH4, CO,
N20 and NOx, compared to IPCC (2006) approach that is based on dry matter rather than elemental
composition.
Table 5-31: Agricultural Crop Production (kt of Product)
Crop
1990
2005
2013
2014
Maize	296,065	371,256	436,565	453,524
Rice	9,543	11,751	10,894	12,380
Wheat	79,805	68,077	67,388	62,602
Barley	9,281	5,161	4,931	5,020
Oats	5,969	2,646	1,806	2,042
Other Small Grains	2,651	2,051	1,902	2,492
Sorghum	23,687	14,382	18,680	18,436
Cotton	4,605	6,106	3,982	4,396
Grass Hay	44,150	49,880	45,588	46,852
Legume Hay	90,360	91,819	79,669	82,844
Peas	51	660	599	447
Sunflower	1,015	1,448	987	907
Tobacco	1,154	337	481	542
Vegetables	0	1,187	1,844	2,107
Chickpeas	0	5	0	0
Dry Beans	467	1,143	1,110	1,087
Lentils	0	101	72	76
Peanuts	1,856	2,176	2,072	2,735
Soybeans	56,612	86,980	94,756	110,560
Potatoes	18,924	20,026	20,234	19,175
Sugarbeets	24,951	25,635	31,890	31,737
Note: The amount of crop production has not been compiled for 2015 to 2019 so a
data splicing method is used to estimate emissions for this portion of the	time
series.
The area burned is determined based on an analysis of remote sensing products (McCarty et al. 2009, 2010, 2011).
The presence of fires has been analyzed at 3600 survey locations in the NRI from 1990 to 2002 with LANDFIRE data
products developed from 30 m Landsat imagery (LANDFIRE 2014), and from 2003 through 2014 using 1 km
Moderate Resolution Imaging Spectroradiometer imagery (MODIS) Global Fire Location Product (MCD14ML) using
combined observations from Terra and Aqua satellites (Giglio et al. 2006). A sample of states are included in the
analysis with high, medium and low burning rates for agricultural residues, including Arkansas, California, Florida,
Indiana, Iowa and Washington. The area burned is determined directly from the analysis for these states.
28 NASS yields are used to derive mass of fuel values because IPCC (2006) only provides default values for 4 of the 21 crops
included in the Inventory.
Agriculture 5-55

-------
For other states within the conterminous United States, the area burned for the 1990 through 2014 portion of the
time series is estimated from a logistical regression model that has been developed from the data collected from
the remote sensing products for the six states. The logistical regression model is used to predict occurrence of fire
events. Several variables are tested in the logistical regression including a) the historical level of burning in each
state (high, medium or low levels of burning) based on an analysis by McCarty et al. (2011), b) year that state laws
limit burning of fields, in addition to c) mean annual precipitation and mean annual temperature from a 4-
kilometer gridded product from the PRISM Climate Group (2015). A K-fold model fitting procedure is used due to
low frequency of burning and likelihood that outliers could influence the model fit. Specifically, the model is
trained with a random selection of sample locations and evaluated with the remaining sample. This process is
repeated ten times to select a model that is most common among the set of ten, and avoid models that appear to
be influenced by outliers due to the random draw of survey locations for training the model. In order to address
uncertainty, a Monte Carlo analysis is used to sample the parameter estimates for the logistical regression model
and produce one thousand estimates of burning for each crop in the remaining forty-two states included in this
Inventory. State-level area burned data are divided by state-level crop area data to estimate the percent of crop
area burned by crop type for each state. Table 5-32 shows the resulting percentage of crop residue burned at the
national scale by crop type. State-level estimates are also available upon request.
Table 5-32: U.S. Average Percent Crop Area Burned by Crop (Percent)
Crop
1990
2005
2013
2014
Maize
+%
+%
+%
+%
Rice
8%
8%
4%
6%
Wheat
1%
2%
2%
1%
Barley
1%
+%
1%
1%
Oats
1%
1%
2%
1%
Other Small Grains
1%
1%
1%
1%
Sorghum
1%
1%
1%
1%
Cotton
1%
1%
1%
1%
Grass Hay
+%
+%
+%
+%
Legume Hay
+%
+%
+%
+%
Peas
+%
+%
+%
+%
Sunflower
+%
+%
+%
+%
Tobacco
2%
2%
3%
3%
Vegetables
0%
+%
+%
+%
Chickpeas
0%
1%
0%
0%
Dry Beans
1%
1%
+%
+%
Lentils
0%
+%
+%
+%
Peanuts
3%
3%
3%
3%
Soybeans
+%
+%
1%
1%
Potatoes
+%
+%
+%
+%
Sugarbeets
+%
+%
+%
+%
+ Does not exceed 0.5 percent
Additional parameters are needed to estimate the amount of burning, including residue: crop ratios, dry matter
fractions, carbon fractions, nitrogen fractions and combustion efficiency. Residue: crop product mass ratios,
residue dry matter fractions, and the residue N contents are obtained from several sources (IPCC 2006 and sources
at bottom of Table 5-33). The residue C contents for all crops are based on IPCC (2006) default value for
herbaceous biomass. The combustion efficiency is assumed to be 90 percent for all crop types
(IPCC/UNEP/OECD/IEA 1997). See Table 5-33 for a summary of the crop-specific conversion factors. Emission ratios
and mole ratio conversion factors for all gases are based on the Revised 1996 IPCC Guidelines
(IPCC/UNEP/OECD/IEA 1997) (see Table 5-34).
5-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 5-33: Parameters for Estimating Emissions from Field Burning of Agricultural Residues
Crop
Residue/Crop
Ratio
Dry
Matter
Fraction
Carbon
Fraction
Nitrogen
Fraction
Combustion
Efficiency
(Fraction)
Maize
0.707
0.56
0.47
0.01
0.90
Rice
1.340
0.89
0.47
0.01
0.90
Wheat
1.725
0.89
0.47
0.01
0.90
Barley
1.181
0.89
0.47
0.01
0.90
Oats
1.374
0.89
0.47
0.01
0.90
Other Small Grains
1.777
0.88
0.47
0.01
0.90
Sorghum
0.780
0.60
0.47
0.01
0.90
Cotton
7.443
0.93
0.47
0.01
0.90
Grass Hay
0.208
0.90
0.47
0.02
0.90
Legume Hay
0.290
0.67
0.47
0.01
0.90
Peas
1.677
0.91
0.47
0.01
0.90
Sunflower
1.765
0.88
0.47
0.01
0.90
Tobacco
0.300
0.87
0.47
0.01
0.90
Vegetables
0.708
0.08
0.47
0.01
0.90
Chickpeas
1.588
0.91
0.47
0.01
0.90
Dry Beans
0.771
0.90
0.47
0.01
0.90
Lentils
1.837
0.91
0.47
0.02
0.90
Peanuts
1.600
0.94
0.47
0.02
0.90
Soybeans
1.500
0.91
0.47
0.01
0.90
Potatoes
0.379
0.25
0.47
0.02
0.90
Sugarbeets
0.196
0.22
0.47
0.02
0.90
Notes:
Chickpeas: IPCC (2006), Table 11.2; values are for Beans & pulses.
Cotton: Combined sources (Heitholt et al. 1992; Halevy 1976; Wells and Meredith 1984; Sadras and
Wilson 1997; Pettigrew and Meredith 1997; Torbert and Reeves 1994; Gerik et al. 1996; Brouder
and Cassmen 1990; Fritschi et al. 2003; Pettigrew et al. 2005; Bouquet and Breitenbeck 2000;
Mahroni and Aharonov 1964; Bange and Milroy 2004; Hollifield et al. 2000; Mondino et al. 2004;
Wallach etal. 1978).
Lentils: IPCC (2006), Table 11.2; Beans & pulses.
Peas: IPCC (2006), Table 11.2; values are for Beans & pulses.
Peanuts: IPCC (2006); Table 11.2; Root ratio and belowground N content values are for Root crops,
other.
Sugarbeets: IPCC (2006); Table 11.2; values are for Tubers.
Sunflower: IPCC (2006), Table 11.2; values are for Grains.
Sugarcane: combined sources (Wiedenfels 2000, Dua and Sharma 1976; Singels & Bezuidenhout
2002; Stirling et al. 1999; Sitompul et al. 2000).
Tobacco: combined sources (Beyaert 1996; Moustakas and Ntzanis 2005; Crafts-Brandner et al. 1994;
Hopkinson 1967; Crafts-Brandner et al. 1987).
Vegetables (Combination of carrots, lettuce/cabbage, melons, onions, peppers and tomatoes):
Carrots: McPharlin et al. (1992); Gibberd et al. (2003); Reid and English (2000); Peach et al. (2000);
see IPCC Tubers for R:S and N fraction.
Lettuce, cabbage: combined sources (Huett and Dettman 1991; De Pinheiro Henriques & Marcelis
2000; Huett and Dettman 1989; Peach et al. 2000; Kage et al. 2003; Tan et al. 1999; Kumar et al.
1994; MacLeod et al. 1971; Jacobs et al. 2004; Jacobs et al. 2001; Jacobs et al. 2002); values from
IPCC Grains used for N fraction.
Melons: Valantin et al. (1999); squash for R:S; IPCC Grains for N fraction.
Onion: Peach et al. (2000), Halvorson et al. (2002); IPCC (2006) Tubers for N fraction.
Peppers: combined sources (Costa and Gianquinto 2002; Marcussi et al. 2004; Tadesse et al. 1999;
Diaz-Perez et al. 2008); IPCC Grains for N fraction.
Agriculture 5-57

-------
Tomatoes: Scholberg et al. (2000a,b); Akintoye et al. (2005); values for AGR-N and BGR-N are from
Grains.
Table 5-34: Greenhouse Gas Emission Ratios and Conversion Factors
Gas
Emission Ratio
Conversion Factor
CH4:C
0.005a
16/12
CO:C
0.060a
28/12
N20:N
0.007b
44/28
NOx:N
0.121b
30/14
a Mass of C compound released (units of C) relative to
mass of total C released from burning (units of C).
b Mass of N compound released (units of N) relative to
mass of total N released from burning (units of N).
For this Inventory, new activity data on the burned areas have not been analyzed for 2015 to 2019. To complete
the emissions time series, a linear extrapolation of the trend is applied to estimate the emissions in the last five
years of the inventory. Specifically, a linear regression model with autoregressive moving-average (ARMA) errors is
used to estimate the trend in emissions over time from 1990 through 2014, and the trend is used to approximate
the CH4, N20, CO and NOx for the last five years in the time series from 2015 to 2019 (Brockwell and Davis 2016).
The Tier 2 method described previously will be applied to recalculate the emissions for the last five years in the
time series (2015 to 2019) in a future Inventory.
Uncertainty and Time-Series Consistency
Emissions are estimated using a linear regression model with autoregressive moving-average (ARMA) errors for
2019. The linear regression ARMA model produced estimates of the upper and lower bounds to quantify
uncertainty (Table 5-35), and the results are summarized in Table 5-35. Methane emissions from field burning of
agricultural residues in 2019 are between 0.35 and 0.50 MMT C02 Eq. at a 95 percent confidence level. This
indicates a range of 18 percent below and 18 percent above the 2019 emission estimate of 0.43 MMT C02 Eq.
Nitrous oxide emissions are between 0.16 and 0.22 MMT C02 Eq., or approximately 17 percent below and 17
percent above the 2019 emission estimate of 0.19 MMT C02 Eq.
Table 5-35: Approach 2 Quantitative Uncertainty Estimates for ChU and N2O Emissions from
Field Burning of Agricultural Residues (MMT CO2 Eq. and Percent)


2019 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMTC02 Eq.)
(MMT C02
Eq.)
(%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Field Burning of Agricultural
Residues
ch4
0.4
0.35
0.50
-18% 18%
Field Burning of Agricultural
Residues
n2o
0.2
0.16
0.22
-17% 17%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Due to data limitations, there are additional uncertainties in agricultural residue burning, particularly the potential
omission of burning associated with Kentucky bluegrass (produced on farms for turf grass installation) and
sugarcane.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies through time are described in the Introduction
and Methodology sections.
5-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
QA/QC and Verification
A source-specific QA/QC plan for field burning of agricultural residues is implemented with Tier 1 analyses,
consistent with the U.S. Inventory QA/QC plan outlined in Annex 8. The previous Inventory included a term for
burning efficiency that is not found in the IPCC/UNEP/OECD/IEA (1997) method. This term has been removed
based on a QA/QC initiated by the UN Expert Review Team. In addition, the combustion efficiency term has been
set to 90 percent to be consistent with the Tier 1 method in IPCC/UNEP/OECD/IEA (1997).
Recalculations Discussion
Methodological recalculations are associated with two methodological revisions, a) removing the burning
efficiency term and b) adopting the combustion efficiency value in IPCC/UNEP/OECD/IEA (1997) (See QA/QC and
Verification Section for more information). As a result of these two revisions, the emissions increased on average
across the time series by 10 percent and 9 percent for CH4 and N20, respectively. The absolute increases in
emissions are 0.4 MMT C02 Eq. and 0.2 MMT C02 Eq. for CH4 and N20, respectively.
Planned Improvements
A key planned improvement is to estimate the emissions associated with field burning of agricultural residues in
the states of Alaska and Hawaii. In addition, a new method is in development that will directly link agricultural
residue burning with the Tier 3 methods that are used in several other source categories, including Agricultural Soil
Management, Cropland Remaining Cropland, and Land Converted to Cropland chapters of the Inventory. The
method is based on simulating burning events directly within the DayCent process-based model framework using
information derived from remote sensing fire products as described in the Methodology section. This
improvement will lead to greater consistency in the methods for across sources, ensuring mass balance of C and N
in the Inventory analysis.
As previously noted in this chapter, remote sensing data were used in combination with a resource survey to
estimate non-C02 emissions and these data did not allow identification of burning of sugarcane (see Annex 5). In
addition, during the Public Review period of this current (1990 through 2019) Inventory, EPA received feedback on
this category/crop type which provided average estimates of emissions of sugarcane burning found in academic
literature. EPA plans to assess the information identified in feedback, as well as other available activity data, as
part of future inventory improvements.
Agriculture 5-59

-------
Land Usi
Forestry
This chapter provides an assessment of the greenhouse gas fluxes resulting from land use and land-use change in
the United States.1 The Intergovernmental Panel on Climate Change's 2006IPCC Guidelines for National
Greenhouse Gas Inventories (IPCC 2006) recommends reporting fluxes according to changes within and
conversions between all land-use types including: Forest Land, Cropland, Grassland, Wetlands, and Settlements (as
well as Other Land).
The greenhouse gas flux from Forest Land Remaining Forest Land is reported for all forest ecosystem carbon (C)
pools (i.e., aboveground biomass, belowground biomass, dead wood, litter, and mineral and organic soils),
harvested wood pools, and non-carbon dioxide (non-C02) emissions from forest fires, the application of synthetic
nitrogen fertilizers to forest soils, and the draining of organic soils. Fluxes from Land Converted to Forest Land are
included for aboveground biomass, belowground biomass, dead wood, litter, and C stock changes from mineral
soils, while C stock changes from drained organic soils and the non-C02 emissions from Land Converted to Forest
Land are included in the fluxes from Forest Land Remaining Forest Land as it is not currently possible to separate
the fluxes by conversion category.
Fluxes are reported for four agricultural land use/land-use change categories: Cropland Remaining Cropland, Land
Converted to Cropland, Grassland Remaining Grassland, and Land Converted to Grassland. The reported
greenhouse gas fluxes from these agricultural lands include changes in soil organic C stocks in mineral and organic
soils due to land use and management, and for the subcategories of Forest Land Converted to Cropland and Forest
Land Converted to Grassland, the changes in aboveground biomass, belowground biomass, dead wood, and litter C
stocks are also reported. The greenhouse gas flux from Grassland Remaining Grassland also includes estimates of
non-C02 emissions from grassland fires occurring on both Grassland Remaining Grassland and Land Converted to
Grassland.
Fluxes from Wetlands Remaining Wetlands include changes in C stocks and methane (CH4) and nitrous oxide (N20)
emissions from managed peatlands, aboveground and belowground biomass, dead organic matter soil C stock
changes and CH4 emissions from coastal wetlands, as well as N20 emissions from aquaculture. Estimates for Land
Converted to Wetlands include aboveground and belowground biomass, dead organic matter and soil C stock
changes, and CH4 emissions from land converted to vegetated coastal wetlands.
Fluxes from Settlements Remaining Settlements include changes in C stocks from organic soils, N20 emissions from
nitrogen fertilizer additions to soils, and C02fluxes from settlement trees and landfilled yard trimmings and food
scraps. The reported greenhouse gas flux from Land Converted to Settlements includes changes in C stocks in
mineral and organic soils due to land use and management for all land use conversions to settlements, and the C
1 The term "flux" is used to describe the exchange of C02 to and from the atmosphere, with net flux being either positive or
negative depending on the overall balance. Removal and long-term storage of C02 from the atmosphere is also referred to as
"carbon sequestration."
Land Use, Land-Use Change, and Forestry 6-1

-------
stock changes in aboveground biomass, belowground biomass, dead wood, and litter are also included for the
subcategory Forest Land Converted to Settlements.
In 2019 the land use, land-use change, and forestry (LULUCF) sector resulted in a net increase in C stocks (i.e., net
C02 removals) of 812.7 MMT C02 Eq. (221.6 MMT C).2 This represents an offset of approximately 12.4 percent of
total (i.e., gross) greenhouse gas emissions in 2019. Emissions of CH4 and N20 from LULUCF activities in 2019 were
23.5 MMT C02 Eq. and represent 0.4 percent of total greenhouse gas emissions.3 In 2019 the overall net flux from
LULUCF resulted in a removal of 789.2 MMT C02 Eq. Emissions, removals and net greenhouse gas flux from LULUCF
are summarized in Figure 6-1 and Table 6-1 by land-use and category, and Table 6-2 and Table 6-3 by gas in MMT
C02 Eq. and kt, respectively. Trends in LULUCF sources and sinks over the 1990 to 2019 time series are shown in
Figure 6-2.
Figure 6-1: 2019 LULUCF Chapter Greenhouse Gas Sources and Sinks
Forest Land Remaining Forest Land
1

Settlements Remaining Settlements


Land Converted to Forest Land


Land Converted to Grassland
¦

Cropland Remaining Cropland
¦

Wetlands Remaining Wetlands
|

Land Converted to Wetlands

l< 0.51
Non-COz Emissions from Peatlands Remaining Peatlands

l< 0.5|
Non-C02 Emissions from Drained Organic Soils

l< 0.51
CH4 Emissions from Land Converted to Coastal Wetlands

l< 0.5|
N2O Emissions from Forest Soils

l< 0.51
Non-C02 Emissions from Grassland Fires


N2O Emissions from Settlement Soils


Non-C02 Emissions from Coastal Wetlands Remaining Coastal Wetlands


Grassland Remaining Grassland

1
Non-C02 Emissions from Forest Fires

¦
Land Converted to Cropland
Carbon Stock Change

Land Converted to Settlements
¦ Non-C02 Emissions
u
(250) (200) (150) (100) (50) 0 50 100
MMT CO2 Eq.
Note: Parentheses indicate net sequestration.
2	LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining Settlements,
and Land Converted to Settlements.
3	LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
6-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 6-2: Trends in Emissions and Removals (Net CO2 Flux) from Land Use, Land-Use
Change, and Forestry3
400
300
200
100
-100
-200
-300
-400
-500
-600
-700
-800
-900
-1,000
I Land Converted to Settlements
I Land Converted to Cropland
Grassland Remaining Grassland
Land Converted to Wetlands
Wetlands Remaining Wetlands
Cropland Remaining Cropland
Land Converted to Grassland
Land Converted to Forest Land
Settlements Remaining Settlements
Forest Land Remaining Forest Land
I Net Emission (Sources and Sinks)
Milium
ai 01 Ol CTi en 01	01 os en
CTioicjioicncncBoioicri
rMrMrsirMfMrMrMfMf>JfVJf>jrM(MfM(MfMfM(M(N(M
1 In Figure 6-2, the values above stacked bars represent only non-C02 LULUCF emissions. LULUCF emissions include the CH4 and
N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils, Grassland Fires, and Coastal
Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal Wetlands; and N20 emissions from
Forest Soils and Settlement Soils.
Table 6-1: Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and
Forestry (MMT CO2 Eq.)
Land-Use Category
Forest Land Remaining Forest Land
Changes in Forest Carbon Stocks3
Non-C02 Emissions from Forest Firesb
N20 Emissions from Forest Soilsc
Non-C02 Emissions from Drained Organic
Soilsd
Land Converted to Forest Land
Changes in Forest Carbon Stocks0
Cropland Remaining Cropland
Changes in Mineral and Organic Soil
Carbon Stocks
Land Converted to Cropland
Changes in all Ecosystem Carbon Stocks'
Grassland Remaining Grassland
Changes in Mineral and Organic Soil
Carbon Stocks
Non-C02 Emissions from Grassland Fires5
Land Converted to Grassland
Changes in all Ecosystem Carbon Stocks'
785.9
650.6
(787.6)
(671.4)
2016
2017
2018
2019
(715.7)
(721.9)
5.6
0.5
0.1
(99.0)
(99.0)
(22.7)
(22.7)
54.4
54.4
10.4
9.8
0.6
(24.0)
(24.0)
(640.9)
(659.7)
18.3
0.5
0.1
(99.1)
(99.1)
(22.3)
(22.3)
54.6
54.6
11.9
11.3
0.6
(24.4)
(24.4)
(682.4)
(698.6)
15.7
0.5
0.1
(99.1)
(99.1)
(16.6)
(16.6)
54.3
54.3
12.3
11.7
0.6
(24.1)
(24.1)
(675.5)
(691.8)
15.7
0.5
0.1
(99.1)
(99.1)
(14.5)
(14.5)
54.2
54.2
15.1
14.5
0.6
(23.2)
(23.2)
Land Use, Land-Use Change, and Forestry 6-3

-------
Wetlands Remaining Wetlands
(3.5)
(2.6)
(4.1)
(4.1)
(4.0)
(4.0)
(4.0)
Changes in Organic Soil Carbon Stocks in







Peatlands
1.1
1.1
0.8
0.7
0.8
0.8
0.8
Changes in Biomass, DOM, and Soil







Carbon Stocks in Coastal Wetlands
(8.5)
(7.6)
(8.8)
(8.8)
(8.8)
(8.8)
(8.8)
CH4 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
3.7
3.8
3.8
3.8
3.8
3.8
3.8
N20 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
0.1
0.2
0.1
0.1
0.1
0.1
0.1
Non-C02 Emissions from Peatlands







Remaining Peatlands
+
+
+
+
+
+
+
Land Converted to Wetlands
0.7
0.7
0.2
0.2
0.2
0.2
0.2
Changes in Biomass, DOM, and Soil







Carbon Stocks
0.4
0.4
(0.1)
(+)
(+)
(+)
(+)
CH4 Emissions from Land Converted to







Coastal Wetlands
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Settlements Remaining Settlements
(107.6)
(113.5)
(123.7)
(121.5)
(121.4)
(121.2)
(121.7)
Changes in Organic Soil Carbon Stocks
11.3
12.2
15.7
16.0
16.0
15.9
15.9
Changes in Settlement Tree Carbon







Stocks
(96.4)
(117.4)
(130.4)
(129.8)
(129.8)
(129.8)
(129.8)
Changes in Yard Trimming and Food







Scrap Carbon Stocks in Landfills
(24.5)
(11.4)
(11.1)
(10.0)
(9.8)
(9.8)
(10.2)
N20 Emissions from Settlement Soilsh
2.0
3.1
2.2
2.2
2.3
2.4
2.4
Land Converted to Settlements
62.9
85.0
80.1
79.4
79.3
79.3
79.2
Changes in all Ecosystem Carbon Stocks'
62.9
85.0
80.1
79.4
79.3
79.3
79.2
LULUCF Emissions'
7.9
16.8
27.8
13.2
26.0
23.4
23.5
LULUCF Carbon Stock Change1
(908.7)
(804.8)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
LULUCF Sector NetTotalk
(900.8)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
a Includes the net changes to carbon stocks stored in all forest ecosystem pools and harvested wood products.
b Estimates include CH4 and N20 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
c Estimates include N20 emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
d Estimates include CH4 and N20 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
6 Includes the net changes to carbon stocks stored in all forest ecosystem pools.
f Includes changes in mineral and organic soil carbon stocks for all land use conversions to cropland, grassland, and
settlements, respectively. Also includes aboveground/belowground biomass, dead wood, and litter carbon stock changes
for conversion of forest land to cropland, grassland, and settlements, respectively.
5 Estimates include CH4 and N20 emissions from fires on both Grassland Remaining Grassland and Land Converted to
Grass/and.
h Estimates include N20 emissions from N fertilizer additions on both Settlements Remaining Settlements and Land
Converted to Settlements because it is not possible to separate the activity data at this time.
' LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
1 LULUCF Carbon Stock Change includes any C stock gains and losses from all land use and land use conversion categories.
k The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes in units of MMT C02 Eq.
The C stock change and emissions of CH4 and N20 from LULUCF are summarized in Table 6-2 (MMT C02 Eq.) and
Table 6-3 (kt). Total C sequestration in the LULUCF sector decreased by approximately 10.6 percent between 1990
and 2019. This decrease was primarily due to a decline in the rate of net C accumulation in Forest Land and
6-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Cropland Remaining Cropland, as well as an increase in emissions from Land Converted to Settlements4
Specifically, there was a net C accumulation in Settlements Remaining Settlements, which increased from 1990 to
2019, while the net C accumulation in Forest Land Remaining Forest Land and Cropland Remaining Cropland
slowed over this period. Net C accumulation remained steady from 1990 to 2019 in Land Converted to Forest Land,
Land Converted to Cropland, Wetlands Remaining Wetlands, and Land Converted to Wetlands, while net C
accumulation fluctuated in Grassland Remaining Grassland. Net C accumulation from Land Converted to Grassland
increased from 1990 to 2019.
Forest fires were the largest source of CH4 emissions from LULUCF in 2019, totaling 9.5 MMT C02 Eq. (379 kt of
CH4). Coastal Wetlands Remaining Coastal Wetlands resulted in CH4 emissions of 3.8 MMT C02 Eq. (153 kt of CH4).
Grassland fires resulted in CH4 emissions of 0.3 MMT C02 Eq. (12 kt of CH4). Land Converted to Wetlands resulted
in CH4 emissions of 0.2 MMT C02 Eq. (7 kt of CH4). Drained Organic Soils on forest lands and Peatlands Remaining
Peatlands resulted in CH4 emissions of less than 0.05 MMT C02 Eq. each.
For N20 emissions, forest fires were also the largest source from LULUCF in 2019, totaling 6.2 MMT C02 Eq. (21 kt
of N20). Nitrous oxide emissions from fertilizer application to settlement soils in 2019 totaled to 2.4 MMT C02 Eq.
(8 kt of N20). This represents an increase of 20.2 percent since 1990. Additionally, the application of synthetic
fertilizers to forest soils in 2019 resulted in N20 emissions of 0.5 MMT C02 Eq. (2 kt of N20). Nitrous oxide
emissions from fertilizer application to forest soils have increased by 455.1 percent since 1990, but still account for
a relatively small portion of overall emissions. Grassland fires resulted in N20 emissions of 0.3 MMT C02 Eq. (1 kt of
N20). Coastal Wetlands Remaining Coastal Wetlands and drained organic soils on forest lands resulted in N20
emissions of 0.1 MMT C02 Eq. each (less than 0.5 kt of N20), and Peatlands Remaining Peatlands resulted in N20
emissions of less than 0.05 MMT C02 Eq.
Table 6-2: Emissions and Removals from Land Use, Land-Use Change, and Forestry by Gas
(MMT COz Eq.)
Gas/Land-Use Category
1990
2005
2015
2016
2017
2018
2019
Carbon Stock Change3
(908.7)
(804.8)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
Forest Land Remaining Forest Land
(787.6)
(661.5)
(671.4)
(721.9)
(659.7)
(698.6)
(691.8)
Land Converted to Forest Land
(98.2)
(98.7)
(98.9)
(99.0)
(99.1)
(99.1)
(99.1)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.8)
(22.7)
(22.3)
(16.6)
(14.5)
Land Converted to Cropland
51.8
52.2
56.1
54.4
54.6
54.3
54.2
Grassland Remaining Grassland
8.3
10.0
13.1
9.8
11.3
11.7
14.5
Land Converted to Grassland
(6.2)
(40.1)
(23.9)
(24.0)
(24.4)
(24.1)
(23.2)
Wetlands Remaining Wetlands
(7.4)
(6.5)
(8.0)
(8.0)
(8.0)
(8.0)
(8.0)
Land Converted to Wetlands
0.4
0.4
(0.1)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(109.6)
(116.6)
(125.9)
(123.8)
(123.7)
(123.6)
(124.1)
Land Converted to Settlements
62.9
85.0
80.1
79.4
79.3
79.3
79.2
ch4
5.0
9.3
16.6
7.7
15.3
13.8
13.8
Forest Land Remaining Forest Land:







Forest Firesb
0.9
5.0
12.2
3.4
11.0
9.5
9.5
Wetlands Remaining Wetlands: Coastal







Wetlands Remaining Coastal Wetlands
3.7
3.8
3.8
3.8
3.8
3.8
3.8
Grassland Remaining Grassland:







Grassland Firesc
0.1
0.3
0.3
0.3
0.3
0.3
0.3
Land Converted to Wetlands: Land







Converted to Coastal Wetlands
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Forest Land Remaining Forest Land:







Drained Organic Soilsd
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
4 Carbon sequestration estimates are net figures. The C stock in a given pool fluctuates due to both gains and losses. When
losses exceed gains, the C stock decreases, and the pool acts as a source. When gains exceed losses, the C stock increases, and
the pool acts as a sink; also referred to as net C sequestration or removal.
Land Use, Land-Use Change, and Forestry 6-5

-------
n2o
3.0
7.5
11.3
5.5
10.6
9.7
9.7
Forest Land Remaining Forest Land:







Forest Firesb
0.6
3.3
8.1
2.2
7.3
6.2
6.2
Settlements Remaining Settlements:







Settlement Soils0
2.0
3.1
2.2
2.2
2.3
2.4
2.4
Forest Land Remaining Forest Land:







Forest Soils'
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Grassland Remaining Grassland:







Grassland Firesc
0.1
0.3
0.3
0.3
0.3
0.3
0.3
Wetlands Remaining Wetlands: Coastal







Wetlands Remaining Coastal Wetlands
0.1
0.2
0.1.
0.1
0.1
0.1
0.1
Forest Land Remaining Forest Land:







Drained Organic Soilsd
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
LULUCF Carbon Stock Change3
(908.7)
(804.S)
(791.7)
(856.0)
(792.0)
(824.9)
(812.7)
LULUCF Emissions8
7.9
16.8
27.8
13.2
26.0
23.4
23.5
LULUCF Sector NetTotalh
(900.S)
(788.1)
(763.8)
(842.8)
(766.1)
(801.4)
(789.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
a LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining
Grassland, Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements
Remaining Settlements, and Land Converted to Settlements.
b Estimates include CH4 and N20 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
c Estimates include CH4 and N20 emissions from fires on both Grassland Remaining Grassland and Land Converted to
Grassland.
d Estimates include CH4 and N20 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
0 Estimates include N20 emissions from N fertilizer additions on both Settlements Remaining Settlements and Land
Converted to Settlements.
f Estimates include N20 emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
5 LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
h The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes in units of MMT C02 Eq.
Table 6-3: Emissions and Removals from Land Use, Land-Use Change, and Forestry by Gas
(kt)
Gas/Land-Use Category
1990
2005
2015
2016
2017
2018
2019
Carbon Stock Change (C02)a
(908,722)
(804,823)
(791,695)
(855,998)
(792,046)
(824,885)
(812,695)
Forest Land Remaining Forest







Land
(787,559)
(661,530)
(671,413)
(721,870)
(659,737)
(698,628)
(691,782)
Land Converted to Forest







Land
(98,170)
(98,673)
(98,943)
(99,005)
(99,075)
(99,079)
(99,080)
Cropland Remaining Cropland
(23,176)
(29,002)
(12,826)
(22,724)
(22,290)
(16,595)
(14,539)
Land Converted to Cropland
51,765
52,160
56,051
54,401
54,563
54,265
54,225
Grassland Remaining







Grassland
8,315
10,024
13,149
9,754
11,278
11,686
14,506
Land Converted to Grassland
(6,248)
(40,081)
(23,927)
(24,023)
(24,396)
(24,149)
(23,165)
Wetlands Remaining







Wetlands
(7,399)
(6,549)
(8,021)
(8,046)
(7,954)
(7,991)
(8,011)
Land Converted to Wetlands
449
439
(56)
(47)
(38)
(28)
(19)
Settlements Remaining
(109,567)
(116,642)
(125,854)
(123,790)
(123,707)
(123,638)
(124,062)
6-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Settlements
Land Converted to
Settlements
CH4
Forest Land Remaining Forest
Land: Forest Firesb
Wetlands Remaining
Wetlands: Coastal Wetlands
Remaining Coastal Wetlands
Grassland Remaining
Grassland: Grassland Firesc
Land Converted to Wetlands:
Land Converted to Coastal
Wetlands
Forest Land Remaining Forest
Land: Drained Organic Soilsd
Wetlands Remaining
Wetlands: Peatlands
Remaining Peatlands
n2o
Forest Land Remaining Forest
Land: Forest Firesb
Settlements Remaining
Settlements: Settlement
Soils6
Forest Land Remaining Forest
Land: Forest Soils'
Grassland Remaining
Grassland: Grassland Firesc
Wetlands Remaining
Wetlands: Coastal Wetlands
Remaining Coastal Wetlands
Forest Land Remaining Forest
Land: Drained Organic Soilsd
Wetlands Remaining
Wetlands: Peatlands
Remaining Peatlands	
62,867
198
35
149
3
10
1
+
10
85,032
372
198
151
13
10
1
+
25
11
10
2
1
80,145 79,350 79,310 79,271 79,233
663	308	614	552	552
489
152
13
135
153
11
440
153
12
+
38
27
+
18
+
36
24
379
153
12
7
1
+
32
21
379
153
12
7
1
+
32
21
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.5 kt.
a LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Converted to Grassland, Wetlands Remaining Wetlands, Land Converted to Wetlands, Settlements Remaining
Settlements, and Land Converted to Settlements.
b Estimates include CH4 and N20 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest
Land.
c Estimates include CH4 and N20 emissions from fires on both Grassland Remaining Grassland and Land Converted to Grassland.
d Estimates include CH4 and N20 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
0 Estimates include N20 emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted
to Settlements.
f Estimates include N20 emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
Each year, some emission and sink estimates in the LULUCF sector of the Inventory are recalculated and revised
with improved methods and/or data. In general, recalculations are made to the U.S. greenhouse gas emission
estimates either to incorporate new methodologies or, most commonly, to update recent historical data. These
improvements are implemented consistently across the previous Inventory's time series (i.e., 1990 to 2018) to
ensure that the trend is accurate. Of the updates implemented, the most significant include (1) Forest Lands: use
Land Use, Land-Use Change, and Forestry 6-7

-------
of new data from the National Forest Inventory (NFI), compiling population estimates of carbon stocks and stock
changes using NFI data from each U.S. state and summing over all states to obtain the national estimates, refined
estimates in the Digital General Soil Map, and new data on area burned from the Monitoring Trends in Burn
Severity (MTBS) data product; and (2) Coastal Wetlands: including belowground biomass carbon stock changes,
and use of new data from the Coastal Change Analysis Program (C-CAP), the USDA's Soil Survey Geographic
Database (SSURGO) and fisheries data from NOAA. Together, these updates for 2018 increased total sequestration
of C02 by 25.0 MMT C02 Eq. (3 percent) and decreased total non-C02 emissions by 2.7 MMT C02 Eq. (10 percent),
compared to the previous Inventory (i.e., 1990 to 2018). For more information on specific methodological updates,
please see the Recalculations discussion within the respective source category section of this chapter.
Emissions and removals reported in the LULUCF chapter include those from all states, however, for Hawaii and
Alaska some land use and land use changes are not included (see chapter sections on "Uncertainty and Time-Series
Consistency" and "Planned Improvements" for more details). In addition, U.S. Territories are not included. EPA
continues to review available data on an ongoing basis to include emissions from territories and DC in a future
inventories to the extent they are occurring (e.g., see Box 6-2). See Annex 5 for more information on EPA's
assessment of the emissions and removals not included in this Inventory.
Box 6-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals
In following the United Nations Framework Convention on Climate Change (UNFCCC) requirement under Article
4.1 to develop and submit national greenhouse gas emission inventories, the gross emissions total presented in
this report for the United States excludes emissions and removals from LULUCF. The LULUCF Sector Net Total
presented in this report for the United States includes emissions and removals from LULUCF. All emissions and
removals estimates are calculated using internationally-accepted methods provided by the IPCC in the 2006
IPCC Guidelines for National Greenhouse Gas Inventories (2006 IPCC Guidelines) and the 2013 Supplement.
Additionally, the calculated emissions and removals in a given year for the United States are presented in a
common manner in line with the UNFCCC reporting guidelines for the reporting of inventories under this
international agreement.5 The use of consistent methods to calculate emissions and removals by all nations
providing their inventories to the UNFCCC ensures that these reports are comparable. The presentation of
emissions and removals provided in the Land Use Land-Use Change and Forestry chapter does not preclude
alternative examinations, but rather, this Chapter presents emissions and removals in a common format
consistent with how countries are to report Inventories under the UNFCCC. The report itself, and this chapter,
follows this standardized format, and provides an explanation of the application of methods used to calculate
emissions and removals.
6.1 Representation of the U.S. Land Base
A national land-use representation system that is consistent and complete, both temporally and spatially, is
needed in order to assess land use and land-use change status and the associated greenhouse gas fluxes over the
Inventory time series. This system should be consistent with IPCC (2006), such that all countries reporting on
national greenhouse gas fluxes to the UNFCCC should: (1) describe the methods and definitions used to determine
areas of managed and unmanaged lands in the country (Table 6-4), (2) describe and apply a consistent set of
definitions for land-use categories over the entire national land base and time series (i.e., such that increases in
the land areas within particular land-use categories are balanced by decreases in the land areas of other categories
unless the national land base is changing) (Table 6-5), and (3) account for greenhouse gas fluxes on all managed
^ See .
6-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
lands. The IPCC (2006, Vol. IV, Chapter 1) considers all anthropogenic greenhouse gas emissions and removals
associated with land use and management to occur on managed land, and all emissions and removals on managed
land should be reported based on this guidance (See IPCC (2010), Ogle et al. (2018) for further discussion).
Consequently, managed land serves as a proxy for anthropogenic emissions and removals. This proxy is intended
to provide a practical framework for conducting an inventory, even though some of the greenhouse gas emissions
and removals on managed land are influenced by natural processes that may or may not be interacting with the
anthropogenic drivers. Guidelines for factoring out natural emissions and removals may be developed in the
future, but currently the managed land proxy is considered the most practical approach for conducting an
inventory in this sector (IPCC 2010). This section of the Inventory has been developed in order to comply with this
guidance.
Three databases are used to track land management in the United States and are used as the basis to classify
United States land area into the thirty-six IPCC land-use and land-use change categories (Table 6-5) (IPCC 2006).
The three primary databases are the U.S. Department of Agriculture (USDA) National Resources Inventory (NRI),6
the USDA Forest Service (USFS) Forest Inventory and Analysis (FIA)7 Database, and the Multi-Resolution Land
Characteristics Consortium (MRLC) National Land Cover Dataset (NLCD).8
The total land area included in the United States Inventory is 936 million hectares across the 50 states.9
Approximately 886 million hectares of this land base is considered managed and 50 million hectares is unmanaged,
which has not changed much over the time series of the Inventory (Table 6-5). In 2019, the United States had a
total of 282 million hectares of managed Forest Land (0.03 percent decrease compared to 1990). There are 162
million hectares of cropland (7.2 percent decrease compared to 1990), 337 million hectares of managed Grassland
(0.01 percent increase compared to 1990), 39 million hectares of managed Wetlands (1.8 percent increase
compared to 1990), 45 million hectares of Settlements (34 percent increase compared to 1990), and 22 million
hectares of managed Other Land (2.4 percent increase compared to 1990) (Table 6-5).
Wetlands are not differentiated between managed and unmanaged with the exception of remote areas in Alaska,
and so are reported mostly as managed.10 In addition, C stock changes are not currently estimated for the entire
managed land base, which leads to discrepancies between the managed land area data presented here and in the
subsequent sections of the Inventory (e.g., Grassland Remaining Grassland within interior Alaska).1112 There are
also discrepancies in the inventory emissions data and the land representation section because new FIA data were
used in the inventory analysis, but were not incorporated into the land representation analysis due to timing of
data availability and resources to complete the analysis. The land representation analysis will incorporate the new
time series of FIA data int the next Inventory. In addition, planned improvements are under development to
estimate C stock changes and greenhouse gas emissions on all managed land and ensure consistency between the
total area of managed land in the land-representation description and the remainder of the Inventory.
6	NRI data are available at .
7	FIA data are available at .
8	NLCD data are available at  and MRLC is a consortium of several U.S. government agencies.
9	The current land representation does not include areas from U.S. Territories, but there are planned improvements to include
these regions in future Inventories. U.S. Territories represent approximately 0.1 percent of the total land base for the United
States. See Box 6-2.
10	According to the IPCC (2006), wetlands are considered managed if they are created through human activity, such as dam
construction, or the water level is artificially altered by human activity. Distinguishing between managed and unmanaged
wetlands in the conterminous United States and Alaska is difficult due to limited data availability. Wetlands are not
characterized within the NRI with information regarding water table management. As a result, all Wetlands in the conterminous
United States and Hawaii are reported as managed. See the Planned Improvements section of the Inventory for future
refinements to the Wetland area estimates.
11	Other discrepancies occur because the coastal wetlands analysis is based on another land use product (NOAA C-CAP) that is
not currently incorporated into the land representation analysis for this section, which relies on the NRI and NLCD for wetland
areas. EPA anticipates addressing these discrepancies in the next Inventory.
12	These "managed area" discrepancies also occur in the Common Reporting Format (CRF) tables submitted to the UNFCCC.
Land Use, Land-Use Change, and Forestry 6-9

-------
Dominant land uses vary by region, largely due to climate patterns, soil types, geology, proximity to coastal
regions, and historical settlement patterns (Figure 6-3). Forest Land tends to be more common in the eastern
United States, mountainous regions of the western United States, and Alaska. Cropland is concentrated in the mid-
continent region of the United States, and Grassland is more common in the western United States and Alaska.
Wetlands are fairly ubiquitous throughout the United States, though they are more common in the upper Midwest
and eastern portions of the country, as well as coastal regions. Settlements are more concentrated along the
coastal margins and in the eastern states.
Table 6-4: Managed and Unmanaged Land Area by Land-Use Categories for All 50 States
(Thousands of Hectares)
Land Use Categories
1990
2005
2015
2016
2017
2018
2019a
Managed Lands
886,515
886,513
886,513
886,513
886,513
886,513
886,513
Forest
281,621
281,681
281,945
281,796
281,652
281,546
281,546
Croplands
174,471
165,727
161,929
161,933
161,933
161,933
161,933
Grasslands
336,840
337,621
336,529
336,657
336,781
336,863
336,863
Settlements
33,446
40,469
44,799
44,795
44,797
44,797
44,797
Wetlands
38,422
39,017
39,076
39,089
39,108
39,132
39,132
Other
21,715
21,997
22,236
22,243
22,243
22,243
22,243
Unmanaged Lands
49,681
49,684
49,683
49,683
49,683
49,683
49,683
Forest
9,243
8,829
8,208
8,208
8,208
8,208
8,208
Croplands
0
0
0
0
0
0
0
Grasslands
25,530
25,962
26,608
26,608
26,608
26,608
26,608
Settlements
0
0
0
0
0
0
0
Wetlands
4,166
4,166
4,165
4,165
4,165
4,165
4,165
Other
10,742
10,727
10,701
10,701
10,701
10,701
10,701
Total Land Areas
936,196
936,196
936,196
936,196
936,196
936,196
936,196
Forest
290,864
290,510
290,153
290,004
289,860
289,754
289,754
Croplands
174,471
165,727
161,929
161,933
161,933
161,933
161,933
Grasslands
362,370
363,583
363,138
363,266
363,389
363,471
363,471
Settlements
33,446
40,469
44,799
44,795
44,797
44,797
44,797
Wetlands
42,589
43,183
43,241
43,254
43,273
43,297
43,297
Other
32,457
32,725
32,937
32,944
32,944
32,944
32,944
a Land use data were not updated in this Inventory and the data for 2019 were assumed to be the same as in 2018. New land
use activity data will be incorporated and the time series will be updated in the next Inventory.
Table 6-5: Land Use and Land-Use Change for the U.S. Managed Land Base for All 50 States
(Thousands of Hectares)
Land-Use & Land-Use
Change Categories3
1990
2005
2015
2016
2017
2018
2019b
Total Forest Land
281,621
281,681
281,945
281,796
281,652
281,546
281,546
FF
280,393
280,207
280,528
280,529
280,380
280,274
280,274
CF
169
167
139
134
135
135
135
GF
919
1,162
1,125
989
992
992
992
WF
77
28
25
25
25
25
25
SF
12
24
27
26
26
26
26
OF
50
93
100
93
93
93
93
Total Cropland
174,471
165,727
161,929
161,933
161,933
161,933
161,933
CC
162,163
150,304
148,880
148,885
148,884
148,884
148,884
FC
182
86
58
58
58
58
58
GC
11,738
14,820
12,609
12,609
12,609
12,609
12,609
WC
118
178
104
104
104
104
104
SC
75
100
99
99
99
99
99
OC
195
239
179
179
179
179
179
Total Grassland
336,840
337,621
336,529
336,657
336,781
336,863
336,863
6-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
GG
327,446
315,161
316,287
316,408
316,502
316,622
316,622
FG
593
560
547
553
583
545
545
CG
8,237
17,523
16,600
16,600
16,600
16,600
16,600
WG
176
542
308
308
308
308
308
SG
43
509
346
346
346
346
346
OG
345
3,328
2,442
2,442
2,442
2,442
2,442
Total Wetlands
38,422
39,017
39,076
39,089
39,108
39,132
39,132
WW
37,860
37,035
37,602
37,616
37,634
37,658
37,658
FW
83
59
54
54
54
54
54
CW
132
566
440
440
440
440
440
GW
297
1,187
836
836
836
836
836
SW
0
38
25
25
25
25
25
OW
50
133
118
118
118
118
118
Total Settlements
33,446
40,469
44,799
44,795
44,797
44,797
44,797
SS
30,585
31,522
38,210
38,210
38,210
38,210
38,210
FS
310
549
544
539
541
541
541
CS
1,237
3,602
2,452
2,452
2,452
2,452
2,452
GS
1,255
4,499
3,352
3,352
3,352
3,352
3,352
WS
4
61
46
46
46
46
46
OS
54
235
197
197
197
197
197
Total Other Land
21,715
21,997
22,236
22,243
22,243
22,243
22,243
00
20,953
18,231
19,000
19,007
19,007
19,007
19,007
FO
41
70
90
90
90
90
90
CO
301
590
678
678
678
678
678
GO
391
2,965
2,331
2,331
2,331
2,331
2,331
WO
26
121
121
121
121
121
121
SO
2
20
16
16
16
16
16
Grand Total
886,515
886,513
886,513
886,513
886,513
886,513
886,513
Notes: All land areas reported in this table are considered managed. A planned improvement is underway to deal
with an exception for Wetlands, which based on the definitions for the current U.S. Land Representation
assessment includes both managed and unmanaged lands. U.S. Territories have not been classified into land
uses and are not included in the U.S. Land Representation Assessment. See the Planned Improvements section
for discussion on plans to include territories in future Inventories. In addition, C stock changes are not currently
estimated for the entire land base, which leads to discrepancies between the managed land area data
presented here and in the subsequent sections of the Inventory (see land use chapters e.g., Forest Land
Remaining Forest Land for more information). Totals may not sum due to independent rounding.
a The abbreviations are "F" for Forest Land, "C" for Cropland, "G" for Grassland, "W" for Wetlands, "S" for
Settlements, and "0" for Other Lands. Lands remaining in the same land-use category are identified with the
land-use abbreviation given twice (e.g., "FF" is Forest Land Remaining Forest Land), and land-use change
categories are identified with the previous land use abbreviation followed by the new land-use abbreviation
(e.g., "CF" is Cropland Converted to Forest Land).
b Land use data were not updated in this Inventory and the data for 2019 were assumed to be the same as in
2018. New land use activity data will be incorporated and the time series will be updated in the next Inventory.
Land Use, Land-Use Change, and Forestry 6-11

-------
Figure 6-3: Percent of Total Land Area for Each State in the General Land-Use Categories for
2019
Forest Lands	Croplands
Hawaii
Grasslands
Wetlands
11 - 30
31 - 50
-i Cvft
~ 11 - 30
31 - 50
Alaska
Hawaii
Alaska
Settlements
Other Lands
6-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
IPCC Approaches for Representing Land Areas
IPCC (2006) describes three approaches for representing land areas. Approach 1 provides data on the total area for
each individual land-use category, but does not provide detailed information on changes of area between
categories and is not spatially explicit other than at the national or regional level. With Approach 1, total net
conversions between categories can be detected, but not the individual changes (i.e., additions and/or losses)
between the land-use categories that led to those net changes. Approach 2 introduces tracking of individual land-
use changes between the categories (e.g., Forest Land to Cropland, Cropland to Forest Land, and Grassland to
Cropland), using survey samples or other forms of data, but does not provide spatially-explicit location data.
Approach 3 extends Approach 2 by providing spatially-explicit location data, such as surveys with spatially
identified sample locations and maps derived from remote sensing products. The three approaches are not
presented as hierarchical tiers and are not mutually exclusive.
According to IPCC (2006), the approach or mix of approaches selected by an inventory agency should reflect
calculation needs and national circumstances. For this analysis, the NRI, FIA, and the NLCD have been combined to
provide a complete representation of land use for managed lands. These data sources are described in more detail
later in this section. NRI, FIA and NLCD are Approach 3 data sources that provide spatially-explicit representations
of land use and land-use conversions. Lands are treated as remaining in the same category (e.g., Cropland
Remaining Cropland) if a land-use change has not occurred in the last 20 years. Otherwise, the land is classified in a
land-use change category based on the current use and most recent use before conversion to the current use (e.g.,
Cropland Converted to Forest Land).
Definitions of Land Use in the United States
Managed and Unmanaged Land
The United States definition of managed land is similar to the general definition of managed land provided by the
IPCC (2006), but with some additional elaboration to reflect national circumstances. Based on the following
definitions, most lands in the United States are classified as managed:
•	Managed Land: Land is considered managed if direct human intervention has influenced its condition.
Direct intervention occurs mostly in areas accessible to human activity and includes altering or
maintaining the condition of the land to produce commercial or non-commercial products or services; to
serve as transportation corridors or locations for buildings, landfills, or other developed areas for
commercial or non-commercial purposes; to extract resources or facilitate acquisition of resources; or to
provide social functions for personal, community, or societal objectives where these areas are readily
accessible to society.13
•	Unmanaged Land: All other land is considered unmanaged. Unmanaged land is largely comprised of areas
inaccessible to society due to the remoteness of the locations. Though these lands may be influenced
13 Wetlands are an exception to this general definition, because these lands, as specified by IPCC (2006), are only considered
managed if they are created through human activity, such as dam construction, or the water level is artificially altered by
human activity. Distinguishing between managed and unmanaged wetlands in the United States is difficult due to limited data
availability. Wetlands are not characterized within the NRI with information regarding water table management or origin (i.e.,
constructed rather than natural origin). Therefore, unless wetlands are converted into cropland or grassland, it is not possible
to know if they are artificially created or if the water table is managed based on the use of NRI data. As a result, most wetlands
are reported as managed with the exception of wetlands in remote areas of Alaska, but emissions from managed wetlands are
only reported for coastal regions and peatlands due to insufficient activity data to estimate emissions and limited resources to
improve the inventory. See the Planned Improvements section of the Inventory for future refinements to the wetland area
estimates.
Land Use, Land-Use Change, and Forestry 6-13

-------
indirectly by human actions such as atmospheric deposition of chemical species produced in industry or
C02 fertilization, they are not influenced by a direct human intervention.14
In addition, land that is previously managed remains in the managed land base for 20 years before re-classifying
the land as unmanaged in order to account for legacy effects of management on C stocks. Unmanaged land is also
re-classified as managed over time if anthropogenic activity is introduced into the area based on the definition of
managed land.
Land-Use Categories
As with the definition of managed lands, IPCC (2006) provides general non-prescriptive definitions for the six main
land-use categories: Forest Land, Cropland, Grassland, Wetlands, Settlements and Other Land. In order to reflect
national circumstances, country-specific definitions have been developed, based predominantly on criteria used in
the land-use surveys for the United States. Specifically, the definition of Forest Land is based on the FIA definition
of forest,15 while definitions of Cropland, Grassland, and Settlements are based on the NRI.16The definitions for
Other Land and Wetlands are based on the IPCC (2006) definitions for these categories.
•	Forest Land: A land-use category that includes areas at least 120 feet (36.6 meters) wide and at least one
acre (0.4 hectare) in size with at least 10 percent cover (or equivalent stocking) by live trees including land
that formerly had such tree cover and that will be naturally or artificially regenerated. Trees are woody
plants having a more or less erect perennial stem(s) capable of achieving at least 3 inches (7.6 cm) in
diameter at breast height, or 5 inches (12.7 cm) diameter at root collar, and a height of 16.4 feet (5 m) at
maturity in situ. Forest Land includes all areas recently having such conditions and currently regenerating
or capable of attaining such condition in the near future. Forest Land also includes transition zones, such
as areas between forest and non-forest lands that have at least 10 percent cover (or equivalent stocking)
with live trees and forest areas adjacent to urban and built-up lands. Unimproved roads and trails,
streams, and clearings in forest areas are classified as forest if they are less than 120 feet (36.6 m) wide or
an acre (0.4 ha) in size. However, land is not classified as Forest Land if completely surrounded by urban
or developed lands, even if the criteria are consistent with the tree area and cover requirements for
Forest Land. These areas are classified as Settlements. In addition, Forest Land does not include land that
is predominantly under an agricultural land use (Oswalt et al. 2014).
•	Cropland: A land-use category that includes areas used for the production of adapted crops for harvest;
this category includes both cultivated and non-cultivated lands. Cultivated crops include row crops or
close-grown crops and also pasture in rotation with cultivated crops. Non-cultivated cropland includes
continuous hay, perennial crops (e.g., orchards) and horticultural cropland. Cropland also includes land
with agroforestry, such as alley cropping and windbreaks,17 if the dominant use is crop production,
assuming the stand or woodlot does not meet the criteria for Forest Land. Lands in temporary fallow or
enrolled in conservation reserve programs (i.e., set-asides18) are also classified as Cropland, as long as
these areas do not meet the Forest Land criteria. Roads through Cropland, including interstate highways,
state highways, other paved roads, gravel roads, dirt roads, and railroads are excluded from Cropland
area estimates and are, instead, classified as Settlements.
•	Grassland: A land-use category on which the plant cover is composed principally of grasses, grass-like
plants (i.e., sedges and rushes), forbs, or shrubs suitable for grazing and browsing, and includes both
14	There are some areas, such as Forest Land and Grassland in Alaska that are classified as unmanaged land due to the
remoteness of their location.
15	See , page 22.
16	See .
17	Currently, there is no data source to account for biomass C stock change associated with woody plant growth and losses in
alley cropping systems and windbreaks in cropping systems, although these areas are included in the Cropland land base.
18	A set-aside is cropland that has been taken out of active cropping and converted to some type of vegetative cover, including,
for example, native grasses or trees, but is still classified as cropland based on national circumstances.
6-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
pastures and native rangelands. This includes areas where practices such as clearing, burning, chaining,
and/or chemicals are applied to maintain the grass vegetation. Land is also categorized as Grassland if
there have been three or fewer years of continuous hay production.19 Savannas, deserts, and tundra are
considered Grassland.20 Drained wetlands are considered Grassland if the dominant vegetation meets the
plant cover criteria for Grassland. Woody plant communities of low forbs, shrubs and woodlands, such as
sagebrush, mesquite, chaparral, mountain shrubland, and pinyon-juniper, are also classified as Grassland
if they do not meet the criteria for Forest Land. Grassland includes land managed with agroforestry
practices, such as silvopasture and windbreaks, if the land is principally grass, grass-like plants, forbs, and
shrubs suitable for grazing and browsing, and assuming the stand or woodlot does not meet the criteria
for Forest Land. Roads through Grassland, including interstate highways, state highways, other paved
roads, gravel roads, dirt roads, and railroads are excluded from Grassland and are, instead, classified as
Settlements.
•	Wetlands: A land-use category that includes land covered or saturated by water for all or part of the year,
in addition to lakes, reservoirs, and rivers. Managed Wetlands are those where the water level is
artificially changed, or were created by human activity. Certain areas that fall under the managed
Wetlands definition are included in other land uses based on the IPCC guidance and national
circumstances, including lands that are flooded for most or just part of the year in Croplands (e.g., rice
cultivation and cranberry production), Grasslands (e.g., wet meadows dominated by grass cover) and
Forest Lands (e.g., Riparian Forests near waterways).
•	Settlements: A land-use category representing developed areas consisting of units equal to or greater
than 0.25 acres (0.1 ha) that includes residential, industrial, commercial, and institutional land;
construction sites; public administrative sites; railroad yards; cemeteries; airports; golf courses; sanitary
landfills; sewage treatment plants; water control structures and spillways; parks within urban and built-up
areas; and highways, railroads, and other transportation facilities. Also included are all tracts that may
meet the definition of Forest Land, and tracts of less than 10 acres (4.05 ha) that may meet the definitions
for Cropland, Grassland, or Other Land but are completely surrounded by urban or built-up land, and so
are included in the Settlements category. Rural transportation corridors located within other land uses
(e.g., Forest Land, Cropland, and Grassland) are also included in Settlements.
•	Other Land: A land-use category that includes bare soil, rock, ice, and all land areas that do not fall into
any of the other five land-use categories. Following the guidance provided by the IPCC (2006), C stock
changes and non-C02 emissions are not estimated for Other Lands because these areas are largely devoid
of biomass, litter and soil C pools. However, C stock changes and non-C02 emissions are estimated for
Land Converted to Other Land during the first 20 years following conversion to account for legacy effects.
Land-Use Data Sources, r»?h.Tiption and Application to U.S.
Land Area Classification
U.S. Land-Use Data Sources
The three main sources for land-use data in the United States are the NRI, FIA, and the NLCD (Table 6-6). These
data sources are combined to account for land use in all 50 states. FIA and NRI data are used when available for an
area because these surveys contain additional information on management, site conditions, crop types, biometric
measurements, and other data that are needed to estimate C stock changes, N20, and CH4 emissions on those
19	Areas with four or more years of continuous hay production are Cropland because the land is typically more intensively
managed with cultivation, greater amounts of inputs, and other practices. Occasional harvest of hay from grasslands typically
does not involve cultivation or other intensive management practices.
20	2006 IPCC Guidelines do not include provisions to separate desert and tundra as land-use categories.
Land Use, Land-Use Change, and Forestry 6-15

-------
lands. If NRI and FIA data are not available for an area, however, then the NLCD product is used to represent the
land use.
Table 6-6: Data Sources Used to Determine Land Use and Land Area for the Conterminous
		 ——¦'
NRI
FIA
NLCD
Forest Land
Conterminous
United States
Non-Federal
Federal
Hawaii
Non-Federal
Federal
Alaska
Non-Federal
Federal
Croplands, Grasslands, Other Lands, Settlements, and Wetlands
Conterminous
United States
Non-Federal	•
Federal
Hawaii
Non-Federal	•
Federal
Alaska
Non-Federal
Federal
National Resources Inventory
For the Inventory, the NRI is the official source of data for land use and land use change on non-federal lands in the
conterminous United States and Hawaii, and is also used to determine the total land base for the conterminous
United States and Hawaii. The NRI is a statistically-based survey conducted by the USDA Natural Resources
Conservation Service and is designed to assess soil, water, and related environmental resources on non-federal
lands. The NRI has a stratified multi-stage sampling design, where primary sample units are stratified on the basis
of county and township boundaries defined by the United States Public Land Survey (Nusser and Goebel 1997).
Within a primary sample unit (typically a 160 acre [64.75 ha] square quarter-section), three sample points are
selected according to a restricted randomization procedure. Each point in the survey is assigned an area weight
(expansion factor) based on other known areas and land-use information (Nusser and Goebel 1997). The NRI
survey utilizes data derived from remote sensing imagery and site visits in order to provide detailed information on
land use and management, particularly for Croplands and Grasslands (i.e., agricultural lands), and is used as the
basis to account for C stock changes in agricultural lands (except federal Grasslands). The NRI survey was
conducted every 5 years between 1982 and 1997, but shifted to annualized data collection in 1998. The land use
between five-year periods from 1982 and 1997 are assumed to be the same for a five-year time period if the land
use is the same at the beginning and end of the five-year period (Note: most of the data has the same land use at
the beginning and end of the five-year periods). If the land use had changed during a five-year period, then the
change is assigned at random to one of the five years. For crop histories, years with missing data are estimated
based on the sequence of crops grown during years preceding and succeeding a missing year in the NRI history.
This gap-filling approach allows for development of a full time series of land-use data for non-federal lands in the
conterminous United States and Hawaii. This Inventory incorporates data through 2015 from the NRI. The land use
patterns are assumed to remain the same from 2016 through 2019 for this Inventory, but the time series will be
updated when new data are released.
6-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Forest Inventory and Analysis
The FIA program, conducted by the USFS, is the official source of data on Forest Land area and management data
for the Inventory and is another statistically-based survey for the conterminous United States in addition to the
including southeast and south-central coastal Alaska. FIA engages in a hierarchical system of sampling, with
sampling categorized as Phases 1 through 3, in which sample points for phases are subsets of the previous phase.
Phase 1 refers to collection of remotely-sensed data (either aerial photographs or satellite imagery) primarily to
classify land into forest or non-forest and to identify landscape patterns like fragmentation and urbanization.
Phase 2 is the collection of field data on a network of ground plots that enable classification and summarization of
area, tree, and other attributes associated with forest-land uses. Phase 3 plots are a subset of Phase 2 plots where
data on indicators of forest health are measured. Data from all three phases are also used to estimate C stock
changes for Forest Land. Historically, FIA inventory surveys have been conducted periodically, with all plots in a
state being measured at a frequency of every five to 14 years. A new national plot design and annual sampling
design was introduced by the FIA program in 1998 and is now used in all states. Annualized sampling means that a
portion of plots throughout each state is sampled each year, with the goal of measuring all plots once every five to
seven years in the eastern United States and once every ten years in the western United States. See Annex 3.13 to
see the specific survey data available by state. The most recent year of available data varies state by state (range of
most recent data is from 2015 through 2018; see Table A-203 in Annex 3.13).
National Land Cover Dataset
As noted above, while the NRI survey sample covers the conterminous United States and Hawaii, land use data are
only collected on non-federal lands. In addition, FIA only records data for forest land across the land base in the
conterminous United States and Alaska.21 Consequently, gaps exist in the land representation when the datasets
are combined, such as federal grassland operated by Bureau of Land Management (BLM), USDA, and National Park
Service, as well as Alaska.22 The NLCD is used to account for land use on federal lands in the conterminous United
States and Hawaii, in addition to federal and non-federal lands in Alaska with the exception of Forest Lands in
Alaska.
NLCD products provide land-cover for 1992, 2001, 2004, 2006, 2008, 2011, 2013, and 2016 in the conterminous
United States (Yang et al. 2018; Fry et al. 2011; Homer et al. 2007, 2015), and also for Alaska in 2001 and 2011 and
Hawaii in 2001. A Land Cover Change Product is also available for Alaska from 2001 to 2011. A NLCD change
product is not available for Hawaii because data are only available for one year, i.e., 2001. The NLCD products are
based primarily on Landsat Thematic Mapper imagery at a 30-meter resolution, and the land cover categories have
been aggregated into the 36 IPCC land-use categories for the conterminous United States and Alaska, and into the
six IPCC land-use categories for Hawaii. The land use patterns are assumed to remain the same after the last year
of data in the time series, which is 2001 for Hawaii, 2016 for the conterminous United States and 2011 for Alaska,
but the time series will be updated when new data are released.
For the conterminous United States, the aggregated maps of IPCC land-use categories derived from the NLCD
products were used in combination with the NRI database to represent land use and land-use change for federal
lands, with the exception of forest lands, which are based on FIA. Specifically, NRI survey locations designated as
federal lands were assigned a land use/land-use change category based on the NLCD maps that had been
aggregated into the IPCC categories. This analysis addressed shifts in land ownership across years between federal
or non-federal classes as represented in the NRI survey (i.e., the ownership is classified for each survey location in
the NRI). The sources of these additional data are discussed in subsequent sections of the report.
21	FIA does collect some data on non-forest land use, but these are held in regional databases versus the national database.
The status of these data is being investigated.
22	The NRI survey program does not include U.S. Territories with the exception of non-federal lands in Puerto Rico. The FIA
program recently began implementing surveys of forest land in U.S. Territories and those data will be used in the years ahead.
Furthermore, NLCD does not include coverage for all U.S. Territories.
Land Use, Land-Use Change, and Forestry 6-17

-------
Managed Land Designation
Lands are designated as managed in the United States based on the definition provided earlier in this section. The
following criteria are used in order to apply the definition in an analysis of managed land:
•	All Croplands and Settlements are designated as managed so only Grassland, Forest Land, Wetlands or
Other Lands may be designated as unmanaged land;23
•	All Forest Lands with active fire protection are considered managed;
•	All Forest Lands designated for timber harvests are considered managed;
•	All Grasslands are considered managed at a county scale if there are grazing livestock in the county;
•	Other areas are considered managed if accessible based on the proximity to roads and other
transportation corridors, and/or infrastructure;
•	Protected lands maintained for recreational and conservation purposes are considered managed (i.e.,
managed by public and/or private organizations);
•	Lands with active and/or past resource extraction are considered managed; and
•	Lands that were previously managed but subsequently classified as unmanaged, remain in the managed
land base for 20 years following the conversion to account for legacy effects of management on C stocks.
The analysis of managed lands, based on the criteria listed above, is conducted using a geographic information
system (Ogle et al. 2018). Lands that are used for crop production or settlements are determined from the NLCD
(Fry et al. 2011; Homer et al. 2007; Homer et al. 2015). Forest Lands with active fire management are determined
from maps of federal and state management plans from the National Atlas (U.S. Department of Interior 2005) and
Alaska Interagency Fire Management Council (1998). It is noteworthy that all forest lands in the conterminous
United States have active fire protection, and are therefore designated as managed regardless of accessibility or
other criteria. In addition, forest lands with timber harvests are designated as managed based on county-level
estimates of timber products in the U.S. Forest Service Timber Products Output Reports (U.S. Department of
Agriculture 2012). Timber harvest data do lead to additional designation of managed forest land in Alaska. The
designation of grasslands as managed is based on grazing livestock population data at the county scale from the
USDA National Agricultural Statistics Service (U.S. Department of Agriculture 2015). Accessibility is evaluated based
on a 10-km buffer surrounding road and train transportation networks using the ESRI Data and Maps product (ESRI
2008), and a 10-km buffer surrounding settlements using NLCD.
Lands maintained for recreational purposes are determined from analysis of the Protected Areas Database (U.S.
Geological Survey 2012). The Protected Areas Database includes lands protected from conversion of natural
habitats to anthropogenic uses and describes the protection status of these lands. Lands are considered managed
that are protected from development if the regulations allow for extractive or recreational uses or suppression of
natural disturbance. Lands that are protected from development and not accessible to human intervention,
including no suppression of disturbances or extraction of resources, are not included in the managed land base.
Multiple data sources are used to determine lands with active resource extraction: Alaska Oil and Gas Information
System (Alaska Oil and Gas Conservation Commission 2009), Alaska Resource Data File (U.S. Geological Survey
2012), Active Mines and Mineral Processing Plants (U.S. Geological Survey 2005), and Coal Production and
Preparation Report (U.S. Energy Information Administration 2011). A buffer of 3,300 and 4,000 meters is
established around petroleum extraction and mine locations, respectively, to account for the footprint of
operation and impacts of activities on the surrounding landscape. The buffer size is based on visual analysis of
disturbance to the landscape for approximately 130 petroleum extraction sites and 223 mines. After applying the
criteria identified above, the resulting managed land area is overlaid on the NLCD to estimate the area of managed
land by land use for both federal and non-federal lands in Alaska. The remaining land represents the unmanaged
23 All wetlands are considered managed in this Inventory with the exception of remote areas in Alaska. Distinguishing between
managed and unmanaged wetlands in the conterminous United States and Hawaii is difficult due to limited data availability.
Wetlands are not characterized within the NRI with information regarding water table management. Regardless, a planned
improvement is underway to subdivide managed and unmanaged wetlands.
6-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
land base. The resulting spatial product is also used to identify NRI survey locations that are considered managed
and unmanaged for the conterminous United States and Hawaii.24
Approach for Combining Data Sources
The managed land base in the United States has been classified into the 36 IPCC land-use/land-use conversion
categories (Table 6-5) using definitions developed to meet national circumstances, while adhering to IPCC
guidelines (20 06).25 In practice, the land was initially classified into land-use subcategories within the NRI, FIA, and
NLCD datasets, and then aggregated into the 36 broad land use and land-use change categories identified in IPCC
(2006).
All three datasets provide information on forest land areas in the conterminous United States, but the area data
from FIA serve as the official dataset for Forest Land. Therefore, another step in the analysis is to address the
inconsistencies in the representation of the Forest Land among the three databases. NRI and FIA have different
criteria for classifying Forest Land in addition to different sampling designs, leading to discrepancies in the resulting
estimates of Forest Land area on non-federal land in the conterminous United States. Similarly, there are
discrepancies between the NLCD and FIA data for defining and classifying Forest Land on federal lands. Any change
in Forest Land Area in the NRI and NLCD also requires a corresponding change in other land use areas because of
the dependence between the Forest Land area and the amount of land designated as other land uses, such as the
amount of Grassland, Cropland, and Wetlands (i.e., areas for the individual land uses must sum to the total
managed land area of the country).
FIA is the main database for forest statistics, and consequently, the NRI and NLCD are adjusted to achieve
consistency with FIA estimates of Forest Land in the conterminous United States. Adjustments are made in the
Forest Land Remaining Forest Land, Land Converted to Forest Land, and Forest Land converted to other uses (i.e.,
Grassland, Cropland, Settlements, Other Lands, and Wetlands). All adjustments are made at the state scale to
address the discrepancies in areas associated with Forest Land and conversions to and from Forest Land. There are
three steps in this process. The first step involves adjustments to Land Converted to Forest Land (Grassland,
Cropland, Settlements, Other Lands, and Wetlands), followed by a second step in which there are adjustments in
Forest Land converted to another land use (i.e., Grassland, Cropland, Settlements, Other Lands, and Wetlands),
and finally the last step is to adjust Forest Land Remaining Forest Land.
In the first step, Land Converted to Forest Land in the NRI and NLCD are adjusted to match the state-level
estimates in the FIA data for non-federal and federal Land Converted to Forest Land, respectively. FIA data have
not provided specific land-use categories that are converted to Forest Land in the past, but rather a sum of all Land
Converted to Forest Land.26 The NRI and NLCD provide information on specific land use conversions, such as
Grassland Converted to Forest Land. Therefore, adjustments at the state level to NRI and NLCD are made
proportional to the amount of specific land use conversions into Forest Land for the state, prior to any
adjustments. For example, if 50 percent of the land use change to Forest Land is associated with Grassland
Converted to Forest Land in a state according to NRI or NLCD, then half of the discrepancy with FIA data in the area
of Land Converted to Forest Land is addressed by increasing or decreasing the area in Grassland Converted to
Forest Land. Moreover, any increase or decrease in Grassland Converted to Forest Land in NRI or NLCD is
addressed by a corresponding change in the area of Grassland Remaining Grassland, so that the total amount of
managed area is not changed within an individual state.
In the second step, state-level areas are adjusted in the NRI and NLCD to address discrepancies with FIA data for
Forest Land converted to other uses. Similar to Land Converted to Forest Land, FIA have not provided information
24	The exception is cropland and settlement areas in the NRI, which are classified as managed, regardless of the managed land
base derived from the spatial analysis described in this section.
25	Definitions are provided in the previous section.
26	The FIA program has started to collect data on the specific land uses that are converted to Forest Land, which will be further
investigated and incorporated into a future Inventory.
Land Use, Land-Use Change, and Forestry 6-19

-------
on the specific land-use changes in the past,27 and so areas associated with Forest Land conversion to other land
uses in NRI and NLCD are adjusted proportional to the amount of area in each conversion class in these datasets.
In the final step, the area of Forest Land Remaining Forest Land in a given state according to the NRI and NLCD is
adjusted to match the FIA estimates for non-federal and federal land, respectively. It is assumed that the majority
of the discrepancy in Forest Land Remaining Forest Land is associated with an under- or over-prediction of
Grassland Remaining Grassland and Wetland Remaining Wetland in the NRI and NLCD. This step also assumes that
there are no changes in the land use conversion categories. Therefore, corresponding increases or decreases are
made in the area estimates of Grasslands Remaining Grasslands and Wetlands Remaining Wetlands from the NRI
and NLCD. This adjustment balances the change in Forest Land Remaining Forest Land area, which ensures no
change in the overall amount of managed land within an individual state. The adjustments are based on the
proportion of land within each of these land-use categories at the state level according to NRI and NLCD (i.e., a
higher proportion of Grassland led to a larger adjustment in Grassland area).
The modified NRI data are then aggregated to provide the land-use and land-use change data for non-federal lands
in the conterminous United States, and the modified NLCD data are aggregated to provide the land use and land-
use change data for federal lands. Data for all land uses in Hawaii are based on NRI for non-federal lands and on
NLCD for federal lands. Land use data in Alaska are based on the NLCD data after adjusting this dataset to be
consistent with forest land areas in the FIA (Table 6-6). The result is land use and land-use change data for the
conterminous United States, Hawaii, and Alaska.
A summary of the details on the approach used to combine data sources for each land use are described below.
•	Forest Land: Land representation for both non-federal and federal forest lands in the conterminous
United States and Alaska are based on the FIA. FIA is used as the basis for both Forest Land area data as
well as to estimate C stocks and fluxes on Forest Land in the conterminous United States and Alaska. FIA
does have survey plots in Alaska that are used to determine the C stock changes, and the associated area
data for this region are harmonized with the NLCD using the methods described above. NRI is used in the
current report to provide Forest Land areas on non-federal lands in Hawaii, and NLCD is used for federal
lands. FIA data is being collected in Hawaii and U.S. Territories, however there is insufficient data to make
population estimates for this Inventory.
•	Cropland: Cropland is classified using the NRI, which covers all non-federal lands within 49 states
(excluding Alaska), including state and local government-owned land as well as tribal lands. NRI is used as
the basis for both Cropland area data as well as to estimate soil C stocks and fluxes on Cropland. NLCD is
used to determine Cropland area and soil C stock changes on federal lands in the conterminous United
States and Hawaii. NLCD is also used to determine croplands in Alaska, but C stock changes are not
estimated for this region in the current Inventory.
•	Grassland: Grassland on non-federal lands is classified using the NRI within 49 states (excluding Alaska),
including state and local government-owned land as well as tribal lands. NRI is used as the basis for both
Grassland area data as well as to estimate soil C stocks and non-C02 greenhouse emissions on Grassland.
Grassland area and soil C stock changes are determined using the classification provided in the NLCD for
federal land within the conterminous United States. NLCD is also used to estimate the areas of federal and
non-federal grasslands in Alaska, and the federal grasslands in Hawaii, but the current Inventory does not
include C stock changes in these areas.
•	Wetlands: NRI captures wetlands on non-federal lands within 49 states (excluding Alaska), while the land
representation data for federal wetlands and wetlands in Alaska are based on the NLCD.28
27	The FIA program has started to collect data on specific land uses following conversion from Forest Land, which will be further
investigated and incorporated into a future Inventory.
28	This analysis does not distinguish between managed and unmanaged wetlands except for remote areas in Alaska, but there
is a planned improvement to subdivide managed and unmanaged wetlands for the entire land base.
6-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
•	Settlements: NRI captures non-federal settlement area in 49 states (excluding Alaska). If areas of Forest
Land or Grassland under 10 acres (4.05 ha) are contained within settlements or urban areas, they are
classified as Settlements (urban) in the NRI database. If these parcels exceed the 10 acres (4.05 ha)
threshold and are Grassland, they are classified as Grassland by NRI. Regardless of size, a forested area is
classified as non-forest by FIA if it is located within an urban area. Land representation for settlements on
federal lands and Alaska is based on the NLCD.
•	Other Land: Any land that is not classified into one of the previous five land-use categories, is categorized
as Other Land using the NRI for non-federal areas in the conterminous United States and Hawaii and using
the NLCD for the federal lands in all regions of the United States and for non-federal lands in Alaska.
Some lands can be classified into one or more categories due to multiple uses that meet the criteria of more than
one definition. However, a ranking has been developed for assignment priority in these cases. The ranking process
is from highest to lowest priority based on the following order:
Settlements > Cropland > Forest Land > Grassland > Wetlands > Other Land
Settlements are given the highest assignment priority because they are extremely heterogeneous with a mosaic of
patches that include buildings, infrastructure, and travel corridors, but also open grass areas, forest patches,
riparian areas, and gardens. The latter examples could be classified as Grassland, Forest Land, Wetlands, and
Cropland, respectively, but when located in close proximity to settlement areas, they tend to be managed in a
unique manner compared to non-settlement areas. Consequently, these areas are assigned to the Settlements
land-use category. Cropland is given the second assignment priority, because cropping practices tend to dominate
management activities on areas used to produce food, forage, or fiber. The consequence of this ranking is that
crops in rotation with pasture are classified as Cropland, and land with woody plant cover that is used to produce
crops (e.g., orchards) is classified as Cropland, even though these areas may also meet the definitions of Grassland
or Forest Land, respectively. Similarly, Wetlands are considered Croplands if they are used for crop production,
such as rice or cranberries. Forest Land occurs next in the priority assignment because traditional forestry practices
tend to be the focus of the management activity in areas with woody plant cover that are not croplands (e.g.,
orchards) or settlements (e.g., housing subdivisions with significant tree cover). Grassland occurs next in the
ranking, while Wetlands and then Other Land complete the list.
The assignment priority does not reflect the level of importance for reporting greenhouse gas emissions and
removals on managed land, but is intended to classify all areas into a discrete land use category. Currently, the
IPCC does not make provisions in the guidelines for assigning land to multiple uses. For example, a wetland is
classified as Forest Land if the area has sufficient tree cover to meet the stocking and stand size requirements.
Similarly, wetlands are classified as Cropland if they are used for crop production, such as rice, or as Grassland if
they are composed principally of grasses, grass-like plants (i.e., sedges and rushes), forbs, or shrubs suitable for
grazing and browsing. Regardless of the classification, emissions and removals from these areas should be included
in the Inventory if the land is considered managed, and therefore impacted by anthropogenic activity in
accordance with the guidance provided by the IPCC (2006).
QA/QC and Verification
The land base derived from the NRI, FIA, and NLCD was compared to the Topological^ Integrated Geographic
Encoding and Referencing (TIGER) survey (U.S. Census Bureau 2010). The United States Census Bureau gathers
data on the population and economy, and has a database of land areas for the country. The area estimates of land-
use categories, based on NRI, FIA, and NLCD, are derived from remote sensing data instead of the land survey
approach used by the United States Census Survey. The Census does not provide a time series of land-use change
data or land management information, which is needed for estimating greenhouse gas emissions from land use
and land use change. Regardless, the Census does provide sufficient information to provide a check on the
Inventory data. There are 46 million more hectares of land in the United States according to the Census, compared
to the total area estimate of 936 million hectares derived from the combined NRI, FIA, and NLCD data. Much of this
difference is associated with open waters in coastal regions and the Great Lakes, which is included in the TIGER
Survey of the Census, but not included in the land representation using the NRI, FIA and NLCD. There is only a 0.4
Land Use, Land-Use Change, and Forestry 6-21

-------
percent difference when open water in coastal regions is removed from the TIGER data. General QC procedures for
data gathering and data documentation also were applied consistent with the QA/QC and Verification Procedures
described in Annex 8.
Recalculations Discussion
No recalculations were performed for the 1990 through 2018 portion of the time series, thus the land use areas for
2019 are assumed the same as 2018.
Planned Improvements
The next (i.e., 1990 through 2020) Inventory will be improved by using new NRI, FIA and possibly NLCD data to
update the time series for land representation, providing consistency between the total area of managed land in
the land representation section and the remainder of the Inventory. Another key planned improvement for the
Inventory is to fully incorporate area data by land-use type for U.S. Territories. Fortunately, most of the managed
land in the United States is included in the current land-use data, but a complete reporting of all lands in the
United States is a key goal for the near future. Preliminary land-use area data for U.S. Territories by land-use
category are provided in Box 6-2.
Box 6-2: Preliminary Estimates of Land Use in U.S. Territories
Several programs have developed land cover maps for U.S. Territories using remote sensing imagery, including
the Gap Analysis Program, Caribbean Land Cover project, National Land Cover Dataset, USFS Pacific Islands
Imagery Project, and the National Oceanic and Atmospheric Administration (NOAA) Coastal Change Analysis
Program (C-CAP). Land-cover data can be used to inform a land-use classification if there is a time series to
evaluate the dominate practices. For example, land that is principally used for timber production with tree cover
over most of the time series is classified as forest land even if there are a few years of grass dominance
following timber harvest. These products were reviewed and evaluated for use in the national Inventory as a
step towards implementing a planned improvement to include U.S. Territories in the land representation for the
Inventory. Recommendations are to use the NOAA C-CAP Regional Land Cover Database for the smaller island
Territories (U.S. Virgin Islands, Guam, Northern Marianas Islands, and American Samoa) because this program is
ongoing and therefore will be continually updated. The C-CAP product does not cover the entire territory of
Puerto Rico so the NLCD was used for this area. The final selection of land-cover products for these territories is
still under discussion. Results are presented below (in hectares). The total land area of all U.S. Territories is 1.05
million hectares, representing 0.1 percent of the total land base for the United States (see Table 6-7).
Table 6-7: Total Land Area (Hectares) by Land-Use Category for U.S. Territories

Puerto Rico
U.S. Virgin
Islands
Guam
Northern
Marianas
Islands
American
Samoa
Total
Cropland
19,712
138
236
289
389
20,764
Forest Land
404,004
13,107
24,650
25,761
15,440
482,962
Grasslands
299,714
12,148
15,449
13,636
1,830
342,777
Other Land
5,502
1,006
1,141
5,186
298
13,133
Settlements
130,330
7,650
11,146
3,637
1,734
154,496
Wetlands
24,525
4,748
1,633
260
87
31,252
Total
883,788
38,796
54,255
48,769
19,777
1,045,385
Note: Totals may not sum due to independent rounding.
Methods in the 2013 Supplement to the 2006 Guidelines for National Greenhouse Gas Inventories: Wetlands (IPCC
2014) have been applied to estimate emissions and removals from coastal wetlands. Specifically, greenhouse gas
6-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
emissions from coastal wetlands have been developed for the Inventory using the NOAA C-CAP land cover product.
The NOAA C-CAP product is not used directly in the land representation analysis, however, so a planned
improvement for the next (i.e., 1990 through 2020) Inventory is to reconcile the coastal wetlands data from the C-
CAP product with the wetlands area data provided in the NRI, FIA and NLCD. In addition, the current Inventory
does not include a classification of managed and unmanaged wetlands, except for remote areas in Alaska.
Consequently, there is a planned improvement to classify managed and unmanaged wetlands for the
conterminous United States and Hawaii, and more detailed wetlands datasets will be evaluated and integrated
into the analysis to meet this objective.
6.2 Forest Land Remaining Forest Land (CRF
Category 4A1)
Changes in Forest Carbon Stocks (CRF Category 4A1)
Delineation of Carbon Pools
For estimating carbon (C) stocks or stock change (flux), C in forest ecosystems can be divided into the following five
storage pools (IPCC 2006):
•	Aboveground biomass, which includes all living biomass above the soil including stem, stump, branches,
bark, seeds, and foliage. This category includes live understory.
•	Belowground biomass, which includes all living biomass of coarse living roots greater than 2 millimeters
(mm) diameter.
•	Dead wood, which includes all non-living woody biomass either standing, lying on the ground (but not
including litter), or in the soil.
•	Litter, which includes all duff, humus, and fine woody debris above the mineral soil and includes woody
fragments with diameters of up to 7.5 cm.
•	Soil organic C (SOC), including all organic material in soil to a depth of 1 meter but excluding the coarse
roots of the belowground pools.
In addition, there are two harvested wood pools included when estimating C flux:
•	Harvested wood products (HWP) in use.
•	HWP in solid waste disposal sites (SWDS).
Forest Carbon Cycle
Carbon is continuously cycled among the previously defined C storage pools and the atmosphere as a result of
biogeochemical processes in forests (e.g., photosynthesis, respiration, decomposition, and disturbances such as
fires or pest outbreaks) and anthropogenic activities (e.g., harvesting, thinning, and replanting). As trees
photosynthesize and grow, C is removed from the atmosphere and stored in living tree biomass. As trees die and
otherwise deposit litter and debris on the forest floor, C is released to the atmosphere and is also transferred to
the litter, dead wood, and soil pools by organisms that facilitate decomposition.
The net change in forest C is not equivalent to the net flux between forests and the atmosphere because timber
harvests do not cause an immediate flux of all harvested biomass C to the atmosphere. Instead, harvesting
transfers a portion of the C stored in wood to a "product pool." Once in a product pool, the C is emitted over time
as C02 in the case of decomposition and as C02, CH4, N20, CO, and NOxwhen the wood product combusts. The rate
Land Use, Land-Use Change, and Forestry 6-23

-------
of emission varies considerably among different product pools. For example, if timber is harvested to produce
energy, combustion releases C immediately, and these emissions are reported for information purposes in the
Energy sector while the harvest (i.e., the associated reduction in forest C stocks) and subsequent combustion are
implicitly estimated in the Land Use, Land-Use Change, and Forestry (LULUCF) sector (i.e., the portion of harvested
timber combusted to produce energy does not enter the HWP pools). Conversely, if timber is harvested and used
as lumber in a house, it may be many decades or even centuries before the lumber decays and C is released to the
atmosphere. If wood products are disposed of in SWDS, the C contained in the wood may be released many years
or decades later, or may be stored almost permanently in the SWDS. These latter fluxes, with the exception of CH4
from wood in SWDS, which is included in the Waste sector, are also estimated in the LULUCF sector.
Net Change in Carbon Stocks within Forest Land of the United States
This section describes the general method for quantifying the net changes in C stocks in the five C storage pools
and two harvested wood pools (a more detailed description of the methods and data is provided in Annex 3.13).
The underlying methodology for determining C stock and stock change relies on data from the national forest
inventory (NFI) conducted by the Forest Inventory and Analysis (FIA) program within the USDA Forest Service. The
annual NFI is implemented across all U.S. forest lands within the conterminous 48 states and Alaska and
inventories have been initiated in Hawaii and some of the U.S. Territories. The methods for estimation and
monitoring are continuously improved and these improvements are reflected in the C estimates (Domke et al.
2016; Domke et al. 2017). First, the total C stocks are estimated for each C storage pool at the individual NFI plot,
next the annual net changes in C stocks for each pool are estimated, and then the changes in stocks are summed
for all pools to estimate total net flux at the population level (e.g., U.S. state). Changes in C stocks from
disturbances, such natural disturbances (e.g., wildfires, insects/disease, wind) or harvesting, are included in the net
changes (See Box 6-3 for more information). For instance, an inventory conducted after a fire implicitly includes
only the C stocks remaining on the NFI plot. The IPCC (2006) recommends estimating changes in C stocks from
forest lands according to several land-use types and conversions, specifically Forest Land Remaining Forest Land
and Land Converted to Forest Land, with the former being lands that have been forest lands for 20 years or longer
and the latter being lands (i.e., croplands, grassland, wetlands, settlements and other lands) that have been
converted to forest lands for less than 20 years. The methods and data used to delineate forest C stock changes by
these two categories continue to improve and in order to facilitate this delineation, a combination of modeling
approaches for carbon estimation were used in this Inventory.
Forest Area in the United States
Approximately 32 percent of the U.S. land area is estimated to be forested based on the U.S. definition of forest
land as provided in Section Representation of the U.S. Land Base. All annual NFI plots included in the public FIA
database as of August 2020 (which includes data collected through 2019) were used in this Inventory. Since area
estimates for some land use categories were not updated in the Land Representation in the current Inventory
there are differences in the area estimates reported in this section and those reported in Section 6.1
Representation of the U.S. Land Base. The NFIs from each of the conterminous 48 states (CONUS; USDA Forest
Service 2020a, 2020b) and Alaska comprise an estimated 279 million hectares of forest land that are considered
managed and are included in the current Inventory. Some differences also exist in forest land area estimates from
the latest update to the Resources Planning Act (RPA) Assessment (Oswalt et al. 2019) and the forest land area
estimates included in this report, which are based on the annual NFI data through 2019 for all states (USDA Forest
Service 2020b, Nelson et al. 2020). Sufficient annual NFI data are not yet available for Hawaii and the U.S.
Territories to include them in them in this section of the Inventory but estimates of these areas are included in
Oswalt et al. (2019). While Hawaii and U.S. Territories have relatively small areas of forest land and thus may not
substantially influence the overall C budget for forest land, these regions will be added to the forest C estimates as
sufficient data become available. Since HI was not included in this section of the current Inventory there are small
differences in the area estimates reported in this section and those reported in Section 6 Representation of the
6-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
U.S. Land Base.29 Agroforestry systems that meet the definition of forest land are also not currently included in the
current Inventory since they are not explicitly inventoried (i.e., classified as an agroforestry system) by either the
FIA program or the Natural Resources Inventory (NRI)30 of the USDA Natural Resources Conservation Service (Perry
etal. 2005).
An estimated 67 percent (208 million hectares) of U.S. forests in Alaska, and Hawaii and the conterminous United
States are classified as timberland, meaning they meet minimum levels of productivity and have not been removed
from production. Approximately ten percent of Alaska forest land and 73 percent of forest land in the
conterminous United States are classified as timberland. Of the remaining non-timberland, nearly 33 million
hectares are reserved forest lands (withdrawn by law from management for production of wood products) and 102
million hectares are lower productivity forest lands (Oswalt et al. 2019). Historically, the timberlands in the
conterminous 48 states have been more frequently or intensively surveyed than the forest land removed from
production because it does not meet the minimum level of productivity.
Since the late 1980s, gross forest land area in Alaska, Hawaii, and the conterminous United States has increased by
about 13 million hectares (Oswalt et al. 2019) with the southern region of the United States containing the most
forest land (Figure 6-4). A substantial portion of this accrued forest land is from the conversion of abandoned
croplands to forest (e.g., Woodall et al. 2015b). Estimated forest land area in the CONUS and Alaska represented
here is stable but there are substantial conversions as described in Section 6 Representation of the U.S. Land Base
and each of the land conversion sections for each land use category (e.g., Land Converted to Cropland, Land
Converted to Grassland). The major influences on the net C flux from forest land across the 1990 to 2019 time
series are management activities, natural disturbance, and the ongoing impacts of current and previous land-use
conversions. These activities affect the net flux of C by altering the amount of C stored in forest ecosystems and
also the area converted to forest land. For example, intensified management of forests that leads to an increased
rate of growth of aboveground biomass (and possible changes to the other C storage pools) may increase the
eventual biomass density of the forest, thereby increasing the uptake and storage of C in the aboveground biomass
pool.31 Though harvesting forests removes much of the C in aboveground biomass (and possibly changes C density
in other pools), on average, the estimated volume of annual net growth in aboveground tree biomass in the
conterminous United States is about double the volume of annual removals on timberlands (Oswalt et al. 2019).
The net effects of forest management and changes in Forest Land Remaining Forest Land are captured in the
estimates of C stocks and fluxes presented in this section.
29	See Annex 3.13, Table A-214 for annual differences between the forest area reported in Section 6 Representation of the U.S.
Land Base and Section 6.2 Forest Land Remaining Forest Land.
30	The Natural Resources Inventory of the USDA Natural Resources Conservation Service is described in Section 6
Representation of the U.S. Land Base.
31	The term "biomass density" refers to the mass of live vegetation per unit area. It is usually measured on a dry-weight basis.
A carbon fraction of 0.5 is used to convert dry biomass to C (USDA Forest Service 2020d).
Land Use, Land-Use Change, and Forestry 6-25

-------
Figure 6-4: Changes in Forest Area by Region for Forest Land Remaining Forest Land in the
conterminous United States and Alaska (1990-2019)
100-1
s 90 H
0
2
| 80-
c
0
1	70H
CD
Ł
™ 60 H
8
k-
o
50H
40
• South
¦ North
Pacific
Coast
Rocky
Mountain
| I i I I | I I I I | I I
1990 1995 2000
I I | I I I I | I I I I | I I I I
2005 2010 2015
Year
South
Paahc
Coast
Kocky
Mountain
Forest Carbon Stocks and Stock Change
In Forest Land Remaining Forest Land, forest management practices, the regeneration of forest areas cleared more
than 20 years prior to the reporting year, and timber harvesting have resulted in net uptake (i.e., net sequestration
or accumulation) of C each year from 1990 through 2019. The rate of forest clearing in the 17th century following
European settlement had slowed by the late 19th century. Through the later part of the 20th century many areas of
previously forested land in the United States were allowed to revert to forests or were actively reforested. The
impacts of these land-use changes still influence C fluxes from these forest lands. More recently, the 1970s and
1980s saw a resurgence of federally-sponsored forest management programs (e.g., the Forestry Incentive
Program) and soil conservation programs (e.g., the Conservation Reserve Program), which have focused on tree
planting, improving timber management activities, combating soil erosion, and converting marginal cropland to
forests. In addition to forest regeneration and management, forest harvests and natural disturbance have also
affected net C fluxes. Because most of the timber harvested from U.S. forest land is used in wood products, and
many discarded wood products are disposed of in SWDS rather than by incineration, significant quantities of C in
harvested wood are transferred to these long-term storage pools rather than being released rapidly to the
atmosphere (Skog 2008). Maintaining current harvesting practices and regeneration activities on these forested
lands, along with continued input of harvested products into the HWP pool, C stocks in the Forest Land Remaining
Forest Land category are likely to continue to increase in the near term, though possibly at a lower rate. Changes in
6-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
C stocks in the forest ecosystem and harvested wood pools associated with Forest Land Remaining Forest Land
were estimated to result in net uptake of 691.8 MMT C02 Eq. (188.7 MMT C) in 2019 (Table 6-8, Table 6-9, Table A-
211, A-212 and state-level estimates in A-215). The estimated net uptake of C in the Forest Ecosystem was 583.3
MMT C02 Eq. (159.1 MMT C) in 2019 (Table 6-8 and Table 6-9). The majority of this uptake in 2019, 394.0 MMT
C02 Eq. (107.4 MMT C), was from aboveground biomass. Overall, estimates of average C density in forest
ecosystems (including all pools) increased consistently over the time series with an average of approximately 191
MT C ha 1 from 1990 to 2019. This was calculated by dividing the Forest Land area estimates by Forest Ecosystem C
Stock estimates for every year (see Table 6-10 and Table A-213) and then calculating the mean across the entire
time series, i.e., 1990 through 2019. The increasing forest ecosystem C density when combined with relatively
stable forest area results in net C accumulation over time. Aboveground live biomass is responsible for the
majority of net C uptake among all forest ecosystem pools (Figure 6-5). These increases may be influenced in some
regions by reductions in C density or forest land area due to natural disturbances (e.g., wildfire, weather,
insects/disease), particularly in Alaska. The inclusion of all managed forest land in Alaska has increased the
interannual variability in carbon stock change estimates over the time series and much of this variability can be
attributed to severe fire years. The distribution of carbon in forest ecosystems in Alaska is substantially different
from forests in the CONUS. In Alaska, more than 12 percent of forest ecosystem C is stored in the litter carbon pool
whereas in the CONUS only 6 percent of the total ecosystem C stocks are in the litter pool. Much of the litter
material in forest ecosystems is combusted during fire (IPCC 2006) which is why there are substantial C losses in
this pool during severe fire years (Figure 6-5, A-229).
The estimated net uptake of C in HWP was 108.5 MMT C02 Eq. (29.6 MMT C) in 2019 (Table 6-8,Table 6-9, Table A-
211, and A-212). The majority of this uptake, 69.3 MMT C02 Eq. (18.9 MMT C), was from wood and paper in SWDS.
Products in use were an estimated 39.2 MMT C02 Eq. (10.7 MMT C) in 2019.
Table 6-8: Net CO2 Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land
and Harvested Wood Pools (MMT CO2 Eq.)
Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Forest Ecosystem	(663.8)
Aboveground Biomass	(456.4)
Belowground Biomass	(103.7)
Dead Wood	(97.3)
Litter	(8.1)
Soil (Mineral)	1.5
Soil (Organic)	(0.6)
Drained Organic Soil3	0.8
Harvested Wood	(123.8)
Products in Use	(54.8)
SWDS	(69.0)
(555.5)
(401.3)
(92.0)
(93.5)
32.2
(1.5)
(0.2)
0.8
(106.0)
(42.6)
(63.4)
(582.7)
(414.2)
(92.6)
(98.7)
30.5
(7.3)
(1.1)
0.8
(88.7)
(24.6)
(64.1)
(629.5)
(421.3)
(95.0)
(105.1)
(3.2)
(6.8)
1.2
0.8
(92.4)
(27.8)
(64.6)
(564.0)
(395.1)
(89.2)
(97.1)
0.2
14.3
2.1
0.8
(95.7)
(30.3)
(65.5)
(599.8)
(402.4)
(90.9)
(101.7)
(2.3)
(4.5)
1.2
0.8
(98.8)
(31.5)
(67.2)
(583.3)
(394.0)
(89.2)
(99.3)
(0.5)
(2.4)
1.2
0.8
(108.5)
(39.2)
(69.3)
Total Net Flux
(787.6)
(661.5)
(671.4) (721.9) (659.7) (698.6) (691.8)
Notes: Forest ecosystem C stock changes do not include forest stocks in U.S. Territories because managed
forest land for U.S. Territories is not currently included in Section 6 Representation of the U.S. Land Base.
The forest ecosystem C stock changes do not include Hawaii because there is not sufficient NFI data to
support inclusion at this time. However, managed forest land area for Hawaii is included in Section 6
Representation of the U.S. Land Base so there are small differences in the forest land area estimates in this
Section and Section 6. See Annex 3.13, Table A-214 for annual differences between the forest area
Land Use, Land-Use Change, and Forestry 6-27

-------
reported in Section 6 Representation of the U.S. Land Base and Section 6.2 Forest Land Remaining Forest
Land. The forest ecosystem C stock changes do not include trees on non-forest land (e.g., agroforestry
systems and settlement areas—see Section 6.10 Settlements Remaining Settlements for estimates of C
stock change from settlement trees). Forest ecosystem C stocks on managed forest land in Alaska were
compiled using the gain-loss method as described in Annex 3.13. Parentheses indicate net C uptake (i.e., a
net removal of C from the atmosphere). Total net flux is an estimate of the actual net flux between the
total forest C pool and the atmosphere. Harvested wood estimates are based on results from annual
surveys and models. Totals may not sum due to independent rounding.
a These estimates include C stock changes from drained organic soils from both Forest Land Remaining Forest
Land and Land Converted to Forest Land. See the section below on C02, CH4, and N20 Emissions from
Drained Organic Soils for the methodology used to estimate the C02 emissions from drained organic soils.
Also, Table 6-20 and Table 6-21 for non-C02 emissions from drainage of organic soils from both Forest Land
Remaining Forest Land and Land Converted to Forest Land.
Table 6-9: Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land
and Harvested Wood Pools (MMT C)
Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Forest Ecosystem
(181.0)
(151.5)
(158.9)
(171.7)
(153.8)
(163.6)
(159.1)
Aboveground Biomass
(124.5)
(109.5)
(113.0)
(114.9)
(107.7)
(109.7)
(107.4)
Belowground Biomass
(28.3)
(25.1)
(25.3)
(25.9)
(24.3)
(24.8)
(24.3)
Dead Wood
(26.5)
(25.5)
(26.9)
(28.7)
(26.5)
(27.7)
(27.1)
Litter
(2.2)
8.8
8.3
(0.9)
0.1
(0.6)
(0.1)
Soil (Mineral)
0.4
(0.4)
(2.0)
(1.9)
3.9
(1.2)
(0.7)
Soil (Organic)
(0.2)
(0.1)
(0.3)
0.3
0.6
0.3
0.3
Drained Organic Soil3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Harvested Wood
(33.8)
(28.9)
(24.2)
(25.2)
(26.1)
(26.9)
(29.6)
Products in Use
(14.9)
(11.6)
(6.7)
(7.6)
(8.3)
(8.6)
(10.7)
SWDS
(18.8)
(17.3)
(17.5)
(17.6)
(17.9)
(18.3)
(18.9)
Total Net Flux
(214.8)
(180.4)
(183.1)
(196.9)
(179.9)
(190.5)
(188.7)
Notes: Forest ecosystem C stock changes do not include forest stocks in U.S. Territories because managed
forest land for U.S. Territories is not currently included in Section 6 Representation of the U.S. Land Base. The
forest ecosystem C stock changes do not include Hawaii because there is not sufficient NFI data to support
inclusion at this time. However, managed forest land area for Hawaii is included in 6 Representation of the
U.S. Land Base so there are small differences in the forest land area estimates in this Section and Section 6.
See Annex 3.13, Table A-214 for annual differences between the forest area reported in Section 6
Representation of the U.S. Land Base and Section 6.2 Forest Land Remaining Forest Land. The forest
ecosystem C stock changes do not include trees on non-forest land (e.g., agroforestry systems and settlement
areas—see Section 6.10 Settlements Remaining Settlements for estimates of C stock change from settlement
trees). Forest ecosystem C stocks on managed forest land in Alaska were compiled using the gain-loss method
as described in Annex 3.13. Parentheses indicate net C uptake (i.e., a net removal of C from the atmosphere).
Total net flux is an estimate of the actual net flux between the total forest C pool and the atmosphere.
Harvested wood estimates are based on results from annual surveys and models. Totals may not sum due to
independent rounding.
a These estimates include carbon stock changes from drained organic soils from both Forest Land Remaining
Forest Land and Land Converted to Forest Land. See the section below on C02, CH4, and N20 Emissions from
Drained Organic Soils for the methodology used to estimate the C flux from drained organic soils. Also, see
Table 6-20 and Table 6-21 for greenhouse gas emissions from non-C02 gases changes from drainage of organic
soils from Forest Land Remaining Forest Land and Land Converted to Forest Land.
6-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Stock estimates for forest ecosystem and harvested wood C storage pools are presented in Table 6-10. Together,
the estimated aboveground biomass and soil C pools account for a large proportion of total forest ecosystem C
stocks. Forest land area estimates are also provided in Table 6-10, but these do not precisely match those in
Section 6 Representation of the U.S. Land Base for Forest Land Remaining Forest Land. This is because the forest
land area estimates in Table 6-10 only include managed forest land in the conterminous 48 states and Alaska while
the area estimates in Section 6 include all managed forest land in Hawaii. Differences also exist because forest land
area estimates are based on the latest NFI data through 2019 and woodland areas previously included as forest
land have been separated and included in the Grassland categories in this Inventory.32
Table 6-10: Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C)

1990
2005
2016
2017
2018
2019
2020
Forest Area (1,000 ha)
279,661
279,491
279,533
279,511
279,483
279,386
279,289
Carbon Pools (MMT C)







Forest Ecosystem
50,913
53,489
55,284
55,456
55,610
55,774
55,933
Aboveground Biomass
11,810
13,584
14,820
14,935
15,043
15,152
15,260
Belowground Biomass
2,319
2,723
3,004
3,030
3,054
3,079
3,103
Dead Wood
2,049
2,446
2,743
2,771
2,798
2,825
2,852
Litter
3,656
3,655
3,636
3,637
3,637
3,638
3,638
Soil (Mineral)
25,145
25,145
25,147
25,149
25,145
25,146
25,147
Soil (Organic)
5,934
5,936
5,935
5,934
5,934
5,933
5,933
Harvested Wood
1,895
2,353
2,591
2,616
2,642
2,669
2,699
Products in Use
1,249
1,447
1,497
1,505
1,513
1,521
1,532
SWDS
646
906
1,094
1,112
1,129
1,148
1,167
Total C Stock
52,808
55,842
57,875
58,072
58,252
58,443
58,632
Notes: Forest area and C stock estimates include all Forest Land Remaining Forest Land in the conterminous 48
states and Alaska. Forest ecosystem C stocks do not include forest stocks in U.S. Territories because managed
forest land for U.S. Territories is not currently included in Section 6 Representation of the U.S. Land Base. The
forest ecosystem C stocks do not include Hawaii because there is not sufficient NFI data to support inclusion at this
time. However, managed forest land area for Hawaii is included in Section 6 Representation of the U.S. Land Base
so there are small differences in the forest land area estimates in this Section and Section 6. See Annex 3.13, Table
A-214 for annual differences between the forest area reported in Section 6 Representation of the U.S. Land Base
and Section 6.2 Forest Land Remaining Forest Land. The forest ecosystem C stocks do not include trees on non-
forest land (e.g., agroforestry systems and settlement areas—see Section 6.10 Settlements Remaining Settlements
for estimates of C stock change from settlement trees). Forest ecosystem C stocks on managed forest land in
Alaska were compiled using the gain-loss method as described in Annex 3.13. Harvested wood product stocks
include exports, even if the logs are processed in other countries, and exclude imports. Harvested wood estimates
are based on results from annual surveys and models. Totals may not sum due to independent rounding.
Population estimates compiled using FIA data are assumed to represent stocks as of January 1 of the inventory
year. Flux is the net annual change in stock. Thus, an estimate of flux for 2019 requires estimates of C stocks for
2019 and 2020.
32 See Annex 3.13, Table A-214 for annual differences between the forest area reported in Section 6 Representation of the U.S.
Land Base and Section 6.2 Forest Land Remaining Forest Land.
Land Use, Land-Use Change, and Forestry 6-29

-------
Figure 6-5; Estimated Net Annual Changes in C Stocks for All C Pools in Forest Land
Remaining Forest Land in the Conterminous U.S. and Alaska (1990-2019)
2 -
1 &
"65
O
20 H
o-
-20-
-40-
-60-
o>~ -80-
c m
¦I S? -100-
<0 03
I "5 -120-
"8 I -140-
iS to
to § -160-
|| -180-
-200-
-220
I
1990
A.
1995
I i I I | i
2000
I I
I
2005
Year
i i i | i
2010
2015
¦	All forest ecosystem pools
Aboveground biomass
¦	Belowground biomass
Dead wood
¦	Litter
Soil (mineral)
Soil (organic)
Drained Organic Soil
Harvested Wood Products (HWP)
Products in use
Solid waste disposal sites
Total net change
(forest ecosystem + HWP)
Box 6-3: C02 Emissions from Forest Fires
As stated previously, the forest inventory approach implicitly includes all C losses due to disturbances such as
forest fires, because only C remaining in the forest is estimated. Net C stock change is estimated by subtracting
consecutive C stock estimates. A forest fire disturbance removes C from the forest. The inventory data on which
net C stock estimates are based already reflect this C loss. Therefore, estimates of net annual changes in C
stocks for U.S. forest iand already includes C02 emissions from forest fires occurring in the conterminous states
as well as the portion of managed forest lands in Alaska. Because it is of interest to quantify the magnitude of
C02 emissions from fire disturbance, these separate estimates are highlighted here. Note that these C02
estimates are based on the same methodology as applied for the non-C02 greenhouse gas emissions from forest
fires that are also quantified in a separate section below as required by IPCC Guidance and UNFCCC reporting
requirements.
The IPCC (2006) methodology with U.S.-specific data on annual area burned, potential fuel availability, and fire-
specific severity and combustion were combined with IPCC default factors as needed to estimate C02 emissions
from forest fires. The latest information on area burned is used to compile fire emissions for the United States.
At the time this Inventory was compiled data were limited for 2018 and data for 2019 fires were unavailable;
6-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
therefore 2017 or 2018, the most recent available estimate, is applied to 2019. The 2019 estimates will be
updated in subsequent reports as fire data become available. Estimated C02 emissions for wildfires in the
conterminous 48 states and in Alaska as well as prescribed fires in 2019 were 126 MMT C02 per year (Table
6-11). This estimate is an embedded component of the net annual forest C stock change estimates provided
previously (i.e., Table 6-9), but this separate approach to estimate C02 emissions is necessary in order to
associate these emissions with fire. See the discussion in Annex 3.13 for more details on this methodology. Note
that in Alaska a portion of the forest lands are considered unmanaged, therefore the estimates for Alaska
provided in Table 6-11 include only managed forest land within the state, which is consistent with C stock
change estimates provided above.
Table 6-11: Estimates of CO2 (MMT per Year) Emissions from Forest Fires in the
Conterminous 48 States and Alaska3

C02 emitted from




Wildfires in the
C02 emitted from
C02 emitted from


Conterminous 48
Wildfires in
Prescribed Fires
Total C02 emitted
Year
States (MMTyr1)
AlaskafMMTyr1)
(MMTyr1)
(MMTyr1)
1990
6.2
5.3
0.2
11.7

2005
20.5
44.1
1.5
66.2

2015
116.4
40.7
6.1
163.2
2016
33.7
1.7
9.7
45.2
2017
137.2
1.5
8.4
147.1
2018
115.5
2.9
8.0
126.4
2019b
115.5
2.9
8.0
126.4
Note: Totals may not sum due to independent rounding.
a These emissions have already been included in the estimates of net annual changes in C stocks, which include the
amount sequestered minus any emissions, including the assumption that combusted wood may continue to
decay through time.
b Some fire data for 2018 and all of 2019 were unavailable when these estimates were summarized; therefore 2017
or 2018, the most recent available estimate, is applied to 2019.
Methodology
The methodology described herein is consistent with IPCC (2006). Forest ecosystem C stocks and net annual C
stock change were determined according to the stock-difference method for the CONUS, which involved applying
C estimation factors to annual forest inventories across time to obtain C stocks and then subtracting between the
years to obtain the stock change. The gain-loss method was used to estimate C stocks and net annual C stock
changes in Alaska. The approaches for estimating carbon stocks and stock changes on Forest Land Remaining
Forest Land are described in Annex 3.13. All annual NFI plots available in the public FIA database (USDA Forest
Service 2020b) were used in the current Inventory. Additionally, NFI plots established and measured in 2014 as
part of a pilot inventory in interior Alaska were also included in this report as were plots established and measured
in 2015 and 2016 as part of the operational NFI in interior Alaska. Some of the data from the pilot and operational
NFI in interior Alaska are not yet available in the public FIA database. Only plots which meet the definition of forest
land (see Section 6 Representation of the U.S. Land Base) are measured in the NFI, as part of the pre-field process
in the FIA program, all plots or portions of plots (i.e., conditions) are classified into a land use category. This land
use information on each forest and non-forest plot was used to estimate forest land area and land converted to
and from forest land over the time series. The estimates in this section of the report are based on land use
information from the NFI and they may differ with the other land use categories where area estimates reported in
the Land Representation were not updated (see Section 6.1 Representation of the U.S. Land Base) Further, HI was
not included in this section of the current Inventory so that also contributes to small differences in the area
Land Use, Land-Use Change, and Forestry 6-31

-------
estimates reported in this section and those reported in Section 6.1 Representation of the U.S. Land Base (See
Annex 3.13 for details on differences). To implement the stock-difference approach, forest Land conditions in the
CONUS were observed on NFI plots at time t0 and at a subsequent time ti=t0+s, where s is the time step (time
measured in years) and is indexed by discrete (e.g., 5 year) forest age classes. The inventory from t0 was then
projected from ti to 2019. This projection approach requires simulating changes in the age-class distribution
resulting from forest aging and disturbance events and then applying C density estimates for each age class to
obtain population estimates for the nation. To implement the gain-loss approach in Alaska, forest land conditions
in Alaska were observed on NFI plots from 2004 to 2017. Plot-level data from the NFI were harmonized with
auxiliary data describing climate, forest structure, disturbance, and other site-specific conditions to develop non-
parametric models to predict carbon stocks by forest ecosystem carbon pool as well as fluxes over the entire
inventory period, 1990 to 2019. First, carbon stocks for each forest ecosystem carbon pool were predicted for the
year 2016 for all base intensity NFI plot locations (representing approximately 2,403 ha) in coastal southeast and
southcentral Alaska and for 1/5 intensity plots in interior Alaska (representing 12,015 ha). Next, the
chronosequence of sampled NFI plots and auxiliary information (e.g., climate, forest structure, disturbance, and
other site-specific data) were used to predict annual gains and losses by forest ecosystem carbon pool. The annual
gains and losses were then combined with the stock estimates and disturbance information to compile plot- and
population-level carbon stocks and fluxes for each year from 1990 to 2019. To estimate C stock changes in
harvested wood, estimates were based on factors such as the allocation of wood to various primary and end-use
products as well as half-life (the time at which half of the amount placed in use will have been discarded from use)
and expected disposition (e.g., product pool, SWDS, combustion). An overview of the different methodologies and
data sources used to estimate the C in forest ecosystems within the conterminous states and Alaska and harvested
wood products for all of the United States is provided below. See Annex 3.13 and Domke et al. (In prep) for details
and additional information related to the methods and data.
Forest Ecosystem Carbon from Forest Inventory
The United States applied the compilation approach described in Woodall et al. (2015a) for the current Inventory
which removes the older periodic inventory data, which may be inconsistent with annual inventory data, from the
estimation procedures and enables the delineation of forest C accumulation by forest growth, land use change,
and natural disturbances such as fire. Development will continue on a system that attributes changes in forest C to
disturbances and delineates Land Converted to Forest Land from Forest Land Remaining Forest Land. As part of this
development, C pool science will continue and will be expanded to improve the estimates of C stock transfers from
forest land to other land uses and include techniques to better identify land use change (see the Planned
Improvements section below).
Unfortunately, the annual FIA inventory system does not extend into the 1970s, necessitating the adoption of a
system to estimate carbon stocks prior to the establishment of the annual forest inventory. The estimation of
carbon stocks prior to the annual national forest inventory consisted of a modeling framework comprised of a
forest dynamics module (age transition matrices) and a land use dynamics module (land area transition matrices).
The forest dynamics module assesses forest uptake, forest aging, and disturbance effects (e.g., disturbances such
as wind, fire, and floods identified by foresters on inventory plots). The land use dynamics module assesses C stock
transfers associated with afforestation and deforestation (Woodall et al. 2015b). Both modules are developed
from land use area statistics and C stock change or C stock transfer by age class. The required inputs are estimated
from more than 625,000 forest and non-forest observations recorded in the FIA national database (U.S. Forest
Service 2020a, b, c). Model predictions prior to the annual inventory period are constructed from the estimation
system using the annual estimates. The estimation system is driven by the annual forest inventory system
conducted by the FIA program (Frayer and Furnival 1999; Bechtold and Patterson 2005; USDA Forest Service
2020d, 2020a). The FIA program relies on a rotating panel statistical design with a sampling intensity of one 674.5
m2 ground plot per 2,403 ha of land and water area. A five-panel design, with 20 percent of the field plots typically
measured each year within a state, is used in the eastern United States and a ten-panel design, with typically 10
percent of the field plots measured each year within a state, is used in the western United States. The
interpenetrating hexagonal design across the U.S. landscape enables the sampling of plots at various intensities in
a spatially and temporally unbiased manner. Typically, tree and site attributes are measured with higher sample
6-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
intensity while other ecosystem attributes such as downed dead wood are sampled during summer months at
lower intensities. The first step in incorporating FIA data into the estimation system is to identify annual inventory
datasets by state. Inventories include data collected on permanent inventory plots on forest lands and were
organized as separate datasets, each representing a complete inventory, or survey, of an individual state at a
specified time. Many of the annual inventories reported for states are represented as "moving window" averages,
which mean that a portion—but not all—of the previous year's inventory is updated each year (USDA Forest
Service 2020d). Forest C estimates are organized according to these state surveys, and the frequency of surveys
varies by state.
Using this FIA data, separate estimates were prepared for the five C storage pools identified by IPCC (2006) and
described above. All estimates were based on data collected from the extensive array of permanent, annual forest
inventory plots and associated models (e.g., live tree belowground biomass) in the United States (USDA Forest
Service 2020b, 2020c). Carbon conversion factors were applied at the disaggregated level of each inventory plot
and then appropriately expanded to population estimates.
Carbon in Biomass
Live tree C pools include aboveground and belowground (coarse root) biomass of live trees with diameter at breast
height (dbh) of at least 2.54 cm at 1.37 m above the litter. Separate estimates were made for above- and
belowground biomass components. If inventory plots included data on individual trees, aboveground and
belowground (coarse roots) tree C was based on Woodall et al. (2011a), which is also known as the component
ratio method (CRM), and is a function of tree volume, species, and diameter. An additional component of foliage,
which was not explicitly included in Woodall et al. (2011a), was added to each tree following the same CRM
method.
Understory vegetation is a minor component of biomass, which is defined in the FIA program as all biomass of
undergrowth plants in a forest, including woody shrubs and trees less than 2.54 cm dbh. For this Inventory, it was
assumed that 10 percent of total understory C mass is belowground (Smith et al. 2006). Estimates of C density
were based on information in Birdsey (1996) and biomass estimates from Jenkins et al. (2003). Understory biomass
represented over 1 percent of C in biomass, but its contribution rarely exceeded 2 percent of the total carbon
stocks or stock changes across all forest ecosystem C pools each year.
Carbon in Dead Organic Matter
Dead organic matter was initially calculated as three separate pools—standing dead trees, downed dead wood,
and litter—with C stocks estimated from sample data or from models as described below. The standing dead tree C
pool includes aboveground and belowground (coarse root) biomass for trees of at least 12.7 cm dbh. Calculations
followed the basic method applied to live trees (Woodall et al. 2011a) with additional modifications to account for
decay and structural loss (Domke et al. 2011; Harmon et al. 2011). Downed dead wood estimates are based on
measurement of a subset of FIA plots for downed dead wood (Domke et al. 2013; Woodall and Monleon 2008;
Woodall et al. 2013). Downed dead wood is defined as pieces of dead wood greater than 7.5 cm diameter, at
transect intersection, that are not attached to live or standing dead trees. This includes stumps and roots of
harvested trees. To facilitate the downscaling of downed dead wood C estimates from the state-wide population
estimates to individual plots, downed dead wood models specific to regions and forest types within each region
are used. Litter C is the pool of organic C (also known as duff, humus, and fine woody debris) above the mineral
soil and includes woody fragments with diameters of up to 7.5 cm. A subset of FIA plots are measured for litter C.
A modeling approach, using litter C measurements from FIA plots (Domke et al. 2016) was used to estimate litter C
for every FIA plot used in the estimation framework.
Carbon in Forest Soil
Soil carbon is the largest terrestrial C sink with much of that C in forest ecosystems. The FIA program has been
consistently measuring soil attributes as part of the annual inventory since 2001 and has amassed an extensive
inventory of soil measurement data on forest land in the conterminous United States and coastal Alaska (O'Neill et
al. 2005). Observations of mineral and organic soil C on forest land from the FIA program and the International Soil
Land Use, Land-Use Change, and Forestry 6-33

-------
Carbon Monitoring Network were used to develop and implement a modeling approach that enabled the
prediction of mineral and organic (i.e., undrained organic soils) soil C to a depth of 100 cm from empirical
measurements to a depth of 20 cm and included site-, stand-, and climate-specific variables that yield predictions
of soil C stocks specific to forest land in the United States (Domke et al. 2017). This new approach allowed for
separation of mineral and organic soils, also referred to as Histosols, in the Forest Land Remaining Forest Land
category. Note that mineral and organic (i.e., undrained organic soils) soil C stock changes are reported to a depth
of 100 cm for Forest Land Remaining Forest Land to remain consistent with past reporting in this category,
however for consistency across land-use categories mineral (e.g., cropland, grassland, settlements) soil C is
reported to a depth of 30 cm in Section 6.3 Land Converted to Forest Land. Estimates of C stock changes from
organic soils shown in Table 6-8 and Table 6-9 include separately the emissions from drained organic forest soils,
the methods used to develop these estimates can be found in the Drained Organic Soils section below.
Harvested Wood Carbon
Estimates of the HWP contribution to forest C sinks and emissions (hereafter called "HWP contribution") were
based on methods described in Skog (2008) using the WOODCARB II model. These methods are based on IPCC
(2006) guidance for estimating the HWP contribution. IPCC (2006) provides methods that allow for reporting of
HWP contribution using one of several different methodological approaches: Production, stock change and
atmospheric flow, as well as a default method that assumes there is no change in HWP C stocks (see Annex 3.13
for more details about each approach). The United States uses the production approach to report HWP
contribution. Under the production approach, C in exported wood was estimated as if it remains in the United
States, and C in imported wood was not included in the estimates. Though reported U.S. HWP estimates are based
on the production approach, estimates resulting from use of the two alternative approaches, the stock change and
atmospheric flow approaches, are also presented for comparison (see Annex 3.13). Annual estimates of change
were calculated by tracking the annual estimated additions to and removals from the pool of products held in end
uses (i.e., products in use such as housing or publications) and the pool of products held in SWDS. The C loss from
harvest is reported in the Forest Ecosystem component of the Forest Land Remaining Forest Land and Land
Converted to Forest Land sections and for information purposes in the Energy sector, but the non-C02 emissions
associated with biomass energy are included in the Energy sector emissions (see Chapter 3).
Solidwood products include lumber and panels. End-use categories for solidwood include single and multifamily
housing, alteration and repair of housing, and other end uses. There is one product category and one end-use
category for paper. Additions to and removals from pools were tracked beginning in 1900, with the exception of
additions of softwood lumber to housing, which began in 1800. Solidwood and paper product production and
trade data were taken from USDA Forest Service and other sources (Hair and Ulrich 1963; Hair 1958; USDC Bureau
of Census 1976; Ulrich 1985,1989; Steer 1948; AF&PA 2006a, 2006b; Howard 2003, 2007, Howard and Jones 2016,
Howard and Liang 2019). Estimates for disposal of products reflects the change over time in the fraction of
products discarded to SWDS (as opposed to burning or recycling) and the fraction of SWDS that were in sanitary
landfills versus dumps.
There are five annual HWP variables that were used in varying combinations to estimate HWP contribution using
any one of the three main approaches listed above. These are:
(IA)	annual change of C in wood and paper products in use in the United States,
(IB)	annual change of C in wood and paper products in SWDS in the United States,
(2A) annual change of C in wood and paper products in use in the United States and other countries where the
wood came from trees harvested in the United States,
(2B) annual change of C in wood and paper products in SWDS in the United States and other countries where
the wood came from trees harvested in the United States,
(3)	C in imports of wood, pulp, and paper to the United States,
(4)	C in exports of wood, pulp and paper from the United States, and
6-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
(5) C in annual harvest of wood from forests in the United States.
The sum of variables 2A and 2B yielded the estimate for HWP contribution under the production estimation
approach. A key assumption for estimating these variables that adds uncertainty in the estimates was that
products exported from the United States and held in pools in other countries have the same half-lives for
products in use, the same percentage of discarded products going to SWDS, and the same decay rates in SWDS as
they would in the United States.
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis placed bounds on the flux estimates for forest ecosystems through a
combination of sample-based and model-based approaches to uncertainty for forest ecosystem C02 flux using IPCC
Approach 1 (Table 6-12 and Table A-215 for state-level uncertainties). A Monte Carlo Stochastic Simulation of the
methods described above, and probabilistic sampling of C conversion factors, were used to determine the HWP
uncertainty using IPCC Approach 2. See Annex 3.13 for additional information. The 2019 net annual change for
forest C stocks was estimated to be between -766.6 and -618.3 MMT C02 Eq. around a central estimate of-691.8
MMT C02 Eq. at a 95 percent confidence level. This includes a range of -651.8 to -514.9 MMT C02 Eq. around a
central estimate of-583.3 MMT C02 Eq. for forest ecosystems and -138.5 to -81.7 MMT C02 Eq. around a central
estimate of-108.5 MMT C02 Eq. for HWP. An error in the code used to compile the components of forest
ecosystem uncertainty was identified while aggregating estimates and uncertainties by individual state in this
Inventory. The error has been corrected in Table 6-12 resulting in a reduction in the uncertainty range relative to
the flux estimate for forest ecosystems and thus, total forest uncertainty.
Table 6-12: Quantitative Uncertainty Estimates for Net CO2 Flux from Forest Land
Remaining ForestLanch Changes in Forest C Stocks (MMT CO2 Eq. and Percent)
2019 Flux Estimate Uncertainty Range Relative to Flux Estimate
(MMTCOz Eq.)	(MMT C02 Eq.)	(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Forest Ecosystem C Pools3
C02
(583.3)
(651.8)
(514.9)
-11.7%
11.7%
Harvested Wood Products'5
C02
(108.5)
(138.5)
(81.7)
-27.7%
24.7%
Total Forest
C02
(691.8)
(766.6)
(618.3)
-10.8%
10.6%
Note: Parentheses indicate negative values or net uptake. Totals may not sum due to independent rounding
a Range of flux estimates predicted through a combination of sample-based and model-based uncertainty for a 95 percent
confidence interval, IPCC Approach 1.
b Range of flux estimates predicted by Monte Carlo stochastic simulation for a 95 percent confidence interval, IPCC
Approach 2.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
As discussed above, the FIA program has conducted consistent forest surveys based on extensive statistically-
based sampling of most of the forest land in the conterminous United States, dating back to 1952. The FIA program
includes numerous quality assurance and quality control (QA/QC) procedures, including calibration among field
crews, duplicate surveys of some plots, and systematic checking of recorded data. Because of the statistically-
based sampling, the large number of survey plots, and the quality of the data, the survey databases developed by
the FIA program form a strong foundation for C stock estimates. Field sampling protocols, summary data, and
detailed inventory databases are archived and are publicly available on the Internet (USDA Forest Service 2020d).
General quality control procedures were used in performing calculations to estimate C stocks based on survey
data. For example, the C datasets, which include inventory variables such as areas and volumes, were compared to
Land Use, Land-Use Change, and Forestry 6-35

-------
standard inventory summaries such as the forest resource statistics of Oswalt et al. (2019) or selected population
estimates generated from the FIA database, which are available at an FIA internet site (USDA Forest Service
2020b). Agreement between the C datasets and the original inventories is important to verify accuracy of the data
used.
Estimates of the HWP variables and the HWP contribution under the production estimation approach use data
from U.S. Census and USDA Forest Service surveys of production and trade and other sources (Hair and Ulrich
1963; Hair 1958; USDC Bureau of Census 1976; Ulrich 1985,1989; Steer 1948; AF&PA 2006a, 2006b; Howard 2003,
2007, Howard and Jones 2016, Howard and Liang 2019). Factors to convert wood and paper to units of C are based
on estimates by industry and Forest Service published sources (see Annex 3.13). The WOODCARB II model uses
estimation methods suggested by IPCC (2006). Estimates of annual C change in solidwood and paper products in
use were calibrated to meet two independent criteria. The first criterion is that the WOODCARB II model estimate
of C in houses standing in 2001 needs to match an independent estimate of C in housing based on U.S. Census and
USDA Forest Service survey data. Meeting the first criterion resulted in an estimated half-life of about 80 years for
single family housing built in the 1920s, which is confirmed by other U.S. Census data on housing. The second
criterion is that the WOODCARB II model estimate of wood and paper being discarded to SWDS needs to match
EPA estimates of discards used in the Waste sector each year over the period 1990 to 2000 (EPA 2006). These
criteria help reduce uncertainty in estimates of annual change in C in products in use in the United States and, to a
lesser degree, reduce uncertainty in estimates of annual change in C in products made from wood harvested in the
United States. In addition, WOODCARB II landfill decay rates have been validated by ensuring that estimates of CH4
emissions from landfills based on EPA (2006) data are reasonable in comparison to CH4 estimates based on
WOODCARB II landfill decay rates.
Recalculations Discussion
The methods used in the current Inventory to compile estimates for forest ecosystem carbon stocks and stock
changes and HWPs from 1990 through 2019 are consistent with those used in the previous (1990 through 2018)
Inventory. However, population estimates of carbon stocks and stock changes were compiled using NFI data from
each U.S. state and national estimates were compiled by summing over all states. In past Inventories population
estimates were compiled by 4 geographic regions (Figure 6-4) and summed over all regions to compile national
estimates. This improvement in resolution resulted in minor changes in the estimates over the time series due to
rounding. Also, the state-level disaggregation contributed to identifying an error in the compilation of the Alaska
time series data resulting in a one-year misalignment in carbon stock changes for this state in comparison to the
1990 through 2018 Inventory. This error has been corrected resulting in differences in each year of the time series
(i.e., 1990 to 2018), given the one-year misalignment, with substantial differences in major fire years in Alaska.
New NFI data contributed to minor decreases in forest land area and stock changes, particularly in the
Intermountain West region (Table 6-13). Soil carbon stocks decreased in the latest Inventory relative to the
previous Inventory and this change can be attributed to refinements in the Digital General Soil Map of the United
States (STATSG02) dataset where soil orders may have changed in the updated data product (Table 6-13). This
resulted in a structural change in the soil organic carbon estimates for mineral and organic soils across the entire
time series (Table 6-8).
Table 6-13: Recalculations of Forest Area (1,000 ha) and C Stocks in Forest Land Remaining
Forest Land and Harvested Wood Pools (MMT C)
Previous Estimate
Year 2019,
2020 Inventory
Current Estimate
Year 2019,
2021 Inventory
Current Estimate
Year 2020,
2021 Inventory
Forest Area (1000 ha)
Carbon Pools (MMT C)
Forest
279,682
56,051
14,989
3,081
2,777
3,641
279,386
55,774
15,152
3,079
2,825
3,638
279,289
55,933
15,260
3,103
2,852
3,638
Aboveground Biomass
Belowground Biomass
Dead Wood
Litter
6-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Soil (Mineral)
25,638
25,146
25,147
Soil (Organic)
5,926
5,933
5,933
Harvested Wood
2,669
2,669
2,699
Products in Use
1,521
1,521
1,532
SWDS
1,148
1,148
1,167
Total Stock
58,720
58,443
58,632
Note: Totals may not sum due to independent rounding.


Table 6-14: Recalculations of Net C Flux from Forest Ecosystem Pools in Forest Land
Remaining Forest Land and Harvested Wood Pools (MMT C)


Previous Estimate
Current Estimate
Current Estimate

Year 2018,
Year 2018,
Year 2019,
Carbon Pool (MMT C)
2020 Inventory
2021 Inventory
2021 Inventory
Forest
(153.9)
(163.6)
(159.1)
Aboveground Biomass
(105.1)
(109.7)
(107.4)
Belowground Biomass
(24.2)
(24.8)
(24.3)
Dead Wood
(23.6)
(27.7)
(27.1)
Litter
(0.8)
(0.6)
(0.1)
Soil (Mineral)
(0.9)
(1.2)
(0.7)
Soil (Organic)
0.4
0.3
0.3
Drained organic soil
0.2
0.2
0.2
Harvested Wood
(26.9)
(26.9)
(29.6)
Products in Use
(8.6)
(8.6)
(10.7)
SWDS
(18.3)
(18.3)
(18.9)
Total Net Flux
(180.9)
(190.5)
(188.7)
Note: Totals may not sum due to independent rounding.
Planned Improvements
Reliable estimates of forest C stocks and changes across the diverse ecosystems of the United States require a high
level of investment in both annual monitoring and associated analytical techniques. Development of improved
monitoring/reporting techniques is a continuous process that occurs simultaneously with annual Inventory
submissions. Planned improvements can be broadly assigned to the following categories: development of a robust
estimation and reporting system, individual C pool estimation, coordination with other land-use categories, and
annual inventory data incorporation.
While this Inventory submission includes C change by Forest Land Remaining Forest Land and Land Converted to
Forest Land and C stock changes for all IPCC pools in these two categories, there are many improvements that are
still necessary. The estimation approach used for the CONUS in the current Inventory for the forest land category
operates at the state scale, whereas previously the western United States and southeast and southcentral coastal
Alaska operated at a regional scale. While this is an improvement over previous Inventories and led to improved
estimation and separation of land use categories in the current Inventory, research is underway to leverage all FIA
data and auxiliary information (i.e., remotely sensed information) to operate at finer spatial and temporal scales.
As in past submissions, emissions and removals associated with natural (e.g., wildfire, insects, and disease) and
human (e.g., harvesting) disturbances are implicitly included in the report given the design of the annual NFI, but
not explicitly estimated. In addition to integrating auxiliary information into the estimation framework and
leveraging all NFI plot measurements, alternative estimators are also being evaluated which will eliminate latency
in population estimates from the NFI, improve annual estimation and characterization of interannual variability,
facilitate attribution of fluxes to particular activities, and allow for easier harmonization of NFI data with auxiliary
data products. The transparency and repeatability of estimation and reporting systems will be improved through
the dissemination of open source code (e.g., R programming language) in concert with the public availability of the
annual NFI (USDA Forest Service 2020b). Also, several FIA database processes are being institutionalized to
increase efficiency and QA/QC in reporting and further improve transparency, completeness, consistency,
accuracy, and availability of data used in reporting. Finally, a combination of approaches were used to estimate
uncertainty associated with C stock changes in the Forest Land Remaining Forest Land category in this report.
Land Use, Land-Use Change, and Forestry 6-37

-------
There is research underway investigating more robust approaches to total uncertainty (Clough et al. 2016), which
will be considered in future Inventory reports.
The modeling framework used to estimate downed dead wood within the dead wood C pool will be updated
similar to the litter (Domke et al. 2016) and soil C pools (Domke et al. 2017). Finally, components of other pools,
such as C in belowground biomass (Russell et al. 2015) and understory vegetation (Russell et al. 2014; Johnson et
al. 2017), are being explored but may require additional investment in field inventories before improvements can
be realized with the Inventory report.
The foundation of forest C estimation and reporting is the annual NFI. The ongoing annual surveys by the FIA
program are expected to improve the accuracy and precision of forest C estimates as new state surveys become
available (USDA Forest Service 2020b). With the exception of Wyoming and western Oklahoma, all other states in
the CONUS now have sufficient annual NFI data to consistently estimate C stocks and stock changes for the future
using the state-level compilation system. The FIA program continues to install permanent plots in Alaska as part of
the operational NFI and as more plots are added to the NFI they will be used to improve estimates for all managed
forest land in Alaska. The methods used to include all managed forest land in Alaska will be used in the years ahead
for Hawaii and U.S. Territories as forest C data become available (only a small number of plots from Hawaii are
currently available from the annualized sampling design). To that end, research is underway to incorporate all NFI
information (both annual and periodic data) and the dense time series of remotely sensed data in multiple
inferential frameworks for estimating greenhouse gas emissions and removals as well as change detection and
attribution across the entire reporting period and all managed forest land in the United States. Leveraging this
auxiliary information will aid not only the interior Alaska effort but the entire inventory system. In addition to fully
inventorying all managed forest land in the United States, the more intensive sampling of fine woody debris, litter,
and SOC on a subset of FIA plots continues and will substantially improve resolution of C pools (i.e., greater sample
intensity; Westfall et al. 2013) as this information becomes available (Woodall et al. 2011b). Increased sample
intensity of some C pools and using annualized sampling data as it becomes available for those states currently not
reporting are planned for future submissions. The NFI sampling frame extends beyond the forest land use category
(e.g., woodlands, which fall into the grasslands land use category, and urban areas, which fall into the settlements
land use category) with inventory-relevant information for trees outside of forest land. These data will be utilized
as they become available in the NFI.
Non-C02 Emissions from Forest Fires
Emissions of non-C02 gases from forest fires were estimated using U.S.-specific data for annual area of forest
burned, potential fuel availability, and fire severity as well as the default IPCC (2006) emissions and some
combustion factors applied to the IPCC methodology. In 2019, emissions from this source were estimated to be 9.5
MMT C02 Eq. of CH4 and 6.2 MMT C02 Eq. of N20 (Table 6-15; kt units provided in Table 6-16). The estimates of
non-C02 emissions from forest fires include wildfires and prescribed fires in the conterminous 48 states and all
managed forest land in Alaska.
Table 6-15: N011-CO2 Emissions from Forest Fires (MMT CO2 Eq.)a
Gas
1990
2005
2015
2016
2017
2018
2019b
ch4
0.9
5.0
12.2
3.4
11.0
9.5
9.5
n2o
0.6
3.3
8.1
2.2
7.3
6.2
6.2
Total
1.5
8.2
20.3
5.6
18.3
15.7
15.7
Note: Totals may not sum due to independent rounding
a These estimates include Non-C02 Emissions from Forest Fires on Forest Land Remaining
Forest Land and Land Converted to Forest Land.
bSome fire data for 2018 and all of 2019 were unavailable when these estimates were
summarized; therefore 2017 or 2018, the most recent available estimate, is applied to
2019.
Table 6-16: N011-CO2 Emissions from Forest Fires (kt)a
6-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Gas
1990
2005
2015
2016
2017
2018
2019b
ch4
35
198
489
135
440
379
379
n2o
2
11
27
7
24
21
21
CO
800
4511
11136
3080
10036
8626
8626
NOx
22
126
312
87
281
242
242
a These estimates include Non-C02 Emissions from Forest Fires on Forest Land Remaining
Forest Land and Land Converted to Forest Land.
bSome fire data for 2018 and all of 2019 were unavailable when these estimates were
summarized; therefore 2017 or 2018, the most recent available estimate, is applied to
2019.
Methodology
Non-C02 emissions from forest fires—primarily CH4 and N20 emissions—were calculated following IPCC (2006)
methodology, which included a combination of U.S. specific data on area burned, potential fuel available for
combustion, and estimates of combustion based on fire severity along with IPCC default combustion and emission
factors. The estimates were calculated according to Equation 2.27 of IPCC (2006) (Volume 4, Chapter 2), which is:
Emissions = Area burned x Fuel available x Combustion factor x Emission Factor x 10"3
where forest area burned is based on Monitoring Trends in Burn Severity (MTBS, Eidenshink et al. 2007 and 2015)
and National Land Cover (NLCD, Homer et al. 2015) data. Fuel estimates are based on current C density estimates
obtained from FIA plot data, combustion is partly a function of burn severity, and emission factors are from IPCC
(2006, Volume 4, Chapter 2). See Annex 3.13 for further details.
Uncertainty and Time-Series Consistency
In order to quantify the uncertainties for non-C02 emissions from wildfires and prescribed burns, a Monte Carlo
(IPCC Approach 2) sampling approach was employed to propagate uncertainty based on the model and data
applied for U.S. forest land. See IPCC (2006) and Annex 3.13 for the quantities and assumptions employed to
define and propagate uncertainty. The results of the Approach 2 quantitative uncertainty analysis are summarized
in Table 6-17.
Table 6-17: Quantitative Uncertainty Estimates of Non-C02 Emissions from Forest Fires
(MMT CO2 Eq. and Percent)3
Source
Gas
2019 Emission Estimate
(MMTC02 Eq.)
Uncertainty Range Relative to Emission Estimateb
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Non-C02 Emissions from
Forest Fires
ch4
9.5
7.9
11.2
-16%
18%
Non-C02 Emissions from
Forest Fires
n2o
6.2
5.5
7.1
-13%
14%
a These estimates include Non-C02 Emissions from Forest Fires on Forest Land Remaining Forest Land and Land
Converted to Forest Land.
b Range of flux estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
Land Use, Land-Use Change, and Forestry 6-39

-------
QA/QC and Verification
Tier 1 and Tier 2 QA/QC activities were conducted consistent with the U.S. QA/QC plan. Source-specific quality
control measures for estimating non-C02 emissions from forest fires included checking input data, documentation,
and calculations to ensure data were properly handled through the inventory process. The QA/QC procedures did
not reveal any inaccuracies or incorrect input values.
Recalculations Discussion
The methods used in the current (1990 through 2019) Inventory to compile estimates of non-C02 emissions from
forest fires are consistent with those used in the previous 1990 through 2018 Inventory. The recalculations
reflected updates to MTBS and NFI data. These recalculations resulted in a 16 percent decrease in total emissions
for 2018 as compared to the previous Inventory.
Planned Improvements
Continuing improvements are planned for developing better fire and site-specific estimates for forest area burned,
potential fuel available, and combustion. The goal is to develop easy to apply models based on readily available
data to characterize the site and fire for the over twenty thousand fires in the MTBS data. The results will be less
reliant on wide regional values or IPCC defaults. Spatially relating potential fuel availability to more localized forest
structure is the best example of this. An additional future consideration is to apply the forest inventory data to
identify and quantify the likely small additional contribution of fires that are below the minimum size threshold for
the MTBS data.
N20 Emissions from N Additions to Forest Soils
Of the synthetic nitrogen (N) fertilizers applied to soils in the United States, no more than one percent is applied to
forest soils. Application rates are similar to those occurring on cropland soils, but in any given year, only a small
proportion of total forested land receives N fertilizer. This is because forests are typically fertilized only twice
during their approximately 40-year growth cycle (once at planting and once midway through their life cycle). While
the rate of N fertilizer application for the area of forests that receives N fertilizer in any given year is relatively high,
the annual application rate is quite low over the entire area of forest land.
N additions to soils result in direct and indirect N20 emissions. Direct emissions occur on-site due to the N
additions. Indirect emissions result from fertilizer N that is transformed and transported to another location
through volatilization in the form of ammonia [NH3] and nitrogen oxide [NOx], in addition to leaching and runoff of
nitrates [N03], and later converted into N20 at the off-site location. The indirect emissions are assigned to forest
land because the management activity leading to the emissions occurred in forest land.
Direct soil N20 emissions from Forest Land Remaining Forest Land and Land Converted to Forest Land33 in 2019
were 0.3 MMT C02 Eq. (1 kt), and the indirect emissions were 0.1 MMT C02 Eq. (0.4 kt). Total emissions for 2019
were 0.5 MMT C02 Eq. (2 kt) and have increased by 455 percent from 1990 to 2019. Total forest soil N20 emissions
are summarized in Table 6-18.
Table 6-18: N2O Fluxes from Soils in Forest Land Remaining Forest Land and Land Converted
to Forest Land (MMT CO2 Eq. and kt N2O)
1990 2005 2014 2015 2016 2017 2018 2019
Direct N20 Fluxes from Soils
MMT C02 Eq.	0.1	0.3
kt N20	+	1
0.3 0.3 0.3 0.3 0.3 0.3
111111
33 The N20 emissions from Land Converted to Forest Land are included with Forest Land Remaining Forest Land because it is
not currently possible to separate the activity data by land use conversion category.
6-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Indirect N20 Fluxes from Soils
MMTC02 Eq.
kt N20
+
+
O
+ i->
o
+ i->
o
+ i->
o
+ i->
o
+ i->
o
+ i->
o
+ i->
Total








MMT C02 Eq.
0.1
0.5
0.5
0.5
0.5
0.5
0.5
0.5
kt N20
+
2
2
2
2
2
2
2
Note: Totals may not sum due to independent rounding. The N20 emissions from Land Converted to
Forest Land are included with Forest Land Remaining Forest Land because it is not currently possible to
separate the activity data by land use conversion category.
+ Does not exceed 0.05 MMT C02 Eq. or 0.5 kt.
Methodology
The IPCC Tier 1 approach is used to estimate N20 from soils within Forest Land Remaining Forest Land and Land
Converted to Forest Land. According to U.S. Forest Service statistics for 1996 (USDA Forest Service 2001),
approximately 75 percent of trees planted are for timber, and about 60 percent of national total harvested forest
area is in the southeastern United States. Although southeastern pine plantations represent the majority of
fertilized forests in the United States, this Inventory also incorporated N fertilizer application to commercial
Douglas-fir stands in western Oregon and Washington. For the Southeast, estimates of direct N20 emissions from
fertilizer applications to forests are based on the area of pine plantations receiving fertilizer in the southeastern
United States and estimated application rates (Albaugh et al. 2007; Fox et al. 2007). Fertilizer application is rare for
hardwoods and therefore not included in the inventory (Binkley et al. 1995). For each year, the area of pine
receiving N fertilizer is multiplied by the weighted average of the reported range of N fertilization rates (121 lbs. N
per acre). Area data for pine plantations receiving fertilizer in the Southeast are not available for 2005 through
2019, so data from 2004 are used for these years. For commercial forests in Oregon and Washington, only fertilizer
applied to Douglas-fir is addressed in the inventory because the vast majority (approximately 95 percent) of the
total fertilizer applied to forests in this region is applied to Douglas-fir (Briggs 2007). Estimates of total Douglas-fir
area are multiplied by the portion of fertilized area to obtain annual area estimates of fertilized Douglas-fir stands.
Similar to the Southeast, data are not available for 2005 through 2019, so data from 2004 are used for these years.
The annual area estimates are multiplied by the typical rate used in this region (200 lbs. N per acre) to estimate
total N applied (Briggs 2007), and the total N applied to forests is multiplied by the IPCC (2006) default emission
factor of one percent to estimate direct N20 emissions.
For indirect emissions, the volatilization and leaching/runoff N fractions for forest land are calculated using the
IPCC default factors of 10 percent and 30 percent, respectively. The amount of N volatilized is multiplied by the
IPCC default factor of one percent to estimate the amount of volatilized N that is converted to N20 off-site. The
amount of N leached/runoff is multiplied by the IPCC default factor of 0.075 percent to estimate the amount of
leached/runoff N that is converted to N20 off-site. The resulting estimates are summed to obtain total indirect
emissions.
Uncertainty and Time-Series Consistency
The amount of N20 emitted from forests depends not only on N inputs and fertilized area, but also on a large
number of variables, including organic C availability, oxygen gas partial pressure, soil moisture content, pH,
temperature, and tree planting/harvesting cycles. The effect of the combined interaction of these variables on N20
emissions is complex and highly uncertain. IPCC (2006) does not incorporate any of these variables into the default
methodology, except variation in estimated fertilizer application rates and areas of forest land receiving N
fertilizer. All forest soils are treated equivalently under this methodology. Furthermore, only applications of
synthetic N fertilizers to forest are captured in this inventory, so applications of organic N fertilizers are not
estimated. However, the total quantity of organic N inputs to soils in the United States is included in the inventory
for Agricultural Soil Management (Section 5.4) and Settlements Remaining Settlements (Section 6.10).
Land Use, Land-Use Change, and Forestry 6-41

-------
Uncertainties exist in the fertilization rates, annual area of forest lands receiving fertilizer, and the emission
factors. Fertilization rates are assigned a default level34 of uncertainty at ±50 percent, and area receiving fertilizer
is assigned a ±20 percent according to expert knowledge (Binkley 2004). The uncertainty ranges around the 2004
activity data are directly applied to the 2019 emission estimates. IPCC (2006) provided estimates for the
uncertainty associated with direct and indirect N20 emission factor for synthetic N fertilizer application to soils.
Uncertainty is quantified using simple error propagation methods (IPCC 2006). The results of the quantitative
uncertainty analysis are summarized in Table 6-19. Direct N20 fluxes from soils in 2019 are estimated to be
between 0.1 and 1.1 MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 59 percent below and
211 percent above the emission estimate of 0.3 MMT C02 Eq. for 2019. Indirect N20 emissions in 2019 are 0.1
MMT C02 Eq. and have a range are between 0.02 and 0.4 MMT C02 Eq., which is 86 percent below to 238 percent
above the emission estimate for 2019.
Table 6-19: Quantitative Uncertainty Estimates of N2O Fluxes from Soils in Forest Land
Remaining Forest Land and Land Converted to Forest Land (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMTC02 Eq.)
Uncertainty Range Relative to Emission Estimate
(MMT C02 Eq.) (%)
Forest Land Remaining Forest


Lower
Upper
Lower
Upper
Land


Bound
Bound
Bound
Bound
Direct N20 Fluxes from Soils
N20
0.3
0.1
1.1
-59%
+211%
Indirect N20 Fluxes from Soils
n2o
0.1
+
0.4
-86%
+238%
Note: Totals may not sum due to independent rounding
+ Does not exceed 0.05 MMT C02 Eq.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
QA/QC and Verification
The spreadsheet containing fertilizer applied to forests and calculations for N20 and uncertainty ranges are
checked and verified based on the sources of these data.
Recalculations Discussion
No recalculations were performed for the 1990 to 2018 estimates.
C02, CH4, and N20 Emissions from Drained Organic Soils35
Drained organic soils on forest land are identified separately from other forest soils largely because mineralization
of the exposed or partially dried organic material results in continuous C02 and N20 emissions (IPCC 2006). In
addition, the 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands
(IPCC 2014) calls for estimating CH4 emissions from these drained organic soils and the ditch networks used to
drain them.
Organic soils are identified on the basis of thickness of organic horizon and percent organic matter. All organic soils
are assumed to have originally been wet, and drained organic soils are further characterized by drainage or the
34	Uncertainty is unknown for the fertilization rates so a conservative value of ±50 percent is used in the analysis.
35	Estimates of C and C02 emissions from drained organic soils are described in this section but reported in Table 6-8 and Table
6-9 for both Forest Land Remaining Forest Land and Land Converted to Forest Land in order to allow for reporting of all C stock
changes on forest lands in a complete and comprehensive manner.
6-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
process of artificially lowering the soil water table, which exposes the organic material to drying and the associated
emissions described in this section. The land base considered here is drained inland organic soils that are
coincident with forest area as identified by the NFI of the USDA Forest Service (USDA Forest Service 2020b).
The estimated area of drained organic soils on forest land is 70,849 ha and did not change over the time series
based on the data used to compile the estimates in the current Inventory. These estimates are based on
permanent plot locations of the NFI (USDA Forest Service 2020b) coincident with mapped organic soil locations
(STATSG02 2016), which identifies forest land on organic soils. Forest sites that are drained are not explicitly
identified in the data, but for this estimate, planted forest stands on sites identified as mesic or xeric (which are
identified in USDA Forest Service 2020c, d) are labeled "drained organic soil" sites.
Land use, region, and climate are broad determinants of emissions as are more site-specific factors such as
nutrient status, drainage level, exposure, or disturbance. Current data are limited in spatial precision and thus lack
site specific details. At the same time, corresponding emissions factor data specific to U.S. forests are similarly
lacking. Tier 1 estimates are provided here following IPCC (2014). Total annual non-C02 emissions on forest land
with drained organic soils in 2019 are estimated as 0.1 MMT C02 Eq. per year (Table 6-20; kt units provided in
Table 6-21).
The Tier 1 methodology provides methods to estimate C emission as C02 from three pathways: direct emissions
primarily from mineralization; indirect, or off-site, emissions associated with dissolved organic carbon releasing
C02 from drainage waters; and emissions from (peat) fires on organic soils. Data about forest fires specifically
located on drained organic soils are not currently available; as a result, no corresponding estimate is provided
here. Non-C02 emissions provided here include CH4 and N20. Methane emissions generally associated with anoxic
conditions do occur from the drained land surface but the majority of these emissions originate from ditches
constructed to facilitate drainage at these sites. Emission of N20 can be significant from these drained organic soils
in contrast to the very low emissions from wet organic soils.
Table 6-20: N011-CO2 Emissions from Drained Organic Forest Soilsa'b (MMT CO2 Eq.)
Source
1990
2005
2015
2016
2017
2018
2019
ch4
+
+
+
+
+
+
+
n2o
0.1
0.1
0.1
0.1
0. 1
0.1
0.1
Total
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a This table includes estimates from Forest Land Remaining Forest Land and Land Converted to
Forest Land.
b Estimates of C and C02 emissions from drained organic soils are described in this section but
reported in Table 6-8 and Table 6-9 for both Forest Land Remaining Forest Land and Land
Converted to Forest Land in order to allow for reporting of all C stock changes on forest lands in
a complete and comprehensive manner.
Table 6-21: Non-C02 Emissions from Drained Organic Forest Soilsa'b(kt)
Source
1990
2005
2015
2016
2017
2018
2019
ch4
0.6
0.6
0.6
0.6
0.6
0.6
0.6
n2o
+
+
+
+
+
+
+
+ Does not exceed 0.5 kt.
a This table includes estimates from Forest Land Remaining Forest Land and Land Converted to
Forest Land.
b Estimates of C and C02 emissions from drained organic soils are described in this section but
reported in Table 6-8 and Table 6-9 for both Forest Land Remaining Forest Land and Land
Converted to Forest Land in order to allow for reporting of all C stock changes on forest lands in a
complete and comprehensive manner.
Land Use, Land-Use Change, and Forestry 6-43

-------
Methodology
The Tier 1 methods for estimating C02, CH4 and N20 emissions from drained inland organic soils on forest lands
follow IPCC (2006), with extensive updates and additional material presented in the 2013 Supplement to the 2006
IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands (IPCC 2014). With the exception of quantifying
area of forest on drained organic soils, which is user-supplied, all quantities necessary for Tier 1 estimates are
provided in Chapter 2, Drained Inland Organic Soils of IPCC (2014).
Estimated area of drained organic soils on forest land is 70,849 ha based on analysis of the permanent NFI of the
USDA Forest Service and did not change over the time series. The most recent plot data per state within the
inventories were used in a spatial overlay with the STATSG02 (2016) soils data, and forest plots coincident with the
soil order histosol were selected as having organic soils. Information specific to identifying "drained organic" are
not in the inventory data so an indirect approach was employed here. Specifically, artificially regenerated forest
stands (inventory field STDORGCD=l) on mesic or xeric sites (inventory field 11
-------
these drained soils in contrast to the very low emissions from wet organic soils. Calculations are according to
Equation 2.7 and Table 2.5, which provide the estimate as kg N per year.
Uncertainty and Time-Series Consistency
Uncertainties are based on the sampling error associated with forest area of drained organic soils and the
uncertainties provided in the Chapter 2 (IPCC 2014) emissions factors (Table 6-23). The estimates and resulting
quantities representing uncertainty are based on the IPCC Approach 1-error propagation. However, probabilistic
sampling of the distributions defined for each emission factor produced a histogram result that contained a mean
and 95 percent confidence interval. The primary reason for this approach was to develop a numerical
representation of uncertainty with the potential for combining with other forest components. The methods and
parameters applied here are identical to previous inventories, but input values were resampled for this inventory,
which results in minor changes in the less significant digits in the resulting estimates, relative to past values. The
total non-C02 emissions in 2019 from drained organic soils on Forest Land Remaining Forest Land and Land
Converted to Forest Land were estimated to be between 0.004 and 0.236 MMT C02 Eq. around a central estimate
of 0.106 MMT C02 Eq. at a 95 percent confidence level.
Table 6-23: Quantitative Uncertainty Estimates for Non-C02 Emissions on Drained Organic
Forest Soils (MMT CO2 Eq. and Percent)3
2019 Emission
Source Estimate Uncertainty Range Relative to Emission Estimate
	(MMTCOz Eq.)	(MMTCOz Eq.)	(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
ch4
+
+
+
-70%
80%
n2o
0.1
+
0.2
-100%
128%
Total
0.1
+
0.2
-96%
122%
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Range of flux estimates predicted through a combination of sample-based and IPCC defaults for a 95
percent confidence interval, IPCC Approach 1.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
QA/QC and Verification
IPCC (2014) guidance cautions of a possibility of double counting some of these emissions. Specifically, the off-site
emissions of dissolved organic C from drainage waters may be double counted if soil C stock and change is based
on sampling and this C is captured in that sampling. Double counting in this case is unlikely since plots identified as
drained were treated separately in this chapter. Additionally, some of the non-C02 emissions may be included in
either the Wetlands or sections on N20 emissions from managed soils. These paths to double counting emissions
are unlikely here because these issues are taken into consideration when developing the estimates and this
chapter is the only section directly including such emissions on forest land.
Recalculations Discussion
No recalculations were performed for the 1990 to 2018 estimates.
Planned Improvements
Additional data will be compiled to update estimates of forest areas on drained organic soils as new reports are
made available and new geospatial products become available.
Land Use, Land-Use Change, and Forestry 6-45

-------
6.3 Land Converted to Forest Land (CRF Source
Category 4A2)
The C stock change estimates for Land Converted to Forest Land that are provided in this Inventory include all
forest land in an inventory year that had been in another land use(s) during the previous 20 years.36 For example,
cropland or grassland converted to forest land during the past 20 years would be reported in this category.
Converted lands are in this category for 20 years as recommended in the 2006IPCC Guidelines (IPCC 2006), after
which they are classified as Forest Land Remaining Forest Land. Estimates of C stock changes from all pools (i.e.,
aboveground and belowground biomass, dead wood, litter and soils), as recommended by IPCC (2006), are
included in the Land Converted to Forest Land category of this Inventory.
Area of Land Converted to Forest in the United States37
Land conversion to and from forests has occurred regularly throughout U.S. history. The 1970s and 1980s saw a
resurgence of federally-sponsored forest management programs (e.g., the Forestry Incentive Program) and soil
conservation programs (e.g., the Conservation Reserve Program), which have focused on tree planting, improving
timber management activities, combating soil erosion, and converting marginal cropland to forests. Recent
analyses suggest that net accumulation of forest area continues in areas of the United States, in particular the
northeastern United States (Woodall et al. 2015b). Specifically, the annual conversion of land from other land-use
categories (i.e., Cropland, Grassland, Wetlands, Settlements, and Other Lands) to Forest Land resulted in a fairly
continuous net annual accretion of Forest Land area from over the time series at an average rate of 1.0 million ha
year1.
Over the 20-year conversion period used in the Land Converted to Forest Land category, the conversion of
cropland to forest land resulted in the largest source of C transfer and uptake, accounting for approximately 40
percent of the uptake annually. Estimated C uptake has remained relatively stable over the time series across all
conversion categories (see Table 6-24). The net flux of C from all forest pool stock changes in 2019 was -99.1 MMT
C02 Eq. (-27.0 MMT C) (Table 6-24 and Table 6-25).
Mineral soil C stocks increase slightly over the time series for Land Converted to Forest Land. The small gains are
associated with Cropland Converted to Forest Land, Settlements Converted to Forest Land, and Other Land
Converted to Forest Land. Much of this conversion is from soils that are more intensively used under annual crop
production or settlement management, or are conversions from other land, which has little to no soil C. In
contrast, Grassland Converted to Forest Land leads to a loss of soil C across the time series, which negates some of
the gain in soil C with the other land use conversions. Managed pasture to Forest Land is the most common
conversion. This conversion leads to a loss of soil C because pastures are mostly improved in the United States with
fertilization and/or irrigation, which enhances C input to soils relative to typical forest management activities.
36	The annual NFI data used to compile estimates of carbon transfer and uptake in this section are based on 5- to 10-yr
remeasurements so the exact conversion period was limited to the remeasured data over the time series.
37	The estimates reported in this section only include the 48 conterminous states in the United States. Land use conversion to
forest in Alaska and Hawaii were not included. Since area estimates for some land use categories were not updated in the Land
Representation in the current Inventory there are differences in the area estimates reported in this section and those reported
in Section 6.1 Representation of the U.S. Land Base. See Annex 3.13, Table A-214 for annual differences between the forest
area reported in Section 6 Representation of the U.S. Land Base and Section 6.3 Land Converted to Forest Land.
6-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 6-24: Net CO2 Flux from Forest C Pools in Land Converted to Forest Land by Land Use
Change Category (MMT CO2 Eq.)
Land Use/Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Cropland Converted to Forest Land
(39.8)
(39.7)
(39.7)
(39.8)
(39.8)
(39.8)
(39.8)
Aboveground Biomass
(23.0)
(23.0)
(23.0)
(23.0)
(23.0)
(23.0)
(23.0)
Belowground Biomass
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
Dead Wood
(5.0)
(5.0)
(5.1)
(5.1)
(5.1)
(5.1)
(5.1)
Litter
(7.0)
(7.0)
(7.0)
(7.0)
(7.0)
(7.0)
(7.0)
Mineral Soil
(0.3)
(0.3)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Grassland Converted to Forest Land
(10.3)
(10.3)
(10.4)
(10.4)
(10.5)
(10.5)
(10.5)
Aboveground Biomass
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
Belowground Biomass
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
Dead Wood
(0.9)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
Litter
(3.8)
(3.9)
(3.9)
(4.0)
(4.0)
(4.0)
(4.0)
Mineral Soil
0.2
0.3
0.3
0.3
0.3
0.3
0.3
Other Land Converted to Forest Land
(10.2)
(10.8)
(11.0)
(11.0)
(11.0)
(11.0)
(11.0)
Aboveground Biomass
(4.7)
(4.7)
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
Belowground Biomass
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
Dead Wood
(1.4)
(1.4)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
Litter
(2.6)
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
Mineral Soil
(0.6)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Settlements Converted to Forest Land
(34.6)
(34.5)
(34.5)
(34.5)
(34.5)
(34.5)
(34.5)
Aboveground Biomass
(21.1)
(21.1)
(21.1)
(21.1)
(21.1)
(21.1)
(21.1)
Belowground Biomass
(4.1)
(4.0)
(4.0)
(4.0)
(4.0)
(4.0)
(4.0)
Dead Wood
(3.9)
(3.9)
(3.9)
(3.9)
(4.0)
(4.0)
(4.0)
Litter
(5.4)
(5.4)
(5.4)
(5.4)
(5.4)
(5.4)
(5.4)
Mineral Soil
(0.1)
+
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Wetlands Converted to Forest Land
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
Aboveground Biomass
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
Belowground Biomass
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Dead Wood
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
Litter
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
Mineral Soil
+
+
+
+
+
+
+
Total Aboveground Biomass Flux
(54.9)
(54.9)
(55.0)
(55.0)
(55.1)
(55.1)
(55.1)
Total Belowground Biomass Flux
(10.6)
(10.7)
(10.7)
(10.7)
(10.7)
(10.7)
(10.7)
Total Dead Wood Flux
(11.7)
(11.7)
(11.8)
(11.8)
(11.8)
(11.8)
(11.8)
Total Litter Flux
(20.1)
(20.2)
(20.3)
(20.3)
(20.4)
(20.4)
(20.4)
Total Mineral Soil Flux
(0.8)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Total Flux
(98.2)
(98.7)
(98.9)
(99.0)
(99.1)
(99.1)
(99.1)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake. Forest ecosystem C stock changes
from land conversion in Alaska are currently included in the Forest Land Remaining Forest Land section because there is
insufficient data to separate the changes at this time. Forest ecosystem C stock changes from land conversion do not include U.S.
Territories because managed forest land in U.S. Territories is not currently included in Section 6 Representation of the U.S. Land
Base. The forest ecosystem C stock changes from land conversion do not include Hawaii because there is insufficient NFI data to
support inclusion at this time. See Annex 3.13, Table A-218 for annual differences between the forest area reported in Section 6
Representation of the U.S. Land Base and Section 6.3 Land Converted to Forest Land. Since area estimates for some land use
categories were not updated in the Land Representation in the current Inventory there are differences in the area estimates
reported in this section and those reported in Section 6 Representation of the U.S. Land Base. The forest ecosystem C stock
changes from land conversion do not include trees on non-forest land (e.g., agroforestry systems and settlement areas—see
Section 6.10 Settlements Remaining Settlements for estimates of C stock change from settlement trees). It is not possible to
separate emissions from drained organic soils between Forest Land Remaining Forest Land and Land Converted to Forest Land so
estimates for all organic soils are included in Table 6-8 and Table 6-9 of the Forest Land Remaining Forest Land section of the
Inventory.
Land Use, Land-Use Change, and Forestry 6-47

-------
Table 6-25: Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use
Change Category (MMT C)
Land Use/Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Cropland Converted to Forest Land
(10.9)
(10.8)
(10.8)
(10.8)
(10.9)
(10.9)
(10.9)
Aboveground Biomass
(6.3)
(6.3)
(6.3)
(6.3)
(6.3)
(6.3)
(6.3)
Belowground Biomass
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
Dead Wood
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
Litter
(1.9)
(1.9)
(1.9)
(1.9)
(1.9)
(1.9)
(1.9)
Mineral Soil
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Grassland Converted to Forest Land
(2.8)
(2.8)
(2.8)
(2.8)
(2.9)
(2.9)
(2.9)
Aboveground Biomass
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
Belowground Biomass
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Dead Wood
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Litter
(1.0)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Mineral Soil
+
0.1
0.1
0.1
0.1
0.1
0.1
Other Land Converted to Forest Land
(2.8)
(3.0)
(3.0)
(3.0)
(3.0)
(3.0)
(3.0)
Aboveground Biomass
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
Belowground Biomass
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Dead Wood
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
Litter
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
Mineral Soil
(0.2)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Settlements Converted to Forest Land
(9.4)
(9.4)
(9.4)
(9.4)
(9.4)
(9.4)
(9.4)
Aboveground Biomass
(5.8)
(5.8)
(5.7)
(5.7)
(5.7)
(5.7)
(5.7)
Belowground Biomass
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Dead Wood
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Litter
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
Mineral Soil
+
+
+
+
+
+
+
Wetlands Converted to Forest Land
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
Aboveground Biomass
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
Belowground Biomass
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Dead Wood
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Litter
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Mineral Soil
+
+
+
+
+
+
+
Total Aboveground Biomass Flux
(15.0)
(15.0)
(15.0)
(15.0)
(15.0)
(15.0)
(15.0)
Total Belowground Biomass Flux
(2.9)
(2.9)
(2.9)
(2.9)
(2.9)
(2.9)
(2.9)
Total Dead Wood Flux
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
Total Litter Flux
(5.5)
(5.5)
(5.5)
(5.5)
(5.6)
(5.6)
(5.6)
Total Mineral Soil Flux
(0.2)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Total Flux
(26.8)
(26.9)
(27.0)
(27.0)
(27.0)
(27.0)
(27.0)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake. Forest ecosystem C stock
changes from land conversion in Alaska are currently included in the Forest Land Remaining Forest Land section
because there is not sufficient data to separate the changes at this time. Forest ecosystem C stock changes from
land conversion do not include U.S. Territories because managed forest land in U.S. Territories is not currently
included in Section 6 Representation of the U.S. Land Base. The forest ecosystem C stock changes from land
conversion do not include Hawaii because there is not sufficient NFI data to support inclusion at this time. See
Annex 3.13, Table A-218 for annual differences between the forest area reported in Section 6 Representation of the
U.S. Land Base and Section 6.3 Land Converted to Forest Land. Since area estimates for some land use categories
were not updated in the Land Representation in the current Inventory there are differences in the area estimates
reported in this section and those reported in Section 6.1 Representation of the U.S. Land Base The forest
ecosystem C stock changes from land conversion do not include trees on non-forest land (e.g., agroforestry systems
and settlement areas—see Section 6.10 Settlements Remaining Settlements for estimates of C stock change from
settlement trees). It is not possible to separate emissions from drained organic soils between Forest Land Remaining
Forest Land and Land Converted to Forest Land so estimates for organic soils are included in Table 6-8 and Table 6-9
of the Forest Land Remaining Forest Land section of the Inventory.
+ Absolute value does not exceed 0.05 MMT C.
6-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
The following section includes a description of the methodology used to estimate stock changes in all forest C
pools for Land Converted to Forest Land. National Forest Inventory data and IPCC (2006) defaults for reference C
stocks were used to compile separate estimates for the five C storage pools. Estimates for Aboveground and
Belowground Biomass, Dead Wood and Litter were based on data collected from the extensive array of
permanent, annual NFI plots and associated models (e.g., live tree belowground biomass estimates) in the United
States (USDA Forest Service 2020b, 2020c). Carbon conversion factors were applied at the individual plot and then
appropriately expanded to population estimates. To ensure consistency in the Land Converted to Forest Land
category where C stock transfers occur between land-use categories, all soil estimates are based on methods from
Ogle et al. (2003, 2006) and IPCC (2006).
The methods used for estimating carbon stocks and stock changes in the Land Converted to Forest Land are
consistent with those used for Forest Land Remaining Forest Land. For land use conversion, IPCC (2006) default
biomass C stocks removed due to land use conversion from Croplands and Grasslands were used in the year of
conversion on individual plots. All annual NFI plots available through May 2019 were used in this Inventory. Forest
Land conditions were observed on NFI plots at time t0 and at a subsequent time ti=t0+s, where s is the time step
(time measured in years) and is indexed by discrete (e.g., 5 year) forest age classes. The inventory from t0 was then
projected from ti to 2019. This projection approach requires simulating changes in the age-class distribution
resulting from forest aging and disturbance events and then applying C density estimates for each age class to
obtain population estimates for the nation.
Carbon in Biomass
Live tree C pools include aboveground and belowground (coarse root) biomass of live trees with diameter at breast
height (dbh) of at least 2.54 cm at 1.37 m above the forest floor. Separate estimates were made for above and
belowground biomass components. If inventory plots included data on individual trees, above- and belowground
tree C was based on Woodall et al. (2011a), which is also known as the component ratio method (CRM), and is a
function of volume, species, and diameter. An additional component of foliage, which was not explicitly included in
Woodall et al. (2011a), was added to each tree following the same CRM method.
Understory vegetation is a minor component of biomass and is defined as all biomass of undergrowth plants in a
forest, including woody shrubs and trees less than 2.54 cm dbh. For the current Inventory, it was assumed that 10
percent of total understory C mass is belowground (Smith et al. 2006). Estimates of C density were based on
information in Birdsey (1996) and biomass estimates from Jenkins et al. (2003). Understory biomass represented
over one percent of C in biomass, but its contribution rarely exceeded 2 percent of the total.
Biomass losses associated with conversion from Grassland and Cropland to Forest Land were assumed to occur in
the year of conversion. To account for these losses, IPCC (2006) defaults for aboveground and belowground
biomass on Grasslands and aboveground biomass on Croplands were subtracted from sequestration in the year of
the conversion. For all other land use (i.e., Other Lands, Settlements, Wetlands) conversions to Forest Land no
biomass loss data were available and no IPCC (2006) defaults currently exist to include transfers, losses, or gains of
carbon in the year of the conversion so none were incorporated for these conversion categories. As defaults or
country-specific data become available for these conversion categories they will be incorporated.
Carbon in Dead Organic Matter
Dead organic matter was initially calculated as three separate pools—standing dead trees, downed dead wood,
and litter—with C stocks estimated from sample data or from models. The standing dead tree C pool includes
aboveground and belowground (coarse root) biomass for trees of at least 12.7 cm dbh. Calculations followed the
basic method applied to live trees (Woodall et al. 2011a) with additional modifications to account for decay and
structural loss (Domke et al. 2011; Harmon et al. 2011). Downed dead wood estimates are based on measurement
of a subset of FIA plots for downed dead wood (Domke et al. 2013; Woodall and Monleon 2008; Woodall et al.
2013). Downed dead wood is defined as pieces of dead wood greater than 7.5 cm diameter, at transect
Land Use, Land-Use Change, and Forestry 6-49

-------
intersection, that are not attached to live or standing dead trees. This includes stumps and roots of harvested
trees. To facilitate the downscaling of downed dead wood C estimates from the state-wide population estimates to
individual plots, downed dead wood models specific to regions and forest types within each region are used. Litter
C is the pool of organic C (also known as duff, humus, and fine woody debris) above the mineral soil and includes
woody fragments with diameters of up to 7.5 cm. A subset of FIA plots are measured for litter C. A modeling
approach, using litter C measurements from FIA plots (Domke et al. 2016) was used to estimate litter C for every
FIA plot used in the estimation framework.
Mineral Soil Carbon Stock Changes
A Tier 2 method is applied to estimate mineral soil C stock changes for Land Converted to Forest Land (Ogle et al.
2003, 2006; IPCC 2006). For this method, land is stratified by climate, soil types, land use, and land management
activity, and then assigned reference carbon levels and factors for the forest land and the previous land use. The
difference between the stocks is reported as the stock change under the assumption that the change occurs over
20 years. Reference C stocks have been estimated from data in the National Soil Survey Characterization Database
(USDA-NRCS 1997), and U.S.-specific stock change factors have been derived from published literature (Ogle et al.
2003, 2006). Land use and land use change patterns are determined from a combination of the Forest Inventory
and Analysis Dataset (FIA), the 2015 National Resources Inventory (NRI) (USDA-NRCS 2018), and National Land
Cover Dataset (NLCD) (Yang et al. 2018). See Annex 3.12 (Methodology for Estimating N20 Emissions, CH4
Emissions and Soil Organic C Stock Changes from Agricultural Soil Management) for more information about this
method. Note that soil C in this Inventory is reported to a depth of 100 cm in the Forest Land Remaining Forest
Land category (Domke et al. 2017) while other land-use categories report soil C to a depth of 30 cm. However, to
ensure consistency in the Land Converted to Forest Land category where C stock transfers occur between land-use
categories, soil C estimates were based on a 30 cm depth using methods from Ogle et al. (2003, 2006) and IPCC
(2006), as described in Annex 3.12. For consistency, the same methods are also used for land use conversions to
Cropland, Grasslands and Settlements in this Inventory.
A quantitative uncertainty analysis placed bounds on the flux estimates for Land Converted to Forest Land through
a combination of sample-based and model-based approaches to uncertainty for forest ecosystem C02 Eq. flux
(IPCC Approach 1). Uncertainty estimates for forest pool C stock changes were developed using the same
methodologies as described in the Forest Land Remaining Forest Land section for aboveground and belowground
biomass, dead wood, and litter. The exception was when IPCC default estimates were used for reference C stocks
in certain conversion categories (i.e., Cropland Converted to Forest Land and Grassland Converted to Forest Land).
In those cases, the uncertainties associated with the IPCC (2006) defaults were included in the uncertainty
calculations. IPCC Approach 2 was used for mineral soils and is described in the Cropland Remaining Cropland
section.
Uncertainty estimates are presented in Table 6-26 for each land conversion category and C pool. Uncertainty
estimates were obtained using a combination of sample-based and model-based approaches for all non-soil C
pools (IPCC Approach 1) and a Monte Carlo approach (IPCC Approach 2) was used for mineral soil. Uncertainty
estimates were combined using the error propagation model (IPCC Approach 1). The combined uncertainty for all
C stocks in Land Converted to Forest Land ranged from 11 percent below to 11 percent above the 2019 C stock
change estimate of-99.1 MMT C02 Eq.
Table 6-26: Quantitative Uncertainty Estimates for Forest C Pool Stock Changes (MMT CO2
Eq. per Year) in 2019 from Land Converted to Forest Land by Land Use Change
Uncertainty and Time-Series Consistency
Land Use/Carbon Pool
2019 Flux
Estimate
Uncertainty Range Relative to Flux Range1
(MMT CP2 Eq.)
(MMTC02 Eq.)
(%)
6-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Cropland Converted to Forest Land
(39.8)
(48.6)
(31.0)
-22%
22%
Aboveground Biomass
(23.0)
(31.6)
(14.4)
-38%
37%
Belowground Biomass
(4.5)
(5.6)
(3.4)
-24%
24%
Dead Wood
(5.1)
(6.3)
(3.8)
-24%
24%
Litter
(7.0)
(8.1)
(6.0)
-15%
15%
Mineral Soils
(0.2)
(0.5)
0.1
-134%
134%
Grassland Converted to Forest Land
(10.5)
(12.9)
(8.1)
23%
23%
Aboveground Biomass
(4.8)
(6.2)
(3.4)
-29%
29%
Belowground Biomass
(1.0)
(1.3)
(0.7)
-29%
29%
Dead Wood
(1.0)
(1.1)
(0.8)
-16%
16%
Litter
(4.0)
(4.5)
(3.4)
-14%
14%
Mineral Soils
0.3
(0.1)
0.6
-135%
135%
Other Lands Converted to Forest Land
(11.0)
(13.4)
(8.7)
-21%
21%
Aboveground Biomass
(4.8)
(6.9)
(2.7)
-44%
44%
Belowground Biomass
(0.9)
(1.3)
(0.5)
-47%
47%
Dead Wood
(1.5)
(2.0)
(0.9)
-38%
38%
Litter
(2.7)
(3.4)
(2.1)
-23%
23%
Mineral Soils
(1.1)
(1.9)
(0.4)
-64%
64%
Settlements Converted to Forest Land
(34.5)
(41.0)
(28.0)
-19%
19%
Aboveground Biomass
(21.1)
(27.3)
(14.9)
-29%
29%
Belowground Biomass
(4.0)
(5.4)
(2.7)
-33%
33%
Dead Wood
(4.0)
(5.1)
(2.8)
-29%
29%
Litter
(5.4)
(6.3)
(4.5)
-17%
17%
Mineral Soils
(0.1)
(0.1)
(0.0)
-40%
40%
Wetlands Converted to Forest Land
(3.2)
(3.4)
(3.1)
-5%
5%
Aboveground Biomass
(1.4)
(1.5)
(1.2)
-10%
10%
Belowground Biomass
(0.3)
(0.3)
(0.2)
-12%
12%
Dead Wood
(0.4)
(0.4)
(0.3)
-11%
11%
Litter
(1.2)
(1.3)
(1.2)
-5%
5%
Mineral Soils
+
+
+
NA
NA
Total: Aboveground Biomass
(55.1)
(66.0)
(44.2)
-20%
20%
Total: Belowground Biomass
(10.7)
(12.5)
(8.9)
-17%
17%
Total: Dead Wood
(11.8)
(13.6)
(10.0)
-15%
15%
Total: Litter
(20.4)
(22.0)
(18.8)
-8%
8%
Total: Mineral Soils
(1.1)
(1.7)
(0.6)
-49%
49%
Total: Lands Converted to Forest Lands
(99.1)
(110.4)
(87.8)
-11%
11%
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake. It is not possible to separate
emissions from drained organic soils between Forest Land Remaining Forest Land and Land Converted to Forest Land so
estimates for organic soils are included in Table 6-8 and Table 6-9 of the Forest Land Remaining Forest Land section of the
Inventory.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
NA (Not Applicable)
a Range of flux estimate for 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
See QA/QC and Verification sections under Forest Land Remaining Forest Land and for mineral soil estimates
Cropland Remaining Cropland.
Land Use, Land-Use Change, and Forestry 6-51

-------
Recalculations Discussion
The approach for estimating carbon stock changes in Land Converted to Forest Land is consistent with the methods
used for Forest Land Remaining Forest Land and is described in Annex 3.13. The Land Converted to Forest Land
estimates in this Inventory are based on the land use change information in the annual NFI. All conversions are
based on empirical estimates compiled using plot remeasurements from the NFI, IPCC (2006) default biomass C
stocks removed from Croplands and Grasslands in the year of conversion on individual plots and the Tier 2 method
for estimating mineral soil C stock changes (Ogle et al. 2003, 2006; IPCC 2006). All annual NFI plots available
through August 2020 were used in this Inventory. This is the second year that remeasurement data from the
annual NFI were available throughout the CONUS (with the exception of Wyoming and western Oklahoma) to
estimate land use conversion. The availability of remeasurement data from the annual NFI allowed for consistent
plot-level estimation of C stocks and stock changes for Forest Land Remaining Forest Land and the Land Converted
to Forest Land categories. Estimates in the previous Inventory were based on state-level carbon density estimates
and a combination of NRI data and NFI data in the eastern United States. The refined analysis in this Inventory
resulted in changes in the Land Converted to Forest Land categories. Overall, the Land Converted to Forest Land C
stock changes decreased by 10 percent in 2018 between the previous Inventory and the current Inventory (Table
6-27). This decrease is directly attributed to the incorporation of annual NFI data into the compilation system. In
the previous Inventory, Grasslands Converted to Forest Land represented the largest transfer and uptake of C
across the land use conversion categories. In this Inventory, Cropland Converted to Forest Land represented the
largest transfer and uptake of C across the land use change categories followed by Settlements Converted to Forest
Land (Table 6-27).
Table 6-27: Recalculations of the Net C Flux from Forest C Pools in Land Converted to Forest
Land by Land Use Change Category (MMT C)
Conversion category
and Carbon pool (MMT C)
2018 Estimate,
Previous Inventory
2018 Estimate,
Current Inventory
2019 Estimate,
Current Inventory
Cropland Converted to Forest Land
(12.6)
(10.9)
(10.9)
Aboveground Biomass
(7.2)
(6.3)
(6.3)
Belowground Biomass
(1.4)
(1.2)
(1.2)
Dead Wood
(1.6)
(1.4)
(1.4)
Litter
(2.3)
(1.9)
(1.9)
Mineral soil
(0.1)
(0.1)
(0.1)
Grassland Converted to Forest Land
(2.6)
(2.9)
(2.9)
Aboveground Biomass
(1.2)
(1.3)
(1.3)
Belowground Biomass
(0.3)
(0.3)
(0.3)
Dead Wood
(0.2)
(0.3)
(0.3)
Litter
(1.0)
(1.1)
(1.1)
Mineral soil
0.1
0.1
0.1
Other Land Converted to Forest Land
(4.1)
(3.0)
(3.0)
Aboveground Biomass
(1.7)
(1.3)
(1.3)
Belowground Biomass
(0.3)
(0.2)
(0.2)
Dead Wood
(0.5)
(0.4)
(0.4)
Litter
(1.1)
(0.7)
(0.7)
Mineral soil
(0.3)
(0.3)
(0.3)
Settlements Converted to Forest Land
(10.6)
(9.4)
(9.4)
Aboveground Biomass
(6.4)
(5.7)
(5.7)
Belowground Biomass
(1.2)
(1.1)
(1.1)
Dead Wood
(1.2)
(1.1)
(1.1)
Litter
(1.7)
(1.5)
(1.5)
Mineral soil
+
+
+
Wetlands Converted to Forest Land
(0.2)
(0.9)
(0.9)
Aboveground Biomass
(0.1)
(0.4)
(0.4)
Belowground Biomass
(0.0)
(0.1)
(0.1)
Dead Wood
(0.0)
(0.1)
(0.1)
Litter
(0.1)
(0.3)
(0.3)
Mineral soil
+
+
+
6-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Total Aboveground Biomass Flux
(16.6)
(15.0)
(15.0)
Total Belowground Biomass Flux
(3.2)
(2.9)
(2.9)
Total Dead Wood Flux
(3.7)
(3.2)
(3.2)
Total Litter Flux
(6.3)
(5.6)
(5.6)
Total SOC (mineral) Flux
(0.3)
(0.3)
(0.3)
Total Flux
(30.2)
(27.0)
(27.0)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake.
+ Absolute value does not exceed 0.05 MMT C.
Planned Improvements
There are many improvements necessary to improve the estimation of carbons stock changes associated with land
use conversion to forest land over the entire time series. First, soil C has historically been reported to a depth of
100 cm in the Forest Land Remaining Forest Land category (Domke et al. 2017) while other land-use categories
(e.g., Grasslands and Croplands) report soil carbon to a depth of 30 cm. To ensure greater consistency in the Land
Converted to Forest Land category where C stock transfers occur between land-use categories, all mineral soil
estimates in the Land Converted to Forest Land category in this Inventory are based on methods from Ogle et al.
(2003, 2006) and IPCC (2006). Methods have recently been developed (Domke et al. 2017) to estimate soil C to
depths of 20, 30, and 100 cm in the Forest Land category using in situ measurements from the Forest Inventory
and Analysis program within the USDA Forest Service and the International Soil Carbon Network. In subsequent
Inventories, a common reporting depth will be defined for all land use conversion categories and Domke et al.
(2017) will be used in the Forest Land Remaining Forest Land and Land Converted to Forest Land categories to
ensure consistent reporting across all forest land. Third, due to the 5 to 10-year remeasurement periods within the
FIA program and limited land use change information available over the entire time series, estimates presented in
this section may not reflect the entire 20-year conversion history. Work is underway to integrate the dense time
series of remotely sensed data into a new estimation system, which will facilitate land conversion estimation over
the entire time series.
6.4 Cropland Remaining Cropland (CRF
Category 4B1)
Carbon (C) in cropland ecosystems occurs in biomass, dead organic matter, and soils. However, C storage in
cropland biomass and dead organic matter is relatively ephemeral and does not need to be reported according to
the IPCC (2006), with the exception of C stored in perennial woody crop biomass, such as citrus groves and apple
orchards, in addition to the biomass, downed wood and dead organic matter in agroforestry systems. Within soils,
C is found in organic and inorganic forms of C, but soil organic C is the main source and sink for atmospheric C02 in
most soils. IPCC (2006) recommends reporting changes in soil organic C stocks due to agricultural land-use and
management activities for mineral and organic soils.38
Well-drained mineral soils typically contain from 1 to 6 percent organic C by weight, whereas mineral soils with
high water tables for substantial periods of a year may contain significantly more C (NRCS 1999). Conversion of
mineral soils from their native state to agricultural land uses can cause up to half of the soil organic C to be lost to
the atmosphere due to enhanced microbial decomposition. The rate and ultimate magnitude of C loss depends on
subsequent management practices, climate and soil type (Ogle et al. 2005). Agricultural practices, such as clearing,
drainage, tillage, planting, grazing, crop residue management, fertilization, application of biosolids (i.e., treated
38 Carbon dioxide emissions associated with liming and urea application are also estimated but are included in the Liming and
Urea Fertilization sections of the Agriculture chapter of the Inventory.
Land Use, Land-Use Change, and Forestry 6-53

-------
sewage sludge) and flooding, can modify both organic matter inputs and decomposition, and thereby result in a
net C stock change (Paustian et al. 1997a; Lai 1998; Conant et al. 2001; Ogle et al. 2005; Griscom et al. 2017; Ogle
et al. 2019). Eventually, the soil can reach a new equilibrium that reflects a balance between C inputs (e.g.,
decayed plant matter, roots, and organic amendments such as manure and crop residues) and C loss through
microbial decomposition of organic matter (Paustian et al. 1997b).
Organic soils, also referred to as Histosols, include all soils with more than 12 to 20 percent organic C by weight,
depending on clay content (NRCS 1999; Brady and Weil 1999). The organic layer of these soils can be very deep
(i.e., several meters), and form under inundated conditions that results in minimal decomposition of plant
residues. When organic soils are prepared for crop production, they are drained and tilled, leading to aeration of
the soil that accelerates both the decomposition rate and C02 emissions.39 Due to the depth and richness of the
organic layers, C loss from drained organic soils can continue over long periods of time, which varies depending on
climate and composition (i.e., decomposability) of the organic matter (Armentano and Menges 1986). Due to
deeper drainage and more intensive management practices, the use of organic soils for annual crop production
leads to higher C loss rates than drainage of organic soils in grassland or forests (IPCC 2006).
Cropland Remaining Cropland includes all cropland in an Inventory year that has been cropland for a continuous
time period of at least 20 years. This determination is based on the United States Department of Agriculture
(USDA) National Resources Inventory (NRI) for non-federal lands (USDA-NRCS 2018a) and the National Land Cover
Dataset for federal lands (Yang et al. 2018; Homer et al. 2007; Fry et al. 2011; Homer et al. 2015). Cropland
includes all land that is used to produce food and fiber, forage that is harvested and used as feed (e.g., hay and
silage), in addition to cropland that has been enrolled in the Conservation Reserve Program (CRP)40 (i.e.,
considered set-aside cropland).
Cropland in Alaska is not included in the Inventory, but is a relatively small amount of U.S. cropland area
(approximately 28,700 hectares). Some miscellaneous croplands are also not included in the Inventory due to
limited understanding of greenhouse gas emissions from these management systems (e.g., aquaculture). This leads
to a small discrepancy between the managed area in Cropland Remaining Cropland (see Table 6-31 in Planned
Improvements for more details on the land area discrepancies) and the cropland area included in the Inventory
analysis. Improvements are underway to include croplands in Alaska as part of future C inventories.
Land-use and land management of mineral soils are the largest contributor to total net C stock change, especially
in the early part of the time series (see Table 6-28 and Table 6-29). In 2019, mineral soils are estimated to
sequester 47.4 MMT C02 Eq. from the atmosphere (12.9 MMT C). This rate of C storage in mineral soils represents
about a 18 percent decrease in the rate since the initial reporting year of 1990. Carbon dioxide emissions from
organic soils are 32.9 MMT C02 Eq. (9.0 MMT C) in 2019, which is a 6 percent decrease compared to 1990. In total,
United States agricultural soils in Cropland Remaining Cropland sequestered approximately 14.5 MMT C02 Eq. (4.0
MMT C) in 2019.
Table 6-28: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT
COz Eq.)
Soil Type
1990
2005
2015
2016
2017
2018
2019
Mineral Soils
(58.2)
(62.4)
(44.9)
(54.3)
(55.1)
(49.4)
(47.4)
Organic Soils
35.0
33.4
32.1
31.6
32.8
32.8
32.9
Total Net Flux
(23.2)
(29.0)
(12.8)
(22.7)
(22.3)
(16.6)
(14.5)
39	N20 emissions from drained organic soils are included in the Agricultural Soil Management section of the Agriculture chapter
of the Inventory.
40	The Conservation Reserve Program (CRP) is a land conservation program administered by the Farm Service Agency (FSA). In
exchange for a yearly rental payment, farmers enrolled in the program agree to remove environmentally sensitive land from
agricultural production and plant species that will improve environmental health and quality. Contracts for land enrolled in CRP
are 10 to 15 years in length. The long-term goal of the program is to re-establish valuable land cover to help improve water
quality, prevent soil erosion, and reduce loss of wildlife habitat.
6-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Notes: Totals may not sum due to independent rounding. Parentheses indicate net
sequestration.
Table 6-29: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT
C)
Soil Type
1990
2005
2015
2016
2017
2018
2019
Mineral Soils
(15.9)
(17.0)
(12.3)
(14.8)
(15.0)
(13.5)
(12.9)
Organic Soils
9.5
9.1
8.8
8.6
8.9
8.9
9.0
Total Net Flux
(6.3)
(7.9)
(3.5)
(6.2)
(6.1)
(4.5)
(4.0)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net
sequestration.
Soil organic C stocks increase in Cropland Remaining Cropland largely due to conservation tillage (i.e., reduced- and
no-till practices), land set-aside from production in the Conservation Reserve Program, annual crop production
with hay or pasture in rotations, and manure amendments. However, there is a decline in the net amount of C
sequestration (i.e., 2019 is 18 percent less than 1990), and this decline is due to lower sequestration rates in set-
aside lands, less impact of manure amendments and annual crop production with hay and pasture in rotation. Soil
organic C losses from drainage of organic soils are relatively stable across the time series with a small decline
associated with the land base declining for Cropland Remaining Cropland on organic soils since 1990.
The spatial variability in the 2015 annual soil organic C stock changes41 are displayed in Figure 6-6 and Figure 6-7
for mineral and organic soils, respectively. Isolated areas with high rates of C accumulation occur throughout the
agricultural land base in the United States, but there are more concentrated areas. In particular, higher rates of net
C accumulation in mineral soils occur in the Corn Belt region, which is the region with the largest amounts of
conservation tillage, along with moderate rates of CRP enrollment. The regions with the highest rates of emissions
from drainage of organic soils occur in the Southeastern Coastal Region (particularly Florida), upper Midwest and
Northeast surrounding the Great Lakes, and isolated areas along the Pacific Coast (particularly California), which
coincides with the largest concentrations of organic soils in the United States that are used for agricultural
production.
41 Only national-scale emissions are estimated for 2016 to 2019 in this Inventory using the surrogate data method, and
therefore the fine-scale emission patterns in this map are based on inventory data from 2015.
Land Use, Land-Use Change, and Forestry 6-55

-------
Figure 6-6: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural
Management within States, 2015, Cropland Remaining Cropland
~ -1 to 1
Note: Only national-scale soil organic C stock changes are estimated for 2016 to 2019 in the current Inventory
using a surrogate data method, and therefore the fine-scale emission patterns in this map are based on
inventory data from 2015. Negative values represent a net increase in soil organic C stocks, and positive values
represent a net decrease in soil organic C stocks.
6-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 6-7: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural
Management within States, 2015, Cropland Remaining Cropland
¦ >40
Note: Only national-scale soil organic C stock changes are estimated for 2016 to 2019 in the current Inventory
using a surrogate data method, and therefore the fine-scale emission patterns in this map are based on
inventory data from 2015.
Methodology
The following section includes a description of the methodology used to estimate changes in soil organic C stocks
for Cropland Remaining Cropland, including (1) agricultural land-use and management activities on mineral soils;
and (2) agricultural land-use and management activities on organic soils. Carbon dioxide emissions and removals42
due to changes in mineral soil organic C stocks are estimated using a Tier 3 method for the majority of annual
crops (Ogle et al. 2010), A Tier 2 IPCC method is used for the remaining crops not included in the Tier 3 method
(see list of crops in the Mineral Soil Carbon Stock Changes section below) (Ogle et al. 2003, 2006). In addition, a
Tier 2 method is used for very gravelly, cobbly, or shaley soils (i.e., classified as soils that have greater than 35
percent of soil volume comprised of gravel, cobbles, or shale, regardless of crop). Emissions from organic soils are
estimated using a Tier 2 IPCC method. While a combination of Tier 2 and 3 methods are used to estimate C stock
changes across most of the time series, a surrogate data method has been applied to estimate stock changes in the
last few years of the Inventory. Stock change estimates based on surrogate data will be recalculated in a future
Inventory report using the Tier 2 and 3 methods when data become available.
Soil organic C stock changes on non-federal lands are estimated for Cropland Remaining Cropland (as well as
agricultural land falling into the IPCC categories Land Converted to Cropland, Grassland Remaining Grassland, and
Land Converted to Grassland) according to land-use histories recorded in the USDA NRI survey (USDA-NRCS
2018a). The NRI is a statistically-based sample of all non-federal land, and includes approximately 489,178 survey
locations in agricultural land for the conterminous United States and Hawaii. Each survey location is associated
with an "expansion factor" that allows scaling of C stock changes from NRI survey locations to the entire country
42 Removals occur through uptake of C02 into crop and forage biomass that is later incorporated into soil C pools.
Land Use, Land-Use Change, and Forestry 6-57

-------
(i.e., each expansion factor represents the amount of area that is expected to have the same land-
use/management history as the sample point). Land-use and some management information (e.g., crop type, soil
attributes, and irrigation) are collected for each NRI point on a 5-year cycle beginning from 1982 through 1997. For
cropland, data has been collected for 4 out of 5 years during each survey cycle (i.e., 1979 through 1982,1984
through 1987,1989 through 1992, and 1994 through 1997). In 1998, the NRI program began collecting annual
data, and the annual data are currently available through 2015 (USDA-NRCS 2018a). NRI survey locations are
classified as Cropland Remaining Cropland in a given year between 1990 and 2015 if the land use has been
cropland for a continuous time period of at least 20 years. NRI survey locations are classified according to land-use
histories starting in 1979, and consequently the classifications are based on less than 20 years from 1990 to 1998.
This may have led to an overestimation of Cropland Remaining Cropland in the early part of the time series to the
extent that some areas are converted to cropland between 1971 and 1978.
Mineral Soil Carbon Stock Changes
An IPCC Tier 3 model-based approach (Ogle et al. 2010) is applied to estimate organic C stock changes for mineral
soils on the majority of land that is used to produce annual crops and forage crops that are harvested and used as
feed (e.g., hay and silage) in the United States. These crops include alfalfa hay, barley, corn, cotton, grass hay,
grass-clover hay, oats, peanuts, potatoes, rice, sorghum, soybeans, sugar beets, sunflowers, tobacco and wheat,
but is not applied to estimate organic C stock changes from other crops or rotations with other crops. The model-
based approach uses the DayCent biogeochemical model (Parton et al. 1998; Del Grosso et al. 2001, 2011) to
estimate soil organic C stock changes, soil nitrous oxide (N20) emissions from agricultural soil management, and
methane (CH4) emissions from rice cultivation. Carbon and N dynamics are linked in plant-soil systems through the
biogeochemical processes of microbial decomposition and plant production (McGill and Cole 1981). Coupling the
two source categories (i.e., agricultural soil C and N20) in a single inventory analysis ensures that there is a
consistent treatment of the processes and interactions between C and N cycling in soils.
The remaining crops on mineral soils are estimated using an IPCC Tier 2 method (Ogle et al. 2003), including some
vegetables, perennial/horticultural crops, and crops that are rotated with these crops. The Tier 2 method is also
used for very gravelly, cobbly, or shaley soils (greater than 35 percent by volume), and soil organic C stock changes
on federal croplands. Mineral soil organic C stocks are estimated using a Tier 2 method for these areas because the
DayCent model, which is used for the Tier 3 method, has not been fully tested for estimating C stock changes
associated with these crops and rotations, as well as cobbly, gravelly, or shaley soils. In addition, there is
insufficient information to simulate croplands on federal lands using DayCent.
A surrogate data method is used to estimate soil organic C stock changes from 2016 to 2019 at the national scale
for land areas included in the Tier 2 and Tier 3 methods. Specifically, linear regression models with autoregressive
moving-average (ARMA) errors (Brockwell and Davis 2016) are used to estimate the relationship between
surrogate data and the 1990 to 2015 stock change data that are derived using the Tier 2 and 3 methods. Surrogate
data for these regression models include corn and soybean yields from USDA-NASS statistics,43 and weather data
from the PRISM Climate Group (PRISM 2018). See Box 6-4 for more information about the surrogate data method.
Stock change estimates for 2016 to 2019 will be recalculated in future inventories when new NRI data are
available.
Box 6-4: Surrogate Data Method
Time series extension is needed because there are typically gaps at the end of the time series. This is mainly
because the NRI, which provides critical data for estimating greenhouse gas emissions and removals, does not
release new activity data every year.
A surrogate data method has been used to impute missing emissions at the end of the time series for soil
organic C stock changes in Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining
Grassland, and Land Converted to Grassland. A linear regression model with autoregressive moving-average
43 See .
6-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
(ARMA) errors (Brockwell and Davis 2016) is used to estimate the relationship between the surrogate data and
the modeled 1990 to 2015 emissions data that has been compiled using the inventory methods described in this
section. The model to extend the time series is given by
Y = xp + Ł,
where Y is the response variable (e.g., soil organic carbon), xp contains specific surrogate data depending on the
response variable, and Ł is the remaining unexplained error. Models with a variety of surrogate data were
tested, including commodity statistics, weather data, or other relevant information. Parameters are estimated
from the emissions data for 1990 to 2015 using standard statistical techniques, and these estimates are used to
predict the missing emissions data for 2016 to 2019.
A critical issue with application of splicing methods is to adequately account for the additional uncertainty
introduced by predicting emissions rather than compiling the full inventory. Consequently, uncertainty will
increase for years with imputed estimates based on the splicing methods, compared to those years in which the
full inventory is compiled. This added uncertainty is quantified within the model framework using a Monte Carlo
approach. The approach requires estimating parameters for results in each iteration of the Monte Carlo analysis
for the full inventory (i.e., the surrogate data model is refit with the emissions estimated in each Monte Carlo
iteration from the full inventory analysis with data from 1990 to 2015), estimating emissions from each model
and deriving confidence intervals combining uncertainty across all iterations. This approach propagates
uncertainties through the calculations from the original inventory and the surrogate data method. Furthermore,
the 95% confidence intervals are estimated using the 3 sigma rules assuming a unimodal density (Pukelsheim
1994).
Tier 3 Approach. Mineral soil organic C stocks and stock changes are estimated to a 30 cm depth using the
DayCent biogeochemical44 model (Parton et al. 1998; Del Grosso et al. 2001, 2011), which simulates cycling of C, N,
and other nutrients in cropland, grassland, forest, and savanna ecosystems. The DayCent model utilizes the soil C
modeling framework developed in the Century model (Parton et al. 1987,1988,1994; Metherell et al. 1993), but
has been refined to simulate dynamics at a daily time-step. Input data on land use and management are specified
at a daily resolution and include land-use type, crop/forage type, and management activities (e.g., planting,
harvesting, fertilization, manure amendments, tillage, irrigation, cover crops, and grazing; more information is
provided below). The model simulates net primary productivity (NPP) using the NASA-CASA production algorithm
MODIS Enhanced Vegetation Index (EVI) products, MOD13Q1 and MYD13Q1, for most croplands45 (Potter et al.
1993, 2007). The model simulates soil temperature and water dynamics, using daily weather data from a 4-
kilometer gridded product developed by the PRISM Climate Group (2018), and soil attributes from the Soil Survey
Geographic Database (SSURGO) (Soil Survey Staff 2019). This method is more accurate than the Tier 1 and 2
approaches provided by the IPCC (2006) because the simulation model treats changes as continuous over time as
opposed to the simplified discrete changes represented in the default method (see Box 6-5 for additional
information).
Box 6-5: Tier 3 Approach for Soil C Stocks Compared to Tier 1 or 2 Approaches
A Tier 3 model-based approach is used to estimate soil organic C stock changes for the majority of agricultural
land with mineral soils. This approach results in a more complete and accurate estimation of soil organic C stock
44	Biogeochemical cycles are the flow of chemical elements and compounds between living organisms and the physical
environment.
45	NPP is estimated with the NASA-CASA algorithm for most of the cropland that is used to produce major commodity crops in
the central United States from 2000 to 2015. Other regions and years prior to 2000 are simulated with a method that
incorporates water, temperature and moisture stress on crop production (see Metherell et al. 1993), but does not incorporate
the additional information about crop condition provided with remote sensing data.
Land Use, Land-Use Change, and Forestry 6-59

-------
changes and entails several fundamental differences from the IPCC Tier 1 or 2 methods, as described below.
1)	The IPCC Tier 1 and 2 methods are simplified approaches for estimating soil organic C stock changes
and classify land areas into discrete categories based on highly aggregated information about climate
(six regions), soil (seven types), and management (eleven management systems) in the United States.
In contrast, the Tier 3 model incorporates the same variables (i.e., climate, soils, and management
systems) with considerably more detail both temporally and spatially, and captures multi-dimensional
interactions through the more complex model structure.
2)	The IPCC Tier 1 and 2 methods have a coarser spatial resolution in which data are aggregated to soil
types in climate regions, of which there about 30 combinations in the United States. In contrast, the
Tier 3 model simulates soil C dynamics at about 350,000 individual NRI survey locations in crop fields
and grazing lands.
The IPCC Tier 1 and 2 methods use a simplified approach for estimating changes in C stocks that assumes a step-
change from one equilibrium level of the C stock to another equilibrium level. In contrast, the Tier 3 approach
simulates a continuum of C stock changes that may reach a new equilibrium over an extended period of time
depending on the environmental conditions (i.e., a new equilibrium often requires hundreds to thousands of
years to reach). More specifically, the DayCent model, which is used in the United States Inventory, simulates
soil C dynamics (and C02 emissions and uptake) on a daily time step based on C emissions and removals from
plant production and decomposition processes. These changes in soil organic C stocks are influenced by
multiple factors that affect primary production and decomposition, including changes in land use and
management, weather variability and secondary feedbacks between management activities, climate, and soils.
Historical land-use patterns and irrigation histories are simulated with DayCent based on the 2015 USDA NRI
survey (USDA-NRCS 2018a). Additional sources of activity data are used to supplement the activity data from the
NRI. The USDA-NRCS Conservation Effects and Assessment Project (CEAP) provides data on a variety of cropland
management activities, and is used to inform the inventory analysis about tillage practices, mineral fertilization,
manure amendments, cover cropping management, as well as planting and harvest dates (USDA-NRCS 2018b;
USDA-NRCS 2012). CEAP data are collected at a subset of NRI survey locations, and currently provide management
information from approximately 2002 to 2006. These data are combined with other datasets in an imputation
analysis that extend the time series from 1990 to 2015. This imputation analysis is comprised of three steps: a)
determine the trends in management activity across the time series by combining information across several
datasets (discussed below), b) use an artificial neural network to determine the likely management practice at a
given NRI survey location (Cheng and Titterington 1994), and c) assign management practices from the CEAP
survey to the specific NRI locations using predictive mean matching methods that is adapted to reflect the trending
information (Little 1988, van Buuren 2012). The artificial neural network is a machine learning method that
approximates nonlinear functions of inputs and searches through a very large class of models to impute an initial
value for management practices at specific NRI survey locations. The predictive mean matching method identifies
the most similar management activity recorded in the CEAP survey that matches the prediction from the artificial
neural network. Predictive mean matching ensures that imputed management activities are realistic for each NRI
survey location, and not odd or physically unrealizable results that could be generated by the artificial neural
network. There are six complete imputations of the management activity data using these methods.
To determine trends in mineral fertilization and manure amendments from 1979 to 2015, CEAP data are combined
with information on fertilizer use and rates by crop type for different regions of the United States from the USDA
Economic Research Service. The data collection program was known as the Cropping Practices Surveys through
1995 (USDA-ERS 1997), and is now part of a data collection program known as the Agricultural Resource
Management Surveys (ARMS) (USDA-ERS 2018). Additional data on fertilization practices are compiled through
other sources particularly the National Agricultural Statistics Service (USDA-NASS 1992,1999, 2004). The donor
survey data from CEAP contain both mineral fertilizer rates and manure amendment rates, so that the selection of
a donor via predictive mean matching yields the joint imputation of both rates. This approach captures the
relationship between mineral fertilization and manure amendment practices for U.S. croplands based directly on
the observed patterns in the CEAP survey data.
6-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
To determine the trends in tillage management from 1979 to 2015, CEAP data are combined with Conservation
Technology Information Center data between 1989 and 2004 (CTIC 2004) and USDA-ERS Agriculture Resource
Management Surveys (ARMS) data from 2002 to 2015 (Claasen et al. 2018). CTIC data are adjusted for long-term
adoption of no-till agriculture (Towery 2001). It is assumed that the majority of agricultural lands are managed
with full tillage prior to 1985. For cover crops, CEAP data are combined with information from 2011 to 2016 in the
USDA Census of Agriculture (USDA-NASS 2012, 2017). It is assumed that cover cropping was minimal prior to 1990
and the rates increased linearly over the decade to the levels of cover crop management derived from the CEAP
survey.
Uncertainty in the C stock estimates from DayCent associated with management activity includes input uncertainty
due to missing management data in the NRI survey, which is imputed from other sources as discussed above;
model uncertainty due to incomplete specification of C and N dynamics in the DayCent model algorithms and
associated parameterization; and sampling uncertainty associated with the statistical design of the NRI survey. To
assess input uncertainty, The C and N dynamics at each NRI survey location are simulated six times using the
imputation product and other model driver data. Uncertainty in parameterization and model algorithms are
determined using a structural uncertainty estimator as described in Ogle et al. (2007, 2010). Sampling uncertainty
is assessed using the NRI replicate sampling weights.
Carbon stocks and 95 percent confidence intervals are estimated for each year between 1990 and 2015 using the
DayCent model. However, note that the areas have been modified in the original NRI survey through the process in
which the Forest Inventory and Analysis (FIA) survey data and the National Land Cover Dataset (Homer et al. 2007;
Fry et al. 2011; Homer et al. 2015) are harmonized with the NRI data. This process ensures that the areas of Forest
Land Remaining Forest Land and Land Converted to Forest Land are consistent with other land use categories while
maintaining a consistent time series for the total land area of the United States. For example, if the FIA estimate
less Cropland Converted to Forest Land than the NRI, then the amount of area for this land use conversion is
reduced in the NRI dataset and re-classified as Cropland Remaining Cropland (See Section 6, Representation of the
U.S. Land Base for more information). Further elaboration on the methodology and data used to estimate stock
changes from mineral soils are described in Annex 3.12.
Soil organic C stock changes from 2016 to 2019 are estimated using a surrogate data method that is described in
Box 6-4. Future Inventories will be updated with new NRI activity data when the data are made available, and the
time series from 2016 to 2019 will be recalculated.
Tier 2 Approach. In the IPCC Tier 2 method, data on climate, soil types, land-use, and land management activity
are used to classify land area and apply appropriate factors to estimate soil organic C stock changes to a 30 cm
depth (Ogle et al. 2003, 2006). The primary source of activity data for land use, crop and irrigation histories is the
2015 NRI survey (USDA-NRCS 2018a). Each NRI survey location is classified by soil type, climate region, and
management condition using data from other sources. Survey locations on federal lands are included in the NRI,
but land use and cropping history are not compiled for these locations in the survey program (i.e., NRI is restricted
to data collection on non-federal lands). Therefore, land-use patterns for the NRI survey locations on federal lands
are based on the National Land Cover Database (NLCD) (Yang et al. 2018; Fry et al. 2011; Homer et al. 2007; Homer
et al. 2015).
Additional management activities needed for the Tier 2 method are based on the imputation product described for
the Tier 3 approach, including tillage practices, mineral fertilization, and manure amendments that are assigned to
NRI survey locations. The one exception are activity data on wetland restoration of Conservation Reserve Program
land that are obtained from Euliss and Gleason (2002). Climate zones in the United States are classified using mean
precipitation and temperature (1950 to 2000) variables from the WorldClim data set (Hijmans et al. 2005) and
potential evapotranspiration data from the Consortium for Spatial Information (CGIAR-CSI) (Zomer et al. 2008,
2007) (Figure A-9). IPCC climate zones are then assigned to NRI survey locations.
Reference C stocks are estimated using the National Soil Survey Characterization Database (NRCS 1997) with
cultivated cropland as the reference condition, rather than native vegetation as used in IPCC (2006). Soil
measurements under agricultural management are much more common and easily identified in the National Soil
Survey Characterization Database (NRCS 1997) than are soils under a native condition, and therefore cultivated
cropland provides a more robust sample for estimating the reference condition. Country-specific C stock change
Land Use, Land-Use Change, and Forestry 6-61

-------
factors are derived from published literature to determine the impact of management practices on soil organic C
storage (Ogle et al. 2003, 2006). The factors represent changes in tillage, cropping rotations, intensification, and
land-use change between cultivated and uncultivated conditions. However, country-specific factors associated
with organic matter amendments are not estimated due to an insufficient number of studies in the United States
to analyze the impacts. Instead, factors from IPCC (2006) are used to estimate the effect of those activities.
Changes in soil organic C stocks for mineral soils are estimated 1,000 times for 1990 through 2015, using a Monte
Carlo stochastic simulation approach and probability distribution functions for the country-specific stock change
factors, reference C stocks, and land-use activity data (Ogle et al. 2003; Ogle et al. 2006). Further elaboration on
the methodology and data used to estimate stock changes from mineral soils are described in Annex 3.12.
Soil organic C stock changes from 2016 to 2019 are estimated using a surrogate data method that is described in
Box 6-4. As with the Tier 3 method, future Inventories will be updated with new NRI activity data when the data
are made available, and the time series will be recalculated (see Planned Improvements section).
Organic Soil Carbon Stock Changes
Annual C emissions from drained organic soils in Cropland Remaining Cropland are estimated using the Tier 2
method provided in IPCC (2006), with country-specific C loss rates (Ogle et al. 2003) rather than default IPCC rates.
The final estimates include a measure of uncertainty as determined from a Monte Carlo Simulation with 1,000
iterations. Emissions are based on the land area data for drained organic soils from 1990 to 2015 for Cropland
Remaining Cropland in the 2015 NRI (USDA-NRCS 2018a). Further elaboration on the methodology and data used
to estimate stock changes from organic soils are described in Annex 3.12.
A surrogate data method was used to estimate annual C emissions from organic soils from 2016 to 2019 as
described in Box 6-4 of this section. Estimates for 2016 to 2019 will be recalculated in future Inventories when new
NRI data are available.
Uncertainty and Time-Series Consistency
Uncertainty is quantified for changes in soil organic C stocks associated with Cropland Remaining Cropland
(including both mineral and organic soils). Uncertainty estimates are presented in Table 6-30 for each subsource
(mineral and organic soil C stocks) and the methods that are used in the Inventory analyses (i.e., Tier 2 and Tier 3).
Uncertainty for the Tier 2 and 3 approaches is derived using a Monte Carlo approach (see Annex 3.12 for further
discussion). For 2016 to 2019, additional uncertainty is propagated through the Monte Carlo Analysis that is
associated with the surrogate data method. Soil organic C stock changes from the Tier 2 and 3 approaches are
combined using the simple error propagation method provided by the IPCC (2006). The combined uncertainty is
calculated by taking the square root of the sum of the squares of the standard deviations of the uncertain
quantities.
The combined uncertainty for soil organic C stocks in Cropland Remaining Cropland ranges from 601 percent below
to 601 percent above the 2019 stock change estimate of-14.5 MMT C02 Eq. The large relative uncertainty around
the 2019 stock change estimate is mostly due to variation in soil organic C stock changes that is not explained by
the surrogate data method, leading to high prediction error with this splicing method.
Table 6-30: Approach 2 Quantitative Uncertainty Estimates for Soil C Stock Changes
occurring within Cropland Remaining Cropland (MMT CO2 Eq. and Percent)
Source
2019 Flux Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Flux Estimate3
(MMTCOz Eq.)	(%)
Mineral Soil C Stocks: Cropland Remaining
Cropland, Tier 3 Inventory Methodology
Mineral Soil C Stocks: Cropland Remaining
Cropland, Tier 2 Inventory Methodology

Lower
Upper
Lower
Upper

Bound
Bound
Bound
Bound
(41.5)
(126.6)
43.5
-205%
205%
(5.9)
(12.5)
0.8
-114%
114%
6-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Organic Soil C Stocks: Cropland Remaining
Cropland, Tier 2 Inventory Methodology	°	°
Combined Uncertainty for Flux associated
with Agricultural Soil Carbon Stock Change in	(14.5)	(102.0)	72.9	-601%	601%
Cropland Remaining Cropland	
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation with a 95 percent confidence interval.
Uncertainty is also associated with lack of reporting of agricultural woody biomass and dead organic matter C stock
changes. However, woody biomass C stock changes are likely minor in perennial crops, such as orchards and nut
plantations. There will be removal and replanting of tree crops each year, but the net effect on biomass C stock
changes is probably minor because the overall area and tree density is relatively constant across time series. In
contrast, agroforestry practices, such as shelterbelts, riparian forests and intercropping with trees, may have more
significant changes over the Inventory time series, compared to perennial woody crops, at least in some regions of
the United States, but there are currently no datasets to evaluate the trends. Changes in litter C stocks are also
assumed to be negligible in croplands over annual time frames, although there are certainly significant changes at
sub-annual time scales across seasons. This trend may change in the future, particularly if crop residue becomes a
viable feedstock for bioenergy production.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
Quality control measures included checking input data, model scripts, and results to ensure data are properly
handled throughout the inventory process. Inventory reporting forms and text are reviewed and revised as needed
to correct transcription errors. Results from the DayCent model are compared to field measurements and soil
monitoring sites associated with the NRI (Spencer et al. 2011), and a statistical relationship has been developed to
assess uncertainties in the predictive capability of the model (Ogle et al. 2007). The comparisons include 72 long-
term experiment sites and 142 NRI soil monitoring network sites, with 948 observations across all of the sites (see
Annex 3.12 for more information).
Recalculations Discussion
There are no recalculations in the time series from the previous Inventory.
Planned Improvements
A key improvement for a future Inventory will be to incorporate additional management activity data from the
USDA-NRCS Conservation Effects Assessment Project survey. This survey has compiled new data in recent years
that will be available for the Inventory analysis by next year. The latest land use data will also be incorporated from
the USDA National Resources Inventory and related management data from USDA-ERS ARMS surveys.
There are several other planned improvements underway related to the plant production module. Crop
parameters associated with temperature effects on plant production will be further improved in DayCent with
additional model calibration. Senescence events following grain filling in crops, such as wheat, are being modified
based on recent model algorithm development, and will be incorporated. There will also be further testing and
parameterization of the DayCent model to reduce the bias in model predictions for grasslands, which was
discovered through model evaluation by comparing output to measurement data from 72 experimental sites and
142 NRI soil monitoring network sites (See QA/QC and Verification section).
Land Use, Land-Use Change, and Forestry 6-63

-------
Improvements are underway to simulate crop residue burning in the DayCent model based on the amount of crop
residues burned according to the data that are used in the Field Burning of Agricultural Residues source category
(see Section 5.7). This improvement will more accurately represent the C inputs to the soil that are associated with
residue burning.
A review of available data on biosolids (i.e., treated sewage sludge) application will be undertaken to improve the
distribution of biosolids application on croplands, grasslands and settlements.
In the future, the Inventory will include an analysis of C stock changes in Alaska for cropland, using the Tier 2
method for mineral and organic soils that is described earlier in this section. This analysis will initially focus on land
use change, which typically has a larger impact on soil organic C stock changes than management practices, but
will be further refined over time to incorporate management data. See Table 6-31 for the amount of managed area
in Cropland Remaining Cropland that is not included in the Inventory, which is less than one thousand hectares per
year. This includes the area in Alaska and also other miscellaneous cropland areas, such as aquaculture.
Many of these improvements are expected to be completed for the 1990 through 2020 Inventory (i.e., 2021
submission to the UNFCCC). However, the timeline may be extended if there are insufficient resources to fund all
or part of these planned improvements.
Table 6-31: Area of Managed Land in Cropland Remaining Cropland that is not included in
the current Inventory (Thousand Hectares)
Area (Thousand Hectares)
Year
Managed Land
Inventory
Not Included in
Inventory
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
162,163
161,721
161,252
159,449
157,732
157,054
156,409
155,767
152,016
151,135
150,981
150,471
150,175
150.843
150,645
150,304
149,791
150,032
149,723
149,743
149,343
148.844
148,524
149,018
149,492
148,880
162,134
161,692
161,223
159,420
157,703
157,025
156,380
155,738
151,987
151,105
150,952
150,442
150,146
150.814
150,616
150,275
149,762
150,003
149,694
149,714
149,314
148.815
148,495
148,989
149,463
148,851
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
29
ND
ND
ND
ND
ND
ND
6-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2018	ND	ND	ND
	2019	ND	ND	ND	
Note: NRI data are not available after 2015, and so these years are designated as ND (No data).
6.5 Land Converted to Cropland (CRF Category
4B2)	
Land Converted to Cropland includes all cropland in an inventory year that had been in another land use(s) during
the previous 20 years (USDA-NRCS 2018), and used to produce food or fiber, or forage that is harvested and used
as feed (e.g., hay and silage). For example, grassland or forest land converted to cropland during the past 20 years
would be reported in this category. Recently converted lands are retained in this category for 20 years as
recommended by IPCC (2006). This Inventory includes all croplands in the conterminous United States and Hawaii,
but does not include a minor amount of Land Converted to Cropland in Alaska. Some miscellaneous croplands are
also not included in the Inventory due to limited understanding of greenhouse gas dynamics in management
systems (e.g., aquaculture). Consequently, there is a discrepancy between the total amount of managed area in
Land Converted to Cropland (see Section 6 Representation of the U.S. Land Base) and the cropland area included in
the Inventory. Improvements are underway to include croplands in Alaska and miscellaneous croplands in future C
inventories (see Table 6-35 in the Planned Improvements section for more details on the land area discrepancies).
Land-use change can lead to large losses of C to the atmosphere, particularly conversions from forest land
(Houghton et al. 1983; Houghton and Nassikas 2017). Moreover, conversion of forest to another land use (i.e.,
deforestation) is one of the largest anthropogenic sources of emissions to the atmosphere globally, although this
source may be declining according to a recent assessment (Tubiello et al. 2015).
The 2006 IPCC Guidelines recommend reporting changes in biomass, dead organic matter and soil organic C stocks
with land use change. All soil organic C stock changes are estimated and reported for Land Converted to Cropland,
but reporting of C stock changes for aboveground and belowground biomass, dead wood, and litter pools is limited
to Forest Land Converted to Cropland.46
Forest Land Converted to Cropland is the largest source of emissions from 1990 to 2019, accounting for
approximately 87 percent of the average total loss of C among all of the land use conversions in Land Converted to
Cropland. The pattern is due to the large losses of biomass and dead organic matter C for Forest Land Converted to
Cropland. The next largest source of emissions is Grassland Converted to Cropland accounting for approximately 17
percent of the total emissions (Table 6-32 and Table 6-33).
The net change in total C stocks for 2019 led to C02 emissions to the atmosphere of 54.2 MMT C02 Eq. (14.8 MMT
C), including 28.3 MMT C02 Eq. (7.7 MMT C) from aboveground biomass C losses, 5.6 MMT C02 Eq. (1.5 MMT C)
from belowground biomass C losses, 5.5 MMT C02 Eq. (1.5 MMT C) from dead wood C losses, 8.1 MMT C02 Eq.
(2.2 MMT C) from litter C losses, 3.0 MMT C02 Eq. (0.8 MMT C) from mineral soils and 3.7 MMT C02 Eq. (1.0 MMT
C) from drainage and cultivation of organic soils. Emissions in 2019 are 5 percent higher than emissions in the
initial reporting year, i.e., 1990.
46 Changes in biomass C stocks are not currently reported land use conversions to cropland except for forest land converted to
cropland, but this is a planned improvement for a future inventory. Note: changes in dead organic matter are assumed to
negligible for other land use conversions to cropland, except forest land.
Land Use, Land-Use Change, and Forestry 6-65

-------
Table 6-32: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Land Converted to Croplandby Land Use Change Category (MMT CO2 Eq.)

1990
2005
2015
2016
2017
2018
2019
Grassland Converted to Cropland
6.9
7.5
10.2
8.5
8.7
8.5
8.4
Mineral Soils
4.1
4.0
6.9
5.2
5.4
5.1
5.1
Organic Soils
2.7
3.5
3.3
3.3
3.3
3.3
3.3
Forest Land Converted to Cropland
46.3
46.8
47.5
47.6
47.6
47.6
47.6
Aboveground Live Biomass
27.3
27.7
28.2
28.3
28.3
28.3
28.3
Belowground Live Biomass
5.4
5.5
5.6
5.6
5.6
5.6
5.6
Dead Wood
5.4
5.5
5.5
5.5
5.5
5.5
5.5
Litter
7.7
7.9
8.1
8.1
8.1
8.1
8.1
Mineral Soils
0.4
0.2
0.1
0.1
0.1
0.1
0.1
Organic Soils
0.1
0.1
0.0
0.0
0.0
0.0
0.0
Other Lands Converted to Cropland
(2.2)
(2.9)
(2.0)
(2.1)
(2.2)
(2.2)
(2.3)
Mineral Soils
(2.3)
(2.9)
(2.0)
(2.1)
(2.2)
(2.2)
(2.3)
Organic Soils
0.2
0.1
0.0
0.0
0.0
0.0
0.0
Settlements Converted to Cropland
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Mineral Soils
(0.1)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Organic Soils
+
+
0.0
0.0
0.0
0.0
0.0
Wetlands Converted to Cropland
0.8
0.9
0.5
0.5
0.6
0.6
0.6
Mineral Soils
0.3
0.3
0.2
0.2
0.2
0.2
0.2
Organic Soils
0.6
0.6
0.3
0.3
0.3
0.4
0.4
Aboveground Live Biomass
27.3
27.7
28.2
28.3
28.3
28.3
28.3
Belowground Live Biomass
5.4
5.5
5.6
5.6
5.6
5.6
5.6
Dead Wood
5.4
5.5
5.5
5.5
5.5
5.5
5.5
Litter
7.7
7.9
8.1
8.1
8.1
8.1
8.1
Total Mineral Soil Flux
2.3
1.3
5.0
3.3
3.4
3.1
3.0
Total Organic Soil Flux
3.7
4.3
3.7
3.7
3.7
3.7
3.7
Total Net Flux
51.8
52.2
56.1
54.4
54.6
54.3
54.2
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 6-33: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Land Converted to Cropland {MMT C)

1990
2005
2015
2016
2017
2018
2019
Grassland Converted to Cropland
1.9
2.0
2.8
2.3
2.4
2.3
2.3
Mineral Soils
1.1
1.1
1.9
1.4
1.5
1.4
1.4
Organic Soils
0.7
1.0
0.9
0.9
0.9
0.9
0.9
Forest Land Converted to Cropland
12.6
12.8
13.0
13.0
13.0
13.0
13.0
Aboveground Live Biomass
7.4
7.6
7.7
7.7
7.7
7.7
7.7
Belowground Live Biomass
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Dead Wood
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Litter
2.1
2.2
2.2
2.2
2.2
2.2
2.2
Mineral Soils
0.1
0.0
0.0
0.0
0.0
0.0
0.0
Organic Soils
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Other Lands Converted to Cropland
(0.6)
(0.8)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
Mineral Soils
(0.6)
(0.8)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
Organic Soils
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Settlements Converted to Cropland
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
Mineral Soils
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
Organic Soils
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Wetlands Converted to Cropland
0.2
0.3
0.1
0.1
0.2
0.2
0.2
Mineral Soils
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Organic Soils
0.2
0.2
0.1
0.1
0.1
0.1
0.1
Aboveground Live Biomass
7.4
7.6
7.7
7.7
7.7
7.7
7.7
Belowground Live Biomass
1.5
1.5
1.5
1.5
1.5
1.5
1.5
6-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Dead Wood
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Litter
2.1
2.2
2.2
2.2
2.2
2.2
2.2
Total Mineral Soil Flux
0.6
0.4
1.4
0.9
0.9
0.8
0.8
Total Organic Soil Flux
1.0
1.2
1.0
1.0
1.0
1.0
1.0
Total Net Flux
14.1
14.2
15.3
14.8
14.9
14.8
14.8
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C.
Methodology
The following section includes a description of the methodology used to estimate C stock changes for Land
Converted to Cropland, including (1) loss of aboveground and belowground biomass, dead wood and litter C with
conversion of forest lands to croplands, as well as (2) the impact from all land use conversions to cropland on
mineral and soil organic C stocks.
Biomass, Dead Wood and Litter Carbon Stock Changes
A Tier 2 method is applied to estimate biomass, dead wood, and litter C stock changes for Forest Land Converted to
Cropland. Estimates are calculated in the same way as those in the Forest Land Remaining Forest Land category
using data from the USDA Forest Service, Forest Inventory and Analysis (FIA) program (USDA Forest Service 2020).
However, there are no country-specific data for cropland biomass, so default biomass values (IPCC 2006) were
used to estimate the carbon stocks for the new cropland (litter and dead wood carbon stocks were assumed to be
zero since no reference C density estimates exist for croplands). The difference between the stocks is reported as
the stock change under the assumption that the change occurred in the year of the conversion. If FIA plots include
data on individual trees, aboveground and belowground C density estimates are based on Woodall et al. (2011).
Aboveground and belowground biomass estimates also include live understory which is a minor component of
biomass defined as all biomass of undergrowth plants in a forest, including woody shrubs and trees less than 2.54
cm dbh. For this Inventory, it was assumed that 10 percent of total understory C mass is belowground (Smith et al.
2006). Estimates of C density are based on information in Birdsey (1996) and biomass estimates from Jenkins et al.
(2003).
For dead organic matter, if FIA plots include data on standing dead trees, standing dead tree C density is estimated
following the basic method applied to live trees (Woodall et al. 2011) with additional modifications to account for
decay and structural loss (Domke et al. 2011; Harmon et al. 2011). If FIA plots include data on downed dead wood,
downed dead wood C density is estimated based on measurements of a subset of FIA plots for downed dead wood
(Domke et al. 2013; Woodall and Monleon 2008). Downed dead wood is defined as pieces of dead wood greater
than 7.5 cm diameter, at transect intersection, that are not attached to live or standing dead trees. This includes
stumps and roots of harvested trees. To facilitate the downscaling of downed dead wood C estimates from the
state-wide population estimates to individual plots, downed dead wood models specific to regions and forest types
within each region are used. Litter C is the pool of organic C (also known as duff, humus, and fine woody debris)
above the mineral soil and includes woody fragments with diameters of up to 7.5 cm. A subset of FIA plots are
measured for litter C. If FIA plots include litter material, a modeling approach using litter C measurements from FIA
plots is used to estimate litter C density (Domke et al. 2016). See Annex 3.13 for more information about reference
C density estimates for forest land and the compilation system used to estimate carbon stock changes from forest
land.
Soil Carbon Stock Changes
Soil organic stock changes are estimated for Land Converted to Cropland according to land-use histories recorded
in the 2015 USDA NRI survey for non-federal lands (USDA-NRCS 2018). Land-use and some management
information (e.g., crop type, soil attributes, and irrigation) had been collected for each NRI point on a 5-year cycle
beginning in 1982. In 1998, the NRI program began collecting annual data, which are currently available through
2015 (USDA-NRCS 2018). NRI survey locations are classified as Land Converted to Cropland in a given year between
1990 and 2015 if the land use is cropland but had been another use during the previous 20 years. NRI survey
Land Use, Land-Use Change, and Forestry 6-67

-------
locations are classified according to land-use histories starting in 1979, and consequently the classifications are
based on less than 20 years from 1990 to 1998, which may have led to an underestimation of Land Converted to
Cropland in the early part of the time series to the extent that some areas are converted to cropland from 1971 to
1978. For federal lands, the land use history is derived from land cover changes in the National Land Cover Dataset
(Yang et al. 2018; Homer et al. 2007; Fry et al. 2011; Homer et al. 2015).
Mineral Soil Carbon Stock Changes
An IPCC Tier 3 model-based approach (Ogle et al. 2010) is applied to estimate C stock changes from 1990 to 2015
for mineral soils on the majority of land that is used to produce annual crops and forage crops that are harvested
and used as feed (e.g., hay and silage) in the United States. These crops include alfalfa hay, barley, corn, cotton,
grass hay, grass-clover hay, oats, peanuts, potatoes, rice, sorghum, soybeans, sugar beets, sunflowers, tobacco,
and wheat. Soil organic C stock changes on the remaining mineral soils are estimated with the IPCC Tier 2 method
(Ogle et al. 2003), including land used to produce some vegetables and perennial/horticultural crops and crops
rotated with these crops; land on very gravelly, cobbly, or shaley soils (greater than 35 percent by volume); and
land converted from another land use or federal ownership.47
For the years 2016 to 2019, a surrogate data method is used to estimate soil organic C stock changes at the
national scale for land areas included in the Tier 2 and Tier 3 methods. Specifically, linear regression models with
autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) are used to estimate the relationship
between surrogate data and the 1990 to 2015 stock change data from the Tier 2 and 3 methods. Surrogate data
for these regression models include corn and soybean yields from USDA-NASS statistics,48 and weather data from
the PRISM Climate Group (PRISM 2015). See Box 6-4 in the Methodology section of Cropland Remaining Cropland
for more information about the surrogate data method. Stock change estimates for 2016 to 2019 will be
recalculated in future inventories when new NRI data are available.
Tier 3 Approach. For the Tier 3 method, mineral soil organic C stocks and stock changes are estimated using the
DayCent biogeochemical model (Parton et al. 1998; Del Grosso et al. 2001, 2011). The DayCent model utilizes the
soil C modeling framework developed in the Century model (Parton et al. 1987,1988,1994; Metherell et al. 1993),
but has been refined to simulate dynamics at a daily time-step. National estimates are obtained by using the
model to simulate historical land-use change patterns as recorded in the USDA NRI survey (USDA-NRCS 2018).
Carbon stocks and 95 percent confidence intervals are estimated for each year between 1990 and 2015. See the
Cropland Remaining Cropland section and Annex 3.12 for additional discussion of the Tier 3 methodology for
mineral soils.
Soil organic C stock changes from 2016 to 2019 were estimated using the surrogate data method described in Box
6-4 of the Methodology section in Cropland Remaining Cropland. Future inventories will be updated with new
activity data when the data are made available, and the time series will be recalculated (see the Planned
Improvements section in Cropland Remaining Cropland).
Tier 2 Approach. For the mineral soils not included in the Tier 3 analysis, soil organic C stock changes are estimated
using a Tier 2 Approach, as described in the Tier 2 Approach for mineral soils in Cropland Remaining Cropland. This
includes application of the surrogate data method that is described in Box 6-4of the Methodology section in
Cropland Remaining Cropland. As with the Tier 3 method, future inventories will be updated with new NRI activity
data when the data are made available, and the time series will be recalculated.
Organic Soil Carbon Stock Changes
Annual C emissions from drained organic soils in Land Converted to Cropland are estimated using the Tier 2
method provided in IPCC (2006), with country-specific C loss rates (Ogle et al. 2003) as described in the Cropland
47	Federal land is not a land use, but rather an ownership designation that is treated as grassland for purposes of these
calculations. The specific land use on federal lands is not identified in the NRI survey (USDA-NRCS 2015).
48	See .
6-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Remaining Cropland section for organic soils. Further elaboration on the methodology is also provided in Annex
3.12.
The Inventory analysis includes application of the surrogate data method that is described in Box 6-4 of the
Methodology section in Cropland Remaining Cropland. Estimates will be recalculated in future Inventories when
new NRI data are available.
Uncertainty and Time-Series Consistency
The uncertainty analysis for biomass, dead wood and litter C losses with Forest Land Converted to Cropland is
conducted in the same way as the uncertainty assessment for forest ecosystem C flux associated with Forest Land
Remaining Forest Land. Sample and model-based error are combined using simple error propagation methods
provided by the IPCC (2006) by taking the square root of the sum of the squares of the standard deviations of the
uncertain quantities. For additional details, see the Uncertainty Analysis in Annex 3.13.
The uncertainty analyses for mineral soil organic C stock changes using the Tier 3 and Tier 2 methodologies are
based on a Monte Carlo approach that is described in Cropland Remaining Cropland (Also see Annex 3.12 for
further discussion). The uncertainty for annual C emission estimates from drained organic soils in Land Converted
to Cropland is estimated using a Monte Carlo approach, which is also described in the Cropland Remaining
Cropland section. For 2016 to 2019, there is additional uncertainty propagated through the Monte Carlo Analysis
associated with the surrogate data method, which is also described in Cropland Remaining Cropland.
Uncertainty estimates are presented in Table 6-34 for each subsource (i.e., biomass C stocks, dead wood C stocks,
litter C stocks, soil organic C stocks for mineral and organic soils) and the method applied in the Inventory analysis
(i.e., Tier 2 and Tier 3). Uncertainty estimates for the total C stock changes for biomass, dead organic matter and
soils are combined using the simple error propagation methods provided by the IPCC (2006), as discussed in the
previous paragraph. The combined uncertainty for total C stocks in Land Converted to Cropland ranged from 102
percent below to 103 percent above the 2019 stock change estimate of 54.2 MMT C02 Eq. The large relative
uncertainty in the 2019 estimate is mostly due to variation in soil organic C stock changes that is not explained by
the surrogate data method, leading to high prediction error with this splicing method.
Table 6-34: Approach 2 Quantitative Uncertainty Estimates for Soil, Dead Organic Matter
and Biomass C Stock Changes occurring within Land Converted to Cropland (MMT CO2 Eq.
and Percent)
2018 Flux Estimate Uncertainty Range Relative to Flux Estimate3
(MMT C02 Eq.)	(MMTCOz Eq.)	(%)


Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Grassland Converted to Cropland
8.4
(31.1)
48.0
-468%
468%
Mineral Soil C Stocks: Tier 3
0.8
(38.5)
40.2
-4767%
4767%
Mineral Soil C Stocks: Tier 2
4.3
1.3
7.3
-70%
70%
Organic Soil C Stocks: Tier 2
3.3
0.8
5.8
-75%
75%
Forest Land Converted to Cropland
47.6
8.8
87.2
-81%
83%
Aboveground Live Biomass
28.3
(7.7)
64.7
-127%
129%
Belowground Live Biomass
5.6
(1.5)
12.8
-127%
129%
Dead Wood
5.5
(1.6)
13.3
-129%
141%
Litter
8.1
(2.3)
19.4
-129%
140%
Mineral Soil C Stocks: Tier 2
0.1
+
0.3
-136%
136%
Organic Soil C Stocks: Tier 2
+
(0.1)
0.1
-1350%
1350%
Other Lands Converted to Cropland
(2.3)
(3.7)
(0.9)
-61%
61%
Mineral Soil C Stocks: Tier 2
(2.3)
(3.7)
(0.9)
-61%
61%
Organic Soil C Stocks: Tier 2
+
+
+
+
+
Settlements Converted to Cropland
(0.1)
(0.3)
+
-111%
111%
Mineral Soil C Stocks: Tier 2
(0.2)
(0.3)
+
-86%
86%
Organic Soil C Stocks: Tier 2
+
+
0.1
-84%
84%
Land Use, Land-Use Change, and Forestry 6-69

-------
Wetlands Converted to Croplands	0.6	+	1.2	-96%	96%
Mineral Soil C Stocks: Tier 2 0.2 + 0.5 -105% 105%
Organic Soil C Stocks: Tier 2	04	(02)	09	-141%	141%
Total: Land Converted to Cropland
54.2
(1.2)
110.2
-102%
103%
Aboveground Live Biomass
28.3
(7.7)
64.7
-127%
129%
Belowground Live Biomass
5.6
(1.5)
12.8
-127%
129%
Dead Wood
5.5
(1.6)
13.3
-129%
141%
Litter
8.1
(2.3)
19.4
-129%
140%
Mineral Soil C Stocks: Tier 3
0.8
(38.5)
40.2
-4767%
4767%
Mineral Soil C Stocks: Tier 2
2.2
(1.1)
5.5
-151%
151%
Organic Soil C Stocks: Tier 2
3.7
1.2
6.3
-68%
68%
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Uncertainty is also associated with lack of reporting of agricultural biomass and dead organic matter C stock
changes. Biomass C stock changes are likely minor in perennial crops, such as orchards and nut plantations, given
the small amount of change in land that is used to produce these commodities in the United States. In contrast,
agroforestry practices, such as shelterbelts, riparian forests and intercropping with trees, may have led to larger
changes in biomass C stocks at least in some regions of the United States. However, there are currently no datasets
to evaluate the trends. Changes in dead organic matter C stocks are assumed to be negligible with conversion of
land to croplands with the exception of forest lands, which are included in this analysis. This assumption will be
further explored in a future Inventory.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
See the QA/QC and Verification section in Cropland Remaining Cropland for information on QA/QC steps.
Recalculations Discussion
Differences in biomass, dead wood and litter C stock changes in Forest Land Converted to Cropland can be
attributed to incorporation of the latest FIA data for 1990 to 2019. As a result of these new data, Land Converted
to Cropland has a marginally smaller reported loss of C compared to the previous Inventory, estimated at an
average of 1.7 MMT C02 Eq. over the time series. This represents a 3 percent decline in losses of C for Land
Converted to Cropland compared to the previous Inventory.
Planned Improvements
Planned improvements are underway to include an analysis of C stock changes in Alaska for cropland, using the
Tier 2 method for mineral and organic soils that is described earlier in this section. This analysis will initially focus
on land use change, which typically has a larger impact on soil organic C stock changes than management
practices, but will be further refined over time to incorporate management data that drive C stock changes on
long-term cropland. See Table 6-35 for the amount of managed area in Land Converted to Cropland that is not
included in the Inventory, which is less than one thousand hectares per year. This includes the area in Alaska and
other miscellaneous cropland areas, such as aquaculture. Additional planned improvements are discussed in the
Planned Improvements section of Cropland Remaining Cropland.
6-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 6-35: Area of Managed Land in Land Converted to Cropland that is not included in the
current Inventory (Thousand Hectares)
Area (Thousand Hectares)
Year
Managed Land
Inventory
Not Included in
Inventory
1990
12,308
12,308
<1
1991
12,654
12,654
<1
1992
12,943
12,943
<1
1993
14,218
14,218
<1
1994
15,400
15,400
<1
1995
15,581
15,581
<1
1996
15,888
15,888
<1
1997
16,073
16,073
<1
1998
17,440
17,440
<1
1999
17,819
17,819
<1
2000
17,693
17,693
<1
2001
17,600
17,600
<1
2002
17,487
17,487
<1
2003
16,257
16,257
<1
2004
15,317
15,317
<1
2005
15,424
15,424
<1
2006
15,410
15,410
<1
2007
14,923
14,923
<1
2008
14,399
14,399
<1
2009
13,814
13,814
<1
2010
13,905
13,905
<1
2011
14,186
14,186
<1
2012
14,429
14,429
<1
2013
13,752
13,752
<1
2014
13,050
13,050
<1
2015
13,049
13,049
<1
2016
ND
ND
ND
2017
ND
ND
ND
2018
ND
ND
ND
2019
ND
ND
ND
Note: NRI data are not available after 2015 so these years are designated as ND (No data).
6.6 Grassland Remaining Grassland (CRF
Category 4C1)
Carbon (C) in grassland ecosystems occurs in biomass, dead organic matter, and soils. Soils are the largest pool of C
in grasslands, and have the greatest potential for longer-term storage or release of C. Biomass and dead organic
matter C pools are relatively ephemeral compared to the soil C pool, with the exception of C stored in tree and
shrub biomass that occurs in grasslands. The 2006IPCC Guidelines recommend reporting changes in biomass, dead
organic matter and soil organic C stocks with land use and management. C stock changes for aboveground and
belowground biomass, dead wood and litter pools are reported for woodlands (i.e., a subcategory of grasslands),
and may be extended to include agroforestry management associated with grasslands in the future. For soil
Land Use, Land-Use Change, and Forestry 6-71

-------
organic C, the 2006IPCCGuidelines (IPCC 2006) recommend reporting changes due to (1) agricultural land-use and
management activities on mineral soils, and (2) agricultural land-use and management activities on organic soils.49
Grassland Remaining Grassland includes all grassland in an Inventory year that had been grassland for a
continuous time period of at least 20 years (USDA-NRCS 2018). Grassland includes pasture and rangeland that are
primarily, but not exclusively used for livestock grazing. Rangelands are typically extensive areas of native
grassland that are not intensively managed, while pastures are typically seeded grassland (possibly following tree
removal) that may also have additional management, such as irrigation or interseeding of legumes. Woodlands are
also considered grassland and are areas of continuous tree cover that do not meet the definition of forest land
(See Land Representation section for more information about the criteria for forest land). The current Inventory
includes all grasslands in the conterminous United States and Hawaii, but does not include approximately 50
million hectares of Grassland Remaining Grassland in Alaska. This leads to a discrepancy with the total amount of
managed area in Grassland Remaining Grassland (see Table 6-39 in Planned Improvements for more details on the
land area discrepancies) and the grassland area included in the Inventory analysis.
In Grassland Remaining Grassland, there has been considerable variation in C stocks between 1990 and 2019.
These changes are driven by variability in weather patterns and associated interaction with land management
activity. Moreover, changes are small on a per hectare rate basis across the time series even in the years with a
larger total change in stocks. The net change in total C stocks for 2019 led to net C02 emissions to the atmosphere
of 14.5 MMT C02 Eq. (4.0 MMT C), including 1.3 MMT C02 Eq. (0.4 MMT C) from net losses of aboveground
biomass C, 0.1 MMT C02 Eq. (<0.05 MMT C) from net losses in belowground biomass C, 2.3 MMT C02 Eq. (0.6
MMT C) from net losses in dead wood C, 0.2 MMT C02 Eq. (<0.05 MMT C) from net gains in litter C, 5.5 MMT C02
Eq. (1.5 MMT C) from net losses in mineral soil organic C, and 5.4 MMT C02 Eq. (1.5 MMT C) from losses of C due
to drainage and cultivation of organic soils (Table 6-36 and Table 6-37). Losses of carbon are 74 percent higher in
2019 compared to 1990, but as noted previously, stock changes are highly variable from 1990 to 2019, with an
average annual change of 8.7 MMT C02 Eq. (2.4 MMT C).
Table 6-36: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Grassland Remaining Grassland (MMT CO2 Eq.)
Soil Type
1990
2005
2015
2016
2017
2018
2019
Aboveground Live Biomass
1.4
1.4
1.4
1.4
1.4
1.3
1.3
Belowground Live Biomass
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Dead Wood
2.8
2.7
2.4
2.4
2.4
2.4
2.3
Litter
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Mineral Soils
(2.2)
0.8
4.0
0.7
2.2
2.7
5.5
Organic Soils
6.3
5.2
5.4
5.4
5.4
5.4
5.4
Total Net Flux
8.3
10.0
13.1
9.8
11.3
11.7
14.5
Table 6-37: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Grassland Remaining Grassland (MMT C)
Soil Type
1990
2005
2015
2016
2017
2018
2019
Aboveground Live Biomass
0.4
0.4
0.4
0.4
0.4
0.4
0.4
Belowground Live Biomass
+
+
+
+
+
+
+
Dead Wood
0.8
0.7
0.7
0.7
0.7
0.6
0.6
Litter
+
+
+
+
+
+
+
Mineral Soils
(0.6)
0.2
1.1
0.2
0.6
0.7
1.5
Organic Soils
1.7
1.4
1.5
1.5
1.5
1.5
1.5
Total Net Flux
2.3
2.7
3.6
2.7
3.1
3.2
4.0
49 C02 emissions associated with liming and urea fertilization are also estimated but included in the Agriculture chapter of the
report.
6-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The spatial variability in soil organic C stock changes for 201550 is displayed in Figure 6-8 for mineral soils and in
Figure 6-9 for organic soils. Although relatively small on a per-hectare basis, grassland soils gained C in isolated
areas that mostly occurred in pastures of the eastern United States. For organic soils, the regions with the highest
rates of emissions coincide with the largest concentrations of organic soils used for managed grassland, including
the Southeastern Coastal Region (particularly Florida), upper Midwest and Northeast, and a few isolated areas
along the Pacific Coast,
Figure 6-8: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural
Management within States, 2015, Grassland Remaining Grassland
Note: Only national-scale soil organic C stock changes are estimated for 2016 to 2019 in the current Inventory using a
surrogate data method, and therefore the fine-scale emission patterns in this map are based on inventory data from
2015. Negative values represent a net increase in soil organic C stocks, and positive values represent a net decrease in
soil organic C stocks.
50 Only national-scale emissions are estimated for 2016 to 2019 in the current Inventory using the surrogate data method, and
therefore the fine-scale emission patterns in this map are based on inventory data from 2015.
Land Use, Land-Use Change, and Forestry 6-73

-------
Figure 6-9: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural
Management within States, 2015, Grassland Remaining Grassland
¦ > 40
Note: Only national-scale soil organic carbon stock changes are estimated for 2016 to 2019 in the current Inventory
using a surrogate data method, and therefore the fine-scale emission patterns in this map are based on inventory
data from 2015.
Methodology
The following section includes a description of the methodology used to estimate C stock changes for Grassland
Remaining Grassland, including (1) aboveground and belowground biomass, dead wood and litter C for woodlands,
as well as (2) soil organic C stocks for mineral and organic soils.
Biomass, Dead Wood and Litter Carbon Stock Changes
The methodology is consistent with IPCC (2006). Woodlands are lands that do not meet the definition of forest
land or agroforestry (see Section 6 Representation of the U.S. Land Base), but include woody vegetation with C
storage in aboveground and belowground biomass, dead wood and litter C (IPCC 2006) as described in the Forest
Land Remaining Forest Land section. Carbon stocks and net annual C stock change were determined according to
the stock-difference method for the CONUS, which involved applying C estimation factors to annual forest
inventories across time to obtain C stocks and then subtracting the values between years to estimate the stock
changes. The methods for estimating carbon stocks and stock changes for woodlands in Grassland Land Remaining
Grassland are consistent with those in the Forest Land Remaining Forest Land section and are described in Annex
3.13. All annual National Forest Inventory (NFI) plots available in the public FIA database (USDA Forest Service
2020) were used in the current Inventory. While the NFI is an all-lands inventory, only those plots that meet the
definition of forest land are typically measured. However, in some cases, particularly in the Central Plains and
Southwest United States, woodlands have been measured as part of the survey. This analysis is limited to those
plots and is not considered a comprehensive assessment of trees outside of forest land that meet the definition of
grassland.
6-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Soil Carbon Stock Changes
The following section includes a brief description of the methodology used to estimate changes in soil organic C
stocks for Grassland Remaining Grassland, including: (1) agricultural land-use and management activities on
mineral soils; and (2) agricultural land-use and management activities on organic soils. Further elaboration on the
methodologies and data used to estimate stock changes from mineral and organic soils are provided in the
Cropland Remaining Cropland section and Annex 3.12.
Soil organic C stock changes are estimated for Grassland Remaining Grassland on non-federal lands according to
land use histories recorded in the 2015 USDA NRI survey (USDA-NRCS 2018). Land-use and some management
information (e.g., grass type, soil attributes, and irrigation) were originally collected for each NRI survey location
on a 5-year cycle beginning in 1982. In 1998, the NRI program began collecting annual data, and the annual data
are currently available through 2015 (USDA-NRCS 2015). NRI survey locations are classified as Grassland Remaining
Grassland in a given year between 1990 and 2015 if the land use had been grassland for 20 years. NRI survey
locations are classified according to land-use histories starting in 1979, and consequently the classifications are
based on less than 20 years from 1990 to 1998. This may have led to an overestimation of Grassland Remaining
Grassland in the early part of the time series to the extent that some areas are converted to grassland between
1971 and 1978. For federal lands, the land use history is derived from land cover changes in the National Land
Cover Dataset (Yang et al. 2018; Homer et al. 2007; Fry et al. 2011; Homer et al. 2015).
Mineral Soil Carbon Stock Changes
An IPCC Tier 3 model-based approach (Ogle et al. 2010) is applied to estimate C stock changes from 1990 to 2015
for most mineral soils in Grassland Remaining Grassland. The C stock changes for the remaining soils are estimated
with an IPCC Tier 2 method (Ogle et al. 2003), including gravelly, cobbly, or shaley soils (greater than 35 percent by
volume), the additional stock changes associated with biosolids (i.e., treated sewage sludge) amendments, and
federal land.51
A surrogate data method is used to estimate soil organic C stock changes from 2016 to 2019 at the national scale
for land areas included in the Tier 2 and Tier 3 methods. Specifically, linear regression models with autoregressive
moving-average (ARMA) errors (Brockwell and Davis 2016) are used to estimate the relationship between
surrogate data and the 1990 to 2015 emissions data from the Tier 2 and 3 methods. Surrogate data for these
regression models are based on weather data from the PRISM Climate Group (PRISM Climate Group 2018). See
Box 6-4 in the Methodology section of Cropland Remaining Cropland for more information about the surrogate
data method. Stock change estimates for 2016 to 2019 will be recalculated in future inventories when new NRI
data are available.
Tier 3 Approach. Mineral soil organic C stocks and stock changes for Grassland Remaining Grassland are estimated
using the DayCent biogeochemical52 model (Parton et al. 1998; Del Grosso et al. 2001, 2011), as described in
Cropland Remaining Cropland. The DayCent model utilizes the soil C modeling framework developed in the
Century model (Parton et al. 1987,1988,1994; Metherell et al. 1993), but has been refined to simulate dynamics
at a daily time-step. Historical land-use patterns and irrigation histories are simulated with DayCent based on the
2015 USDA NRI survey (USDA-NRCS 2018).
The amount of manure produced by each livestock type is calculated for managed and unmanaged waste
management systems based on methods described in Section 5.2 Manure Management and Annex 3.11. Manure N
deposition from grazing animals (i.e., PRP manure) is an input to the DayCent model to estimate the influence of
PRP manure on C stock changes for lands included in the Tier 3 method. Carbon stocks and 95 percent confidence
51	Federal land is not a land use, but rather an ownership designation that is treated as grassland for purposes of these
calculations. The specific land use on federal lands is not identified in the NRI survey (USDA-NRCS 2015).
52	Biogeochemical cycles are the flow of chemical elements and compounds between living organisms and the physical
environment.
Land Use, Land-Use Change, and Forestry 6-75

-------
intervals are estimated for each year between 1990 and 2015 using the NRI survey data. Further elaboration on
the Tier 3 methodology and data used to estimate C stock changes from mineral soils are described in Annex 3.12.
Soil organic C stock changes from 2016 to 2019 were estimated using a surrogate data method described in Box
6-4 of the Methodology section in Cropland Remaining Cropland. Future Inventories will be updated with new
activity data when the data are made available, and the time series will be recalculated (see the Planned
Improvements section in Cropland Remaining Cropland).
Tier 2 Approach. The Tier 2 approach is based on the same methods described in the Tier 2 portion of Cropland
Remaining Cropland section for mineral soils, with the exception of the manure N deposition from grazing animals
(i.e., PRP manure), and the land use and management data that are used in the Inventory for federal grasslands.
First, the PRP N manure is included in the Tier 2 method that is not deposited on lands included in the Tier 3
method. Second, the NRI (USDA-NRCS 2018) provides land use and management histories for all non-federal lands,
and is the basis for the Tier 2 analysis for these areas. However, NRI does not provide land use information on
federal lands. The land use data for federal lands is based on the National Land Cover Database (NLCD) (Yang et al.
2018; Fry et al. 2011; Homer et al. 2007; Homer et al. 2015). In addition, the Bureau of Land Management (BLM)
manages some of the federal grasslands, and compiles information on grassland condition through the BLM
Rangeland Inventory (BLM 2014). To estimate soil organic C stock changes from federal grasslands, rangeland
conditions in the BLM data are aligned with IPCC grassland management categories of nominal, moderately
degraded, and severely degraded in order to apply the appropriate emission factors. Further elaboration on the
Tier 2 methodology and data used to estimate C stock changes from mineral soils are described in Annex 3.12.
The time series of stock changes for non-federal and federal lands has been extended from 2016 to 2019 using a
surrogate data method described in Box 6-4 of the Methodology section in Cropland Remaining Cropland.
Additional Mineral C Stock Change Calculations
A Tier 2 method is used to adjust annual C stock change estimates for mineral soils between 1990 and 2019 to
account for additional C stock changes associated with biosolids (i.e., treated sewage sludge) amendments.
Estimates of the amounts of biosolids N applied to agricultural land are derived from national data on biosolids
generation, disposition, and N content (see Section 7.2, Wastewater Treatment for a detailed discussion of the
methodology for estimating treated sewage sludge available for land application application). Although biosolids
can be added to land managed for other land uses, it is assumed that agricultural amendments only occur in
Grassland Remaining Grassland. Total biosolids generation data for 1988,1996, and 1998, in dry mass units, are
obtained from EPA (1999) and estimates for 2004 are obtained from an independent national biosolids survey
(NEBRA 2007). These values are linearly interpolated to estimate values for the intervening years, and linearly
extrapolated to estimate values for years since 2004. Nitrogen application rates from Kellogg et al. (2000) are used
to determine the amount of area receiving biosolids amendments. The soil organic C storage rate is estimated at
0.38 metric tons C per hectare per year for biosolids amendments to grassland as described above. The stock
change rate is based on country-specific factors and the IPCC default method (see Annex 3.12 for further
discussion).
Organic Soil Carbon Stock Changes
Annual C emissions from drained organic soils in Grassland Remaining Grassland are estimated using the Tier 2
method in IPCC (2006), which utilizes country-specific C loss rates (Ogle et al. 2003) rather than default IPCC rates.
For more information, see the Cropland Remaining Cropland section for organic soils and Annex 3.12.
A surrogate data method was used to estimate annual C emissions from organic soils from 2016 to 2019 as
described in Box 6-4 of the Methodology section in Cropland Remaining Cropland. Estimates for 2016 to 2019 will
be updated in future Inventories when new NRI data are available.
6-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Uncertainty and Time-Series Consistency
The uncertainty analysis for biomass, dead wood and litter C losses with Forest Land Converted to Cropland is
conducted in the same way as the uncertainty assessment for forest ecosystem C flux associated with Forest Land
Remaining Forest Land. Sample and model-based error are combined using simple error propagation methods
provided by the IPCC (2006) by taking the square root of the sum of the squares of the standard deviations of the
uncertain quantities. For additional details, see the Uncertainty Analysis in Annex 3.13.
Uncertainty analysis for mineral soil organic C stock changes using the Tier 3 and Tier 2 methodologies are based
on a Monte Carlo approach that is described in the Cropland Remaining Cropland section and Annex 3.12. The
uncertainty for annual C emission estimates from drained organic soils in Grassland Remaining Grassland is
estimated using a Monte Carlo approach, which is also described in the Cropland Remaining Cropland section. For
2016 to 2019, there is additional uncertainty propagated through the Monte Carlo Analysis associated with the
surrogate data method.
Uncertainty estimates are presented in Table 6-38 for each subsource (i.e., soil organic C stocks for mineral and
organic soils) and the method applied in the Inventory analysis (i.e., Tier 2 and Tier 3). Uncertainty estimates from
the Tier 2 and 3 approaches are combined using the simple error propagation methods provided by the IPCC
(2006), i.e., by taking the square root of the sum of the squares of the standard deviations of the uncertain
quantities.
The combined uncertainty for soil organic C stocks in Grassland Remaining Grassland ranges from more than 1,066
percent below and above the 2019 stock change estimate of 14.5 MMT C02 Eq. The large relative uncertainty is
mostly due to variation in soil organic C stock changes that is not explained by the surrogate data method, leading
to high prediction error with this splicing method.
Table 6-38: Approach 2 Quantitative Uncertainty Estimates for C Stock Changes Occurring
Within Grassland Remaining Grassland (MMT CO2 Eq. and Percent)
Source
2019 Flux Estimate
Uncertainty Range Relative to Flux Estimate3
(MMT C02 Eq.)
(MMT CO;
2 Eq.)
(%)



Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Woodland Biomass:





Aboveground live biomass
1.3
0.9
1.7
-31%
31%
Belowground live biomass
0.1
0.1
0.1
16%
16%
Dead wood
2.3
1.8
2.8
-22%
22%
Litter
(0.2)
(0.3)
+
-105%
105%
Mineral Soil C Stocks Grassland Remaining





Grassland, Tier 3 Methodology
5.7
(148.9)
160.3
-2,712%
2,712%
Mineral Soil C Stocks: Grassland Remaining





Grassland, Tier 2 Methodology
+
(0.9)
0.9
-5,287%
5,287%
Mineral Soil C Stocks: Grassland Remaining





Grassland, Tier 2 Methodology (Change in Soil





C due to Biosolids [i.e., Treated Sewage





Sludge] Amendments)
(0.2)
(0.3)
(0.1)
-50%
50%
Organic Soil C Stocks: Grassland Remaining





Grassland, Tier 2 Methodology
5.4
1.2
9.6
-77%
77%
Combined Uncertainty for Flux Associated





with Carbon Stock Changes Occurring in





Grassland Remaining Grassland
14.5
(140.2)
169.2
-1,066%
1,066%
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Land Use, Land-Use Change, and Forestry 6-77

-------
Uncertainty is also associated with a lack of reporting on biomass, dead wood and litter C stock changes for
agroforestry systems. Changes in biomass and dead organic matter C stocks are assumed to be negligible in other
grasslands, largely comprised of herbaceous biomass, although there are certainly significant changes at sub-
annual time scales across seasons.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
Recalculations are associated with new FIA data from 1990 to 2018 on biomass, dead wood and litter C stocks in
woodlands, and updated estimates for mineral soils from 2016 to 2018 using revised surrogate data. As a result of
these new data, Grassland Remaining Grassland has a small loss of C compared to the previous Inventory,
estimated at an average reduction in losses of 0.59 MMT C02 Eq. over the time series. This represents a 14 percent
decrease in losses of C for Grassland Remaining Grassland compared to the previous Inventory.
Grasslands in Alaska are not currently included in the Inventory. This is a significant planned improvement and
estimates are expected to be available in a future Inventory contingent on funding availability. Table 6-39 provides
information on the amount of managed area in Alaska that is Grassland Remaining Grassland, which includes
about 50 million hectares per year. For information about other improvements, see the Planned Improvements
section in Cropland Remaining Cropland.
Table 6-39: Area of Managed Land in Grassland Remaining Grassland's Alaska that is not
included in the current Inventory (Thousand Hectares)
QA/QC and Verification
See the QA/QC and Verification section in Cropland Remaining Cropland.
Recalculations Discussion
Planned Improvements
Area (Thousand Hectares)
Year
Not Included in
Managed Land Inventory	Inventory
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
327,446	277,406	50,040
326,959	276,918	50,040
326,462	276,422	50,040
324,524	274,484	50,040
322,853	272,813	50,040
322,015	271,975	50,040
321,164	271,123	50,040
320,299	270,259	50,040
318,214	268,174	50,040
317,341	267,301	50,040
316,242	266,202	50,040
315,689	265,649	50,040
315,232	265,192	50,040
315,442	265,403	50,039
315,459	265,421	50,038
315,161	265,123	50,038
314,841	264,804	50,037
6-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2007
314,786
264,749
50,036
2008
314,915
264,878
50,037
2009
315,137
265,099
50,037
2010
314,976
264,942
50,035
2011
314,662
264,627
50,035
2012
314,466
264,413
50,053
2013
315,301
265,239
50,062
2014
316,242
266,180
50,062
2015
316,287
266,234
50,053
2016
ND
ND
ND
2017
ND
ND
ND
2018
ND
ND
ND
2019
ND
ND
ND
Note: NRI data are not available after 2015, and these years are designated as ND (No data).
Additionally, a review of available data on biosolids (i.e., treated sewage sludge) application will be undertaken to
improve the distribution of biosolids application on croplands, grasslands and settlements.
N011-CO2 Emissions from Grassland Fires (CRF Source Category
4C1)
Fires are common in grasslands, and are thought to have been a key feature shaping the evolution of the grassland
vegetation in North America (Daubenmire 1968; Anderson 2004). Fires can occur naturally through lightning
strikes, but are also an important management practice to remove standing dead vegetation and improve forage
for grazing livestock. Woody and herbaceous biomass will be oxidized in a fire, although in this section the current
focus is primarily on herbaceous biomass.53 Biomass burning emits a variety of trace gases including non-C02
greenhouse gases such as CH4 and N20, as well as CO and NOx that can become greenhouse gases when they react
with other gases in the atmosphere (Andreae and Merlet 2001). IPCC (2006) recommends reporting non-C02
greenhouse gas emissions from all wildfires and prescribed burning occurring in managed grasslands.
Biomass burning in grassland of the United States (Including burning emissions in Grassland Remaining Grassland
and Land Converted to Grassland) is a relatively small source of emissions, but it has increased by over 300 percent
since 1990. In 2019, CH4 and N20 emissions from biomass burning in grasslands were 0.3 MMT C02 Eq. (12 kt) and
0.3 MMT C02 Eq. (1 kt), respectively. Annual emissions from 1990 to 2019 have averaged approximately 0.3 MMT
C02 Eq. (12 kt) of CH4 and 0.3 MMT C02 Eq. (1 kt) of N20 (see Table 6-40 and Table 6-41).
Table 6-40: ChU and N2O Emissions from Biomass Burning in Grassland (MMT CO2 Eq.)
1990

2005

2015 2016 2017 2018 2019
CH4 0.1
N20 0.1

0.3
0.3

0.3 0.3 0.3 0.3 0.3
0.3 0.3 0.3 0.3 0.3
Total Net Flux 0.2

0.7

0.7 0.6 0.6 0.6 0.6
Table 6-41: ChU, N2O, CO, and NOx Emissions from Biomass Burning in Grassland (kt)
1990

2005

2015
2016
2017
2018
2019
ch4
3

13

13
12
12
12
12
n2o
+

1

1
1
1
1
1
CO
84

358

356
324
345
331
341
NOx
5

22

21
20
21
20
20
+ Does not exceed 0.5 kt.
53 A planned improvement is underway to incorporate woodland tree biomass into the Inventory.
Land Use, Land-Use Change, and Forestry 6-79

-------
Methodology
The following section includes a description of the methodology used to estimate non-C02 greenhouse gas
emissions from biomass burning in grassland, including (1) determination of the land base that is classified as
managed grassland; (2) assessment of managed grassland area that is burned each year, and (3) estimation of
emissions resulting from the fires. For this Inventory, the IPCC Tier 1 method is applied to estimate non-C02
greenhouse gas emissions from biomass burning in grassland from 1990 to 2014 (IPCC 2006). A data splicing
method is used to estimate the emissions in 2015 to 2019, which is discussed later in this section.
The land area designated as managed grassland is based primarily on the National Resources Inventory (NRI)
(Nusser and Goebel 1997; USDA-NRCS 2015). NRI has survey locations across the entire United States, but does not
classify land use on federally-owned areas, and so survey locations on federal lands are designated as grassland
using land cover data from the National Land Cover Dataset (NLCD) (Fry et al. 2011; Homer et al. 2007; Homer et
al. 2015) (see Section 6 Representation of the U.S. Land Base).
The area of biomass burning in grasslands (Grassland Remaining Grassland and Land Converted to Grassland) is
determined using 30-m fire data from the Monitoring Trends in Burn Severity (MTBS) program for 1990 through
2014.54 NRI survey locations on grasslands are designated as burned in a year if there is a fire within a 500 m of the
survey point according to the MTBS fire data. The area of biomass burning is estimated from the NRI spatial
weights and aggregated to the country (Table 6-42).
Table 6-42: Thousands of Grassland Hectares Burned Annually

Thousand
Year
Hectares
1990
317
2005
1,343
2014
1,659
2015
NE
2016
NE
2017
NE
2018
NE
2019
NE
Notes: Burned area was not
estimated (NE) for 2015 to 2019
but will be updated in a future
Inventory.
For 1990 to 2014, the total area of grassland burned is multiplied by the IPCC default factor for grassland biomass
(4.1 tonnes dry matter per ha) (IPCC 2006) to estimate the amount of combusted biomass. A combustion factor of
1 is assumed in this Inventory, and the resulting biomass estimate is multiplied by the IPCC default grassland
emission factors for CH4 (2.3 g CH4 per kg dry matter), N20 (0.21 g CH4 per kg dry matter), CO (65 g CH4 per kg dry
matter) and NOx (3.9 g CH4 per kg dry matter) (IPCC 2006). The Tier 1 analysis is implemented in the Agriculture
and Land Use National Greenhouse Gas Inventory (ALU) software (Ogle et al. 2016).55
A linear extrapolation of the trend in the time series is applied to estimate the emissions for 2015 to 2019 because
new activity data have not been compiled for these years. Specifically, a linear regression model with
54	See .
55	See .
6-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) is used to estimate the trend in
emissions over time from 1990 to 2014, and the trend is used to approximate the 2015 to 2019 emissions. The Tier
1 method described previously will be applied to recalculate the 2015 to 2019 emissions in a future Inventory.
Uncertainty and Time-Series Consistency
Emissions are estimated using a linear regression model with ARMA errors for 2015 to 2019. The linear regression
ARMA model produced estimates of the upper and lower bounds of the emission estimate and the results are
summarized in Table 6-43. Methane emissions from Biomass Burning in Grassland for 2018 are estimated to be
between approximately 0.0 and 0.7 MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 100
percent below and 146 percent above the 2019 emission estimate of 0.3 MMT C02 Eq. Nitrous oxide emissions are
estimated to be between approximately 0.0 and 0.8 MMT C02 Eq., or approximately 100 percent below and 146
percent above the 2019 emission estimate of 0.3 MMT C02 Eq.
Table 6-43: Uncertainty Estimates for Non-C02 Greenhouse Gas Emissions from Biomass
Burning in Grassland (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)

Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Grassland Burning
Grassland Burning
ch4
n2o
0.3
0.3
+ 0.7
+ 0.8
-100%
-100%
146%
146%
+ Does not exceed 0.05 MMT C02 Eq.
a Range of emission estimates predicted by linear regression time-series model for a 95 percent confidence interval.
Uncertainty is also associated with lack of reporting of emissions from biomass burning in grassland of Alaska.
Grassland burning emissions could be relatively large in this region of the United States, and therefore extending
this analysis to include Alaska is a planned improvement for the Inventory. There is also uncertainty due to lack of
reporting combustion of woody biomass, and this is another planned improvement.
There were no methodological recalculations in this Inventory, but data splicing methods to extend the time series
for another year were applied in a manner to be consistent with the previous Inventory. Details on the
emission/removal trends and methodologies through time are described in more detail in the Introduction and
Methodology sections.
QA/QC and Verification
Quality control measures included checking input data, model scripts, and results to ensure data are properly
handled throughout the inventory process. Inventory reporting forms and text are reviewed and revised as needed
to correct transcription errors. Quality control identified problems with input data for common reporting format
tables in the spreadsheets, which have been corrected.
Recalculations Discussion
There are no recalculations in the time series from the previous Inventory.
Planned Improvements
A splicing data method is applied to estimate emissions in the latter part of the time series, which introduces
additional uncertainty in the emissions data. Therefore, a key improvement for the next Inventory will be to
update the time series with new activity data from the Monitoring Trends in Burn Severity program and recalculate
the emissions. Two other planned improvements have been identified for this source category, including a)
Land Use, Land-Use Change, and Forestry 6-81

-------
incorporation of country-specific grassland biomass factors, and b) extending the analysis to include Alaska. In the
current Inventory, biomass factors are based on a global default for grasslands that is provided by the IPCC (2006).
There is considerable variation in grassland biomass, however, which would affect the amount of fuel available for
combustion in a fire. Alaska has an extensive area of grassland and includes tundra vegetation, although some of
the areas are not managed. There has been an increase in fire frequency in boreal forest of the region (Chapin et
al. 2008), and this may have led to an increase in burning of neighboring grassland areas. There is also an effort
under development to incorporate grassland fires into DayCent model simulations. Both improvements are
expected to reduce uncertainty and produce more accurate estimates of non-C02 greenhouse gas emissions from
grassland burning.
6.7 Land Converted to Grassland (CRF Category
4C2)	
Land Converted to Grassland includes all grassland in an Inventory year that had been in another land use(s) during
the previous 20 years (USDA-NRCS 20 18).56 For example, cropland or forest land converted to grassland during the
past 20 years would be reported in this category. Recently converted lands are retained in this category for 20
years as recommended by IPCC (2006). Grassland includes pasture and rangeland that are used primarily but not
exclusively for livestock grazing. Rangelands are typically extensive areas of native grassland that are not
intensively managed, while pastures are typically seeded grassland (possibly following tree removal) that may also
have additional management, such as irrigation or interseeding of legumes. This Inventory includes all grasslands
in the conterminous United States and Hawaii, but does not include Land Converted to Grassland in Alaska.
Consequently, there is a discrepancy between the total amount of managed area for Land Converted to Grassland
(see Table 6-47 in Planned Improvements) and the grassland area included in the inventory analysis.
Land use change can lead to large losses of C to the atmosphere, particularly conversions from forest land
(Houghton et al. 1983, Houghton and Nassikas 2017). Moreover, conversion of forest to another land use (i.e.,
deforestation) is one of the largest anthropogenic sources of emissions to the atmosphere globally, although this
source may be declining according to a recent assessment (Tubiello et al. 2015).
IPCC (2006) recommends reporting changes in biomass, dead organic matter, and soil organic C stocks due to land
use change. All soil organic C stock changes are estimated and reported for Land Converted to Grassland, but there
is limited reporting of other pools in this Inventory. Losses of aboveground and belowground biomass, dead wood
and litter C from Forest Land Converted to Grassland are reported, but these C stock changes are not estimated for
other land use conversions to grassland.57
The largest C losses with Land Converted to Grassland are associated with aboveground biomass, belowground
biomass, and litter C losses from Forest Land Converted to Grassland (see Table 6-44 and Table 6-45). These three
pools led to net emissions in 2019 of 8.6, 2.1, and 4.6 MMT C02 Eq. (2.4,0.6, and 1.3 MMT C), respectively. Land
use and management of mineral soils in Land Converted to Grassland led to an increase in soil organic C stocks,
estimated at 39.8 MMT C02 Eq. (10.9 MMT C) in 2019. The gains are primarily associated with conversion of Other
Land, which have relatively low soil organic C stocks, to Grassland that tend to have conditions suitable for storing
larger amounts of C in soils, and also due to conversion of Cropland to Grassland that leads to less intensive
management of the soil. Drainage of organic soils for grassland management led to C02 emissions to the
56	NRI survey locations are classified according to land-use histories starting in 1979, and consequently the classifications are
based on less than 20 years from 1990 to 2001. This may have led to an underestimation of Land Converted to Grassland in the
early part of the time series to the extent that some areas are converted to grassland between 1971 and 1978.
57	Changes in biomass C stocks are not currently reported for other conversions to grassland (other than forest land), but this is
a planned improvement for a future Inventory. Note: changes in dead organic matter are assumed to negligible for other land
use conversions (i.e., other than forest land) to grassland based on the Tier 1 method in IPCC (2006).
6-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
atmosphere of 1.8 MMT C02 Eq. (0.5 MMT C). The total net C stock change in 2019 for Land Converted to
Grassland is estimated as a gain of 23.2 MMT C02 Eq. (6.3 MMT C), which represents an increase in C stock
changes of 271 percent compared to the initial reporting year of 1990.
Table 6-44: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Grassland (MMT CO2 Eq.)

1990
2005
2015
2016
2017
2018
2019
Cropland Converted to







Grassland
(18.3)
(23.5)
(15.5)
(19.9)
(20.2)
(20.3)
(19.8)
Mineral Soils
(18.9)
(25.0)
(16.9)
(21.3)
(21.6)
(21.6)
(21.1)
Organic Soils
0.6
1.5
1.4
1.4
1.4
1.3
1.3
Forest Land Converted to







Grassland
16.3
16.3
15.1
14.8
14.9
14.9
14.8
Aboveground Live Biomass
9.9
9.6
8.8
8.6
8.6
8.6
8.6
Belowground Live Biomass
2.4
2.4
2.2
2.1
2.1
2.1
2.1
Dead Wood
(0.7)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
Litter
4.8
4.8
4.6
4.6
4.6
4.6
4.6
Mineral Soils
(0.1)
(0.1)
(0.1)
(0.2)
(0.1)
(0.2)
(0.2)
Organic Soils
+
0.2
0.2
0.2
0.2
0.2
0.2
Other Lands Converted to







Grassland
(4.2)
(31.7)
(22.8)
(18.3)
(18.3)
(18.1)
(17.6)
Mineral Soils
(4.2)
(31.7)
(22.9)
(18.4)
(18.4)
(18.2)
(17.7)
Organic Soils
+
+
0.1
0.1
0.1
0.1
0.1
Settlements Converted to







Grassland
(0.2)
(1.4)
(1.0)
(0.9)
(0.9)
(0.9)
(0.8)
Mineral Soils
(0.2)
(1.4)
(1.0)
(0.9)
(0.9)
(0.9)
(0.8)
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted to







Grassland
0.1
0.2
0.3
0.2
0.2
0.2
0.2
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
0.1
0.2
0.3
0.3
0.2
0.2
0.2
Aboveground Live Biomass
9.9
9.6
8.8
8.6
8.6
8.6
8.6
Belowground Live Biomass
2.4
2.4
2.2
2.1
2.1
2.1
2.1
Dead Wood
(0.7)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
Litter
4.8
4.8
4.6
4.6
4.6
4.6
4.6
Total Mineral Soil Flux
(23.4)
(58.2)
(40.8)
(40.7)
(41.1)
(40.8)
(39.8)
Total Organic Soil Flux
0.8
1.9
1.9
1.9
1.9
1.9
1.8
Total Net Flux
(6.2)
(40.1)
(23.9)
(24.0)
(24.4)
(24.1)
(23.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
Table 6-45: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Grassland (MMT C)
	1990	2005	2015 2016 2017 2018 2019
Cropland Converted to
Grassland
Mineral Soils
Organic Soils
Forest Land Converted to
Grassland
Aboveground Live Biomass
Belowground Live Biomass
Dead Wood
Litter
Mineral Soils
Organic Soils
Land Use, Land-Use Change, and Forestry 6-83
(5.0)
(6.4)
(4.2)
(5.4)
(5.5)
(5.5)
(5.4)
(5.2)
(6.8)
(4.6)
(5.8)
(5.9)
(5.9)
(5.8)
0.2
0.4
0.4
0.4
0.4
0.4
0.4
4.4
4.4
4.1
4.0
4.1
4.1
4.0
2.7
2.6
2.4
2.4
2.4
2.4
2.4
0.7
0.6
0.6
0.6
0.6
0.6
0.6
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
1.3
1.3
1.3
1.3
1.3
1.3
1.3
+
+
+
+
+
+
+
+
+
0.1
0.1
0.1
0.1
0.1

-------
Other Lands Converted to







Grassland
(1.1)
(8.6)
(6.2)
(5.0)
(5.0)
(4.9)
(4.8)
Mineral Soils
(1.2)
(8.6)
(6.3)
(5.0)
(5.0)
(5.0)
(4.8)
Organic Soils
+
+
+
+
+
+
+
Settlements Converted to







Grassland
+
(0.4)
(0.3)
(0.2)
(0.2)
(0.2)
(0.2)
Mineral Soils
+
(0.4)
(0.3)
(0.2)
(0.2)
(0.2)
(0.2)
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted to







Grassland
+
0.1
0.1
0.1
0.1
0.1
0.1
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Aboveground Live Biomass
2.7
2.6
2.4
2.4
2.4
2.4
2.4
Belowground Live Biomass
0.7
0.6
0.6
0.6
0.6
0.6
0.6
Dead Wood
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Litter
1.3
1.3
1.3
1.3
1.3
1.3
1.3
Total Mineral Soil Flux
(6.4)
(15.9)
(11.1)
(11.1)
(11.2)
(11.1)
(10.9)
Total Organic Soil Flux
0.2
0.5
0.5
0.5
0.5
0.5
0.5
Total Net Flux
(1.7)
(10.9)
(6.5)
(6.6)
(6.7)
(6.6)
(6.3)
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
+ Does not exceed 0.05 MMT C.
Methodology
The following section includes a description of the methodology used to estimate C stock changes for Land
Converted to Grassland, including (1) loss of aboveground and belowground biomass, dead wood and litter C with
conversion of Forest Land Converted to Grassland, as well as (2) the impact from all land use conversions to
grassland on mineral and organic soil organic C stocks.
Biomass, Dead Wood, and Litter Carbon Stock Changes
A Tier 3 method is applied to estimate biomass, dead wood and litter C stock changes for Forest Land Converted to
Grassland. Estimates are calculated in the same way as those in the Forest Land Remaining Forest Land category
using data from the USDA Forest Service, Forest Inventory and Analysis (FIA) program (USDA Forest Service 2018).
There are limited data on the herbaceous grassland C stocks following conversion so default biomass estimates
(IPCC 2006) for grasslands were used to estimate C stock changes (Note: litter and dead wood C stocks were
assumed to be zero following conversion because no reference C density estimates exist for grasslands). The
difference between the stocks is reported as the stock change under the assumption that the change occurred in
the year of the conversion.
The amount of biomass C that is lost abruptly with Forest Land Converted to Grasslands is estimated based on the
amount of C before conversion and the amount of C following conversion according to remeasurements in the FIA
program. This approach is consistent with IPCC (2006) that assumes there is an abrupt change during the first year,
but does not necessarily capture the slower change over the years following conversion until a new steady is
reached. It was determined that using an IPCC Tier I approach that assumes all C is lost in the year of conversion
for Forest Land Converted to Grasslands in the West and Great Plains states does not accurately characterize the
transfer of C in woody biomass during abrupt or gradual land use change. To estimate this transfer of C in woody
biomass, state-specific C densities for woody biomass remaining on these former forest lands following conversion
to grasslands were developed and included in the estimation of C stock changes from Forest Land Converted to
Grasslands in the West and Great Plains states. A review of the literature in grassland and rangeland ecosystems
(Asner et al. 2003; Huang et al. 2009; Tarhouni et al. 2016), as well as an analysis of FIA data, suggests that a
conservative estimate of 50 percent of the woody biomass C density was lost during conversion from Forest Land
to Grasslands. This estimate was used to develop state-specific C density estimates for biomass, dead wood, and
litter for Grasslands in the West and Great Plains states and these state-specific C densities were applied in the
6-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
compilation system to estimate the C losses associated with conversion from forest land to grassland in the West
and Great Plains states. Further, losses from forest land to what are often characterized as woodlands are included
in this category using FIA plot re-measurements and the methods and models described hereafter.
If FIA plots include data on individual trees, aboveground and belowground C density estimates are based on
Woodall et al. (2011). Aboveground and belowground biomass estimates also include live understory which is a
minor component of biomass defined as all biomass of undergrowth plants in a forest, including woody shrubs and
trees less than 2.54 cm dbh. For this Inventory, it was assumed that 10 percent of total understory C mass is
belowground (Smith et al. 2006). Estimates of C density are based on information in Birdsey (1996) and biomass
estimates from Jenkins et al. (2003).
If FIA plots include data on standing dead trees, standing dead tree C density is estimated following the basic
method applied to live trees (Woodall et al. 2011) with additional modifications to account for decay and structural
loss (Domke et al. 2011; Harmon et al. 2011). If FIA plots include data on downed dead wood, downed dead wood
C density is estimated based on measurements of a subset of FIA plots for downed dead wood (Domke et al. 2013;
Woodall and Monleon 2008). Downed dead wood is defined as pieces of dead wood greater than 7.5 cm diameter
that are not attached to live or standing dead trees at transect intersection. This includes stumps and roots of
harvested trees. To facilitate the downscaling of downed dead wood C estimates from the state-wide population
estimates to individual plots, downed dead wood models specific to regions and forest types within each region
are used. Litter C is the pool of organic C (also known as duff, humus, and fine woody debris) above the mineral
soil and includes woody fragments with diameters of up to 7.5 cm. A subset of FIA plots is measured for litter C. If
FIA plots include litter material, a modeling approach using litter C measurements from FIA plots is used to
estimate litter C density (Domke et al. 2016). See Annex 3.13 for more information about reference C density
estimates for forest land.
Soil Carbon Stock Changes
Soil organic C stock changes are estimated for Land Converted to Grassland according to land use histories
recorded in the 2015 USDA NRI survey for non-federal lands (USDA-NRCS 2018). Land use and some management
information (e.g., crop type, soil attributes, and irrigation) were originally collected for each NRI survey locations
on a 5-year cycle beginning in 1982. In 1998, the NRI Program began collecting annual data, and the annual data
are currently available through 2015 (USDA-NRCS 2018). NRI survey locations are classified as Land Converted to
Grassland in a given year between 1990 and 2015 if the land use is grassland but had been classified as another
use during the previous 20 years. NRI survey locations are classified according to land use histories starting in
1979, and consequently the classifications are based on less than 20 years from 1990 to 1998. This may have led to
an underestimation of Land Converted to Grassland in the early part of the time series to the extent that some
areas are converted to grassland between 1971 and 1978. For federal lands, the land use history is derived from
land cover changes in the National Land Cover Dataset (Yang et al. 2018; Homer et al. 2007; Fry et al. 2011; Homer
et al. 2015).
Mineral Soil Carbon Stock Changes
An IPCC Tier 3 model-based approach (Ogle et al. 2010) is applied to estimate C stock changes in mineral soils for
most of the area in Land Converted to Grassland. C stock changes on the remaining area are estimated with an
IPCC Tier 2 approach (Ogle et al. 2003), including prior cropland used to produce vegetables, tobacco, and
perennial/horticultural crops; land areas with very gravelly, cobbly, or shaley soils (greater than 35 percent by
volume); and land converted to grassland from another land use other than cropland.
A surrogate data method is used to estimate soil organic C stock changes from 2016 to 2019 at the national scale
for land areas included in the Tier 2 and Tier 3 methods. Specifically, linear regression models with autoregressive
moving-average (ARMA) errors (Brockwell and Davis 2016) are used to estimate the relationship between
surrogate data and the 1990 to 2015 emissions data that are derived using the Tier 2 and 3 methods. Surrogate
data for these regression models includes weather data from the PRISM Climate Group (PRISM Climate Group
2018). See Box 6-4 in the Methodology section of Cropland Remaining Cropland for more information about the
Land Use, Land-Use Change, and Forestry 6-85

-------
surrogate data method. Stock change estimates for 2016 to 2019 will be recalculated in future inventories when
new NRI data are available.
Tier 3 Approach. Mineral soil organic C stocks and stock changes are estimated using the DayCent biogeochemical58
model (Parton et al. 1998; Del Grosso et al. 2001, 2011). The DayCent model utilizes the soil C modeling framework
developed in the Century model (Parton et al. 1987,1988,1994; Metherell et al. 1993), but has been refined to
simulate dynamics at a daily time-step. Historical land use patterns and irrigation histories are simulated with
DayCent based on the 2015 USDA NRI survey (USDA-NRCS 2018). Carbon stocks and 95 percent confidence
intervals are estimated for each year between 1990 and 2015. See the Cropland Remaining Cropland section and
Annex 3.12 for additional discussion of the Tier 3 methodology for mineral soils.
Soil organic C stock changes from 2016 to 2019 were estimated using a surrogate data method described in Box
6-4 of the Methodology section in Cropland Remaining Cropland. Future inventories will be updated with new
activity data when the data are made available, and the time series will be recalculated (see the Planned
Improvements section in Cropland Remaining Cropland).
Tier 2 Approach. For the mineral soils not included in the Tier 3 analysis, soil organic C stock changes are estimated
using a Tier 2 Approach, as described in the Tier 2 Approach for mineral soils in Grassland Remaining Grassland
and Annex 3.12. This analysis includes application of the surrogate data method that is described in Box 6-4 of the
Methodology section in Cropland Remaining Cropland. As with the Tier 3 method, future Inventories will be
updated with new NRI activity data when the data are made available, and the time series will be recalculated.
Organic Soil Carbon Stock Changes
Annual C emissions from drained organic soils in Land Converted to Grassland are estimated using the Tier 2
method provided in IPCC (2006), with country-specific C loss rates (Ogle et al. 2003) as described in the Cropland
Remaining Cropland section and Annex 3.12 for organic soils. A surrogate data method is used to estimate annual
C emissions from organic soils from 2016 to 2019 as described in Box 6-4 of the Methodology section in Cropland
Remaining Cropland. Estimates for 2016 to 2019 will be recalculated in future Inventories when new NRI data are
available.
Uncertainty and Time-Series Consistency
The uncertainty analysis for biomass, dead wood and litter C losses with Forest Land Converted to Grassland is
conducted in the same way as the uncertainty assessment for forest ecosystem C flux in the Forest Land Remaining
Forest Land category. Sample and model-based error are combined using simple error propagation methods
provided by the IPCC (2006), by taking the square root of the sum of the squares of the standard deviations of the
uncertain quantities. For additional details see the Uncertainty Analysis in Annex 3.13.
The uncertainty analyses for mineral soil organic C stock changes using the Tier 3 and Tier 2 methodologies are
based on a Monte Carlo approach that is described in the Cropland Remaining Cropland section and Annex 3.12.
The uncertainty for annual C emission estimates from drained organic soils in Land Converted to Grassland is
estimated using a Monte Carlo approach, which is also described in the Cropland Remaining Cropland section. For
2016 to 2019, there is additional uncertainty propagated through the Monte Carlo Analysis associated with a
surrogate data method, which is also described in Cropland Remaining Cropland.
Uncertainty estimates are presented in Table 6-46 for each subsource (i.e., biomass C stocks, mineral and organic C
stocks in soils) and the method applied in the inventory analysis (i.e., Tier 2 and Tier 3). Uncertainty estimates from
the Tier 2 and 3 approaches are combined using the simple error propagation methods provided by the IPCC
(2006), as discussed in the previous paragraph. The combined uncertainty for total C stocks in Land Converted to
Grassland ranges from 146 percent below to 148 percent above the 2019 stock change estimate of 23.2 MMT C02
58 Biogeochemical cycles are the flow of chemical elements and compounds between living organisms and the physical
environment.
6-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Eq. The large relative uncertainty around the 2019 stock change estimate is partly due to large uncertainties in
biomass and dead organic matter C losses with Forest Land Conversion to Grassland. The large relative uncertainty
is also partly due to variation in soil organic C stock changes that is not explained by the surrogate data method,
leading to high prediction error with the splicing method.
Table 6-46: Approach 2 Quantitative Uncertainty Estimates for Soil, Dead Organic Matter
and Biomass C Stock Changes occurring within Land Converted to Grassland(MMT CO2 Eq.
and Percent)
Source
2019 Flux Estimate3
Uncertainty Range Relative to Flux Estimate3
(MMT C02 Eq.)
(MMTCOz
Eq.)
(%)



Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Cropland Converted to Grassland
(19.8)
(51.1)
11.5
-158%
158%
Mineral Soil C Stocks: Tier 3
(15.0)
(46.0)
15.9
-206%
206%
Mineral Soil C Stocks: Tier 2
(6.1)
(10.4)
(1.8)
-70%
70%
Organic Soil C Stocks: Tier 2
1.3
+
2.7
-101%
101%
Forest Land Converted to Grassland
14.8
4.2
27.0
-71%
82%
Aboveground Live Biomass
8.6
(0.4)
19.3
-105%
123%
Belowground Live Biomass
2.1
(0.1)
4.8
-105%
127%
Dead Wood
(0.6)
(1.9)
+
-226%
106%
Litter
4.6
(0.2)
10.0
-105%
116%
Mineral Soil C Stocks: Tier 2
(0.2)
(0.4)
+
-99%
99%
Organic Soil C Stocks: Tier 2
0.2
+
0.4
-110%
110%
Other Lands Converted to Grassland
(17.6)
(24.4)
(10.8)
-39%
39%
Mineral Soil C Stocks: Tier 2
(17.7)
(24.5)
(10.9)
-39%
39%
Organic Soil C Stocks: Tier 2
0.1
+
0.2
-146%
146%
Settlements Converted to Grassland
(0.8)
(1.1)
(0.5)
-37%
37%
Mineral Soil C Stocks: Tier 2
(0.8)
(1.1)
(0.5)
-36%
36%
Organic Soil C Stocks: Tier 2
+
+
+
-293%
293%
Wetlands Converted to Grasslands
0.2
(0.1)
0.5
-137%
137%
Mineral Soil C Stocks: Tier 2
+
(0.1)
+
-130%
130%
Organic Soil C Stocks: Tier 2
0.2
+
0.5
-111%
111%
Total: Land Converted to Grassland
(23.2)
(56.9)
11.1
-146%
148%
Aboveground Live Biomass
8.6
(0.4)
19.3
-105%
123%
Belowground Live Biomass
2.1
(0.1)
4.8
-105%
127%
Dead Wood
(0.6)
(1.9)
+
-226%
106%
Litter
4.6
(0.2)
10.0
-105%
116%
Mineral Soil C Stocks: Tier 3
(15.0)
(46.0)
15.9
-206%
206%
Mineral Soil C Stocks: Tier 2
(24.8)
(32.9)
(16.7)
-33%
33%
Organic Soil C Stocks: Tier 2
1.8
0.5
3.2
-76%
76%
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Uncertainty is also associated with a lack of reporting on biomass, dead wood and litter C stock changes for
agroforestry systems. However, there are currently no datasets to evaluate the trends. Changes in biomass and
dead organic matter C stocks are assumed to be negligible with the exception of forest lands, which are included in
this analysis in other grasslands. This assumption will be further explored in a future Inventory.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
Land Use, Land-Use Change, and Forestry 6-87

-------
QA/QC and Verification
See the QA/QC and Verification section in Cropland Remaining Cropland and Grassland Remaining Grassland for
information on QA/QC steps. Errors were found in the estimation of uncertainty due to incorrect cell references in
spreadsheets. The errors were corrected through quality control review of the Inventory.
Recalculations Discussion
Recalculations are associated with new FIA data from 1990 to 2018 on biomass, dead wood and litter C stocks in
Forest Land Converted to Grassland, and updated estimates for mineral soils from 2016 to 2018 using additional
surrogate data. As a result, Land Converted to Grassland has a smaller reported change in C stocks compared to
the previous Inventory, estimated at 0.17 MMT C02 Eq. on average over the time series. This represents a 1
percent decrease in C stock changes for Land Converted to Grassland compared to the previous Inventory.
Planned Improvements
The amount of biomass C that is lost abruptly or the slower changes that continue to occur over a decade or longer
with Forest Land Converted to Grasslands will be further refined in a future Inventory. The current values are
estimated based on the amount of C before conversion and an estimated level of C left after conversion based on
limited plot data from the FIA and published literature for the Western United States and Great Plains Regions. The
amount of C left after conversion will be further investigated with additional data collection, particularly in the
Western United States and Great Plains, including tree biomass, understory biomass, dead wood and litter C pools.
In addition, biomass C stock changes will be estimated for Cropland Converted to Grassland, and other land use
conversions to grassland, to the extent that data are available.
An additional planned improvement for the Land Converted to Grassland category is to develop an inventory of C
stock changes for grasslands in Alaska. Table 6-47 provides information on the amount of managed area in Alaska
that is Land Converted to Grassland, which is as high as 54 thousand hectares in 2011.59 Note that areas of Land
Converted to Grassland in Alaska for 1990 to 2001 are classified as Grassland Remaining Grassland because land
use change data are not available until 2002. For information about other improvements, see the Planned
Improvements section in Cropland Remaining Cropland and Grassland Remaining Grassland.
Table 6-47: Area of Managed Land in Land Converted to Grasslands Alaska that is not
included in the current Inventory (Thousand Hectares)
Area (Thousand Hectares)
Not Included in
Year
Managed Land
Inventory
Inventory
1990
9,394
9,394
0
1991
9,485
9,485
0
1992
9,691
9,691
0
1993
11,566
11,566
0
1994
13,378
13,378
0
1995
13,994
13,994
0
1996
14,622
14,622
0
59 All of the Land Converted to Grassland according to the land representation is included in the inventory from 1990 through
2001 for the conterminous United States. However, there are no data to evaluate land use change in Alaska for this time
period, and so the balance of the managed area that may be converted to grassland in these years is included in Grassland
Remaining Grassland section. This gap in land use change data for Alaska will be addressed in a future Inventory.
6-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
1997
15,162
15,162
0
1998
19,052
19,052
0
1999
19,931
19,931
0
2000
20,859
20,859
0
2001
21,968
21,968
0
2002
22,395
22,392
3
2003
22,015
22,008
7
2004
22,557
22,547
10
2005
22,460
22,447
13
2006
22,718
22,702
16
2007
22,450
22,428
21
2008
22,685
22,661
24
2009
22,608
22,581
26
2010
22,664
22,634
29
2011
22,805
22,750
54
2012
22,643
22,596
47
2013
21,472
21,439
33
2014
20,195
20,163
33
2015
20,242
20,210
33
2016
ND
ND
ND
2017
ND
ND
ND
2018
ND
ND
ND
2019
ND
ND
ND
Note: NRI data are not available after 2015, and these years are designated as ND (No data).
6.8 Wetlands Remaining Wetlands (CRF
Category 4D1)
Wetlands Remaining Wetlands includes all wetland in an Inventory year that had been classified as wetland for the
previous 20 years, and in this Inventory the flux estimates include Peatlands and Coastal Wetlands.
Peatlands Remaining Peatlands
Emissions from Managed Peatlands
Managed peatlands are peatlands that have been cleared and drained for the production of peat. The production
cycle of a managed peatland has three phases: land conversion in preparation for peat extraction (e.g., clearing
surface biomass, draining), extraction (which results in the emissions reported under Peatlands Remaining
Peatlands), and abandonment, restoration, rewetting, or conversion of the land to another use.
Carbon dioxide emissions from the removal of biomass and the decay of drained peat constitute the major
greenhouse gas flux from managed peatlands. Managed peatlands may also emit CH4 and N20. The natural
production of CH4 is largely reduced but not entirely shut down when peatlands are drained in preparation for
peat extraction (Strack et al. 2004 as cited in the 2006IPCC Guidelines). Drained land surface and ditch networks
contribute to the CH4 flux in peatlands managed for peat extraction. Methane emissions were considered
insignificant under the IPCC Tier 1 methodology (IPCC 2006), but are included in the emissions estimates for
Peatlands Remaining Peatlands consistent with the 2013 Supplement to the 2006 IPCC Guidelines for National
Greenhouse Gas Inventories: Wetlands (IPCC 2013). Nitrous oxide emissions from managed peatlands depend on
Land Use, Land-Use Change, and Forestry 6-89

-------
site fertility. In addition, abandoned and restored peatlands continue to release greenhouse gas emissions.
Although methodologies are provided for rewetted organic soils (which includes rewetted/restored peatlands) in
IPCC (2013) guidelines, information on the areal extent of rewetted/restored peatlands in the United States is
currently unavailable. The current Inventory estimates C02, CH4and N20 emissions from peatlands managed for
peat extraction in accordance with IPCC (2006 and 2013) guidelines.
C02, N20, and CH4 Emissions from Peatlands Remaining Peatlands
IPCC (2013) recommends reporting C02, N20, and CH4 emissions from lands undergoing active peat extraction (i.e.,
Peatlands Remaining Peatlands) as part of the estimate for emissions from managed wetlands. Peatlands occur
where plant biomass has sunk to the bottom of water bodies and water-logged areas and exhausted the oxygen
supply below the water surface during the course of decay. Due to these anaerobic conditions, much of the plant
matter does not decompose but instead forms layers of peat over decades and centuries. In the United States,
peat is extracted for horticulture and landscaping growing media, and for a wide variety of industrial, personal
care, and other products. It has not been used for fuel in the United States for many decades. Peat is harvested
from two types of peat deposits in the United States: sphagnum bogs in northern states (e.g., Minnesota) and
wetlands in states further south (e.g., Florida). The peat from sphagnum bogs in northern states, which is nutrient-
poor, is generally corrected for acidity and mixed with fertilizer. Production from more southerly states is relatively
coarse (i.e., fibrous) but nutrient-rich.
IPCC (2006 and 2013) recommend considering both on-site and off-site emissions when estimating C02 emissions
from Peatlands Remaining Peatlands using the Tier 1 approach. Current methodologies estimate only on-site N20
and CH4 emissions, since off-site N20 estimates are complicated by the risk of double-counting emissions from
nitrogen fertilizers added to horticultural peat, and off-site CH4 emissions are not relevant given the non-energy
uses of peat, so methodologies are not provided in IPCC (2013) guidelines.
On-site emissions from managed peatlands occur as the land is cleared of vegetation and the underlying peat is
exposed to sun and weather. As this occurs, some peat deposit is lost and C02 is emitted from the oxidation of the
peat. Since N20 emissions from saturated ecosystems tend to be low unless there is an exogenous source of
nitrogen, N20 emissions from drained peatlands are dependent on nitrogen mineralization and therefore on soil
fertility. Peatlands located on highly fertile soils contain significant amounts of organic nitrogen in inactive form.
Draining land in preparation for peat extraction allows bacteria to convert the nitrogen into nitrates which leach to
the surface where they are reduced to N20, and contributes to the activity of methanogens and methanotrophs
that result in CH4 emissions (Blodau 2002; Treat et al. 2007 as cited in IPCC 2013). Drainage ditches, which are
constructed to drain the land in preparation for peat extraction, also contribute to the flux of CH4 through in situ
production and lateral transfer of CH4 from the organic soil matrix (IPCC 2013).
Off-site C02 emissions from managed peatlands occur from waterborne carbon losses and the horticultural and
landscaping use of peat. Dissolved organic carbon from water drained off peatlands reacts within aquatic
ecosystems and is converted to C02, which is then emitted to the atmosphere (Billet et al. 2004 as cited in IPCC
2013). During the horticultural and landscaping use of peat, nutrient-poor (but fertilizer-enriched) peat tends to be
used in bedding plants and in greenhouse and plant nursery production, whereas nutrient-rich (but relatively
coarse) peat is used directly in landscaping, athletic fields, golf courses, and plant nurseries. Most (nearly 94
percent) of the C02 emissions from peat occur off-site, as the peat is processed and sold to firms which, in the
United States, use it predominantly for the aforementioned horticultural and landscaping purposes.
Total emissions from Peatlands Remaining Peatlands were estimated to be 0.8 MMT C02 Eq. in 2019 (see Table
6-48 and Table 6-49) comprising 0.8 MMT C02 Eq. (778 kt) of C02, 0.004 MMT C02 Eq. (0.16 kt) of CH4 and 0.0005
MMT C02 Eq. (0.002 kt) of N20. Total emissions in 2019 were about 2.7 percent less than total emissions in 2018.
Total emissions from Peatlands Remaining Peatlands have fluctuated between 0.7 and 1.3 MMT C02 Eq. across the
time series with a decreasing trend from 1990 until 1993, followed by an increasing trend until reaching peak
emissions in 2000. After 2000, emissions generally decreased until 2006 and then increased until 2009. The trend
reversed in 2009 and total emissions have generally decreased between 2009 and 2019. Carbon dioxide emissions
from Peatlands Remaining Peatlands have fluctuated between 0.7 and 1.3 MMT C02 across the time series, and
6-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
these emissions drive the trends in total emissions. Methane and N20 emissions remained close to zero across the
time series. Nitrous oxide emissions showed a decreasing trend from 1990 until 1995, followed by an increasing
trend through 2001. Nitrous oxide emissions decreased between 2001 and 2006, followed by a leveling off
between 2008 and 2010, and a general decline between 2011 and 2019. Methane emissions decreased from 1990
until 1995, followed by an increasing trend through 2000, a period of fluctuation through 2010, and a general
decline between 2010 and 2019 (emissions rose slightly from 2016 to 2017 but resumed the downward trend
since).
Table 6-48: Emissions from Peatlands Remaining Peatiands (MMT CO2 Eq.)
Gas
1990
2005
2015
2016
2017
2018
2019
C02
1.1
1.1
0.8
0.8
0.8
0.7
0.8
Off-site
1.0
1.0
0.7
0.7
0.7
0.7
0.7
On-site
0.1
0.1
+
+
0.1
0.1
0.1
CH4 (On-site)
+
+
+
+
+
+
+
N20 (On-site)
+
+
+
+
+
+
+
Total
1.1
1.1
0.8
0.7
0.8
0.8
0.8
Note: These numbers are based on U.S. production data in accordance with Tier 1 guidelines, which
does not take into account imports, exports, and stockpiles (i.e., apparent consumption). Off-site N20
emissions are not estimated to avoid double-counting N20 emitted from the fertilizer that the peat is
mixed with prior to horticultural use (see IPCC 2006). Totals may not sum due to independent
rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 6-49: Emissions from Peatiands Remaining Peatiands (kt)
Gas
1990
2005
2015
2016
2017
2018
2019
co2
1,055
1,101
755
733
829
795
778
Off-site
985
1,030
706
686
774
744
727
On-site
70
71
49
47
55
51
50
CH4 (On-site)
+
+
+
+
+
+
+
N20 (On-site)
+
+
+
+
+
+
+
Note: These numbers are based on U.S. production data in accordance with Tier 1 guidelines, which
does not take into account imports, exports, and stockpiles (i.e., apparent consumption). Off-site N20
emissions are not estimated to avoid double-counting N20 emitted from the fertilizer that the peat is
mixed with prior to horticultural use (see IPCC 2006). Totals may not sum due to independent
rounding.
+ Does not exceed 0.5 kt
Methodology
Off-Site CO2 Emissions
Carbon dioxide emissions from domestic peat production were estimated using a Tier 1 methodology consistent
with IPCC (2006). Off-site C02 emissions from Peatiands Remaining Peatiands were calculated by apportioning the
annual weight of peat produced in the United States (Table 6-50) into peat extracted from nutrient-rich deposits
and peat extracted from nutrient-poor deposits using annual percentage-by-weight figures. These nutrient-rich
and nutrient-poor production values were then multiplied by the appropriate default C fraction conversion factor
taken from IPCC (2006) in order to obtain off-site emission estimates. For the lower 48 states, both annual
percentages of peat type by weight and domestic peat production data were sourced from estimates and industry
statistics provided in the Minerals Yearbook and Mineral Commodity Summaries from the U.S. Geological Survey
(USGS 1995 through 2017; USGS 2018; USGS 2020). To develop these data, the U.S. Geological Survey (USGS; U.S.
Bureau of Mines prior to 1997) obtained production and use information by surveying domestic peat producers.
On average, about 75 percent of the peat operations respond to the survey; and USGS estimates data for non-
respondents on the basis of prior-year production levels (Apodaca 2011).
Land Use, Land-Use Change, and Forestry 6-91

-------
The Alaska estimates rely on reported peat production from the annual Alaska's Mineral Industry reports (DGGS
1993 through 2015). Similar to the U.S. Geological Survey, the Alaska Department of Natural Resources, Division of
Geological & Geophysical Surveys (DGGS) solicits voluntary reporting of peat production from producers for the
Alaska's Mineral Industry report. However, the report does not estimate production for the non-reporting
producers, resulting in larger inter-annual variation in reported peat production from Alaska depending on the
number of producers who report in a given year (Szumigala 2011). In addition, in both the lower 48 states and
Alaska, large variations in peat production can also result from variations in precipitation and the subsequent
changes in moisture conditions, since unusually wet years can hamper peat production. The methodology
estimates Alaska emissions separately from lower 48 emissions because the state conducts its own mineral survey
and reports peat production by volume, rather than by weight (Table 6-51). However, volume production data
were used to calculate off-site C02 emissions from Alaska applying the same methodology but with volume-specific
C fraction conversion factors from IPCC (2006).60 Peat production was not reported for 2015 in Alaska's Mineral
Industry 2014 report (DGGS 2015); and reliable data are not available beyond 2012, so Alaska's peat production in
2013 through 2019 (reported in cubic yards) was assumed to be equal to the 2012 value.
Consistent with IPCC (2013) guidelines, off-site C02 emissions from dissolved organic carbon were estimated based
on the total area of peatlands managed for peat extraction, which is calculated from production data using the
methodology described in the On-Site C02 Emissions section below. Carbon dioxide emissions from dissolved
organic C were estimated by multiplying the area of peatlands by the default emission factor for dissolved organic
C provided in IPCC (2013).
The apparent consumption of peat, which includes production plus imports minus exports plus the decrease in
stockpiles, in the United States is over time the amount of domestic peat production. However, consistent with the
Tier 1 method whereby only domestic peat production is accounted for when estimating off-site emissions, off-site
C02 emissions from the use of peat not produced within the United States are not included in the Inventory. The
United States has largely imported peat from Canada for horticultural purposes; in 2018, imports of sphagnum
moss (nutrient-poor) peat from Canada represented 96 percent of total U.S. peat imports (USGS 2018). Most peat
produced in the United States is reed-sedge peat, generally from southern states, which is classified as nutrient-
rich by IPCC (2006). Higher-tier calculations of C02 emissions from apparent consumption would involve
consideration of the percentages of peat types stockpiled (nutrient-rich versus nutrient-poor) as well as the
percentages of peat types imported and exported.
Table 6-50: Peat Production of Lower 48 States (kt)
Type of Deposit
1990
2005
2015
2016
2017
2018
2019
Nutrient-Rich
595.1
657.6
405.0
388.1
423.3
416.7
423.0
Nutrient-Poor
55.4
27.4
50.1
52.9
74.7
62.3
47.0
Total Production
692.0
685.0
455.0
441.0
498.0
479.0
470.0
Sources: United States Geological Survey (USGS) (1991-2017) Minerals Yearbook: Peat (1994-2016); United
States Geological Survey (USGS) (2018) Minerals Yearbook: Peat - Tables-only release (2018); United States
Geological Survey (USGS) (2020) Mineral Commodity Summaries: Peat (2020).
Table 6-51: Peat Production of Alaska (Thousand Cubic Meters)

1990
2005
2015
2016
2017
2018
2019
Total Production
49.7
47.8
93.1
93.1
93.1
93.1
93.1
Sources: Division of Geological & Geophysical Surveys (DGGS), Alaska Department of Natural Resources
(1997-2015) Alaska's Mineral Industry Report (1997-2014).
60 Peat produced from Alaska was assumed to be nutrient poor; as is the case in Canada, "where deposits of high-quality [but
nutrient poor] sphagnum moss are extensive" (USGS 2008).
6-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
On-site CO2 Emissions
IPCC (2006) suggests basing the calculation of on-site emission estimates on the area of peatlands managed for
peat extraction differentiated by the nutrient type of the deposit (rich versus poor). Information on the area of
land managed for peat extraction is currently not available for the United States, but consistent with IPCC (2006),
an average production rate for the industry was applied to derive an area estimate. In a mature industrialized peat
industry, such as exists in the United States and Canada, the vacuum method can extract up to 100 metric tons per
hectare per year (Cleary et al. 2005 as cited in IPCC 2006).61 The area of land managed for peat extraction in the
lower 48 states of the United States was estimated using nutrient-rich and nutrient-poor production data and the
assumption that 100 metric tons of peat are extracted from a single hectare in a single year, see Table 6-52. The
annual land area estimates were then multiplied by the IPCC (2013) default emission factor in order to calculate
on-site C02 emission estimates.
Production data are not available by weight for Alaska. In order to calculate on-site emissions resulting from
Peatlands Remaining Peatlands in Alaska, the production data by volume were converted to weight using annual
average bulk peat density values, and then converted to land area estimates using the assumption that a single
hectare yields 100 metric tons, see Table 6-53. The IPCC (2006) on-site emissions equation also includes a term
that accounts for emissions resulting from the change in C stocks that occurs during the clearing of vegetation
prior to peat extraction. Area data on land undergoing conversion to peatlands for peat extraction is also
unavailable for the United States. However, USGS records show that the number of active operations in the United
States has been declining since 1990; therefore, it seems reasonable to assume that no new areas are being
cleared of vegetation for managed peat extraction. Other changes in C stocks in living biomass on managed
peatlands are also assumed to be zero under the Tier 1 methodology (IPCC 2006 and 2013).
Table 6-52: Peat Production Area of Lower 48 States (Hectares)

1990a
2005
2015
2016
2017
2018
2019
Nutrient-Rich
5,951
i 6,576 :
4,050
3,881
4,233
4,167
4,230
Nutrient-Poor
554
274
501
529
747
623
470
Total Production
6,920
6,850
4,550
4,410
4,980
4,790
4,700
a A portion of the production in 1990 is of unknown nutrient type, resulting in a total
production value greater than the sum of nutrient-rich and nutrient-poor.
Sources: Calculated using peat production values in Table 6-50, an assumed yield of 100
metric tons per hectare per year.
Table 6-53: Peat Production Area of Alaska (Hectares)

1990
2005
2015
2016
2017
2018
2019
Nutrient-Rich
0
0
0
0
0
0
0
Nutrient-Poor
286
104
209
201
333
212
212
Total Production
286
104
209
201
333
212
212
Sources: Calculated using peat production values in Table 6-51, an assumed yield of 100 metric
tons per hectare per year.
On-site N2O Emissions
IPCC (2006) suggests basing the calculation of on-site N20 emission estimates on the area of nutrient-rich
peatlands managed for peat extraction. These area data are not available directly for the United States, but the on-
site C02 emissions methodology above details the calculation of area data from production data. In order to
61 The vacuum method is one type of extraction that annually "mills" or breaks up the surface of the peat into particles, which
then dry during the summer months. The air-dried peat particles are then collected by vacuum harvesters and transported from
the area to stockpiles (IPCC 2006).
Land Use, Land-Use Change, and Forestry 6-93

-------
estimate N20 emissions, the area of nutrient-rich Peatlands Remaining Peatlands was multiplied by the
appropriate default emission factor taken from IPCC (2013).
On-site CH4 Emissions
IPCC (2013) also suggests basing the calculation of on-site CH4 emission estimates on the total area of peatlands
managed for peat extraction. Area data is derived using the calculation from production data described in the On-
site C02 Emissions section above. In order to estimate CH4 emissions from drained land surface, the area of
Peatlands Remaining Peatlands was multiplied by the emission factor for direct CH4 emissions taken from IPCC
(2013). In order to estimate CH4 emissions from drainage ditches, the total area of peatland was multiplied by the
default fraction of peatland area that contains drainage ditches, and the appropriate emission factor taken from
IPCC (2013). See Table 6-54 for the calculated area of ditches and drained land.
Table 6-54: Peat Production (Hectares)

1990
2005
2015
2016
2017
2018
2019
Lower 48 States
Area of Drained Land
6,574
6,508
4,323
4,190
4,731
4,551
4,465
Area of Ditches
346
343
228
221
249
240
235
Total Production
6,920
6,850
4,550
4,410
4,980
4,790
4,700
Alaska
Area of Drained Land
272
99
198
191
317
202
202
Area of Ditches
14
5
10
10
17
11
11
Total Production
286
104
209
201
333
212
212
Sources: Calculated using peat production values in Tables Table 6-50 and Table 6-51, an assumed yield of 100 metric tons
per hectare per year, and an assumed value of 5 percent ditch area.
Uncertainty and Time-Series Consistency
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the uncertainty of C02, CH4, and N20
emissions from Peatlands Remaining Peatlands for 2019, using the following assumptions:
•	The uncertainty associated with peat production data was estimated to be ± 25 percent (Apodaca 2008)
and assumed to be normally distributed.
•	The uncertainty associated with peat production data stems from the fact that the USGS receives data
from the smaller peat producers but estimates production from some larger peat distributors. The peat
type production percentages were assumed to have the same uncertainty values and distribution as the
peat production data (i.e., ± 25 percent with a normal distribution).
•	The uncertainty associated with the reported production data for Alaska was assumed to be the same as
for the lower 48 states, or ± 25 percent with a normal distribution. It should be noted that the DGGS
estimates that around half of producers do not respond to their survey with peat production data;
therefore, the production numbers reported are likely to underestimate Alaska peat production
(Szumigala 2008).
•	The uncertainty associated with the average bulk density values was estimated to be ± 25 percent with a
normal distribution (Apodaca 2008).
•	IPCC (2006 and 2013) gives uncertainty values for the emissions factors for the area of peat deposits
managed for peat extraction based on the range of underlying data used to determine the emission
factors. The uncertainty associated with the emission factors was assumed to be triangularly distributed.
•	The uncertainty values surrounding the C fractions were based on IPCC (2006) and the uncertainty was
assumed to be uniformly distributed.
•	The uncertainty values associated with the fraction of peatland covered by ditches was assumed to be ±
100 percent with a normal distribution based on the assumption that greater than 10 percent coverage,
the upper uncertainty bound, is not typical of drained organic soils outside of The Netherlands (IPCC
2013).
6-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 6-55. Carbon dioxide
emissions from Peatlands Remaining Peatlands in 2019 were estimated to be between 0.7 and 0.9 MMT C02 Eq. at
the 95 percent confidence level. This indicates a range of 16 percent below to 16 percent above the 2019 emission
estimate of 0.8 MMT C02 Eq. Methane emissions from Peatlands Remaining Peatlands in 2019 were estimated to
be between 0.002 and 0.007 MMT C02 Eq. This indicates a range of 59 percent below to 78 percent above the
2019 emission estimate of 0.004 MMT C02 Eq. Nitrous oxide emissions from Peatlands Remaining Peatlands in
2017 were estimated to be between 0.0003 and 0.0009 MMT C02 Eq. at the 95 percent confidence level. This
indicates a range of 56 percent below to 56 percent above the 2019 emission estimate of 0.0006 MMT C02 Eq.
Table 6-55: Approach 2 Quantitative Uncertainty Estimates for CO2, Cm, and N2O Emissions
from Peatlands Remaining Peatlands (MMT CO2 Eq. and Percent)


2019 Emission




Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMT CO? Eq.)
(MMTCOz
Eq.)

(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Peatlands Remaining Peatlands
C02
0.8
0.7
0.9
-16%
16%
Peatlands Remaining Peatlands
ch4
+
+
+
-59%
78%
Peatlands Remaining Peatlands
n2o
+
+
+
-56%
56%
+ Does not exceed 0.05 MMT C02 Eq.
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
A QA/QC analysis was performed to review input data and calculations, and no issues were identified. In addition,
the emission trends were analyzed to ensure they reflected activity data trends.
Recalculations Discussion
The emission estimates for Peatlands Remaining Peatlands were updated for 2019 using the Peat section of the
Mineral Commodity Summaries 2018, Mineral Commodity Summaries 2019 and Mineral Commodity Summaries
2020. The 2018 edition updated 2013 data and the 2019 edition updated 2014 data for the lower 48 states. The
2020 edition updated 2017 data and provided peat type production estimates for 2018 and 2019. Although Alaska
peat production data for 2017 were unavailable, 2014 data are available in the Alaska's Mineral Industry 2014
report. However, the reported values represented an apparent 98 percent decrease in production since 2012. Due
to the uncertainty of the most recent data, 2013, 2014, 2015, 2016, 2017, 2018, and 2019 values were assumed to
be equal to the 2012 value, seen in the Alaska's Mineral Industry 2013 report. If updated data are available for the
next inventory cycle, this will result in a recalculation in the next Inventory report.
Planned Improvements
In order to further improve estimates of C02, N20, and CH4 emissions from Peatlands Remaining Peatlands, future
efforts will investigate if improved data sources exist for determining the quantity of peat harvested per hectare
and the total area undergoing peat extraction.
Efforts will also be made to identify a new source for Alaska peat production. The current source has not been
reliably updated since 2012 and Alaska Department of Natural Resources indicated future publication of data has
been discontinued.
The implied emission factors will be calculated and included in this chapter for future Inventories. The N20
emissions calculation uses different land areas than the C02 and CH4 emission calculations, so estimating the
Land Use, Land-Use Change, and Forestry 6-95

-------
implied emission factor per total land area is not appropriate and are not generated in the CRF tables. The
inclusion of implied emission factors in this chapter will provide another method of QA/QC and verification.
The 2006IPCC Guidelines do not cover all wetland types; they are restricted to peatlands drained and managed for
peat extraction, conversion to flooded lands, and some guidance for drained organic soils. They also do not cover
all of the significant activities occurring on wetlands (e.g., rewetting of peatlands). Since this Inventory only
includes Peatlands Remaining Peatlands, additional wetland types and activities found in the 2013 IPCC
Supplement will be reviewed to determine if they apply to the United States. For those that do, available data will
be investigated to allow for the estimation of greenhouse gas fluxes in future inventory years.
Coastal Wetlands Remaining Coastal Wetlands
This Inventory recognizes Wetlands as a "land-use that includes land covered or saturated for all or part of the
year, in addition to areas of lakes, reservoirs, and rivers." Consistent with ecological definitions of wetlands,62 the
United States has historically included under the category of Wetlands those coastal shallow water areas of
estuaries and bays that lie within the extent of the Land Representation.
Guidance on quantifying greenhouse gas emissions and removals on Coastal Wetlands is provided in the 2013
Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands (Wetlands
Supplement), which recognizes the particular importance of vascular plants in sequestering C02 from the
atmosphere within biomass, dead organic material (DOM; including litter and dead wood stocks) and soils. Thus,
the Wetlands Supplement provides specific guidance on quantifying emissions and removals on organic and
mineral soils that are covered or saturated for part of the year by tidal fresh, brackish or saline water and are
vegetated by vascular plants and may extend seaward to the maximum depth of vascular plant vegetation. The
United States calculates emissions and removals based upon the stock change method for soil carbon and the gain-
loss method for biomass and DOM. Presently, this Inventory does not calculate the lateral flux of carbon to or from
any land use. Lateral transfer of organic carbon to coastal wetlands and to marine sediments within U.S. waters is
the subject of ongoing scientific investigation.
The United States recognizes both Vegetated Wetlands and Unvegetated Open Water as Coastal Wetlands. Per
guidance provided by the Wetlands Supplement, sequestration of carbon into biomass, DOM and soil carbon pools
is recognized only in Vegetated Coastal Wetlands and does not occur in Unvegetated Open Water Coastal
Wetlands. The United States takes the additional step of recognizing that stock losses occur when Vegetated
Coastal Wetlands are converted to Unvegetated Open Water Coastal Wetlands.
This Inventory includes all privately-owned and publicly-owned coastal wetlands (i.e., mangroves and tidal marsh)
along the oceanic shores on the conterminous United States, but does not include Coastal Wetlands Remaining
Coastal Wetlands in Alaska or Hawaii. Seagrasses are not currently included within the Inventory due to insufficient
data on distribution, change through time and carbon (C) stocks or C stock changes as a result of anthropogenic
influence.
Under the Coastal Wetlands Remaining Coastal Wetlands category, the following emissions and removals are
quantified in this chapter:
1)	Carbon stock changes and CH4 emissions on Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands,
2)	Carbon stock changes on Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal
Wetlands,
3)	Carbon stock changes on Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal
Wetlands, and
4)	Nitrous Oxide Emissions from Aquaculture in Coastal Wetlands.
62 See ; accessed October 2020.
6-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Vegetated coastal wetlands hold C in all five C pools (i.e., aboveground, belowground, dead organic matter [DOM;
dead wood and litter], and soil) though typically soil C and, to a lesser extent, aboveground and belowground
biomass are the dominant pools, depending on wetland type (i.e., forested vs. marsh). Vegetated Coastal Wetlands
are net accumulators of C as soils accumulate C under anaerobic soil conditions and in plant biomass. Emissions
from soil C and biomass stocks occur when Vegetated Coastal Wetlands are converted to Unvegetated Open Water
Coastal Wetlands (e.g., when Vegetated Coastal Wetlands are lost due to subsidence, channel cutting through
Vegetated Coastal Wetlands), but are still recognized as Coastal Wetlands in this Inventory. These C stock losses
resulting from conversion to Unvegetated Open Water Coastal Wetlands can cause the release of many years of
accumulated soil C, as well as the standing stock of biomass C. Conversion of Unvegetated Open Water Coastal
Wetlands to Vegetated Coastal Wetlands initiates the building of C stocks within soils and biomass. In applying the
Wetlands Supplement methodologies for CH4 emissions, coastal wetlands in salinity conditions less than half that
of sea water are sources of CH4 as result of slow decomposition of organic matter under lower salinity brackish and
freshwater, anaerobic conditions. Conversion of Vegetated Coastal Wetlands to or from Unvegetated Open Water
Coastal Wetlands do not result in a change in salinity condition and are assumed to have no impact on CH4
emissions. The Wetlands Supplement provides methodologies to estimate N20 emissions from coastal wetlands
that occur due to aquaculture. The N20 emissions from aquaculture result from the N derived from consumption
of the applied food stock that is then excreted as N load available for conversion to N20. While N20 emissions can
also occur due to anthropogenic N loading from the watershed and atmospheric deposition, these emissions are
not reported here to avoid double-counting of indirect N20 emissions with the Agricultural Soils Management,
Forest Land and Settlements categories.
The Wetlands Supplement provides methodologies for estimating C stock changes and CH4emissions from
mangroves, tidal marshes and seagrasses. Depending upon their height and area, C stock changes from mangroves
may be reported under the Forest Land category or under Coastal Wetlands. If mangrove stature is 5 m or greater
or if there is evidence that trees can obtain that height, mangroves are reported under the Forest Land category.
Mangrove forests that are less than 5 m are reported under Coastal Wetlands. All other non-drained, intact coastal
marshes are intended to be reported under Coastal Wetlands.
Because of human activities and level of regulatory oversight, all coastal wetlands within the conterminous United
States are included within the managed land area described in Section 6, and as such estimates of C stock changes,
emissions of CH4, and emissions of N20 from aquaculture are included in this Inventory. At the present stage of
inventory development, Coastal Wetlands are not explicitly shown in the Land Representation analysis while work
continues to harmonize data from NOAA's Coastal Change Analysis Program (C-CAP)63 with NRI, FIA and NLDC data
used to compile the Land Representation. However, a check was undertaken to confirm that Coastal Wetlands
recognized by C-CAP represented a subset of Wetlands recognized by the NRI for marine coastal states.
The greenhouse gas fluxes for all four wetland categories described above are summarized in Table 6-56. Coastal
Wetlands Remaining Coastal Wetlands are generally a net C sink, with the fluxes ranging from -3.7 to -4.8 MMT
C02 Eq. across the majority of the time series, however, between 2006 and 2010 they were a net source of
emissions (ranging from of 5.2 to 5.5 MMT C02 Eq.), resulting from large loss of vegetated coastal wetlands to
open water due to hurricanes (Table 6-56). Recognizing removals of C02to soil of 10.2 MMT C02 Eq. and CH4
emissions of 3.8 MMT C02 Eq. in 2019, Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands are a
net sink of 6.4 MMT C02 Eq. Loss of coastal wetlands, primarily in the Mississippi Delta as a result of hurricane
impacts and sediment diversion and other human impacts, recognized as Vegetated Coastal Wetlands Converted
to Unvegetated Coastal Wetlands, drive an emission of 1.5 MMT C02 Eq. primarily from soils. Building of new
wetlands from open water, recognized as Unvegetated Coastal Wetlands Converted to Vegetated Coastal, results
each year in removal of 0.1 MMT C02 Eq. Aquaculture is a minor industry in the United States, resulting in an
emission of N20 across the time series of between 0.01 to 0.2 MMT C02 Eq. In all, Coastal Wetlands are a net sink
of 4.8 MMT C02 Eq. in 2019.
63 See ; accessed August 2020.
Land Use, Land-Use Change, and Forestry 6-97

-------
Table 6-56: Emissions and Removals from Coastal Wetlands Remaining Coastal Wetlands
(MMT COz Eq.)
Land Use/Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Vegetated Coastal Wetlands







Remaining Vegetated Coastal







Wetlands
(6.5)
(6.4)
(6.4)
(6.4)
(6.4)
(6.4)
(6.4)
Biomass C Flux
(+)
0.1
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
Soil C Flux
(10.2)
(10.2)
(10.2)
(10.2)
(10.2)
(10.2)
(10.2)
Net CH4 Flux
3.7
3.8
3.8
3.8
3.8
3.8
3.8
Vegetated Coastal Wetlands







Converted to Unvegetated Open







Water Coastal Wetlands
1.8
2.6
1.5
1.5
1.5
1.5
1.5
Biomass C Flux
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Dead Organic Matter C Flux
+
+
+
+
+
+
+
Soil C Flux
1.7
2.5
1.5
1.5
1.5
1.5
1.5
Unvegetated Open Water Coastal







Wetlands Converted to Vegetated







Coastal Wetlands
(+)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Biomass C Flux
(+)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Dead Organic Matter C Flux
(+)
(+)
0
0
0
0
0
Soil C Flux
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Net N20 Flux from Aquaculture in







Coastal Wetlands
0.1
0.2
0.1
0.1
0.1
0.1
0.1
Total Biomass C Flux
+
0.1
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Total Dead Organic Matter C Flux
(+)
(+)
+
+
+
+
+
Total Soil C Flux
(8.5)
(7.7)
(8.7)
(8.7)
(8.7)
(8.7)
(8.7)
Total CH4 Flux
3.7
3.8
3.8
3.8
3.8
3.8
3.8
Total N20 Flux
0.1
0.2
0.1
0.1
0.1
0.1
0.1
Total Flux
(4.6)
(3.7)
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
Note: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
Emissions and Removals from Vegetated Coastal Wetlands
Remaining Vegetated Coastal Wetlands
The conterminous United States currently has 2.98 million hectares of intertidal Vegetated Coastal Wetlands
Remaining Vegetated Coastal Wetlands comprised of tidally influenced palustrine emergent marsh (659,178 ha),
palustrine scrub shrub (132,934 ha) and estuarine emergent marsh (1,894,854 ha), estuarine scrub shrub (93,555
ha) and estuarine forest (195,646 ha). Mangroves fall under both estuarine forest and estuarine scrub shrub
categories depending upon height. Dwarf mangroves, found in Texas, do not attain the height status to be
recognized as Forest Land, and are therefore always classified within Vegetated Coastal Wetlands. Vegetated
Coastal Wetlands Remaining Vegetated Coastal Wetlands are found in cold temperate (53,976 ha), warm
temperate (895,976 ha), subtropical (1,963,565 ha) and Mediterranean (62,649 ha) climate zones.
Soils are the largest C pool in Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands, reflecting long-
term removal of atmospheric C02 by vegetation and transfer into the soil pool in the form of decaying organic
matter. Soil C emissions are not assumed to occur in coastal wetlands that remain vegetated. This Inventory
includes changes in biomass C stocks along with soils. Methane emissions from decomposition of organic matter in
anaerobic conditions are significant at salinity less than half that of sea water. Mineral and organic soils are not
differentiated in terms of C stock changes or CH4 emissions.
Table 6-57 through Table 6-59 below summarize nationally aggregated biomass and soil C stock changes and CH4
emissions on Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands. Intact Vegetated Coastal
Wetlands Remaining Vegetated Coastal Wetlands hold a total biomass C stock of 35.9 MMT C. Removals from
6-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
biomass C stocks in 2019 were 0.05 MMT C02 Eq. (0.01 MMT C), which has increased over the time series (Table
6-57 and Table 6-58). Carbon dioxide emissions from biomass in Vegetated Coastal Wetlands Remaining Vegetated
Coastal Wetlands between 2002 and 2011 are not inherently typical and are a result of coastal wetland loss over
time. Most of the coastal wetland loss has occurred in palustrine and estuarine emergent wetlands. Vegetated
coastal wetlands maintain a large C stock within the top 1 meter of soil (estimated to be 800 MMT C) to which C
accumulated at a rate of 10.2 MMT C02 Eq. (2.8 MMT C) in 2019, a value that has remained relatively constant
across the reporting period. For Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands, methane
emissions of 3.8 of MMT C02 Eq. (153 kt CH4) in 2019 (Table 6-59) offset C removals resulting in a net removal of
6.4 MMT C02 Eq. in 2019; this rate has been relatively consistent across the reporting period. Dead organic matter
stock changes are not calculated in Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands since this
stock is considered to be in a steady state when using Tier 1 methods (IPCC 2014). Due to federal regulatory
protection, loss of Vegetated Coastal Wetlands through human activities slowed considerably in the 1970s and the
current annual rates of C stock change and CH4 emissions are relatively constant over time.
Table 6-57: Net CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands (MMT CO2 Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
Biomass Flux
Soil Flux
(+)
(10.2)
0.1
(10.2)
(0.05)
(10.2)
(0.05)
(10.2)
(0.05)
(10.2)
(0.05)
(10.2)
(0.05)
(10.2)
Total C Stock Change
(10.2)
(10.2)
(10.2)
(10.2)
(10.2)
(10.2)
(10.2)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
Table 6-58: Net CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands (MMT C)
Year
1990
2005
2015
2016
2017
2018
2019
Biomass Flux
(+)
+
(+)
(+)
(+)
(+)
(+)
Soil Flux
(2.8)
(2.8)
(2.8)
(2.8)
(2.8)
(2.8)
(2.8)
Total C Stock Change
(2.8)
(2.8)
(2.8)
(2.8)
(2.8)
(2.8)
(2.8)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.05 MMT C.
Table 6-59: CH4 Emissions from Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands (MMT COz Eq. and kt CH4)
Year
1990
2005
2015
2016
2017
2018
2019
Methane Emissions (MMT C02 Eq.)
Methane Emissions (kt CH4)
3.7
149
I 3.8
; 151
3.8
152
3.8
153
3.8
153
3.8
153
3.8
153
Methodology
The following section includes a description of the methodology used to estimate changes in biomass C stocks, soil
C stocks and emissions of CH4 for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands. Dead
organic matter is not calculated for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands since it is
assumed to be in steady state (IPCC 2014).
Biomass Carbon Stock Changes
Above- and below ground biomass C Stocks for palustrine (freshwater) and estuarine (saline) marshes are
estimated for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands on land below the elevation of
high tides (taken to be mean high water spring tide elevation) and as far seawards as the extent of intertidal
vascular plants according to the national LiDAR dataset, the national network of tide gauges and land use histories
recorded in the 1996, 2001, 2006, 2010, and 2016 (new to this year's inventory) NOAA C-CAP surveys (NOAA OCM
Land Use, Land-Use Change, and Forestry 6-99

-------
2020). C-CAP areas are calculated at the state/territory level and summed according to climate zone to national
values. Federal and non-federal lands are represented. Trends in land cover change are extrapolated to 1990 and
2019 from these datasets. Based upon NOAA C-CAP, coastal wetlands are subdivided into palustrine and estuarine
classes and further subdivided into emergent marsh, scrub shrub and forest classes (Table 6-60). Biomass is not
sensitive to soil organic content but is differentiated based on climate zone. Aboveground biomass carbon stocks
for non-forested wetlands data are derived from a national assessment combining field plot data and aboveground
biomass mapping by remote sensing (Byrd et al. 2017; Byrd et al. 2018; Byrd et al. 2020). The aboveground
biomass carbon stock for estuarine forested wetlands (dwarf mangroves that are not classified as forests due to
their stature) is derived from a meta-analysis by Lu and Megonigal (2017;
Table 6-61). Root to shoot ratios from the Wetlands Supplement (Table 6-62; IPCC 2014) were used to account for
belowground biomass, which were multiplied by the aboveground carbon stock. Above- and belowground values
were summed to obtain total biomass carbon stocks. Biomass C stock changes per year for wetlands remaining
wetlands were determined by calculating the difference in area between that year and the previous year to
calculate gain/loss of area for each climate type, which was multiplied by the mean biomass for that climate type.
Table 6-60: Area of Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands,
Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands, and
Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands (ha)
Year	 1990	2005	2015 2016	2017 2018 2019
Vegetated Coastal Wetlands
Remaining Vegetated Coastal
Wetlands
Vegetated Coastal Wetlands
Converted to Unvegetated Open
Water Coastal Wetlands
Unvegetated Open Water Coastal
Wetlands Converted to
Vegetated Coastal Wetlands
2,985,512 2,988,258
1,720	2,515
953	1,775
2,971,102 2,972,368 2;
1,488 1,488
2,406 2,406
'3,634 2,974,900 2,976,166
1,488 1,488 1,488
2,406 2,406 2,406
Table 6-61: Aboveground Biomass Carbon Stocks for Vegetated Coastal Wetlands (t C ha1)


Climate Zone


Wetland Type
Cold Temperate
Warm Temperate
Subtropical
Mediterranean
Palustrine Scrub/Shrub Wetland
3.25
3.17
2.24
4.69
Palustrine Emergent Wetland
3.25
3.17
2.24
4.69
Estuarine Forested Wetland
3.05
3.10
17.83
3.44
Estuarine Scrub/Shrub Wetland
3.05
3.05
2.43
3.44
Estuarine Emergent Wetland
3.05
3.10
2.43
3.44
All data from Byrd et al. (2017, 2018 and 2020) except for subtropical estuarine forested wetlands, which is
from Lu and Megonigal (2017).
Table 6-62: Root to Shoot Ratios for Vegetated Coastal Wetlands


Climate Zone


Wetland Type
Cold Temperate
Warm Temperate
Subtropical
Mediterranean
Palustrine Scrub/Shrub Wetland
1.15
1.15
3.65
3.63
Palustrine Emergent Wetland
1.15
1.15
3.65
3.63
Estuarine Forested Wetland
1.15
1.15
0.96
3.63
Estuarine Scrub/Shrub Wetland
2.11
2.11
3.65
3.63
Estuarine Emergent Wetland
2.11
2.11
3.65
3.63
All values from IPCC (2014).
6-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Soil Carbon Stock Changes
Soil C stock changes are estimated for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands for
both mineral and organic soils. Soil C stock changes, stratified by climate zones and wetland classes, are derived
from a synthesis of peer-reviewed literature (Table 6-63; Lynch 1989; Orson et al. 1990; Kearny & Stevenson 1991;
Roman et al. 1997; Craft et al. 1998; Orson et al. 1998; Merrill 1999; Hussein et al. 2004; Church et al. 2006; Koster
et al. 2007; Callaway et al. 2012 a & b; Bianchi et al. 2013; Crooks et al. 2014; Weston et al. 2014; Villa & Mitsch
2015; Marchio et al. 2016; Noe et al. 2016).
Tier 2 level estimates of soil C removals associated with annual soil C accumulation on managed Vegetated Coastal
Wetlands Remaining Vegetated Coastal Wetlands were developed with country-specific soil C removal factors
multiplied by activity data of land area for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands.
The methodology follows Eq. 4.7, Chapter 4 of the Wetlands Supplement, and is applied to the area of Vegetated
Coastal Wetlands Remaining Vegetated Coastal Wetlands on an annual basis. To estimate soil C stock changes, no
differentiation is made between organic and mineral soils since currently no statistical evidence supports
disaggregation (Holmquist et al. 2018).
Table 6-63: Annual Soil Carbon Accumulation Rates for Vegetated Coastal Wetlands (t C ha1
yr1)
Climate Zone	Cold Temperate Warm Temperate Subtropical Mediterranean
Palustrine Scrub/Shrub Wetland
1.01
1.54
0.45
0.85
Palustrine Emergent Wetland
1.01
1.54
0.45
0.85
Estuarine Forested Wetland
1.01
0.82
0.87
0.85
Estuarine Scrub/Shrub Wetland
1.01
0.82
1.09
0.85
Estuarine Emergent Wetland
2.17
0.82
1.09
0.85
All data from Lu and Megonigal (2017)64
Soil Methane Emissions
Tier 1 estimates of CH4 emissions for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands are
derived from the same wetland map used in the analysis of wetland soil C fluxes, produced from C-CAP, LiDAR and
tidal data, in combination with default CH4 emission factors provided in Table 4.14 of the Wetlands Supplement.
The methodology follows Equation 4.9, Chapter 4 of the Wetlands Supplement; Tier 1 emissions factors are
multiplied by the area of freshwater (palustrine) coastal wetlands. The CH4 fluxes applied are determined based on
salinity; only palustrine wetlands are assumed to emit CH4. Estuarine coastal wetlands in the C-CAP classification
include wetlands with salinity less than 18 ppt, a threshold at which methanogenesis begins to occur (Poffenbarger
et al. 2011), but the dataset currently does not differentiate estuarine wetlands based on their salinities and as a
result CH4 emissions from estuarine wetlands are not included at this time.
Uncertainty and Time-Series Consistency
Underlying uncertainties in the estimates of soil and biomass C stock changes and CH4 emissions include
uncertainties associated with Tier 2 literature values of soil C stocks, biomass C stocks and CH4 flux, assumptions
that underlie the methodological approaches applied and uncertainties linked to interpretation of remote sensing
data. Uncertainty specific to Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands include
differentiation of palustrine and estuarine community classes, which determines the soil C stock and CH4 flux
applied. Uncertainties for soil and biomass C stock data for all subcategories are not available and thus
assumptions were applied using expert judgment about the most appropriate assignment of a C stock to a
disaggregation of a community class. Because mean soil and biomass C stocks for each available community class
are in a fairly narrow range, the same overall uncertainty was assigned to each, respectively (i.e., applying
64 See ; accessed October 2020.
Land Use, Land-Use Change, and Forestry 6-101

-------
approach for asymmetrical errors, where the largest uncertainty for any one soil C stock referenced using
published literature values for a community class; uncertainty approaches provide that if multiple values are
available for a single parameter, the highest uncertainty value should be applied to the propagation of errors; IPCC
2000). Uncertainty for root to shoot ratios, which are used for quantifying belowground biomass, are derived from
the 2013 Wetlands Supplement. Uncertainties for CH4 flux are the Tier 1 default values reported in the 2013 IPCC
Wetlands Supplement. Overall uncertainty of the NOAA C-CAP remote sensing product is 15 percent. This is in the
range of remote sensing methods (±10-15 percent; IPCC 2003). However, there is significant uncertainty in salinity
ranges for tidal and non-tidal estuarine wetlands and activity data used to apply CH4 flux emission factors
(delineation of an 18 ppt boundary) that will need significant improvement to reduce uncertainties. Details on the
emission/removal trends and methodologies through time are described in more detail in the introduction and the
Methodology section. The combined uncertainty was calculated using the IPCC Approach 1 method of summing
the squared uncertainty for each individual source (C-CAP, soil, biomass and CH4) and taking the square root of
that total.
Uncertainty estimates are presented in Table 6-64 and Table 6-65 for each subsource (i.e., soil C, biomass C and
CH4 emissions). The combined uncertainty across all subsources is +/-36.6 percent, which is primarily driven by the
uncertainty in the CH4 estimates because there is high variability in CH4 emissions when the salinity is less than 18
ppt. In 1990, the total flux was -6.5 MMT C02 Eq., with lower and upper estimates of-8.8 and -4.1 MMT C02 Eq. In
2019, the total flux was -6.4 MMT C02 Eq., with lower and upper estimates of-8.7 and -4.1 MMT C02 Eq.
Table 6-64: IPCC Approach 1 Quantitative Uncertainty Estimates for C Stock Changes and
CH4 Emissions occurring within Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlandsin 1990 (MMT CO2 Eq. and Percent)
Source
Gas
1990 Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Estimate
(MMT C02 Eq.) (%)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Biomass C Stock Change
Soil C Stock Change
CH4 emissions
C02
C02
ch4
(0.01)
(10.2)
3.7
(0.01)
(12.0)
2.6
(0.01)
(8.4)
4.8
-24.1%
-17.8%
-29.8%
24.1%
17.8%
29.8%
Total Flux

(6.5)
(8.8)
(4.1)
-36.6%
36.6%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding. + Absolute value
does not exceed 0.05 MMT C02 Eq.
Table 6-65: IPCC Approach 1 Quantitative Uncertainty Estimates for C Stock Changes and
CH4 Emissions occurring within Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlandsin 2019 (MMT CO2 Eq. and Percent)
Source
Gas
2019 Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Estimate
(MMT C02 Eq.) (%)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Biomass C Stock Change
Soil C Stock Change
CH4 emissions
C02
C02
CH4
(0.05)
(10.2)
3.8
(0.06)
(12.0)
2.7
(+)
(8.4)
5.0
-24.1%
-17.8%
-29.8%
24.1%
17.8%
29.8%
Total Flux

(6.4)
(8.7)
(4.1)
-36.6%
36.6%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding. + Absolute value
does not exceed 0.05 MMT C02 Eq.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
6-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
QA/QC and Verification
NOAA provided National LiDAR Dataset, tide data, and C-CAP land cover and land cover change mapping, all of
which are subject to agency internal QA/QC assessment. Acceptance of final datasets into archive and
dissemination are contingent upon the product compilation being compliant with mandatory QA/QC requirements
(McCombs et al. 2016). QA/QC and verification of soil C stock datasets have been provided by the Smithsonian
Environmental Research Center and Coastal Wetland Inventory team leads who reviewed summary tables against
reviewed sources. Biomass C stocks are derived from peer-review literature and reviewed by the U.S. Geological
Survey prior to publishing, by the peer-review process during publishing, and by the Coastal Wetland Inventory
team leads before inclusion in this Inventory. A team of two evaluated and verified there were no computational
errors within the calculation worksheets. Soil and biomass C stock change data are based upon peer-reviewed
literature and CH4 emission factors derived from the Wetlands Supplement.
Recalculations Discussion
As part of the addition of a 2016 C-CAP dataset, the previous datasets (1996, 2001, 2006, and 2011) were refined
and reanalyzed using the improved methods, software and techniques that were used for the 2016 dataset to
create the most accurate and representative product. Additionally, coastal wetland areas were calculated at the
state/territory level and summed based on climate zone to the national level. This change was implemented so
that emissions and removals could be calculated at the state/territory level.
The addition of the 2016 C-CAP dataset resulted in changes in area calculations for 1990 to 1995 and 2011 to 2018.
Previously, the average change across all C-CAP time periods (1996 through 2011) was used for years outside of
the C-CAP dataset. For 1990 to 1995 and 2017 to 2019, the C-CAP change data for the period closest to a given
year were used instead of the average change value. This is because there was largescale wetland loss between
2001 and 2006 due to major hurricanes in 2005, which resulted in elevated wetland loss that was not consistent
across years. For 1990 through 1995, the 1996 to 2000 C-CAP change data were used, and the 2011 to 2016 C-CAP
change data were used for 2017 through 2019. This resulted in the following area changes in 1990 (64,848 ha
increase in palustrine emergent marsh, 626 ha increase in palustrine scrub shrub marshes, 28,419 increase in
estuarine emergent marshes, 8,140 ha decrease in estuarine scrub shrub marshes, and 14,347 ha increase in
estuarine forested wetlands) and 2018 (54,438 ha increase in palustrine emergent marsh, 9,320 ha decrease in
palustrine scrub shrub marshes, 57,966 increase in estuarine emergent marshes, 4,388 ha decrease in estuarine
scrub shrub marshes, and 3,513 ha increase in estuarine forested wetlands).
Although this change does not affect the area or flux calculations presented here, a new organic soil geospatial
layer was derived from CONUS distributions of histosols in the USDA's Soil Survey Geographic Database (SSURGO).
The dataset is complete except for a large portion of the Florida Everglades did not contain data. Incorporating this
new dataset resulted in a large change in extents between organic and mineral soils, with less area of organic soils
nationally than reported previously.
A corrigendum was published by Byrd et al. (2020) for non-forested, emergent wetland aboveground biomass
values initially presented in Byrd et al. (2018). The updated analyses resulted in increases in aboveground biomass
carbon stocks across all wetland types and climate zones (except for the Puget Sound, which had a 0.11C ha 1
decrease); the average aboveground biomass increased from 1.93 to 3.021 C ha"1.
Belowground biomass carbon stock changes are now included for coastal wetlands and result in increased biomass
C stock fluxes. These values are based on Tier 2 root to shoot ratios from the Wetlands Supplement (IPCC 2014).
These recalculations did not result in a net change in removals for Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands; however, within individual subsource categories soil C accumulation increased by 0.3
MMT C02 Eq. in both 1990 and 2018 from -9.9 to -10.2 MMT C02 Eq., and net biomass carbon stocks increased by
0.01 MMT C02 Eq. in 1990 and changed towards an emission of 0.2 MMT C02 Eq. in 2018 as compared to the
previous Inventory.
Land Use, Land-Use Change, and Forestry 6-103

-------
Planned Improvements
Administered by the Smithsonian Environmental Research Center, the Coastal Wetland Carbon Research
Coordination Network has established a U.S. country-specific database of soil C stock and biomass estimates for
coastal wetlands.65 This dataset will be updated periodically. Refined error analysis combining land cover change
and C stock estimates will be provided as new data are incorporated. Through this work, a model is in
development to represent updated changes in soil C stocks for estuarine emergent wetlands.
Work is currently underway to examine the feasibility of incorporating seagrass soil and biomass C stocks into the
Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands estimates. Additionally, investigation into
quantifying the distribution, area, and emissions resulting from impounded waters (i.e., coastal wetlands where
tidal connection to the ocean has been restricted or eliminated completely) is underway.
Emissions from Vegetated Coastal Wetlands Converted to
negetated Open Water Coastal Wetlands
Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands is a source of emissions
from soil, biomass, and DOM C stocks. An estimated 1,488 ha of Vegetated Coastal Wetlands were converted to
Unvegetated Open Water Coastal Wetlands in 2019, which largely occurred within estuarine and palustrine
emergent wetlands. Prior to 2006, annual conversion to unvegetated open water coastal wetlands was higher than
current rates: 1,720 between 1990 and 2000 and 2,515 ha between 2001 and 2005. The Mississippi Delta
represents more than 40 percent of the total coastal wetland of the United States, and over 90 percent of the area
of Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands. The drivers of coastal
wetlands loss include legacy human impacts on sediment supply through rerouting river flow, direct impacts of
channel cutting on hydrology, salinity and sediment delivery, and accelerated subsidence from aquafer extraction.
Each of these drivers directly contributes to wetland erosion and subsidence, while also reducing the resilience of
the wetland to build with sea-level rise or recover from hurricane disturbance. Over recent decades, the rate of
Mississippi Delta wetland loss has slowed, though episodic mobilization of sediment occurs during hurricane
events (Couvillion et al. 2011; Couvillion et al. 2016). The most recent land cover analysis between the 2006 and
2011C-CAP surveys coincides with two such events, hurricanes Katrina and Rita (both making landfall in the late
summer of 2005), that occurred between these C-CAP survey dates. The subsequent 2016 C-CAP survey
determined that erosion rates had slowed.
Shallow nearshore open water within the U.S. Land Representation is recognized as falling under the Wetlands
category within this Inventory. While high resolution mapping of coastal wetlands provides data to support IPCC
Approach 2 methods for tracking land cover change, the depth in the soil profile to which sediment is lost is less
clear. This Inventory adopts the Tier 1 methodological guidance from the Wetlands Supplement for estimating
emissions following the methodology for excavation (see Methodology section, below) when Vegetated Coastal
Wetlands are converted to Unvegetated Open Water Coastal Wetlands, assuming aim depth of disturbed soil.
This 1 m depth of disturbance is consistent with estimates of wetland C loss provided in the literature and the
Wetlands Supplement (Crooks et al. 2009; Couvillion et al. 2011; Delaune and White 2012; IPCC 2014). The same
assumption on depth of soils impacted by erosion has been applied here. It is a reasonable Tier 1 assumption,
based on experience, but estimates of emissions are sensitive to the depth to which the assumed disturbances
have occurred (Holmquist et al. 2018). A Tier 1 assumption is also adopted in that all mobilized C is immediately
returned to the atmosphere (as assumed for terrestrial land use categories), rather than redeposited in long-term
C storage. The science is currently under evaluation to adopt more refined emissions factors for mobilized coastal
wetland C based upon the geomorphic setting of the depositional environment.
In 2019, there were 1,488 ha of Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal
Wetlands across all wetland types and climates, which resulted in 1.5 MMT C02 Eq. (0.4 MMT C) and 0.06 MMT
65 See ; accessed October 2020.
6-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
C02 Eq. (0.02 MMT C) lost through soil and biomass, respectively, while DOM C stock loss was present it was
minimal (Table 6-60, Table 6-66, and Table 6-67). Across the reporting period, the area of vegetated coastal
wetlands converted to unvegetated open water coastal wetlands was greatest between the 2006 to 2011 C-CAP
reporting period (11,373 ha) and has decreased since then to current levels (Table 6-60).
Table 6-66: Net CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands (MMT CO2 Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
Biomass Flux
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Dead Organic Matter Flux
+
+
+
+
+
+
+
Soil Flux
1.7
2.5
1.5
1.5
1.5
1.5
1.5
Total C Stock Change
1.8
2.6
1.5
1.5
1.5
1.5
1.5
Note: Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
Table 6-67: Net CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands (MMT C)
Year
1990
2005
2015
2016
2017
2018
2019
Biomass Flux
+
+
+
+
+
+
+
Dead Organic Matter Flux
+
+
+
+
+
+
+
Soil Flux
0.5
0.7
0.4
0.4
0.4
0.4
0.4
Total C Stock Change
0.5
0.7
0.4
0.4
0.4
0.4
0.4
Note: Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.05 MMT C.
Methodology
The following section includes a brief description of the methodology used to estimate changes in soil, biomass
and DOM C stocks for Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands.
Biomass Carbon Stock Changes
Biomass C stock changes for palustrine and estuarine marshes are estimated for Vegetated Coastal Wetlands
Converted to Unvegetated Open Water Coastal Wetlands on lands below the elevation of high tides (taken to be
mean high water spring tide elevation) within the U.S. Land Representation according to the national LiDAR
dataset, the national network of tide gauges and land use histories recorded in the 1996, 2001, 2006, 2010, and
2016 NOAA C-CAP surveys; the latter is new to this analysis. C-CAP areas are calculated at the state/territory level
and summed according to climate zone to national values. Publicly-owned and privately-owned lands are
represented. Trends in land cover change are extrapolated to 1990 and 2019 from these datasets. The C-CAP
database provides peer reviewed country-specific mapping to support IPCC Approach 3 quantification of coastal
wetland distribution, including conversion to and from open water. Biomass C stocks are not sensitive to soil
organic content but are differentiated based on climate zone. Non-forested aboveground biomass C stock data are
derived from a national assessment combining field plot data and aboveground biomass mapping by remote
sensing (Byrd et al. 2017; Byrd et al. 2018; Byrd et al. 2020). The aboveground biomass carbon stock for estuarine
forested wetlands (dwarf mangroves that are not classified as forests due to their stature) is derived from a meta-
analysis by Lu and Megonigal (201766; Table 6-61). Aboveground biomass C stock data for all subcategories are not
available and thus assumptions were applied using expert judgment about the most appropriate assignment of a C
stock to a disaggregation of a community class. Root to shoot ratios from the Wetlands Supplement were used to
account for belowground biomass, which were multiplied by the aboveground carbon stock (Table 6-62; IPCC
2014). Above- and belowground values were summed to obtain total biomass carbon stocks. Conversion to open
66 See ; accessed October 2020.
Land Use, Land-Use Change, and Forestry 6-105

-------
water results in emissions of all biomass C stocks during the year of conversion; therefore, emissions are calculated
by multiplying the C-CAP derived area of vegetated coastal wetlands lost that year in each climate zone by its
mean biomass.
Dead Organic Matter
Dead organic matter (DOM) carbon stocks, which include litter and dead wood stocks for subtropical estuarine
forested wetlands, are an emission from Vegetated Coastal Wetlands Converted to Unvegetated Open Water
Coastal Wetlands across all years in the time series. Data on DOM carbon stocks are not currently available for
either palustrine or estuarine scrub/shrub wetlands for any climate zone. Data for estuarine forested wetlands in
other climate zones are not included since there is no estimated loss of these forests to unvegetated open water
coastal wetlands across any year based on C-CAP data. For subtropical estuarine forested wetlands, Tier 1
estimates of mangrove DOM were used (IPCC 2014). Trends in land cover change are derived from the NOAA C-
CAP dataset and extrapolated to cover the entire 1990 to 2019 time series. Conversion to open water results in
emissions of all DOM C stocks during the year of conversion; therefore, emissions are calculated by multiplying the
C-CAP derived area of vegetated coastal wetlands lost that year by its Tier 1 DOM C stock.
Soil Carbon Stock Changes
Soil C stock changes are estimated for Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal
Wetlands. Country-specific soil C stocks were updated in 2018 based upon analysis of an assembled dataset of
1,959 cores from across the conterminous United States (Holmquist et al. 2018). This analysis demonstrated that it
was not justified to stratify C stocks based upon mineral or organic soil classification, climate zone, or wetland
classes; therefore, a single soil C stock of 2701C ha 1 was applied to all classes. Following the Tier 1 approach for
estimating C02 emissions with extraction provided within the Wetlands Supplement, soil C loss with conversion of
Vegetated Coastal Wetlands to Unvegetated Open Water Coastal Wetlands is assumed to affect soil C stock to
one-meter depth (Holmquist et al. 2018) with all emissions occurring in the year of wetland conversion, and
multiplied by activity data of vegetated coastal wetland area converted to unvegetated open water wetlands. The
methodology follows Eq. 4.6 in the Wetlands Supplement.
Soil Methane Emissions
A Tier 1 assumption has been applied that salinity conditions are unchanged and hence CH4 emissions are assumed
to be zero with conversion of Vegetated Coastal Wetlands to Unvegetated Open Water Coastal Wetlands.
Uncertainty and Time-Series Consistency
Underlying uncertainties in estimates of soil and biomass C stock changes are associated with country-specific (Tier
2) literature values of these stocks, while the uncertainties with the Tier 1 estimates are associated with
subtropical estuarine forested wetland DOM stocks. Assumptions that underlie the methodological approaches
applied and uncertainties linked to interpretation of remote sensing data are also included in this uncertainty
assessment. The IPCC default assumption of 1 m of soil erosion with anthropogenic activities was adopted to
provide standardization in U.S. tidal carbon accounting (Holmquist et al. 2018). This depth of potentially erodible
tidal wetland soil has not been comprehensively addressed since most soil cores analyzed were shallow (e.g., less
than 50 cm) and do not necessarily reflect the depth to non-wetland soil or bedrock (Holmquist et al. 2018).
Uncertainty specific to coastal wetlands include differentiation of palustrine and estuarine community classes,
which determines the soil C stock applied. Because mean soil and biomass C stocks for each available community
class are in a fairly narrow range, the same overall uncertainty was assigned to each (i.e., applying approach for
asymmetrical errors, where the largest uncertainty for any one soil C stock referenced using published literature
values for a community class; if multiple values are available for a single parameter, the highest uncertainty value
should be applied to the propagation of errors; IPCC 2000). For aboveground biomass C stocks, the mean standard
error was very low and largely influenced by the uncertainty associated with the estimated map area (Byrd et al.
2018). Uncertainty for root to shoot ratios, which are used for quantifying belowground biomass, are derived from
the Wetlands Supplement. Uncertainty for subtropical estuarine forested wetland DOM stocks was derived from
6-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
those listed for the Tier 1 estimates (IPCC 2014). Overall uncertainty of the NOAA C-CAP remote sensing product is
15 percent. This is in the range of remote sensing methods (±10-15 percent; IPCC 2003). The combined uncertainty
was calculated by summing the squared uncertainty for each individual source (C-CAP, soil, biomass, and DOM)
and taking the square root of that total.
Details on the emission/removal trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
Uncertainty estimates are presented in Table 6-69 and Table 6-69 for each subsource (i.e., soil C, biomass C, and
DOM emissions). The combined uncertainty across all subsources is +/-32.0 percent, which is driven by the
uncertainty in the soil C estimates. In 1990, the total C flux was 1.8 MMT C02 Eq., with lower and upper estimates
of 1.2 and 2.3 MMT C02 Eq. In 2019, the total C flux was 1.5 MMT C02 Eq., with lower and upper estimates of 1.0
and 2.0 MMT C02 Eq.
Table 6-68: Approach 1 Quantitative Uncertainty Estimates for CO2 Flux Occurring within
Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands in
1990 (MMT CO2 Eq. and Percent)
Source
1990 Flux Estimate
Uncertainty Range Relative to Flux Estimate
(MMTCOz Eq.)
(MMT CO
2 Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Biomass C Stock
0.07
0.06
0.09
-24.1%
24.1%
Dead Organic Matter C Stock
+
+
+
-25.8%
25.8%
Soil C Stock
1.7
1.4
2.0
-15.0%
15.0%
Total Flux
1.8
1.2
2.3
-32.0%
32.0%
Note: Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.0005 MMT C02 Eq.
Table 6-69: Approach 1 Quantitative Uncertainty Estimates for CO2 Flux Occurring within
Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands in
2019 (MMT CO2 Eq. and Percent)
Source
2019 Flux Estimate
Uncertainty Range Relative to Flux Estimate
(MMTCOz Eq.)
(MMT CO
2 Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Biomass C Stock
0.06
0.05
0.08
-24.1%
24.1%
Dead Organic Matter C Stock
0.0005
+
0.001
-25.8%
25.8%
Soil C Stock
1.5
1.3
1.7
-15.0%
15.0%
Total Flux
1.5
1.0
2.0
-32.0%
32.0%
Note: Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.0005 MMT C02 Eq.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
Data provided by NOAA (i.e., National LiDAR Dataset, NOS Tide Data, and C-CAP land cover and land cover change
mapping) undergo internal agency QA/QC procedures. Acceptance of final datasets into archive and dissemination
are contingent upon assurance that the data product is compliant with mandatory NOAA QA/QC requirements
(McCombs et al. 2016). QA/QC and Verification of the soil C stock dataset have been provided by the Smithsonian
Environmental Research Center and by the Coastal Wetlands project team leads who reviewed the estimates
Land Use, Land-Use Change, and Forestry 6-107

-------
against primary scientific literature. Biomass C stocks are derived from peer-review literature and reviewed by the
U.S. Geological Survey prior to publishing, by the peer-review process during publishing, and by the Coastal
Wetland Inventory team leads before inclusion in the Inventory. For subtropical estuarine forested wetlands, Tier 1
estimates of mangrove DOM were used (IPCC 2014) Land cover estimates were assessed to ensure that the total
land area did not change over the time series in which the inventory was developed, and were verified by a second
QA team. A team of two evaluated and verified there were no computational errors within the calculation
worksheets.
Recalculations Discussion
For a full discussion on the recalculations implemented for this Inventory, please see the Recalculations section for
Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands. For Vegetated Coastal Wetlands Converted
to Unvegetated Open Water Wetlands, these changes resulted in a reduction of 3,107 and 3,340 ha of vegetated
coastal wetlands converted to unvegetated open water coastal wetlands in 1990 and 2018, respectively. In
comparison to the previous Inventory, this decreased emissions by 3.1 and 3.3 MMT C02 Eq. (64.0 and 68.2
percent) in 1990 and 2018, respectively. Between 2006 and 2010, the period with the largest area conversion, an
increase in 1,664 ha per year occurred (from 9,709 to 11,373 ha for the previous and current Inventories,
respectively), resulting in a 1.7 and 0.4 MMT C02 Eq. increase in soil and biomass C emissions, respectively (17 and
496 percent); DOM emissions decreased by 0.0001 MMT C02 Eq. (3.6 percent). The large increase in biomass
emissions is driven by the incorporation of belowground biomass in addition to updates to the C-CAP dataset; the
small change in DOM emissions reflects the minimal reduction in the areal extent of estuarine forested wetlands
mapped with C-CAP recalculations.
Planned Improvements
The depth of soil carbon affected by conversion of vegetated coastal wetlands converted to unvegetated coastal
wetlands will be updated from the IPCC default assumption of 1 m of soil erosion when mapping and modeling
advancements can quantitatively improve accuracy and precision. Until the time where these more detailed and
spatially distributed data are available, the IPCC default assumption that the top 1 m of soil is disturbed by
anthropogenic activity will be applied.
More detailed research is in development that provides a longer-term assessment and more highly refined rates of
wetlands loss across the Mississippi Delta (e.g., Couvillion et al. 2016). The Mississippi Delta is the largest extent of
coastal wetlands in the US. Higher resolution imagery analysis would improve quantification of conversation to
open water, which occurs not only at the edge of the marsh but also within the interior. Improved mapping could
provide a more refined regional Approach 2-3 land representation to support the national-scale assessment
provided by C-CAP.
An approach for calculating the fraction of remobilized coastal wetland soil C returned to the atmosphere as C02 is
currently under review and may be included in future reports.
Research by USGS is investigating higher resolution mapping approaches to quantify conversion of coastal
wetlands is also underway. Such approaches may form the basis for a full Approach 3 land representation
assessment in future years. C-CAP data harmonization with the National Land Cover Dataset (NLCD) will be
incorporated into a future iteration of the inventory.
Stock Changes from Unvegetated Open Water Coastal
Wetlands Converted to Vegetated Coastal Wetlands
Open water within the U.S. land base, as described in the 6 Land Representation section, is recognized as Coastal
Wetlands within this Inventory. The appearance of vegetated tidal wetlands on lands previously recognized as
open water reflects either the building of new vegetated marsh through sediment accumulation or the transition
from other lands uses through an intermediary open water stage as flooding intolerant plants are displaced and
6-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
then replaced by wetland plants. Biomass, DOM and soil C accumulation on Unvegetated Open Water Coastal
Wetlands Converted to Vegetated Coastal Wetlands begins with vegetation establishment.
Within the United States, conversion of Unvegetated Open Water Coastal Wetlands to Vegetated Coastal
Wetlands is predominantly due to engineered activities, which include active restoration of wetlands (e.g.,
wetlands restoration in San Francisco Bay), dam removals or other means to reconnect sediment supply to the
nearshore (e.g., Atchafalaya Delta, Louisiana, Couvillion et al. 2011). Wetlands restoration projects have been
ongoing in the United States since the 1970s. Early projects were small, a few hectares in size. By the 1990s,
restoration projects, each hundreds of hectares in size, were becoming common in major estuaries. In several
coastal areas e.g., San Francisco Bay, Puget Sound, Mississippi Delta and south Florida, restoration activities are in
planning and implementation phases, each with the goal of recovering tens of thousands of hectares of wetlands.
In 2019, 2,407 ha of unvegetated open water coastal wetlands were converted to vegetated coastal wetlands
across all wetland types and climates, which has steadily increased over the reporting period (Table 6-59). This
resulted in 0.007 MMT C02 Eq. (0.002 MMT C) and 0.1 MMT C02 Eq. (0.03 MMT C) sequestered in soil and
biomass, respectively (Table 6-70 and Table 6-71). The soil C stock has increased during the Inventory's reporting
period, likely due to increasing vegetated coastal wetland restoration over time. While DOM C stock increases are
present they are minimal in the early part of the time series and zero in the later because there are no conversions
from unvegetated open water coastal wetlands to subtropical estuarine forested wetlands between 2011 and
2016 (and by proxy through 2019), and that is the only coastal wetland type where DOM data is currently
available.
Throughout the reporting period, the amount of Open Water Coastal Wetlands Converted to Vegetated Coastal
Wetlands has increased over time, reflecting the increase in engineered restoration activities mentioned above.
Table 6-70: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands
Converted to Vegetated Coastal Wetlands (MMT CO2 Eq.)
Year
1990
2005
2015
2016
2017
2018
2019
Biomass C Flux
(0.04)
(0.08)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Dead Organic Matter C Flux
(+)
(+)
0
0
0
0
0
Soil C Flux
(+)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
Total C Stock Change
(0.04)
(0.09)
(0.11)
(0.11)
(0.11)
(0.11)
(0.11)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.005 MMT C02 Eq.
Table 6-71: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands
Year
1990
2005
2015
2016
2017
2018
2019
Biomass C Flux
(0.01)
(0.02)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Dead Organic Matter C Flux
(+)
(+)
0
0
0
0
0
Soil C Flux
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Total C Stock Change
(0.01)
(0.02)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.005 MMT C.
Methodology
The following section includes a brief description of the methodology used to estimate changes in soil, biomass
and DOM C stocks, and CH4 emissions for Unvegetated Open Water Coastal Wetlands Converted to Vegetated
Coastal Wetlands.
Biomass Carbon Stock Changes
Quantification of regional coastal wetland biomass C stock changes for palustrine and estuarine marsh vegetation
are presented for Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands on lands
Land Use, Land-Use Change, and Forestry 6-109

-------
below the elevation of high tides (taken to be mean high water spring tide elevation) according to the national
LiDAR dataset, the national network of tide gauges and land use histories recorded in the 1996, 2001, 2005, 2011,
and 2016 NOAA C-CAP surveys. C-CAP areas are calculated at the state/territory level and summed according to
climate zone to national values. Privately-owned and publicly-owned lands are represented. Trends in land cover
change are extrapolated to 1990 and 2019 from these datasets (Table 6-58). C-CAP provides peer reviewed
country-level mapping of coastal wetland distribution, including conversion to and from open water. Biomass C
stock is not sensitive to soil organic content but differentiated based on climate zone. Data for non-forested
wetlands are derived from a national assessment combining field plot data and aboveground biomass mapping by
remote sensing (Table 6-61; Byrd et al. 2017; Byrd et al., 2018; Byrd et al. 2020). The aboveground biomass carbon
stock for estuarine forested wetlands (dwarf mangroves that are not classified as forests due to their stature) is
derived from a meta-analysis by Lu and Megonigal (20 1767;
Table 6-61). Aboveground biomass C stock data for all subcategories are not available and thus assumptions were
applied using expert judgment about the most appropriate assignment of a C stock to a disaggregation of a
community class. Root to shoot ratios from the Wetlands Supplement were used to account for belowground
biomass, which were multiplied by the aboveground carbon stock (Table 6-62; IPCC 2014). Above- and
belowground values were summed to obtain total biomass carbon stocks.
Conversion of open water to Vegetated Coastal Wetlands results in the establishment of a standing biomass C
stock; therefore, stock changes that occur are calculated by multiplying the C-CAP derived area gained that year in
each climate zone by its mean biomass. While the process of revegetation of unvegetated open water wetlands
can take many years to occur, it is assumed in the calculations that the total biomass is reached in the year of
conversion.
Dead Organic Matter
Dead organic matter (DOM) carbon stocks, which include litter and dead wood stocks, are included for subtropical
estuarine forested wetlands for Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal
Wetlands across all years. Tier 1 default or country-specific data on DOM are not currently available for either
palustrine or estuarine scrub/shrub wetlands for any climate zone. Data for estuarine forested wetlands in other
climate zones are not included since there is no estimated loss of these forests to unvegetated open water coastal
wetlands across any year based on C-CAP data. Tier 1 estimates of subtropical estuarine forested wetland DOM
were used (IPCC 2014). Trends in land cover change are derived from the NOAA C-CAP dataset and extrapolated to
cover the entire 1990 through 2019 time series. Dead organic matter removals are calculated by multiplying the C-
CAP derived area gained that year by its Tier 1 DOM C stock. Similar to biomass C stock gains, gains in DOM can
take many years to occur, but for this analysis, the total DOM stock is assumed to accumulate during the first year
of conversion.
Soil Carbon Stock Change
Soil C stock changes are estimated for Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal
Wetlands. Country-specific soil C removal factors associated with soil C accretion, stratified by climate zones and
wetland classes, are derived from a synthesis of peer-reviewed literature and updated this year based upon
refined review of the dataset (Lynch 1989; Orson et al. 1990; Kearny & Stevenson 1991; Roman et al. 1997; Craft et
al. 1998; Orson et al. 1998; Merrill 1999; Hussein et al. 2004; Church et al. 2006; Koster et al. 2007; Callaway et al.
2012 a & b; Bianchi et al. 2013; Crooks et al. 2014; Weston et al. 2014; Villa & Mitsch 2015; Marchio et al. 2016;
Noe et al. 2016). Soil C stock changes are stratified based upon wetland class (Estuarine, Palustrine) and subclass
(Emergent Marsh, Scrub Shrub). For soil C stock change no differentiation is made for soil type (i.e., mineral,
organic). Soil C removal factors were developed from literature references that provided soil C removal factors
disaggregated by climate region and vegetation type by salinity range (estuarine or palustrine) as identified using
NOAA C-CAP as described above (see Table 6-63 for values).
67 See ; accessed October 2020.
6-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Tier 2 level estimates of C stock changes associated with annual soil C accumulation in Vegetated Coastal Wetlands
were developed using country-specific soil C removal factors multiplied by activity data on Unvegetated Coastal
Wetlands converted to Vegetated Coastal Wetlands. The methodology follows Eq. 4.7, Chapter 4 of the Wetlands
Supplement, and is applied to the area of Unvegetated Coastal Wetlands converted to Vegetated Coastal Wetlands
on an annual basis.
Soil Methane Emissions
A Tier 1 assumption has been applied that salinity conditions are unchanged and hence CH4 emissions are assumed
to be zero with conversion of Vegetated Open Water Coastal Wetlands to Vegetated Coastal Wetlands.
Uncertainty and Time-Series Consistency
Underlying uncertainties in estimates of soil and biomass C stock changes include uncertainties associated with
country-specific (Tier 2) literature values of these C stocks and assumptions that underlie the methodological
approaches applied and uncertainties linked to interpretation of remote sensing data. Uncertainty specific to
coastal wetlands include differentiation of palustrine and estuarine community classes that determines the soil C
stock applied. Because mean soil and biomass C stocks for each available community class are in a fairly narrow
range, the same overall uncertainty was applied to each, respectively (i.e., applying approach for asymmetrical
errors, where the largest uncertainty for any one soil C stock referenced using published literature values for a
community class; uncertainty approaches provide that if multiple values are available for a single parameter, the
highest uncertainty value should be applied to the propagation of errors; IPCC 2000). For aboveground biomass C
stocks, the mean standard error was very low and largely influenced by error in estimated map area (Byrd et al.
2018). Uncertainty for root to shoot ratios, which are used for quantifying belowground biomass (Table 6-62), are
derived from the Wetlands Supplement. Uncertainty for subtropical estuarine forested wetland DOM stocks were
derived from those listed for the Tier 1 estimates (IPCC 2014). Overall uncertainty of the NOAA C-CAP remote
sensing product is 15 percent. This is in the range of remote sensing methods (±10 to 15 percent; IPCC 2003). The
combined uncertainty was calculated by summing the squared uncertainty for each individual source (C-CAP, soil,
biomass, and DOM) and taking the square root of that total.
Uncertainty estimates are presented in Table 6-73 and Table 6-77 for each subsource (i.e., soil C, biomass C and
DOM emissions). The combined uncertainty across all subsources is ±33.4 percent. In 1990, the total C flux was -
0.04 MMT C02 Eq., with lower and upper estimates of -0.06 and -0.03 MMT C02 Eq. In 2019, the total C flux was -
0.1 MMT C02 Eq., with lower and upper estimates of-0.1 and -0.07 MMT C02 Eq.
Table 6-72: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes Occurring
within Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands
in 1990 (MMT CO2 Eq. and Percent)
Source
1990 Flux Estimate
Uncertainty Range
Relative to Flux Estimate
(MMT CO? Eq.)
(MMT C02
Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Biomass C Stock Flux
(0.4)
(0.05)
(0.03)
-20.0%
20.0%
Dead Organic Matter C Stock Flux
(+)
(+)
(+)
-25.8%
25.8%
Soil C Stock Flux
(0.003)
(0.003)
(0.005)
-17.8%
17.8%
Total Flux
(0.04)
(0.06)
(0.03)
-33.4%
33.4%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.0005 MMT C02 Eq.
Land Use, Land-Use Change, and Forestry 6-111

-------
Table 6-73: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes Occurring
within Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands
in 2019 (MMT CO2 Eq. and Percent)
Source
2019 Flux Estimate
Uncertainty Range
Relative to Flux Estimate
(MMT CO? Eq.)
(MMT C02
Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Biomass C Stock Flux
(0.1)
(0.1)
(0.08)
-20.0%
20.0%
Dead Organic Matter C Stock Flux
0
0
0
-25.8%
25.8%
Soil C Stock Flux
(0.007)
(0.008)
(0.005)
-17.8%
17.8%
Total Flux
(0.1)
(0.1)
(0.07)
-33.4%
33.4%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
NOAA provided data (i.e., National LiDAR Dataset, NOS Tide Data, and C-CAP land cover and land cover change
mapping), which undergo internal agency QA/QC assessment procedures. Acceptance of final datasets into the
archive for dissemination are contingent upon assurance that the product is compliant with mandatory NOAA
QA/QC requirements (McCombs et al. 2016). QA/QC and Verification of soil C stock dataset has been provided by
the Smithsonian Environmental Research Center and Coastal Wetlands project team leads who reviewed the
summary tables against primary scientific literature. Aboveground biomass C reference stocks are derived from an
analysis by the Blue Carbon Monitoring project and reviewed by U.S. Geological Survey prior to publishing, the
peer-review process during publishing, and the Coastal Wetland Inventory team leads before inclusion in the
inventory. Root to shoot ratios and DOM data are derived from peer-reviewed literature and undergo review as
per IPCC methodology. Land cover estimates were assessed to ensure that the total land area did not change over
the time series in which the inventory was developed and verified by a second QA team. A team of two evaluated
and verified there were no computational errors within calculation worksheets. Two biogeochemists at the USGS,
also members of the NASA Carbon Monitoring System Science Team, corroborated the simplifying assumption that
where salinities are unchanged CH4 emissions are constant with conversion of Unvegetated Open Water Coastal
Wetlands to Vegetated Coastal Wetlands.
Recalculations Discussion
For discussion on recalculations implemented for this Inventory, please see the Recalculations section for
Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands. For Unvegetated Open Water Coastal Waters
Converted to Vegetated Coastal Wetlands, these changes resulted in a decrease of 541 ha of unvegetated coastal
wetlands converted to vegetated coastal wetlands and an increased removal of 0.02 MMT C02 Eq. (50 percent) in
1990 and a decrease of 913 ha of unvegetated coastal wetlands converted to vegetated coastal wetlands and an
increased removal of 0.09 MMT C02 Eq. (450 percent) in 2018, when compared to the 2020 NIR submission.
Planned Improvements
Administered by the Smithsonian Environmental Research Center, the Coastal Wetland Carbon Research
Coordination Network has established a U.S. country-specific database of published data quantifying soil C stock
and biomass in coastal wetlands. Reference values for soil and biomass C stocks will be updated as new data
emerge. Refined error analysis combining land cover change, soil and biomass C stock estimates will be updated at
those times.
The USGS is investigating higher resolution mapping approaches to quantify conversion of coastal wetlands. Such
approaches may form the basis for a full Approach 3 land representation assessment in future years. C-CAP data
6-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
harmonization with the National Land Cover Dataset (NLCD) will be incorporated into a future iteration of the
inventory.
N20 Emissions from Aquaculture in Coastal Wetlands
Shrimp and fish cultivation in coastal areas increases nitrogen loads resulting in direct emissions of N20. Nitrous
oxide is generated and emitted as a byproduct of the conversion of ammonia (contained in fish urea) to nitrate
through nitrification and nitrate to N2 gas through denitrification (Hu et al. 2012). Nitrous oxide emissions can be
readily estimated from data on fish production (IPCC 2014).
Aquaculture production in the United States has fluctuated slightly from year to year, with resulting N20 emissions
increasing from 0.1 in 1990 to upwards of 0.2 MMT C02 Eq. between 1992 and 2010, and reducing again to 0.1
MMT C02 Eq. between 2015 and 2019 (Table 6-74). Aquaculture production data were updated through 2017;
data through 2019 are not yet available and in this analysis are held constant with 2017 emissions of 0.1 MMT C02
Eq. (0.5 Kt N20).
Table 6-74: N2O Emissions from Aquaculture in Coastal Wetlands (MMT CO2 Eq. and kt N2O)
Year
1990
2005
2015
2016
2017
2018
2019
Emissions (MMT C02 Eq.)
0.1
1 0.2
0.1
0.1
0.1
0.1
0.1
Emissions (kt N20)
0.4
0.6 ¦
0.5
0.5
0.5
0.5
0.5
Methodology
The methodology to estimate N20 emissions from Aquaculture in Coastal Wetlands follows the Tier 1 guidance in
the Wetlands Supplement by applying country-specific fisheries production data and the IPCC Tier 1 default
emission factor.
Each year NOAA Fisheries document the status of U.S. marine fisheries in the annual report of Fisheries of the
United States (National Marine Fisheries Service 2020), from which activity data for this analysis is derived.68 The
fisheries report has been produced in various forms for more than 100 years, primarily at the national level, on
U.S. recreational catch and commercial fisheries landings and values. In addition, data are reported on U.S.
aquaculture production, the U.S. seafood processing industry, imports and exports offish-related products, and
domestic supply and per capita consumption of fisheries products. Within the aquaculture chapter, the mass of
production for catfish, striped bass, tilapia, trout, crawfish, salmon and shrimp are reported. While some of these
fisheries are produced on land and some in open water cages within coastal wetlands, all have data on the
quantity of food stock produced, which is the activity data that is applied to the IPCC Tier 1 default emissions
factor to estimate emissions of N20 from aquaculture. It is not apparent from the data as to the amount of
aquaculture occurring above the extent of high tides on river floodplains. While some aquaculture occurs on
coastal lowland floodplains, this is likely a minor component of tidal aquaculture production because of the need
for a regular source of water for pond flushing. The estimation of N20 emissions from aquaculture is not sensitive
to salinity using IPCC approaches and as such the location of aquaculture ponds within the boundaries of coastal
wetlands does not influence the calculations.
Other open water shellfisheries for which no food stock is provided, and thus no additional N inputs, are not
applicable for estimating N20 emissions (e.g., clams, mussels, and oysters) and have not been included in the
analysis. The IPCC Tier 1 default emissions factor of 0.00169 kg N20-N per kg of fish/shellfish produced is applied to
the activity data to calculate total N20 emissions.
68 See ; accessed October 2020.
Land Use, Land-Use Change, and Forestry 6-113

-------
Uncertainty and Time-Series Consistency
Uncertainty estimates are based upon the Tier 1 default 95 percent confidence interval provided in Table 4.15,
chapter 4 of the Wetlands Supplement for N20 emissions and on expert judgment of the NOAA Fisheries of the
United States fisheries production data. Given the overestimate of fisheries production from coastal wetland areas
due to the inclusion offish production in non-coastal wetland areas, this is a reasonable initial first approximation
for an uncertainty range.
Uncertainty estimates for N20 emissions from aquaculture production are presented in Table 6-75 and Table 6-76
for N20 emissions. The combined uncertainty is ±116 percent. In 1990, the total flux was 0.1 MMT C02 Eq., with
lower and upper estimates of 0.00 and 0.28 MMT C02 Eq. In 2019, the total flux was 0.1 MMT C02 Eq., with lower
and upper estimates of 0.00 and 0.31 MMT C02 Eq.
Table 6-75: Approach 1 Quantitative Uncertainty Estimates for N2O Emissions from
Aquaculture Production in Coastal Wetlands in 1990 (MMT CO2 Eq. and Percent)

1990 Emissions



Estimate
Uncertainty Range Relative to Emissions Estimate3
Source
(MMTC02 Eq.)
(MMT C02 Eq.)
(%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
Combined Uncertainty for N20 Emissions



for Aquaculture Production in Coastal
0.1
0.00 0.28
-116% 116%
Wetlands



a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Table 6-76: Approach 1 Quantitative Uncertainty Estimates for N2O Emissions from
Aquaculture Production in Coastal Wetlands in 2019 (MMT CO2 Eq. and Percent)

2019 Emissions





Estimate
Uncertainty Range Relative to Emissions Estimate3
Source
(MMTCOz Eq.)
(MMT CO?
Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Combined Uncertainty for N20 Emissions





for Aquaculture Production in Coastal
0.1
0.00
0.31
-116%
116%
Wetlands





a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
NOAA provided internal QA/QC review of reported fisheries data. The Coastal Wetlands Inventory team consulted
with the Coordinating Lead Authors of the Coastal Wetlands chapter of the Wetlands Supplement to assess which
fisheries production data to include in estimating emissions from aquaculture. It was concluded that N20 emissions
estimates should be applied to any fish production to which food supplement is supplied be they pond or coastal
open water and that salinity conditions were not a determining factor in production of N20 emissions.
Recalculations Discussion
A NOAA report was released in 2020 that contains updated fisheries data through 2017 (National Marine Fisheries
Service 2020). The new values were applied for 2015, 2016, and 2017 and the 2017 value was applied in 2018 and
6-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
2019. This resulted in an increase of N20 emissions by 0.01 MMT C02 Eq. (0.03 kt N20), a 7.4 percent increase, for
2018 compared to the previous Inventory.
6.9 Land Converted to Wetlands (CRF Source
Category 4D2)
Emissions and Removals from Land Converted to Vegetated
Coastal Wetlands
Land Converted to Vegetated Coastal Wetlands occurs as a result of inundation of unprotected low-lying coastal
areas with gradual sea-level rise, flooding of previously drained land behind hydrological barriers, and through
active restoration and creation of coastal wetlands through removal of hydrological barriers. All other land
categories (i.e., Forest Land, Cropland, Grassland, Settlements and Other Lands) are identified as having some area
converting to Vegetated Coastal Wetlands. Between 1990 and 2019 the rate of annual transition for Land
Converted to Vegetated Coastal Wetlands ranged from 0 to 2,650 ha per year, depending on the type of land
converted.69 Conversion rates from Forest Land were relatively consistent between 1990 and 2010 (ranging
between 2,409 and 2,650 ha) and decreased to 625 ha starting in 2011; the majority of these conversions resulted
in increases in the area of palustrine wetlands, which also initiates CH4 emissions when lands are inundated with
fresh water.70 Little to no conversion of Cropland, Grassland, Settlement, or Other Lands to vegetated coastal
wetlands occurred during the reporting period, with converted areas ranging from 0 to 25 ha per year.
At the present stage of Inventory development, Coastal Wetlands are not explicitly shown in the Land
Representation analysis while work continues harmonizing data from NOAA's Coastal Change Analysis Program (C-
CAP) with NRI, FIA and NLDC data used to compile the Land Representation (NOAA OCM 2020).
In this Inventory, biomass, dead organic material (DOM) and soil C stock changes as well as CH4 emissions are
quantified as a result of the land use conversion to coastal wetlands and the land is assumed to be held in this
category for up to 20 years after which it is classified as Coastal Wetlands Remaining Coastal Wetlands. Estimates
of emissions and removals are based on emission factor data that have been applied to assess changes in each
respective flux for Land Converted to Vegetated Coastal Wetlands. Following conversion to Vegetated Coastal
Wetlands, it is assumed there is a loss of biomass C stocks from the converted Forest Land, Cropland and Grassland
and as well as the loss of DOM C stocks from Forest Land. Converted lands are held in this land category for up to
20 years and the assumption is that the C stock losses from biomass and DOM all occur in the year of conversion.
There are no soil C losses from land use conversion. Carbon stock increases in coastal wetlands as a result of gains
in plant biomass and DOM on these converted lands are also included during the year of transition even though
the entire C stock accrual takes many years to occur. Soil C accumulation and CH4 emissions are quantified using an
annual rate in this Inventory and thus are occurring over the period under which lands are held in this category;
therefore, the soil C removals and CH4 emissions presented for a given year include the cumulative
removals/emissions for the new area that was converted during that year and the area held in this category for the
prior 19 years. At salinities less than half that of seawater, the transition from upland dry soils to wetland soils
results in CH4emissions. The United States calculates emissions and removals based upon stock change.
69	Data from C-CAP; see . Accessed October 2020.
70	Currently, the C-CAP dataset categorizes coastal wetlands as either palustrine (fresh water) or estuarine (presence of saline
water). This classification does not differentiate between estuarine wetlands with salinity < 18 ppt (when methanogenesis
begins to occur) and those that are >18 ppt (where negligible to no CH4 is produced); therefore, it is not possible at this time to
account for CH4 emissions from estuarine wetlands in the Inventory.
Land Use, Land-Use Change, and Forestry 6-115

-------
Conversion to coastal wetlands resulted in a biomass C stock loss of 0.1 MMT C02 Eq. (0.03 MMT C) in 2019 (Table
6-77 and Table 6-78). Loss of forest biomass through conversion of Forest Lands to Vegetated Coastal Wetlands is
the primary driver behind biomass C stock change being a source rather than a sink across the time series.
Conversion of Cropland, Grassland, Settlement and Other Lands result in a net increase in biomass stocks.
Conversion of lands to vegetated coastal wetlands resulted in a DOM loss of 0.02 MMT C02 Eq. (0.006 MMT C) in
2019 (Table 6-77 and Table 6-78), which is driven by the loss of DOM when Forest Land is converted to Vegetated
Coastal Wetlands. This is likely an overestimate of loss because wetlands inherently preserve dead organic
material. Conversion of Cropland, Grassland, Settlement and Other Land results in a net increase in DOM. Once
Tier 1 or 2 DOM values are collated and accounted for in estuarine and palustrine scrub shrub coastal wetlands
and estuarine forested wetlands (in climates other than subtropical), the carbon emissions will decrease. Across all
time periods, soil C accumulation resulting from Lands Converted to Vegetated Coastal Wetlands is a carbon sink
and has ranged between -0.2 and -0.3 MMT C02 Eq. (-0.05 and -0.07 MMT C; Table 6-77 and Table 6-78).
Conversion of lands to coastal wetlands resulted in CH4 emissions of 0.2 MMT C02 Eq. (7.1 kt CH4) in 2019 (Table
6-79). Methane emissions due to the conversion of Lands to Vegetated Coastal Wetlands are largely the result of
Forest Land converting to palustrine emergent and scrub shrub coastal wetlands in warm temperate climates.
Emissions were the highest between 1990 and 2001 (0.2 MMT C02 Eq., 10.0 kt CH4) and have continually
decreased to current levels.
Table 6-77: Net CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal
Wetlands (MMT COz Eq.)
Land Use/Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Cropland Converted to Vegetated







Coastal Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Forest Land Converted to







Vegetated Coastal Wetlands
0.48
0.47
(0.03)
(0.02)
(0.01)
(0.00)
0.01
Biomass C Stock
0.62
0.62
0.13
0.13
0.13
0.13
0.13
Dead Organic Matter C Flux
0.09
0.09
0.02
0.02
0.02
0.02
0.02
Soil C Stock
(0.23)
(0.24)
(0.19)
(0.18)
(0.17)
(0.16)
(0.15)
Grassland Converted to Vegetated







Coastal Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Other Land Converted to Vegetated







Coastal Wetlands
(0.03)
(0.03)
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
Biomass C Stock
(0.01)
(0.02)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
Soil C Stock
(0.01)
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
Settlements Converted to







Vegetated Coastal Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Total Biomass Flux
0.60
0.60
0.12
0.12
0.12
0.12
0.12
Total Dead Organic Matter Flux
0.09
0.09
0.02
0.02
0.02
0.02
0.02
Total Soil C Flux
(0.25)
(0.25)
(0.20)
(0.19)
(0.18)
(0.18)
(0.17)
Total Flux
0.45
0.44
(0.06)
(0.05)
(0.04)
(0.03)
(0.02)
Note: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.005 MMT C02 Eq.
Table 6-78: Net CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal
Wetlands (MMT C)	
Land Use/Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Cropland Converted to Vegetated







Coastal Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
6-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Forest Land Converted to
Vegetated Coastal Wetlands
0.13
0.13
(0.01)
(0.01)
(+)
(+)
(+)
Biomass C Stock
0.17
0.17
0.04
0.04
0.04
0.04
0.04
Dead Organic Matter C Flux
0.03
0.02
0.01
0.01
0.01
0.01
0.01
Soil C Stock
(0.06)
(0.06)
(0.05)
(0.05)
(0.05)
(0.04)
(0.04)
Grassland Converted to Vegetated







Coastal Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Other Land Converted to







Vegetated Coastal Wetlands
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Converted to







Vegetated Coastal Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Biomass C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Soil C Stock
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Total Biomass Flux
0.16
0.16
0.03
0.03
0.03
0.03
0.03
Total Dead Organic Matter Flux
0.03
0.02
0.01
0.01
0.01
0.01
0.01
Total Soil C Flux
(0.07)
(0.07)
(0.06)
(0.05)
(0.05)
(0.05)
(0.05)
Total Flux
0.12
0.12
(0.02)
(0.01)
(0.01)
(0.01)
(0.01)
Note: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Absolute value does not exceed 0.005 MMT C.
Table 6-79: ChU Emissions from Land Converted to Vegetated Coastal Wetlands (MMT CO2
Eq. and kt CH4)
Land Use/Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Cropland Converted to Vegetated







Coastal Wetlands







Methane Emissions (MMT C02 Eq.)
+
+
+
+
+
+
+
Methane Emissions (kt CH4)
+
0.01
0.03
0.04
0.04
0.04
0.04
Forest Land Converted to Vegetated







Coastal Wetlands







Methane Emissions (MMT C02 Eq.)
0.25
0.24
0.20
0.19
0.18
0.17
0.16
Methane Emissions (kt CH4)
9.88
9.74
7.97
7.60
7.22
6.85
6.48
Grassland Converted to Vegetated







Coastal Wetlands







Methane Emissions (MMT C02 Eq.)
+
+
+
+
+
+
+
Methane Emissions (kt CH4)
0.01
0.01
0.05
0.05
0.06
0.07
0.07
Other Land Converted to Vegetated







Coastal Wetlands







Methane Emissions (MMT C02 Eq.)
+
+
0.01
0.01
0.01
0.01
0.01
Methane Emissions (kt CH4)
0.08
0.14
0.33
0.37
0.40
0.43
0.47
Settlements Converted to Vegetated







Coastal Wetlands







Methane Emissions (MMT C02 Eq.)
+
+
+
+
+
+
+
Methane Emissions (kt CH4)
0.01
+
+
+
+
+
+
Total Methane Emissions (MMT C02 Eq.)
0.25
0.25
0.21
0.20
0.19
0.18
0.18
Total Methane Emissions (kt CH4)
9.98
9.91
8.38
8.05
7.72
7.39
7.06
Note: Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.005 MMT C02 Eq. or 0.005 kt CH4.
Land Use, Land-Use Change, and Forestry 6-117

-------
Methodology
The following section provides a description of the methodology used to estimate changes in biomass, dead
organic matter and soil C stocks and CH4 emissions for Land Converted to Vegetated Coastal Wetlands.
Biomass Carbon Stock Changes
Biomass C stocks for Land Converted to Vegetated Coastal Wetlands are estimated for palustrine and estuarine
marshes for land below the elevation of high tides (taken to be mean high water spring tide elevation) and as far
seawards as the extent of intertidal vascular plants within the U.S. Land Representation according to the national
LiDAR dataset, the national network of tide gauges and land use histories recorded in the 1996, 2001, 2005, 2011,
and 2016 NOAA C-CAP surveys (NOAA OCM 2020). Both federal and non-federal lands are represented.
Delineating Vegetated Coastal Wetlands from ephemerally flooded upland Grasslands represents a particular
challenge in remote sensing. Moreover, at the boundary between wetlands and uplands, which may be gradual on
low lying coastlines, the presence of wetlands may be ephemeral depending upon weather and climate cycles and
as such impacts on the emissions and removals will vary over these time frames. Trends in land cover change are
extrapolated to 1990 and 2019 from these datasets using the C-CAP change data closest in date to a given year.
Based upon NOAA C-CAP, wetlands are subdivided into freshwater (Palustrine) and saline (Estuarine) classes and
further subdivided into emergent marsh, scrub shrub and forest classes. Biomass is not sensitive to soil organic
content. Aboveground biomass carbon stocks for non-forested coastal wetlands are derived from a national
assessment combining field plot data and aboveground biomass mapping by remote sensing (Byrd et al. 2017; Byrd
et al. 2018; Byrd et al. 2020). Aboveground biomass C removal data for all subcategories are not available and thus
assumptions were applied using expert judgment about the most appropriate assignment to a disaggregation of a
community class. The aboveground biomass carbon stock for estuarine forested wetlands (dwarf mangroves that
are not classified as forests due to their stature) is derived from a meta-analysis by Lu and Megonigal (201771).
Root to shoot ratios from the Wetlands Supplement were used to account for belowground biomass, which were
multiplied by the aboveground carbon stock (IPCC 2014), and summed with aboveground biomass to obtain total
biomass carbon stocks. Aboveground biomass C stocks for Forest Land, Cropland, and Grassland that are lost with
the conversion to Vegetated Coastal Wetlands were derived from Tier 1 default values (IPCC 2006; IPCC 2019).
Biomass carbon stock changes are calculated by subtracting the biomass C stock values of each land use category
(i.e., Forest Land, Cropland, and Grassland) from those of Vegetated Coastal Wetlands in each climate zone and
multiplying that value by the corresponding C-CAP derived area gained that year in each climate zone. The
difference between the stocks is reported as the stock change under the assumption that the change occurred in
the year of the conversion. The total coastal wetland biomass C stock change is accounted for during the year of
conversion; therefore, no interannual changes are calculated during the remaining years it is in the category.
Dead Organic Matter
Dead organic matter (DOM) carbon stocks, which include litter and dead wood stocks, are accounted for in
subtropical estuarine forested wetlands for Lands Converted to Vegetated Coastal Wetlands across all years. Tier 1
estimates of mangrove DOM C stocks were used for subtropical estuarine forested wetlands (IPCC 2014). Neither
Tier 1 or 2 data on DOM are currently available for either palustrine or estuarine scrub/shrub wetlands for any
climate zone or estuarine forested wetlands in climates other than subtropical climates. Tier 1 DOM C stocks for
Forest Land converted to Vegetated Coastal Wetlands were derived from IPCC (2019) to account for the loss of
DOM that occurs with conversion. Changes in DOM are assumed to negligible for other land use conversions (i.e.,
other than Forest Land) to coastal wetlands based on the Tier 1 method in IPCC (2006). Trends in land cover
change are derived from the NOAA C-CAP dataset and extrapolated to cover the entire 1990 through 2019 time
series. Dead organic matter removals are calculated by multiplying the C-CAP derived area gained that year by the
difference between Tier 1 DOM C stocks for Vegetated Coastal Wetlands and Forest Land. The difference between
71 See ; accessed October 2020
6-118 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
the stocks is reported as the stock change under the assumption that the change occurred in the year of the
conversion. The coastal wetland DOM stock is assumed to be in steady state once established in the year of
conversion; therefore, no interannual changes are calculated.
Soil Carbon Stock Changes
Soil C removals are estimated for Land Converted to Vegetated Coastal Wetlands across all years. Soil C stock
changes, stratified by climate zones and wetland classes, are derived from a synthesis of peer-reviewed literature
(Lynch 1989; Orson et al. 1990; Kearny & Stevenson 1991; Roman et al. 1997; Craft et al. 1998; Orson et al. 1998;
Merrill 1999; Hussein et al. 2004; Church et al. 2006; Koster et al. 2007; Callaway et al. 2012 a & b; Bianchi et al.
2013; Crooks et al. 2014; Weston et al. 2014; Villa & Mitsch 2015; Marchio et al. 2016; Noe et al. 2016). To
estimate soil C stock changes, no differentiation is made for soil type (i.e., mineral, organic). Soil C removal data for
all subcategories are not available and thus assumptions were applied using expert judgment about the most
appropriate assignment to a disaggregation of a community class.
As per IPCC Guidance, Land Converted to Vegetated Coastal Wetlands is assumed to remain in this category for up
to 20 years before transitioning to Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands. Tier 2
level estimates of soil C stock changes associated with annual soil C accumulation from Land Converted to
Vegetated Coastal Wetlands were developed using country-specific soil C removal factors multiplied by activity
data of land area for Land Converted to Vegetated Coastal Wetlands for a given year in addition to the previous 19-
year cumulative area. Guidance from the Wetlands Supplement allows for the rate of soil C accumulation to be
instantaneously equivalent to that in natural settings and that soil C accumulation is initiated when natural
vegetation becomes established; this is assumed to occur in the first year of conversion. No loss of soil carbon as a
result of land conversion to coastal wetlands is assumed to occur. Since the C-CAP coastal wetland area dataset
begins in 1996, the area converted prior to 1996 is assumed to be the same as in 1996. Similarly, the coastal
wetland area data for 2017 through 2019 is assumed to be the same as in 2016. The methodology follows Eq. 4.7,
Chapter 4 of the IPCC Wetlands Supplement, and is applied to the area of Land Converted to Vegetated Coastal
Wetlands on an annual basis.
Soil Methane Emissions
Tier 1 estimates of CH4 emissions for Land Converted to Vegetated Coastal Wetlands are derived from the same
wetland map used in the analysis of wetland soil C fluxes for palustrine wetlands, and are produced from C-CAP,
LiDAR and tidal data, in combination with default CH4 emission factors provided in Table 4.14 of the IPCC Wetlands
Supplement. The methodology follows Eq. 4.9, Chapter 4 of the IPCC Wetlands Supplement. Because Land
Converted to Vegetated Coastal Wetlands is held in this category for up to 20 years before transitioning to
Vegetated Coastal Wetlands Remaining to Vegetated Coastal Wetlands, CH4 emissions in a given year represent
the cumulative area held in this category for that year and the prior 19 years.
Uncertainty and Time-Series Consistency
Underlying uncertainties in estimates of soil C removal factors, biomass change, DOM, and CH4 emissions include
error in uncertainties associated with Tier 2 literature values of soil C removal estimates, biomass stocks, DOM,
and IPCC default CH4 emission factors, uncertainties linked to interpretation of remote sensing data, as well as
assumptions that underlie the methodological approaches applied.
Uncertainty specific to coastal wetlands include differentiation of palustrine and estuarine community classes,
which determines what flux is applied. Because mean soil and biomass C removal for each available community
class are in a fairly narrow range, the same overall uncertainty was assigned to each, respectively (i.e., applying
approach for asymmetrical errors, the largest uncertainty for any soil C stock value should be applied in the
calculation of error propagation; IPCC 2000). Uncertainties for CH4 flux are the Tier 1 default values reported in the
Wetlands Supplement. Overall uncertainty of the NOAA C-CAP remote sensing product is 15 percent. This is in the
range of remote sensing methods (±10-15 percent; IPCC 2003). However, there is significant uncertainty in salinity
ranges for tidal and non-tidal estuarine wetlands and activity data used to estimate the CH4 flux (e.g., delineation
Land Use, Land-Use Change, and Forestry 6-119

-------
of an 18 ppt boundary), which will need significant improvement to reduce uncertainties. The combined
uncertainty was calculated by summing the squared uncertainty for each individual source (C-CAP, soil, biomass,
and DOM) and taking the square root of that total.
Uncertainty estimates are presented in Table 6-80 for each carbon pool and the CH4 emissions. The combined
uncertainty is +/-42.5 percent. In 2019, the total flux was 0.16 MMT C02 Eq., with lower and upper estimates of
0.09 and 0.22 MMT C02 Eq.
Table 6-80: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes occurring
within Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent)
Source
2019 Estimate
Uncertainty Range Relative to Estimate3
(MMTCOz Eq.)
(MMT CO:
> Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Biomass C Stock Flux
0.1
0.1
0.15
-20.0%
20.0%
Dead Organic Matter Flux
0.02
0.02
0.03
-25.8%
25.8%
Soil C Stock Flux
(0.2)
(0.2)
(0.1)
-17.8%
17.8%
Methane Emissions
0.2
0.12
0.23
-29.9%
29.9%
Total Uncertainty
0.16
0.09
0.22
-42.2%
42.2%
Note: Totals may not sum due to independent rounding.
a Range of flux estimates based on error propagation at 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the introduction and Methodology sections.
QA/QC and Verification
NOAA provided National LiDAR Dataset, tide data, and C-CAP land cover and land cover change mapping, all of
which are subject to agency internal mandatory QA/QC assessment (McCombs et al. 2016). QA/QC and verification
of soil C stock dataset has been provided by the Smithsonian Environmental Research Center and Coastal Wetland
Inventory team leads. Biomass C stocks are derived from peer-review literature, reviewed by U.S. Geological
Survey prior to publishing, by the peer-review process during publishing, and by the Coastal Wetland Inventory
team leads prior to inclusion in the inventory and from IPCC reports. As a QC step, a check was undertaken
confirming that Coastal Wetlands recognized by C-CAP represent a subset of Wetlands recognized by the NRI for
marine coastal states. Land cover estimates were assessed to ensure that the total land area did not change over
the time series in which the inventory was developed, and verified by a second QA team. A team of two evaluated
and verified there were no computational errors within the calculation worksheets. Soil C stock,
emissions/removals data are based upon peer-reviewed literature and CH4 emission factors are derived from the
Wetlands Supplement.
Recalculations Discussion
As part of the addition of a 2016 C-CAP dataset, the previous datasets (1996, 2001, 2006, and 2011) were refined
and reanalyzed using improved methods, software and techniques that were deemed important to incorporate in
order to create the most accurate and representative product.
The addition of the 2016 C-CAP dataset resulted in changes in area calculations for 1990 to 1995 and 2011 through
2018, all years that previously used a change value that was the average change across all C-CAP time periods
(1996 through 2011). For 1990 through 1995, the 1996 to 2000 C-CAP change data were used, and the 2011 to
2016 C-CAP change data were used for 2017 through 2019. This improvement results in fewer years where coastal
wetland area is interpolated.
6-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Although this change does not affect the area calculations presented here, a new organic soil geospatial layer was
derived from CONUS distributions of histosols in the USDA's Soil Survey Geographic Database (SSURGO). The
dataset is complete except for a large portion of the Florida Everglades that did not contain data. Incorporating
this new dataset resulted in a large change in extents between organic and mineral soils, with far less area of
organic soils nationally than previously recorded.
A corrigendum was published by Byrd et al. (2020) for non-forested aboveground biomass values initially
presented in Byrd et al. (2018). The updated analyses resulted in increases in aboveground biomass carbon stocks
across all wetland types and climate zones (except for the Puget Sound, which had a 0.11 C ha 1 decrease); the
average aboveground biomass increased from 1.93 to 3.021 C ha"1.
Belowground biomass carbon stocks are now included for coastal wetlands. They are based on Tier 2 root to shoot
ratios from the Wetlands Supplement (IPCC 2014). This resulted in a 0.05 and 0.0001 MMT C02 Eq. (0.01 and
0.00003 MMT C) increase in 1990 and 2018, respectively, compared to the previous Inventory.72
Dead organic matter carbon fluxes are now included for estuarine forested wetlands and for Forest Land
converted to Vegetated Coastal Wetlands. This update resulted in an increase in emissions of 0.09 and 0.02 MMT
C02 Eq. (0.02 and 0.005 MMT C) in 1990 and 2018, respectively.
An update was made to how biomass C fluxes are calculated when lands are converted to Vegetated Coastal
Wetlands. The loss of the converted land's former carbon stocks is now included, which applies to Forest Land,
Grassland, and Cropland, and results in net emissions from biomass from coastal wetlands converted from Forest
Land. The total loss and subsequent coastal wetland biomass gain is assumed to occur in the first year of
conversion, although the loss and gain of C actually occurs over a longer period of time. This update, in addition to
inclusion of belowground biomass and the corrigendum to aboveground biomass C stocks, resulted in an increase
in emissions of 0.63 and 0.15 MMT C02 Eq. (0.17 and 0.04 MMT C) in 1990 and 2018, respectively.
Soil C accumulation and CH4 emissions are now calculated as the cumulative flux from the area converted each
year in addition to the prior 19 years. This results in greater soil C removals and CH4 emissions across the entire
reporting period. Soil C removals increased by 0.24 and 0.15 MMT C02 Eq. (0.06 and 0.04 MMT C) in 1990 and
2018, respectively. Methane emissions increased by 0.24 and 0.16 MMT C02 Eq. (9.4 and 6.5 kt CH4) in 1990 and
2018, respectively.
Planned Improvements
Administered by the Smithsonian Environmental Research Center, the Coastal Wetland Carbon Research
Coordination Network has established a U.S. country-specific database of soil C stocks and biomass for coastal
wetlands.73 This dataset will be updated periodically. Refined error analysis combining land cover change and C
stock estimates will be provided as new data are incorporated. Through this work, a model is in development to
represent changes in soil C stocks and will be incorporated into the 2022 Inventory submission.
Currently, the only coastal wetland conversion that is reported in the Inventory is Lands Converted to Vegetated
Coastal Wetlands. The next (2022) submission will include C stock change data for Lands Converted to Unvegetated
Open Water Coastal Wetlands.
72	These values only include changes in the coastal wetland biomass stock emissions factors and do not include the updates of
accounting for other land uses' biomass stock loss.
73	See ; accessed October 2020.
Land Use, Land-Use Change, and Forestry 6-121

-------
6.10 Settlements Remaining Settlements
(CRF Category 4E1)
Soil Carbon Stock Changes (CRF Category 4E1)
Soil organic C stock changes for Settlements Remaining Settlements occur in both mineral and organic soils.
However, the United States does not estimate changes in soil organic C stocks for mineral soils in Settlements
Remaining Settlements. This approach is consistent with the assumption of the Tier 1 method in the 2006IPCC
Guidelines (IPCC 2006) that inputs equal outputs, and therefore the soil organic C stocks do not change. This
assumption may be re-evaluated in the future if funding and resources are available to conduct an analysis of soil
organic C stock changes for mineral soils in Settlements Remaining Settlements.
Drainage of organic soils is common when wetland areas have been developed for settlements. Organic soils, also
referred to as Histosols, include all soils with more than 12 to 20 percent organic C by weight, depending on clay
content (NRCS 1999; Brady and Weil 1999). The organic layer of these soils can be very deep (i.e., several meters),
and form under inundated conditions that results in minimal decomposition of plant residues. Drainage of organic
soils leads to aeration of the soil that accelerates decomposition rate and C02 emissions.74 Due to the depth and
richness of the organic layers, C loss from drained organic soils can continue over long periods of time, which
varies depending on climate and composition (i.e., decomposability) of the organic matter (Armentano and
Menges 1986).
Settlements Remaining Settlements includes all areas that have been settlements for a continuous time period of
at least 20 years according to the 2015 United States Department of Agriculture (USDA) National Resources
Inventory (NRI) (USDA-NRCS 2018)75 or according to the National Land Cover Dataset (NLCD) for federal lands
(Yang et al. 2018; Fry et al. 2011; Homer et al. 2007, 2015). The Inventory includes settlements on privately-owned
lands in the conterminous United States and Hawaii. Alaska and the small amount of settlements on federal lands
are not included in this Inventory even though these areas are part of the U.S. managed land base. This leads to a
discrepancy with the total amount of managed area in Settlements Remaining Settlements (see Section 6
Representation of the U.S. Land Base) and the settlements area included in the Inventory analysis. There is a
planned improvement to include C02 emissions from drainage of organic soils in settlements of Alaska and federal
lands as part of a future Inventory.
C02 emissions from drained organic soils in settlements are 15.9 MMT C02 Eq. (4.3 MMT C) in 2019 (See Table 6-81
and 6-82). Although the flux is relatively small, the amount has increased by over 41 percent since 1990 due to an
increase in area of drained organic soils in settlements.
Table 6-81: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements
(MMT COz Eq.)
Soil Type	1990	2005	2015 2016 2017 2018 2019
Organic Soils	11.3	12.2	15.7 16.0 16.0 15.9 15.9
74	N20 emissions from soils are included in the N20 Emissions from Settlement Soils section.
75	NRI survey locations are classified according to land-use histories starting in 1979, and consequently the classifications are
based on less than 20 years from 1990 to 1998. This may have led to an overestimation of Settlements Remaining Settlements
in the early part of the time series to the extent that some areas are converted to settlements between 1971 and 1978.
6-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 6-82: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements
(MMT C)
Soil Type
1990
2005
2015
2016
2017
2018
2019
Organic Soils
3.1
3.3
4.3
4.4
4.4
4.3
4.3
Methodology
An IPCC Tier 2 method is used to estimate soil organic C stock changes for organic soils in Settlements Remaining
Settlements (IPCC 2006). Organic soils in Settlements Remaining Settlements are assumed to be losing C at a rate
similar to croplands due to deep drainage, and therefore emission rates are based on country-specific values for
cropland (Ogle et al. 2003).
The land area designated as settlements is based primarily on the 2018 NRI (USDA-NRCS 2018) with additional
information from the NLCD (Yang et al. 2018; Fry et al. 2011; Homer et al. 2007, 2015). It is assumed that all
settlement area on organic soils is drained, and those areas are provided in Table 6-83 (See Section 6,
Representation of the U.S. Land Base for more information). The area of drained organic soils is estimated from
the NRI spatial weights and aggregated to the country (Table 6-83). The area of land on organic soils in Settlements
Remaining Settlements has increased from 220 thousand hectares in 1990 to over 303 thousand hectares in 2015.
The area of land on organic soils are not currently available from NRI for Settlements Remaining Settlements after
2015.
Table 6-83: Thousands of Hectares of Drained Organic Soils in Settlements Remaining
Settlements
Year
Area
(Thousand Hectares)
1990
220
2005
235
2014
291
2015
303
2016
ND
2017
ND
2018
ND
2019
ND
Note: No NRI data are available after 2015,
designated as ND (No data)
To estimate C02 emissions from drained organic soils across the time series from 1990 to 2015, the total area of
organic soils in Settlements Remaining Settlements is multiplied by the country-specific emission factors for
Cropland Remaining Cropland under the assumption that there is deep drainage of the soils. The emission factors
are 11.2 MT C per ha in cool temperate regions, 14.0 MT C per ha in warm temperate regions, and 14.3 MT C per
ha in subtropical regions (see Annex 3.12 for more information).
A linear extrapolation of the trend in the time series is applied to estimate the emissions from 2016 to 2019
because NRI activity data are not available for these years to determine the area of drained organic soils in
Settlements Remaining Settlements. Specifically, a linear regression model with autoregressive moving-average
(ARMA) errors (Brockwell and Davis 2016) is used to estimate the trend in emissions over time from 1990 to 2015,
and in turn, the trend is used to approximate the 2016 to 2019 emissions. The Tier 2 method described previously
will be applied in future inventories to recalculate the estimates beyond 2015 as activity data become available.
Land Use, Land-Use Change, and Forestry 6-123

-------
Uncertainty and Time-Series Consistency
Uncertainty for the Tier 2 approach is derived using a Monte Carlo approach, along with additional uncertainty
propagated through the Monte Carlo Analysis for 2016 to 2019 based on the linear time series model. The results
of the Approach 2 Monte Carlo uncertainty analysis are summarized in Table 6-84. Soil C losses from drained
organic soils in Settlements Remaining Settlements for 2019 are estimated to be between 7.5 and 24.3 MMT C02
Eq. at a 95 percent confidence level. This indicates a range of 53 percent below and 53 percent above the 2019
emission estimate of 15.9 MMT C02 Eq.
Table 6-84: Uncertainty Estimates for CO2 Emissions from Drained Organic Soils in
Settlements Remaining Settlements (MMT CO2 Eq. and Percent)


2019 Emission


Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMTCOz Eq.)
(MMT CO? Eq.)
(%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Organic Soils
C02
15.9
7.5 24.3
-53% 53%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in the
Introduction and Methodology sections.
QA/QC and Verification
Quality control measures included checking input data, model scripts, and results to ensure data are properly
handled throughout the inventory process. Inventory reporting forms and text are reviewed and revised as needed
to correct transcription errors. No errors were found in this Inventory.
Recalculations Discussion
There were no recalculations to the 1990 through 2018 time series in this Inventory.
Planned Improvements
This source will be updated to include C02 emissions from drainage of organic soils in settlements of Alaska and
federal lands in order to provide a complete inventory of emissions for this category. See Table 6-85 for the
amount of managed land area in Settlements Remaining Settlements that is not included in the Inventory due to
these omissions. The managed settlements area that is not included in the Inventory is in the range of 150 to 160
thousand hectares each year. These improvements will be made as funding and resources are available to expand
the inventory for this source category.
Table 6-85: Area of Managed Land in Settlements Remaining Settlements that is not
included in the current Inventory (Thousand Hectares)
Area (Thousand Hectares)
SRS Managed Land SRS Area Included SRS Area Not Included
Year	Area (Section 6)	in Inventory	in Inventory
1990	30,585	30,425	159
1991	30,589	30,430	159
1992	30,593	30,434	159
1993	30,505	30,346	159
1994	30,423	30,264	159
6-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
1995
30,365
30,206
159
1996
30,316
30,157
158
1997
30,264
30,105
158
1998
30,200
30,041
159
1999
30,144
29,992
152
2000
30,101
29,949
152
2001
30,041
29,889
152
2002
30,034
29,882
152
2003
30,530
30,378
152
2004
31,011
30,859
152
2005
31,522
31,370
152
2006
31,964
31,812
152
2007
32,469
32,317
152
2008
33,074
32,922
152
2009
33,646
33,494
152
2010
34,221
34,069
152
2011
34,814
34,662
152
2012
35,367
35,215
152
2013
36,308
36,156
152
2014
37,281
37,129
152
2015
38,210
38,058
152
2016
ND
ND
ND
2017
ND
ND
ND
2018
ND
ND
ND
2019
ND
ND
ND
Note: NRI data are not available after 2015, and these years are designated as ND (No data).
Changes In Carbon Stocks in Settlement Trees (CRF Source
Category 4E1)
Settlements are land uses where human populations and activities are concentrated. In these areas, the
anthropogenic impacts on tree growth, stocking and mortality are particularly pronounced (Nowak 2012) in
comparison to forest lands where non-anthropogenic forces can have more significant impacts. Trees in
settlement areas of the United States are estimated to account for an average annual net sequestration of 115.9
MMT C02 Eq. (31.6 MMT C) over the period from 1990 through 2019. Net C sequestration from settlement trees in
2019 is estimated to be 129.8 MMT C02 Eq. (35.4 MMT C) (Table 6-86). Dominant factors affecting carbon flux
trends for settlement trees are changes in the amount of settlement area (increasing sequestration due to more
land and trees) and net changes in tree cover (e.g., tree losses vs tree gains through planting and natural
regeneration), with percent tree cover trending downward recently. In addition, changes in species composition,
tree sizes and tree densities affect base C flux estimates. Annual sequestration increased by 35 percent between
1990 and 2019 due to increases in settlement area and changes in tree cover.
Trees in settlements often grow faster than forest trees because of their relatively open structure (Nowak and
Crane 2002). Because tree density in settlements is typically much lower than in forested areas, the C storage per
hectare of land is in fact smaller for settlement areas than for forest areas. Also, percent tree cover in settlement
areas are less than in forests and this tree cover varies significantly across the United States (e.g., Nowak and
Greenfield 2018a). To quantify the C stored in settlement trees, the methodology used here requires analysis per
unit area of tree cover, rather than per unit of total land area (as is done for Forest Lands).
Table 6-86: Net Flux from Trees in Settlements Remaining Settlements (MMT CO2 Eq. and
MMT C)a
Year	MMTCQ2 Eq.	MMT C
1990	(96.4)	(26.3)
Land Use, Land-Use Change, and Forestry 6-125

-------
2005
(117.4)
(32.0)
2014
2015
2016
2017
2018
2019
(129.4)
(130.4)
(129.8)
(129.8)
(129.8)
(129.8)
(35.3)
(35.6)
(35.4)
(35.4)
(35.4)
(35.4)
Note: Parentheses indicate net sequestration.
a These estimates include net C02 and C flux from
trees on Settlements Remaining Settlements and
Land Converted to Settlements as it is not possible
to report on these separately at this time.
To estimate net carbon sequestration in settlement areas, three types of data are required for each state:
1.	Settlement area
2.	Percent tree cover in settlement areas
3.	Carbon sequestration density per unit of tree cover
Settlement Area
Settlements area is defined in Section 6 Representation of the U.S. Land Base as a land-use category representing
developed areas. The data used to estimate settlement area within Section 6 comes from the NRI as updated
through 2015 with the extension of the time series through 2018 based on assuming the settlements area is the
same as 2015, while harmonizing these data with the FIA dataset, which are available through 2018, and the NLCD
dataset, which is available through 2016. Settlement areas for 2019 are held constant with the 2018 values. This
process of combining the datasets extends the time series to ensure that there is a complete and consistent
representation of land use data for all source categories in the LULUCF sector. Annual estimates of C02 flux (Table
6-86) were developed based on estimates of annual settlement area and tree cover derived from NLCD developed
lands. Developed land, which was used to estimate tree cover in settlement areas, is about six percent higher than
the area categorized as Settlements in the Representation of the U.S. Land Base developed for this report.
Percent Tree Cover in Settlement Areas
Percent tree cover in settlement area by state is needed to convert settlement land area to settlement tree cover
area. Converting to tree cover area is essential as tree cover, and thus carbon estimates, can vary widely among
states in settlement areas due to variations in the amount of tree cover (e.g., Nowak and Greenfield 2018a).
However, since the specific geography of settlement area is unknown because they are based on NRI sampling
methods, NLCD developed land was used to estimate the percent tree cover to be used in settlement areas. NLCD
developed classes 21-24 (developed, open space (21), low intensity (22), medium intensity (23), and high intensity
(24)) were used to estimate percent tree cover in settlement area by state (U.S. Department of Interior 2018;
MRLC 2013).
a)	"Developed, Open Space - areas with a mixture of some constructed materials, but mostly vegetation in
the form of lawn grasses. Impervious surfaces account for less than 20 percent of total cover. These areas
most commonly include large-lot single-family housing units, parks, golf courses, and vegetation planted
in developed settings for recreation, erosion control, or aesthetic purposes." Plots designated as either
park, recreation, cemetery, open space, institutional or vacant land were classified as Developed Open
Space.
b)	"Developed, Low Intensity - areas with a mixture of constructed materials and vegetation. Impervious
surfaces account for 20 to 49 percent of total cover. These areas most commonly include single-family
Methodology
6-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
housing units." Plots designated as single family or low-density residential land were classified as
Developed, Low Intensity.
c)	"Developed, Medium Intensity - areas with a mixture of constructed materials and vegetation.
Impervious surfaces account for 50 to 79 percent of the total cover. These areas most commonly include
single-family housing units." Plots designated as medium density residential, other urban or mixed urban
were classified as Developed, Medium Intensity.
d)	"Developed High Intensity - highly developed areas where people reside or work in high numbers.
Examples include apartment complexes, row houses and commercial/industrial. Impervious surfaces
account for 80 to 100 percent of the total cover." Plots designated as either commercial, industrial, high
density residential, downtown, multi-family residential, shopping, transportation or utility were classified
as Developed, High Intensity.
As NLCD is known to underestimate tree cover (Nowak and Greenfield 2010), photo-interpretation of tree cover
within NLCD developed lands was conducted for the years of c. 2011 and 2016 using 1,000 random points to
determine an average adjustment factor for NLCD tree cover estimates in developed land and determine recent
tree cover changes. This photo-interpretation of change followed methods detailed in Nowak and Greenfield
(2018b). Percent tree cover (%TC) in settlement areas by state was estimated as:
%TC in state = state NLCD %TC x national photo-interpreted %TC / national NLCD %TC
Percent tree cover in settlement areas by year was set as follows:
•	1990 to 2011: used 2011 NLCD tree cover adjusted with 2011 photo-interpreted values
•	2012 to 2015: used 2011 NLCD tree cover adjusted with photo-interpreted values, which were
interpolated from values between 2011 and 2016
•	2016 to 2019: used 2011 NLCD tree cover adjusted with 2016 photo-interpreted values
Carbon Sequestration Density per Unit of Tree Cover
Methods for quantifying settlement tree biomass, C sequestration, and C emissions from tree mortality and
decomposition were taken directly from Nowak et al. (2013), Nowak and Crane (2002), and Nowak (1994). In
general, net C sequestration estimates followed three steps, each of which is explained further in the paragraphs
below. First, field data from cities and urban areas within entire states were used to estimate C in tree biomass
from field data on measured tree dimensions. Second, estimates of annual tree growth and biomass increment
were generated from published literature and adjusted for tree condition, crown competition, and growing season
to generate estimates of gross C sequestration in settlement trees for all 50 states and the District of Columbia.
Third, estimates of C emissions due to mortality and decomposition were subtracted from gross C sequestration
estimates to obtain estimates of net C sequestration. Carbon storage, gross and net sequestration estimates were
standardized per unit tree cover based on tree cover in the study area.
Settlement tree carbon estimates are based on published literature (Nowak et al. 2013; Nowak and Crane 2002;
Nowak 1994) as well as newer data from the i-Tree database76 and Forest Service urban forest inventory data
(e.g., Nowak et al. 2016, 2017) (Table 6-87). These data are based on collected field measurements in several U.S.
cities between 1989 and 2017. Carbon storage and sequestration in these cities were estimated using the U.S.
Forest Service's i-Tree Eco model (Nowak et al. 2008). This computer model uses standardized field data from
randomly located plots, along with local hourly air pollution and meteorological data to quantify urban forest
structure, monetary values of the urban forest, and environmental effects, including total C stored and annual C
sequestration (Nowak et al. 2013).
In each city, a random sample of plots were measured to assess tree stem diameter, tree height, crown height and
crown width, tree location, species, and canopy condition. The data for each tree were used to estimate total dry-
76 See .
Land Use, Land-Use Change, and Forestry 6-127

-------
weight biomass using allometric models, a root-to-shoot ratio to convert aboveground biomass estimates to whole
tree biomass, and wood moisture content. Total dry weight biomass was converted to C by dividing by two (50
percent carbon content). An adjustment factor of 0.8 was used for open grown trees to account for settlement
trees having less aboveground biomass for a given stem diameter than predicted by allometric models based on
forest trees (Nowak 1994). Carbon storage estimates for deciduous trees include only C stored in wood. Estimated
C storage was divided by tree cover in the area to estimate carbon storage per square meter of tree cover.
Table 6-87: Carbon Storage (kg C/m2 tree cover), Gross and Net Sequestration (kg C/m2
tree cover/year) and Tree Cover (percent) among Sampled U.S. Cities (see Nowak et al.
2013)
Sequestration
City
Storage
SE
Gross
SE
Net
SE
Ratio3
Tree
Cover
SE
Adrian, Ml
12.17
1.88
0.34
0.04
0.13
0.07
0.36
22.1
2.3
Albuquerque, NM
5.61
0.97
0.24
0.03
0.20
0.03
0.82
13.3
1.5
Arlington, TX
6.37
0.73
0.29
0.03
0.26
0.03
0.91
22.5
0.3
Atlanta, GA
6.63
0.54
0.23
0.02
0.18
0.03
0.76
53.9
1.6
Austin, TX
3.57
0.25
0.17
0.01
0.13
0.01
0.73
30.8
1.1
Baltimore, MD
10.30
1.24
0.33
0.04
0.20
0.04
0.59
28.5
1.0
Boise, ID
7.33
2.16
0.26
0.04
0.16
0.06
0.64
7.8
0.2
Boston, MA
7.02
0.96
0.23
0.03
0.17
0.02
0.73
28.9
1.5
Camden, NJ
11.04
6.78
0.32
0.20
0.03
0.10
0.11
16.3
9.9
Casper, WY
6.97
1.50
0.22
0.04
0.12
0.04
0.54
8.9
1.0
Chester, PA
8.83
1.20
0.39
0.04
0.25
0.05
0.64
20.5
1.7
Chicago (region), IL
9.38
0.59
0.38
0.02
0.26
0.02
0.70
15.5
0.3
Chicago, IL
6.03
0.64
0.21
0.02
0.15
0.02
0.70
18.0
1.2
Corvallis, OR
10.68
1.80
0.22
0.03
0.20
0.03
0.91
32.6
4.1
El Paso, TX
3.93
0.86
0.32
0.05
0.23
0.05
0.72
5.9
1.0
Freehold, NJ
11.50
1.78
0.31
0.05
0.20
0.05
0.64
31.2
3.3
Gainesville, FL
6.33
0.99
0.22
0.03
0.16
0.03
0.73
50.6
3.1
Golden, CO
5.88
1.33
0.23
0.05
0.18
0.04
0.79
11.4
1.5
Grand Rapids, Ml
9.36
1.36
0.30
0.04
0.20
0.05
0.65
23.8
2.0
Hartford, CT
10.89
1.62
0.33
0.05
0.19
0.05
0.57
26.2
2.0
Houston, TX
4.55
0.48
0.31
0.03
0.25
0.03
0.83
18.4
1.0
Indiana15
8.80
2.68
0.29
0.08
0.27
0.07
0.92
20.1
3.2
Jersey City, NJ
4.37
0.88
0.18
0.03
0.13
0.04
0.72
11.5
1.7
Kansas'5
7.42
1.30
0.28
0.05
0.22
0.04
0.78
14.0
1.6
Kansas City (region),









MO/KS
7.79
0.85
0.39
0.04
0.26
0.04
0.67
20.2
1.7
Lake Forest Park, WA
12.76
2.63
0.49
0.07
0.42
0.07
0.87
42.4
0.8
Las Cruces, NM
3.01
0.95
0.31
0.14
0.26
0.14
0.86
2.9
1.0
Lincoln, NE
10.64
1.74
0.41
0.06
0.35
0.06
0.86
14.4
1.6
Los Angeles, CA
4.59
0.51
0.18
0.02
0.11
0.02
0.61
20.6
1.3
Milwaukee, Wl
7.26
1.18
0.26
0.03
0.18
0.03
0.68
21.6
1.6
Minneapolis, MN
4.41
0.74
0.16
0.02
0.08
0.05
0.52
34.1
1.6
Moorestown, NJ
9.95
0.93
0.32
0.03
0.24
0.03
0.75
28.0
1.6
Morgantown, WV
9.52
1.16
0.30
0.04
0.23
0.03
0.78
39.6
2.2
Nebraska15
6.67
1.86
0.27
0.07
0.23
0.06
0.84
15.0
3.6
New York, NY
6.32
0.75
0.33
0.03
0.25
0.03
0.76
20.9
1.3
North Dakota15
7.78
2.47
0.28
0.08
0.13
0.08
0.48
2.7
0.6
Oakland, CA
5.24
0.19
NA
NA
NA
NA
NA
21.0
0.2
Oconomowoc, Wl
10.34
4.53
0.25
0.10
0.16
0.06
0.65
25.0
7.9
Omaha, NE
14.14
2.29
0.51
0.08
0.40
0.07
0.78
14.8
1.6
Philadelphia, PA
8.65
1.46
0.33
0.05
0.29
0.05
0.86
20.8
1.8
Phoenix, AZ
3.42
0.50
0.38
0.04
0.35
0.04
0.94
9.9
1.2
Roanoke, VA
9.20
1.33
0.40
0.06
0.27
0.05
0.67
31.7
3.3
Sacramento, CA
7.82
1.57
0.38
0.06
0.33
0.06
0.87
13.2
1.7
6-128 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
San Francisco, CA
9.18
2.25
0.24
0.05
0.22
0.05
0.92
16.0
2.6
Scranton, PA
9.24
1.28
0.40
0.05
0.30
0.04
0.74
22.0
1.9
Seattle, WA
9.59
0.98
0.67
0.06
0.55
0.05
0.82
27.1
0.4
South Dakotab
3.14
0.66
0.13
0.03
0.11
0.02
0.87
16.5
2.2
Syracuse, NY
9.48
1.08
0.30
0.03
0.22
0.04
0.72
26.9
1.3
Tennessee15
6.47
0.50
0.34
0.02
0.30
0.02
0.89
37.7
0.8
Washington, DC
8.52
1.04
0.26
0.03
0.21
0.03
0.79
35.0
2.0
Woodbridge, NJ
8.19
0.82
0.29
0.03
0.21
0.03
0.73
29.5
1.7
SE-Standard Error
NA - Not Available
a Ratio of net to gross sequestration
b Statewide assessment of urban areas
To determine gross sequestration rates, tree growth rates need to be estimated. Base growth rates were
standardized for open-grown trees in areas with 153 days of frost-free length based on measured data on tree
growth (Nowak et al. 2013). These growth rates were adjusted to local tree conditions based on length of frost-
free season, crown competition (as crown competition increased, growth rates decreased), and tree condition (as
tree condition decreased, growth rates decreased). Annual growth rates were applied to each sampled tree to
estimate gross annual sequestration - that is, the difference in C storage estimates between year 1 and year (x + 1)
represents the gross amount of C sequestered. These annual gross C sequestration rates for each tree were then
scaled up to city estimates using tree population information. Total C sequestration was divided by total tree cover
to estimate a gross carbon sequestration density (kg C/m2 of tree cover/year). The area of assessment for each city
or state was defined by its political boundaries; parks and other forested urban areas were thus included in
sequestration estimates.
Where gross C sequestration accounts for all C sequestered, net C sequestration for settlement trees considers C
emissions associated with tree death and removals. The third step in the methodology estimates net C emissions
from settlement trees based on estimates of annual mortality, tree condition, and assumptions about whether
dead trees were removed from the site. Estimates of annual mortality rates by diameter class and condition class
were obtained from a study of street-tree mortality (Nowak 1986). Different decomposition rates were applied to
dead trees left standing compared with those removed from the site. For removed trees, different rates were
applied to the removed/aboveground biomass in contrast to the belowground biomass (Nowak et al. 2002). The
estimated annual gross C emission rates for each plot were then scaled up to city estimates using tree population
information.
The full methodology development is described in the underlying literature, and key details and assumptions were
made as follows. The allometric models applied to the field data for the Nowak methodology for each tree were
taken from the scientific literature (see Nowak 1994, Nowak et al. 2002), but if no allometric model could be found
for the particular species, the average result for the genus or botanical relative was used. The adjustment (0.8) to
account for less live tree biomass in open-grown urban trees was based on information in Nowak (1994).
Measured tree growth rates for street (Frelich 1992; Fleming 1988; Nowak 1994), park (deVries 1987), and forest
(Smith and Shifley 1984) trees were standardized to an average length of growing season (153 frost free days) and
adjusted for site competition and tree condition. Standardized growth rates of trees of the same species or genus
were then compared to determine the average difference between standardized street tree growth and
standardized park and forest growth rates. Crown light exposure (CLE) measurements (number of sides and/or top
of tree exposed to sunlight) were used to represent forest, park, and open (street) tree growth conditions. Local
tree base growth rates were then calculated as the average standardized growth rate for open-grown trees
multiplied by the number of frost-free days divided by 153. Growth rates were then adjusted for CLE. The CLE
adjusted growth rate was then adjusted based on tree condition to determine the final growth rate. Assumptions
for which dead trees would be removed versus left standing were developed specific to each land use and were
based on expert judgment of the authors. Decomposition rates were based on literature estimates (Nowak et al.
2013).
Estimates of gross and net sequestration rates for each of the 50 states and the District of Columbia (Table 6-88)
were compiled in units of C sequestration per unit area of tree canopy cover. These rates were used in conjunction
with estimates of state settlement area and developed land percent tree cover data to calculate each state's
Land Use, Land-Use Change, and Forestry 6-129

-------
annual net C sequestration by urban trees. This method was described in Nowak et al. (2013) and has been
modified here to incorporate developed land percent tree cover data.
Net annual C sequestration estimates were obtained for all 50 states and the District of Columbia by multiplying
the gross annual emission estimates by 0.73, the average ratio for net/gross sequestration (Table 6-88). However,
state specific ratios were used where available.
State Carbon Sequestration Estimates
The gross and net annual C sequestration values for each state were multiplied by each state's settlement area of
tree cover, which was the product of the state's settlement area and the state's tree cover percentage based on
NLCD developed land. The model used to calculate the total carbon sequestration amounts for each state, can be
written as follows:
Net state annual C sequestration (t C/yr) = Gross state sequestration rate (t C/ha/yr) x Net to Gross state
sequestration ratio x state settlement Area (ha) x % state tree cover in settlement area
The results for all 50 states and the District of Columbia are given in Table 6-88. This approach is consistent with
the default IPCC Gain-Loss methodology in IPCC (2006), although sufficient field data are not yet available to
separately determine interannual gains and losses in C stocks in the living biomass of settlement trees. Instead, the
methodology applied here uses estimates of net C sequestration based on modeled estimates of decomposition,
as given by Nowak et al. (2013).
Table 6-88: Estimated Annual C Sequestration (Metric Tons C/Year), Tree Cover (Percent),
and Annual C Sequestration per Area of Tree Cover (kg C/m2/ year) for settlement areas in
United States by State and the District of Columbia (2019)




Gross Annual
Net Annual
Net: Gross




Sequestration
Sequestration
Annual

Gross Annual
Net Annual
Tree
per Area of
per Area of
Sequestration
State
Sequestration
Sequestration
Cover
Tree Cover
Tree Cover
Ratio
Alabama
2,060,001
1,501,070
53.5
0.376
0.274
0.73
Alaska
111,722
81,409
47.4
0.169
0.123
0.73
Arizona
172,750
125,878
4.6
0.388
0.283
0.73
Arkansas
1,266,164
922,622
48.9
0.362
0.264
0.73
California
2,007,869
1,463,083
16.9
0.426
0.311
0.73
Colorado
142,719
103,996
8.0
0.216
0.157
0.73
Connecticut
618,683
450,818
58.7
0.262
0.191
0.73
Delaware
97,533
71,070
24.4
0.366
0.267
0.73
DC
11,995
8,741
25.1
0.366
0.267
0.73
Florida
4,322,610
3,149,776
40.3
0.520
0.379
0.73
Georgia
3,411,478
2,485,857
56.3
0.387
0.282
0.73
Hawaii
285,700
208,182
41.7
0.637
0.464
0.73
Idaho
59,611
43,437
7.4
0.201
0.146
0.73
Illinois
662,891
483,032
15.5
0.310
0.226
0.73
Indiana
472,905
437,275
17.1
0.274
0.254
0.92
Iowa
177,692
129,480
8.6
0.263
0.191
0.73
Kansas
290,461
226,027
10.8
0.310
0.241
0.78
Kentucky
926,269
674,949
36.8
0.313
0.228
0.73
Louisiana
1,512,145
1,101,861
47.0
0.435
0.317
0.73
Maine
394,471
287,441
55.5
0.242
0.176
0.73
Maryland
818,044
596,088
40.1
0.353
0.257
0.73
Massachusetts
1,002,723
730,659
57.2
0.278
0.203
0.73
Michigan
1,343,325
978,847
34.7
0.241
0.175
0.73
Minnesota
313,364
228,340
13.1
0.251
0.183
0.73
Mississippi
1,518,448
1,106,454
57.3
0.377
0.275
0.73
Missouri
850,492
619,732
23.2
0.313
0.228
0.73
Montana
48,911
35,640
4.9
0.201
0.147
0.73
6-130 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Nebraska
98,584
83,192
7.3
0.261
0.220
0.84
Nevada
41,181
30,008
4.8
0.226
0.165
0.73
New Hampshire
363,989
265,229
59.3
0.238
0.174
0.73
New Jersey
904,868
659,355
40.7
0.321
0.234
0.73
New Mexico
177,561
129,384
10.2
0.288
0.210
0.73
New York
1,531,415
1,115,903
39.9
0.263
0.192
0.73
North Carolina
3,064,797
2,233,239
54.1
0.341
0.249
0.73
North Dakota
18,492
8,787
1.8
0.244
0.116
0.48
Ohio
1,248,841
909,999
28.2
0.271
0.198
0.73
Oklahoma
699,044
509,376
22.1
0.364
0.265
0.73
Oregon
682,468
497,297
39.9
0.265
0.193
0.73
Pennsylvania
1,794,939
1,307,927
40.2
0.267
0.195
0.73
Rhode Island
121,940
88,855
50.0
0.283
0.206
0.73
South Carolina
1,801,029
1,312,364
53.8
0.370
0.269
0.73
South Dakota
29,489
25,573
2.9
0.258
0.224
0.87
Tennessee
1,591,278
1,422,789
41.1
0.332
0.297
0.89
Texas
4,239,494
3,089,211
28.5
0.403
0.294
0.73
Utah
118,880
86,625
11.7
0.235
0.172
0.73
Vermont
176,564
128,658
50.6
0.234
0.170
0.73
Virginia
1,968,537
1,434,422
52.9
0.321
0.234
0.73
Washington
1,063,871
775,216
37.6
0.282
0.206
0.73
West Virginia
699,320
509,577
64.1
0.264
0.192
0.73
Wisconsin
697,863
508,515
25.9
0.246
0.180
0.73
Wyoming
29,984
21,849
4.7
0.199
0.145
0.73
Total
48,065,406
35,405,113




Uncertainty and Time-Series Consistency
Uncertainty associated with changes in C stocks in settlement trees includes the uncertainty associated with
settlement area, percent tree cover in developed land and how well it represents percent tree cover in settlement
areas, and estimates of gross and net C sequestration for each of the 50 states and the District of Columbia. A 10
percent uncertainty was associated with settlement area estimates based on expert judgment. Uncertainty
associated with estimates of percent settlement tree coverage for each of the 50 states was based on standard
error associated with the photo-interpretation of national tree cover in developed lands. Uncertainty associated
with estimates of gross and net C sequestration for each of the 50 states and the District of Columbia was based on
standard error estimates for each of the state-level sequestration estimates (Table 6-89). These estimates are
based on field data collected in each of the 50 states and the District of Columbia, and uncertainty in these
estimates increases as they are scaled up to the national level.
Additional uncertainty is associated with the biomass models, conversion factors, and decomposition assumptions
used to calculate C sequestration and emission estimates (Nowak et al. 2002). These results also exclude changes
in soil C stocks, and there is likely some overlap between the settlement tree C estimates and the forest tree C
estimates (e.g., Nowak et al. 2013). Due to data limitations, urban soil flux is not quantified as part of this analysis,
while reconciliation of settlement tree and forest tree estimates will be addressed through the land-representation
effort described in the Planned Improvements section of this chapter.
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the overall uncertainty of the
sequestration estimate in 2019. The results of this quantitative uncertainty analysis are summarized in Table 6-89.
The change in C stocks in Settlement Trees in 2019 was estimated to be between -195.4 and -62.2 MMT C02 Eq. at
a 95 percent confidence level. This analysis indicates a range of 51 percent more sequestration to 52 percent less
sequestration than the 2019 flux estimate of-129.8 MMT C02 Eq.
Land Use, Land-Use Change, and Forestry 6-131

-------
Table 6-89: Approach 2 Quantitative Uncertainty Estimates for Net CO2 Flux from Changes
in C Stocks in Settlement Trees (MMT CO2 Eq. and Percent)
Source Gas
2019 Flux Estimate
Uncertainty Range Relative to Flux Estimate
(MMTCOz Eq.)
(MMT CO? Eq.)
(%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
Changes in C Stocks in
CO2
Settlement Trees
(129.8)
(195.42) (62.22)
-51% 52%
Note: Parentheses indicate negative values or net sequestration.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
Tier 1 and Tier 2 QA/QC activities were conducted consistent with the U.S. QA/QC plan. Source-specific quality
control measures for settlement trees included checking input data, documentation, and calculations to ensure
data were properly handled through the inventory process. Errors that were found during this process were
corrected as necessary.
Recalculations Discussion
There were no recalculations to the 1990 through 2018 time series in this Inventory.
Planned Improvements
A consistent representation of the managed land base in the United States is discussed in Section 6 Representation
of the U.S. Land Base, and discusses a planned improvement by the USDA Forest Service to reconcile the overlap
between Settlement Trees and the forest land categories. Estimates for Settlement Trees are based on tree cover in
settlement areas. What needs to be determined is how much of this settlement area tree cover might also be
accounted for in "forest" area assessments as some of these forests may fall within settlement areas. For example,
"forest" as defined by the USDA Forest Service Forest Inventory and Analysis (FIA) program fall within urban areas.
Nowak et al. (2013) estimates that 1.5 percent of forest plots measured by the FIA program fall within land
designated as Census urban, suggesting that approximately 1.5 percent of the C reported in the Forest source
category might also be counted in the urban areas. The potential overlap with settlement areas is unknown. Future
research may also enable more complete coverage of changes in the C stock of trees for all settlements land.
To provide more accurate emissions estimates in the future, the following actions will be taken:
a)	Photo-interpret settlement tree cover in 2021 to update tree cover estimates and trends
b)	Update photo-interpretation for settlement areas using 2016 NLCD developed land information
c)	Develop spatially explicit and spatially continuous representations of land to eliminate the overlap
between forest and settlement areas, as well as allow for improved estimates in "settlement areas"
N^O Emissions from Settlement Soils fCRF Source Category
4E1)
Of the synthetic N fertilizers applied to soils in the United States, approximately 1 to 2 percent are currently
applied to lawns, golf courses, and other landscaping within settlement areas, and contributes to soil N20
emissions. The area of settlements is considerably smaller than other land uses that are managed with fertilizer,
particularly cropland soils, and therefore, settlements account for a smaller proportion of total synthetic fertilizer
6-132 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
application in the United States. In addition to synthetic N fertilizers, a portion of surface applied biosolids (i.e.,
treated sewage sludge) is used as an organic fertilizer in settlement areas, and drained organic soils (i.e., soils with
high organic matter content, known as Histosols) also contribute to emissions of soil N20.
N additions to soils result in direct and indirect N20 emissions. Direct emissions occur on-site due to the N
additions in the form of synthetic fertilizers and biosolids as well as enhanced mineralization of N in drained
organic soils. Indirect emissions result from fertilizer and biosolids N that is transformed and transported to
another location in a form other than N20 (i.e., ammonia [NH3] and nitrogen oxide [NOx] volatilization, nitrate
[N03 ] leaching and runoff), and later converted into N20 at the off-site location. The indirect emissions are
assigned to settlements because the management activity leading to the emissions occurred in settlements.
Total N20 emissions from soils in Settlements Remaining Settlements77 are 2.4 MMT C02 Eq. (8 kt of N20) in 2019.
There is an overall increase of 20 percent from 1990 to 2019 due to an expanding settlement area leading to more
synthetic N fertilizer applications that peaked in the mid-2000s. Inter-annual variability in these emissions is
directly attributable to variability in total synthetic fertilizer consumption, area of drained organic soils, and
biosolids applications in the United States. Emissions from this source are summarized in Table 6-90.
Table 6-90: N2O Emissions from Soils in Settlements Remaining Settlements (MMT CO2 Eq.
and kt N2O)

1990
2005
2015
2016
2017
2018
2019
MMTCO2 Eq.







Direct N20 Emissions from Soils
1.6
2.5
1.8
1.9
2.0
2.0
2.0
Synthetic Fertilizers
0.8
1.6
0.8
0.9
1.0
1.0
1.1
Biosolids
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Drained Organic Soils
0.6
0.7
0.8
0.8
0.8
0.8
0.8
Indirect N20 Emissions from Soils
0.4
0.6
0.3
0.3
0.4
0.4
0.4
Total
2.0
3.1
2.2
2.2
2.3
2.4
2.4
kt N20







Direct N20 Emissions from Soils
6
9
6
6
7
7
7
Synthetic Fertilizers
3
6
3
3
3
4
4
Biosolids
1
1
1
1
1
1
1
Drained Organic Soils
•J
2
3
3
3
3
3
Indirect N20 Emissions from Soils
1
2
1
1
1
1
1
Total
7
11
7
8
8
8
8
Methodology
For settlement soils, the IPCC Tier 1 approach is used to estimate soil N20 emissions from synthetic N fertilizer,
biosolids additions, and drained organic soils. Estimates of direct N20 emissions from soils in settlements are based
on the amount of N in synthetic commercial fertilizers applied to settlement soils, the amount of N in biosolids
applied to non-agricultural land and surface disposal (see Section 7.2—Wastewater Treatment for a detailed
discussion of the methodology for estimating biosolids available for non-agricultural land application), and the area
of drained organic soils within settlements.
Nitrogen applications to settlement soils are estimated using data compiled by the USGS (Brakebill and Gronberg
2017). The USGS estimated on-farm and non-farm fertilizer use is based on sales records at the county level from
1987 through 2012 (Brakebill and Gronberg 2017). Non-farm N fertilizer is assumed to be applied to settlements
and forest lands; values for 2013 through 2019 are based on 2012 values adjusted for annual total N fertilizer sales
in the United States because there are no activity data on non-farm application after 2012. Settlement application
is calculated by subtracting forest application from total non-farm fertilizer use. The total amount of fertilizer N
77 Estimates of Soil N20 for Settlements Remaining Settlements include emissions from Land Converted to Settlements because
it was not possible to separate the activity data.
Land Use, Land-Use Change, and Forestry 6-133

-------
applied to settlements is multiplied by the IPCC default emission factor (1 percent) to estimate direct N20
emissions (IPCC 2006) for 1990 to 2012.
Biosolids applications are derived from national data on biosolids generation, disposition, and N content (see
Section 7.2, Wastewater Treatment for further detail). The total amount of N resulting from these sources is
multiplied by the IPCC default emission factor for applied N (one percent) to estimate direct N20 emissions (IPCC
2006) for 1990 to 2019.
The IPCC (2006) Tier 1 method is also used to estimate direct N20 emissions due to drainage of organic soils in
settlements at the national scale. Estimates of the total area of drained organic soils are obtained from the 2015
NRI (USDA-NRCS 2018) using soils data from the Soil Survey Geographic Database (SSURGO) (Soil Survey Staff
2011). To estimate annual emissions from 1990 to 2015, the total area is multiplied by the IPCC default emission
factor for temperate regions (IPCC 2006). This Inventory does not include soil N20 emissions from drainage of
organic soils in Alaska and federal lands, although this is a planned improvement for a future Inventory.
For indirect emissions, the total N applied from fertilizer and biosolids is multiplied by the IPCC default factors of
10 percent for volatilization and 30 percent for leaching/runoff to calculate the amount of N volatilized and the
amount of N leached/runoff. The amount of N volatilized is multiplied by the IPCC default factor of one percent for
the portion of volatilized N that is converted to N20 off-site and the amount of N leached/runoff is multiplied by
the IPCC default factor of 0.075 percent for the portion of leached/runoff N that is converted to N20 off-site. The
resulting estimates are summed to obtain total indirect emissions from 1990 to 2015 for fertilizer and from 1990
to 2019 for biosolids.
A linear extrapolation of the trend in the time series is applied to estimate the direct and indirect N20 emissions
for fertilizer and drainage of organic soils from 2016 to 2019 because N fertilizer inputs and area data for these two
sources have not been compiled for the latter part of the time series. Specifically, a linear regression model with
autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) is used to estimate the trend in
emissions over time from 1990 to 2015, and in turn, the trend is used to approximate the 2016 to 2019 emissions.
The time series will be recalculated for the years beyond 2015 in a future inventory with the methods described
above for 1990 to 2015. This Inventory does incorporate updated activity data on biosolids application in
settlements through 2019.
Uncertainty and Time-Series Consistency
The amount of N20 emitted from settlement soils depends not only on N inputs and area of drained organic soils,
but also on a large number of variables that can influence rates of nitrification and denitrification, including organic
C availability; rate, application method, and timing of N input; oxygen gas partial pressure; soil moisture content;
pH; temperature; and irrigation/watering practices. The effect of the combined interaction of these variables on
N20 emissions is complex and highly uncertain. The IPCC default methodology does not explicitly incorporate any
of these variables, except variation in the total amount of fertilizer N and biosolids application, which in turn, leads
to uncertainty in the results.
Uncertainties exist in both the fertilizer N and biosolids application rates in addition to the emission factors.
Uncertainty in fertilizer N application is assigned a default level of ±50 percent.78 Uncertainty in the area of
drained organic soils is based on the estimated variance from the NRI survey (USDA-NRCS 2018). For 2016 to 2019,
there is also additional uncertainty associated with the fit of the linear regression model for the data splicing
methods.
For biosolids, there is uncertainty in the amounts of biosolids applied to non-agricultural lands and used in surface
disposal. These uncertainties are derived from variability in several factors, including: (1) N content of biosolids; (2)
total sludge applied in 2000; (3) wastewater existing flow in 1996 and 2000; and (4) the biosolids disposal practice
78 No uncertainty is provided with the USGS fertilizer consumption data (Brakebill and Gronberg 2017) so a conservative ±50
percent is used in the analysis. Biosolids data are also assumed to have an uncertainty of ±50 percent.
6-134 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
distributions to non-agricultural land application and surface disposal. In addition, there is uncertainty in the direct
and indirect emission factors that are provided by IPCC (2006).
Uncertainty is propagated through the calculations of N20 emissions from fertilizer N and drainage of organic soils
based on a Monte Carlo analysis. The results are combined with the uncertainty in N20 emissions from the
biosolids application using simple error propagation methods (IPCC 2006). The results are summarized in Table
6-91. Direct N20 emissions from soils in Settlements Remaining Settlements in 2019 are estimated to be between
1.4 and 2.9 MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 29 percent below to 41 percent
above the 2019 emission estimate of 2.0 MMT C02 Eq. Indirect N20 emissions in 2019 are between 0.2 and 0.5
MMT C02 Eq., ranging from 39 percent below to 39 percent above the estimate of 0.4 MMT C02 Eq.
Table 6-91: Quantitative Uncertainty Estimates of N2O Emissions from Soils in Settlements
Remaining Settlements (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emissions
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTC02Eq.) (%)
Settlements Remaining
Settlements


Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Direct N20 Emissions from Soils
N20
2.0
1.4
2.9
-29% 41%
Indirect N20 Emissions from Soils
n2o
0.4
0.2
0.5
-39% 39%
Note: These estimates include direct and indirect N20 emissions from Settlements Remaining Settlements and Land
Converted to Settlements because it was not possible to separate the activity data.
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in the
Introduction and Methodology sections.
QA/QC and Verification
The spreadsheet containing fertilizer, drainage of organic soils, and biosolids applied to settlements and
calculations for N20 and uncertainty ranges have been checked. An error was found in the uncertainty calculation
that was corrected.
Recalculations Discussion
There were no recalculations to the 1990 through 2018 time series in this Inventory.
Planned Improvements
This source will be extended to include soil N20 emissions from drainage of organic soils in settlements of Alaska
and federal lands in order to provide a complete inventory of emissions for this category. Data on fertilizer amount
and area of drained organic soils will be compiled to update emissions estimates from 2016 to 2019 in a future
Inventory.
Changes in Yard Trimmings and Food Scrap Carbon Stock' in
Landfills (CRF Category 4E1)
In the United States, yard trimmings (i.e., grass clippings, leaves, and branches) and food scraps account for a
significant portion of the municipal waste stream, and a large fraction of the collected yard trimmings and food
scraps are put in landfills. Carbon (C) contained in landfilled yard trimmings and food scraps can be stored for very
long periods.
Land Use, Land-Use Change, and Forestry 6-135

-------
Carbon storage estimates within the Inventory are associated with particular land uses. For example, harvested
wood products are reported under Forest Land Remaining Forest Land because these wood products originated
from the forest ecosystem. Similarly, C stock changes in yard trimmings and food scraps are reported under
Settlements Remaining Settlements because the bulk of the C, which comes from yard trimmings, originates from
settlement areas. While the majority of food scraps originate from cropland and grassland, in this Inventory they
are reported with the yard trimmings in the Settlements Remaining Settlements section. Additionally, landfills are
considered part of the managed land base under settlements (see Section 6.1 Representation of the U.S. Land
Base), and reporting these C stock changes that occur entirely within landfills fits most appropriately within the
Settlements Remaining Settlements section.
Both the estimated amount of yard trimmings collected annually and the fraction that is landfilled have been
declining. In 1990, over 53 million metric tons (wet weight) of yard trimmings and food scraps are estimated to
have been generated (i.e., put at the curb for collection to be taken to disposal sites or to composting facilities)
(EPA 2019). Since then, programs banning or discouraging yard trimmings disposal have led to an increase in
backyard composting and the use of mulching mowers, and a consequent estimated 0.5 percent decrease between
1990 and 2019 in the tonnage of yard trimmings generated (i.e., collected for composting or disposal in landfills).
At the same time, an increase in the number of municipal composting facilities has reduced the proportion of
collected yard trimmings that are discarded in landfills—from 72 percent in 1990 to 25 percent in 2019. The net
effect of the reduction in generation and the increase in composting is a 66 percent decrease in the quantity of
yard trimmings disposed of in landfills since 1990.
Food scrap generation has grown by an estimated 70 percent since 1990, and while the proportion of total food
scraps generated that are eventually discarded in landfills has decreased slightly, from an estimated 82 percent in
1990 to 75 percent in 2019, the tonnage disposed of in landfills has increased considerably (by an estimated 57
percent) due to the increase in food scrap generation. Although the total tonnage of food scraps disposed of in
landfills has increased from 1990 to 2019, the difference in the amount of food scraps added from one year to the
next generally decreased, and consequently the annual carbon stock net changes from food scraps have generally
decreased as well (as shown in Table 6-92 and Table 6-93). As described in the Methodology section, the carbon
stocks are modeled using data on the amount of food scraps landfilled since 1960. These food scraps decompose
over time, producing CH4 and C02. Decomposition happens at a higher rate initially, then decreases. As
decomposition decreases, the carbon stock becomes more stable. Because the cumulative carbon stock left in the
landfill from previous years is (1) not decomposing as much as the carbon introduced from food scraps in a single
more recent year; and (2) is much larger than the carbon introduced from food scraps in a single more recent year,
the total carbon stock in the landfill is primarily driven by the more stable 'older' carbon stock, thus resulting in
less annual change in later years.
Overall, the decrease in the landfill disposal rate of yard trimmings has more than compensated for the increase in
food scrap disposal in landfills, and the net result is a decrease in annual net change landfill C storage from 24.5
MMT C02 Eq. (6.7 MMT C) in 1990 to 10.2 MMT C02 Eq. (2.8 MMT C) in 2019 (Table 6-92 and Table 6-93).
Table 6-92: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills
(MMT COz Eq.)
Carbon Pool
1990
2005
2015
2016
2017
2018
2019
Yard Trimmings
(20.1)
: (7.5)
(7.2)
(6.3)
(6.3)
(6.4)
(6.4)
Grass
(1.7)
(0.6)
(0.6)
(0.5)
(0.5)
(0.6)
(0.6)
Leaves
(8.7)
5 (3.4)
(3.4)
(3.0)
(3.0)
(3.0)
(3.0)
Branches
(9.8)
i (3.4)
(3.2)
(2.8)
(2.8)
(2.8)
(2.8)
Food Scraps
(4.4)
(3.9)
(3.9)
(3.7)
(3.5)
(3.4)
(3.8)
Total Net Flux
(24.5)
(11.4)
(11.1)
(10.0)
(9.8)
(9.8)
(10.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
Table 6-93: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills
(MMT C)
Carbon Pool	1990	2005	2015 2016 2017 2018 2019
6-136 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Yard Trimmings
(5.5)
(2.0)
(2.0)
(1.7)
(1.7)
(1.7)
(1.7)
Grass
(0.5)
(0.2)
(0.2)
(0.1)
(0.1)
(0.2)
(0.2)
Leaves
(2.4)
(0.9)
(0.9)
(0.8)
(0.8)
(0.8)
(0.8)
Branches
(2.7)
(0.9)
(0.9)
(0.8)
(0.8)
(0.8)
(0.8)
Food Scraps
(1.2)
(1.1)
(1.1)
(1.0)
(1.0)
(0.9)
(1.0)
Total Net Flux
(6.7)
(3.1)
(3.0)
(2.7)
(2.7)
(2.7)
(2.8)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
Methodology
When wastes of biogenic origin (such as yard trimmings and food scraps) are landfilled and do not completely
decompose, the C that remains is effectively removed from the C cycle. Empirical evidence indicates that yard
trimmings and food scraps do not completely decompose in landfills (Barlaz 1998, 2005, 2008; De la Cruz and
Barlaz 2010), and thus the stock of C in landfills can increase, with the net effect being a net atmospheric removal
of C. Estimates of net C flux resulting from landfilled yard trimmings and food scraps were developed by estimating
the change in landfilled C stocks between inventory years and are based on methodologies presented for the Land
Use, Land-Use Change, and Forestry sector in IPCC (2003) and the 2006IPCC Guidelines for National Greenhouse
Gas Inventories (IPCC 2006). Carbon stock estimates were calculated by determining the mass of landfilled C
resulting from yard trimmings and food scraps discarded in a given year; adding the accumulated landfilled C from
previous years; and subtracting the mass of C that was landfilled in previous years and has since decomposed and
been emitted as C02 and CH4.
To determine the total landfilled C stocks for a given year, the following data and factors were assembled: (1) The
composition of the yard trimmings; (2) the mass of yard trimmings and food scraps discarded in landfills; (3) the C
storage factor of the landfilled yard trimmings and food scraps; and (4) the rate of decomposition of the
degradable C. The composition of yard trimmings was assumed to be 30 percent grass clippings, 40 percent leaves,
and 30 percent branches on a wet weight basis (Oshins and Block 2000). The yard trimmings were subdivided,
because each component has its own unique adjusted C storage factor (i.e., moisture content and C content) and
rate of decomposition. The mass of yard trimmings and food scraps disposed of in landfills was estimated by
multiplying the quantity of yard trimmings and food scraps discarded by the proportion of discards managed in
landfills. Data on discards (i.e., the amount generated minus the amount diverted to centralized composting
facilities) for both yard trimmings and food scraps were taken primarily from Advancing Sustainable Materials
Management: Facts and Figures 2017 (EPA 2019), which provides data for 1960,1970,1980,1990, 2000, 2005,
2010, 2015, 2016, and 2017. To provide data for some of the missing years, detailed backup data were obtained
from the 2012, 2013, and 2014, and 2015 versions of the Advancing Sustainable Materials Management: Facts and
Figures reports (EPA 2019), as well as historical data tables that EPA developed for 1960 through 2012 (EPA 2016).
Remaining years in the time series for which data were not provided were estimated using linear interpolation.
Since the Advancing Sustainable Materials Management: Facts and Figures reports for 2018 and 2019 were
unavailable, landfilled material generation, recovery, and disposal data for 2018 and 2019 were set equal to 2017
values.
The amount of C disposed of in landfills each year, starting in 1960, was estimated by converting the discarded
landfilled yard trimmings and food scraps from a wet weight to a dry weight basis, and then multiplying by the
initial (i.e., pre-decomposition) C content (as a fraction of dry weight). The dry weight of landfilled material was
calculated using dry weight to wet weight ratios (Tchobanoglous et al. 1993, cited by Barlaz 1998) and the initial C
contents and the C storage factors were determined by Barlaz (1998, 2005, 2008) (Table 6-94).
The amount of C remaining in the landfill for each subsequent year was tracked based on a simple model of C fate.
As demonstrated by Barlaz (1998, 2005, 2008), a portion of the initial C resists decomposition and is essentially
persistent in the landfill environment. Barlaz (1998, 2005, 2008) conducted a series of experiments designed to
measure biodegradation of yard trimmings, food scraps, and other materials, in conditions designed to promote
decomposition (i.e., by providing ample moisture and nutrients). After measuring the initial C content, the
materials were placed in sealed containers along with methanogenic microbes from a landfill. Once decomposition
was complete, the yard trimmings and food scraps were re-analyzed for C content; the C remaining in the solid
Land Use, Land-Use Change, and Forestry 6-137

-------
sample can be expressed as a proportion of the initial C (shown in the row labeled "C Storage Factor, Proportion of
Initial C Stored (%)" in Table 6-94).
The modeling approach applied to simulate U.S. landfill C flows builds on the findings of Barlaz (1998, 2005, 2008).
The proportion of C stored is assumed to persist in landfills. The remaining portion is assumed to degrade over
time, resulting in emissions of CH4 and C02. (The CH4 emissions resulting from decomposition of yard trimmings
and food scraps are reported in the Waste chapter.) The degradable portion of the C is assumed to decay
according to first-order kinetics. The decay rates for each of the materials are shown in Table 6-94.
The first-order decay rates, k, for each waste component are derived from De la Cruz and Barlaz (2010):
•	De la Cruz and Barlaz (2010) calculate first-order decay rates using laboratory data published in Eleazer et
al. (1997), and a correction factor,/, is calculated so that the weighted average decay rate for all
components is equal to the EPA AP-42 default decay rate (0.04) for mixed MSW for regions that receive
more than 25 inches of rain annually (EPA 1995). Because AP-42 values were developed using landfill data
from approximately 1990, De la Cruz and Barlaz used 1990 waste composition for the United States from
EPA's Characterization of Municipal Solid Waste in the United States: 1990 Update (EPA 1991) to calculate
/. De la Cruz and Barlaz multiplied this correction factor by the Eleazer et al. (1997) decay rates of each
waste component to develop field-scale first-order decay rates.
•	De la Cruz and Barlaz (2010) also use other assumed initial decay rates for mixed MSW in place of the AP-
42 default value based on different types of environments in which landfills in the United States are
located, including dry conditions (less than 25 inches of rain annually, k=0.02) and bioreactor landfill
conditions (moisture is controlled for rapid decomposition, /c=0.12).
Similar to the methodology in the Landfills section of the Inventory (Section 7.1), which estimates CH4 emissions,
the overall MSW decay rate is estimated by partitioning the U.S. landfill population into three categories based on
annual precipitation ranges of: (1) Less than 20 inches of rain per year, (2) 20 to 40 inches of rain per year, and (3)
greater than 40 inches of rain per year. These correspond to overall MSW decay rates of 0.020, 0.038, and 0.057
year"1, respectively. De la Cruz and Barlaz (2010) calculate component-specific decay rates corresponding to the
first value (0.020 year"1), but not for the other two overall MSW decay rates.
To maintain consistency between landfill methodologies across the Inventory, EPA developed correction factors (/)
for decay rates of 0.038 and 0.057 year"1 through linear interpolation. A weighted national average component-
specific decay rate is calculated by assuming that waste generation is proportional to population (the same
assumption used in the landfill methane emission estimate), based on population data from the 2000 U.S. Census.
The percent of census population is calculated for each of the three categories of annual precipitation (noted in
the previous paragraph); the population data are used as a surrogate for the number of landfills in each annual
precipitation category. The component-specific decay rates are shown in Table 6-94.
De la Cruz and Barlaz (2010) also use other assumed initial decay rates for mixed MSW in place of the AP-42
default value based on different types of environments in which landfills in the United States are located, including
dry conditions (less than 25 inches of rain annually, k=0.02) and bioreactor landfill conditions (moisture is
controlled for rapid decomposition, /c=0.12).
For each of the four materials (grass, leaves, branches, food scraps), the stock of C in landfills for any given year is
calculated according to Equation 1:
LFC,,t= Ł Wi,n x (1 - MG) x ICOx {[CSi x fCG\ + [(1 - (CS,x ICC])) x e-^-")]}
n
where,
t
Year for which C stocks are being estimated (year),
Waste type for which C stocks are being estimated (grass, leaves, branches, food
scraps),
Stock of C in landfills in year t, for waste / (metric tons),
Mass of waste / disposed of in landfills in year n (metric tons, wet weight),
6-138 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
n	=	Year in which the waste was disposed of (year, where 1960 < n < t),
MCi	=	Moisture content of waste / (percent of water),
CSi	=	Proportion of initial C that is stored for waste / (percent),
ICCi	=	Initial C content of waste / (percent),
e	=	Natural logarithm, and
k	=	First-order decay rate for waste/', (year-1).
For a given year t, the total stock of C in landfills (TLFCt) is the sum of stocks across all four materials (grass, leaves,
branches, food scraps). The annual flux of C in landfills (Ft) for year t is calculated in as the change in C stock
compared to the preceding year according to Equation 2:
Ft = TLFCt- TLFCt- u
Thus, as seen in Equation 1, the C placed in a landfill in year n is tracked for each year t through the end of the
inventory period. For example, disposal of food scraps in 1960 resulted in depositing about 1,135,000 metric tons
of C in landfills. Of this amount, 16 percent (179,000 metric tons) is persistent; the remaining 84 percent (956,000
metric tons) is degradable. By 1965, more than half of the degradable portion (507,000 metric tons) decomposes,
leaving a total of 628,000 metric tons (the persistent portion, plus the remainder of the degradable portion).
Continuing the example, by 2019, the total food scraps C originally disposed of in 1960 had declined to 179,000
metric tons (i.e., virtually all degradable C had decomposed). By summing the C remaining from 1960 with the C
remaining from food scraps disposed of in subsequent years (1961 through 2019), the total landfill C from food
scraps in 2019 was 47.3 million metric tons. This value is then added to the C stock from grass, leaves, and
branches to calculate the total landfill C stock in 2019, yielding a value of 280.1 million metric tons (as shown in
Table 6-95). In the same way total net flux is calculated for forest C and harvested wood products, the total net flux
of landfill C for yard trimmings and food scraps for a given year (Table 6-93) is the difference in the landfill C stock
for the following year (2020 C stock was forecast using 1990 to 2019 C stocks) and the stock in the current year.
For example, the net change in 2019 shown in Table 6-93 (2.8 MMT C) is equal to the stock in 2020 (282.9 MMT C)
minus the stock in 2019 (280.1 MMT C). The C stocks calculated through this procedure are shown in Table 6-95.
Table 6-94: Moisture Contents, C Storage Factors (Proportions of Initial C Sequestered),
Initial C Contents, and Decay Rates for Yard Trimmings and Food Scraps in Landfills
Yard Trimmings
Variable		;	 Food Scraps
Grass	Leaves Branches
Moisture Content (% H20)
70
30
10
70
C Storage Factor, Proportion of Initial C




Stored (%)
53
85
77
16
Initial C Content (%)
45
46
49
51
Decay Rate (year-1)
0.313
0.179
0.015
0.151
Note: The decay rates are presented as weighted averages based on annual precipitation categories
and population residing in each precipitation category.
Table 6-95: C Stocks in Yard Trimmings and Food Scraps in Landfills (MMT C)
Carbon Pool
1990
2005
2015
2016
2017
2018
2019
2020
Yard Trimmings
156.0
203.1
225.7
227.7
229.4
261.1
232.8
234.6
Branches
14.6
18.1
20.2
20.3
20.5
20.6
20.8
20.9
Leaves
66.7
87.4
97.7
98.6
99.4
100.2
101.0
101.9
Grass
74.7
97.7
107.8
108.7
109.5
110.2
111.0
111.8
Food Scraps
17.9
33.2
43.3
44.4
45.4
46.3
47.3
48.3
Total Carbon Stocks
173.9
236.3
269.0
272.0
274.8
277.4
280.1
282.9
Note: Totals may not sum due to independent rounding.
a 2020 C stock estimate was forecasted using 1990 to 2019 data.
Land Use, Land-Use Change, and Forestry 6-139

-------
To develop the 2020 C stock estimate, estimates of yard trimming and food scrap carbon stocks were forecasted
for 2020, based on data from 1990 through 2019. These forecasted values were used to calculate net changes in
carbon stocks for 2019. Excel's FORECAST.ETS function was used to predict a 2020 value using historical data via an
algorithm called "Exponential Triple Smoothing." This method determined the overall trend and provided
appropriate carbon stock estimates for 2020.
Uncertainty and Time-Series Consistency
The uncertainty analysis for landfilled yard trimmings and food scraps includes an evaluation of the effects of
uncertainty for the following data and factors: disposal in landfills per year (tons of C), initial C content, moisture
content, decay rate, and proportion of C stored. The C storage landfill estimates are also a function of the
composition of the yard trimmings (i.e., the proportions of grass, leaves and branches in the yard trimmings
mixture). There are respective uncertainties associated with each of these factors.
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the overall uncertainty of the
sequestration estimate for 2019. The results of the Approach 2 quantitative uncertainty analysis are summarized in
Table 6-96. Total yard trimmings and food scraps C02 flux in 2019 was estimated to be between -16.3 and -4.4
MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 61 percent below to 57 percent above the
2019 flux estimate of-10.2 MMT C02 Eq.
Table 6-96: Approach 2 Quantitative Uncertainty Estimates for CO2 Flux from Yard
Trimmings and Food Scraps in Landfills (MMT CO2 Eq. and Percent)

2019 Flux



Source Gas
Estimate
Uncertainty Range Relative to Flux Estimate3

(MMTCOz Eq.)
(MMTCOz
Eq.)
(%)


Lower
Upper
Lower Upper


Bound
Bound
Bound Bound
Yard Trimmings and Food
c CO2
Scraps
(10.2)
(16.3)
(4.4)
-61% 57%
Note: Parentheses indicate negative values or net C sequestration.
a Range of flux estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above.
QA/QC and Verification
Tier 1 and Tier 2 QA/QC activities were conducted consistent with the U.S. QA/QC plan. Source-specific quality
control measures for Landfilled Yard Trimmings and Food Scraps included checking that input data were properly
transposed within the spreadsheet, checking calculations were correct, and confirming that all activity data and
calculations documentation was complete and updated to ensure data were properly handled through the
inventory process.
Order of magnitude checks and checks of time-series consistency were performed to ensure data were updated
correctly and any changes in emissions estimates were reasonable and reflected changes in activity data. An
annual change trend analysis was also conducted to ensure the validity of the emissions estimates. Errors that
were found during this process were corrected as necessary.
Recalculations Discussion
The current Inventory has been revised to reflect updated data from the most recent Advancing Sustainable
Materials Management: Facts and Figures report. Recalculations based on these updates resulted in 2.1 percent
change in the annual carbon stocks and sequestration values as compared to the previous inventory values. The
6-140 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
largest changes occurred in the most recent years: a 7.7 percent increase in sequestration in 2015, a 16.3 percent
increase in sequestration in 2016, an 18.4 percent increase in sequestration in 2017, and a 19.0 percent increase in
sequestration in 2018. Large changes can be attributed to updates to 2015, 2016, and 2017 yard trimmings and
food scraps landfilled values reported in Advancing Sustainable Materials Management: Facts and Figures 2017
(EPA 2019). A large increase in sequestration in 2018 can be attributed to updated generation values as well - 2018
landfill data were unavailable and were reported as 2017 values. An increase of less than 0.1 percent occurred in
2014 due to a small increase in calculated leaf C mass and grass C mass values for 2015 between the current and
previous Inventory.
Planned Improvements
Future work is planned to evaluate the consistency between the estimates of C storage described in this chapter
and the estimates of landfill CH4 emissions described in the Waste chapter. For example, the Waste chapter does
not distinguish landfill CH4 emissions from yard trimmings and food scraps separately from landfill CH4 emissions
from total bulk (i.e., municipal solid) waste, which includes yard trimmings and food scraps. In future years, as time
and resources allow, EPA will further evaluate both categories to ensure consistency.
In addition, data from recent peer-reviewed literature will be evaluated that may modify the default C storage
factors, initial C contents, and decay rates for yard trimmings and food scraps in landfills. Based upon this
evaluation, changes may be made to the default values.
EPA will also investigate updates to the decay rate estimates for food scraps, leaves, grass, and branches. Currently
the inventory calculations use 2010 U.S. Census data. EPA will evaluate using decay rates that vary over time based
on Census data changes over time.
Yard waste composition will also be investigated to determine if changes need to be made based on changes in
residential practices, a review of available literature will be conducted to determine if there are changes in the
allocation of yard trimmings. For example, leaving grass clippings in place is becoming a more common practice,
thus reducing the percentage of grass clippings in yard trimmings disposed in landfills. In addition, agronomists
may be consulted for determining the mass of grass per acre on residential lawns to provide an estimate of total
grass generation for comparison with Inventory estimates.
Finally, EPA will review available data to ensure all types of landfilled yard trimmings and food scraps are being
included in Inventory estimates, such as debris from road construction and commercial food waste not included in
other chapter estimates.
6.11 Land Converted to Settlements (CRF
Category 4E2)
Land Converted to Settlements includes all settlements in an Inventory year that had been in another land use(s)
during the previous 20 years (USDA-NRCS 2015).79 For example, cropland, grassland or forest land converted to
settlements during the past 20 years would be reported in this category. Converted lands are retained in this
category for 20 years as recommended by IPCC (2006). This Inventory includes all settlements in the conterminous
United States and Hawaii, but does not include settlements in Alaska. Areas of drained organic soils on settlements
in federal lands are also not included in this Inventory. Consequently, there is a discrepancy between the total
79 NRI survey locations are classified according to land-use histories starting in 1979, and consequently the classifications are
based on less than 20 years from 1990 to 2001. This may have led to an underestimation of Land Converted to Settlements in
the early part of the time series to the extent that some areas are converted to settlements from 1971 to 1978.
Land Use, Land-Use Change, and Forestry 6-141

-------
amount of managed area for Land Converted to Settlements (see Section 6.1 Representation of the U.S. Land Base)
and the settlements area included in the Inventory analysis.
Land use change can lead to large losses of carbon (C) to the atmosphere, particularly conversions from forest land
(Houghton et al. 1983). Moreover, conversion of forest to another land use (i.e., deforestation) is one of the largest
anthropogenic sources of emissions to the atmosphere globally (Schimel 1995), although this source may be
declining globally according to a recent assessment (Tubiello et al. 2015).
IPCC (2006) recommends reporting changes in biomass, dead organic matter, and soil organic C stocks due to land
use change. All soil organic C stock changes are estimated and reported for Land Converted to Settlements, but
there is limited reporting of other pools in this Inventory. Loss of aboveground and belowground biomass, dead
wood and litter C are reported for Forest Land Converted to Settlements, but not for other land use conversions to
settlements.
Forest Land Converted to Settlements is the largest source of emissions from 1990 to 2019, accounting for
approximately 76 percent of the average total loss of C among all of the land use conversions in Land Converted to
Settlements. Losses of aboveground and belowground biomass, dead wood and litter C losses in 2019 are 36.9, 7.2,
6.7, and 9.9 MMT C02 Eq., respectively (10.1, 2.0,1.8, and 2.7 MMT C). Mineral and organic soils also lost 16.2 and
2.4 MMT C02 Eq. in 2019 (4.4 and 0.6 MMT C). The total net flux is 79.2 MMT C02 Eq. in 2019 (21.6 MMT C), which
is a 26 percent increase in C02 emissions compared to the emissions in the initial reporting year of 1990 (Tables 6-
97 and 6-98). The main driver of net emissions for this source category is the conversion of forest land to
settlements, with large losses of biomass, deadwood and litter C.
Table 6-97: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Settlements (MMT CO2 Eq.)

1990
2005
2015
2016
2017
2018
2019
Cropland Converted to







Settlements
3.4
9.8
6.2
6.0
6.0
5.9
5.9
Mineral Soils
2.8
8.4
5.3
5.2
5.2
5.2
5.1
Organic Soils
0.6
1.3
0.8
0.8
0.8
0.8
0.8
Forest Land Converted to







Settlements
54.6
59.9
63.0
62.9
62.9
62.9
62.9
Aboveground Live Biomass
32.5
35.1
36.9
36.9
36.9
36.9
36.9
Belowground Live Biomass
6.3
6.8
7.2
7.2
7.2
7.2
7.2
Dead Wood
5.8
6.3
6.7
6.7
6.7
6.7
6.7
Litter
8.7
9.4
9.9
9.9
9.9
9.9
9.9
Mineral Soils
1.1
2.0
2.0
1.9
1.9
1.9
1.9
Organic Soils
0.2
0.3
0.3
0.3
0.3
0.3
0.3
Grassland Converted to







Settlements
5.2
16.3
11.9
11.3
11.3
11.3
11.3
Mineral Soils
4.6
14.9
11.0
10.4
10.4
10.4
10.4
Organic Soils
0.6
1.4
0.9
0.9
0.9
0.9
0.9
Other Lands Converted to







Settlements
(0.4)
(1.4)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
Mineral Soils
(0.4)
(1.6)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
Organic Soils
+
0.2
0.1
0.1
0.1
0.1
0.1
Wetlands Converted to







Settlements
+
0.5
0.4
0.4
0.4
0.4
0.4
Mineral Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Organic Soils
+
0.4
0.3
0.3
0.3
0.3
0.3
Total Aboveground Biomass Flux
32.5
35.1
36.9
36.9
36.9
36.9
36.9
Total Belowground Biomass Flux
6.3
6.8
7.2
7.2
7.2
7.2
7.2
Total Dead Wood Flux
5.8
6.3
6.7
6.7
6.7
6.7
6.7
Total Litter Flux
8.7
9.4
9.9
9.9
9.9
9.9
9.9
Total Mineral Soil Flux
8.1
23.8
17.0
16.3
16.2
16.2
16.2
Total Organic Soil Flux
1.4
3.6
2.5
2.4
2.4
2.4
2.4
Total Net Flux
62.9
85.0
80.1
79.4
79.3
79.3
79.2
6-142 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Note: Parentheses indicate negative values or net C sequestration.
+ Absolute value does not exceed 0.05 MMT C02 Eq.
Table 6-98: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Settlements (MMT C)

1990
2005
2015
2016
2017
2018
2019
Cropland Converted to







Settlements
0.9
2.7
1.7
1.6
1.6
1.6
1.6
Mineral Soils
0.8
2.3
1.5
1.4
1.4
1.4
1.4
Organic Soils
0.2
0.4
0.2
0.2
0.2
0.2
0.2
Forest Land Converted to







Settlements
14.9
16.3
17.2
17.1
17.1
17.1
17.1
Aboveground Live Biomass
8.9
9.6
10.1
10.1
10.1
10.1
10.1
Belowground Live Biomass
1.7
1.9
2.0
2.0
2.0
2.0
2.0
Dead Wood
1.6
1.7
1.8
1.8
1.8
1.8
1.8
Litter
2.4
2.6
2.7
2.7
2.7
2.7
2.7
Mineral Soils
0.3
0.5
0.5
0.5
0.5
0.5
0.5
Organic Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Grassland Converted to







Settlements
1.4
4.4
3.2
3.1
3.1
3.1
3.1
Mineral Soils
1.3
4.1
3.0
2.8
2.8
2.8
2.8
Organic Soils
0.2
0.4
0.3
0.2
0.2
0.2
0.2
Other Lands Converted to







Settlements
(0.1)
(0.4)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Mineral Soils
(0.1)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.3)
Organic Soils
+
+
+
+
+
+
0.0
Wetlands Converted to







Settlements
+
0.1
0.1
0.1
0.1
0.1
0.1
Mineral Soils
+
+
+
+
+
+
0.0
Organic Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Total Aboveground Biomass







Flux
8.9
9.6
10.1
10.1
10.1
10.1
10.1
Total Belowground Biomass







Flux
1.7
1.9
2.0
2.0
2.0
2.0
2.0
Total Dead Wood Flux
1.6
1.7
1.8
1.8
1.8
1.8
1.8
Total Litter Flux
2.4
2.6
2.7
2.7
2.7
2.7
2.7
Total Mineral Soil Flux
2.2
6.5
4.6
4.4
4.4
4.4
4.4
Total Organic Soil Flux
0.4
1.0
0.7
0.7
0.7
0.6
0.6
Total Net Flux
17.1
23.2
21.9
21.6
21.6
21.6
21.6
Note: Parentheses indicate negative values or net C sequestration.
+ Absolute value does not exceed 0.05 MMT C.
fviet had ©logy
The following section includes a description of the methodology used to estimate C stock changes for Land
Converted to Settlements, including (1) loss of aboveground and belowground biomass, dead wood and litter C
with conversion of forest lands to settlements, as well as (2) the impact from all land use conversions to
settlements on soil organic C stocks in mineral and organic soils.
Biomass, Dead Wood, and Litter Carbon Stock Changes
A Tier 2 method is applied to estimate biomass, dead wood, and litter C stock changes for Forest Land Converted to
Settlements. Estimates are calculated in the same way as those in the Forest Land Remaining Forest Land category
using data from the USDA Forest Service, Forest Inventory and Analysis (FIA) program (USDA Forest Service 2020),
however there is no country-specific data for settlements so the biomass, litter, and dead wood carbon stocks on
these converted lands were assumed to be zero. The difference between the stocks is reported as the stock
Land Use, Land-Use Change, and Forestry 6-143

-------
change under the assumption that the change occurred in the year of the conversion. If FIA plots include data on
individual trees, aboveground and belowground C density estimates are based on Woodall et al. (2011).
Aboveground and belowground biomass estimates also include live understory, which is a minor component of
biomass defined as all biomass of undergrowth plants in a forest, including woody shrubs and trees less than 2.54
cm dbh. For this Inventory, it was assumed that 10 percent of total understory C mass is belowground (Smith et al.
2006). Estimates of C density are based on information in Birdsey (1996) and biomass estimates from Jenkins et al.
(2003). If FIA plots include data on standing dead trees, standing dead tree C density is estimated following the
basic method applied to live trees (Woodall et al. 2011) with additional modifications to account for decay and
structural loss (Domke et al. 2011; Harmon et al. 2011). If FIA plots include data on downed dead wood, downed
dead wood C density is estimated based on measurements of a subset of FIA plots for downed dead wood (Domke
et al. 2013; Woodall and Monleon 2008). Downed dead wood is defined as pieces of dead wood greater than 7.5
cm diameter, at transect intersection, that are not attached to live or standing dead trees. This includes stumps
and roots of harvested trees. To facilitate the downscaling of downed dead wood C estimates from the state-wide
population estimates to individual plots, downed dead wood models specific to regions and forest types within
each region are used. Litter C is the pool of organic C (also known as duff, humus, and fine woody debris) above
the mineral soil and includes woody fragments with diameters of up to 7.5 cm. A subset of FIA plots is measured
for litter C. If FIA plots include litter material, a modeling approach using litter C measurements from FIA plots is
used to estimate litter C density (Domke et al. 2016). See Annex 3.13 for more information about reference C
density estimates for forest land and the compilation system used to estimate carbon stock changes from forest
land.
Soil Carbon Stock Changes
Soil organic C stock changes are estimated for Land Converted to Settlements according to land-use histories
recorded in the 2015 USDA NRI survey for non-federal lands (USDA-NRCS 2018). Land use and some management
information were originally collected for each NRI survey location on a 5-year cycle beginning in 1982. In 1998, the
NRI program began collecting annual data, and the annual data are currently available through 2015 (USDA-NRCS
2018).
NRI survey locations are classified as Land Converted to Settlements in a given year between 1990 and 2015 if the
land use is settlements but had been classified as another use during the previous 20 years. NRI survey locations
are classified according to land-use histories starting in 1979, and consequently the classifications are based on less
than 20 years from 1990 to 1998. This may have led to an underestimation of Land Converted to Settlements in the
early part of the time series to the extent that some areas are converted to settlement between 1971 and 1978.
For federal lands, the land use history is derived from land cover changes in the National Land Cover Dataset (Yang
et al. 2018; Fry et al. 2011; Homer et al. 2007, 2015).
Mineral Soil Carbon Stock Changes
An IPCC Tier 2 method (Ogle et al. 2003) is applied to estimate C stock changes for Land Converted to Settlements
on mineral soils from 1990 to 2015. Data on climate, soil types, land-use, and land management activity are used
to classify land area and apply appropriate stock change factors (Ogle et al. 2003, 2006). Reference C stocks are
estimated using the National Soil Survey Characterization Database (USDA-NRCS 1997) with cultivated cropland as
the reference condition, rather than native vegetation as used in IPCC (2006). Soil measurements under
agricultural management are much more common and easily identified in the National Soil Survey Characterization
Database (USDA-NRCS 1997) than are soils under a native condition, and therefore cultivated cropland provide a
more robust sample for estimating the reference condition. Country-specific C stock change factors are derived
from published literature to determine the impact of management practices on soil organic C storage (Ogle et al.
2003, Ogle et al. 2006). However, there are insufficient data to estimate a set of land use, management, and input
factors for settlements. Moreover, the 2015 NRI survey data (USDA-NRCS 2018) do not provide the information
needed to assign different land use subcategories to settlements, such as turf grass and impervious surfaces, which
is needed to apply the Tier 1 factors from the IPCC guidelines (2006). Therefore, the United States has adopted a
land use factor of 0.7 to represent a net loss of soil organic C with conversion to settlements under the assumption
that there are additional soil organic C losses with land clearing, excavation and other activities associated with
6-144 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
development. More specific factor values can be derived in future inventories as data become available. See Annex
3.12 for additional discussion of the Tier 2 methodology for mineral soils.
A linear extrapolation of the trend in the time series is applied to estimate soil organic C stock changes from 2016
to 2019 because NRI activity data are not available for these years. Specifically, a linear regression model with
autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) is used to estimate the trend in stock
changes over time from 1990 to 2015, and in turn, the trend is used to approximate stock changes from 2016 to
2019. The Tier 2 method described previously will be applied to recalculate the 2016 to 2019 emissions in a future
Inventory.
Organic Soil Carbon Stock Changes
Annual C emissions from drained organic soils in Land Converted to Settlements are estimated using the Tier 2
method provided in IPCC (2006). The Tier 2 method assumes that organic soils are losing C at a rate similar to
croplands, and therefore uses the country-specific values for cropland (Ogle et al. 2003). To estimate C02
emissions from 1990 to 2015, the area of organic soils in Land Converted to Settlements is multiplied by the Tier 2
emission factor, which is 11.2 MT C per ha in cool temperate regions, 14.0 MT C per ha in warm temperate regions
and 14.3 MT C per ha in subtropical regions (See Annex 3.12 for more information). Similar to the mineral soil
organic C stocks changes, a linear extrapolation of the trend in the time series is applied to estimate the emissions
from 2016 to 2019 because NRI activity data are not available for these years to determine the area of Land
Converted to Settlements.
Uncertainty and Time-Series Consistency
The uncertainty analysis for C losses with Forest Land Converted to Settlements is conducted in the same way as
the uncertainty assessment for forest ecosystem C flux in the Forest Land Remaining Forest Land category. Sample
and model-based error are combined using simple error propagation methods provided by the IPCC (2006), i.e., by
taking the square root of the sum of the squares of the standard deviations of the uncertain quantities. For
additional details, see the Uncertainty Analysis in Annex 3.13. The uncertainty analysis for mineral soil organic C
stock changes and annual C emission estimates from drained organic soils in Land Converted to Settlements is
estimated using a Monte Carlo approach, which is also described in the Cropland Remaining Cropland section.
Uncertainty estimates are presented in Table 6-99 for each subsource (i.e., biomass C, dead wood, litter, soil
organic C in mineral soil and organic soils) and the method applied in the inventory analysis (i.e., Tier 2 and Tier 3).
Uncertainty estimates from the Tier 2 and 3 approaches are combined using the simple error propagation methods
provided by the IPCC (2006), i.e., as described in the previous paragraph. There are also additional uncertainties
propagated through the analysis associated with the data splicing methods applied to estimate soil organic C stock
changes from 2016 to 2019. The combined uncertainty for total C stocks in Land Converted to Settlements ranges
from 33 percent below to 34 percent above the 2019 stock change estimate of 79.2 MMT C02 Eq.
Table 6-99: Approach 2 Quantitative Uncertainty Estimates for Soil, Dead Organic Matter
and Biomass C Stock Changes occurring within Land Converted to Settlements (MMT CO2 Eq.
and Percent)
2019 Flux Estimate Uncertainty Range Relative to Flux Estimate3
Source	(MMT CP2 Eq.)	(MMTCQ2 Eq.)	(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Cropland Converted to Settlements
5.9
2.3
9.5
-61%
61%
Mineral Soil C Stocks
5.1
1.6
8.7
-69%
69%
Organic Soil C Stocks
0.8
0.1
1.4
-82%
82%
Forest Land Converted to Settlements
62.9
38.5
87.4
-39%
39%
Aboveground Biomass C Stocks
36.9
14.0
59.9
-62%
62%
Belowground Biomass C Stocks
7.2
2.7
11.7
-62%
62%
Dead Wood
6.7
3.5
10.9
-47%
62%
Land Use, Land-Use Change, and Forestry 6-145

-------
Litter
9.9
3.7
16.0
-62%
62%
Mineral Soil C Stocks
1.9
1.4
2.4
-27%
28%
Organic Soil C Stocks
0.3
0.1
0.5
-68%
69%
Grassland Converted to Settlements
11.3
6.6
15.9
-41%
41%
Mineral Soil C Stocks
10.4
5.8
15.0
-44%
44%
Organic Soil C Stocks
0.9
0.1
1.6
-85%
85%
Other Lands Converted to Settlements
(1.2)
(1.9)
(0.4)
-62%
62%
Mineral Soil C Stocks
(1.3)
(2.0)
(0.6)
-55%
55%
Organic Soil C Stocks
0.1
(0.1)
0.3
-160%
160%
Wetlands Converted to Settlements
0.4
0.1
0.9
-142%
142%
Mineral Soil C Stocks
0.1
+
0.1
-96%
96%
Organic Soil C Stocks
0.3
(0.2)
0.8
-172%
172%
Total: Land Converted to Settlements
79.2
52.8
105.9
-33%
34%
Aboveground Biomass C Stocks
36.9
14.0
59.9
-62%
62%
Belowground Biomass C Stocks
7.2
2.7
11.7
-62%
62%
Dead Wood
6.7
3.5
10.9
-47%
62%
Litter
9.9
3.7
16.0
-62%
62%
Mineral Soil C Stocks
16.2
10.3
22.0
-36%
36%
Organic Soil C Stocks
2.4
(6.1)
10.8
-356%
356%
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2019. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
QA/QC and Verification
Quality control measures included checking input data, model scripts, and results to ensure data are properly
handled throughout the inventory process. Inventory reporting forms and text are reviewed and revised as needed
to correct transcription errors. No errors were found in this Inventory.
Recalculations Discussion
There were no recalculations in this Inventory.
Planned Improvements
A planned improvement for the Land Converted to Settlements category is to develop an inventory of mineral soil
organic C stock changes in Alaska and losses of C from drained organic soils in federal lands. This includes C stock
changes for biomass, dead organic matter and soils. See Table 6-100 for the amount of managed land area in Land
Converted to Settlements that is not included in the Inventory due to these omissions. The managed area that is
not included in the Inventory ranges between 0 and about 600 thousand hectares depending on the year.
There are plans to improve classification of trees in settlements and to include transfer of biomass from forest land
to those areas in this category. There are also plans to extend the Inventory to included C losses associated with
drained organic soils in settlements occurring on federal lands
New land representation data will also be compiled, and the time series recalculated for the latter years in the
time series that are estimated using data splicing methods in this Inventory. These improvements will be made as
funding and resources are available to expand the inventory for this source category.
6-146 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 6-100: Area of Managed Land in Settlements Remaining Settlements that is not
included in the current Inventory (Thousand Hectares)

Area (Thousand Hectares)



LCS Area
LCS Area Not

LCS Managed Land
Included in
Included in
Year
Area (Section 6.1)
Inventory
Inventory
1990
2,861
2,861
0
1991
3,238
3,238
0
1992
3,592
3,592
0
1993
4,178
4,107
72
1994
4,777
4,630
147
1995
5,384
5,161
223
1996
5,927
5,658
269
1997
6,520
6,174
346
1998
7,065
6,650
416
1999
7,577
7,116
461
2000
8,095
7,568
528
2001
8,544
7,947
597
2002
8,886
8,284
602
2003
8,941
8,335
606
2004
8,957
8,345
612
2005
8,947
8,341
606
2006
8,959
8,352
607
2007
8,902
8,295
607
2008
8,722
8,111
610
2009
8,541
7,930
611
2010
8,335
7,725
611
2011
8,108
7,498
611
2012
7,918
7,298
620
2013
7,504
6,932
572
2014
7,087
6,586
501
2015
6,589
6,165
424
2016
ND
ND
ND
2017
ND
ND
ND
2018
ND
ND
ND
2019
ND
ND
ND
Note: NRI data are not available after 2015, and these years are designated as ND (No data).
6.12 Other Land Remaining Other Land (CRF
Category 4F1)
Land use is constantly occurring, and areas under a number of differing land-use types remain in their respective
land-use type each year, just as other land can remain as other land. While the magnitude of Other Land
Remaining Other Land is known (see Table 6-5), research is ongoing to track C pools in this land use. Until such
time that reliable and comprehensive estimates of C for Other Land Remaining Other Land can be produced, it is
not possible to estimate C02, CH4 or N20 fluxes on Other Land Remaining Other Land at this time.
Land Use, Land-Use Change, and Forestry 6-147

-------
6.13 Land Converted to Other Land (CRF
Category 4F2)
Land-use change is constantly occurring, and areas under a number of differing land-use types are converted to
other land each year, just as other land is converted to other uses. While the magnitude of these area changes is
known (see Table 6-5), research is ongoing to track C across Other Land Remaining Other Land and Land Converted
to Other Land. Until such time that reliable and comprehensive estimates of C across these land-use and land-use
change categories can be produced, it is not possible to separate C02, CH4 or N20 fluxes on Land Converted to
Other Land from fluxes on Other Land Remaining Other Land at this time.
6-148 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
7. Waste
Waste management and treatment activities are sources of greenhouse gas emissions (see Figure 7-1 and Figure
7-2). Landfills accounted for approximately 17.4 percent of total U.S. anthropogenic methane (CH4) emissions in
2019, the third largest contribution of any CH4 source in the United States. Additionally, wastewater treatment and
discharge, composting of organic waste, and stand-alone anaerobic digestion accounted for approximately 2.8
percent, 0.3 percent, and less than 0.1 percent of U.S. CH4 emissions, respectively. Nitrous oxide (N20) emissions
from the discharge of wastewater treatment effluents into aquatic environments were estimated, as were N20
emissions from the treatment process itself. Nitrous oxide emissions from composting were also estimated.
Together, these waste activities account for 6.2 percent of total U.S. N20 emissions. Nitrogen oxides (NOx), carbon
monoxide (CO), and non-CH4 volatile organic compounds (NMVOCs) are emitted by waste activities and are
addressed separately at the end of this chapter. A summary of greenhouse gas emissions from the Waste chapter
is presented in Table 7-1 and Table 7-2.
Figure 7-1: 2019 Waste Chapter Greenhouse Gas Sources
Wastewater Treatment
Anaerobic Digestion at
Biogas Facilities
Composting
Landfills
Waste as a Portion of All
Emissions
I Energy
¦ Agriculture
IPPU
Waste
114
0 10 20 30 40 50 60 70 80 90 100 110 120
MMT CO2 Eq.
Waste 7-1

-------
Figure 7-2: Trends in Waste Chapter Greenhouse Gas Sources
50 ¦ Anaerobic Digestion at Biogas Facilities
¦	Composting
¦	Wastewater T reatment
„ ¦ Landfills
Overall, in 2019, waste activities generated emissions of 163.7 MMT C02 Eq., or 2.5 percent of total U.S.
greenhouse gas emissions.
Table 7-1: Emissions from Waste (MMT CO2 Eq.)
Gas/Source
1990

2005

2015
2016
2017
2018
2019
ch4
197.1

153.4

132.5
129.2
130.5
132.9
135.3
Landfills
176.6

131.4

111.4
108.0
109.4
112.1
114.5
Wastewater Treatment
20.2

20.1

18.8
18.7
18.5
18.4
18.4
Composting
0.4

1.9

2.1
2.3
2.4
2.3
2.3
Anaerobic Digestion at Biogas









Facilities
+

0.1

0.2
0.2
0.2
0.2
0.2
n2o
19.0

24.6

27.3
27.9
28.6
28.2
28.4
Wastewater Treatment
18.7

23.0

25.4
25.9
26.4
26.1
26.4
Composting
0.3

1.7

1.9
2.0
2.2
2.0
2.0
Total
216.2

178.0

159.8
157.1
159.0
161.1
163.7
+ Does not exceed 0.05 MMT C02 Eq.
Note: Totals may not sum due to independent rounding.
Table 7-2: Emissions from Waste (kt)
Gas/Source
1990

2005

2014
2015
2016
2017
2018
ch4
7,885

6,135

5,301
5,166
5,218
5,317
5,414
Landfills
7,063

5,255

4,456
4,321
4,375
4,482
4,580
Wastewater Treatment
806

803

753
747
739
737
736
Composting
15

75

85
91
98
90
91
Anaerobic Digestion at Biogas









Facilities
1

2

7
7
7
7
7
n2o
64

83

91
94
96
94
95
Wastewater Treatment
63

77

85
87
89
88
88
Composting
1

6

6
7
7
7
7
Note: Totals may not sum due to independent rounding.
7-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Carbon dioxide (C02), CH4, and N20 emissions from the incineration of waste are accounted for in the Energy
sector rather than in the Waste sector because almost all incineration of municipal solid waste (MSW) in the
United States occurs at waste-to-energy facilities where useful energy is recovered. Similarly, the Energy sector
also includes an estimate of emissions from burning waste tires and hazardous industrial waste, because virtually
all of the combustion occurs in industrial and utility boilers that recover energy. The incineration of waste in the
United States in 2019 resulted in 11.8 MMT C02 Eq. emissions, more than half of which is attributable to the
combustion of plastics. For more details on emissions from the incineration of waste, see Section 7.5.
Each year, some emission and sink estimates in the Inventory are recalculated and revised with improved methods
and/or data. In general, recalculations are made to the U.S. greenhouse gas emission estimates either to
incorporate new methodologies or, most commonly, to update recent historical data. These improvements are
implemented consistently across the previous Inventory's time series (i.e., 1990 to 2018) to ensure that the trend
is accurate. Based on the availability of updated methodological guidance from 2019 Refinement (IPCC 2019), EPA
revised the methodologies used to estimate CH4 and N20 emissions from domestic wastewater treatment as well
as the methodology used to estimate CH4 emissions from industrial wastewater treatment. EPA also added N20
emission estimates from industrial wastewater treatment using a methodology based on the 2019 Refinement
(IPCC 2019). EPA also added emissions estimates from stand-alone anaerobic digestion to the Waste Chapter. For
more information on specific methodological updates, please see the Recalculations for each category, in this
chapter.
Due to lack of data availability, EPA is not able to estimate emissions associated with sludge generated from the
treatment of industrial wastewater or the amount of CH4 flared at composting sites. Emissions reported in the
Waste chapter for landfills, wastewater treatment, and stand-alone anaerobic digestion include those from all 50
states, including Hawaii and Alaska, as well as from the District of Columbia. Emissions from wastewater treatment
include most U.S. Territories except for Pacific Islands. Those emission are likely insignificant as those Pacific
Islands have no permanent population . Emissions for composting include all 50 states, including Hawaii and
Alaska, but not U.S. Territories. Composting emissions from U.S. Territories are assumed to be small. Similarly, we
are not aware of any anerobic digestion at biogas facilities in U.S. territories, but will review this on an ongoing
basis to include these emissions if they are occurring. See Annex 5 for more information on EPA's assessment of
the sources not included in this inventory.
Box 7-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to Greenhouse Gas Reporting Data
In following the United Nations Framework Convention on Climate Change (UNFCCC) requirement under Article
4.1 to develop and submit national greenhouse gas emission inventories, the emissions and removals presented
in this report and this chapter are organized by source and sink categories and calculated using internationally-
accepted methods provided by the Intergovernmental Panel on Climate Change (IPCC) in the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories (2006 IPCC Guidelines) and its supplements and
refinements. Additionally, the calculated emissions and removals in a given year for the United States are
presented in a common format in line with the UNFCCC reporting guidelines for the reporting of inventories
under this international agreement. The use of consistent methods to calculate emissions and removals by all
nations providing their inventories to the UNFCCC ensures that these reports are comparable. The presentation
of emissions and sinks provided in the Waste chapter do not preclude alternative examinations, but rather, this
chapter presents emissions and removals in a common format consistent with how countries are to report
Inventories under the UNFCCC. The report itself, and this chapter, follows this standardized format, and
provides an explanation of the application of methods used to calculate emissions and removals from waste
management and treatment activities.
EPA also collects greenhouse gas emissions data from individual facilities and suppliers of certain fossil fuels and
industrial gases through its Greenhouse Gas Reporting Program (GHGRP). The GHGRP applies to direct
greenhouse gas emitters, fossil fuel suppliers, industrial gas suppliers, and facilities that inject C02 underground
for sequestration or other reasons and requires reporting by sources or suppliers in 41 industrial categories.
Waste 7-3

-------
Annual reporting is at the facility level, except for certain suppliers of fossil fuels and industrial greenhouse
gases. In general, the threshold for reporting is 25,000 metric tons or more of C02 Eq. per year. See Annex 9
"Use of EPA Greenhouse Gas Reporting Program in Inventory" for more information.
Waste Data from EPA's Greenhouse Gas Reporting Program
EPA uses annual GHGRP facility-level data in the Landfills category to compile the national estimate of emissions
from Municipal Solid Waste (MSW) landfills (see section 7.1 of this chapter for more information). EPA uses
directly reported GHGRP data for net CH4 emissions from MSW landfills for the years 2010 to 2019 of the
Inventory. MSW landfills subject to the GHGRP began collecting data in 2010. These data are also used to
recalculate emissions from MSW landfills for the years 2005 to 2009 to ensure time-series consistency.
7.1 Landfills (CRF Source Category 5A1)
In the United States, solid waste is managed by landfilling, recovery through recycling or composting, and
combustion through waste-to-energy facilities. Disposing of solid waste in modern, managed landfills is the most
used waste management technique in the United States. More information on how solid waste data are collected
and managed in the United States is provided in Box 7-3. The municipal solid waste (MSW) and industrial waste
landfills referred to in this section are all modern landfills that must comply with a variety of regulations as
discussed in Box 7-2. Disposing of waste in illegal dumping sites is not considered to have occurred in years later
than 1980 and these sites are not considered to contribute to net emissions in this section for the timeframe of
1990 to the current Inventory year. MSW landfills, or sanitary landfills, are sites where MSW is managed to prevent
or minimize health, safety, and environmental impacts. Waste is deposited in different cells and covered daily with
soil; many have environmental monitoring systems to track performance, collect leachate, and collect landfill gas.
Industrial waste landfills are constructed in a similar way as MSW landfills, but are used to dispose of industrial
solid waste, such as RCRA Subtitle D wastes (e.g., non-hazardous industrial solid waste defined in Title 40 of the
Code of Federal Regulations or CFR in section 257.2), commercial solid wastes, or conditionally exempt small-
quantity generator wastes (EPA 2016a).
After being placed in a landfill, organic waste (such as paper, food scraps, and yard trimmings) is initially
decomposed by aerobic bacteria. After the oxygen has been depleted, the remaining waste is available for
consumption by anaerobic bacteria, which break down organic matter into substances such as cellulose, amino
acids, and sugars. These substances are further broken down through fermentation into gases and short-chain
organic compounds that form the substrates for the growth of methanogenic bacteria. These methane (CH4)
producing anaerobic bacteria convert the fermentation products into stabilized organic materials and biogas
consisting of approximately 50 percent biogenic carbon dioxide (C02) and 50 percent CH4, by volume. Landfill
biogas also contains trace amounts of non-methane organic compounds (NMOC) and volatile organic compounds
(VOC) that either result from decomposition byproducts or volatilization of biodegradable wastes (EPA 2008).
Box 7-2: Description of a Modern, Managed Landfill in the United States
Modern, managed landfills are well-engineered facilities that are located, designed, operated, and monitored to
ensure compliance with federal, state, and tribal regulations. A modern, managed landfill is EPA's interpretation
of the IPCC's terminology of a managed solid waste disposal site. Municipal solid waste (MSW) landfills must be
designed to protect the environment from contaminants which may be present in the solid waste stream.
Additionally, many new landfills collect and destroy landfill gas through flares or landfill gas-to-energy projects.
Requirements for affected MSW landfills may include:
•	Siting requirements to protect sensitive areas (e.g., airports, floodplains, wetlands, fault areas, seismic
impact zones, and unstable areas);
•	Design requirements for new landfills to ensure that Maximum Contaminant Levels (MCLs) will not be
7-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
exceeded in the uppermost aquifer (e.g., composite liners and leachate collection systems);
•	Leachate collection and removal systems;
•	Operating practices (e.g., daily and intermediate cover, receipt of regulated hazardous wastes, use of
landfill cover material, access options to prevent illegal dumping, use of a collection system to prevent
stormwater run-on/run-off, record-keeping);
•	Air monitoring requirements (explosive gases);
•	Groundwater monitoring requirements;
•	Closure and post-closure care requirements (e.g., final cover construction); and
•	Corrective action provisions.
Specific federal regulations that affected MSW landfills must comply with include the 40 CFR Part 258 (Subtitle
D of RCRA), or equivalent state regulations and the NSPS 40 CFR Part 60 Subpart WWW and XXX.1 Additionally,
state and tribal requirements may exist.
Methane and C02 are the primary constituents of landfill gas generation and emissions. However, the 2006IPCC
Guidelines set an international convention to not report biogenic C02 from activities in the Waste sector (IPCC
2006). Net carbon dioxide flux from carbon stock changes in landfills are estimated and reported under the Land
Use, Land-Use Change, and Forestry (LULUCF) sector (see Chapter 6 of this Inventory). Additionally, emissions of
NMOC and VOC are not estimated because they are emitted in trace amounts. Nitrous oxide (N20) emissions from
the disposal and application of sewage sludge on landfills are also not explicitly modeled as part of greenhouse gas
emissions from landfills. Nitrous oxide emissions from sewage sludge applied to landfills as a daily cover or for
disposal are expected to be relatively small because the microbial environment in an anaerobic landfill is not very
conducive to the nitrification and denitrification processes that result in N20 emissions. Furthermore, the 2006
IPCC Guidelines did not include a methodology for estimating N20 emissions from solid waste disposal sites
"because they are not significant." Therefore, only CH4 generation and emissions are estimated for landfills under
the Waste sector.
Methane generation and emissions from landfills are a function of several factors, including: (1) the total amount
and composition of waste-in-place, which is the total waste landfilled annually over the operational lifetime of a
landfill; (2) the characteristics of the landfill receiving waste (e.g., size, climate, cover material); (3) the amount of
CH4 that is recovered and either flared or used for energy purposes; and (4) the amount of CH4 oxidized as the
landfill gas - that is not collected by a gas collection system - passes through the cover material into the
atmosphere. Each landfill has unique characteristics, but all managed landfills employ similar operating practices,
including the application of a daily and intermediate cover material over the waste being disposed of in the landfill
to prevent odor and reduce risks to public health. Based on recent literature, the specific type of cover material
used can affect the rate of oxidation of landfill gas (RTI 2011). The most used cover materials are soil, clay, and
sand. Some states also permit the use of green waste, tarps, waste derived materials, sewage sludge or biosolids,
and contaminated soil as a daily cover. Methane production typically begins within the first year after the waste is
disposed of in a landfill and will continue for 10 to 60 years or longer as the degradable waste decomposes over
time.
In 2019, landfill CH4 emissions were approximately 114.5 MMT C02 Eq. (4,580 kt), representing the third largest
source of CH4 emissions in the United States, behind enteric fermentation and natural gas systems. Emissions from
MSW landfills accounted for approximately 87 percent of total landfill emissions (99.4 MMT C02 Eq.), while
industrial waste landfills accounted for the remainder (15.1 MMT C02 Eq). Estimates of operational MSW landfills
in the United States have ranged from 1,700 to 2,000 facilities (EPA 2019a; EPA 2019c; Waste Business Journal
[WBJ] 2016; WBJ 2010). More recently, the Environment Research & Education Foundation (EREF) conducted a
nationwide analysis of MSW management and counted 1,540 operational MSW landfills in 2013 (EREF 2016).
Conversely, there are approximately 3,200 MSW landfills in the United States that have been closed since 1980 (for
which a closure data is known, (EPA 2019a; WBJ 2010). While the number of active MSW landfills has decreased
1 For more information regarding federal MSW landfill regulations, see
.
Waste 7-5

-------
significantly over the past 20 years, from approximately 6,326 in 1990 to as few as 1,540 in 2013, the average
landfill size has increased (EREF 2016; EPA 2019b; BioCycle 2010). Regarding industrial waste landfills, the WBJ
database (WBJ 2016) includes approximately 1,200 landfills accepting industrial and/or construction and
demolition debris for 2016 (WBJ 2016). Only 169 facilities with industrial waste landfills met the reporting
threshold under Subpart TT (Industrial Waste Landfills) of EPA's Greenhouse Gas Reporting Program (GHGRP
codified in 40 CFR part 98), indicating that there may be several hundred industrial waste landfills that are not
required to report under EPA's GHGRP.
The annual amount of MSW generated and subsequently disposed in MSW landfills varies annually and depends
on several factors (e.g., the economy, consumer patterns, recycling and composting programs, inclusion in a
garbage collection service). The estimated annual quantity of waste placed in MSW landfills increased 10 percent
from approximately 205 MMT in 1990 to 226 MMT in 2000 and then decreased by 5.7 percent to 213 MMT in
2019 (see Annex 3.14, Table A-221). The total amount of MSW generated is expected to increase as the U.S.
population continues to grow, but the percentage of waste landfilled may decline due to increased recycling and
composting practices. Net CH4 emissions from MSW landfills have decreased since 1990 (see Table 7-3 and Table
7-4).
The estimated quantity of waste placed in industrial waste landfills (from the pulp and paper, and food processing
sectors) has remained relatively steady since 1990, ranging from 9.7 MMT in 1990 to 10.3 MMT in 2019 (see Annex
3.14, Table A-221). CH4 emissions from industrial waste landfills have also remained at similar levels recently,
ranging from 14.4 MMT C02 Eq. in 2005 to 15.1 MMT C02 Eq. in 2019 when accounting for both CH4 generation
and oxidation.
EPA's Landfill Methane Outreach Program (LMOP) collects information on landfill gas energy projects currently
operational or under construction throughout the United States. LMOP's project and technical database contains
certain information on the gas collection and control systems in place at landfills that are a part of the program,
which can include the amount of landfill gas collected and flared. In 2020, LMOP identified 9 new landfill gas-to-
energy (LFGE) projects (EPA 2020a) that began operation. While the amount of landfill gas collected and
combusted continues to increase, the rate of increase in collection and combustion no longer exceeds the rate of
additional CH4 generation from the amount of organic MSW landfilled as the U.S. population grows (EPA 2020b).
Landfill gas collection and control is not accounted for at industrial waste landfills in this chapter (see the
Methodology discussion for more information).
Table 7-3: ChU Emissions from Landfills (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
MSW CH4 Generation3
205.3
-
-
-
-
-
-
Industrial CH4 Generation
12.1
16.0
16.6
16.6
16.7
16.7
16.7
MSW CH4 Recovered
(17.9)
-
-
-
-
-
-
MSW CH4 Oxidized
(18.5)
-
-
-
-
-
-
Industrial CH4 Oxidized
(1.2)
(1.6)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
MSW net CH4 Emissions







(GHGRP)
-
117.0
96.4
93.1
94.4
97.0
99.4
Industrial CH4 Emissions'5
10.9
14.4
15.0
15.0
15.0
15.0
15.1
Total
176.6
131.4
111.4
108.0
109.4
112.1
114.5
7-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
a MSW CH4 generation is not estimated after 2005 because the directly reported net CH4 emissions from the GHGRP
are used.
b Methane recovery is not calculated for industrial landfills because this is not a common practice in the United
States. Only 1 landfill of 169 that report to Subpart TT (Industrial Waste Landfills) of the GHGRP had an active gas
collection and control system during the year 2019 (EPA 2020b).
Not applicable due to methodology change.
Note: Totals may not sum due to independent rounding. Parentheses indicate negative values. For years 1990 to
2004, the Inventory methodology for MSW landfills uses the first order decay methodology. A methodological
change occurs in year 2005. For years 2005 to 2019, directly reported net CH4 emissions from the GHGRP data plus a
scale-up factor are used to account for emissions from landfill facilities that are not subject to the GHGRP. These
data incorporate CH4 recovered and oxidized for MSW landfills. As such, CH4 generation and CH4 recovery are not
calculated separately. See the Time-Series Consistency section of this chapter for more information.
Table 7-4: ChU Emissions from Landfills (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
MSW CH4 Generation3
8,214
-
-
-
-
-
-
Industrial CH4 Generation
484
638
665
666
667
668
669
MSW CH4 Recovered
(851)
-
-
-
-
-
-
MSWCH4 Oxidized
(736)
-
-
-
-
-
-
Industrial CH4 Oxidized
(48)
(64)
(66)
(67)
(67)
(67)
(67)
MSW net CH4 Emissions







(GHGRP)
-
4,681
3,858
3,722
3,775
3,881
3,978
Industrial net CH4 Emissions'5
436
575
598
599
600
601
602
Total
7,063
5,255
4,456
4,321
4,375
4,482
4,580
a MSW CH4 generation is not estimated after 2005 because the directly reported net CH4 emissions from the GHGRP
are used.
b Methane recovery is not calculated for industrial landfills because this is not a common practice in the United
States. Only 1 landfill of 169 that report to Subpart TT (Industrial Waste Landfills) of the GHGRP had an active gas
collection and control system during the year 2019 (EPA 2020b).
Not applicable due to methodology change.
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values. For years 1990 to
2004, the Inventory methodology for MSW landfills uses the first order decay methodology. A methodological
change occurs in year 2005. For years 2005 to 2019, directly reported net CH4 emissions from the GHGRP data plus a
scale-up factor are used to account for emissions from landfill facilities that are not subject to the GHGRP. These
data incorporate CH4 recovered and oxidized for MSW landfills. As such, CH4 generation and CH4 recovery are not
calculated separately. See the Time-Series Consistency section of this chapter for more information.
Methodology
Methodology Applied for MSW Landfills
A combination of IPCC Tier 2 and 3 approaches (IPCC 2006) are used to calculate emissions from MSW Landfills.
Methane emissions from landfills are estimated using two primary methods. The first method uses the first order
decay (FOD) model as described by the 2006 IPCC Guidelines to estimate CH4 generation. The amount of CH4
recovered and combusted from MSW landfills is subtracted from the CH4 generation and is then adjusted with an
oxidation factor. The oxidation factor represents the amount of CH4 in a landfill that is oxidized to C02 as it passes
through the landfill cover (e.g., soil, clay, geomembrane). This method is presented below and is similar to
Equation HH-6 in 40 CFR Part 98.343 for MSW landfills, and Equation TT-6 in 40 CFR Part 98.463 for industrial
waste landfills.
CH4,msw= (GCH4 - Rn) * (1 - OX)
where,
Waste 7-7

-------
CH4,msw	= Net CH4 emissions from solid waste
Gch4,msw = CH4 generation from MSW landfills, using emission factors for DOC, k, MCF, F from IPCC
(2006) and other peer-reviewed sources
R	= CH4 recovered and combusted
Ox	= CH4 oxidized from MSW landfills before release to the atmosphere, using Ox values from
IPCC (2006) and other peer-reviewed or scientifically-validated literature (40 CFR Part
98)
The second method used to calculate CH4 emissions from landfills, also called the back-calculation method, is
based on directly measured amounts of recovered CH4 from the landfill gas and is expressed below and by
Equation HH-8 in 40 CFR Part 98.343. The two parts of the equation consider the portion of CH4 in the landfill gas
that is not collected by the landfill gas collection system, and the portion that is collected. First, the recovered CH4
is adjusted with the collection efficiency of the gas collection and control system and the fraction of hours the
recovery system operated in the calendar year. This quantity represents the amount of CH4 in the landfill gas that is
not captured by the collection system; this amount is then adjusted for oxidation. The second portion of the
equation adjusts the portion of CH4 in the collected landfill gas with the efficiency of the destruction device(s), and
the fraction of hours the destruction device(s) operated during the year.
CH4,solid Waste = [(7—7	r) x(l - OX) + R x (l - (DE X fDest))]
\CE x j rec '
where,
CH4,solid waste = Net CH4 emissions from solid waste
R	= Quantity of recovered CH4 from Equation HH-4 of EPA's GHGRP
CE	= Collection efficiency estimated at the landfill, considering system coverage,
operation, and cover system materials from Table HH-3 of EPA's GHGRP. If area
by soil cover type information is not available, the default value of 0.75 should
be used, (percent)
fREc	= fraction of hours the recovery system was operating (percent)
OX	= oxidation factor (percent)
DE	= destruction efficiency (percent)
fDest = fraction of hours the destruction device was operating (fraction)
The current Inventory uses both methods to estimate CH4 emissions across the time series within EPA's Waste
Model, as summarized in Figure 7-3 below. This chapter provides a summary of the methods, activity data, and
parameters used. Additional step-wise explanations to generate the net emissions are provided in Annex 3.14.
7-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 7-3: Methodologies Used Across the Time Series to Compile the U.S. Inventory of
Emission Estimates for MSW Landfills

Annex Steps 1-3
Annex Step 4
Annex Step 5
Annex Step 6
Method
US-specific first-order decay (FOD)
model
Back-casted EPA
GFIGRP reported
net methane
EPA GHGRP
reported net
methane emissions
EPA GHGRP
reported net
methane emissions


emissions



1990 - 2004

2005 - 2009
2010 - 2016
2017 - Present
Parameters
IPCC 2006 Emission
Factors:
•	DOC = 0.20
•	MCF = 1
•	DOC, = 0.5
•	OX = 0.10
•	DE = 0.99
Activity Data:
• National waste
generation data
multiplied by the
national disposal
factor
•	Back-casted
GFIGRP emissions
from 2010 to the
current reporting
year1-2
•	9% scale-up factor
applied to GFIGRP
emissions
•	Net GHGRP
emissions2
•	9% scale-up factor
applied to GHGRP
emissions
• Net GHGRP
emissions2
11% scale-up
factor applied to
GHGRP emissions
1	The back-casted emissions are calculated using directly reported net methane emissions for GHGRP reporting years
2010 to 2019 (the current reporting year). The back-casted emissions are subject to change in each Inventory based on
new reporting year reports and re-submitted greenhouse gas reports for previous years. This method is compatible with
the IPCC 2006 Guidelines because facilities reporting to the GHGRP either use the FOD method, or directly measured
methane recovery data with default emission factors either directly included in the IPCC 2006 Guidelines or scientifically
validated through peer review.
2	Emission factors used by facilities reporting to GHGRP Subpart HH are facility-specific defaults derived from peer-
reviewed literature and the IPCC 2006 Guidelines.
The Waste Model is a spreadsheet developed by the IPCC for purposes of estimating methane emissions from solid
waste disposal sites, adapted to the United States by the inclusion and usage of U.S.-specific parameters. The
Waste Model contains activity and waste generation information from both the MSW and Industrial landfill sectors
and estimates the amount of CH4 emissions from each sector for each year of the time series, using both methods.
Prior to the 1990 through 2015 Inventory, only the FOD method was used. Methodological changes were made to
the 1990 through 2015 Inventory to incorporate higher tier data (i.e., CH4 emissions as directly reported to EPA's
GHGRP), which cannot be directly applied to earlier years in the time series without significant bias. The technique
used to merge the directly reported GHGRP data with the previous methodology is described as the overlap
technique in the Time-Series Consistency chapter of the 2006 IPCC Guidelines. Additional details on the technique
used is included in the Time Series Consistency section of this chapter, Annex 3.14, and a technical memorandum
(RTI 2017).
A summary of the methodology used to generate the current 1990 through 2018 Inventory estimates for MSW
landfills is as follows and is also illustrated in Annex Figure A-18:
• 1940 through 1989: These years are included for historical waste disposal amounts. Estimates of the
annual quantity of waste landfilled for 1960 through 1988 were obtained from EPA's Anthropogenic
Methane Emissions in the United States, Estimates for 1990: Report to Congress (EPA 1993) and an
extensive landfill survey by the EPA's Office of Solid Waste in 1986 (EPA 1988). Although waste placed in
landfills in the 1940s and 1950s contributes very little to current CH4 generation, estimates for those years
were included in the FOD model for completeness in accounting for CH4 generation rates and are based
on the population in those years and the per capita rate for land disposal for the 1960s. For the Inventory
calculations, wastes landfilled prior to 1980 were broken into two groups: wastes disposed in managed,
anaerobic landfills (Methane Conversion Factor, MCF, of 1) and those disposed in uncategorized solid
waste disposal waste sites (MCF of 0.6) (IPCC 2006). Uncategorized sites represent those sites for which
limited information is known about the management practices. All calculations after 1980 assume waste
Waste 7-9

-------
is disposed in managed, anaerobic landfills. The FOD method was applied to estimate annual CH4
generation. Methane recovery amounts were then subtracted, and the result was then adjusted with a 10
percent oxidation factor to derive the net emissions estimates. A detailed explanation of the methods
used are presented in Annex 3.14 Step 1.
•	1990 through 2004: The Inventory time series begins in 1990. The FOD method is exclusively used for this
group of years. The national total of waste generated (based on state-specific landfill waste generation
data) and a national average disposal factor for 1989 through 2004 were obtained from the State of
Garbage (SOG) survey every two years (i.e., 2002, 2004 as published in BioCycle 2006). In-between years
were interpolated based on population growth. For years 1989 to 2000, directly reported total MSW
generation data were used; for other years, the estimated MSW generation (excluding construction and
demolition waste and inerts) were presented in the reports and used in the Inventory. The FOD method
was applied to estimate annual CH4 generation. Landfill-specific CH4 recovery amounts were then
subtracted from CH4 generation and the result was adjusted with a 10 percent oxidation factor to derive
the net emissions estimates. A detailed explanation of the methods used are presented in Annex 3.14
Steps 1 through 3.
•	2005 through 2009: Emissions for these years are estimated using net CH4 emissions that are reported by
landfill facilities under EPA's GHGRP. Because not all landfills in the United States are required to report to
EPA's GHGRP, a 9 percent scale-up factor is applied to the GHGRP emissions for completeness. Supporting
information, including details on the technique used to estimate emissions for 2005 to 2009, to develop
the scale-up factor, and to ensure time-series consistency by incorporating the directly reported GHGRP
emissions is presented in Annex 3.14 Step 4 and in RTI 2018a. A single oxidation factor is not applied to
the annual CH4 generated as is done for 1990 to 2004 because the GHGRP emissions data are used, which
already take oxidation into account. The GHGRP allows facilities to use varying oxidation factors (i.e., 0,
10, 25, or 35 percent) depending on their facility-specific calculated CH4 flux rate. The average oxidation
factor from the GHGRP facilities is 19.5 percent (from reporting years 2011 to 2017). A detailed
explanation of the methods used to develop the back-casted emissions and revised scale-up factor are
presented in Annex 3.14 Step 4.
•	2010 through 2016: Net CH4 emissions as directly reported to the GHGRP are used with a 9 percent scale-
up factor to account for landfills that are not required to report to the GHGRP. A combination of the FOD
method and the back-calculated CH4 emissions were used by the facilities reporting to the GHGRP.
Landfills reporting to the GHGRP without gas collection and control apply the FOD method, while most
landfills with landfill gas collection and control apply the back-calculation method. As noted above,
GHGRP facilities use a variety of oxidation factors. The average oxidation factor from the GHGRP facilities
is 19.5 percent. A detailed explanation of the methods used to develop the revised scale-up factor are
presented in Annex 3.14 Step 5.
•	2017 through 2019: The same methodology is applied as for 2010 through 2016 where a scale-up factor is
applied to account for landfills that are not required to report to the GHGRP. The scale-up factor was
revised for the current (1990 to 2019) Inventory to incorporate facilities that have stopped reporting to
the GHGRP, new additions to the 2020 LMOP Database (EPA 2020a), and corrections to the underlying
database of non-reporting landfills used to develop the 9 percent scale-up factor that were identified. For
2017 to 2019, a scale-up factor of 11 percent is applied annually to the GHGRP net reported CH4
emissions. A detailed explanation of the methods used to develop the revised scale-up factor are
presented in Annex 3.14 Step 6.
Supporting information, including details on the techniques used to ensure time-series consistency by
incorporating the directly reported GHGRP emissions is presented in the Time-Series Consistency section of this
chapter and in Annex 3.14.
7-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology Applied for Industrial Waste Landfills
Emissions from industrial waste landfills are estimated using a Tier 2 approach (IPCC 2006) and a tailored (country-
specific) IPCC waste model. Activity data used are industrial production data (ERG 2020) for two sectors (pulp and
paper manufacturing, and food and beverage manufacturing) to which country-specific default waste disposal
factors are applied (a separate disposal factor for each sector). The disposal factors, as described below, are based
on scientifically reviewed data, and are the same across the entire time series. The emission factors are based on
those recommended by the IPCC 2006 Guidelines and are the same across the entire time series.
The FOD equation from IPCC 2006 is used via the waste model to estimate methane emissions:
CH4JND = (ficHA ~ Yin-lRn) * (1— ox)
where,
CH4,solid waste = Net CH4 emissions from solid waste
GCH4,ind	= CH4 generation from industrial waste landfills, using production data multiplied by a
disposal factor and emission factors for DOC, k, MCF, F (IPCC 2006)
R	= CH4 recovered and combusted (no recovery is assumed for industrial waste landfills)
OX	= CH4 oxidized from industrial waste landfills before release to the atmosphere (using the
IPCC 2006 Guideline value for OX of 0.10)
The activity data used in the emission calculations are production data (e.g., the amount of meat, poultry,
vegetables processed; the amount of paper produced) versus disposal data. There are currently no facility-specific
data sources that track and report the amount and type of waste disposed of in the universe of industrial waste
landfills in the United States. EPA's GHGRP provides some insight into waste disposal in industrial waste landfills
but is not comprehensive. Data reported to the GHGRP on industrial waste landfills suggests that most of the
organic waste which would result in methane emissions is disposed at pulp and paper and food processing
facilities. Of the 168 facilities that reported to Subpart TT of the GHGRP in 2019, 92 (54 percent) are in the North
American Industrial Classification System (NAICS) for Pulp, Paper, and Wood Products (NAICS 321 and 322) and 12
(7 percent) are in Food Manufacturing (NAICS 311).
Based on this limited information, the Inventory methodology assumes most of the organic waste placed in
industrial waste landfills originates from the food processing (meat, vegetables, fruits) and pulp and paper sectors,
thus estimates of industrial landfill emissions focused on these two sectors. EPA validated this assumption through
an analysis of the Subpart TT of the GHGRP in the 2016 reporting year (RTI 2018b). The Subpart TT waste disposal
information for pulp and paper facilities correlates well with the activity data currently used to estimate Inventory
emissions; however, the waste disposal information in Subpart TT related to food and beverage facilities are
approximately an order of magnitude different than the Inventory disposal estimates for the entire time series.
EPA conducted a literature review in 2020 to investigate other sources of industrial food waste, which is briefly
described in the Planned Improvements section, and decided to maintain the currently used methodology for the
1990-2019 Inventory due to questions around data availability across the 1990 to 2019 time series, the
completeness and representativeness of other estimates and methodologies, and the level of effort required to
reproduce and/or merge estimates across the 1990 to 2019 time series.
The composition of waste disposed of in industrial waste landfills is expected to be more consistent in terms of
composition and quantity than that disposed of in MSW landfills. The amount of waste landfilled is assumed to be
a fraction of production that is held constant over the time series as explained in Annex 3.14.
Landfill CH4 recovery is not accounted for in industrial waste landfills. Data collected through EPA's GHGRP for
industrial waste landfills (Subpart TT) show that only one of the 168 facilities, or 1 percent of facilities, have active
gas collection systems (EPA 2020b). However, because EPA's GHGRP is not a national database and comprehensive
data regarding gas collection systems have not been published for industrial waste landfills, assumptions regarding
a percentage of landfill gas collection systems, or a total annual amount of landfill gas collected for the non-
reporting industrial waste landfills have not been made for the Inventory methodology.
Waste 7-11

-------
The amount of CH4 oxidized by the landfill cover at industrial waste landfills was assumed to be 10 percent of the
CH4 generated (IPCC 2006; Mancinelli and McKay 1985; Czepiel et al. 1996) for all years.
Box 7-3: Nationwide Municipal Solid Waste Data Sources
Municipal solid waste (MSW) generated in the United States can be managed through a variety of methods.
MSW that is not recycled, composted (and later land applied), combusted with energy recovery, or digested is
assumed to be landfilled. In addition to these management pathways, waste or excess food from the food
manufacturing and processing sector may be disposed through the sewerage network, used for animal feed,
donated for human consumption, and rendered or recycled into biofuels in the case of animal by-products, fats,
oils and greases.
There have been three main sources for nationwide solid waste management data in the United States that the
Inventory has used:
•	The BioCycle and Earth Engineering Center of Columbia University's SOG in America surveys [no longer
published];
•	The EPA's Advancing Sustainable Materials Management: Facts and Figures reports; and
•	The EREF's MSW Generation in the United States reports.
The SOG surveys and, most recently EREF, collected state-reported data on the amount of waste generated and
the amount of waste managed via different management options: landfilling, recycling, composting, and
combustion. These data sources used a 'bottom-up' method. The survey asked for actual tonnages instead of
percentages in each waste category (e.g., residential, commercial, industrial, construction and demolition,
organics, tires) for each waste management option. If such a breakdown was not available, the survey asked for
total tons landfilled. The data were adjusted for imports and exports across state lines so that the principles of
mass balance were adhered to for completeness, whereby the amount of waste managed did not exceed the
amount of waste generated. The SOG and EREF reports present survey data aggregated to the state level.
The EPA Advancing Sustainable Materials Management: Facts and Figures reports use a materials flow
methodology, commonly referred to as a 'top-down' methodology, which relies heavily on a mass balance
approach. Data are gathered from industry associations, key businesses, similar industry sources, and
government agencies (e.g., the Department of Commerce and the U.S. Census Bureau) and are used to estimate
tons of materials and products generated, recycled, combusted with energy recovery, other food management
pathways, or landfilled nationwide. The amount of MSW generated is estimated by estimating production and
then adjusting these values by addressing the imports and exports of produced materials to other countries.
MSW that is not recycled or composted is assumed to be combusted or landfilled, except for wasted food,
which uses a different methodology and includes nine different management pathways. The 2018 Facts and
Figures Report (U.S. EPA 2020) uses a methodology that expanded the number of management pathways to
include: animal feed; bio-based materials/biochemical processing (i.e., rendering); codigestion/anaerobic
digestion; composting/aerobic processes; combustion; donation; land application; landfill; and
sewer/wastewater treatment.
In this Inventory, emissions from solid waste management are presented separately by waste management
option, except for recycling of waste materials. Emissions from recycling are attributed to the stationary
combustion of fossil fuels that may be used to power on-site recycling machinery and are presented in the
stationary combustion chapter in the Energy sector, although the emissions estimates are not called out
separately. Emissions from solid waste disposal in landfills and the composting of solid waste materials are
presented in the Landfills and Composting sections in the Waste sector of this report. Emissions from anaerobic
digesters are presented in three different sections depending on the digester category. Emissions from on-farm
digesters are included in the Agriculture sector; emissions from digesters at wastewater treatment plants
emissions from stand-alone digesters are presented in separate sections in the Waste sector of this report. In
the United States, almost all incineration of MSW occurs at waste-to-energy (WTE) facilities or industrial
facilities where useful energy is recovered, and thus emissions from waste incineration are accounted for in the
Incineration chapter of the Energy sector of this report.
7-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Uncertainty and Time-Series Consistency
Several types of uncertainty are associated with the estimates of CH4 emissions from MSW and industrial waste
landfills when the FOD method is applied directly for 1990 to 2004 in the Waste Model and, to some extent, in the
GHGRP methodology. The approach used in the MSW emission estimates assumes that the CH4 generation
potential (U) and the rate of decay that produces CH4from MSW, as determined from several studies of CH4
recovery at MSW landfills, are representative of conditions at U.S. MSW landfills. When this top-down approach is
applied at the nationwide level, the uncertainties are assumed to be less than when applying this approach to
individual landfills and then aggregating the results to the national level. In other words, the FOD method as
applied in this Inventory is not facility-specific modeling and while this approach may over- or under-estimate CH4
generation at some landfills if used at the facility-level, the result is expected to balance out because it is being
applied nationwide.
There is a high degree of uncertainty associated with the FOD model, particularly when a homogeneous waste
composition and hypothetical decomposition rates are applied to heterogeneous landfills (IPCC 2006). There is less
uncertainty in EPA's GHGRP data because this methodology is facility-specific, uses directly measured CH4 recovery
data (when applicable), and allows for a variety of landfill gas collection efficiencies, destruction efficiencies,
and/or oxidation factors to be used.
Uncertainty also exists in the scale-up factors (both 9 percent and 11 percent) applied for years 2005 to 2016 and
2017 to 2019, respectively, and in the back-casted emissions estimates for 2005 to 2009. As detailed in RTI
(2018a), limited information is available for landfills that do not report to the GHGRP. RTI developed an initial list
of landfills that do not report to the GHGRP with the intent of quantifying the total waste-in-place for these
landfills that would add up to the scale-up factor. Input was provided by industry, LMOP, and additional EPA
support. However, many gaps existed in the initial development of this Non-Reporting Landfills Database.
Assumptions were made for hundreds of landfills to estimate their waste-in-place and the subsequent scale-up
factors. The waste-in-place estimated for each landfill is likely not 100 percent accurate and should be considered
a reasonable estimate. Additionally, a simple methodology was used to back-cast emissions for 2005 to 2009 using
the GHGRP-reported emissions from 2010 to 2019. This methodology does not factor in annual landfill to landfill
changes in landfill CH4 generation and recovery. Because of this, an uncertainty factor of 25 percent is applied to
the scale-up factor and years (emission estimates) the scale-up factor is applied to.
With regard to the time series and as stated in 2006 IPCC Guidelines Volume 1: Chapter 5 Time-Series Consistency
(IPCC 2006), "the time series is a central component of the greenhouse gas inventory because it provides
information on historical emissions trends and tracks the effects of strategies to reduce emissions at the national
level. All emissions in a time series should be estimated consistently, which means that as far as possible, the time
series should be calculated using the same method and data sources in all years" (IPCC 2006). This chapter,
however, recommends against back-casting emissions back to 1990 with a limited set of data and instead provides
guidance on techniques to splice, or join methodologies together. One of those techniques is referred to as the
overlap technique. The overlap technique is recommended when new data becomes available for multiple years.
This was the case with the GHGRP data for MSW landfills, where directly reported CH4 emissions data became
available for more than 1,200 MSW landfills beginning in 2010. The GHGRP emissions data had to be merged with
emissions from the FOD method to avoid a drastic change in emissions in 2010, when the datasets were combined.
EPA also had to consider that according to IPCC's good practice, efforts should be made to reduce uncertainty in
Inventory calculations and that, when compared to the GHGRP data, the FOD method presents greater
uncertainty.
In evaluating the best way to combine the two datasets, EPA considered either using the FOD method from 1990
to 2009, or using the FOD method for a portion of that time and back-casting the GHGRP emissions data to a year
where emissions from the two methodologies aligned. Plotting the back-casted GHGRP emissions against the
emissions estimates from the FOD method showed an alignment of the data in 2004 and later years which
facilitated the use of the overlap technique while also reducing uncertainty. A detailed explanation and a chart
showing the estimates across the time series considering the two method options is included in Annex 3.14. EPA
Waste 7-13

-------
ultimately decided to back-cast the GHGRP emissions from 2009 to 2005 only, to merge the datasets and adhere to
the IPCC Good Practice Guidance for ensuring time series consistency.
Aside from the uncertainty in estimating landfill CH4 generation, uncertainty also exists in the estimates of the
landfill gas oxidized at MSW landfills. Facilities directly reporting to EPA's GHGRP can use oxidation factors ranging
from 0 to 35 percent, depending on their facility-specific CH4 flux. As recommended by the 2006 IPCC Guidelines
for managed landfills, a 10 percent default oxidation factor is applied in the Inventory for both MSW landfills
(those not reporting to the GHGRP and for the years 1990 to 2004 when GHGRP data are not available) and
industrial waste landfills regardless of climate, the type of cover material, and/or presence of a gas collection
system.
Another significant source of uncertainty lies with the estimates of CH4 recovered by flaring and gas-to-energy
projects at MSW landfills that are sourced from the Inventory's CH4 recovery databases (used for years 1990 to
2004). Four CH4 recovery databases are used to estimate nationwide CH4 recovery for MSW landfills for 1990 to
2004; whereas directly reported CH4 emissions, which accounts for CH4 recovery, is used for facilities reporting to
the GHGRP for years 2005 to 2018. The GHGRP MSW landfills database was added as a fourth recovery database
starting with the 1990 through 2013 Inventory report (two years before the full GHGRP data set started being used
for net CH4 emissions for the Inventory). Relying on multiple databases for a complete picture introduces
uncertainty because the coverage and characteristics of each database differs, which increases the chance of
double counting avoided emissions. Additionally, the methodology and assumptions that go into each database
differ. For example, the flare database assumes the midpoint of each flare capacity at the time it is sold and
installed at a landfill; the flare may be achieving a higher capacity, in which case the flare database would
underestimate the amount of CH4 recovered.
The LFGE database was updated annually until 2015. The flare database was populated annually until 2015 by the
voluntary sharing of flare sales data by select vendors, which likely underestimated recovery for landfills not
included in the three other recovery databases used by the Inventory. The EIA database has not been updated
since 2006 and has, for the most part, been replaced by the GHGRP MSW landfills database. To avoid double
counting and to use the most relevant estimate of CH4 recovery for a given landfill, a hierarchical approach is used
among the four databases. GHGRP data and the EIA data are given precedence because facility data were directly
reported; the LFGE data are given second priority because CH4 recovery is estimated from facility-reported LFGE
system characteristics; and the flare data are given the lowest priority because this database contains minimal
information about the flare, no site-specific operating characteristics, and includes smaller landfills not included in
the other three databases (Bronstein et al. 2012). The coverage provided across the databases most likely
represents the complete universe of landfill CH4 gas recovery; however, the number of unique landfills between
the four databases does differ.
The 2006 IPCC Guidelines default value of 10 percent for uncertainty in recovery estimates was used for two of the
four recovery databases in the uncertainty analysis where metering of landfill gas was in place (for about 64
percent of the CH4 estimated to be recovered). This 10 percent uncertainty factor applies to the LFGE database; 12
percent to the EIA database; and 1 percent for the GHGRP MSW landfills dataset because of the supporting
information provided and rigorous verification process. For flaring without metered recovery data (the flare
database), a much higher uncertainty value of 50 percent is used. The compounding uncertainties associated with
the four databases in addition to the uncertainties associated with the FOD method and annual waste disposal
quantities leads to the large upper and lower bounds for MSW landfills presented in Table 7-5.
The lack of landfill-specific information regarding the number and type of industrial waste landfills in the United
States is a primary source of uncertainty with respect to the industrial waste generation and emission estimates.
The approach used here assumes that most of the organic waste disposed of in industrial waste landfills that
would result in CH4 emissions consists of waste from the pulp and paper and food processing sectors. However,
because waste generation and disposal data are not available in an existing data source for all U.S. industrial waste
landfills, a straight disposal factor is applied over the entire time series to the amount produced to determine the
amounts disposed. Industrial waste facilities reporting under EPA's GHGRP do report detailed waste stream
information, and these data have been used to improve, for example, the DOC value used in the Inventory
methodology for the pulp and paper sector. A 10 percent oxidation factor is also applied to CH4 generation
7-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
estimates for industrial waste landfills and carries the same amount of uncertainty as with the factor applied to
CH4 generation for MSW landfills.
The results of the 2006IPCC Guidelines Approach 2 quantitative uncertainty analysis are summarized in Table 7-5.
There is considerable uncertainty for the MSW landfills estimates due to the many data sources used, each with its
own uncertainty factor.
Table 7-5: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Landfills
(MMT CO2 Eq. and Percent)	


2019 Emission




Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


(MMT CO? Eq.)
(MMTCOz
Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Total Landfills
ch4
114.5
88.0
140.0
-22%
+23%
MSW
ch4
99.4
74.1
124.5
-25%
+25%
Industrial
ch4
15.1
10.4
18.9
-31%
+25%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Individual uncertainty factors are applied to activity data and emission factors in the Monte Carlo analysis.
QA/QC and Verification
General quality assurance/quality control (QA/QC) procedures were applied consistent with the U.S. Inventory
QA/QC plan, which is in accordance with Vol. 1, Chapter 6 of 2006 IPCC Guidelines (see Annex 8 for more details).
QA/QC checks are performed for the transcription of the published data set (e.g., EPA's GHGRP dataset) used to
populate the Inventory data set in terms of completeness and accuracy against the reference source. Additionally,
all datasets used for this category have been checked to ensure they are of appropriate quality and are
representative of U.S. conditions. The primary calculation spreadsheet is tailored from the 2006 IPCC Guidelines
waste model and has been verified previously using the original, peer-reviewed IPCC waste model. All model input
values and calculations were verified by secondary QA/QC review. Stakeholder engagements sessions in 2016 and
2017 were used to gather input on methodological improvements and facilitate an external expert review on the
methodology, activity data, and emission factors.
Category-specific checks include the following:
•	Evaluation of the secondary data sources used as inputs to the Inventory dataset to ensure they are
appropriately collected and are reliable;
•	Cross-checking the data (activity data and emissions estimates) with previous years to ensure the data are
reasonable, and that any significant variation can be explained through the activity data;
•	Conducting literature reviews to evaluate the appropriateness of country-specific emission factors (e.g.,
DOC values, precipitation zones with respect to the application of the k values) given findings from recent
peer-reviewed studies; and
•	Reviewing secondary datasets to ensure they are nationally complete and supplementing where
necessary (e.g., using a scale-up factor to account for emissions from landfills that do not report to EPA's
GHGRP).
A primary focus of the QA/QC checks in past Inventories was to ensure that CH4 recovery estimates were not
double-counted and that all LFGE projects and flares were included in the respective project databases. QA/QC
checks performed in the past for the recovery databases were not performed in this Inventory, because new data
were not added to the recovery databases in this Inventory year.
For the GHGRP data, EPA verifies annual facility-level reports through a multi-step process (e.g., combination of
electronic checks and manual reviews by staff) to identify potential errors and ensure that data submitted to EPA
Waste 7-15

-------
are accurate, complete, and consistent.2 Based on the results of the verification process, EPA follows up with
facilities to resolve mistakes that may have occurred. The post-submittals checks are consistent with several
general and category-specific QC procedures, including range checks, statistical checks, algorithm checks, and year-
to-year checks of reported data and emissions. For the MSW Landfills sector, under Subpart HH of the GHGRP,
MSW Landfills with gas collection are required to report emissions from their site using both a forward- (using a
first order decay model as a basis) and back-calculating (using parameters specific to the landfill itself, such as
measured recovery and collection efficiency of the landfill gas) methodology. Reporters can choose which of these
two methodologies they believe best represents the emissions at their landfill and are required to submit that
value as their total Subpart HH emissions. Facilities are generally not expected to switch between the two
equations each year, as the emissions calculated using each method can vary greatly and can have a significant
effect on emission trends for that landfill, and potentially the entire MSW Landfill sector under the GHGRP. Key
checks are in place to assure that emissions are trending in a sensible way year over year for each reporting
landfill.
For the current (1990 to 2019) Inventory, the scale-up factor was revised from 9 percent to 11 percent resulting
from additional QC checks performed on the underlying 2016 Non-Reporting Landfills Database used to develop
the 9 percent scale-up factor, the addition of the total waste-in-place for the 194 landfills no longer reporting to
Subpart HH, changes to the waste-in-place for some landfills in the 2020 LMOP Database, and the increase in
estimated annual waste disposed between 2016 and 2018 for all non-reporting landfills in the database. Overall,
the estimated waste-in-place for non-reporting landfills increased by approximately 274 million MT. The estimates
of waste-in-place for the non-reporting landfills should be considered best estimates based on available data from
the 2020 LMOP Database (EPA 2020a) and the 2016 WBJ Directory (WBJ 2016). No efforts were made in
developing the 2018 Non-Reporting Landfills Database to contact facilities to verify the information included in
either source database.
Additional QC checks on the 2016 Non-Reporting Landfills Database increased the total waste-in-place estimated
for 2016 by 38 million MT. Specifically, QC checks and corrections made to the underlying 2016 Non-Reporting
Landfills Database resulted in an increase of 38 million tons of waste-in-place resulting from a formula error that
under-estimated the waste-in-place for some landfills with a permitted end year after 2016, especially for those
landfills that had reported closure dates in 2030 or later. The year that the waste-in-place data were from in the
2017 LMOP Database, a primary source used, was not pulled into the 2016 Non-Reporting Landfills Database, thus
the methodology assumed that all waste-in-place data were from 2016. The waste-in-place year is now included in
the 2018 Non-Reporting Landfills Database, allowing for a more realistic waste-in-place value to be estimated by
landfill.
Several quality control checks were performed on the 2018 Non-Reporting Landfills Database used to calculate the
11 percent scale-up factor. Specific checks included a 10 percent check of the data carried over from the 2016 WBJ
Directory and the LMOP Databases, randomly checking formula calculations, comparing the 2017 and 2020 LMOP
Databases for changes in waste-in-place, and sorting the estimated waste-in-place column from largest to smallest
to identify errors in the larger landfills.
Recalculations Discussion
Revisions to the individual facility reports submitted to EPA's GHGRP can be made at any time and a portion of
facilities have revised their reports since 2010 for various reasons, resulting in changes to the total net CH4
emissions for MSW landfills. These recalculations increased net emissions for MSW landfills from 2005 to 2015 by
less than 0.5 percent when compared to the previous Inventory report. Each Inventory year, the back-casted
emissions for 2005 to 2009 will be recalculated using the most recently verified data from the GHGRP. Changes in
these data result in changes to the back-casted emissions. The impact of the revisions to the GHGRP Subpart HH
2 See .
7-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
annual greenhouse gas reports resubmitted for 2015 to 2018 slightly increased or decreased total Subpart HH
reported net emissions by +/- 0.3 percent in the years the Subpart HH data are applied (i.e., 2005 to 2019). No
Subpart HH reports were resubmitted for the 2010 to 2014 reporting years that resulted in net emission changes.
These changes resulted in changes to the net Inventory emissions ranging from -0.03 percent to +0.06 percent. A
change in net Subpart HH reported emissions results in the same percentage change in the Inventory emissions for
that year.
The scale-up factor was also reassessed as a planned improvement for the current (1990 to 2019) Inventory.
Results from this effort increased the scale-up factor from 9 percent to 11 percent. The scale-up factor increased
because of the inclusion of 194 GHGRP Subpart HH facilities that have off-ramped, a calculation error identified for
some non-reporting landfills when developing the 9 percent scale-up factor in 2016, and changes to the estimated
waste-in-place for all non-reporting landfills since 2016. The 9 percent scale-up factor is being retained and used
for 2005-2016 and the 11 percent is being used for 2017 to 2019.
Using the 11 percent scale-up factor, in addition to revisions to the previously submitted GHGRP reports between
2015 to 2018 ultimately increased net CH4 emissions by 1.6 percent in 2017 (1.7 MMT C02 Eq.) and 1.4 percent in
2018 (1.5 MMT C02 Eq.) compared to the previous (1990 to 2018) Inventory.
Planned Improvements
EPA has received recommendations from industry stakeholders regarding the DOC values and decay rates (k value)
required to be used in the GHGRP calculations based on recent trends in the composition of waste disposed in
MSW landfills. Stakeholders have suggested that newer, more up-to-date default values for both k and DOC in the
GHGRP could then be reflected in the 2005 and later years of the Inventory. In response, EPA is developing a
multivariate analysis using publicly available Subpart HH GHGRP data, solving for optimized DOC and k values
across the more than 1,100 landfills reporting to the program. The results of this analysis could help inform future
GHGRP rulemaking where changes could be made to the default DOC and k values contained within Subpart HH,
which could then be carried over to the Inventory emissions estimates for MSW landfills upon promulgation of any
revisions to 40 CFR part 98.
EPA is investigating the k values for the three climate types (dry, moderate, and wet) against new data and other
landfill gas models, and how they are applied to the percentage of the population assigned to these climate types.
EPA will also assess the uncertainty factor applied to these k values in the Waste Model.
With respect to the scale-up factor, EPA will periodically assess the impact to the waste-in-place and emissions
data from facilities that have resubmitted annual reports during any reporting years, are new reporting facilities,
and from facilities that have stopped reporting to the GHGRP to ensure national estimates are as complete as
possible. Facilities may stop reporting to the GHGRP when they meet the "off-ramp" provisions (reported less than
15,000 metric tons of C02 equivalent emissions for 3 consecutive years or less than 25,000 metric tons of C02
equivalent emissions for 5 consecutive years). As was the case with this Inventory, if warranted, EPA will revise the
scale-up factor to reflect newly acquired information to ensure completeness of the Inventory. The methodology
applied to develop the scale-up factor will be reviewed for the 1990 to 2020 inventory cycle to determine whether
the total waste-in-place for all landfills in the Non-Reporting Landfills Database should be factored into the scale-
up factor, or a subset of years (e.g., 25, 30, 50) depending on the duration of operation.
EPA began investigating the prevalence of food-related waste deposited into industrial waste landfills in 2020 and
will record the findings from this exercise in a memorandum. The resources identified with the most relevant data
for the Inventory include the EPA's 2020 Wasted Food Measurement Methodology Scoping Memo (EPA 2020c);
the Food Waste Reduction Alliance survey reports on the industrial food manufacturing sector conducted to date
for 2012 (BSR 2013), 2013 (BSR 2014), and 2015 (FWRA 2016); and one peer-reviewed journal article entitled,
Assessing U.S. food wastage and opportunities for reduction (Dou et al. 2016). EPA's wasted food measurement
methodology includes estimates for industrial food waste based on others' research estimates, but industrial food
waste estimates will not be incorporated into the EPA's Advancing Sustainable Materials reports because industrial
waste is beyond the scope of these reports. Dou et al. (2016) primarily used findings from the Food Waste
Reduction Alliance surveys, which received survey data from a handful of facilities and may not be representative
Waste 7-17

-------
of the entire U.S. food and beverage sector. EPA has decided to maintain the currently applied methodology to
estimate emissions from the industrial food and beverage sector for the current Inventory cycle.
Additionally, with the recent publication of the 2019 Refinement to the 2006IPCC Guidelines for National
Greenhouse Gas Inventories (IPCC 2019), EPA will begin to update applicable emission factors, methodologies, and
assumptions underlying emission estimates for landfills and make any applicable changes during the next (1990 to
2020) Inventory cycle per the 2019 Refinement.
Box 7-4: Overview of U.S. Solid Waste Management Trends
As shown in Figure 7-4 and Figure 7-5, landfilling of MSW is currently and has been the most common waste
management practice. A large portion of materials in the waste stream are recovered for recycling and
composting, which is becoming an increasingly prevalent trend throughout the country. Materials that are
composted and recycled would have previously been disposed in a landfill.
Figure 7-4: Management of Municipal Solid Waste in the United States, 2018
Management of MSW in the United States
MSW to WTE
12%
Other Food_
Management
6%
La ndfil led
50%
Composted
8%
Recycled
24%
Source: EPA(2020d)
Note: 2018 is the latest year of available data. Data taken from Table 35 of EPA 2020a. MSW to WTE is combustion with
energy recovery (WTE = waste to energy).
7-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Figure 7-5: MSW Management Trends from 1990 to 2018
Landfilling
	
		
160
140
120
100
SO
60
40	j**1	_	Energy Recovery
20	— — — — —
_ — — — ™" ""	Composting
Recycling
Combustion with
Other Food Management
0	(2018 only)
^ ^ ^ ¦$" ^ ^ ^ ^
Recycling — — Composting	-Other Food Management	Combustion with Energy Recovery — — — LandfilSng
Source: EPA (2020d). The EPA Advancing Sustainable Materials Management reports only present data for select years, thus several reports
were used in the compilation of this figure. All data were taken from Table 35 in EPA 2020d for 1990, 2000, 2015, 2017 and 2018. Data were
taken from Table 35 in EPA 2019c for 2010 and 2016. Data were taken from EPA 2018 for 2014. Data were taken from Table 35 of EPA
2016b for 2012 and 2013. Data were taken from Table 30 of EPA 2014 for 2008 and 2011. The reports with data available for years prior to
EPA 2012 can be provided upon request but are no longer on the EPA's Advancing Sustainable Materials Management web site <
https://www.epa.gov/facts-and-figures-about-materials-waste-and-recycling/advancing-sustainable-materials-management>.
Note: 2018 is the latest year of available data. Only one year of data (2018) is available for the 'Other Food Management' category.
Table 7-6 presents a typical composition of waste disposed of at a typical MSW landfill in the United States over
time. It is important to note that the actual composition of waste entering each landfill will vary from that
presented in Table 7-6. Due to China's recent ban on accepting certain kinds of solid waste by the end of 2017
(WTO 2017), inclusive of some paper and paperboard waste, plastic waste, and other miscellaneous inorganic
wastes, there has been a slight increase in the disposal of paper and paperboard and plastic wastes since 2017
(Table 7-6). EPA expects these numbers to continuing increasing until new markets for recycling of these goods
are identified.
Understanding how the waste composition changes over time, specifically for the degradable waste types (i.e.,
those types known to generate CH4 as they break down in a modern MSW landfill), is important for estimating
greenhouse gas emissions. Increased diversion of degradable materials so that they are not disposed of in
landfills reduces the CH4 generation potential and CH4 emissions from landfills. For certain degradable waste
types (i.e., paper and paperboard), the amounts discarded have decreased over time due to an increase in
waste diversion through recycling and composting (see Table 7-6 and Figure 7-6). As shown in Figure 7-6, the
diversion of food scraps has been consistently low since 1990 because most cities and counties do not practice
curbside collection of these materials, although the quantity has been slowly increasing in recent years. Neither
Table 7-6 nor Figure 7-6 reflect the frequency of backyard composting of yard trimmings and food waste
because this information is largely not collected nationwide and is hard to estimate.
Table 7-6: Materials Discarded in the Municipal Waste Stream by Waste Type from 1990
to 2018 (Percent)
Waste Type 1990

2005

2015 2016 2017 2018
Paper and
Paperboard 30.0%
Glass 6.0%
Metals 7.2%

24.7%
5.8%
7.9%

13.3% 12.7% 13.1% 11.8%
5.0% 4.9% 4.9% 5.2%
9.5% 9.8% 9.9% 9.5%
Waste 7-19

-------
Plastics
9.5%

16.4%

18.9%
18.9%
19.2%
18.5%
Rubber and Leather
3.2%

2.9%

3.3%
3.4%
3.5%
3.4%
Textiles
2.9%

5.3%

7.7%
8.0%
8.0%
7.7%
Wood
6.9%

7.5%

8.0%
8.8%
8.7%
8.3%
Other
1.4%

1.8%

2.2%
2.2%
2.2%
2.0%
Food Scraps
13.6%

18.5%

22.0%
22.1%
22.0%
24.1%
Yard Trimmings
17.6%

7.0%

7.8%
6.9%
6.2%
7.2%
Miscellaneous








Inorganic Wastes
1.7%

2.2%

2.3%
2.3%
2.3%
2.3%
Source: EPA (2020d)
Figure 7-6: Percent of Degradable Materials Diverted from Landfills from 1990 to 2018
(Percent)
90%
ro
Ł 80%
¦a
.
Note: The data shown in this chart are for recycling of paper and paperboard, composting of food scraps and yard trimmings, and
alternative management pathways for the Other Food Management (non-disposal) category. The Other Food Management (non-disposal)
category is a new addition and only one year of data are available for 2018 (28% of the food waste generated was beneficially reused or
managed using a method that was not landfilling, recycling, or composting).
7.2 Wastewater Treatment and Discharge (CRF
Source Category 5D)
Wastewater treatment and discharge processes are sources of anthropogenic methane (CH4) and nitrous oxide
(N20) emissions. Wastewater from domestic and industrial sources is treated to remove soluble organic matter,
7-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
suspended solids, nutrients, pathogenic organisms, and chemical contaminants.3 Treatment of domestic
wastewater may either occur on site, most commonly through septic systems, or off site at centralized treatment
systems, most commonly at publicly owned treatment works (POTWs). In the United States, approximately 18
percent of domestic wastewater is treated in septic systems or other on-site systems, while the rest is collected
and treated centrally (U.S. Census Bureau 2017). Treatment of industrial wastewater may occur at the industrial
plant using package or specially designed treatment plants, or be collected and transferred off site for co-
treatment with domestic wastewater in centralized treatment systems.
Centralized Treatment. Centralized wastewater treatment systems use sewer systems to collect and transport
wastewater to the treatment plant. Sewer collection systems provide an environment conducive to the formation
of CH4, which can be substantial depending on the configuration and operation of the collection system (Guisasola
et al. 2008). Recent research has shown that at least a portion of CH4 formed within the collection system enters
the centralized system where it contributes to CH4 emissions from the treatment system (Foley et al. 2015).
The treatment plant may include a variety of processes, ranging from physical separation of material that readily
settles out (typically referred to as primary treatment), to treatment operations that use biological processes to
convert and remove contaminants (typically referred to as secondary treatment), to advanced treatment for
removal of targeted pollutants, such as nutrients (typically referred to as tertiary treatment). Not all wastewater
treatment plants conduct primary treatment prior to secondary treatment, and not all plants conduct advanced or
tertiary treatment.
Soluble organic matter is generally removed using biological processes in which microorganisms consume the
organic matter for maintenance and growth. Microorganisms can biodegrade soluble organic material in
wastewater under aerobic or anaerobic conditions, where the latter condition produces CH4. The resulting biomass
(sludge) is removed from the effluent prior to discharge to the receiving stream and may be further biodegraded
under aerobic or anaerobic conditions, such as anaerobic sludge digestion. Sludge can be produced from both
primary and secondary treatment operations. Some wastewater may also be treated using constructed (or semi-
natural) wetland systems, though it is much less common in the United States and represents a relatively small
portion of wastewater treated centrally (<0.1 percent) (ERG 2016). Constructed wetlands are a coupled anaerobic-
aerobic system and may be used as the primary method of wastewater treatment, or are more commonly used as
a final treatment step following settling and biological treatment. Constructed wetlands develop natural processes
that involve vegetation, soil, and associated microbial assemblages to trap and treat incoming contaminants (IPCC
2014). Constructed wetlands do not produce secondary sludge (sewage sludge).
The generation of N20 may also result from the treatment of wastewater during both nitrification and
denitrification of the nitrogen (N) present, usually in the form of urea, proteins, and ammonia. Ammonia N is
converted to nitrate (N03) through the aerobic process of nitrification. Denitrification occurs under
anoxic/anaerobic conditions, whereby anaerobic or facultative organisms reduce oxidized forms of nitrogen (e.g.,
nitrite, nitrate) in the absence of free oxygen to produce nitrogen gas (N2). Nitrous oxide is generated as a by-
product of nitrification, or as an intermediate product of denitrification. No matter where N20 is formed it is
typically stripped (i.e., transferred from the liquid stream to the air) in aerated parts of the treatment process.
Stripping also occurs in non-aerated zones at rates lower than in aerated zones.
On-site Treatment. The vast majority of on-site systems in the United States are septic systems composed of a
septic tank, generally buried in the ground, and a soil dispersion system. Solids and dense materials contained in
the incoming wastewater (influent) settle in the septic tank as sludge. Floatable material (scum) is also retained in
the tank. The sludge that settles on the bottom of the tank undergoes anaerobic digestion. Partially treated water
is discharged in the soil dispersal system. The solid fraction accumulates and remains in the tank for several years,
during which time it degrades anaerobically. The gas produced from anaerobic sludge digestion (mainly CH4 and
biogenic C02) rises to the liquid surface and is typically released through vents. The gas produced in the effluent
dispersal system (mainly N20 and biogenic C02) is released through the soil.
3 Throughout the Inventory, emissions from domestic wastewater also include any commercial and industrial wastewater
collected and co-treated with domestic wastewater.
Waste 7-21

-------
Discharge. Dissolved CH4 and N20 that is present in wastewater discharges to aquatic environments has the
potential to be released (Short et al. 2014; Short et al. 2017), and the addition of organic matter or nitrogen from
wastewater discharges is generally expected to increase CH4 and N20 emissions from these environments. Where
organic matter is released to slow-moving aquatic systems, such as lakes, estuaries, and reservoirs, CH4 emissions
are expected to be higher. Similarly, in the case of discharge to nutrient-impacted or hypoxic waters, N20
emissions can be significantly higher.
The principal factor in determining the CH4 generation potential of wastewater is the amount of degradable
organic material in the wastewater. Common parameters used to measure the organic component of the
wastewater are the biochemical oxygen demand (BOD) and chemical oxygen demand (COD). Under the same
conditions, wastewater with higher COD (or BOD) concentrations will generally yield more CH4 than wastewater
with lower COD (or BOD) concentrations. BOD represents the amount of oxygen that would be required to
completely consume the organic matter contained in the wastewater through aerobic decomposition processes,
while COD measures the total material available for chemical oxidation (both biodegradable and non-
biodegradable). The BOD value is most commonly expressed in milligrams of oxygen consumed per liter of sample
during 5 days of incubation at 20°C, or BOD5. Throughout the rest of this chapter, the term "BOD" refers to BOD5.
Because BOD is an aerobic parameter, it is preferable to use COD to estimate CH4 production, since CH4 is
produced only in anaerobic conditions. Where present, biogas recovery and flaring operations reduce the amount
of CH4 generated that is actually emitted. The principal factor in determining the N20 generation potential of
wastewater is the amount of N in the wastewater. The variability of N in the influent to the treatment system, as
well as the operating conditions of the treatment system itself, also impact the N20 generation potential.
In 2019, CH4 emissions from domestic wastewater treatment and discharge were estimated to be 10.3 MMT C02
Eq. (410 kt CH4) and 1.8 MMT C02 Eq. (72 kt CH4), respectively. Emissions remained fairly steady from 1990
through 2002 but have decreased since that time due to decreasing percentages of wastewater being treated in
anaerobic systems, generally including reduced use of on-site septic systems and central anaerobic treatment
systems (EPA 1992,1996, 2000, and 2004a; U.S. Census Bureau 2017). In 2019, CH4 emissions from industrial
wastewater treatment and discharge were estimated to be 6.4 MMT C02 Eq. (254 kt CH4). Industrial emission
sources have generally increased across the time series through 1994 and then fluctuated up and down with
production changes associated with the treatment of wastewater from the pulp and paper manufacturing, meat
and poultry processing, fruit and vegetable processing, starch-based ethanol production, petroleum refining, and
brewery industries. Industrial wastewater emissions have seen an uptick since 2017. Table 7-7 and Table 7-8
provide CH4 emission estimates from domestic and industrial wastewater treatment.
With respect to N20, emissions from domestic wastewater treatment and discharge in 2019 were estimated to be
20.5 MMT C02 Eq. (69 kt N20) and 5.3 MMT C02 Eq. (18 kt N20), respectively. Total N20 emissions from domestic
wastewater were estimated to be 25.8 MMT C02 Eq. (87 kt N20). Nitrous oxide emissions from wastewater
treatment processes gradually increased across the time series as a result of increasing U.S. population and protein
consumption. In 2019, N20 emissions from industrial wastewater treatment were estimated to be 0.6 MMT C02
Eq. (2 kt N20). Industrial emission sources have gradually increased across the time series with production changes
associated with the treatment of wastewater from the pulp and paper manufacturing, meat and poultry
processing, petroleum refining, and brewery industries. Table 7-7 and Table 7-8 provide N20 emission estimates
from domestic wastewater treatment.
Table 7-7: ChU and N2O Emissions from Domestic and Industrial Wastewater Treatment
(MMT COz Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
ch4
20.2
20.1
18.8
18.7
18.5
18.4
18.4
Domestic Treatment
13.5
13.0
11.2
10.9
10.5
10.4
10.3
Domestic Effluent
1.2
1.2
1.8
1.8
1.8
1.8
1.8
Industrial3 Treatment
4.9
5.4
5.4
5.6
5.7
5.8
5.9
Industrial3 Effluent
0.5
0.5
0.4
0.4
0.4
0.4
0.4
n2o
18.7
23.0
25.4
25.9
26.4
26.1
26.4
7-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Domestic T reatment
13.6
17.7
19.5
19.9
20.4
20.3
20.6
Domestic Effluent
4.7
4.8
5.4
5.4
5.4
5.3
5.3
Industrial15 Treatment
0.4
0.4
0.5
0.5
0.5
0.5
0.5
Industrial15 Effluent
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Total
38.9
43.1
44.2
44.6
44.9
44.5
44.8
Note: Totals may not sum due to independent rounding.
a Industrial activity for CH4 includes the pulp and paper manufacturing, meat and poultry
processing, fruit and vegetable processing, starch-based ethanol production, petroleum refining,
and breweries industries.
b Industrial activity for N20 includes the pulp and paper manufacturing, meat and poultry
processing, starch-based ethanol production, and petroleum refining.
Table 7-8: ChU and N2O Emissions from Domestic and Industrial Wastewater Treatment (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
ch4
806
803
753
747
739
737
736
Domestic Treatment
540
518
447
434
421
416
410
Domestic Effluent
49
49
71
72
72
72
72
Industrial3 Treatment
196
215
217
223
228
232
236
Industrial3 Effluent
21
19
18
18
18
18
18
n2o
63
77
85
87
89
88
88
Domestic Treatment
46
59
65
67
69
68
69
Domestic Effluent
16
16
18
18
18
18
18
Industrial15 Treatment
1
1
2
2
2
2
2
Industrial15 Effluent
+
+
+
+
+
+
+
+ Does not exceed 0.5 kt.
Note: Totals may not sum due to independent rounding.
a Industrial activity for CH4 includes the pulp and paper manufacturing, meat and poultry processing,
fruit and vegetable processing, starch-based ethanol production, petroleum refining, and breweries
industries.
b Industrial activity for N20 includes the pulp and paper manufacturing, meat and poultry
processing, starch-based ethanol production, and petroleum refining.
Methodology
The methodologies presented in IPCC (2019) form the basis of the CH4 and N20 emission estimates for both
domestic and industrial wastewater treatment and discharge. Domestic wastewater treatment follows the IPCC
Tier 1 methodology, while domestic wastewater discharge follows IPCC Tier 2 discharge methodology and emission
factors. Industrial wastewater treatment and discharge follow IPCC Tier 1 methodologies.
IPCC (2019) updates, supplements, and elaborates the 2006 IPCC Guidelines where gaps or out-of-date science
have been identified. EPA used these methodologies to improve completeness and include sources of greenhouse
gas emissions that have not been previously estimated, such as N20 emissions from industrial wastewater
treatment, and to improve emission estimates for other sources, such as emissions from wastewater discharge and
centralized wastewater treatment.
Domestic Wastewater CH4 Emission Estimates
Domestic wastewater CH4 emissions originate from both septic systems and from centralized treatment systems.
Within these centralized systems, CH4 emissions can arise from aerobic systems that liberate dissolved CH4 that
formed within the collection system or that are designed to have periods of anaerobic activity (e.g., constructed
wetlands and facultative lagoons), anaerobic systems (anaerobic lagoons and anaerobic reactors), and from
Waste 7-23

-------
anaerobic sludge digesters when the captured biogas is not completely combusted. Emissions will also result from
the discharge of treated effluent from centralized wastewater plants to waterbodies where carbon accumulates in
sediments (typically slow-moving systems, such as lakes, reservoirs, and estuaries). The systems with emissions
estimates are:
•	Septic systems (A);
•	Centralized treatment aerobic systems (B), including aerobic systems (other than constructed wetlands)
(Bl), constructed wetlands only (B2), and constructed wetlands used as tertiary treatment (B3);
•	Centralized anaerobic systems (C);
•	Anaerobic sludge digesters (D); and
•	Centralized wastewater treatment effluent (E).
Methodological equations for each of these systems are presented in the subsequent subsections; total domestic
CH4 emissions are estimated as follows:
Total Domestic CH4 Emissions from Wastewater Treatment and Discharge (kt) = A+ B + C + D + E
Table 7-9 presents domestic wastewater CH4 emissions for both septic and centralized systems, including
anaerobic sludge digesters and emissions from centralized wastewater treatment effluent, in 2019.
Table 7-9: Domestic Wastewater ChU Emissions from Septic and Centralized Systems (2019,
kt, MMT CO2 Eq. and Percent)

CH4 Emissions (kt)
CH4 Emissions
(MMTCOz Eq.)
% of Domestic
Wastewater CH4
Septic Systems (A)
232
5.8
48.1
Centrally-Treated Aerobic Systems (B)
36
0.9
7.5
Centrally-Treated Anaerobic Systems (C)
134
3.3
27.7
Anaerobic Sludge Digesters (D)
8.1
0.2
1.7
Centrally-Treated Wastewater Effluent (E)
72
1.8
15.0
Total
482
12.1
100
Emissions from Septic Systems:
Methane emissions from septic systems were estimated by multiplying the U.S. population by the percent of
wastewater treated in septic systems (about 18 percent in 2019; U.S. Census Bureau 2017) and an emission factor
and then converting the result to kt/year.
U.S. population data were taken from the U.S. Census Bureau International Database (U.S. Census Bureau 2020)
and include the populations of the United States, American Samoa, Guam, Northern Mariana Islands, Puerto Rico,
and the U.S. Virgin Islands. Table 7-12 presents U.S. population for 1990 through 2019. The fraction of the U.S.
population using septic systems or centralized treatment systems is based on data from the American Housing
Surveys (U.S. Census Bureau 2017). Methane emissions for septic systems are estimated as follows:
Emissions from Septic Systems (U.S. Specific) = A
= USpop X (Tseptic) X (EFseptic) X 1/109 X 365.25
Table 7-10: Variables and Data Sources for ChU Emissions from Septic Systems
Variable
Variable Description
Units
Inventory Years: Source of
Value
USpop
U.S. population3
Persons
1990-2019: U.S. Census
Bureau (2020)
Tseptic
Percent treated in septic systems3
%
Odd years from 1989 through
2017: U.S. Census Bureau
(2017)
7-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------



Inventory Years: Source of
Variable
Variable Description
Units
Value



Data for intervening years



obtained by linear



interpolation



2018 and 2019: Forecasted



from the rest of the time



series
EFseptic
Methane emission factor - septic systems
(10.7)
g CH4/capita/day
1990-2019: Leverenz et al.
(2010)
1/109
Conversion factor
g to kt
Standard conversion
365.25
Conversion factor
Days in a year
Standard conversion
a Value of activity data varies over the Inventory time series.
Emissions from Centrally Treated Aerobic and Anaerobic Systems:
Methane emissions from POTWs depend on the total organics in wastewater. Table 7-12 presents total BOD5
produced (also referred to as the total organically degradable material in wastewater or TOW) for 1990 through
2019. The BOD5 production rate was determined using BOD generation rates per capita weighted average both
with and without kitchen scraps as well as an estimated percent of housing units that utilize kitchen garbage
disposals. Households with garbage disposals (with kitchen scraps or ground up food scraps) typically have
wastewater with higher BOD than households without garbage disposals due to increased organic matter
contributions (ERG 2018a). The equations are as follows:
Total wastewater BODs produced per capita (U.S. Specific (ERG 2018a), kg/capita/day)
BODgenrate — BODwithoutscrap X (1 - %disposal) + BODwith scraps x (%disposal)
Total organically degradable material in domestic wastewater (I PCC 2019 (Eq. 6.3), Gg/year)
TOW = USpopX BODgenrate X 365.25
Table 7-11: Variables and Data Sources for Organics in Domestic Wastewater
Variable
Variable Description
Units
Inventory Years: Source of
Value
BODgen rate
Total wastewater BOD produced per
capita
kg/capita/day
1990-2019: Calculated
BODwithout scrap
Wastewater BOD produced per capita
without kitchen scraps3
kg/capita/day
1990-2003: Metcalf & Eddy
(2003)
2004-2013: Linear
interpolation
2014-2019: Metcalf & Eddy
(2014)
BODwjth scraps
Wastewater BOD produced per capita
with kitchen scraps3
kg/capita/day
% disposal
Percent of housing units with kitchen
disposal3
%
1990-2013: U.S. Census
Bureau (2013)
2014-2019: Forecasted from
the rest of the time series
TOW
Total wastewater BOD Produced per
Capita3
Gg BOD/year
1990-2019: Calculated, ERG
(2018a)
USpop
U.S. population3
Persons
1990-2019: U.S. Census
Bureau (2020)
365.25
Conversion factor
Days in a year
Standard conversion
a Value of activity data varies over the Inventory time series.
Waste 7-25

-------
Table 7-12: U.S. Population (Millions) and Domestic Wastewater BODs Produced (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Population
253
300
325
327
329
331
334
TOW
8,131
j 9,624
9,736
9,820
9,896
9,971
10,079
Sources: U.S. Census Bureau (2020); ERG (2018a).
Methane emissions from POTWs were estimated by multiplying the total organics in centrally treated wastewater
(total BOD5) produced per capita in the United States by the percent of wastewater treated centrally, or percent
collected (about 82 percent in 2019), the correction factor for additional industrial BOD discharged to the sewer
system, the relative percentage of wastewater treated by aerobic systems (other than constructed wetlands),
constructed wetlands only, and anaerobic systems, and the emission factor4 for aerobic systems, constructed
wetlands only, and anaerobic systems. Methane emissions from constructed wetlands used as tertiary treatment
were estimated by multiplying the flow from treatment to constructed wetlands, wastewater BOD concentration
entering tertiary treatment, constructed wetlands emission factor, and then converting to kt/year.
In the United States, the removal of sludge5 from wastewater reduces the biochemical oxygen demand of the
wastewater that undergoes aerobic treatment. The amount of this reduction (S) is estimated using the default IPCC
methodology (IPCC 2019) and multiplying the amount of sludge removed from wastewater treatment in the
United States by the default factors in IPCC (2019) to estimate the amount of BOD removed based on whether the
treatment system has primary treatment with no anaerobic sludge digestion (assumed to be zero by expert
judgment), primary treatment with anaerobic sludge digestion, or secondary treatment without primary
treatment. The organic component removed from anaerobic wastewater treatment and the amount of CH4
recovered or flared from both aerobic and anaerobic wastewater treatment were set equal to the IPCC default of
zero.
The methodological equations for CH4 emissions from aerobic and anaerobic systems are:
Emissions from Centrally Treated Aerobic Systems (other than Constructed Wetlands) (Bl) + Emissions
from Centrally Treated Aerobic Systems (Constructed Wetlands Only) (B2) + Emissions from Centrally
Treated Aerobic Systems (Constructed Wetlands used as Tertiary Treatment) (B3) = B
where,
Total organics in centralized wastewater treatment (I PCC 2019 (Eq. 6.3A), Gg BOD/year)
TOWcentralized= TOW X Tcentralized X Icollected
Table 7-13: Variables and Data Sources for Organics in Centralized Domestic Wastewater
Variable
Variable Description
Units
Inventory Years: Source of Value
Centrally Treated Organics (Gg BOD/year)
TOWcentrauzed
Total organics in centralized
wastewater treatment
Gg BOD/year
1990-2019: Calculated
TOW
Total wastewater BOD Produced per
Capita3
Gg BOD/year
1990-2019: Calculated, ERG (2018a)
Tcentralized
Percent collected3
%
1990-2017: U.S. Census Bureau (2017)
Data for intervening years obtained
by linear interpolation
4	Emission factors are calculated by multiplying the maximum CH4-producing capacity of domestic wastewater (B0, 0.6 kg
CH4/kg BOD) and the appropriate methane correction factors (MCF) for aerobic (0.03) and anaerobic (0.8) systems (IPCC 2019)
and constructed wetlands (0.4) (IPCC 2014).
5	Throughout this document, the term "sludge" refers to the solids separated during the treatment of municipal wastewater.
The definition includes domestic septage. "Biosolids" refers to treated sewage sludge that meets the EPA pollutant and
pathogen requirements for land application and surface disposal.
7-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------



2018-2019: Forecasted from the rest
of the time series
IcOLLECTED
Correction factor for additional
industrial BOD discharged (1.25)
No units
1990-2019: IPCC (2019)
a Value of this activity data varies over the time series.
Organic component removed from aerobic wastewater treatment (IPCC 2019 (Eq. 6.3B), Gg/year)
Saerobic — Smass x [(% aerobic w/primary x Krem,aer_prim) + (% aerobic w/out primary x Krem,aer_noprim) +
(%aerobic+digestion x K rem,aer _digest)] X 1000
Emissions from Centrally Treated Aerobic Systems (other than Constructed Wetlands) (IPCC 2019 (Eq. 6.1),
kt CH4/year) = B1
= [(TOWcENTRALIZEd) X (% aerobiCOTCw) - Saerobic] X EFaerobic - Raerobic
Table 7-14: Variables and Data Sources for ChU Emissions from Centrally Treated Aerobic
Systems (Other than Constructed Wetlands)
Variable Variable Description
Units
Inventory Years: Source of Value
Emissions from Centrally Treated Aerobic Systems (Other than Constructed Wetlands) (kt CH4/year)
Saerobic
Organic component removed from aerobic
wastewater treatment
Gg BOD/year
1990-2019: Calculated
Smass
Raw sludge removed from wastewater
treatment as dry mass3
Tg dry weight/year
1988: EPA (1993c); EPA (1999)
1990-1995: Calculated based on
sewage sludge production change
per year EPA (1993c); EPA (1999);
Beecher et al. (2007)
1996: EPA (1999)
2004: Beecher etal. (2007)
Data for intervening years obtained
by linear interpolation
2005-2019: Forecasted from the
rest of the time series
% aerobicoTcw
Percent of flow to aerobic systems, other than
wetlands3
%
1990,1991: Set equal to 1992
1992,1996, 2000, 2004: EPA (1992,
1996, 2000, 2004a), respectively
Data for intervening years obtained
by linear interpolation.
2005-2019: Forecasted from the
rest of the time series
% aerobic
w/primary
Percent of aerobic systems with primary
treatment and no anaerobic sludge digestion (0)
%
% aerobic w/out
primary
Percent of aerobic systems without primary
treatment
%
%aerobic+digestion
Percent of aerobic systems with primary and
anaerobic sludge digestion
%
Krem,aer_prim
Sludge removal factor for aerobic treatment
plants with primary treatment (mixed primary
and secondary sludge, untreated or treated
aerobically) (0.8)
kg BOD/kg sludge
1990-2019: IPCC (2019)
Krem,aer_noprim
Sludge removal factor for aerobic wastewater
treatment plants without separate primary
treatment (1.16)
kg BOD/kg sludge
Krem,aer_digest
Sludge removal factor for aerobic treatment
plants with primary treatment and anaerobic
sludge digestion (mixed primary and secondary
sludge, treated anaerobically) (1)
kg BOD/kg sludge
1000
Conversion factor
metric tons to
kilograms
EF aerobic
Emission factor - aerobic systems (0.018)
kg CH4/kg BOD
Waste 7-27

-------
Variable
Variable Description
Units
Inventory Years: Source of Value
^aerobic
Amount CH4 recovered or flared from aerobic
wastewater treatment (0)
kg CH4/year

a Value of this activity data varies over the time series.
Constructed wetlands exhibit both aerobic and anaerobic treatment (partially anaerobic treatment) but are
referred to in this chapter as aerobic systems. Constructed wetlands may be used as the sole treatment unit at a
centralized wastewater treatment plant or may serve as tertiary treatment after simple settling and biological
treatment. Emissions from all constructed wetland systems were included in the estimates of emissions from
centralized wastewater treatment plant processes and effluent from these plants. Methane emissions equations
from constructed wetlands used as sole treatment were previously described. Methane emissions from
constructed wetlands used as tertiary treatment were estimated by multiplying the flow from treatment to
constructed wetlands, wastewater BOD concentration entering tertiary treatment, constructed wetlands emission
factor, and then converting to kt/year.
For constructed wetlands, an IPCC default emission factor for surface flow wetlands was used. This is the most
conservative factor for constructed wetlands and was recommended by IPCC (2014) when the type of constructed
wetland is not known. A BOD5 concentration of 30 mg/L was used for wastewater entering constructed wetlands
used as tertiary treatment based on U.S. secondary treatment standards for POTWs. These standards are based on
plants generally utilizing simple settling and biological treatment (EPA 2013). Constructed wetlands do not have
secondary sludge removal.
Emissions from Centrally Treated Aerobic Systems (Constructed Wetlands Only) (IPCC 2014 (Eq. 6.1), kt
CH4/year) = B2
= [(TOWcentralized) X (%aerobiccw)] X (EFcw)
Emissions from Centrally Treated Aerobic Systems (Constructed Wetlands used as Tertiary Treatment) (U.S.
Specific, ktCH4/year) = B3
= [(POTW_flow_CW) X (BODcwjnf) X 3.785 X (EFcw)] X 1/106 X 365.25
Table 7-15: Variables and Data Sources for ChU Emissions from Centrally Treated Aerobic
Systems (Constructed Wetlands)
Variable Variable Description
Units
Inventory Years: Source of Value
Emissions from Constructed Wetlands Only (kt CH4/year)
TOWcentralized
Total organics in centralized
wastewater treatment
Gg
BOD/year
1990-2019: Calculated
% aerobiccw
Flow to aerobic systems,
constructed wetlands used as sole
treatment / total flow to POTWs.
%
1990,1991: Set equal to 1992
1992, 1996, 2000, 2004, 2008, 2012:
EPA (1992, 1996, 2000, 2004a, 2008b,
and 2012)
Data for intervening years obtained by
linear interpolation.
2013-2019: Forecasted from the rest
of the time series
EFcw
Emission factor for constructed
wetlands
kg CH4/kg
BOD
1990-2019: IPCC (2014)
Emissions from Constructed Wetlands used as Tertiary Treatment (kt CH4/year)
POTW_flow_CW
Wastewater flow to POTWs that
use constructed wetlands as
tertiary treatment3
MGD
1990,1991: Set equal to 1992
1992, 1996, 2000, 2004, 2008, 2012:
EPA (1992, 1996, 2000, 2004a, 2008b,
and 2012)
Data for intervening years obtained by
linear interpolation.
7-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Variable
Variable Description
Units
Inventory Years: Source of Value



2013-2019: Forecasted from the rest
of the time series
BODcw.inf
BOD concentration in wastewater
entering the constructed wetland
(30)
mg/L
1990-2019: EPA (2013)
3.785
Conversion factor
liters to
gallons
Standard conversion
EFcw
Emission factor for constructed
wetlands (0.24)
kg CH4/kg
BOD
1990-2019: IPCC (2014)
1/106
Conversion factor
kg to kt
Standard conversion
365.25
Conversion factor
Days in a
year
Standard conversion
Data sources and methodologies for centrally treated anaerobic systems are similar to those described for aerobic
systems, other than constructed wetlands. See discussion above.
Emissions from Centrally Treated Anaerobic Systems (IPCC 2019 (Eq. 6.1), ktCH4/year) = C
— [(TOWcENTRALIZEd) X (% anaerobic) - Sanaerobic] X EFanaerobic " Ranaerobic
Table 7-16: Variables and Data Sources for ChU Emissions from Centrally Treated Anaerobic
Systems
Variable
Variable Description
Units
Inventory Years: Source of
Value
Emissions from Centrally Treated Anaerobic Systems (kt CH4/year)
TOWcentrauzed
Total organics in centralized wastewater
treatment
Gg BOD/year
1990-2019: Calculated
% anaerobic
Percent centralized wastewater that is
anaerobically treated3
%
1990,1991: Set equal to 1992
1992, 1996, 2000, 2004: EPA
(1992, 1996, 2000, 2004a),
respectively
Data for intervening years
obtained by linear
interpolation.
2005-2019: Forecasted from
the rest of the time series
^anaerobic
Organic component removed from anaerobic
wastewater treatment (0)
Gg/year
1990-2019: IPCC (2019)
EF anaerobic
Emission factor for anaerobic reactors/deep
lagoons (0.48)
kg CH4/kg BOD
Ranaerobic
Amount CH4 recovered or flared from
anaerobic wastewater treatment (0)
kg CH4/year
Emissions from Anaerobic Sludge Digesters:
Total CH4 emissions from anaerobic sludge digesters were estimated by multiplying the wastewater influent flow
to POTWs with anaerobic sludge digesters, the cubic feet of digester gas generated per person per day divided by
the flow to POTWs, the fraction of CH4 in biogas, the density of CH4, one minus the destruction efficiency from
burning the biogas in an energy/thermal device and then converting the results to kt/year.
Emissions from Anaerobic Sludge Digesters (U.S. Specific, ktCH4/year) = D
= [(POTW_flow_AD) x (biogas gen)/(100)] x 0.0283 x (FRAC_CH4) x 365.25 x (662) x (1-DE) x 1/109
Waste 7-29

-------
Table 7-17: Variables and Data Sources for Emissions from Anaerobic Sludge Digesters
Variable
Variable Description
Units
Inventory years: Source of
Value
Emissions from Anaerobic Sludge Digesters (kt CH4/year)
POTW_flow_AD
POTW Flow to Facilities with Anaerobic
Sludge Digesters
MGD
1990,1991: Set equal to 1992
1992,1996, 2000, 2004: EPA
(1992, 1996, 2000, and 2004a),
respectively
Data for intervening years
obtained by linear
interpolation.
2005-2019: Forecasted from
the rest of the time series
biogas gen
Gas Generation Rate (1.0)
ft3/capita/day
1990-2019: Metcalf & Eddy
(2014)
100
Per Capita POTW Flow (100)
gal/capita/day
1990-2019: Ten-State
Standards (2004)
0.0283
Conversion factor
ft3 to m3
Standard Conversion
frac_ch4
Proportion of Methane in Biogas (0.65)
No units
1990-2019: Metcalf & Eddy
(2014)
365.25
Conversion factor
Days in a year
Standard conversion
662
Density of Methane (662)
g CH4/m3 CH4
1990-2019: EPA (1993a)
DE
Destruction Efficiency (99% converted
to fraction)
No units
1990-2019: EPA (1998); CAR
(2011); Sullivan (2007); Sullivan
(2010); and UNFCCC (2012)
1/109
Conversion factor
g to kt
Standard conversion
Emissions from Discharge of Centralized Treatment Effluent:
Methane emissions from the discharge of wastewater treatment effluent were estimated by multiplying the total
BOD of the discharged wastewater effluent by an emission factor associated with the location of the discharge.
The BOD in treated effluent was determined by multiplying the total organics in centrally treated wastewater by
the percent of wastewater treated in primary, secondary, and tertiary treatment, and the fraction of organics
remaining after primary treatment (one minus the fraction of organics removed from primary treatment,
secondary treatment, and tertiary treatment).
Emissions from Centrally Treated Systems Discharge (U.S. Specific, ktCH4/year) = E
= (TOWrlE X EFrle) + (TOWother X EFother)
where,
Total organics in centralized treatment effluent (IPCC 2019 (Eq. 6.3D), Gg BOD/year) = TOWEFFtreat.cENTRALizED
= [TOWcentralized X % primary X (l-TOWrem.PRiMARy)] + [TOWcentralized X % secondary X (1-
TOWrem.SECONDARy)] + [TOWCENTRALIZED X % tertiary X (l-TOWrem,TERTIARY)]
Total organics in effluent discharged to reservoirs, lakes, or estuaries (U.S. Specific, Gg BOD/year) = TOWrle
= TOWEFFtreat.CENTRALIZED X PerCentRLE
Total organics in effluent discharged to other waterbodies (U.S. Specific, Gg BOD/year) = TOWother
= TOWEFFtreat.CENTRALIZED X PerCentother
7-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 7-18: Variables and Data Sources for ChU Emissions from Centrally Treated Systems
Discharge
Variable
Variable Description
Units
Source of Value




TOWEFFtreat.CENTRAUZED
Total organics in centralized treatment effluent
Gg
BOD/year
1990-2019:
Calculated
TOWcentrauzed
Total organics in centralized wastewater treatment
Gg
BOD/year
1990-2019:
Calculated
% primary
Percent of primary domestic centralized treatment
%
1990,1991: Set
% secondary
Percent of secondary domestic centralized treatment
%
equal to 1992.



1992, 1996, 2000,
2004, 2008, 2012:
EPA (1992, 1996,
2000, 2004a, 2008,
and 2012),
respectively
% tertiary
Percent of tertiary domestic centralized treatment
%
Data for
intervening years
obtained by linear
interpolation.
2013-2019:
Forecasted from
the rest of the time
series
TOWrem.PRIMARY
Fraction of organics removed from primary domestic
centralized treatment
No units

TOWrerTlisECONDARY
Fraction of organics removed from secondary domestic
centralized treatment
No units
1990-2019: IPCC
(2019)
TOWrem.TERTIARY
Fraction of organics removed from tertiary domestic
centralized treatment
No units

TOWrle
Total organics in effluent discharged to reservoirs, lakes, and
estuaries
Gg
BOD/year
1990-2019:
TOWother
Total organics in effluent discharge to other waterbodies
Gg
BOD/year
Calculated
EFrle
Emission factor (discharge to reservoirs/lakes/estuaries)
kg CH4/kg
BOD
1990-2019: IPCC
EFother
Emission factor (discharge to other waterbodies)
kg CH4/kg
BOD
(2019)
PercentRLE
% discharged to reservoirs, lakes, and estuaries
%
1990-2010: Set
Percentother
% discharged to other waterbodies
%
equal to 2010
2010: ERG (2021)
2011: Obtained by
linear interpolation
2012: ERG (2021)
2013-2019: Set
equal to 2012
Industrial Wastewater CH4 Emission Estimates
Industrial wastewater CH4 emissions originate from on-site treatment systems, typically comprised of biological
treatment operations. The collection systems at an industrial plant are not as extensive as domestic wastewater
sewer systems; therefore, it is not expected that dissolved CH4 will form during collection. However, some
treatment systems are designed to have anaerobic activity (e.g., anaerobic reactors or lagoons), or may
periodically have anaerobic conditions form (facultative lagoons or large stabilization basins). Emissions will also
result from discharge of treated effluent to waterbodies where carbon accumulates in sediments (typically slow-
moving systems, such as lakes, reservoirs, and estuaries).
Waste 7-31

-------
Industry categories that are likely to produce significant CH4 emissions from wastewater treatment were identified
and included in the Inventory. The main criteria used to identify U.S. industries likely to generate CH4 are whether
they generate high volumes of wastewater, whether there is a high organic wastewater load, and whether the
wastewater is treated using methods that result in CH4 emissions. The top six industries that meet these criteria
are pulp and paper manufacturing; meat and poultry processing; vegetables, fruits, and juices processing; starch-
based ethanol production; petroleum refining; and breweries. Wastewater treatment and discharge emissions for
these sectors for 2019 are displayed in Table 7-19 below. Further discussion of wastewater treatment for each
industry is included below.
Table 7-19: Total Industrial Wastewater ChU Emissions by Sector (2019, MMT CO2 Eq. and
Percent)

CH4 Emissions
% of Industrial
Industry
(MMTCOz Eq.)
Wastewater CH4
Meat & Poultry
5.0
78.5
Pulp & Paper
0.7
11.4
Fruit & Vegetables
0.2
3.6
Ethanol Refineries
0.2
2.5
Breweries
0.1
2.2
Petroleum Refineries
0.1
1.8
Total
6.4
100
Note: Totals may not sum due to independent rounding.
Emissions from Industrial Wastewater Treatment Systems:
The general IPCC equation to estimate methane emissions from each type of treatment system used for each
industrial category is:
where,
CH4 (industrial sector)
i
TOW,
Si
EF
Ri
CH4 (industrial sector) = [(TOW, - S,) x EF -R,]
= Total CH4 emissions from industrial sector wastewater treatment (kg/year)
= Industrial sector
= Total organics in wastewater for industrial sector / (kg COD/year)
= Organic component removed from aerobic wastewater treatment for
industrial sector / (kg COD/year)
= System-specific emission factor (kg CH4/kg COD)
= Methane recovered for industrial sector I (kg CH4/year)
The general IPCC equation to estimate the total organics in wastewater (TOW) for each industrial category is:
TOWi = Pi x Wi x CODi
where,
TOW,
i
Pi
Wi
COD,
= Total organically degradable material in wastewater for industry I (kg
COD/yr)
= Industrial sector
= Total industrial product for industrial sector / (t/yr)
= Wastewater generated (m3/t product)
= Chemical oxygen demand (industrial degradable organic component in
wastewater) (kg COD/m3)
7-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The annual industry production is shown in Table 7-20, and the average wastewater outflow and the organics
loading in the outflow is shown in Table 7-21. For some industries, U.S.-specific data on organics loading is
reported as BOD rather than COD. In those cases, an industry-specific COD:BOD ratio is used to convert the
organics loading to COD.
The amount of organics treated in each type of wastewater treatment system was determined using the percent of
wastewater in the industry that is treated on site and whether the treatment system is anaerobic, aerobic or
partially anaerobic.
Table 7-22 presents the industrial wastewater treatment activity data used in the calculations and described in
detail in ERG (2008a), ERG (2013a), ERG (2013b), and ERG (2021). For CH4 emissions, wastewater treated in
anaerobic lagoons or reactors was categorized as "anaerobic", wastewater treated in aerated stabilization basins
or facultative lagoons were classified as "ASB" (meaning there may be pockets of anaerobic activity), and
wastewater treated in aerobic systems such as activated sludge systems were classified as "aerobic/other."
The amount of organic component removed from aerobic wastewater treatment as a result of sludge removal
(Saerobic) was either estimated as an industry-specific percent removal, if available, or as an estimate of sludge
produced by the treatment system and IPCC default factors for the amount of organic component removed (Krem),
using one of the following equations. Table 7-23 presents the sludge variables used for industries with aerobic
wastewater treatment operations (i.e., pulp and paper, fruit/vegetable processing, and petroleum refining).
Spuip.asb = TOWpuip x % removal w/primary
where,
Spuip.asb	= Organic component removed from pulp and paper wastewater during
primary treatment before treatment in aerated stabilization basins (Gg
COD/yr)
TOWpuip	= Total organically degradable material in pulp and paper wastewater (Gg
COD/yr)
% removal w/primary = Percent reduction of organics in pulp and paper wastewater associated with
sludge removal from primary treatment (%)
Saerobic — Smass X Krem XlO ^
where,
Saerobic	= Organic component removed from fruit and vegetable or petroleum refining
wastewater during primary treatment before treatment in aerated
stabilization basins (Gg COD/yr)
Smass	= Raw sludge removed from wastewater treatment as dry mass (kg sludge/yr)
Krem	= Sludge factor (kg BOD/kg sludge)
10"6	= Conversion factor, kilograms to Gigagrams
Smass — (Sprim + Saer) x P XW
where,
Smass	= Raw sludge removed from wastewater treatment as dry mass (kg sludge/yr)
Sprim	= Sludge production from primary sedimentation (kg sludge/m3)
Saer	= Sludge production from secondary aerobic treatment (kg sludge/m3)
P	= Production (t/yr)
W	= Wastewater Outflow (m3/t)
Waste 7-33

-------
Default emission factors6 from IPCC (2019) were used. Information on methane recovery operations varied by
industry. See industry descriptions below.
Table 7-20: U.S. Pulp and Paper, Meat, Poultry, Vegetables, Fruits and Juices, Ethanol,
Breweries, and Petroleum Refining Production (MMT)


Meat
Poultry
Vegetables,





(Live Weight
(Live Weight
Fruits and
Ethanol

Petroleum
Year
Pulp and Paper3
Killed)
Killed)
Juices
Production
Breweries
Refining
1990
83.6
27.3
14.6
38.7
2.5
23.9
702.4
2005
92.4
31.4
25.1
42.9
11.7
23.1
818.6
2015
80.9
32.8
27.7
44.6
44.2
22.4
914.5
2016
79.9
34.2
28.3
43.5
45.8
22.3
926.0
2017
80.3
35.4
28.9
42.9
47.2
21.8
933.5
2018
79.4
36.4
29.4
42.6
48.0
21.5
951.4
2019
78.8
37.4
30.1
43.1
47.2
21.1
940.2
a Pulp and paper production is the sum of market pulp production plus paper and paperboard production.
Sources: Pulp and Paper - FAO (2020a) and FAO (2020b); Meat, Poultry, and Vegetables - USDA (2020a and 2020c);
Ethanol - Cooper (2018) and RFA (2020a and 2020b); Breweries - Beer Institute (2011) and TTB (2020); Petroleum
Refining - EIA (2020).
Table 7-21: U.S. Industrial Wastewater Characteristics Data (2019)
Industry
Wastewater
Wastewater
Wastewater

Outflow (m3/ton)
BOD (g/L)
COD (kg/m3)
COD: BOD Ratio
Pulp and Paper
See Table 7-25
0.3
-
2.5
Meat Processing
5.3
2.8
-
3
Poultry Processing
12.5
1.5
-
3
Fruit/Vegetable Processing
See Table 7-26

-
1.5
Ethanol Production - Wet Mill
10a
1.5
-
2
Ethanol Production - Dry Mill
1.25a
3b
-
2
Petroleum Refining
0.8
-
0.45
2.5
Breweries - Craft
3.09
-
17.6
1.67
Breweries - NonCraft
1.94
-
17.6
1.67
a Units are gallons per gallons ethanol produced.
b Units are COD (g/L).
Sources: Pulp and Paper (BOD, COD:BOD) - Malmberg (2018); Meat and Poultry (Outflow, BOD) - EPA (2002); Meat
and Poultry (COD:BOD) - EPA (1997a); Fruit/Vegetables (Outflow, BOD) - CAST (1995), EPA (1974), EPA (1975);
Fruit/Vegetables (COD:BOD) - EPA (1997a); Ethanol Production - Wet Mill (Outflow) - Donovan (1996), NRBP
(2001), Ruocco (2006a); Ethanol Production - Wet Mill (BOD) - White and Johnson (2003); Ethanol Production -
Dry Mill (Outflow and COD) - Merrick (1998), Ruocco (2006a); Ethanol Production (Dry and Wet, COD:BOD) - EPA
(1997a); Petroleum Refining (Outflow) - ERG (2013b); Petroleum Refining (COD) - Benyahia et al. (2006);
Petroleum Refining (COD:BOD) - EPA (1982); Breweries-Craft BIER (2017); ERG (2018b); Breweries - NonCraft
ERG (2018b); Brewers Association (2016a); Breweries (Craft and NonCraft; COD and COD:BOD) - Brewers
Association (2016b).
6 Emission factors are calculated by multiplying the maximum CH4-producing capacity of wastewater (B0, 0.25 kg CH4/kg COD)
and the appropriate methane correction factors (MCF) for aerobic (0), partially anaerobic (0.2), and anaerobic (0.8) systems
(IPCC 2019).
7-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 7-22: U.S. Industrial Wastewater Treatment Activity Data
% Treated Aerobically
Industry
% Wastewater
% Treated
% Treated

% Treated in
Treated On Site
Anaerobically
Aerobically
% Treated in
ASBs
Other
Aerobic
Pulp and Paper
60
5.2
75.9
38.5
37.4
Meat Processing
33
331
33
0
33
Poultry Processing
25
251
25
0
25
Fruit/Vegetable Processing
11
0
11
5.5
5.5
Ethanol Production - Wet





Mill
33.3
33.3
0
0
0
Ethanol Production - Dry





Mill
75
75
0
0
0
Petroleum Refining
62.1
0
62.1
23.6
38.5
Breweries - Craft
0.5
0.5
0
0
0
Breweries - NonCraft
100
99
1
0
1
1 Wastewater is pretreated in anaerobic lagoons prior to aerobic treatment.
Note: Due to differences in data availability and methodology, zero values in the table are for calculation purposes only and
may indicate unavailable data.
Sources: ERG (2008b); ERG (2013a); ERG (2013b); ERG (2021).
Table 7-23: Sludge Variables for Aerobic Treatment Systems
Variable
Pulp and
Paper
Industry
Fruit/Vegetable
Processing
Petroleum
Refining
Organic reduction associated with sludge removal (%)
Sludge Production (kg/m3)
Primary Sedimentation
Aerobic Treatment
Sludge Factor (kg BOD/kg dry mass sludge)
Aerobic Treatment w/Primary Sedimentation and No Anaerobic
Sludge Digestion
Aerobic Treatment w/out Primary Sedimentation
58
0.15
0.096
0.8
0.096
1.16
Sources: Organic reduction (pulp) - ERG (2008a); Sludge production - Metcalf & Eddy (2003); Sludge factors - IPCC (2019).
Emissions from Discharge of Industrial Wastewater Treatment Effluent:
Methane emissions from discharge of industrial wastewater treatment effluent are estimated by multiplying the
total organic content of the discharged wastewater effluent by an emission factor associated with the discharge:
where,
CH4 Effluentif
TOWeffluent.ind
EFei
CH4 EffluentiND = TOWeffluent,ind X EFeffluent
CH4 emissions from industrial wastewater discharge for inventory year (kg
CH4/year)
Total organically degradable material in wastewater effluent from industry
for inventory year (kg COD/year or kg BOD/year)
Tier 1 emission factor for wastewater discharged to aquatic environments
(0.028 kg CH4/kg COD or 0.068 kg CH4/kg BOD) (IPCC 2019)
Waste 7-35

-------
The COD or BOD in industrial treated effluent (TOWEffluent,industry) was determined by multiplying the total organics
in the industry's untreated wastewater that is treated on site by an industry-specific percent removal where
available or a more general percent removal based on biological treatment for other industries.
Table 7-22 presents the percent of wastewater treated onsite, while Table 7-24 presents the fraction of TOW
removed during treatment.
TOWeffluent,IND — TOWind * %onsite * (1 - TOWrem)
where,
TOWeffluent,ind	= Total organically degradable material in wastewater effluent from industry
for inventory year (kg COD/year or kg BOD/year)
TOWind	= Total organics in untreated wastewater for industry (kg COD/year)
%onsite	= Percent of industry wastewater treated on site (%)
TOWrem	= Fraction of organics removed during treatment
Table 7-24: Fraction of TOW Removed During Treatment by Industry
Industry
TOWrem
Source
Pulp, Paper, and Paperboard
Red Meat and Poultry
Fruits and Vegetables
Ethanol Production
Biomethanator Treatment
Other Treatment
Petroleum Refining
Breweries
0.905
0.85
0.85
0.90
0.85
0.93
0.85
Malmberg (2018)
IPCC (2019), Table 6.6b
IPCC (2019), Table 6.6b
ERG (2008a), ERG (2006b)
IPCC (2019), Table 6.6b
Kenari, Sarrafzadeh, and Tavakoli
(2010)
IPCC (2019), Table 6.6b
Discussion of Industry-Specific Data:
Pulp and Paper. Wastewater treatment for the pulp and paper industry typically includes neutralization, screening,
sedimentation, and flotation/hydrocycloning to remove solids (World Bank 1999; Nemerow and Dasgupta 1991).
Secondary treatment (storage, settling, and biological treatment) mainly consists of lagooning. About 60 percent of
pulp and paper mills have on-site treatment with primary treatment and about half of these also have secondary
treatment (ERG 2008). In the United States, primary treatment is focused on solids removal, equalization,
neutralization, and color reduction (EPA 1993b). The vast majority of pulp and paper mills with on-site treatment
systems use mechanical clarifiers to remove suspended solids from the wastewater. About 10 percent of pulp and
paper mills with treatment systems use settling ponds for primary treatment and these are more likely to be
located at mills that do not perform secondary treatment (EPA 1993b).
Approximately 42 percent of the BOD passes on to secondary treatment, which consists of activated sludge,
aerated stabilization basins, or non-aerated stabilization basins. Pulp and paper mill wastewater treated using
anaerobic ponds or lagoons or unaerated ponds were classified as anaerobic (with an MCF of 0.8). Wastewater
flow treated in systems with aerated stabilization basins or facultative lagoons was classified as partially anaerobic
(with an MCF of 0.2, which is the 2006 IPCC Guidelines-suggested MCF for shallow lagoons). Wastewater flow
treated in systems with activated sludge systems or similarly aerated biological systems was classified as aerobic.
A time series of CH4 emissions for 1990 through 2019 was developed based on paper and paperboard production
data and market pulp production data. Market pulp production values were available directly for 1998, 2000
through 2003, and 2010 through 2018. Where market pulp data were unavailable, a percent of woodpulp that is
market pulp was applied to woodpulp production values from FAOSTAT to estimate market pulp production (FAO
2020a). The percent of woodpulp that is market pulp for 1990 to 1997 was assumed to be the same as 1998,1999
7-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
was interpolated between values for 1998 and 2000, 2000 through 2009 were interpolated between values for
2003 and 2010, and 2019 was forecasted from the rest of the time series. A time series of the overall wastewater
outflow is presented in Table 7-25. Data for 1990 through 1994 varies based on data outlined in ERG (2013a) to
reflect historical wastewater flow. Wastewater generation rates for 1995, 2000, and 2002 were estimated from the
2014 American Forest and Paper Association (AF&PA) Sustainability Report (AF&PA 2014). Wastewater generation
rates for 2004, 2006, 2008, 2010, 2012, and 2014 were estimated from the 2016 AF&PA Sustainability Report
(AF&PA 2016). Data for 2005 and 2016 were obtained from the 2018 AF&PA Sustainability Report (AF&PA 2018).
Data for intervening years were obtained by linear interpolation, while 2017 to 2019 were forecasted from the rest
of the time series. The average BOD concentrations in raw wastewater was estimated to be 0.4 grams BOD/liter
for 1990 to 1998, while 0.3 grams BOD/liter was estimated for 2014 through 2019 (EPA 1997b; EPA 1993b; World
Bank 1999; Malmberg 2018). Data for intervening years were obtained by linear interpolation.
Table 7-25: Wastewater Outflow (m3/ton) for Pulp, Paper, and Paperboard Mills

Wastewater
Year
Outflow (m3/ton)
1990
68
2005
43
2015
40
2016
40
2017
39
2018
38
2019
38
Sources: ERG (2013a), AF&PA (2014), AF&PA
(2016), AF&PA (2018).
Meat and Poultry Processing. The meat and poultry processing industry makes extensive use of anaerobic lagoons
in sequence with screening, fat traps, and dissolved air flotation when treating wastewater on site. Although all
meat and poultry processing facilities conduct some sort of treatment on site, about 33 percent of meat processing
operations (EPA 2002) and 25 percent of poultry processing operations (U.S. Poultry 2006) perform on-site
treatment in anaerobic lagoons. The IPCC default emission factor of 0.2 kg CH4/kg COD for anaerobic lagoons were
used to estimate the CH4 produced from these on-site treatment systems.
Vegetables, Fruits, and Juices Processing. Treatment of wastewater from fruits, vegetables, and juices processing
includes screening, coagulation/settling, and biological treatment (lagooning). The flows are frequently seasonal,
and robust treatment systems are preferred for on-site treatment. About half of the operations that treat and
discharge wastewater use lagoons intended for aerobic operation, but the large seasonal loadings may develop
limited anaerobic zones. In addition, some anaerobic lagoons may also be used (Nemerow and Dasgupta 1991).
Wastewater treated in partially anaerobic systems were assigned the IPCC default emission factor of 0.12 kg
CH4/kg BOD. Outflow and BOD data, presented in Table 7-26, were obtained from CAST (1995) for apples, apricots,
asparagus, broccoli, carrots, cauliflower, cucumbers (for pickles), green peas, pineapples, snap beans, and spinach;
EPA (1974) for potato and citrus fruit processing; and EPA (1975) for all other commodities.
Table 7-26: Wastewater Outflow (m3/ton) and BOD Production (g/L) for U.S. Vegetables,
Fruits, and Juices Production
Organic Content in Untreated
Commodity	Wastewater Outflow (mB/ton)	Wastewater (g BOD/L)
Vegetables
Potatoes	10.27	1.765
Waste 7-37

-------
Commodity
Wastewater Outflow (mB/ton)
Organic Content in Untreated
Wastewater (g BOD/L)
Other Vegetables
9.93
0.755
Fruit


Apples
9.09
8.17
Citrus Fruits
10.11
0.317
Non-citrus Fruits
12.59
1.226
Grapes (for wine)
2.78
1.831
Sources: CAST (1995); EPA (1974); EPA (1975).
Ethanol Production. Ethanol, or ethyl alcohol, is produced primarily for use as a fuel component, but is also used in
industrial applications and in the manufacture of beverage alcohol. Ethanol can be produced from the
fermentation of sugar-based feedstocks (e.g., molasses and beets), starch- or grain-based feedstocks (e.g., corn,
sorghum, and beverage waste), and cellulosic biomass feedstocks (e.g., agricultural wastes, wood, and bagasse).
Ethanol can also be produced synthetically from ethylene or hydrogen and carbon monoxide. However, synthetic
ethanol comprises a very small percent of ethanol production in the United States. Currently, ethanol is mostly
made from sugar and starch crops, but with advances in technology, cellulosic biomass is increasingly used as
ethanol feedstock (DOE 2013).
Ethanol is produced from corn (or other sugar or starch-based feedstocks) primarily by two methods: wet milling
and dry milling. Historically, the majority of ethanol was produced by the wet milling process, but now the majority
is produced by the dry milling process. The dry milling process is cheaper to implement and is more efficient in
terms of actual ethanol production (Rendleman and Shapouri 2007). The wastewater generated at ethanol
production facilities is handled in a variety of ways. Dry milling facilities often combine the resulting evaporator
condensate with other process wastewaters, such as equipment wash water, scrubber water, and boiler blowdown
and anaerobically treat this wastewater using various types of digesters. Wet milling facilities often treat their
steepwater condensate in anaerobic systems followed by aerobic polishing systems. Wet milling facilities may treat
the stillage (or processed stillage) from the ethanol fermentation/distillation process separately or together with
steepwater and/or wash water. Methane generated in anaerobic sludge digesters is commonly collected and
either flared or used as fuel in the ethanol production process (ERG 2006b).
About 33 percent of wet milling facilities and 75 percent of dry milling facilities treat their wastewater
anaerobically. A default emission factor of 0.2 kg CH4/kg COD for anaerobic treatment was used to estimate the
CH4 produced from these on-site treatment systems. The amount of CH4 recovered through the use of
biomethanators was estimated, and a 99 percent destruction efficiency was used. Biomethanators are anaerobic
reactors that use microorganisms under anaerobic conditions to reduce COD and organic acids and recover biogas
from wastewater (ERG 2006b). For facilities using biomethanators, approximately 90 percent of BOD is removed
during on-site treatment (ERG 2006b, 2008). For all other facilities, the removal of organics was assumed to be
equivalent to secondary treatment systems, or 85 percent (IPCC 2019).
Petroleum Refining. Petroleum refining wastewater treatment operations have the potential to produce CH4
emissions from anaerobic wastewater treatment. EPA's Office of Air and Radiation performed an Information
Collection Request (ICR) for petroleum refineries in 2011.7 Facilities that reported using non-aerated surface
impoundments or other biological treatment units (trickling filter, rotating biological contactor), which have the
potential to lead to anaerobic conditions, were assigned the IPCC default emission factor of 0.05 kg CH4/kg COD. In
addition, the wastewater generation rate was determined to be 26.4 gallons per barrel of finished product, or 0.8
m3/ton (ERG 2013b).
7 Available online at .
7-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Breweries. Since 2010, the number of breweries has increased from less than 2,000 to more than 7,000 (Brewers
Association 2020). This increase has primarily been driven by craft breweries, which have increased by over 250
percent during that period. Craft breweries were defined as breweries producing less than six million barrels of
beer per year, and non-craft breweries produce greater than six million barrels. With their large amount of water
use and high strength wastewater, breweries generate considerable CH4 emissions from anaerobic wastewater
treatment. However, because many breweries recover their CH4, their emissions are much lower.
The Alcohol and Tobacco Tax and Trade Bureau (TTB) provides total beer production in barrels per year for
different facility size categories from 2007 to the present (TTB 2020). For years prior to 2007 where TTB data were
not readily available, the Brewers Almanac (Beer Institute 2011) was used, along with an estimated percent of craft
and non-craft breweries based on the breakdown of craft and non-craft for the years 2007 through 2019.
To determine the overall amount of wastewater produced, data on water use per unit of production and a
wastewater-to-water ratio were used from the Benchmarking Report (Brewers Association 2016a) for both craft
and non-craft breweries. Since brewing is a batch process, and different operations have varying organic loads,
full-strength brewery wastewater can vary widely on a day-to-day basis. However, the organic content of brewery
wastewater does not substantially change between craft and non-craft breweries. Some breweries may collect and
discharge high-strength wastewater from particular brewing processes (known as "side streaming") to a POTW,
greatly reducing the organics content of the wastewater that is treated on site. Subsequently, the MCF for
discharge to a POTW was assumed to be zero (ERG 2018b).
Breweries may treat some or all of their wastewater on site prior to discharge to a POTW or receiving water. On-
site treatment operations can include physical treatment (e.g., screening, settling) which are not expected to
contribute to CH4 emissions, or biological treatment, which may include aerobic treatment or pretreatment in
anaerobic reactors (ERG 2018b). The IPCC default emission factor of 0.2 kg CH4/kg COD for anaerobic treatment
and 0 for aerobic treatment were used to estimate the CH4 produced from these on-site treatment systems (IPCC
2006). The amount of CH4 recovered through anaerobic wastewater treatment was estimated, and a 99 percent
destruction efficiency was used (ERG 2018b; Stier J. 2018). Very limited activity data are available on the number
of U.S. breweries that are performing side streaming or pretreatment of wastewater prior to discharge.
Domestic Wastewater N20 Emission Estimates
Domestic wastewater N20 emissions originate from both septic systems and POTWs. Within these centralized
systems, N20 emissions can result from aerobic systems, including systems like constructed wetlands. Emissions
will also result from discharge of centrally treated wastewater to waterbodies with nutrient-impacted/eutrophic
conditions. The systems with emission estimates are:
•	Septic systems (A);
•	Centralized treatment aerobic systems (B), including aerobic systems (other than constructed wetlands)
(Bl), constructed wetlands only (B2), and constructed wetlands used as tertiary treatment (B3);
•	Centralized anaerobic systems (C); and
•	Centralized wastewater treatment effluent (D).
Methodological equations for each of these systems are presented in the subsequent subsections; total domestic
N20 emissions are estimated as follows:
Total Domestic N2O Emissions from Wastewater Treatment and Discharge (kt) = A + B + C+ D
Table 7-27 presents domestic wastewater N20 emissions for both septic and centralized systems, including
emissions from centralized wastewater treatment effluent, in 2019.
Waste 7-39

-------
Table 7-27: Domestic Wastewater N2O Emissions from Septic and Centralized Systems
(2019, kt, MMT CO2 Eq. and Percent)

N20 Emissions (kt)
N20 Emissions
(MMT C02 Eq.)
% of Domestic
Wastewater N20
Septic Systems
3
0.9
3.5
Centrally-Treated Aerobic Systems
66
19.6
76.1
Centrally-Treated Anaerobic Systems
0
0.0
0
Centrally-Treated Wastewater Effluent
18
5.3
20.4
Total
87
25.8
100
Emissions from Septic Systems:
Nitrous oxide emissions from domestic treatment depend on the nitrogen present, in this case, in the form of
protein. Per capita protein consumption (kg protein/person/year) was determined by multiplying per capita annual
food availability data and its protein content. Those data are then adjusted using a factor to account for the
fraction of protein actually consumed. The methodological equations are:
Annual per capita protein supply (U.S. Specific, kg/person/year) = ProteinsuppLY
= ProteinPercapita/1000 x 365.25
Consumed Protein (IPCC 2019 (Eq. 6.10A), kg/person/year) = Protein
= ProteinsuppLY x FPC
Table 7-28: Variables and Data Sources for Protein Consumed
Variable
Variable Description
Units
Inventory Years: Source of
Value
Protein
ProteinsuppLY
Annual per capita protein supply
kg/person/year
1990-2019: Calculated
PrOteinper capita
Daily per capita protein supply3
g/person/day
1990-2019: USDA (2020b)
1000
Conversion factor
gto kg
Standard conversion
365.25
Conversion factor
Days in a year
Standard conversion
FPC
Fraction of Protein Consumed3
kg protein
consumed / kg
protein available
1990-2010: USDA (2020b)
2011-2017: FAO (2020c)
and scaling factor
2018, 2019: Forecasted
from the rest of the time
series
a Value of this activity data varies over the Inventory time series.
Nitrous oxide emissions from septic systems were estimated by multiplying the U.S. population by the percent of
wastewater treated in septic systems (about 18 percent in 2019; U.S. Census Bureau 2017), consumed protein per
capita (kg protein/person/year), the fraction of N in protein, the correction factor for additional nitrogen from
household products, the factor for industrial and commercial co-discharged protein into septic systems, the factor
for non-consumed protein added to wastewater and an emission factor and then converting the result to kt/year.
All factors obtained from IPCC (2019).
U.S. population data were taken from the U.S. Census Bureau International Database (U.S. Census Bureau 2020)
and include the populations of the United States, American Samoa, Guam, Northern Mariana Islands, Puerto Rico,
and the U.S. Virgin Islands. The fraction of the U.S. population using septic systems, as well as centralized
treatment systems (see below), is based on data from American Housing Survey (U.S. Census Bureau 2017). The
methodological equations are:
Total nitrogen entering septic systems (IPCC 2019 (Eq. 10), kg N/year) = TNdom.septic
= (USPOP X TsEPTIc) X Protein X FnPR X NhH X FNON-CON_septic X FlND-COM_septic
7-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Emissions from Septic Systems (IPCC 2019 (Eq. 6.9)) = A
= TNdom. SEPTIC X (EFseptic) X 44/28 X 1/106
Table 7-29: Variables and Data Sources for N2O Emissions from Septic System



Inventory Years: Source of
Variable
Variable Description
Units
Value
Emissions from Septic Systems
TNdom septic
Total nitrogen entering septic systems
kg N/year
1990-2019: Calculated
USpop
U.S. population3
Persons
1990-2019: U.S. Census
Bureau (2020)



Odd years from 1989



through 2017: U.S. Census



Bureau (2017)



Data for intervening years
Tseptic
Percent treated in septic systems3
%
obtained by linear
interpolation
2018 and 2019: Forecasted
from the rest of the time
series
Fnpr
Fraction of nitrogen in protein (0.016)
kg N/kg protein
1990-2019: IPCC (2019)
Nhh
Additional nitrogen from household products (1.17)
No units

F|ND-COM_septic
Factor for Industrial and Commercial Co-Discharged
Protein, septic systems (1)
No units

FNON-CON_septic
Factor for Non-Consumed Protein Added to Wastewater
(1.13)
No units

EFseptic
Emission factor, septic systems (0.0045)
kg N20-N/kg N



Molecular
Standard conversion
44/28
Conversion factor
weight ratio of
N20 to N2

1/106
Conversion factor
kg to kt
Standard conversion
a Value of this activity data varies over the Inventory time series.
Emissions from Centrally Treated Aerobic and Anaerobic Systems:
Nitrous oxide emissions from POTWs depend on the total nitrogen entering centralized wastewater treatment. The
total nitrogen entering centralized wastewater treatment was estimated by multiplying the U.S. population by the
percent of wastewater collected for centralized treatment (about 82 percent in 2019), the consumed protein per
capita, the fraction of N in protein, the correction factor for additional N from household products, the factor for
industrial and commercial co-discharged protein into wastewater treatment, and the factor for non-consumed
protein added to wastewater.
Non-consumed protein in centralized wastewater treatment for the U.S. was determined by dividing the per capita
total Kjeldahl nitrogen (TKN) loading (estimated by multiplying the influent nitrogen concentration by the
wastewater flow to centralized wastewater treatment divided by the population using centralized wastewater
treatment) by the nitrogen from protein (estimated by multiplying the fraction of N in protein [IPCC 2019] by the
annual per capita protein supply [FAO 2020c]).
Factor for Non-Consumed Protein (U.S. Specific) = Fnon-con
= [(Ninf X Flowus X 3.785 X 365.25)/ (USpop X Tcentralized)]/ (ProteinsuppLY X Fnpr)
Total nitrogen entering centralized systems (IPCC 2019 (Eq. 10), kg N/year) = TNdom.central
= (USpop X Tcentralized) X Protein X Fnpr X Nhh X Fnon-con X Find-com
Waste 7-41

-------
Table 7-30: Variables and Data Sources for Non-Consumed Protein and Nitrogen Entering
Centralized Systems
Variable
Variable Description
Units
Inventory Years: Source
of Value
Fnon-con
Factor for U.S. specific non-consumed protein
No units
1990-2019: Calculated
NinF
Influent Nitrogen Concentration (40)
mg/L
1990-2019: Metcalf &
Eddy (2014)
Flowus
Wastewater Flow to Centralized Wastewater
Treatment3
MGD
1990,1991: Set equal to
1992
1992, 1996, 2000, 2004,
2008, 2012: EPA (1992,
1996, 2000, 2004a,
2008, and 2012),
respectively
Data for intervening
years obtained by linear
interpolation
2013-2019: Forecasted
from the rest of the
time series
3.785
Conversion factor
liters to gallons
Standard conversion
365.25
Conversion factor
Days in a year
Standard conversion
USpop
U.S. population3
Persons
1990-2019: U.S. Census
Bureau (2020)
Tcentrauzed
Percent collected3
%
Odd years from 1989
through 2017: U.S.
Census Bureau (2017)
Data for intervening
years obtained by linear
interpolation
2018 and 2019:
Forecasted from the
rest of the time series
ProteinsuppLY
Annual per capita protein supply3
kg/person/year
1990-2019: Calculated
Fnpr
Fraction of nitrogen in protein (0.16)
kg N/kg protein
1990-2019: IPCC (2019)
TNdom central
Total nitrogen entering centralized systems
kg N/year
1990-2019: Calculated
Protein
Consumed protein per capita3
kg/person/year
1990-2019: Calculated
Nhh
Factor for additional nitrogen from household
products (1.17)
No units
1990-2019: IPCC (2019)
Find-com
Factor for Industrial and Commercial Co-
Discharged Protein (1.25)
No units
a Value of this activity data varies over the Inventory time series.
Nitrous oxide emissions from POTWs were estimated by multiplying the total nitrogen entering centralized
wastewater treatment, the relative percentage of wastewater treated by aerobic systems (other than constructed
wetlands) and anaerobic systems, aerobic systems with constructed wetlands as the sole treatment, the emission
factor for aerobic systems and anaerobic systems, and the conversion from N2 to N20.
Table 7-34 presents the data for U.S. population, population served by centralized wastewater treatment plants,
available protein, and protein consumed. The methodological equations are:
Emissions from Centrally Treated Aerobic Systems (other than Constructed Wetlands) (Bl) + Emissions
from Centrally Treated Aerobic Systems (Constructed Wetlands Only) (B2) + Emissions from Centrally
Treated Aerobic Systems (Constructed Wetlands used as Tertiary Treatment) (B3) = B
where,
7-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Emissions from Centrally Treated Aerobic Systems (other than Constructed Wetlands) (IPCC 2019 (Eq. 6.9),
kt N20/year) = B1
= [(TNdom.central) X (% aerobicoTcw)] X EFaerobic X 44/28 X 1/106
Table 7-31: Variables and Data Sources for N2O Emissions from Centrally Treated Aerobic
Systems (Other than Constructed Wetlands)
Variable
Variable Description
Units
Inventory Years: Source
of Value
Emissions from Centrally Treated Aerobic Systems (Other than Constructed Wetlands) (kt N20/year)
TNdom central
Total nitrogen entering centralized systems
kg N/year
1990-2019: Calculated
% aerobicoTcw
Flow to aerobic systems, other than constructed
wetlands only / total flow to POTWsa
%
1990,1991: Set equal to
1992
1992, 1996, 2000, 2004:
EPA (1992, 1996, 2000,
2004a), respectively
Data for intervening
years obtained by linear
interpolation.
2005-2019: Forecasted
from the rest of the time
series
EF aerobic
Emission factor - aerobic systems (0.016)
kg N20-N/kg N
1990-2019: IPCC (2019)
44/28
Conversion factor
Molecular
weight ratio of
N20 to N2
Standard conversion
1/106
Conversion factor
kg to kt
Standard conversion
a Value of this activity data varies over the Inventory time series.
Nitrous oxide emissions from constructed wetlands used as sole treatment include similar data and processes as
aerobic systems other than constructed wetlands. See description above. Nitrous oxide emissions from
constructed wetlands used as tertiary treatment were estimated by multiplying the flow to constructed wetlands
used as tertiary treatment, wastewater N concentration entering tertiary treatment, constructed wetlands
emission factor, and converting to kt/year.
Emissions from Centrally Treated Aerobic Systems (Constructed Wetlands Only) (IPCC 2014 (Eq. 6.9), kt
N20/year) = B2
= [(TNdom.central) X (%aerobiccw)] X EFcw X 44/28 X 1/106
Emissions from Centrally Treated Aerobic Systems (Constructed Wetlands used as Tertiary Treatment) (U.S.
Specific, kt N20/year) = B3
= [(POTW_flow_CW) X (Ncw.inf) X 3.785 X (EFcw)] X 1/106 X 365.25
Table 7-32: Variables and Data Sources for N2O Emissions from Centrally Treated Aerobic
Systems (Constructed Wetlands)
Variable
Variable Description
Units
Inventory Years: Source of
Value
Emissions from Constructed Wetlands Only (kt N20/year)
TNdom central
Total nitrogen entering centralized treatment
kg N/year
1990-2019: Calculated
% aerobiccw
Flow to aerobic systems, constructed wetlands used as sole
treatment / total flow to POTWsa
%
1990,1991: Set equal to
1992
1992, 1996, 2000, 2004,
2008, 2012: EPA (1992,
Waste 7-43

-------



Inventory Years: Source of
Variable
Variable Description
Units
Value



1996, 2000, 2004a, 2008b,



and 2012)



Data for intervening years



obtained by linear



interpolation.



2013-2019: Forecasted



from the rest of the time



series
EFcw
Emission factor for constructed wetlands (0.0013)
kg N20-N/kg N
1990-2019: IPCC (2014)


Molecular
Standard conversion


weight ratio of

44/28
Conversion factor
N20 to N2

1/106
Conversion factor
kg to kt
Standard conversion
Emissions from Constructed Wetlands used as Tertiary Treatment (kt N20/year)



1990,1991: Set equal to



1992



1992, 1996, 2000, 2004,



2008, 2012: EPA (1992,



1996, 2000, 2004a, 2008b,
POTW_flow_CW
Wastewater flow to POTWs that use constructed wetlands
MGD
and 2012)
as tertiary treatmenta
Data for intervening years
obtained by linear
interpolation.
2013-2019: Forecasted
from the rest of the time
series
Ncw.inf
BOD concentration in wastewater entering the constructed
mg/L
1990-2019: Metcalf & Eddy
wetland (25)
(2014)
3.785
Conversion factor
liters to gallons
Standard conversion
EFcw
Emission factor for constructed wetlands (0.0013)
kg N20-N/kg N
1990-2019: IPCC (2014)
1/106
Conversion factor
mg to kg
Standard conversion
365.25
Conversion factor
Days in a year
Standard conversion
a Value of this activity data varies over the Inventory time series.
Data sources and methodologies are similar to those described for aerobic systems, other than constructed
wetlands. See discussion above.
Emissions from Centrally Treated Anaerobic Systems (IPCC2019 (Eq. 6.9), kt N20/year) = C
= [(TNdom.central) X (% anaerobic)] X EFanaerobic X 44/28 X 1/106
Table 7-33: Variables and Data Sources for N2O Emissions from Centrally Treated Anaerobic
Systems
Variable
Variable Description
Units
Inventory Years: Source of
Value
Emissions from Centrally Treated Anaerobic Systems
TNdom_central
Total nitrogen entering centralized
treatment
kg N/year
1990-2019: Calculated
% anaerobic
Percent centralized wastewater that
is anaerobically treated3
%
1990,1991: Set equal to 1992
1992, 1996, 2000, 2004: (EPA
1992, 1996, 2000, 2004a),
respectively
7-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------



Data for intervening years
obtained by linear
interpolation.
2005-2019: Forecasted from
the rest of the time series
EF anaerobic
Emission factor for anaerobic
reactors/deep lagoons (0)
kg N20-N/kg N
1990-2019: IPCC (2019)
44/28
Conversion factor
Molecular weight
ratio of N20 to N2
Standard conversion
1/106
Conversion factor
mg to kg
Standard conversion
a Value of this activity data varies over the Inventory time series.
Table 7-34: U.S. Population (Millions) Fraction of Population Served by Centralized
Wastewater Treatment (percent), Protein Supply (kg/person-year), and Protein Consumed
(kg/person-year)


Centralized WWT


Year
Population
Population (%)
Protein Supply
Protein Consumed
1990
253
75.6
43.1
33.2
2005
300
78.8
44.9
34.7
2015
325
80.1
44.3
34.2
2016
327
81.1
44.7
34.4
2017
329
82.1
44.9
34.6
2018
331
82.0
44.4
34.2
2019
334
82.2
44.4
34.2
Sources: Population - U.S. Census Bureau (2020); WWTP Population - U.S. Census
Bureau (2017); Available Protein - USDA (2020b); Protein Consumed - FAO (2020c).
Emissions from Discharge of Centralized Treatment Effluent:
Nitrous oxide emissions from the discharge of wastewater treatment effluent were estimated by multiplying the
total nitrogen in centrally treated wastewater effluent by the percent of wastewater treated in primary,
secondary, and tertiary treatment and the fraction of nitrogen remaining after primary, secondary, or tertiary
treatment and then multiplying by the percent of wastewater volume routed to waterbodies with nutrient-
impaired/eutrophic conditions and all other waterbodies (ERG 2021) and emission factors for discharge to
impaired waterbodies and other waterbodies from IPCC (2019). The methodological equations are:
Emissions from Centrally Treated Systems Discharge (U.S. Specific) = D
= [(Neffluent.imp X EFimp) + (Nefluent,nonimp X EFnonimp)] X 44/28 X 1/106
where,
Total organics in centralized treatment effluent (IPCC 2019 (Eq. 6.8), kg N/year) = Neffulent.dom
= [TNdom.central8 X % primary X (l-Nrem,PRiMARY)] + [TNdom.central X % secondary X (l-Nrem,sEcoNDARY)] +
[TNdOM.CENTRAL X % tertiary X (l-Nrem,TERTIARY)]
Total nitrogen in effluent discharged to impaired waterbodies (U.S. Specific, kg N/year) = Neffluentjmp
= (Neffulent,dom X PercentiMp)/1000
Total nitrogen in effluent discharged to nonimpaired waterbodies (U.S. Specific, kg N year) = Neffluent.nonimp
8 See emissions from centrally treated aerobic and anaerobic systems for methodological equation calculating TNdom_central-
Waste 7-45

-------
= (Neffluent.dom X PercentNONiMp)/1000
Table 7-35: Variables and Data Sources for N2O Emissions from Centrally Treated Systems
Discharge
Variable
Variable Description
Units
Source of Value
Neffulent.dom
Total organics in centralized treatment effluent
kg N/year
1990-2019:
Calculated
44/28
Conversion factor
Molecular weight
ratio of N20 to N2
Standard
conversion
1/106
Conversion factor
kg to kt
Standard
conversion
TNdom_central
Total nitrogen entering centralized treatment
kg N/year
1990-2019:
Calculated
1000
Conversion factor
kg to kt
Standard
Conversion
% primary
Percent of primary domestic centralized treatment3
%
1990,1991: Set
equal to 1992.
1992, 1996, 2000,
2004, 2008, 2012:
EPA (1992,1996,
2000, 2004a, 2008,
and 2012),
respectively
Data for
intervening years
obtained by linear
interpolation.
2013-2019:
Forecasted from
the rest of the time
series
% secondary
Percent of secondary domestic centralized treatment3
%
% tertiary
Percent of tertiary domestic centralized treatment3
%
Nrem.PRIMARY
Fraction of nitrogen removed from primary domestic
centralized treatment (0.1)
No units
1990-2019: IPCC
(2019)
Nrem.SECONDARY
Fraction of nitrogen removed from secondary domestic
centralized treatment (0.4)
No units
Nrem.TERTIARY
Fraction of nitrogen removed from tertiary domestic
centralized treatment (0.9)
No units
Neffluent.imp
Total nitrogen in effluent discharged to impaired waterbodies
kg N/year
1990-2019:
Calculated
Neffluent.nonimp
Total nitrogen in effluent discharged to nonimpaired
waterbodies
kg N/year
EFimp
EF (discharge to reservoirs/lakes/estuaries) (0.19)
kg N20-N/kg N
1990-2019: IPCC
(2019)
EFlMONIMPr
EF (discharge to other waterbodies) (0.005)
kg N20-N/kg N
PercentiMP
Percent of wastewater discharged to impaired waterbodies3
%
1990-2010: Set
equal to 2010
2010: ERG (2021)
2011: Obtained by
linear interpolation
2012: ERG (2021)
2013-2019: Set
equal to 2012
PercentNONiMP
Percent of wastewater discharged to nonimpaired
waterbodies3
%
a Value for this activity data varies over the Inventory time series.
Industrial Wastewater N20 Emission Estimates
Nitrous oxide emission estimates from industrial wastewater were added to the inventory for the first time and
developed according to the methodology described in the 2019 Refinement. U.S. industry categories that are likely
7-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
to produce significant N20 emissions from wastewater treatment were identified based on whether they generate
high volumes of wastewater, whether there is a high nitrogen wastewater load, and whether the wastewater is
treated using methods that result in N20 emissions. The top four industries that meet these criteria and were
added to the inventory are meat and poultry processing; petroleum refining; pulp and paper manufacturing; and
breweries (ERG 2021). Wastewater treatment and discharge emissions for these sectors for 2019 are displayed in
Table 7-36 below. Table 7-20 contains production data for these industries.
Table 7-36: Total Industrial Wastewater N2O Emissions by Sector (2019, MMT CO2 Eq. and
Percent)

N20 Emissions


(MMTCOz
% of Industrial
Industry
Eq.)
Wastewater N20
Meat & Poultry
0.3
47.3
Petroleum Refineries
0.2
33.2
Pulp & Paper
0.1
19.0
Breweries
+
0.5
Total
0.5
100
+ Does not exceed 0.5 kt.
Note: Totals may not sum due to independent rounding.
Emissions from Industrial Wastewater Treatment Systems:
More recent research has revealed that emissions from nitrification or nitrification-denitrification processes at
wastewater treatment, previously judged to be a minor source, may in fact result in more substantial emissions
(IPCC 2019). N20 is generated as a by-product of nitrification, or as an intermediate product of denitrification.
Therefore, N20 emissions are primarily expected to occur from aerobic treatment systems. To estimate these
emissions, the total nitrogen entering aerobic wastewater treatment for each industry must be calculated. Then,
the emission factor provided by the 2019 Refinement is applied to the portion of wastewater that undergoes
aerobic treatment.
The total nitrogen that enters each industry's wastewater treatment system is a product of the total amount of
industrial product produced, the wastewater generated per unit of product, and the nitrogen expected to be
present in each meter cubed of wastewater (IPCC equation 6.13).
TN,NDi = PtXWtX TNi
where,
TN 1 nDi	- total nitrogen in wastewater for industry / for inventory year, kg TN/year.
/	= industrial sector.
P,	= total industrial product for industrial sector / for inventory year, t/year.
W,	= wastewater generated per unit of production for industrial sector / for inventory year,
m3/t product.
TNi	= total nitrogen in untreated wastewater for industrial sector / for inventory year, kg
TN/m3.
For the four industries of interest, the total production and the total volume of wastewater generated has already
been calculated for CH4 emissions. For these new N20 emission estimates, the total nitrogen in the untreated
wastewater was determined by multiplying the annual industry production, shown in Table 7-20, by the average
wastewater outflow, shown in Table 7-23, and the nitrogen loading in the outflow shown in Table 7-37.
Waste 7-47

-------
Table 7-37: U.S. Industrial Wastewater Nitrogen Data
Industry
Wastewater Total N
(kg N/ m3)
Source for Total N
Pulp and Paper
0.22a
Cabrera (2017)
Meat Processing
0.19
IPCC (2019), Table 6.12
Poultry Processing
0.19
IPCC (2019), Table 6.12
Petroleum Refining
0.051
Kenari et al. (2010)
Breweries - Craft
0.055
IPCC (2019), Table 6.12
Breweries - NonCraft
0.055
IPCC (2019), Table 6.12
a Units are kilograms N per air-dried metric ton of production.
Nitrous oxide emissions from industry wastewater treatment are calculated by applying an emission factor to the
percent of wastewater (and therefore nitrogen) that undergoes aerobic treatment (IPCC Equation 6.11).
N20 PlantsIND = [Łi(Ty x EFtJ x TNindj)] x —
where,
N20 PlantSiND	= N20 emissions from industrial wastewater treatment plants for inventory
year, kg N20/year.
TN inDi	- total nitrogen in wastewater from industry / for inventory year, kg N/year.
Ti,j	= degree of utilization of treatment/discharge pathway or system j, for each
industry / for inventory year.
/	= industrial sector.
j	= each treatment/discharge pathway or system.
EFi,j	= emission factor for treatment/discharge pathway or system j, kg N20-N/kg N.
Table 6.8a in the 2019 Refinement provides 0.016 kg N20-N/kg N as a default
IPCC value for aerobic treatment systems.
44/28	= conversion of kg N20-N into kg N20.
For each industry, the degree of utilization (Ti,j)—the percent of wastewater that undergoes each type of
treatment-was previously determined for CH4 emissions and presented in Table 7-22.
Emissions from Industrial Wastewater Treatment Effluent:
Nitrous oxide emissions from industrial wastewater treatment effluent are estimated by multiplying the total
nitrogen content of the discharged wastewater effluent by an emission factor associated with the location of the
discharge. Where wastewater is discharged to aquatic environments with nutrient-impacted/eutrophic conditions
(i.e., water bodies which are rich in nutrients and very productive in terms of aquatic animal and plant life), or
environments where carbon accumulates in sediments such as lakes, reservoirs, and estuaries, the additional
organic matter in the discharged wastewater is expected to increase emissions.
N2O EffluentiND = Neffluent.ind X EFeffluent X 44/28
where,
N20 Effluent|ND	= N20 emissions from industrial wastewater discharge for inventory year (kg
N20/year)
Neffluent.ind	= Total nitrogen in industry wastewater effluent discharged to aquatic
environments for inventory year (kg N/year)
EFeffluent	=Tier 1 emission factor for wastewater discharged to aquatic environments
(kg N20-N/kg N)
44/28	= Conversion of kg N20-N into kg N20.
7-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The total N in treated effluent was determined through use of a nutrient estimation tool developed by EPA's Office
of Water (EPA 2019). The Nutrient Tool uses known nutrient discharge data within defined industrial sectors or
subsectors, as reported on Discharge Monitoring Reports, to estimate nutrient discharges for facilities within that
sector or subsector that do not have reported nutrient discharges but are likely to discharge nutrients. The
estimation considers, within each sector or subsector, elements such as the median nutrient concentration and
flow, as well as the percent of facilities within the sector or subsector that have reported discharges. Data from
2018 are available for the pulp, paper, and paperboard, meat and poultry processing, and petroleum refining
industries. To complete the time series, an industry-specific percent removal of nitrogen was calculated using the
total nitrogen in untreated wastewater. See Table 7-38.
Because data for breweries was not available, the removal of nitrogen was assumed to be equivalent to secondary
treatment, or 40 percent (IPCC 2019). The Tier 1 emission factor (0.005 kg N20/kg N) from IPCC (2019) was used.
Table 7-38: Industrial Wastewater Nitrogen Discharged in 2018 by Sector (kg N)
Industry
N EffluentiND (kg N)
Industry-Specific
N Removal Factor
Meat & Poultry
8,773,308
0.082
Petroleum Refineries
1,698,953
0.045
Pulp & Paper
18,809,623
1.08
Breweries
1,069,919
NA
a Nitrogen discharged by breweries was estimated as 60 percent of
untreated wastewater nitrogen.
Sources: ERG (2021)
Uncertainty and Time-Series Consistency
The overall uncertainty associated with both the 2019 CH4 and N20 emission estimates from wastewater
treatment and discharge was calculated using the 2006 IPCC Guidelines Approach 2 methodology (IPCC 2006).
Uncertainty associated with the parameters used to estimate CH4 emissions include that of numerous input
variables used to model emissions from domestic wastewater and emissions from wastewater from pulp and
paper manufacturing, meat and poultry processing, fruits and vegetable processing, ethanol production,
petroleum refining, and breweries. Uncertainty associated with the parameters used to estimate N20 emissions
include that of numerous input variables used to model emissions from domestic wastewater and emissions from
wastewater from pulp and paper manufacturing, meat and poultry processing, petroleum refining, and breweries.
Uncertainty associated with centrally treated constructed wetlands parameters including U.S. population served by
constructed wetlands, and emission and conversion factors are from IPCC (2014), whereas uncertainty associated
with POTW flow to constructed wetlands and influent BOD and nitrogen concentrations were based on expert
judgment.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 7-39 and Table 7-40. For
2019, methane emissions from wastewater treatment were estimated to be between 13.3 and 25.4 MMT C02 Eq.
at the 95 percent confidence level (or in 19 out of 20 Monte Carlo Stochastic Simulations). This indicates a range of
approximately 28 percent below to 38 percent above the 2019 emissions estimate of 18.4 MMT C02 Eq. Nitrous
oxide emissions from wastewater treatment were estimated to be between 16.7 and 81.6 MMT C02 Eq., which
indicates a range of approximately 37 percent below to 209 percent above the 2019 emissions estimate of 26.4
MMT C02Eq.
For 1990, methane emissions from wastewater treatment were estimated to be between 14.8 and 27.5 MMT C02
Eq. at the 95 percent confidence level (or in 19 out of 20 Monte Carlo Stochastic Simulations). This indicates a
range of approximately 27 percent below to 37 percent above the 1990 emissions estimate of 20.2 MMT C02 Eq.
Nitrous oxide emissions from wastewater treatment were estimated to be between 12.9 and 60.1 MMT C02 Eq.,
which indicates a range of approximately 31 percent below to 218 percent above the 1990 emissions estimate of
18.7 MMT C02 Eq.
Waste 7-49

-------
Table 7-39: Approach 2 Quantitative Uncertainty Estimates for 2019 Emissions from
Wastewater Treatment (MMT CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Wastewater Treatment
ch4
18.4
13.3
25.4
-28%
+38%
Domestic
ch4
12.1
7.9
17.7
-35%
+47%
Industrial
ch4
6.4
3.8
10.3
-41%
+62%
Wastewater Treatment
n2o
26.4
16.7
81.6
-37%
+209%
Domestic
n2o
25.8
15.7
80.5
-39%
+212%
Industrial
n2o
0.6
0.6
1.7
-2%
197%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Table 7-40: Approach 2 Quantitative Uncertainty Estimates for 1990 Emissions from
Wastewater Treatment (MMT CO2 Eq. and Percent)
1990 Emission Estimate Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)	(MMT C02 Eq.)	(%)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Wastewater Treatment
ch4
20.2
14.8
27.5
-27%
+37%
Domestic
ch4
14.7
10.1
21.6
-31%
+46%
Industrial
ch4
5.4
3.3
8.4
-40%
+52%
Wastewater Treatment
n2o
18.7
12.9
60.1
-31%
+221%
Domestic
n2o
18.3
12.0
59.3
-34%
+224%
Industrial
n2o
0.4
0.4
1.3
2%
218%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2019. Details on the emission trends through time are described in more detail in the Methodology
section, above and Recalculations section below.
QA/QC and Verification
General QA/QC procedures were applied to activity data, documentation, and emission calculations consistent
with the U.S. Inventory QA/QC plan, which is in accordance with Vol. 1 Chapter 6 of 2006IPCC Guidelines (see
Annex 8 for more details). This effort included a general or Tier 1 analysis, including the following checks:
•	Checked for transcription errors in data input;
•	Ensured references were specified for all activity data used in the calculations;
•	Checked a sample of each emission calculation used for the source category;
•	Checked that parameter and emission units were correctly recorded and that appropriate conversion
factors were used;
•	Checked for temporal consistency in time series input data for each portion of the source category;
•	Confirmed that estimates were calculated and reported for all portions of the source category and for all
years;
•	Investigated data gaps that affected trends of emission estimates; and
•	Compared estimates to previous estimates to identify significant changes.
7-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Calculation-related QC (category-specific, Tier 2) was performed for a portion of the domestic wastewater
treatment discharges methodology, which included assessing available activity data to ensure the most complete
publicly data set was used and checking historical trends in the data to assist determination of best methodology
for filling in the time series for data that are not available annually.
All transcription errors identified were corrected and documented. The QA/QC analysis did not reveal any systemic
inaccuracies or incorrect input values.
EPA conducted early engagement and communication with stakeholders on updates prior to the expert review
cycle of the current Inventory. EPA held stakeholder meetings in August of 2020 where EPA provided a
presentation detailing updates made to both domestic wastewater and a portion of industrial wastewater
treatment and requested stakeholder feedback. Stakeholder feedback received is discussed in the Recalculations
Discussion and Planned Improvements sections.
Recalculations Discussion
Population data were updated to reflect revised U.S. Census Bureau datasets which resulted in changes to 2010
through 2018 values (U.S. Census Bureau 2020). Protein data were updated to reflect available protein values
available for 2014 through 2017 (FAO 2020c). Pulp, paper, and paperboard production data were updated to
reflect revised values for 2018 (FAO 2020a). Updated red meat production 2017 and 2018 data, as well as fruits
and vegetables processing production 2016 through 2018 data, were based on revised values (USDA 2020a; USDA
2020c).
EPA revised the domestic wastewater CH4 methodology based on the 2019 Refinement (IPCC 2019): added a
correction factor to account for organics from industrial and commercial contributions to POTWs (1.25); updated
the emission factor for centralized aerobic systems which accounts for loss of dissolved methane formed with in
the collection system (from 0 to 0.018 kg CH4/kg BOD); revised the estimate of organics removed with sludge from
POTWs; added emission estimates from discharge of domestic wastewater to aquatic environments based on type
of receiving water (e.g., reservoir, lake, estuaries); and updated wastewater treatment activity data to align with
the updates to organics removed and emissions from discharge to aquatic environments (ERG 2021). All of these
changes affected the time series from 1990 through 2018. Domestic wastewater treatment and discharge CH4
emissions increased an average of 43 percent over the time series, with the smallest increase of 39.6 percent (4.2
MMT C02 Eq.) in 1997 and largest increase of 48.0 percent (4.3 MMT C02 Eq.) in 2012.
EPA revised the domestic wastewater N20 methodology based on the 2019 Refinement (IPCC 2019): added
emission estimates from septic systems; added a correction factor to account for nitrogen from household
products to POTWs and septic systems (1.17); revised the methodology for treatment plants to account for aerobic
and anaerobic treatment systems; updated the emission factor for centralized aerobic systems (from 0 to 0.016 kg
N20-N/kg N); and revised emission estimates from discharge of domestic wastewater to aquatic environments to
account for the condition of the receiving waterbody (i.e., nutrient-impacted/eutrophic conditions, or not
impacted) (ERG 2021). All of these changes affected the time series from 1990 through 2018. Domestic
wastewater treatment and discharge N20 emissions increased an average 423 percent over the time series, with
the smallest increase of 410 percent (15 MMT C02 Eq.) in 2010 and largest increase of 441 percent (14.9 MMT C02
Eq.) in 1990.
EPA revised the industrial wastewater CH4 methodology based on the 2019 Refinement (IPCC 2019): revised the
estimate of organics removed with sludge; added emission estimates from discharge of industrial wastewater to
aquatic environments using a Tier 1 methodology and default emission factor; and updated wastewater treatment
activity data to align with the updates to emission factor categories (ERG 2021). All of these changes affected the
time series from 1990 through 2018. Industrial wastewater treatment and discharge CH4 emissions increased an
average of 7.7 percent over the time series, with the smallest increase of 5.9 percent (0.3 MMT C02 Eq.) in 2017
and largest increase of 10.2 percent (0.5 MMT C02 Eq.) in 1990.
EPA added industrial wastewater N20 emissions for the first time based on the 2019 Refinement (IPCC 2019)
methodology. EPA identified four categories with the largest potential contribution to include and added estimates
Waste 7-51

-------
associated with treatment plant emissions as well as emissions from the discharge of wastewater. These additions
affected the entire time series.
The cumulative effect of these recalculations had a large impact on the overall wastewater treatment emission
estimates. Over the time series, the average total emissions increased by 118 percent from the previous Inventory.
The changes ranged from the smallest increase, 108 percent (20.1 MMT C02 Eq.), in 1990, to the largest increase,
135 percent (25.8 MMT C02 Eq.), in 2017.
Planned Improvements
EPA implemented revisions based on the 2019 Refinement but notes the following continued improvements:
•	Evaluate the use of POTW BOD effluent discharge data from ICIS-NPDES.9 Currently only half of POTWs
report organics as BOD5 so EPA would need to determine a hierarchy of parameters to appropriately sum
all loads. Using these data could potentially improve the current methane emission estimates from
domestic discharge.
•	Evaluate the use of POTW N effluent discharge data from ICIS-NPDES. Currently only about 80 percent of
POTWs report a form of N so EPA would need to determine an appropriate method to scale to the total
POTW population. EPA is aware of a method for industrial sources and plans to determine if this method
is appropriate for domestic sources.
•	Investigate additional sources for estimating wastewater volume discharged and discharge location for
both domestic and industrial sources. For domestic wastewater, the goal would be to provide additional
data points along the time series, while the goal for industrial wastewater would be to update the Tier 1
discharge methodology to a Tier 2 methodology.
•	Investigate additional sources for domestic wastewater treatment data.
•	Investigate research on methane and nitrous oxide emissions released from domestic treatment
processes, as time and resources allow. This would include a 2012 Water Environment Research
Foundation study depicting a calculation-based method of potential interest.
•	Investigate anaerobic sludge digester and biogas data compiled by the Water Environment Federation
(WEF) in collaboration with other entities as a potential source of updated activity data;
o Due to lack of these data, the United States continues to use another method for estimating
biogas produced. This method uses the standard 100 gallons/capita/day wastewater generation
factor for the United States (Ten State Standards). However, based on stakeholder input, some
regions of the United States use markedly less water due to water conservation efforts so EPA
plans to investigate updated sources for this method as well.
•	Review whether sufficient data exist to develop U.S.-specific CH4 or N20 emission factors for domestic
wastewater treatment systems, including whether emissions should be differentiated for systems that
incorporate biological nutrient removal operations; and
•	Investigate additional data sources for improving the uncertainty of the estimate of N entering municipal
treatment systems.
•	Consider updating the non-consumed protein factor (Fnon-con) for centralized treatment to the default
value in IPCC (2019). The current U.S.-specific factor is likely overestimating the non-consumed protein as
it is based on influent nitrogen (which includes other sources of nitrogen) (i.e., IPCC 2019).
9 ICIS-NPDES refers to EPA's Integrated Compliance Information System - National Pollutant Discharge Elimination System.
7-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
• EPA will continue to look for methods to improve the transparency of the fate of sludge produced in
wastewater treatment.
7.3 Composting (CRF Source Category 5B1)
Composting of organic waste, such as food waste, garden (yard) and park waste, and wastewater treatment sludge
and/or biosolids, is common in the United States. Composting reduces the amount of methane-generating waste
entering landfills, destroys pathogens in the waste, sequesters carbon, and provides a source of organic matter.
Composting can also generate a saleable product and reduce the need for chemical fertilizers when the end
product is used as a fertilizer or soil amendment. If the end product is of lesser quality, it can be disposed of in a
landfill.
Composting naturally converts a large fraction of the degradable organic carbon in the waste material into carbon
dioxide (C02) through aerobic processes without anthropogenic influence. With anthropogenic influences (e.g., at
commercial or large on-site composting operations), anaerobic conditions can be created in sections of the
compost pile when there is excessive moisture or inadequate aeration (or mixing) of the compost pile, resulting in
the formation of methane (CH4). This CH4 is then oxidized to a large extent in the aerobic sections of the compost.
The estimated CH4 released into the atmosphere ranges from less than 1 percent to a few percent of the initial C
content in the material (IPCC 2006). Depending on how well the compost pile is managed, nitrous oxide (N20)
emissions can also be produced. The formation of N20 depends on the initial nitrogen content of the material and
is mostly due to nitrogen oxide (NOx) denitrification during the thermophilic and secondary mesophilic stages of
composting (Cornell 2007). Emissions vary and range from less than 0.5 percent to 5 percent of the initial nitrogen
content of the material (IPCC 2006). Animal manures are typically expected to generate more N20 than, for
example, yard waste, however data are limited.
Even though C02 emissions are generated, they are not included in net greenhouse gas emissions for composting
because they are considered biogenic, or natural occurring. In accordance with the 2006 IPCC Guidelines, only
anthropogenic emissions are included in the emission estimates for composting.
From 1990 to 2019, the amount of waste composted in the United States increased from 3,810 kt to 22,687 kt.
There was some fluctuation in the amount of waste composted between 2006 to 2009 where a peak of 20,049 kt
composted was observed in 2008, which decreased to 18,824 kt composted the following year, presumably driven
by the economic crisis of 2009. Between 2009 and 2017, the amount of waste composted gradually increased by
approximately 7 percent each year. Emissions of CH4 and N20 from composting from 2010 to 2017 have increased
by the same percentage. The past two years (2017 and 2018) are similar to 2016 in the amount of material
composted and emissions, leading one to conclude that 2017 may be a minor outlier.
In 2019, CH4 emissions from composting (see Table 7-41 and Table 7-42) were 2.3 MMT C02 Eq. (90.7 kt), and N20
emissions from composting were 2.0 MMT C02 Eq. (6.8 kt). Emissions have increased steadily from 2010 and have
exhibited a decreasing trend the past two years. The wastes composted primarily include yard trimmings (grass,
leaves, and tree and brush trimmings) and food scraps from the residential and commercial sectors (such as
grocery stores; restaurants; and school, business, and factory cafeterias). The composted waste quantities
reported here do not include small-scale backyard composting and agricultural composting mainly due to lack of
consistent and comprehensive national data. Additionally, it is assumed that backyard composting tends to be a
more naturally managed process with less chance of generating anaerobic conditions and CH4 and N20 emissions.
Agricultural composting is accounted for in Volume 4, Chapter 5 (Cropland) of this Inventory, as most agricultural
composting operations are assumed to then land-apply the resultant compost to soils.
The growth in composting since the 1990s and specifically over the past decade is attributable primarily to the
following factors: (1) the enactment of legislation by state and local governments that discouraged the disposal of
yard trimmings and food waste in landfills, (2) yard trimming collection and yard trimming drop off sites provided
by local solid waste management districts/divisions, (3) an increased awareness of the environmental benefits of
composting, and (4) loans or grant programs to establish or expand composting infrastructure.
Waste 7-53

-------
Most bans or diversion laws on the disposal of yard trimmings were initiated in the early 1990s by state or local
governments (U.S. Composting Council 2010). California, for example, enacted a waste diversion law for organics
including yard trimmings and food scraps in 1999 (AB939) that required jurisdictions to divert 50 percent of the
waste stream by 2000, or be subjected to fines. Currently, 22 states representing about 44 percent of the nation's
population have enacted such legislation (NERC 2020). There are many more initiatives at the metro and municipal
level across the United States. Roughly 4,713 composting facilities exist in the United States with most (57.2
percent) composting yard trimmings only (BioCycle 2017).
In the last decade, bans and diversions for food waste have also become more common. As of April 2019, six states
(California, Connecticut, New York, Massachusetts, Rhode Island, Vermont) and seven municipalities (Austin, TX;
Boulder, CO; Hennepin County, MN; Metro, OR; New York City, NY; San Francisco, CA; Seattle, WA) had
implemented organic waste bans or mandatory recycling laws to help reduce organic waste entering landfills, most
having taken effect after 2013 (Harvard Law School and CET 2019). In most cases, organic waste reduction in
landfills is accomplished by following recycling guidelines, donating excess food for human consumption, or by
sending waste to organics processing facilities (Harvard Law School and CET 2019). An example of an organic waste
ban as implemented by California is the California Mandatory Recycling Law (AB1826), which requires companies
to comply with organic waste recycling procedures if they produce a certain amount of organic waste and took
effect on January 1, 2015 (Harvard Law School and CET 2019). There are a growing number of initiatives to
encourage households and businesses to compost or beneficially reuse food waste, although many states and
municipalities currently have limited resources to address this directly.
Estimates for excess food and food waste at a national scale have been limited, but EPA has recently filled this gap.
EPA completed a thorough mass balance analysis of all management pathways for food waste and excess food in
the Advancing Sustainable Materials Management: 2018 report (EPA 2020d, commonly referred to as the Facts
and Figures reports) using a methodology that expanded the number of management pathways for excess food
and food waste to include:
•	animal feed;
•	bio-based materials/biochemical processing (i.e., rendering);
•	codigestion/anaerobic digestion;
•	composting/aerobic processes;
•	combustion;
•	donation;
•	land application;
•	landfill; and
•	sewer/wastewater treatment.
Approximately 18 million tons of food was diverted from landfills in 2018 (EPA 2020d).
Table 7-41: ChU and N2O Emissions from Composting (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
ch4
0.4
1.9
2.1
2.3
2.4
2.3

2.3
n2o
0.3
1.7
1.9
2.0
2.2
2.0

2.0
Total
0.7
3.6
4.0
4.3
4.6
4.3

4.3
Note: Totals may not sum due to independent rounding.




ible 7-42: ChU and N2O Emissions from Composting (kt)


Activity
1990
2005
2015
2016
2017
2018
2019

ch4
15.2
74.6
84.9
91.1
97.9
90.3
90.7

n2o
1.1
5.6
6.4
6.8
7.3
6.8
6.8

7-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
Methane and N20 emissions from composting depend on factors such as the type of waste composted, the
amount and type of supporting material (such as wood chips and peat) used, temperature, moisture content (e.g.,
wet and fluid versus dry and crumbly), and aeration during the composting process. The methodology assumes all
material composted is done so at commercial or industrial composting facilities with windrow piles (widely used
because they are cost-effective). Data for small-scale, or household composting or other non-windrow type
composting operations are not documented in the national estimates. The methodology assumes the material
composted primarily consists of yard trimmings, food waste, and some paper products.
The emissions shown in Table 7-41 and Table 7-42 were estimated using the IPCC default (Tier 1) methodology
(IPCC 2006), which is the product of an emission factor and the mass of organic waste composted (note: no CH4
recovery is expected to occur at composting operations in the emission estimates presented):
Ei = M x EFj
where,
E, = CH4 or N20 emissions from composting, kt CH4 or N20,
M = mass of organic waste composted in kt,
EF, = emission factor for composting, 41 CH4/kt of waste treated (wet basis) and
0.31 N20/kt of waste treated (wet basis) (IPCC 2006), and
i	= designates either CH4 or N20.
Per IPCC Tier 1 methodology defaults, the emission factors for CH4 and N20 assume a moisture content of 60
percent in the wet waste. (IPCC 2006). While the moisture content of composting feedstock can vary significantly
by type, composting as a process ideally proceeds between 40 to 65 percent moisture (University of Maine 2016;
Cornell Composting 1996).
Estimates of the quantity of waste composted (M, wet weight as generated) are presented in Table 7-43 for select
years. Estimates of the quantity composted for 1990, 2005, and 2015 were taken from EPA's Advancing
Sustainable Materials Management: Facts and Figures 2015 (EPA 2018); the estimates of the quantities composted
for 2016 and 2017 were taken from EPA's Advancing Sustainable Materials Management: 2016 and 2017 Tables
and Figures (EPA 2019); the estimate of the quantity composted for 2018 was taken from Table 35 of EPA's
Advancing Sustainable Materials Management: Facts and Figures 2015 (EPA 2020); and the estimate for 2019 was
extrapolated using the 2018 quantity composted and a ratio of the U.S. population growth between 2018 to 2019,
respectively (U.S. Census Bureau 2019). Note that the EPA's Advancing Sustainable Materials Management: Facts
and Figures reports present quantity of material composted in short tons and the quantities are converted to
metric tons to perform the emission calculations under the IPCC framework. The quantity of waste composted in
the Facts and Figures reports are developed to provide national coverage, but commercial/industrial composting
facilities in Puerto Rico and U.S. territories may not be explicitly included in the mass balance approach used in the
reports. This is a planned improvement as noted below.
Table 7-43: U.S. Waste Composted (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Waste Composted
3,810
18,643
21,219
22,780
24,485
22,580
22,687
Waste 7-55

-------
Uncertainty and Time-Series Consistency
Uncertainty is the difference between a true value and the measured value. Two uncertainty drivers in the
composting emission estimates include the activity data and the emission factors used.
With respect to the activity data used for the composting emission estimates, a true value is fuzzy because data
sources presenting facility-specific data are lacking. The methodology applied for the 1990 to 2019 emissions
estimates uses annually modeled estimates of waste composted sourced from EPA's Advancing Sustainable
Materials Management: Facts and Figures reports. The EPA's Facts and Figures reports uses, at a national level, a
modeled materials flow methodology, which relies on a mass balance approach. Models use a number of
parameters, which themselves may have varying degrees of uncertainty. No facility-specific or direct state data on
commercial composting facilities are used when calculating annual estimates for national quantities of waste
composted. EPA collects data from industrial associations, key businesses and industry sources, the Department of
Commerce and U.S. Census Bureau. This data is then imported into the materials flow model to estimate tonnage
of materials generated, recycled, composted, sent to combustion facilities, beneficially reused, and finally,
landfilled. Using estimates from a modeled materials flow approach introduces uncertainty.
The second large area of uncertainty lies with the emission factors themselves. A variety of different organic
wastes may be composted at the same facility such as yard waste, animal manure, food waste. One emission
factor for methane and nitrous oxide, respectively, are used regardless of the type of waste composted and the
composting method. The estimated uncertainty from the 2006IPCC Guidelines is ±50 percent for the Tier 1
methodology.
Emissions from composting in 2019 were estimated to be between 2.1 and 6.4 MMT C02 Eq., which indicates a
range of 50 percent below to 50 percent above the 2019 emission estimate (see Table 7-44).
Table 7-44: Tier 1 Quantitative Uncertainty Estimates for Emissions from Composting (MMT
CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate
(MMTCOz Eq.)
(MMT CO?
Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound

ch4
2.3
1.1
3.4
-50%
+50%
Composting
n2o
2.0
1.0
3.0
-50%
+50%

Total
4.3
2.1
6.4
-50%
+50%
The same methodological approaches (e.g., one data source for the quantity of mass composted, the same
emission factors) were applied to the entire time series to ensure consistency in emissions from 1990 through
2019. Details on the emission trends through time are described in more detail in the Methodology section, above.
QA/QC and Verification
General QA/QC procedures were applied to data gathering and input, documentation, and calculations consistent
with the U.S. Inventory QA/QC Plan, which is in accordance with Vol. 1 Chapter 6 of 2006 IPCC Guidelines (see
Annex 8 for more details). No errors were found for the current Inventory.
Recalculations Discussion
The quantity of material composted for 2018 was updated with the publication of the EPA's Advancing Sustainable
Materials Management Report (EPA 2020). The quantity of material composted decreased from 24.59 million tons
in the previous Inventory report to 22.6 million tons (or 8.19 percent) for 2018 in the current Inventory report.
Relatedly, total emissions decreased by 8.19 percent or 0.38 MMT C02 Eq. for 2018.
7-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Planned Improvements
In 2017, EPA completed a literature search on emission factors and composting systems and management
techniques that will be documented in a technical memorandum for the next (1990 to 2019) Inventory. The
purpose of this literature review was to compile all published emission factors specific to various composting
systems and composted materials in the United States to determine whether the emission factors used in the
current methodology can be revised or expanded to account for geographical differences and/or differences in
composting systems used. For example, outdoor composting processes in arid regions typically require the
addition of moisture compared to similar composting processes in wetter climates. In general, there is a lack of
facility-specific data on the management techniques and sum of material composted to enable the incorporation
of different emission factors. EPA will continue to seek out more detailed data on composting facilities to enable
this improvement in the future.
Relatedly, EPA has received comments during previous Inventory cycles recommending that calculations for the
composting sector be based on waste subcategories (i.e., leaves, grass and garden debris, food waste) and
category-specific moisture contents. At this time, EPA is not aware of any available datasets which would enable
estimations to be performed at this level of granularity. EPA will continue to search for data which could lead to
the development of subcategory-specific composting emission factors to be used in future Inventory cycles.
Efforts are also being made to improve the completeness of the composting Inventory by incorporating composted
waste from U.S. territories. In 2016, EPA conducted a desk-based investigation into industrial/commercial
composting facilities in the U.S. territories and identified facilities in Puerto Rico. Three facilities are currently
operational, and some operational data and quantities of material composted are available for the past three
years. Additional efforts are being made to collect additional historical information to estimate of the quantity of
waste composted and/or approximate the population (or households) these facilities serve. This data may be
incorporated into the current or future Inventories as a methodological improvement.
7.3 Stand-Alone Anaerobic Digestion (CRF
Source Category 5B2)
Anaerobic digestion is a series of biological processes in the absence of oxygen in which microorganisms break
down organic matter, producing biogas and soil amendments (e.g., compost). The biogas primarily consists of CH4,
biogenic C02, and trace amounts of other gases such as N20 (IPCC 2006) and is often combusted to produce heat
and power, or further processed into renewable natural gas or for use as a transportation fuel. Digester gas
contains approximately 65 percent CH4 (a normal range is 55 percent to 65 percent) and approximately 35 percent
C02 (WEF 2012). Methane emissions may result from a fraction of the biogas that is lost during the process due to
leakages and other unexpected events (0 to 10 percent of the amount of CH4 generated, IPCC 2006), collected
biogas that is not completely combusted, and entrained gas bubbles and residual gas potential in the digested
sludge. Carbon dioxide emissions are biogenic in origin and should be reported as an informational item in the
Energy Sector (IPCC 2006). Volume 5 Chapter 4 of the IPCC 2006 Guidelines notes that at biogas plants where
unintentional CH4 emissions are flared, CH4 emissions are likely to be close to zero.
Anaerobic digesters differ based on the operating temperature, feedstock type and moisture content, and mode of
operation. The operating temperature dictates the microbial communities that live in the digester. Mesophilic
microbes are present at temperatures ranging from 85 to 100 degrees Fahrenheit while thermophilic microbes
thrive at temperatures ranging from 122 to 140 degrees Fahrenheit (WEF 2012). Digesters may process one or
more types of feedstock, including food waste; municipal wastewater solids; livestock manure; industrial
wastewater and residuals; fats, oils, and grease; and other types of organic waste streams. Co-digestion (multiple
feedstocks) is employed to increase methane production in cases where an organic matter type does not break
down easily. In co-digestion, various organic wastes are decomposed in a singular anaerobic digester by using a
Waste 7-57

-------
combination of manure and food waste from restaurants or food processing industry, or a combination of manure
and waste from energy crops or crop residues (EPA 2016). The moisture content of the feedstock (wet or dry)
impacts the amount of biogas generation. Wet anaerobic digesters process feedstock with a solids content less
than 15 percent while dry anaerobic digesters process feedstock with a solids content greater than 15 percent
(EPA 2020). Digesters may also operate in batch or continuous mode, which affects the feedstock loading and
removal. Batch anaerobic digesters are manually loaded with feedstock all at once and then manually emptied
while continuous anaerobic digesters are continuously loaded and emptied with feedstock (EPA 2020).
The three main categories of anaerobic digestion facilities included in national greenhouse gas inventories include
the following:
•	Stand-alone digesters typically manage food waste from different sources, including food and beverage
processing industries. Some stand-alone digesters also co-digest other organics such as yard waste.
•	On-farm digesters that manage organic matter and reduce odor generated by farm animals or crops. On-
farm digesters are found mainly at dairy, swine, and poultry farms where there is the highest potential for
methane production to energy conversion. On-farm digesters also accept food waste as feedstock for co-
digestion.
•	Digesters at water resource recovery facilities (WRRF) to produce biogas through the treatment and
reduction of wastewater solids. Some WRRF facilities may also accept and co-digest food waste.
This section focuses on stand-alone AD facilities. Emissions from on-farm digesters are included Chapter 5
(Agriculture) and AD facilities at WRRFs are included in Section 7.2 (Wastewater Treatment).
From 1990 to 2019, the estimated amount of waste managed by stand-alone digesters in the United States
increased from approximately 866 kt to 10,620 kt, an increase of 92 percent. As described in the Uncertainty and
Time-Series Consistency section, no data sources present the annual amount of waste managed by these facilities
prior to 2015 when the EPA began a comprehensive data collection survey. Thus, the emission estimates in the
early part of the time series are general estimates, extrapolated from data collected later in the time series (i.e.,
2015 and later). The steady increase in the amount of waste processed over the time series is likely driven by
increasing interest in using waste as a renewable energy source.
In 2019, emissions from stand-alone anaerobic digestion facilities were approximately 0.2 MMT C02 Eq. (7 kt) (see
Table 7-45 and Table 7-46).
Table 7-45: ChU Emissions from Stand-Alone Anaerobic Digestion (MMT CO2 Eq.)
Activity
1990
2005
2015
2016
2017
2018
2019
CH4 Generation
+
0.1
0.2
0.2
0.2
0.2
0.2
CH4 Recovered
(+)
(+)
(+)
(+)
(+)
(+)
(+)
CH4 Emissions
+
0.1
0.2
0.2
0.2
0.2
0.2
Note: Totals may not sum due to independent rounding. Parentheses indicate negative values.
+ Does not exceed 0.05 MMT C02 Eq.
Table 7-46: ChU Emissions from Stand-Alone Anaerobic Digestion (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
CH4 Generation
1
3
8
7
8
8
8
CH4 Recovered
(+)
(+)
(0.6)
(0.7)
(0.6)
(0.6)
(0.6)
CH4 Emissions
1
2
7
7
7
7
7
Note: Totals may not sum due to independent rounding. Parentheses indicate negative values.
+ Does not exceed 0.5 kt.
7-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Methodology
Methane from anaerobic digestion depend on factors such as the type of waste managed, the amount and type of
supporting material (such as wood chips and peat) used, temperature, moisture content (e.g., wet and fluid versus
dry and crumbly), and aeration during the digestion process.
The emissions presented in Table 7-45 were estimated largely using the IPCC default (Tier 1) methodology given in
Equation 4.1 below (Volume 5, Chapter 4, IPCC 2006), which is the product of an emission factor and the mass of
organic waste processed. Only CH4 emissions are estimated because N20 emissions are considered negligible (IPCC
2006). Some Tier 2 data are available (annual quantity of waste digested) for the later portion of the time series
(2015 and later).
CH4 Emissions =	x EFt) x 10~3 — R
where,
CH4 Emissions = total CH4 emissions in inventory year, Gg CH4
M, = mass of organic waste treated by biological treatment type /', Gg, see Table 7-29
EF = emission factor for treatment /', g CH4/kg waste treated, 0.8 Mg/Gg CH4
i = anaerobic digestion
R = total amount of CH4 recovered in inventory year, Gg CH4
= Biogas x 0.0283 x minutes/year x biogas CH4 density x CCH4 x 1/10a9 x (1-DE)
Where,
Biogas = the annual amount of biogas produced, standard cubic feet per
minute (scfm)
0.0283 = conversion factor cubic meter/cubic feet
525,600 = minutes per year
662 = CH4 density in biogas (EPA 1993), g CH4/m3 CH4
65% = CCH4, concentration of CH4 in the biogas
1/10A9 = conversion factor, grams to kt
0.99 = destruction efficiency for combustion unit
Per IPCC Tier 1 methodology defaults, the emission factor for CH4 assumes a moisture content of 60 percent in the
wet waste (IPCC 2006). Both liquid and solid wastes are processed by stand-alone digesters and the moisture
content entering a digester may be higher. One emission factor recommended by the IPCC 2006 Guidelines (0.8
Mg/Gg CH4) is applied for the entire time series (IPCC 2006 Volume 5, Chapter 4, Table 4.1).
The annual quantity of waste digested is sourced from recent EPA surveys of anaerobic digestion facilities (US EPA
2018 and 2019). The EPA was granted the authority to survey anaerobic digestion facilities that process food waste
annually through an Information Collection Request (ICR No. 2533.01). The scope includes stand-alone and co-
digestion facilities (on-farm and water resource recovery facilities [WRRF]). Two reports with survey results have
been published to date (the third is expected in late 2021):
¦	Anaerobic Digestion Facilities Processing Food Waste in the United States in 2015: Survey Results (US EPA,
2018)
¦	Anaerobic Digestion Facilities Processing Food Waste in the United States in 2016: Survey Results (US EPA,
2019).
Waste 7-59

-------
These reports present aggregated survey data including the annual quantity of waste processed by digester type
(i.e., stand-alone, on-farm, and WRRF); waste types accepted; biogas generation and end use; and more. The
amount of waste digested as reported in the survey reports were assumed to be in wet weight; the majority of
stand-alone digesters were found to be wet and mesophilic (US EPA 2019).
The annual quantity of waste digested for 1990 to 2014 (only 1990 and 2005 are shown) was estimated by
multiplying the count of estimated operating facilities (as presented in Table 7-47) by the weighted average of
waste digested in 2015 and 2016 collected through EPA's survey data (EPA 2018; EPA 2019). Masked survey
responses of food and non-food waste processed were shared with the Inventory team by the EPA team leading
the EPA AD Data Collection Surveys. This provided an accurate count of the number of facilities that provided
annual quantities of digested waste, which matters for the weighted average. The weighted average applied to the
1990 to 2019 inventory is calculated as follows:
T/17 T , j .	...	n	j (^2016 X ^"aC2016 + ^2015 X ^"aC2015)
Weighted Average Waste Processed =	;			
(Fac2016 + Fac2015)
where:
W = total waste processed in the respective survey year, food and non-food waste
(short tons).
Fac = the number of facilities that reported an amount of waste processed in the
respective survey year. Note the number of facilities that provided an annual
quantity of waste processed data was internally shared and differs from the total
number of facilities that responded to the EPA surveys as presented in EPA 2018
and 2019.
Estimates of the quantity of waste digested (M, wet weight as generated) are presented in Table 7-47 for select
years and the number of facilities that reported annual quantities of waste digested to the EPA survey were 45 and
44 in 2015 and 2016, respectively (masked facility data provided by the EPA AD survey data collection team).
Estimates of the quantity of waste digested for 1990 to 2014 are calculated by multiplying the weighted average of
waste digested from 2015 and 2016 survey data (216,494 short tons) by the count of operating facilities in each
year. This calculation assumes that each facility operates continuously from the first year of operation for the
remainder of the time series. Additional efforts will be made to quantify the number of operating facilities and
estimates of the total waste digested by year for future inventories as described in the Planned Improvements
section. Estimates of the quantity digested for 2015 and 2016 were taken from EPA's AD survey data (EPA 2018;
EPA 2019, respectively). The estimate of waste digested for 2017 to 2019 were extrapolated using the average of
the waste digested from the 2015 and 2016 surveys as a proxy. Planned updates to the waste digested for 2017 to
2019 are described in the Planned Improvements section.
Table 7-47: U.S. Waste Digested (kt)
Activity
1990
2005
2015
2016
2017
2018
2019
Waste Digested1
786
3,142
9,963
9,305
9,634
9,634
9,634
1 The amount of waste digested primarily consists of food waste. The amount processed for all years
is likely an underestimate because the estimates were developed from survey data provided by
operating facilities for 2015 and 2016 (EPA 2018; EPA 2019). Facilities that did not respond to the
EPA surveys are not included and all years except 2015 and 2016 are estimated using assumptions
regarding the number of operating facilities and the weighted average of waste digested.
Additionally, the liquid portion of the waste digested in 2015 and 2016 are not included due to
limited information on the specific waste types to perform the unit conversion to kt. The average
waste digested as reported in EPA 2018 and 2019 is used as a proxy for years 2017 to 2019.
The estimated count of operating facilities is calculated by summing the count of digesters that began operating by
year over the time series. The year a digester began operating is sourced from EPA 2019. This assumes all facilities
are in operation from their first year of operation throughout the remainder of the time series. This is likely an
7-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
overestimate of facilities operating per year but does not necessarily translate to an overestimate in the amount of
waste processed because a weighted average of waste processed for the surveyed facilities is applied to these
years. The number of facilities in 1990 to 2014 are directly used in calculating the emissions, while the directly
reported annual amount of waste processed from the survey data are used for 2015 to 2017.
Table 7-48: Estimated Number of Stand-Alone AD Facilities Operating from 1990-20191
Year
1990
2005
2015
2016
2017
2018
2019
Estimated Count of
Operational Facilities
4
16
56
58
58
60
60
1 The count of operational facilities was visually estimated from Figure 5 in EPA 2020, which presents the count of the first year
of digester operation. The number of operational facilities by year is assumed to be the cumulative total from the prior year.
This method assumes all facilities are operating from 1990, or their first year of operation, to 2019. This is further discussed in
the Uncertainty and Time Series section.
Activity data for the amount of biogas recovered (R in the emission calculation equation) is limited across the time
series. Currently, there are only two data points (2015 and 2016) represented for the entire sector, as reported in
the EPA AD Data Collection Survey reports (EPA 2018 and 2019). The total quantity of collected biogas from the
survey respondents for 2015 and 2016 is reported in standard cubic feet per minute (scfm) as shown in Table 7-49.
Volume 5 Chapter 4 of the 2006IPCC Guidelines notes that only emissions from flaring can be reported under the
waste sector. The top three known uses of the biogas generation for stand-alone digesters are for combined heat
and power (CHP), the production of electricity that is sold to the grid, and using the biogas to fuel boilers and
furnaces to heat the digestor and other facility spaces (EPA 2018; EPA 2019). Thus, no biogas is assumed to be
flared.
Table 7-49: Estimated Biogas Produced and Methane Recovered from Stand-Alone AD
Facilities Operating from 1990-20191
Activity
1990
2005
2015
2016
2017
2018
2019
Total Biogas
Produced (scfm)2
820
3,279
9,176
10,498
11,886
12,296
12,296
R, recovered CH4
from biogas (kt)3
0.05
0.21
0.59
0.67
0.63
0.63
0.63
1 Total biogas produced in standard cubic feet per minute (scfm) was reported in aggregate in the EPA survey data (EPA
2018 and 2019) for 2015 and 2016. The quantities presented in this table are likely underestimates because not all
operational facilities provided a survey response to the EPA AD Data Collection Surveys.
2	Data for all years in the time series except for 2015 and 2016 are extrapolated using the average of the total biogas
collected in 2015 and 2016, divided by the average number of survey responses to generate an average estimate of
biogas collected per facility, which is then multiplied by the total facility count (as shown in Table 7-30).
3	The quantity of CH4 recovered from the biogas produced is estimated for all years except 2015 and 2016, which are taken
from EPA (2018) and (2019).
Uncertainty and Time-Series Consistency
The methodology applied for the 1990 to 2019 emissions estimates should be considered a starting point to build
on in future years. Two years of facility-provided data are available (2015 and 2016) while the rest of the time
series is estimated based off an assumption of facility counts and a weighted average annual waste processed
developed from the two years of survey data. The major limitations, and uncertainty drivers in the emissions
estimates, are related to the uncertainty in assumptions to ensure completeness across the time series and the
limitations in the EPA AD survey data, as described below:
1. The EPA AD survey (EPA 2018; EPA 2019) did not receive a 100 percent response rate, meaning that the
survey data represent a portion, albeit the majority, of stand-alone digesters, annual waste processed,
and biogas recovered. The methodology applied here did not attempt to estimate waste digested by
facilities that did not respond to the survey, which likely underestimates the quantity of waste digested
and CH4emissions.
Waste 7-61

-------
2.	The EPA AD survey data (EPA 2018; EPA 2019) present both food and non-food waste digested. The non-
food waste was reported as liquid (gallons) and solid (tons). The quantity of liquid waste managed is not
included in the estimated quantity of annual waste digested because data on the waste types are not
available to convert the quantity from gallons to tons. This slightly underestimates the quantity of waste
digested and CH4emissions.
3.	The assumption required to estimate the activity data for 1990 to 2014 may overestimate the number of
facilities in operation because it assumes that each facility operates from its start year for the entire time
series (i.e., facility closures are not taken into account). This introduces a large amount of uncertainty in
the estimates compared to years where there is directly reported survey data. It is unclear whether this
under- or over-estimates the quantity of waste digested and CH4 emissions.
The estimated uncertainty from the 2006IPCC Guidelines is ±50 percent for the Tier 1 methodology.
Emissions from stand alone anaerobic digesters in 2019 were estimated to be between 0.1 and 0.3 MMT C02 Eq.,
which indicates a range of 50 percent below to 50 percent above the 2019 emission estimate of each gas (see
Table 7-50).
Table 7-50: Tier 1 Quantitative Uncertainty Estimates for Emissions from Digestion (MMT
CO2 Eq. and Percent)
Source
Gas
2019 Emission Estimate
Uncertainty Range Relative to Emission Estimate
(MMTC02 Eq.)
(MMT C02 Eq.)
(%)




Lower Upper
Lower
Upper



Bound Bound
Bound
Bound
Stand-alone
Anaerobic Digestion
ch4
0.2
0.1 0.3
-50%
+50%
QA/QC and Verification
General QA/QC procedures were applied to data gathering and input, documentation, and calculations consistent
with the U.S. Inventory QA/QC Plan, which is in accordance with Vol. 1 Chapter 6 of 2006 IPCC Guidelines (see
Annex 8 for more details). No errors were found for the current Inventory.
Recalculations Discussion
This is a new source category included for the current (1990 to 2019) Inventory; thus, no recalculations have been
made.
Planned Improvements
Several potential improvements will be investigated for inclusion in future inventory years with the intent of
reducing the uncertainties described in the Uncertainty and Time-Series Consistency section. First, EPA plans to
incorporate survey data from future EPA AD Data Collection Surveys when the survey data are published. The next
report for 2017 is expected to be published in 2021. This addition will change the estimated emissions for 2017
and potentially the weighted average applied to the 1990 to 2014 time series. EPA will pull in survey data for
future years when published. This revision will change emissions estimates for 2018 and 2019.
Second, EPA will re-assess how best to estimate annual waste processed using proxy data for years between the
EPA AD Data Collection Survey reports as needed. The initial methodology described here assumes the same
amount of waste is processed each year from 2017 to 2019.
7-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Third, EPA will conduct additional research to confirm the number of operational facilities by year prior to 2015
and how best to estimate the quantity of waste processed per year by these facilities with the goal of better
estimating the annual quantity of waste digested between 1990 to 2014. Available data will also be compiled for
facilities that did not directly respond to the EPA AD Data Collection surveys for completeness.
Fourth, EPA will investigate the amount of recovered biogas for years prior to 2015 (i.e., the years prior to the EPA
AD Data Collection Surveys). Currently, only two years of data of recovered biogas are available and the primary
purpose will be to understand whether the range of recovered biogas from the 2015 and 2016 survey data are
reflective of earlier years.
Fifth, the uncertainty assessment will be further reviewed to confirm the appropriateness of the uncertainty
factor(s) to be applied.
7.4	Waste Incineration (CRF Source Category
5C1)	
As stated earlier in this chapter, carbon dioxide (C02), nitrous oxide (N20), and methane (CH4) emissions from the
incineration of waste are accounted for in the Energy sector rather than in the Waste sector because almost all
incineration of municipal solid waste (MSW) in the United States occurs at waste-to-energy facilities where useful
energy is recovered. Similarly, the Energy sector also includes an estimate of emissions from burning waste tires
and hazardous industrial waste, because virtually all of the combustion occurs in industrial and utility boilers that
recover energy. The incineration of waste in the United States in 2019 resulted in 11.8 MMT C02 Eq. of emissions,
over half of which (6.6 MMT C02 Eq.) is attributable to the combustion of plastics. For more details on emissions
from the incineration of waste, see Section 3.3 of the Energy chapter.
Additional sources of emissions from waste incineration include non-hazardous industrial waste incineration and
medical waste incineration. As described in Annex 5 of this report, data are not readily available for these sources
and emission estimates are not provided. An analysis of the likely level of emissions was conducted based on a
2009 study of hospital/ medical/ infectious waste incinerator (HMIWI) facilities in the United States (RTI 2009).
Based on that study's information of waste throughput and an analysis of the fossil-based composition of the
waste, it was determined that annual greenhouse gas emissions for medical waste incineration would be below
500 kt C02 Eq. per year and considered insignificant for the purposes of Inventory reporting under the UNFCCC.
More information on this analysis is provided in Annex 5.
7.5	Waste Sources of Precursor Greenhouse
Gases
In addition to the main greenhouse gases addressed above, waste generating and handling processes are also
sources of precursor gases. The reporting requirements of the UNFCCC10 request that information be provided on
precursor greenhouse gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-CH4 volatile organic
compounds (NMVOCs), and sulfur dioxide (S02). These gases are not direct greenhouse gases, but indirectly affect
terrestrial radiation absorption by influencing the formation and destruction of tropospheric and stratospheric
ozone, or, in the case of S02, by affecting the absorptive characteristics of the atmosphere. Additionally, some of
these gases may react with other chemical compounds in the atmosphere to form compounds that are greenhouse
10 See .
Waste 7-63

-------
gases. Total emissions of NOx, CO, and NMVOCs from waste sources for the years 1990 through 2019 are provided
in Table 7-51. Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and Annex 6.3.
Table 7-51: Emissions of NOx, CO, and NMVOC from Waste (kt)
Gas/Source
1990
2005
2015
2016
2017
2018
2019
NOx
+
2
2
1
1
1
1
Landfills
+
2
2
1
1
1
1
Wastewater Treatment
+
0
0
0
0
0
0
Miscellaneous3
+
0
0
0
0
0
0
CO
1
7
7
6
5
5
5
Landfills
1
6
7
6
5
5
5
Wastewater Treatment
+
+
+
+
+
+
+
Miscellaneous3
+
0
0
0
0
0
0
NMVOCs
673
114
63
57
52
52
52
Wastewater Treatment
57
49
27
25
22
22
22
Miscellaneous3
557
43
24
22
20
20
20
Landfills
58
22
12
11
10
10
10
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
a Miscellaneous includes TSDFs (Treatment, Storage, and Disposal Facilities under the Resource Conservation
and Recovery Act [42 U.S.C. § 6924, SWDA § 3004]) and other waste categories.
Methodology
Emission estimates for 1990 through 2019 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2020) and disaggregated based on EPA (2003). Emission
estimates of these gases were provided by sector, using a "top down" estimating procedure—emissions were
calculated either for individual sources or for many sources combined, using basic activity data (e.g., the amount of
raw material processed) as an indicator of emissions. National activity data were collected for individual categories
from various agencies. Depending on the category, these basic activity data may include data on production, fuel
deliveries, raw material processed, etc.
Uncertainty and Time-Series Consistency
No quantitative estimates of uncertainty were calculated for this source category. Methodological recalculations
were applied to the entire time series to ensure time-series consistency from 1990 through 2019. Details on the
emission trends through time are described in more detail in the Methodology section, above.
7-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
The United States does not report any greenhouse gas emissions under the Intergovernmental Panel on Climate
Change (IPCC) "Other" sector.
Other 8-1

-------
3, Recalculations arid Improvements
Each year, many emission and sink estimates in the Inventory of U.S. Greenhouse Gas Emissions and Sinks are
recalculated and revised, as efforts are made to improve the estimates through the use of better methods and/or
data with the goal of improving inventory quality and reducing uncertainties, including the transparency,
completeness, consistency, and overall usefulness of the report. In this effort, the United States follows the 2006
IPCC Guidelines (IPCC 2006), which states, "Both methodological changes and refinements over time are an
essential part of improving inventory quality. It is good practice to change or refine methods when available data
have changed; the previously used method is not consistent with the IPCC guidelines for that category; a category
has become key; the previously used method is insufficient to reflect mitigation activities in a transparent manner;
the capacity for inventory preparation has increased; improved inventory methods become available; and/or for
correction of errors."
In general, when methodological changes have been implemented, the previous Inventory's time series (i.e., 1990
to 2018) will be recalculated to reflect the change, per guidance in IPCC (2006). Changes in historical data are
generally the result of changes in statistical data supplied by other agencies, and do not necessarily impact the
entire time series.
The results of all methodological changes and historical data updates made in the current Inventory are presented
in Figure 9-1, Table 9-1, and Table 9-2. Figure 9-1 presents the impact of recalculations by sector and on net total
emissions across the timeseries. Table 9-1 summarizes the quantitative effect of all changes on U.S. greenhouse
gas emissions by gas across the Energy, Industrial Processes and Product Use (IPPU), Agriculture, and Waste
sectors, while Table 9-2 summarizes the quantitative effect of changes on annual net fluxes from Land Use, Land-
Use Change, and Forestry (LULUCF). Both the figure and tables present results relative to the previously published
Inventory (i.e., the 1990 to 2018 report) in units of million metric tons of carbon dioxide equivalent (MMT C02 Eq.),
To understand the details of any specific recalculation or methodological improvement, see the Recalculations
within each source/sink categories' section found in Chapters 3 through 7 of this report. A discussion of Inventory
improvements in response to review processes is described in Annex 8.
The Inventory includes new categories not included in the previous Inventory that improve completeness of the
national estimates. Specifically, the current report includes methane emissions from anaerobic digestion at biogas
facilities, N20 emissions from industrial wastewater, CF4 emissions from Low Voltage Anode Effect (LVAE) during
aluminum production and change in carbon stocks for belowground biomass in managed coastal wetlands.
The following source and sink categories underwent the most significant methodological and historical data
changes. A brief summary of the recalculations and/or improvements undertaken are provided for these
categories.
• Forest Land Remaining Forest Land: Changes in Forest Carbon Stocks (C02). The methods used in the
current Inventory to compile estimates for forest ecosystem carbon stocks and stock changes and
harvested wood products (HWPs) from 1990 through 2019 are consistent with those used in the 1990
through 2018 Inventory. However, population estimates of carbon stocks and stock changes were
compiled using NFI data from each U.S. state and national estimates were compiled by summing over all
states. In past Inventories, population estimates were compiled by four geographic regions and summed
over all regions to compile national estimates. Also, the state-level disaggregation contributed to
Recalculations and Improvements 9-1

-------
identifying an error in the compilation of the Alaska time series data resulting in a 1-year misalignment in
carbon stock changes for this state in comparison to the 1990 through 2018 Inventory. This error has been
corrected resulting in differences in each year of the time series (i.e., 1990 to 2018), given the one-year
misalignment, with substantial differences in major fire years in Alaska. Soil carbon stocks decreased in
the latest Inventory relative to the previous Inventory and this change can be attributed to refinements in
the Digital General Soil Map of the United States (STATSG02) dataset where soil orders may have changed
in the updated data product. These changes resulted in an average annual increase in C stock change
losses of 42.7 MMT C02 Eq. (6.5 percent), across the 1990 through 2018 time series, relative to the
previous Inventory.
•	Wastewater Treatment (N20). EPA revised the domestic wastewater N20 methodology based on the 2019
Refinement (IPCC 2019): added emission estimates from septic systems; added a correction factor to
account for nitrogen from household products to POTWs and septic systems (1.17); revised the
methodology for treatment plants to account for aerobic and anaerobic treatment systems; updated the
emission factor for centralized aerobic systems (from 0 to 0.016 kg N20-N/kg N); and revised emission
estimates from discharge of domestic wastewater to aquatic environments to account for the condition of
the receiving waterbody (i.e., nutrient-impacted/eutrophic conditions, or not impacted) (ERG 2020). EPA
added industrial wastewater N20 emissions for the first time based on the 2019 Refinement (IPCC 2019)
methodology. These additions are on average 2 percent of wastewater N20 emissions across the entire
time series. The changes to domestic and industrial wastewater affected the time series from 1990
through 2018. Nitrous oxide emissions from wastewater increased an average 435 percent over the time
series, with the smallest increase of 15.3 MMT C02 Eq. (453 percent) in 1990 and largest increase of 21.4
MMT C02 Eq. (412 percent) in 2017.
•	Land Converted to Forest Land: Changes in Carbon Stocks (C02). The Land Converted to Forest Land
estimates in this Inventory are based on the land use change information in the annual National Forest
Inventory (NFI). All conversions are based on empirical estimates compiled using plot remeasurements
from the NFI, IPCC (2006) default biomass C stocks removed from Croplands and Grasslands in the year of
conversion on individual plots and the Tier 2 method for estimating mineral soil C stock changes (Ogle et
al. 2003, 2006; IPCC 2006). The incorporation of the most recent annual NFI data into the compilation
system resulted in a decrease in C stock changes. Overall, the Land Converted to Forest Land C stock
changes decreased by an average of 10.4 percent (11.5 MMT C02 Eq.) over the time series.
•	Non-Energy Use of Fuels (C02). Adjustments were made to activity data, carbon content coefficients, and
heat contents for hydrocarbon gas liquids (HGL) for 1990 to 2018. In previous Inventories, HGL activity
data from 1990 to 2007 were extracted from the American Petroleum Institute's Sales of Natural Gas
Liquids and Liquefied Refinery Gases. Historical HGL activity data from 1990 to 2007 were adjusted to use
ElA's Petroleum Supply Annual tables for consistency with the rest of the time series (i.e., 2008 to 2019).
In addition, the HGL carbon content coefficient for NEU was updated by separating each fuel out by its
natural gas liquid (NGL) and associated olefin to calculate a more accurate and annually variable factor,
and the heat contents for HGL and pentanes plus were updated using updated data from ElA's Monthly
Energy Review (EIA 2020). Non-energy use of petroleum coke consumption was adjusted to account for
leap years when converting from barrels per day to barrels per year. The "miscellaneous products"
category reported by EIA includes miscellaneous products that are not reported elsewhere in the EIA data
set. The miscellaneous products category reported by EIA was assumed to be mostly petroleum refinery
sulfur compounds that do not contain carbon (EIA 2019). Therefore, the carbon content for miscellaneous
products was updated to be zero across the time series. Overall, these changes resulted in an average
annual decrease of 10.9 MMT C02 Eq. (8.7 percent) in carbon emissions from non-energy uses of fossil
fuels for the period 1990 through 2018, relative to the previous Inventory. This decrease is primarily due
to the removal of miscellaneous products, which previously constituted an average of 8.2 percent of total
emissions from 1990 to 2018.
•	Natural Gas Systems (CH4). EPA received information and data related to the Inventory emission
estimates through GHGRP reporting, the annual Inventory formal public notice periods, stakeholder
9-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
feedback on updates under consideration, and new studies. EPA thoroughly evaluated relevant
information available and made several updates to the Inventory, including using revised emission factors
and produced water volumes to calculate produced water emissions (production segment), and using GTI
(2019) along with GTI 2009 study data to calculate customer meter emissions (distribution segment). In
addition, certain sources did not undergo methodological updates, but CH4 estimates changed due to
GHGRP data submission revisions and revisions to other input activity data sets. Overall, the changes
resulted in an average annual increase of 6.6 MMT C02 Eq. (4.2 percent) in methane emissions from
natural gas systems over the time series.
•	Fossil Fuel Combustion (C02). The EIA (2020c) updated energy consumption statistics across the time
series relative to the previous Inventory. As a result of revised natural gas heat contents, EIA updated
natural gas consumption in the residential, commercial, and industrial sectors for 2018. Approximate heat
rates for electricity and the heat content of electricity were revised for natural gas and noncombustible
renewable energy, which impacted electric power energy consumption by sector. EIA also revised sector
allocations for distillate fuel oil, residual fuel oil, and kerosene for 2018, and for propane for 2010 through
2012, 2014, 2017, and 2018, which impacted LPG by sector. EIA revised product supplied totals for crude
oil and petroleum products, which impacted the nonfuel sequestration statistics, particularly for
lubricants for 2018 and LPG for 2010 through 2018 relative to the previous Inventory. These changes
resulted in an average annual decrease 6.4 MMT C02 Eq. (0.1 percent) in C02 emissions from fossil fuel
combustion for the period 1990 through 2018, relative to the previous Inventory.
•	Wastewater Treatment (CH4). EPA revised the domestic wastewater CH4 methodology based on the 2019
Refinement (IPCC 2019): added a correction factor to account for organics from industrial and commercial
contributions to publicly owned treatment works (POTWs) (1.25); updated the emission factor for
centralized aerobic systems which accounts for loss of dissolved methane formed with in the collection
system (from 0 to 0.018 kg CH4/kg BOD); revised the estimate of organics removed with sludge from
POTWs; added emission estimates from discharge of domestic wastewater to aquatic environments based
on type of receiving water (e.g., reservoir, lake, estuaries); and updated wastewater treatment activity
data to align with the updates to organics removed and emissions from discharge to aquatic
environments (ERG 2020). Domestic wastewater treatment and discharge CH4 emissions increased an
average of 43 percent over the time series. The industrial wastewater CH4 methodology was also revised
based on the based on the 2019 Refinement (IPCC 2019) as described in Chapter 7, and contributed to
smaller recalculation impacts, i.e., averaging a 7.7 percent increase over the time series. The changes to
domestic and industrial wastewater affected the time series from 1990 through 2018. These changes
resulted in an average annual increase of 4.6 MMT C02 Eq. (30.1 percent) in methane emissions from
wastewater treatment across the time series.
•	Gasoline and Diesel Fuel Fossil Fuel Combustion (C02). EPA revised distillate fuel oil and motor gasoline
carbon contents, which impacted petroleum emissions in the transportation, residential, commercial, and
industrial sectors. The combined effect of both the diesel fuel and gasoline emission factor update was an
increase in emissions early in the time series and then decreases in emissions in more recent years. For
years 1990 through 2005, the average annual increase in total emissions was about 7 MMT C02 (0.1
percent of emissions). For the years 2006 to 2018 the average annual decrease in total emissions is about
5 MMT C02 (less than 0.1 percent of emissions).
•	Mobile Combustion (CH4). Updates were made to CH4 and N20 emission factors for newer non-road
gasoline and diesel vehicles. Previously, these emission factors were calculated using the updated IPCC
(2006) Tier 3 guidance and the nonroad component of EPA's MOVES2014b model. Methane emission
factors were calculated directly from MOVES. Updated emission factors were developed this year using
EPA engine certification data for non-road small and large spark-ignition (SI) gasoline engines and
compression-ignition diesel engines (2011 and newer), as well as non-road motorcycles (2006 and newer),
SI marine engines (2011 and newer), and diesel marine engines (2000 and newer). The result of these
changes was a net decrease in CH4 emissions from mobile combustion relative to the previous Inventory.
Recalculations and Improvements 9-3

-------
Methane emissions from mobile combustion decreased by an average of 4.5 MMT C02 Eq. (47.5 percent)
throughout the time series.
• Mobile Combustion (N20). Updates were made to CH4 and N20 emission factors for newer non-road
gasoline and diesel vehicles. Previously, these emission factors were calculated using the updated IPCC
(2006) Tier 3 guidance and the nonroad component of EPA's MOVES2014b model. Nitrous oxide emission
factors are calculated using MOVES-Nonroad activity and emission factors in g/kWh by fuel type from the
European Environment Agency. Updated emission factors were developed this year using EPA engine
certification data for non-road small and large spark-ignition (SI) gasoline engines and compression-
ignition diesel engines (2011 and newer), as well as non-road motorcycles (2006 and newer), SI marine
engines (2011 and newer), and diesel marine engines (2000 and newer). The result of these changes was
an increase in N20 emissions from mobile combustion relative to the previous Inventory. Nitrous oxide
emissions from mobile combustion increased by an average of 3.5 MMT C02 Eq. (11.3 percent)
throughout the time series.
Figure 9-1: Impacts from Recalculations to U.S. Greenhouse Gas Emissions by Sector
¦	Net Total Emissions
¦	Agriculture
Energy
Industrial Processes and Product Use
¦	Net COz Flux from LULUCF
Waste
9-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Table 9-1: Revisions to U.S. Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Source
1990
2005
2015
2016
2017
2018
Average
Annual
Change
CO?
(14.8)
2.6
(40.7)
(44.2)
(45.9)
(49.4)
(17.0)
Fossil Fuel Combustion
(8.5)
12.8
(23.5)
(30.9)
(37.8)
(40.4)
(6.4)
Electric Power Sector
NC
0.1
+
+
+
0.1
+
Transportation
+
2.5
(6.0)
(5.4)
(4.8)
(4.1)
1.4
Industrial
(3.2)
2.8
(4.0)
(8.9)
(14.9)
(19.6)
(1.9)
Residential
0.4
1.0
(0.5)
(0.4)
(0.4)
0.8
0.3
Commercial
0.1
0.3
(0.8)
(0.8)
(0.8)
(0.8)
(0.1)
U.S. Territories
(5.8)
6.2
(12.1)
(15.4)
(16.8)
(16.8)
(6.1)
Non-Energy Use of Fuels
(6.8)
(10.6)
(18.6)
(13.8)
(9.6)
(4.8)
(10.9)
Natural Gas Systems
(0.1)
(0.1)
(0.2)
0.2
0.8
(1.1)
(0.1)
Cement Production
NC
NC
NC
NC
NC
(1.4)
+
Lime Production
NC
NC
NC
+
+
(0.1)
+
Other Process Uses of Carbonates
NC
NC
NC
0.5
+
(2.5)
(0.1)
Glass Production
NC
NC
NC
+
+
+
+
Soda Ash Production
NC
NC
NC
NC
NC
NC
NC
Carbon Dioxide Consumption
NC
NC
0.5
0.2
0.1
(0.3)
+
Incineration of Waste
0.1
0.2
0.8
0.6
0.4
0.4
0.3
Titanium Dioxide Production
NC
NC
NC
NC
NC
NC
NC
Aluminum Production
NC
NC
+
NC
NC
NC
+
Iron and Steel Production & Metallurgical Coke







Production
+
+
+
+
+
+
+
Ferroalloy Production
NC
NC
NC
NC
NC
NC
NC
Ammonia Production
NC
+
+
(0.6)
(2.1)
(1.4)
(0.1)
Urea Consumption for Non-Agricultural







Purposes
NC
NC
NC
NC
1.3
2.2
0.1
Phosphoric Acid Production
NC
NC
NC
NC
NC
NC
NC
Petrochemical Production
NC
NC
NC
NC
NC
(0.1)
+
Carbide Production and Consumption
+
+
+
+
+
+
+
Lead Production
NC
NC
NC
NC
NC
NC
NC
Zinc Production
NC
NC
+
(0.1)
(0.1)
+
+
Petroleum Systems
0.1
(0.1)
(0.2)
(1.1)
0.5
0.3
(0.1)
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Magnesium Production and Processing
NC
NC
NC
NC
NC
NC
NC
Liming
NC
NC
NC
NC
NC
(0.9)
+
Urea Fertilization
0.4
0.4
0.6
0.8
0.5
0.6
0.5
International Bunker Fuels0
NC
0.1
+
+
+
+
+
Wood Biomass, Ethanol, and Biodiesel







Consumption
NC
NC
NC
(0.6)
(9.9)
(9.3)
(0.7)
CH4c
2.5
6.6
13.1
18.1
18.1
21.5
7.8
Stationary Combustion
+
+
+
(0.1)
(0.2)
(0.1)
+
Mobile Combustion
(6.5)
(5.6)
(1.0)
(0.9)
(0.8)
(0.7)
(4.5)
Coal Mining
NC
+
+
+
+
+
+
Abandoned Underground Coal Mines
NC
NC
NC
NC
NC
NC
NC
Natural Gas Systems
3.6
6.1
8.0
11.5
9.5
12.6
6.6
Petroleum Systems
2.8
0.6
0.9
0.2
0.7
1.1
1.1
Abandoned Oil and Gas Wells
0.2
0.2
0.2
0.2
0.1
0.2
0.2
Petrochemical Production
NC
NC
NC
NC
NC
NC
NC
Carbide Production and Consumption
NC
NC
NC
NC
NC
NC
NC
Iron and Steel Production & Metallurgical Coke







Production
NC
NC
NC
NC
NC
NC
NC
Ferroalloy Production
NC
NC
NC
NC
NC
NC
NC
Recalculations and Improvements 9-5

-------
Enteric Fermentation
0.5
0.4
0.4
0.4
0.4
0.4
0.4
Manure Management
NC
NC
NC
NC
NC
NC
NC
Rice Cultivation
NC
NC
NC
2.3
2.1
2.2
0.2
Field Burning of Agricultural Residues
+
+
+
+
+
+
+
Landfills
(3.0)
0.1
0.1
+
1.7
1.5
(0.9)
Wastewater Treatment
4.8
4.6
4.3
4.3
4.3
4.2
4.6
Composting
NC
NC
NC
NC
NC
(0.2)
+
Anaerobic Digestion at Biogas Facilities
NC
NC
NC
NC
NC
NC
NC
Incineration of Waste
NC
NC
NC
NC
NC
NC
NC
International Bunker Fuelsa
NC
NC
NC
NC
NC
NC
NC
N2Oc
18.0
23.2
24.4
24.7
25.0
24.7
22.2
Stationary Combustion
+
+
+
+
(0.2)
(0.2)
+
Mobile Combustion
2.7
4.2
3.4
3.4
3.6
3.6
3.5
AdipicAcid Production
+
NC
NC
NC
NC
NC
+
Nitric Acid Production
NC
NC
NC
NC
NC
0.2
+
Manure Management
NC
NC
NC
NC
NC
NC
NC
Agricultural Soil Management
+
0.3
0.4
0.3
0.2
0.1
0.2
Field Burning of Agricultural Residues
+
+
+
+
+
+
+
Wastewater Treatment
15.3
18.6
20.6
21.0
21.4
21.1
18.5
N20 from Product Uses
NC
NC
NC
NC
NC
NC
NC
Caprolactam, Glyoxal, and Glyoxylic Acid







Production
NC
NC
NC
NC
NC
NC
NC
Incineration of Waste
NC
NC
NC
NC
NC
NC
NC
Composting
NC
NC
NC
NC
NC
(0.2)
+
Electronics Industry
NC
NC
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
Petroleum Systems
+
+
+
+
+
+
+
International Bunker Fuels0
NC
NC
NC
NC
NC
NC
NC
HFCs, PFCs, SF6and NF3
+
(1.1)
(2.1)
(2.4)
(2.2)
(2.0)
(0.9)
HFCs
NC
(1.1)
(2.2)
(2.4)
(2.2)
(1.8)
(1.0)
Substitution of Ozone Depleting Substancesd
NC
(1.1)
(2.2)
(2.4)
(2.2)
(1.8)
(1.0)
HCFC-22 Production
NC
NC
NC
NC
NC
NC
NC
Electronics Industry
NC
+
+
+
+
+
+
Magnesium Production and Processing
NC
NC
NC
NC
+
+
+
PFCs
NC
+
0.1
0.1
0.1
0.1
0.1
Aluminum Production
NC
NC
0.1
0.1
0.1
0.1
0.1
Electronics Industry
NC
+
+
+
+
+
+
Substitution of Ozone Depleting Substancesd
NC
NC
NC
NC
NC
NC
NC
sf6
NC
+
+
+
+
(0.2)
+
Electrical Transmission and Distribution
NC
NC
+
+
+
(0.2)
+
Electronics Industry
NC
+
+
+
+
+
+
Magnesium Production and Processing
NC
NC
NC
NC
(0.1)
(0.1)
+
nf3
NC
+
+
+
+
+
+
Electronics Industry
NC
+
+
+
+
+
+
Unspecified Mix of HFCs, NF3, PFCs and SF6
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Net Emissions (Sources and Sinks)
(41.8)
57.9
6.4
(57.6)
(7.1)
(33.1)
19.8
Percent Change
-0.7%
0.9%
0.1%
-1.0%
-0.1%
-0.6%
-0.4%
Notes: Net change in total emissions presented without LULUCF. Totals may not sum due to independent rounding.
NC (No Change)
+ Absolute value does not exceed 0.05 MMT C02 Eq. or 0.05 percent.
* Indicates a new source for the current Inventory year. Emissions from new sources are captured in net emissions and
percent change totals.
a Emissions from International Bunker Fuels are not included in totals.
9-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
b Emissions from Wood Biomass, Ethanol, and Biodiesel Consumption are not included specifically in summing Energy sector
totals. Net carbon fluxes from changes in biogenic carbon reservoirs are accounted for in the estimates for Land Use, Land-
Use Change, and Forestry.
c LULUCF emissions of CH4 and N20 are reported separately from gross emissions totals in Table 9-2. L LULUCF emissions
include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained Organic Soils,
Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to Coastal
Wetlands; and N20 emissions from Forest Soils and Settlement Soils.d Small amounts of PFC emissions also result from this
source.
Table 9-2: Revisions to U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land
Use, Land-Use Change, and Forestry (MMT CO2 Eq.)
Land-Use Category
1990
2005
2015
2016
2017
2018
Average
Annual
Change
Forest Land Remaining Forest Land
(53.7)
17.1
4.8
(64.0)
(12.5)
(38.5)
(42.8)
Changes in Forest Carbon Stocks3
(53.7)
17.1
4.7
(64.0)
(12.0)
(35.4)
(42.7)
Non-C02 Emissions from Forest Firesb
+
+
+
+
(0.5)
(3.1)
(0.1)
N20 Emissions from Forest Soilsc
NC
NC
NC
NC
NC
NC
NC
Non-C02 Emissions from Drained Organic







Soilsd
NC
NC
NC
NC
NC
NC
NC
Land Converted to Forest Land
11.3
11.5
11.6
11.6
11.5
11.5
11.5
Changes in Forest Carbon Stocks6
11.3
11.5
11.6
11.6
11.5
11.5
11.5
Cropland Remaining Cropland
NC
NC
NC
+
+
+
+
Changes in Mineral and Organic Soil







Carbon Stocks
NC
NC
NC
+
+
+
+
Land Converted to Cropland
(2.3)
(1.7)
(1.1)
(1.1)
(1.1)
(1.1)
(1.7)
Changes in all Ecosystem Carbon Stocks'
(2.3)
(1.7)
(1.1)
(1.1)
(1.1)
(1.1)
(1.7)
Grassland Remaining Grassland
(0.8)
(0.7)
(0.5)
0.2
0.4
0.5
(0.6)
Changes in Mineral and Organic Soil







Carbon Stocks
(0.8)
(0.7)
(0.5)
0.2
0.4
0.5
(0.6)
Non-C02 Emissions from Grassland Fires8
NC
NC
+
+
+
+
+
Land Converted to Grassland
0.4
0.2
(0.8)
0.7
0.5
0.5
0.2
Changes in all Ecosystem Carbon Stocks'
0.4
0.2
(0.8)
0.7
0.5
0.5
0.2
Wetlands Remaining Wetlands
(3.0)
(0.6)
(3.4)
(3.4)
(3.3)
(3.3)
(1.3)
Changes in Organic Soil Carbon Stocks in







Peatlands
NC
NC
NC
NC
0.1
0.1
+
Changes in Biomass, DOM, and Soil







Carbon Stocks in Coastal Wetlands
(3.4)
(0.9)
(3.7)
(3.7)
(3.7)
(3.6)
(1.6)
CH4 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
0.3
0.3
0.2
0.2
0.2
0.2
0.3
N20 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
NC
NC
+
+
+
+
+
Non-C02 Emissions from Peatlands







Remaining Peatlands
NC
NC
NC
NC
+
+
+
Land Converted to Wetlands
0.7
0.7
0.2
0.2
0.2
0.2
0.6
Changes in Biomass, DOM, and Soil







Carbon Stocks
0.5
0.5
+
+
+
+
0.4
CH4 Emissions from Land Converted to







Coastal Wetlands
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Settlements Remaining Settlements
NC
NC
0.9
1.9
2.2
2.3
0.3
Changes in Organic Soil Carbon Stocks
NC
NC
NC
+
+
+
+
Changes in Settlement Tree Carbon







Stocks
NC
NC
NC
NC
NC
NC
NC
Changes in Yard Trimming and Food Scrap







Carbon Stocks in Landfills
NC
NC
0.9
1.9
2.2
2.3
0.3
Recalculations and Improvements 9-7

-------
N20 Emissions from Settlement Soilsh
NC
NC
NC
NC
NC
NC
NC
Land Converted to Settlements
NC
NC
NC
+
+
+
+
Changes in all Ecosystem Carbon Stocks'
NC
NC
NC
+
+
+
+
LULUCF Total Net Flux1
(48.0)
26.1
11.2
(54.3)
(2.0)
(25.3)
(34.3)
LULUCF Emissions1
0.6
0.5
0.5
0.4
(0.1)
(2.7)
0.4
LULUCF Sector Totalk
(47.4)
26.6
11.7
(53.9)
(2.1)
(27.9)
(33.9)
Percent Change
-5.6%
3.3%
1.5%
-6.8%
-0.3%
-3.6%
-4.3%
Note: Totals may not sum due to independent rounding.
NC (No Change)
+ Absolute value does not exceed 0.05 MMT C02 Eq. or 0.05 percent.
a Includes the net changes to carbon stocks stored in all forest ecosystem pools and harvested wood products.
b Estimates include CH4 and N20 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
c Estimates include N20 emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
d Estimates include CH4 and N20 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land
Converted to Forest Land.
0	Includes the net changes to carbon stocks stored in all forest ecosystem pools.
f Includes changes in mineral and organic soil carbon stocks for all land use conversions to cropland, grassland, and
settlements, respectively. Also includes aboveground/belowground biomass, dead wood, and litter carbon stock changes
for conversion of forest land to cropland, grassland, and settlements, respectively.
5 Estimates include CH4 and N20 emissions from fires on both Grassland Remaining Grassland and Land Converted to
Grass/and.
h Estimates include N20 emissions from N fertilizer additions on both Settlements Remaining Settlements and Land
Converted to Settlements because it is not possible to separate the activity data at this time.
1	LULUCF Carbon Stock Change includes any C stock gains and losses from all land use and land use conversion categories.
> LULUCF emissions include the CH4 and N20 emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions from Land Converted to
Coastal Wetlands; and N20 emissions from Forest Soils and Settlement Soils.
k The LULUCF Sector Net Total is the net sum of all LULUCF CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
9-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
10. References
Executive Summary
BEA (2020) 2020 Comprehensive Revision of the National Income and Product Accounts: Current-dollar and "real"
GDP, 1929-2020. Bureau of Economic Analysis (BEA), U.S. Department of Commerce, Washington, D.C. Available
online at: .
Duffield, J. (2006) Personal communication. Jim Duffield, Office of Energy Policy and New Uses, U.S. Department of
Agriculture, and Lauren Flinn, ICF International. December 2006.
EIA (2020a) Electricity Generation. Monthly Energy Review, November 2020. Energy Information Administration,
U.S. Department of Energy, Washington, D.C. DOE/EIA-0035(2019/11).
EIA (2020b) Electricity in the United States. Electricity Explained. Energy Information Administration, U.S.
Department of Energy, Washington, D.C. Available online at:
.
EIA (2019) International Energy Statistics 1980-2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EPA (2020a) Acid Rain Program Dataset 1996-2019. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2020b) Greenhouse Gas Reporting Program (GHGRP). 2019 Envirofacts. Subpart HH: Municipal Solid Waste
Landfills and Subpart TT: Industrial Waste Landfills. Available online at: .
EPA (2020c) "1970 - 2019 Average annual emissions, all criteria pollutants in MS Excel." National Emissions
Inventory (NEI) Air Pollutant Emissions Trends Data. Office of Air Quality Planning and Standards, April 2020.
Available online at: .
EPA (1997) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC. October 1997.
FHWA (1996 through 2019) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Available online at:
.
IEA (2020) C02 Emissions from Fossil Fuel Combustion - Overview. International Energy Agency. Available online
at: .
References 10-1

-------
IPCC (2013) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change. [Stocker, T.F., D. Qin, G.-K., Plattner, M.
Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 1535 pp.
IPCC (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. [S. Solomon, D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC (1996) Climate Change 1995: The Science of Climate Change. Intergovernmental Panel on Climate Change.
[J.T. Houghton, LG. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg, and K. Maskell (eds.)]. Cambridge
University Press. Cambridge, United Kingdom.
National Academies of Sciences, Engineering, and Medicine (2018) Improving Characterization of Anthropogenic
Methane Emissions in the United States. Washington, DC: The National Academies Press. Available online at:
.
National Research Council (2010) Verifying Greenhouse Gas Emissions: Methods to Support International Climate
Agreements. Washington, DC: The National Academies Press. Available online at:
.
NOAA/ESRL (2021a) Trends in Atmospheric Carbon Dioxide. Available online at:
. 05 January 2021.
NOAA/ESRL (2021b) Trends in Atmospheric Methane. Available online at:
. 05 January 2021.
NOAA/ESRL (2021c) Nitrous Oxide (N20) hemispheric and global monthly means from the NOAA/ESRL
Chromatograph for Atmospheric Trace Species data from baseline observatories (Barrow, Alaska; Summit,
Greenland; Niwot Ridge, Colorado; Mauna Loa, Hawaii; American Samoa; South Pole). Available online at:
. 05 January 2021.
UNFCCC (2014) Report of the Conference of the Parties on its Nineteenth Session, Held in Warsaw from 11 to 23
November 2013. (FCCC/CP/2013/10/Add.3). January 31, 2014. Available online at:
.
U.S. Census Bureau (2020) U.S. Census Bureau International Database (IDB). Available online at:
.
Introduction
IPCC (2014) Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change [Edenhofer, O., R. Pichs-Madruga, Y.
Sokona, J. Minx, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, J.
Savolainen, S. Schlomer, C. von Stechow, and T. Zwickel (eds.)]. Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, 1435 pp.
IPCC (2013) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M.
Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 1535 pp.
10-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
IPCC (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. [S. Solomon, D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC (2001) Climate Change 2001: The Scientific Basis. Intergovernmental Panel on Climate Change. [J.T. Houghton,
Y. Ding, D.J. Griggs, M. Noguer, P.J. van der Linden, X. Dai, C.A. Johnson, and K. Maskell (eds.)]. Cambridge
University Press. Cambridge, United Kingdom.
IPCC/TEAP (2005) Special Report: Safeguarding the Ozone Layer and the Global Climate System, Chapter 4:
Refrigeration. 2005. Available online at: .
NOAA (2017) Vital Signs of the Planet. Available online at: . Accessed on 9
January 2017.
NOAA/ESRL (2021a) Trends in Atmospheric Carbon Dioxide. Available online at:
. January 12, 2021.
NOAA/ESRL (2021b) Trends in Atmospheric Methane. Available online at:
. January 5, 2021.
NOAA/ESRL (2021c) Nitrous Oxide (N20) hemispheric and global monthly means from the NOAA/ESRL
Chromatograph for Atmospheric Trace Species data from baseline observatories (Barrow, Alaska; Summit,
Greenland; Niwot Ridge, Colorado; Mauna Loa, Hawaii; American Samoa; South Pole). Available online at:
. January 5, 2021.
NOAA/ESRL (2021d) Sulfur Hexafluoride (SF6) hemispheric and global monthly means from the NOAA/ESRL
Chromatograph for Atmospheric Trace Species data from baseline observatories (Barrow, Alaska; Summit,
Greenland; Niwot Ridge, Colorado; Mauna Loa, Hawaii; American Samoa; South Pole). Available online at:
. January 5, 2021.
UNEP/WMO (1999) Information Unit on Climate Change. Framework Convention on Climate Change. Available
online at: .
UNFCCC (2014) Report of the Conference of the Parties on its nineteenth session, held in Warsaw from 11 to 23
November 2013. (FCCC/CP/2013/10/Add.3). January 31, 2014. Available online at:
.
USGCRP (2017) Climate Science Special Report: Fourth National Climate Assessment, Volume I. [Wuebbles, D.J.,
D.W. Fahey, K.A. Hibbard, D.J. Dokken, B.C. Stewart, and T.K. Maycock (eds.)]. U.S. Global Change Research
Program, Washington, DC, USA, 470 pp, doi: 10.7930/J0J964J6. Available online at:
.
WMO/UNEP (2014) Assessment for Decision-Makers: Scientific Assessment of Ozone Depletion: 2014. Available
online at: .
WMO (2015) "Is the Ozone Layer on the Mend? Highlights from the most recent WMO/UNDP Ozone Assessment"
Bulletin no. Vol (64)(1). Available online at: .
References 10-3

-------
Trends in Greenhouse Gas Emissions
BEA (2020) 2019 Comprehensive Revision of the National Income and Product Accounts: Current-dollar and "real"
GDP, 1929-2019. Bureau of Economic Analysis (BEA), U.S. Department of Commerce, Washington, D.C. Available
online at: .
Duffield, J. (2006) Personal communication. Jim Duffield, Office of Energy Policy and New Uses, U.S. Department of
Agriculture, and Lauren Flinn, ICF International. December 2006.
EIA (2020a) Fuel Oil and Kerosene Sales. Energy Information Administration, U.S. Department of Energy,
Washington, D.C. February 2020.
EIA (2020b) Monthly Energy Review, November 2020. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2020/11).
EIA (2018) "In 2017, U.S. electricity sales fell by the greatest amount since the recession" Available online at:
.
EPA (2020a) Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 - 2019.
Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Available online at:
.
EPA (2020b) 1970 - 2019 Average annual emissions, all criteria pollutants in MS Excel. National Emissions Inventory
(NEI) Air Pollutant Emissions Trends Data. Office of Air Quality Planning and Standards, May 2020. Available online
at: .
IPCC (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. [S. Solomon, D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
U.S. Census Bureau (2020) U.S. Census Bureau International Database (IDB). Available online at:
.
USDA (2019) Personal communication. Claudia Hitaj, USDA Economic Research Service, and Vincent Camobreco,
U.S. EPA. September 2019.
Energy
EIA (2020) Monthly Energy Review, November 2020, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/11).
IEA (2020) Energy related C02 emissions, 1990-2019, International Energy Agency, Paris. Available online at:
.
Carbon Dioxide Emissions from Fossil Fuel Combustion
AAR (2008 through 2019) Railroad Facts. Policy and Economics Department, Association of American Railroads,
Washington, D.C. Available online at 
AISI (2004 through 2019) Annual Statistical Report, American Iron and Steel Institute, Washington, D.C.
10-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
APTA (2007 through 2018) Public Transportation Fact Book. American Public Transportation Association,
Washington, D.C. Available online at: .
APTA (2006) Commuter Rail National Totals. American Public Transportation Association, Washington, D.C.
BEA (2018) Table 1.1.6. Real Gross Domestic Product, Chained 2012 Dollars. Bureau of Economic Analysis (BEA),
U.S. Department of Commerce, Washington, D.C. September 2018. Available online at:
.
Benson, D. (2002 through 2004) Unpublished data. Upper Great Plains Transportation Institute, North Dakota State
University and American Short Line & Regional Railroad Association.
Browning, L. (2020). GHG Inventory EF Development Using Certification Data. Technical Memo, September 2020.
Browning (2019) Updated On-highway CH4 and N20 Emission Factors for GHG Inventory. Memorandum from ICF to
Sarah Roberts, Office of Transportation and Air Quality, U.S. Environmental Protection Agency. September 2019.
Browning, L. (2018a). Updated Methodology for Estimating Electricity Use from Highway Plug-In Electric Vehicles.
Technical Memo, October 2018.
Browning, L. (2018b). Updated Non-Highway CH4 and N20 Emission Factors for U.S. GHG Inventory. Technical
Memo, November 2018.
Browning, L. (2017) Updated Methodology for Estimating CH4 and N20 Emissions from Highway Vehicle Alternative
Fuel Vehicles. Technical Memo, October 2017.
Coffeyville Resources Nitrogen Fertilizers (2012) Nitrogen Fertilizer Operations. Available online at:
.
Dakota Gasification Company (2006) C02 Pipeline Route and Designation Information. Bismarck, ND.
DHS (2008) Email Communication. Elissa Kay, Department of Homeland Security and Joe Aamidor, ICF
International. January 11, 2008.
DLA Energy (2020) Unpublished data from the Fuels Automated System (FAS). Defense Logistics Agency Energy,
U.S. Department of Defense. Washington, D.C.
DOC (1991 through 2019) Unpublished Report of Bunker Fuel Oil Laden on Vessels Cleared for Foreign Countries.
Form-563. Foreign Trade Division, Bureau of the Census, U.S. Department of Commerce. Washington, D.C.
DOE (1993 through 2019) Transportation Energy Data Book. Office of Transportation Technologies, Center for
Transportation Analysis, Energy Division, Oak Ridge National Laboratory. ORNL-6978.
DOE (2012) 2010 Worldwide Gasification Database. National Energy Technology Laboratory and Gasification
Technologies Council. Available online at:
. Accessed on 15
March 2012.
DOT (1991 through 2019) Airline Fuel Cost and Consumption. U.S. Department of Transportation, Bureau of
Transportation Statistics, Washington, D.C. DAI-10. Available online at: .
Eastman Gasification Services Company (2011) Project Data on Eastman Chemical Company's Chemicals-from-Coal
Complex in Kingsport, TN.
EIA (2020a) Monthly Energy Review, November 2020, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/11).
EIA (2020b) Quarterly Coal Report: April-June 2020. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0121.
EIA (2020c) Form EIA-923 detailed data with previous form data (EIA-906/920), Energy Information Administration,
U.S. Department of Energy. Washington, DC. DOE/EIA.
References 10-5

-------
EIA (2020d) "Natural gas prices, production, consumption, and exports increased in 2019." Today in Energy.
Available online at: .
EIA (2020e) Electric Power Annual 2019. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. DOE/EIA-O348(17).
EIA (2020f) Natural Gas Annual 2019. Energy Information Administration, U.S. Department of Energy. Washington,
D.C. DOE/EIA-O131(17).
EIA (2020g) Annual Coal Report 2019. Energy Information Administration, U.S. Department of Energy. Washington,
D.C. DOE/EIA-0584.
EIA (2020) Alternative Fuels Data Tables. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. Available online at: .
EIA (2018) "Both natural gas supply and demand have increased from year-ago levels." Today in Energy. Available
online at: .
EIA (2020) International Energy Statistics 1980-2017. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EIA (1991 through 2019) Fuel Oil and Kerosene Sales. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EIA (2009a) Emissions of Greenhouse Gases in the United States 2008, Draft Report. Office of Integrated Analysis
and Forecasting, Energy Information Administration, U.S. Department of Energy. Washington, D.C. DOE-EIA-
0573(2009).
EIA (2009b) Manufacturing Consumption of Energy (MECS) 2006. U.S. Department of Energy, Energy Information
Administration, Washington, D.C. Released July 2009.
EIA (2008) Historical Natural Gas Annual, 1930 - 2008. Energy Information Administration, U.S. Department of
Energy. Washington, D.C.
EIA (2007) Personal Communication. Joel Lou, Energy Information Administration and Aaron Beaudette, ICF
International. Residual and Distillate Fuel Oil Consumption for Vessel Bunkering (Both International and Domestic)
for American Samoa, U.S. Pacific Islands, and Wake Island. October 24, 2007.
EIA (2003) Personal Communication. Kent Forsberg, Energy Information Administration and ICF International.
Distillate Fuel Oil Consumption.
EIA (2001) U.S. Coal, Domestic and International Issues. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. March 2001.
EIA (1990-2001) State Energy Data System. Energy Information Administration, U.S. Department of Energy.
Washington, D.C.
EPA (2021) Acid Rain Program Dataset 1996-2018. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2020c) EPA Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Updated Gasoline and Diesel
Fuel C02 Emission Factors - Memo.
EPA (2019a) Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 - 2018.
Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Available online at:
.
EPA (2019b) MOtor Vehicle Emissions Simulator (MOVES) 2014b. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency, Washington, D.C. Available online at: .
EPA (2018) The Emissions & Generation Resource Integrated Database (eGRID) 2016 Technical Support Document.
Clean Air Markets Division, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington,
10-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
D.C. Available Online at: .
EPA (2010) Carbon Content Coefficients Developed for EPA's Mandatory Reporting Rule. Office of Air and
Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
Erickson, T. (2003) Plains C02 Reduction (PCOR) Partnership. Presented at the Regional Carbon Sequestration
Partnership Meeting Pittsburgh, Pennsylvania, Energy and Environmental Research Center, University of North
Dakota. November 3, 2003.
FAA (2021) Personal Communication between FAA and John Steller, Mausami Desai, and Vincent Camobreco for
aviation emissions estimates from the Aviation Environmental Design Tool (AEDT). March 2021.
FHWA (1996 through 2019) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Available online at:
.
FHWA (2015) Off-Highway and Public-Use Gasoline Consumption Estimation Models Used in the Federal Highway
Administration, Publication Number FHWA-PL-17-012. Available online at:
.
Fitzpatrick, E. (2002) The Weyburn Project: A Model for International Collaboration.
FRB (2019) Industrial Production and Capacity Utilization. Federal Reserve Statistical Release, G.17, Federal
Reserve Board. Available online at: .
Gaffney, J. (2007) Email Communication. John Gaffney, American Public Transportation Association and Joe
Aamidor, ICF International. December 17, 2007.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Marland, G. and A. Pippin (1990) "United States Emissions of Carbon Dioxide to the Earth's Atmosphere by
Economic Activity." Energy Systems and Policy, 14(4):323.
SAIC/EIA (2001) Monte Carlo Simulations of Uncertainty in U.S. Greenhouse Gas Emission Estimates. Final Report.
Prepared by Science Applications International Corporation (SAIC) for Office of Integrated Analysis and Forecasting,
Energy Information Administration, U.S. Department of Energy. Washington, D.C. June 22, 2001.
U.S. Aluminum Association (USAA) (2008 through 2019) U.S. Primary Aluminum Production. U.S. Aluminum
Association, Washington, D.C.
USAF (1998) Fuel Logistics Planning. U.S. Air Force: AFPAM23-221. May 1,1998.
U.S. Census Bureau (2001 through 2011) Current Industrial Reports Fertilizer Materials and Related Products:
Annual Summary. Available online at: .
United States Geological Survey (USGS) (2020a) 2020 Mineral Commodity Summaries: Aluminum. U.S. Geological
Survey, Reston, VA.
USGS (2020b) 2020 Mineral Commodity Summary: Titanium and Titanium Dioxide. U.S. Geological Survey, Reston,
VA.
USGS (2014 through 2020a) Mineral Industry Surveys: Silicon. U.S. Geological Survey, Reston, VA.
USGS (2014 through 2020b) Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA.
USGS (2014 through 2019) Minerals Yearbook: Nitrogen [Advance Release], Available online at:
.
USGS (1991 through 2018) Minerals Yearbook - Iron and Steel Scrap. U.S. Geological Survey, Reston, VA.
References 10-7

-------
USGS (1991 through 2015a) Minerals Yearbook: Manufactured Abrasives Annual Report. U.S. Geological Survey,
Reston, VA. Available online at: .
USGS (1991 through 2015b) Minerals Yearbook: Titanium. U.S. Geological Survey, Reston, VA.
USGS (1991 through 2015c) Minerals Yearbook: Silicon Annual Report. U.S. Geological Survey, Reston, VA. Available
online at: .
USGS (1996 through 2013) Minerals Yearbook: Silicon. U.S. Geological Survey, Reston, VA.
USGS (1995 through 2013) Minerals Yearbook: Lead Annual Report. U.S. Geological Survey, Reston, VA.
USGS (1995,1998, 2000, 2001, 2002, 2007) Minerals Yearbook: Aluminum Annual Report. U.S. Geological Survey,
Reston, VA.
Stationary Combustion (excluding C02)
EIA (2020a) Monthly Energy Review, November 2020. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2020/11).
EIA (2020b) International Energy Statistics 1980-2017. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EPA (2020) EPA Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Updated Gasoline and Diesel
Fuel C02 Emission Factors - Memo.
EPA (2021) Acid Rain Program Dataset 1996-2019. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2019). Motor Vehicle Emissions Simulator (MOVES) 2014b. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency. Available online at: .
EPA (1997) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC. October 1997.
FHWA (1996 through 2019) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Obtained from Tiffany Presmy at FHWA.
ICF (2020) Potential Improvements to Energy Sector Hydrocarbon Gas Liquid Carbon Content Coefficients.
Memorandum from ICF to Vincent Camobreco, U.S. Environmental Protection Agency. December 7, 2020.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
SAIC/EIA (2001) Monte Carlo Simulations of Uncertainty in U.S. Greenhouse Gas Emission Estimates. Final Report.
Prepared by Science Applications International Corporation (SAIC) for Office of Integrated Analysis and Forecasting,
Energy Information Administration, U.S. Department of Energy. Washington, D.C. June 22, 2001.
Mobile Combustion (excluding C02)
AAR (2008 through 2019) Railroad Facts. Policy and Economics Department, Association of American Railroads,
Washington, D.C. Available online at .
ANL (2020) The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model (GREET2020).
Argonne National Laboratory. Available online at: .
ANL (2006) Argonne National Laboratory (2006) GREET model Version 1.7. June 2006.
APTA (2007 through 2019) Public Transportation Fact Book. American Public Transportation Association,
Washington, D.C. Available online at: .
10-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
APTA (2006) Commuter Rail National Totals. American Public Transportation Association, Washington, D.C.
Available online at: .
BEA (1991 through 2015) Unpublished BE-36 survey data. Bureau of Economic Analysis, U.S. Department of
Commerce. Washington, D.C.
Benson, D. (2002 through 2004) Personal communication. Unpublished data developed by the Upper Great Plains
Transportation Institute, North Dakota State University and American Short Line & Regional Railroad Association.
Browning, L. (2020). Updated Methane and Nitrous Oxide Emission Factors for Non-Road Sources and On-road
Motorcycles. Technical Memorandum from ICF International to Sarah Roberts, Office of Transportation and Air
Quality, U.S. Environmental Protection Agency, September 2020.
Browning, L. (2019) Updated On-highway CH4 and N20 Emission Factors for GHG Inventory. Memorandum from ICF
to Sarah Roberts and Justine Geidosch, Office of Transportation and Air Quality, U.S. Environmental Protection
Agency. September 2019.
Browning, L. (2018a). Updated Methodology for Estimating Electricity Use from Highway Plug-In Electric Vehicles.
Technical Memorandum from ICF International to Sarah Roberts and Justine Geidosch, Office of Transportation
and Air Quality, U.S. Environmental Protection Agency. October 2018.
Browning, L. (2018b) Updated Non-Highway CH4 and N20 Emission Factors for U.S. GHG Inventory. Technical
Memorandum from ICF International to Sarah Roberts and Justine Geidosch, Office of Transportation and Air
Quality, U.S. Environmental Protection Agency. November 2018.
Browning, L. (2017) Updated Methodology for Estimating CH4 and N20 Emissions from Highway Vehicle Alternative
Fuel Vehicles. Technical Memorandum from ICF International to Sarah Roberts and Justine Geidosch, Office of
Transportation and Air Quality, U.S. Environmental Protection Agency. October 2017.
Browning, L. (2009) Personal communication with Lou Browning, "Suggested New Emission Factors for Marine
Vessels," ICF International.
Browning, L. (2005) Personal communication with Lou Browning, "Emission control technologies for diesel highway
vehicles specialist," ICF International.
DHS (2008) Email Communication. Elissa Kay, Department of Homeland Security and Joe Aamidor, ICF
International. January 11, 2008.
DLA Energy (2020) Unpublished data from the Defense Fuels Automated Management System (DFAMS). Defense
Energy Support Center, Defense Logistics Agency, U.S. Department of Defense. Washington, D.C.
DOC (1991 through 2019) Unpublished Report of Bunker Fuel Oil Laden on Vessels Cleared for Foreign Countries.
Form-563. Foreign Trade Division, Bureau of the Census, U.S. Department of Commerce. Washington, D.C.
DOE (1993 through 2018) Transportation Energy Data Book. Office of Transportation Technologies, Center for
Transportation Analysis, Energy Division, Oak Ridge National Laboratory. ORNL-6978.
DOT (1991 through 2019) Airline Fuel Cost and Consumption. U.S. Department of Transportation, Bureau of
Transportation Statistics, Washington, D.C. DAI-10. Available online at: .
EIA (2020a) Monthly Energy Review, November 2020, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/11).
EIA (2020f) Natural Gas Annual 2018. Energy Information Administration, U.S. Department of Energy, Washington,
D.C. DOE/EIA-0131(11).
EIA (1991 through 2019) Fuel Oil and Kerosene Sales. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EIA (2016) "Table 3.1: World Petroleum Supply and Disposition." International Energy Annual. Energy Information
Administration, U.S. Department of Energy. Washington, D.C. Available online at:
.
References 10-9

-------
EIA (2011) Annual Energy Review 2010. Energy Information Administration, U.S. Department of Energy,
Washington, D.C. DOE/EIA-0384(2011). October 19, 2011.
EIA (2007) Personal Communication. Joel Lou, Energy Information Administration and Aaron Beaudette, ICF
International. Residual and Distillate Fuel Oil Consumption for Vessel Bunkering (Both International and Domestic)
for American Samoa, U.S. Pacific Islands, and Wake Island. October 24, 2007.
EIA (2002) Alternative Fuels Data Tables. Energy Information Administration, U.S. Department of Energy,
Washington, D.C. Available online at: .
EPA (2019a) Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 - 2018.
Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Available online at:
.
EPA (2019bJ Motor Vehicle Emissions Simulator (MOVES). Office of Transportation and Air Quality, U.S.
Environmental Protection Agency. Available online at: .
EPA (2020c) Confidential Engine Family Sales Data Submitted to EPA by Manufacturers. Office of Transportation
and Air Quality, U.S. Environmental Protection Agency.
EPA (2020d) Annual Certification Test Results Report. Office of Transportation and Air Quality, U.S. Environmental
Protection Agency. Available online at: .
EPA (2016g) "1970 - 2015 Average annual emissions, all criteria pollutants in MS Excel." National Emissions
Inventory (NEI) Air Pollutant Emissions Trends Data. Office of Air Quality Planning and Standards. Available online
at: .
EPA (2000) Mobile6 Vehicle Emission Modeling Software. Office of Mobile Sources, U.S. Environmental Protection
Agency, Ann Arbor, Michigan.
EPA (1999a) Emission Facts: The History of Reducing Tailpipe Emissions. Office of Mobile Sources. May 1999. EPA
420-F-99-017. Available online at: .
EPA (1999b) Regulatory Announcement: EPA's Program for Cleaner Vehicles and Cleaner Gasoline. Office of Mobile
Sources. December 1999. EPA420-F-99-051. Available online at:
.
EPA (1998) Emissions of Nitrous Oxide from Highway Mobile Sources: Comments on the Draft Inventory of U.S.
Greenhouse Gas Emissions and Sinks, 1990-1996. Office of Mobile Sources, Assessment and Modeling Division,
U.S. Environmental Protection Agency. August 1998. EPA420-R-98-009.
EPA (1994a) Automobile Emissions: An Overview. Office of Mobile Sources. August 1994. EPA 400-F-92-007.
Available online at: .
EPA (1994b) Milestones in Auto Emissions Control. Office of Mobile Sources. August 1994. EPA 400-F-92-014.
Available online at: .
EPA (1993) Automobiles and Carbon Monoxide. Office of Mobile Sources. January 1993. EPA 400-F-92-005.
Available online at: .
Esser, C. (2003 through 2004) Personal Communication with Charles Esser, Residual and Distillate Fuel Oil
Consumption for Vessel Bunkering (Both International and Domestic) for American Samoa, U.S. Pacific Islands, and
Wake Island.
FAA (2021) Personal Communication between FAA and John Steller, Mausami Desai and Vincent Camobreco for
aviation emission estimates from the Aviation Environmental Design Tool (AEDT). March 2021.
FHWA (1996 through 2019) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Available online at:
.
10-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
FHWA (2015) Off-Highway and Public-Use Gasoline Consumption Estimation Models Used in the Federal Highway
Administration, Publication Number FHWA-PL-17-012. Available online at:
.
Gaffney, J. (2007) Email Communication. John Gaffney, American Public Transportation Association and Joe
Aamidor, ICF International. December 17, 2007.
HybridCars.com (2019). Monthly Plug-In Electric Vehicle Sales Dashboard, 2010-2018. Available online at
.
ICF (2006a) Revised Gasoline Vehicle EFsfor LEV and Tier 2 Emission Levels. Memorandum from ICF International to
John Davies, Office of Transportation and Air Quality, U.S. Environmental Protection Agency. November 2006.
ICF (2006b) Revisions to Alternative Fuel Vehicle (AFV) Emission Factors for the U.S. Greenhouse Gas Inventory.
Memorandum from ICF International to John Davies, Office of Transportation and Air Quality, U.S. Environmental
Protection Agency. November 2006.
ICF (2004) Update of Methane and Nitrous Oxide Emission Factors for On-Highway Vehicles. Final Report to U.S.
Environmental Protection Agency. February 2004.
ICF (2017b) Updated Non-Highway CH4 and N20 Emission Factors for U.S. GHG Inventory. Memorandum from ICF
to Sarah Roberts and Justine Geidosch, Office of Transportation and Air Quality, U.S. Environmental Protection
Agency. October 2017.
Lipman, T. and M. Delucchi (2002) "Emissions of Nitrous Oxide and Methane from Conventional and Alternative
Fuel Motor Vehicles." Climate Change, 53:477-516.
SAE (2010) Utility Factor Definitions for Plug-In Hybrid Electric Vehicles Using Travel Survey Data. Society of
Automotive Engineers. Report J2841, Available online at:
.
Raillnc (2014 through 2019) RaillncShort line and Regional Traffic Index. Carloads Originated Year-to-Date.
December 2019. Available online at: .
Santoni, G., B. Lee, E. Wood, S. Herndon, R. Miake-Lye, S. Wofsy, J. McManus, D. Nelson, M. Zahniser (2011)
Aircraft emissions of methane and nitrous oxide during the alternative aviation fuel experiment. Environ Sci
Technol. 2011 Aug 15; 45(16):7075-82.
U.S. Census Bureau (2000) Vehicle Inventory and Use Survey. U.S. Census Bureau, Washington, D.C. Database CD-
EC97-VIUS.
Whorton, D. (2006 through 2014) Personal communication, Class II and III Rail energy consumption, American
Short Line and Regional Railroad Association.
Carbon Emitted from Non-Energy Uses of Fossil Fuels
ACC (2020a) "Guide to the Business of Chemistry, 2020," American Chemistry Council.
ACC (2020b) "U.S. Resin Production & Sales 2019 vs. 2018." Available online at:
.
ACC (2019) "U.S. Resin Production & Sales 2018 vs. 2017." Available online at:
.
ACC (2018) "U.S. Resin Production & Sales 2017 vs. 2016." Available online at:
.
ACC (2017) "U.S. Resin Production & Sales 2016 vs. 2015."
ACC (2016) "U.S. Resin Production & Sales 2015 vs. 2014."
References 10-11

-------
ACC (2015) "PIPS Year-End Resin Statistics for 2014 vs. 2013: Production, Sales and Captive Use." Available online
at: .
ACC (2014) "U.S. Resin Production & Sales: 2013 vs. 2012," American Chemistry Council. Available online at:
.
ACC (2013) "U.S. Resin Production & Sales: 2012 vs. 2011," American Chemistry Council. Available online at:
.
ACC (2003-2011) "PIPS Year-End Resin Statistics for 2010: Production, Sales and Captive Use." Available online at:
.
Bank of Canada (2020) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2019) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2018) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2017) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2016) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2014) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2013) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
Bank of Canada (2012) Financial Markets Department Year Average of Exchange Rates. Available online at:
.
EIA (2021) EIA Manufacturing Consumption of Energy (MECS) 2018. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EIA (2020a) Monthly Energy Review, November 2020. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035 (2020/11).
EIA (2020b) Glossary. Energy Information Administration, U.S. Department of Energy, Washington, D.C. Available
online at: .
EIA (2019) Personal communication between EIA and ICF on November 11, 2019.
EIA (2017) EIA Manufacturing Consumption of Energy (MECS) 2014. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EIA (2013) EIA Manufacturing Consumption of Energy (MECS) 2010. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EIA (2010) EIA Manufacturing Consumption of Energy (MECS) 2006. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EIA (2005) EIA Manufacturing Consumption of Energy (MECS) 2002. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
10-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
EIA (2001) EIA Manufacturing Consumption of Energy (MECS) 1998. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EIA (1997) EIA Manufacturing Consumption of Energy (MECS) 1994. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EIA (1994) EIA Manufacturing Consumption of Energy (MECS) 1991. U.S. Department of Energy, Energy Information
Administration, Washington, D.C.
EPA (2021) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2020) "Criteria pollutants National Tier 1 for 1970 - 2019." National Emissions Inventory (NEI) Air Pollutant
Emissions Trends Data. Office of Air Quality Planning and Standards, April 2020. Available online at:
.
EPA (2019) Advancing Sustainable Materials Management: 2016 and 2017 Data Tables. Office of Land and
Emergency Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2018a) Advancing Sustainable Materials Management: Facts and Figures 2015, Assessing Trends in Material
Generation, Recycling and Disposal in the United States. Washington, D.C.
EPA (2018b) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2017) EPA's Pesticides Industry Sales and Usage, 2008 - 2012 Market Estimates. Available online at:

Accessed September 2017.
EPA (2016a) Advancing Sustainable Materials Management: 2014 Facts and Figures Fact Sheet. Office of Solid
Waste and Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2016b) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2015) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2014a) Municipal Solid Waste in the United States: 2012 Facts and Figures. Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2014b) Chemical Data Access Tool (CDAT). U.S. Environmental Protection Agency, June 2014. Available online
at: . Accessed January 2015.
EPA (2013a) Municipal Solid Waste in the United States: 2011 Facts and Figures. Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2013b) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2011) EPA's Pesticides Industry Sales and Usage, 2006 and 2007 Market Estimates. Available online at:
. Accessed January 2012.
EPA (2009) Biennial Reporting System (BRS) Database. U.S. Environmental Protection Agency, Envirofacts
Warehouse. Washington, D.C. Available online at: . Data for 2001-2007
are current as of Sept. 9, 2009.
References 10-13

-------
EPA (2004) EPA's Pesticides Industry Sales and Usage, 2000 and 2001 Market Estimates. Available online at:
. Accessed September 2006.
EPA (2002) EPA's Pesticides Industry Sales and Usage, 1998 and 1999 Market Estimates, Table 3.6. Available online
at: . Accessed July 2003.
EPA (2001) AP 42, Volume I, Fifth Edition. Chapter 11: Mineral Products Industry. Available online at:
.
EPA (2000a) Biennial Reporting System (BRS). U.S. Environmental Protection Agency, Envirofacts Warehouse.
Washington, D.C. Available online at: .
EPA (2000b) Toxics Release Inventory, 1998. U.S. Environmental Protection Agency, Office of Environmental
Information, Office of Information Analysis and Access, Washington, D.C. Available online at:
.
EPA (1999) EPA's Pesticides Industry Sales and Usage, 1996-1997 Market Estimates. Available online at:
.
EPA (1998) EPA's Pesticides Industry Sales and Usage, 1994-1995 Market Estimates. Available online at:
.
FEB (2013) Fiber Economics Bureau, as cited in C&EN (2013) Lackluster Year for Chemical Output: Production
stayed flat or dipped in most world regions in 2012. Chemical &Engineering News, American Chemical Society, 1
July. Available online at: .
FEB (2012) Fiber Economics Bureau, as cited in C&EN (2012) Too Quiet After the Storm: After a rebound in 2010,
chemical production hardly grew in 2011. Chemical & Engineering News, American Chemical Society, 2 July.
Available online at: .
FEB (2011) Fiber Economics Bureau, as cited in C&EN (2011) Output Ramps up in all Regions. Chemical Engineering
News, American Chemical Society, 4 July. Available online at: .
FEB (2010) Fiber Economics Bureau, as cited in C&EN (2010) Output Declines in U.S., Europe. Chemical &
Engineering News, American Chemical Society, 6 July. Available online at: .
FEB (2009) Fiber Economics Bureau, as cited in C&EN (2009) Chemical Output Slipped In Most Regions Chemical &
Engineering News, American Chemical Society, 6 July. Available online at: .
FEB (2007) Fiber Economics Bureau, as cited in C&EN (2007) Gains in Chemical Output Continue. Chemical &
Engineering News, American Chemical Society. July 2, 2007. Available online at: .
FEB (2005) Fiber Economics Bureau, as cited in C&EN (2005) Production: Growth in Most Regions Chemical &
Engineering News, American Chemical Society, 11 July. Available online at: .
FEB (2003) Fiber Economics Bureau, as cited in C&EN (2003) Production Inches Up in Most Countries, Chemical &
Engineering News, American Chemical Society, 7 July. Available online at: .
FEB (2001) Fiber Economics Bureau, as cited in ACS (2001) Production: slow gains in output of chemicals and
products lagged behind U.S. economy as a whole Chemical & Engineering News, American Chemical Society, 25
June. Available online at: .
Financial Planning Association (2006) Canada/US Cross-Border Tools: US/Canada Exchange Rates. Available online
at: . Accessed on August 16, 2006.
Gosselin, Smith, and Hodge (1984) "Clinical Toxicology of Commercial Products." Fifth Edition, Williams & Wilkins,
Baltimore.
ICIS (2016) "Production issues force US melamine plant down" Available online at:
.
10-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
ICIS (2008) "Chemical profile: Melamine" Available online at:
. Accessed November,
2017.
IISRP (2003) "IISRP Forecasts Moderate Growth in North America to 2007" International Institute of Synthetic
Rubber Producers, Inc. New Release. Available online at: .
IISRP (2000) "Synthetic Rubber Use Growth to Continue Through 2004, Says IISRP and RMA" International Institute
of Synthetic Rubber Producers press release.
INEGI (2006) Produccion bruta total de las unidades economicas manufactureras porSubsector, Rama, Subrama y
Clase de actividad. Available online at:
. Accessed
on August 15, 2006.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Marland, G., and R.M. Rotty (1984) "Carbon dioxide emissions from fossil fuels: A procedure for estimation and
results for 1950-1982," Tellus 36b:232-261.
NPRA (2002) North American Wax - A Report Card. Available online at:
.
RMA (2018) 2017 U.S. Scrap Tire Management Summary. Rubber Manufacturers Association, Washington, D.C. July
2018.
RMA (2016) 2015 U.S. Scrap Tire Management Summary. Rubber Manufacturers Association, Washington, D.C.
August 2016.
RMA (2014) 2013 U.S. Scrap Tire Management Summary. Rubber Manufacturers Association, Washington, D.C.
November 2014.
RMA (2011) U.S. Scrap Tire Management Summary: 2005-2009. Rubber Manufacturers Association, Washington,
D.C. October 2011, updated September 2013.
RMA (2009) "Scrap Tire Markets: Facts and Figures-Scrap Tire Characteristics." Rubber Manufacturers
Association., Washington D.C. Available online at:
. Accessed on 17 September
2009.
U.S. Census Bureau (2014) 2012 Economic Census. Available online at:
. Accessed November 2014.
U.S. Census Bureau (2009) Soap and Other Detergent Manufacturing: 2007. Available online at:
.
U.S. Census Bureau (2004) Soap and Other Detergent Manufacturing: 2002. Issued December 2004. EC02-31I-
325611 (RV). Available online at: .
U.S. Census Bureau (1999) Soap and Other Detergent Manufacturing: 1997. Available online at:
.
U.S. International Trade Commission (1990-2019) "Interactive Tariff and Trade DataWeb: Quick Query." Available
online at: . Accessed September 2020.
References 10-15

-------
Incineration of Waste
ArSova, Ljupka, Rob van Haaren, Nora Goldstein, Scott M. Kaufman, and Nickolas J. Themelis (2008) "16th Annual
BioCycle Nationwide Survey: The State of Garbage in America" BioCycle, JG Press, Emmaus, PA. December.
Bahor, B (2009) Covanta Energy's public review comments re: Draft Inventory of U.S. Greenhouse Gas Emissions
and Sinks: 1990-2007. Submitted via email on April 9, 2009 to Leif Hockstad, U.S. EPA.
De Soete, G.G. (1993) "Nitrous Oxide from Combustion and Industry: Chemistry, Emissions and Control." In A. R.
Van Amstel, (ed.) Proc. of the International Workshop Methane and Nitrous Oxide: Methods in National Emission
Inventories and Options for Control, Amersfoort, NL February 3-5,1993.
Energy Recovery Council (2018) Energy Recovery Council. 2018 Directory of Waste to Energy Facilities. Ted
Michaels and Karunya Krishnan. October 2018. Available online at: .
Energy Recovery Council (2009) "2007 Directory of Waste-to-Energy Plants in the United States." Accessed on
September 29, 2009.
EIA (2017) MSW Incineration for Heating or Electrical Generation, December 2017, Energy Information
Administration, U.S. Department of Energy, Washington, DC. DOE/EIA-0035. Available online at:
.
EPA (2020) Advancing Sustainable Materials Management: 2018 Data Tables. Office of Land and Emergency
Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2019) Advancing Sustainable Materials Management: 2016 and 2017 Data Tables. Office of Land and
Emergency Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2018a) Advancing Sustainable Materials Management: 2015 Data Tables. Office of Land and Emergency
Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2018b). Greenhouse Gas Reporting Program Data. Washington, DC: U.S. Environmental Protection Agency.
Available online at: .
EPA (2016) Advancing Sustainable Materials Management: 2014 Fact Sheet. Office of Land and Emergency
Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2015) Advancing Sustainable Materials Management: Facts and Figures 2013 - Assessing Trends in Material
Generation, Recycling and Disposal in the United States. Office of Solid Waste and Emergency Response, U.S.
Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2007, 2008, 2011, 2013, 2014) Municipal Solid Waste in the United States: Facts and Figures. Office of Solid
Waste and Emergency Response, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2006) Solid Waste Management and Greenhouse Gases: A Life-Cycle Assessment of Emissions and Sinks.
Office of Solid Waste and Emergency Response, U.S. Environmental Protection Agency. Washington, D.C.
EPA (2000) Characterization of Municipal Solid Waste in the United States: Source Data on the 1999 Update. Office
of Solid Waste, U.S. Environmental Protection Agency. Washington, D.C. EPA530-F-00-024.
10-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Goldstein, N. and C. Madtes (2001) "13th Annual BioCycle Nationwide Survey: The State of Garbage in America."
BioCycle, JG Press, Emmaus, PA. December 2001.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Kaufman, et al. (2004) "14th Annual BioCycle Nationwide Survey: The State of Garbage in America 2004" Biocycle,
JG Press, Emmaus, PA. January 2004.
RMA (2020) "2019 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association, Washington, DC.
October 2020. Available online at:

RMA (2018) "2017 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association, Washington, DC.
July 2018. Available online at:

RMA (2016) "2015 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association. August 2016.
Available online at: .
RMA (2014) "2013 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association. November 2014.
Available online at: .
RMA (2013) "U.S. Scrap Tire Management Summary 2005-2009." Rubber Manufacturers Association. October
2011; Updated September 2013. Available online at:
.
RMA (2012a) "Rubber FAQs." Rubber Manufacturers Association. Available online at: . Accessed on 19 November 2014.
RMA (2012b) "Scrap Tire Markets: Facts and Figures - Scrap Tire Characteristics." Rubber Manufacturers
Association. Available online at:
. Accessed 18 on January 2012.
Schneider, S. (2007) E-mail between Shelly Schneider of Franklin Associates (a division of ERG) and Sarah Shapiro of
ICF International, January 10, 2007.
Shin, D. (2014) Generation and Disposition of Municipal Solid Waste (MSW) in the United States-A National
Survey. Thesis. Columbia University, Department of Earth and Environmental Engineering, January 3, 2014.
Simmons, et al. (2006) "15th Nationwide Survey of Municipal Solid Waste Management in the United States: The
State of Garbage in America." BioCycle, JG Press, Emmaus, PA. April 2006.
van Haaren, Rob, Themelis, N., and Goldstein, N. (2010) 'The State of Garbage in America." BioCycle, October
2010. Volume 51, Number 10, pg. 16-23.
Coal Mining
AAPG (1984) Coalbed Methane Resources of the United States. AAPG Studies in Geology Series #17.
Creedy, D.P. (1993) Methane Emissions from Coal Related Sources in Britain: Development of a Methodology.
Chemosphere, 26: 419-439.
DMME (2020) DGO Data Information System. Department of Mines, Minerals and Energy of Virginia. Available
online at .
EIA (2020) Annual Coal Report 2019. Table 1. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. DOE/EIA-0584.
References 10-17

-------
El Paso (2009) Shoal Creek Mine Plan, El Paso Exploration & Production.
EPA (2020) Greenhouse Gas Reporting Program (GHGRP) 2019 Envirofacts. Subpart FF: Underground Coal Mines.
Available online at .
EPA (2005) Surface Mines Emissions Assessment. Draft. U.S. Environmental Protection Agency.
EPA (1996) Evaluation and Analysis of Gas Content and Coal Properties of Major Coal Bearing Regions of the United
States. EPA/600/R-96-065. U.S. Environmental Protection Agency.
ERG (2020). Correspondence between ERG and Buchanan Mine.
Geological Survey of Alabama State Oil and Gas Board (GSA) (2020) Well Records Database. Available online at
.
IEA (2020) Key World Energy Statistics. Coal Production, International Energy Agency.
IPCC (2011) Use of Models and Facility-Level Data in Greenhouse Gas Inventories. Report of IPCC Expert Meeting on
Use of Models and Measurements in Greenhouse Gas Inventories 9-11 August 2010, Sydney, Australia. Eds:
Eggleston H.S., Srivastava N.,Tanabe K., Baasansuren J., Fukuda M. IGES.
JWR (2010) No. 4&7 Mines General Area Maps. Walter Energy: Jim Walter Resources.
King, Brian (1994) Management of Methane Emissions from Coal Mines: Environmental, Engineering, Economic and
Institutional Implication of Options. Neil and Gunter Ltd.
McElroy OVS (2020) Marshall County VAM Abatement Project Offset Verification Statement submitted to
California Air Resources Board, December 2020.
MSHA (2020) Data Transparency at MSHA. Mine Safety and Health Administration. Available online at
.
Mutmansky, Jan M. and Yanbei Wang (2000) Analysis of Potential Errors in Determination of Coal Mine Annual
Methane Emissions. Mineral Resources Engineering, 9(4).
Saghafi, Abouna (2013) Estimation of Fugitive Emissions from Open Cut Coal Mining and Measurable Gas Content.
13th Coal Operators' Conference, University of Wollongong, The Australian Institute of Mining and Metallurgy &
Mine Managers Association of Australia. 306-313.
USBM (1986) Results of the Direct Method Determination of the Gas Contents of U.S. Coal Basins. Circular 9067.
U.S. Bureau of Mines.
West Virginia Geological & Economic Survey (WVGES) (2020) Oil & Gas Production Data. Available online at
.
Abam'.vu d Underground Coal Mines
EPA (2004) Methane Emissions Estimates & Methodology for Abandoned Coal Mines in the U.S. Draft Final Report.
Washington, D.C. April 2004.
MSHA (2020) U.S. Department of Labor, Mine Health & Safety Administration, Mine Data Retrieval System.
Available online at: .
Petroleum Systems
API (1992) Global Emissions of Methane from Petroleum Sources. American Petroleum Institute, Health and
Environmental Affairs Department, Report No. DR140, February 1992.
BOEM (2020a) BOEM Platform Structures Online Query. Available online at:
.
10-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
BOEM (2020b) BOEM Oil and Gas Operations Reports - Part A (OGOR-A). Production Data for 1947 to 2019.
Download "Production Data" online at: .
BOEM (2020c) BOEM Oil and Gas Operations Reports - Part A (OGOR-A). Production Data for 1996 to 2019.
Available online at: .
BOEM (2020d) BOEM Oil and Gas Operations Reports - Part B (OGOR-B). Flaring volumes for 1996 to 2019.
Available online at: 
CenSARA (2012) 2011 Oil and Gas Emission Inventory Enhancement Project for CenSARA States. Prepared by
ENVIRON International Corporation and Eastern Research Group, Inc. (ERG). Central States Air Resources Agencies
(CenSARA). December 2012.
EIA (2020) Crude Oil Production. Energy Information Administration.
Enverus Drillinglnfo (2019) March 2019 Download. Dl Desktop® Enverus Drillinglnfo, Inc.
Enverus (2021) March 2021 Download. Enverus, Inc.
EPA (1977) Atmospheric Emissions from Offshore Oil and Gas Development and Production. Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency. Research Triangle Park, NC. PB272268. June 1977.
EPA (1997) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC. October 1997.
EPA (1999) Estimates of Methane Emissions from the U.S. Oil Industry (Draft Report). Prepared by ICF International
Office of Air and Radiation, U.S. Environmental Protection Agency. October 1999.
EPA (2017) 2017 Nonpoint Oil and Gas Emission Estimation Tool, Version 1.2. Prepared for U.S. Environmental
Protection Agency by Eastern Research Group, Inc. (ERG). October 2019.
EPA (2020) Greenhouse Gas Reporting Program. U.S. Environmental Protection Agency. Data reported as of
September 26, 2020.
EPA (2021a) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Produced Water
Emissions (Produced Water memo). U.S. Environmental Protection Agency. April 2021. Available at:
.
EPA (2021b) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Natural Gas and
Petroleum Systems C02 Uncertainty Estimates. U.S. Environmental Protection Agency. April 2021. Available at:
.
EPA/GRI (1996) Methane Emissions from the Natural Gas Industry. Prepared by Radian. U.S. Environmental
Protection Agency. April 1996.
Illinois Office of Oil and Gas Resources Management (2020) State-level petroleum production quantities.
Indiana Division of Oil & Gas (2020) State-level petroleum production quantities.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Kansas Department of Health and Environment (2020) County-level produced water quantities.
Ohio Environmental Protection Agency (2020) Well-level produced water quantities.
Oklahoma Department of Environmental Quality (2020) Well-level produced water quantities.
Pennsylvania Department of Environmental Protection (2020) Well-level produced water quantities.
West Virginia Department of Environmental Protection (2020) State-level petroleum production quantities.
References 10-19

-------
Natural Gas Systems
CenSARA (2012) 2011 Oil and Gas Emission Inventory Enhancement Project for CenSARA States. Prepared by
ENVIRON International Corporation and Eastern Research Group, Inc. (ERG). Central States Air Resources Agencies
(CenSARA). December 2012.
EIA (2020) Natural Gas Gross Withdrawals and Production. Energy Information Administration.
Enverus Drillinglnfo (2019) March 2019 Download. Dl Desktop" Enverus Drillinglnfo, Inc.
Enverus (2021) March 2021 Download. Enverus, Inc.
EPA (1977) Atmospheric Emissions from Offshore Oil and Gas Development and Production. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. PB272268. June 1977.
EPA (2020) Greenhouse Gas Reporting Program- Subpart W-Petroleum and Natural Gas Systems. Environmental
Protection Agency. Data reported as of September 26, 2020.
EPA (2021a) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Natural Gas Customer
Meter Emissions (Customer Meters memo). U.S. Environmental Protection Agency. April 2021. Available at:
.
EPA (2021b) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Produced Water
Emissions (Produced Water memo). U.S. Environmental Protection Agency. April 2021. Available at:
.
EPA (2021c) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Update for Natural Gas and
Petroleum Systems C02 Uncertainty Estimates. U.S. Environmental Protection Agency. April 2021. Available at:
.
GRI/EPA (1996) Methane Emissions from the Natural Gas Industry. Prepared by Harrison, M., T. Shires, J. Wessels,
and R. Cowgill, eds., Radian International LLC for National Risk Management Research Laboratory, Air Pollution
Prevention and Control Division, Research Triangle Park, NC. EPA-600/R-96-080a.
GSI (2019) Long-term Methane Emissions Rate Quantification and Alert System for Natural Gas Storage Wells and
Fields. 2019. GSI Environmental Inc. DOE Report DE-FE0029085.
GTI (2001) Gas Resource Database: Unconventional Natural Gas and Gas Composition Databases. Second Edition.
GRI-01/0136.
GTI (2019) Classification of Methane Emissions from Industrial Meters, Vintage vs Modern Plastic Pipe, and Plastic-
lined Steel and Cast-Iron Pipe. June 2019. Gas Technology Institute and U.S. Department of Energy GTI Project
Number 22070. DOE project Number ED-FE0029061.
Illinois Office of Oil and Gas Resource Management (2020) State-level natural gas production quantities.
Indiana Division of Oil & Gas (2020) State-level natural gas production quantities.
Kansas Department of Health and Environment (2020) County-level produced water quantities.
Lamb, et al. (2015) "Direct Measurements Show Decreasing Methane Emissions from Natural Gas Local
Distribution Systems in the United States." Environmental Science & Technology, Vol. 49 5161-5169.
Lavoie et al. (2017) "Assessing the Methane Emissions from Natural Gas-Fired Power Plants and Oil Refineries."
Environmental Science & Technology. 2017 Mar 21;51(6):3373-3381. doi: 10.1021/acs.est.6b05531.
Ohio Environmental Protection Agency (2020) Well-level produced water quantities.
Oklahoma Department of Environmental Quality (2020) Well-level produced water quantities.
Pennsylvania Department of Environmental Protection (2020) Well-level produced water quantities.
10-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
PHMSA (2019) Gas Distribution Annual Data. Pipeline and Hazardous Materials Safety Administration, U.S.
Department of Transportation, Washington, DC. Available online at: .
West Virginia Department of Environmental Protection (2020) State-level natural gas production quantities.
Zimmerle et al. (2019) "Characterization of Methane Emissions from Gathering Compressor Stations." October
2019. Available at .
Zimmerle, et al. (2015) "Methane Emissions from the Natural Gas Transmission and Storage System in the United
States." Environmental Science and Technology, Vol. 49 9374-9383.
Abam'. vu d Oil and Gas Wells
Alaska Oil and Gas Conservation Commission, Available online at: .
Arkansas Geological & Conservation Commission, "List of Oil & Gas Wells - Data From November 1,1936 to January
1,1955." Available at: .
The Derrick's Handbook of Petroleum: A Complete Chronological and Statistical Review of Petroleum
Developments From 1859 to 1898 (V.l), (1898-1899) (V.2).
Enverus (2021) March 2021 Download. Enverus, Inc.
GRI/EPA (1996) Methane Emissions from the Natural Gas Industry. Prepared by Harrison, M., T. Shires, J. Wessels,
and R. Cowgill, eds., Radian International LLC for National Risk Management Research Laboratory, Air Pollution
Prevention and Control Division, Research Triangle Park, NC. EPA-600/R-96-080a.
Florida Department of Environmental Protection - Oil and Gas Program, Available online at:
.
Geological Survey of Alabama, Oil & Gas Board, Available online at: .
GTI (2001) Gas Resource Database: Unconventional Natural Gas and Gas Composition Databases. Second Edition.
GRI-01/0136.
Kang, et al. (2016) "Identification and characterization of high methane-emitting abandoned oil and gas wells."
PNAS, vol. 113 no. 48, 13636-13641, doi: 10.1073/pnas.l605913113.
Oklahoma Geological Survey. "Oklahoma Oil: Past, Present, and Future." Oklahoma Geology Notes, v. 62 no. 3,
2002 pp. 97-106.
Pennsylvania Department of Environmental Protection, Oil and Gas Reports - Oil and Gas Operator Well Inventory.
Available online at:
http://www.depreportingservices.state.pa.us/ReportServer/Pages/ReportViewer.aspx7/Oil_Gas/OG_Well_lnvento
ry.
Texas Railroad Commission, Oil and Gas Division, "History of Texas Initial Crude Oil, Annual Production and
Producing Wells, Crude Oil Production and Well Counts (since 1935)." Available online at:
.
Townsend-Small, et al. (2016) "Emissions of coalbed and natural gas methane from abandoned oil and gas wells in
the United States." Geophysical Research Letters, Vol. 43,1789-1792.
United States Geological Survey's (USGS) Mineral Resources of the United States Annual Yearbooks, available
online at: .
References 10-21

-------
Virginia Department of Mines Minerals and Energy, "Wells Drilled for Oil and Gas in Virginia prior to 1962.",
Virginia Division of Mineral Resources. Available online at:
.
Energy Sources Precursor Greenhouse Gases
EPA (2019) "1970 - 2018 Average annual emissions, all criteria pollutants in MS Excel." National Emissions
Inventory (NEI) Air Pollutant Emissions Trends Data. Office of Air Quality Planning and Standards, May 2019.
Available online at: .
EPA (2003) E-mail correspondence containing preliminary ambient air pollutant data. Office of Air Pollution and
the Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. December 22, 2003.
EPA (1997) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC. October 1997.
International Bunker Fuels
Anderson, B.E., et al. (2011) Alternative Aviation Fuel Experiment (AAFEX), NASATechnical Memorandum, in press.
ASTM (1989) Military Specification for Turbine Fuels, Aviation, Kerosene Types, NATO F-34 (JP-8) and NATO F-35.
February 10,1989.
Chevron (2000) Aviation Fuels Technical Review (FTR-3). Chevron Products Company, Chapter 2.
DHS (2008) Personal Communication with Elissa Kay, Residual and Distillate Fuel Oil Consumption (International
Bunker Fuels). Department of Homeland Security, Bunker Report. January 11, 2008.
DLA Energy (2020) Unpublished data from the Defense Fuels Automated Management System (DFAMS). Defense
Energy Support Center, Defense Logistics Agency, U.S. Department of Defense. Washington, D.C.
DOC (1991 through 2020) Unpublished Report of Bunker Fuel Oil Laden on Vessels Cleared for Foreign Countries.
Form-563. Foreign Trade Division, Bureau of the Census, U.S. Department of Commerce. Washington, D.C.
DOT (1991 through 2013) Fuel Cost and Consumption. Federal Aviation Administration, Bureau of Transportation
Statistics, U.S. Department of Transportation. Washington, D.C. DAI-10.
EIA (2020) Monthly Energy Review, November 2020, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2020/11).
EPA (2020) EPA Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2019: Updated Gasoline and Diesel
Fuel C02 Emission Factors - Memo.
FAA (2021) Personal Communication between FAA and John Steller, Mausami Desai, and Vincent Camobreco for
aviation emissions estimates from the Aviation Environmental Design Tool (AEDT). March 2021.
IPCC/UNEP/OECD/IEA (1997) Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories. 31
Intergovernmental Panel on Climate Change, United Nations Environment Programme, Organization for Economic
32 Co-Operation and Development, International Energy Agency, Paris, France.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
USAF (1998) Fuel Logistics Planning. U.S. Air Force pamphlet AFPAM23-221, May 1,1998.
10-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Wood Biomass and Biofuel Consumption
EIA (2020a) Monthly Energy Review, November 2020. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2020/11).
EIA (2020b) Biofuels explained: Use of biomass-based diesel fuel. Energy Information Administration, U.S.
Department of Energy. Washington, D.C. Available online at: .
EPA (2021) Acid Rain Program Dataset 1996-2019. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2010) Carbon Content Coefficients Developed for EPA's Mandatory Reporting Rule. Office of Air and
Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
Lindstrom, P. (2006) Personal Communication. Perry Lindstrom, Energy Information Administration and Jean Kim,
ICF International.
Industrial Processes and Product Use
EPA (2014) Greenhouse Gas Reporting Program. Developments on Publication of Aggregated Greenhouse Gas
Data, November 25, 2014. See .
EPA (2002) Quality Assurance/Quality Control and Uncertainty Management Plan for the U.S. Greenhouse Gas
Inventory: Procedures Manual for Quality Assurance/Quality Control and Uncertainty Analysis, U.S. Greenhouse
Gas Inventory Program, U.S. Environmental Protection Agency, Office of Atmospheric Programs, EPA430-R-02-
007B, June 2002.
IPCC (2011) Use of Models and Facility-Level Data in Greenhouse Gas Inventories (Report of IPCC Expert Meeting
on Use of Models and Measurements in Greenhouse Gas Inventories 9-11 August 2010, Sydney, Australia) eds.:
Eggleston H.S., Srivastava N., Tanabe K., Baasansuren J., Fukuda M., Pub. IGES, Japan 2011.
Cement Production
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
EPA Greenhouse Gas Reporting Program (2020) Aggregation of Reported Facility Level Data under Subpart H -
National Level Clinker Production from Cement Production for Calendar Years 2014 through 2019. Office of Air and
Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
U.S. Bureau of Mines (1990 through 1993) Minerals Yearbook: Cement Annual Report. U.S. Department of the
Interior, Washington, D.C.
United States Geological Survey (USGS) (2020) Mineral Commodity Summaries: Cement. U.S. Geological Survey,
Reston, VA. January 2020. Available at: .
USGS (1995 through 2014) Minerals Yearbook - Cement. U.S. Geological Survey, Reston, VA.
USGS (2020) 2017 Minerals Yearbook - Cement. U.S. Geological Survey, Reston, VA. August 2020.
References 10-23

-------
Van Oss (2013a) 1990 through 2012 Clinker Production Data Provided by Hendrik van Oss (USGS) via email on
November 8, 2013.
Van Oss (2013b) Personal communication. Hendrik van Oss, Commodity Specialist of the U.S. Geological Survey
and Gopi Manne, Eastern Research Group, Inc. October 28, 2013.
Lime Production
EPA (2020) Greenhouse Gas Reporting Program (GHGRP). Aggregation of Reported Facility Level Data under
Subpart S-National Lime Production for Calendar Years 2010 through 2019. Office of Air and Radiation, Office of
Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Males, E. (2003) Memorandum from Eric Males, National Lime Association to Mr. William N. Irving & Mr. Leif
Hockstad, Environmental Protection Agency. March 6, 2003.
Miner, R. and B. Upton (2002) Methods for estimating greenhouse gas emissions from lime kilns at kraft pulp mills.
Energy. Vol. 27 (2002), p. 729-738.
Seeger (2013) Memorandum from Arline M. Seeger, National Lime Association to Mr. Leif Hockstad, Environmental
Protection Agency. March 15, 2013.
USGS (2020a) 2020 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2020).
USGS (2020b) (1992 through 2017) Minerals Yearbook: Lime. U.S. Geological Survey, Reston, VA (June 2020).
USGS (2020c) 2018 Minerals Yearbook Annual Tables: Lime. U.S. Geological Survey, Reston, VA (November 2020).
USGS (2020d) Personal communication. Lori E. Apodaca, U.S. Geological Survey and Amanda Chiu, U.S.
Environmental Protection Agency. December 17, 2020
USGS (2019) 2016 Minerals Yearbook: Lime. U.S. Geological Survey, Reston, VA (August 2019).
USGS (2018a) 2018 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2018).
USGS (2018b) 2015 Minerals Yearbook: Lime. U.S. Geological Survey, Reston, VA (March 2018).
USGS (2012) 2012 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2012).
USGS (2011) 2011 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2011).
USGS (2010) 2010 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2010).
USGS (2008) 2008 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2008).
USGS (2007) 2007 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2007).
USGS (2002) 2002 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2002).
USGS (1996) 1996 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 1996).
USGS (1991) 1991 Minerals Yearbook: Lime. U.S. Geological Survey, Reston, VA (1991).
Glass Production
EPA (2009) Technical Support Document for the Glass Manufacturing Sector: Proposed Rule for Mandatory
Reporting of Greenhouse Gases. U.S. Environmental Protection Agency, Washington, D.C.
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
10-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
OIT (2002) Glass Industry of the Future: Energy and Environmental Profile of the U.S. Glass Industry. Office of
Industrial Technologies, U.S. Department of Energy. Washington, D.C.
U.S. Bureau of Mines (1991 and 1993a) Minerals Yearbook: Crushed Stone Annual Report. U.S. Department of the
Interior. Washington, D.C.
United States Geological Survey (USGS) (2017) Minerals Industry Surveys; Soda Ash in January 2017. U.S.
Geological Survey, Reston, VA. March 2017.
USGS (2018) Mineral Industry Surveys: Soda Ash in February 2018. U.S. Geological Survey, Reston, VA. Accessed
September 2018.
USGS (2019) Mineral Industry Surveys: Soda Ash in December 2018. U.S. Geological Survey, Reston, VA. Accessed
September 24, 2019.
USGS (2020) Mineral Industry Surveys: Soda Ash in April 2020. U.S. Geological Survey, Reston, VA. Accessed
November 2020.
USGS (1995 through 2016a) Minerals Yearbook: Crushed Stone Annual Report. U.S. Geological Survey, Reston, VA.
USGS (1995 through 2015b) Minerals Yearbook: Soda Ash Annual Report. U.S. Geological Survey, Reston, VA.
USGS (2020a) Minerals Yearbook: Crushed Stone Annual Report: Advanced Data Release of the 2017 Annual
Tables. U.S. Geological Survey, Reston, VA. August 2020.
Willett (2020a) Personal communication, Jason Willett, U.S. Geological Survey and Amanda Chiu, U.S.
Environmental Protection Agency. November 16, 2020.
Other Process Uses of Carbonates
AISI (2018 through 2020) Annual Statistical Report. American Iron and Steel Institute.
Kostick, D. S. (2012) Personal communication. Dennis S. Kostick of U.S. Department of the Interior - U.S. Geological
Survey, Soda Ash Commodity Specialist with Gopi Manne and Bryan Lange of Eastern Research Group, Inc. October
2012.
U.S. Bureau of Mines (1991 and 1993a) Minerals Yearbook: Crushed Stone Annual Report. U.S. Department of the
Interior. Washington, D.C.
U.S. Bureau of Mines (1990 through 1993b) Minerals Yearbook: Magnesium and Magnesium Compounds Annual
Report. U.S. Department of the Interior. Washington, D.C.
United States Geological Survey (USGS) (2017a) Mineral Industry Surveys: Soda Ash in January 2017. U.S.
Geological Survey, Reston, VA. March 2017.
USGS (2018) Mineral Industry Surveys: Soda Ash in February 2018. U.S. Geological Survey, Reston, VA. Accessed
September 2018.
USGS (2019) Mineral Industry Surveys: Soda Ash in April 2019. U.S. Geological Survey, Reston, VA. July 2019.
USGS (2020a) 2016 Minerals Yearbook: Stone, Crushed [Advanced Release]. U.S. Geological Survey, Reston, VA.
January 2020.
USGS (2020b) 2017 Minerals Yearbook: Soda Ash [Advanced Release]. U.S. Geological Survey, Reston, VA. August
2020.
USGS (2020c) Minerals Yearbook 2017: Stone, Crushed [Advanced Data Release of the 2017 Annual Tables]. U.S.
Geological Survey, Reston, VA. August 2020.
References 10-25

-------
USGS (1995a through 2017) Minerals Yearbook: Crushed Stone Annual Report. U.S. Geological Survey, Reston, VA.
USGS (1994 through 2015b) Minerals Yearbook: Soda Ash Annual Report. U.S. Geological Survey, Reston, VA.
USGS (1995b through 2020) Minerals Yearbook: Magnesium Annual Report. U.S. Geological Survey, Reston, VA.
Willett (2017) Personal communication, Jason Christopher Willett, U.S. Geological Survey and Mausami Desai and
John Steller, U.S. Environmental Protection Agency. March 9, 2017.
Willett (2020) Personal communication, Jason Christopher Willett, U.S. Geological Survey and Amanda Chiu, U.S.
Environmental Protection Agency. November 16, 2020.
Ammonia Production
ACC (2020) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
Bark (2004) Coffeyville Nitrogen Plant. December 15, 2004. Available online at:
.
Coffeyville Resources Nitrogen Fertilizers (2012) Nitrogen Fertilizer Operations. Available online at:
.
Coffeyville Resources Nitrogen Fertilizers (2011) Nitrogen Fertilizer Operations. Available online at:
.
Coffeyville Resources Nitrogen Fertilizers (2010) Nitrogen Fertilizer Operations. Available online at:
.
Coffeyville Resources Nitrogen Fertilizers (2009) Nitrogen Fertilizer Operations. Available online at:
.
Coffeyville Resources Nitrogen Fertilizers, LLC (2005 through 2007a) Business Data. Available online at:
.
Coffeyville Resources Nitrogen Fertilizers (2007b) Nitrogen Fertilizer Operations. Available online at:
.
Coffeyville Resources Energy, Inc. (CVR) (2012) CVR Energy, Inc. 2012 Annual Report. Available online at:
.
CVR (2013) CVR Energy, Inc. 2013 Annual Report. Available online at: .
CVR (2014) CVR Energy, Inc. 2014 Annual Report. Available online at: .
CVR (2015) CVR Energy, Inc. 2015 Annual Report. Available online at: .
EFMA (2000a) Best Available Techniques for Pollution Prevention and Control in the European Fertilizer Industry.
Booklet No. 1 of 8: Production of Ammonium. Available online at:
.
EFMA (2000b) Best Available Techniques for Pollution Prevention and Control in the European Fertilizer Industry.
Booklet No. 5 of 8: Production of Urea and Urea Ammonium Nitrate. Available online at:
.
EPA Greenhouse Gas Reporting Program (2018) Aggregation of Reported Facility Level Data under Subpart G -
Annual Urea Production from Ammonia Manufacturing for Calendar Years 2011-2016. Office of Air and Radiation,
Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA Greenhouse Gas Reporting Program (2020) Aggregation of Reported Facility Level Data under Subpart G -
Annual Urea Production from Ammonia Manufacturing for Calendar Years 2017-2019. Office of Air and Radiation,
Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
10-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
EPA Greenhouse Gas Reporting Program (GHGRP) (2020) Dataset as of September 26. 2020. Available online at:
.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
U.S. Census Bureau (2011) Current Industrial Reports Fertilizer Materials and Related Products: 2010Summary.
Available online at: .
U.S. Census Bureau (2010) Current Industrial Reports Fertilizer Materials and Related Products: 2009 Summary.
Available online at: .
U.S. Census Bureau (2009) Current Industrial Reports Fertilizer Materials and Related Products: 2008 Summary.
Available online at: .
U.S. Census Bureau (2008) Current Industrial Reports Fertilizer Materials and Related Products: 2007Summary.
Available online at: .
U.S. Census Bureau (2007) Current Industrial Reports Fertilizer Materials and Related Products: 2006 Summary.
Available online at: .
U.S. Census Bureau (2006) Current Industrial Reports Fertilizer Materials and Related Products: 2005 Summary.
Available online at: .
U.S. Census Bureau (2004, 2005) Current Industrial Reports Fertilizer Materials and Related Products: Fourth
Quarter Report Summary. Available online at: .
U.S. Census Bureau (1998 through 2003) Current Industrial Reports Fertilizer Materials and Related Products:
Annual Reports Summary. Available online at: .
U.S. Census Bureau (1991 through 1994) Current Industrial Reports Fertilizer Materials Annual Report. Report No.
MQ28B. U.S. Census Bureau, Washington, D.C.
United States Geological Survey (USGS) (2020) 2020 Mineral Commodity Summaries: Nitrogen (Fixed) - Ammonia.
January 2020. Available online at: .
USGS (1994 through 2009) Minerals Yearbook: Nitrogen. Available online at:
.
Urea Consumption for Non-Agricultural Purposes
EFMA (2000) Best Available Techniques for Pollution Prevention and Control in the European Fertilizer Industry.
Booklet No. 5 of 8: Production of Urea and Urea Ammonium Nitrate.
EPA Greenhouse Gas Reporting Program (2018) Aggregation of Reported Facility Level Data under Subpart G -
Annual Urea Production from Ammonia Manufacturing for Calendar Years 2011-2016. Office of Air and Radiation,
Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA Greenhouse Gas Reporting Program (2020) Aggregation of Reported Facility Level Data under Subpart G -
Annual Urea Production from Ammonia Manufacturing for Calendar Years 2017-2019. Office of Air and Radiation,
Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
TFI (2002) U.S. Nitrogen Imports/Exports Table. The Fertilizer Institute. Available online at:
. August 2002.
References 10-27

-------
U.S. Census Bureau (2001 through 2011) Current Industrial Reports Fertilizer Materials and Related Products:
Annual Summary. Available online at: .
U.S. Department of Agriculture (2012) Economic Research Service Data Sets, Data Sets, U.S. Fertilizer
Imports/Exports: Standard Tables. Available online at: .
U.S. ITC (2002) United States International Trade Commission Interactive Tariff and Trade DataWeb, Version 2.5.0.
Available online at: . August 2002.
United States Geological Survey (USGS) (1994 through 2019a) Minerals Yearbook: Nitrogen. Available online at:
.
USGS (2019b and 2020) Minerals Commodity Summaries: Nitrogen (Fixed)-Ammonia. Available online at:
.
Nitric Acid Production
Climate Action Reserve (CAR) (2013) Project Report. Available online at:
. Accessed on 18 January 2013.
Desai (2012) Personal communication. Mausami Desai, U.S. Environmental Protection Agency, January 25, 2012.
EPA (2020) Greenhouse Gas Reporting Program (GHGRP). Aggregation of Reported Facility Level Data under
Subpart V -National Nitric Acid Production for Calendar Years 2010 through 2019. Office of Air and Radiation,
Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
EPA (2013) Draft Nitric Acid Database. U.S. Environmental Protection Agency, Office of Air and Radiation.
September 2010.
EPA (2012) Memorandum from Mausami Desai, U.S. EPA to Mr. Bill Herz, The Fertilizer Institute. November 26,
2012.
EPA (2010) Available and Emerging Technologies for Reducing Greenhouse Gas Emissions from the Nitric Acid
Production Industry. Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. Research
Triangle Park, NC. December 2010. Available online at: .
EPA (1998) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC. February 1998.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
U.S. Census Bureau (2010a) Current Industrial Reports. Fertilizers and Related Chemicals: 2009. "Table 1: Summary
of Production of Principle Fertilizers and Related Chemicals: 2009 and 2008." June, 2010. MQ325B(08)-5. Available
online at: .
U.S. Census Bureau (2010b) Personal communication between Hilda Ward (of U.S. Census Bureau) and Caroline
Cochran (of ICF International). October 26, 2010 and November 5, 2010.
U.S. Census Bureau (2009) Current Industrial Reports. Fertilizers and Related Chemicals: 2008. 'Table 1: Shipments
and Production of Principal Fertilizers and Related Chemicals: 2004 to 2008." June, 2009. MQ325B(08)-5. Available
online at: .
10-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
U.S. Census Bureau (2008) Current Industrial Reports. Fertilizers and Related Chemicals: 2007. 'Table 1: Shipments
and Production of Principal Fertilizers and Related Chemicals: 2003 to 2007." June, 2008. MQ325B(07)-5. Available
online at: .
Adipic Acid Production
ACC (2020) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
C&EN (1995) "Production of Top 50 Chemicals Increased Substantially in 1994." Chemical & Engineering News,
73(15):17. April 10,1995.
C&EN (1994) 'Top 50 Chemicals Production Rose Modestly Last Year." Chemical & Engineering News, 72(15): 13.
April 11,1994.
C&EN (1993) 'Top 50 Chemicals Production Recovered Last Year." Chemical & Engineering News, 71(15):11. April
12,1993.
C&EN (1992) "Production of Top 50 Chemicals Stagnates in 1991." Chemical & Engineering News, 70(15): 17. April
13,1992.
CMR (2001) "Chemical Profile: Adipic Acid." Chemical Market Reporter. July 16, 2001.
CMR (1998) "Chemical Profile: Adipic Acid." Chemical Market Reporter. June 15,1998.
CW (2005) "Product Focus: Adipic Acid." Chemical Week. May 4, 2005.
CW (1999) "Product Focus: Adipic Acid/Adiponitrile." Chemical Week, p. 31. March 10,1999.
Desai (2010, 2011) Personal communication. Mausami Desai, U.S. Environmental Protection Agency and Adipic
Acid Plant Engineers. 2010 and 2011.
EPA (2019, 2020) Greenhouse Gas Reporting Program. Subpart E, S-CEMS, BB, CC, LL Data Set (XLSX) (Adipic Acid
Tab). Office of Air and Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency,
Washington, D.C. Available online at: .
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
EPA (2014 through 2018) Greenhouse Gas Reporting Program. Subpart E, S-CEMS, BB, CC, LL Data Set (XLSX)
(Adipic Acid Tab). Office of Air and Radiation, Office of Atmospheric Programs, U.S. Environmental Protection
Agency, Washington, D.C. Available online at: .
EPA (2010 through 2013) Analysis of Greenhouse Gas Reporting Program data - Subpart E (Adipic Acid), Office of
Air and Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
ICIS (2007) "Adipic Acid." ICIS Chemical Business Americas. July 9, 2007.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Reimer, R.A., Slaten, C.S., Seapan, M., Koch, T.A. and Triner, V.G. (1999) "Implementation of Technologies for
Abatement of N20 Emissions Associated with Adipic Acid Manufacture." Proceedings of the 2nd Symposium on
Non-C02 Greenhouse Gases (NCGG-2), Noordwijkerhout, The Netherlands, 8-10 Sept. 1999, Ed. J. van Ham et al.,
Kluwer Academic Publishers, Dordrecht, pp. 347-358.
Thiemens, M.H., and W.C. Trogler (1991) "Nylon production; an unknown source of atmospheric nitrous oxide."
Science 251:932-934.
References 10-29

-------
Caprolactam, Glyoxal and Glyoxylic Acid Production
ACC (2020) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
AdvanSix (2020). AdvanSix Hopewell Virginia Information Sheet. Retrieved from:
 on September 21, 2020.
BASF (2020). BASF: Freeport, Texas Fact Sheet. Retrieved from

on September 21, 2020.
Cline, D. (2019, September 9). Firm to Clean Up and Market Former Fibrant Site. The Augusta Chronicle. Retrieved
from .
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Carbide Production and Consumption
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
U.S. Census Bureau (2005 through 2020) USITC Trade DataWeb. Available online at: .
United States Geological Survey (USGS) (2020) 2017 Minerals Yearbook: Abrasives, Manufactured [Advance
Release], August 2020. U.S. Geological Survey, Reston, VA. Available online at:
.
USGS (2019a) Mineral Industry Surveys, Manufactured Abrasives in the First Quarter 2019, Table 1, July 2019 U.S.
Geological Survey, Reston, VA. Available online at: .
USGS (2020a) Mineral Industry Surveys, Manufactured Abrasives in the First Quarter 2020, Table 1, June 2020, U.S.
Geological Survey, Reston, VA. Available online at: .
USGS (2017c) USGS 2015 Minerals Yearbook Silicon [Advance Release], November 2017. Table 4. U.S. Geological
Survey, Reston, VA. Available online at: .
USGS (2019) Mineral Commodity Summaries: Abrasives (Manufactured), February 2019. Available online at:
.
USGS (1991a through 2017) Minerals Yearbook: Manufactured Abrasives Annual Report. U.S. Geological Survey,
Reston, VA. Available online at: .
USGS (1991b through 2015) Minerals Yearbook: Silicon Annual Report. U.S. Geological Survey, Reston, VA.
Available online at: .
Titanium Dioxide Production
Gambogi, J. (2002) Telephone communication. Joseph Gambogi, Commodity Specialist, U.S. Geological Survey and
Philip Groth, ICF International. November 2002.
10-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
United States Geological Survey (USGS) (2020) Mineral Commodity Summaries: Titanium and Titanium Dioxide.
U.S. Geological Survey, Reston, Va. January 2020. Available online at:
.
USGS (1991 through 2015) Minerals Yearbook: Titanium. U.S. Geological Survey, Reston, VA.
Sod ^ a* h Production
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
United States Geological Survey (USGS) (2020a) Mineral Commodity Summary: Soda Ash. U.S. Geological Survey,
Reston, VA. Accessed September 2020.
USGS (2020b) Mineral Industry Surveys: Soda Ash in April 2020. U.S. Geological Survey, Reston, VA. Accessed
September 2020.
USGS (2019) Mineral Industry Surveys: Soda Ash in April 2019. U.S. Geological Survey, Reston, VA. Accessed August
2019.
USGS (2018a) Mineral Industry Surveys: Soda Ash in February 2018. U.S. Geological Survey, Reston, VA. Accessed
September 2018.
USGS (2017) Mineral Industry Surveys: Soda Ash in January 2017. U.S. Geological Survey, Reston, VA. March 2017.
USGS (2016) Mineral Industry Surveys: Soda Ash in November 2016. U.S. Geological Survey, Reston, VA. January
2017.
USGS (2015a) Mineral Industry Surveys: Soda Ash in July 2015. U.S. Geological Survey, Reston, VA. September
2015.
USGS (1994 through 2015b, 2018b) Minerals Yearbook: Soda Ash Annual Report. U.S. Geological Survey, Reston,
VA.
USGS (1995c) Trona Resources in the Green River Basin, Southwest Wyoming. U.S. Department of the Interior, U.S.
Geological Survey. Open-File Report 95-476. Wiig, Stephen, Grundy, W.D., Dyni, John R.
Petrochemical Production
ACC (2020) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
AN (2014) About Acrylonitrile: Production. AN Group, Washington, D.C. Available online at:
.
EPA Greenhouse Gas Reporting Program (2020) Aggregation of Reported Facility Level Data under Subpart X -
National Petrochemical Production for Calendar Years 2018 and 2019. Office of Air and Radiation, Office of
Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA Greenhouse Gas Reporting Program (2019) Aggregation of Reported Facility Level Data under Subpart X -
National Petrochemical Production for Calendar Years 2014 through 2017. Office of Air and Radiation, Office of
Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
References 10-31

-------
EPA (2008) Technical Support Document for the Petrochemical Production Sector: Proposed Rule for Mandatory
Reporting of Greenhouse Gases. U.S. Environmental Protection Agency. September 2008.
EPA (2000) Economic Impact Analysis for the Proposed Carbon Black Manufacturing NESHAP, U.S. Environmental
Protection Agency. Research Triangle Park, NC. EPA-452/D-00-003. May 2000.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Johnson, G. L. (2005 through 2010) Personal communication. Greg Johnson of Liskow& Lewis, on behalf of the
International Carbon Black Association (ICBA) and Caroline Cochran, ICF International. September 2010.
Johnson, G. L. (2003) Personal communication. Greg Johnson of Liskow& Lewis, on behalf of the International
Carbon Black Association (ICBA) and Caren Mintz, ICF International. November 2003.
HCf-C*17	tiOfi
ARAP (2010) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. September 10, 2010.
ARAP (2009) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. September 21, 2009.
ARAP (2008) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. October 17, 2008.
ARAP (2007) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. October 2, 2007.
ARAP (2006) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Sally Rand of the U.S. Environmental Protection Agency. July 11, 2006.
ARAP (2005) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. August 9, 2005.
ARAP (2004) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. June 3, 2004.
ARAP (2003) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Sally Rand of the U.S. Environmental Protection Agency. August 18, 2003.
ARAP (2002) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. August 7, 2002.
ARAP (2001) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Deborah Ottinger of the U.S. Environmental Protection Agency. August 6, 2001.
ARAP (2000) Electronic mail communication from Dave Stirpe, Executive Director, Alliance for Responsible
Atmospheric Policy to Sally Rand of the U.S. Environmental Protection Agency. August 13, 2000.
ARAP (1999) Facsimile from Dave Stirpe, Executive Director, Alliance for Responsible Atmospheric Policy to
Deborah Ottinger Schaefer of the U.S. Environmental Protection Agency. September 23,1999.
ARAP (1997) Letter from Dave Stirpe, Director, Alliance for Responsible Atmospheric Policy to Elizabeth Dutrow of
the U.S. Environmental Protection Agency. December 23,1997.
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
10-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
IPCC (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. [S. Solomon, D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
RTI (2008) "Verification of Emission Estimates of HFC-23 from the Production of HCFC-22:Emissions from 1990
through 2006." Report prepared by RTI International for the Climate Change Division. March 2008.
RTI (1997) "Verification of Emission Estimates of HFC-23 from the Production of HCFC-22: Emissions from 1990
through 1996." Report prepared by Research Triangle Institute for the Cadmus Group. November 25,1997; revised
February 16,1998.
UNFCCC (2014) Report of the Conference of the Parties on its nineteenth session, held in Warsaw from 11 to 23
November 2013. United Nations Framework Convention on Climate Change, Warsaw. (FCCC/CP/2013/10/Add.3).
January 31, 2014. Available online at: .
Carbon Dioxide Consumption
ARI (1990 through 2010) C02 Use in Enhanced Oil Recovery. Deliverable to ICF International under Task Order 102,
July 15, 2011.
ARI (2007) C02-E0R: An Enabling Bridge for the Oil Transition. Presented at "Modeling the Oil Transition—a
DOE/EPA Workshop on the Economic and Environmental Implications of Global Energy Transitions." Washington,
D.C. April 20-21, 2007.
ARI (2006) C02-E0R: An Enabling Bridge for the Oil Transition. Presented at "Modeling the Oil Transition—a
DOE/EPA Workshop on the Economic and Environmental Implications of Global Energy Transitions." Washington,
D.C. April 20-21, 2006.
Broadhead (2003) Personal communication. Ron Broadhead, Principal Senior Petroleum Geologist and Adjunct
faculty, Earth and Environmental Sciences Department, New Mexico Bureau of Geology and Mineral Resources,
and Robin Petrusak, ICF International. September 5, 2003.
COGCC (2014) Monthly C02 Produced by County (1999-2009). Available online at:
. Accessed October
2014.
Denbury Resources Inc. (2002 through 2010) Annual Report: 2001 through 2009, Form 10-K. Available online at:
.
Accessed September 2014.
EPA Greenhouse Gas Reporting Program (2020). Aggregation of Reported Facility Level Data under Subpart PP -
National Level C02 Transferred for Food & Beverage Applications for Calendar Years 2010 through 2019. Office of
Air and Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
New Mexico Bureau of Geology and Mineral Resources (2006) Natural Accumulations of Carbon Dioxide in New
Mexico and Adjacent Parts of Colorado and Arizona: Commercial Accumulation of C02. Available online at:
.
References 10-33

-------
Phosphoric Acid Production
EFMA (2000) "Production of Phosphoric Acid." Best Available Techniques for Pollution Prevention and Control in
the European Fertilizer Industry. Booklet 4 of 8. European Fertilizer Manufacturers Association. Available online at:
.
FIPR (2003a) "Analyses of Some Phosphate Rocks." Facsimile Gary Albarelli, the Florida Institute of Phosphate
Research, Bartow, Florida, to Robert Lanza, ICF International. July 29, 2003.
FIPR (2003b) Florida Institute of Phosphate Research. Personal communication. Mr. Michael Lloyd, Laboratory
Manager, FIPR, Bartow, Florida, to Mr. Robert Lanza, ICF International. August 2003.
Golder Associates and M3 Engineering, Bayovar 12 Phosphate Project: Nl 43-101 Updated Pre-Feasibility Study,
Issued June 28, 2016. Available at:
. Accessed on October 7, 2020.
NCDENR (2013) North Carolina Department of Environment and Natural Resources, Title V Air Permit Review for
PCS Phosphate Company, Inc. - Aurora. Available online at:
. Accessed on January 25, 2013.
United States Geological Survey (USGS) (2020) Mineral Commodity Summaries: Phosphate Rock 2020. January
2020. U.S. Geological Survey, Reston, VA. Accessed September 2020. Available online at:
.
USGS (2019) Mineral Commodity Summaries: Phosphate Rock 2019. February 2019. U.S. Geological Survey, Reston,
VA. Accessed August 2019. Available online at: .
USGS (2019b) Communication between Stephen Jasinski (USGS) and EPA on November 15, 2019.
USGS (2018) Mineral Commodity Summaries: Phosphate Rock 2018. January 2018. U.S. Geological Survey, Reston,
VA. Available online at: .
USGS (2017) Mineral Commodity Summaries: Phosphate Rock 2017. January 2017. U.S. Geological Survey, Reston,
VA. Available online at: .
USGS (2016) Mineral Commodity Summaries: Phosphate Rock 2016. January 2016. U.S. Geological Survey, Reston,
VA. Available online at: .
USGS (1994 through 2015b) Minerals Yearbook. Phosphate Rock Annual Report. U.S. Geological Survey, Reston, VA.
USGS (2012) Personal communication between Stephen Jasinski (USGS) and Mausami Desai (EPA) on October 12,
2012.
Iron and Steel Production and Metallurgical Coke Production
American Coke and Coal Chemicals Institute (ACCCI) (2020) U.S. & Canadian Coke Plants as of February 2020.
ACCCI, Washington, D.C. February 2020.
American Iron and Steel Institute (AISI) (2004 through 2020) Annual Statistical Report, American Iron and Steel
Institute, Washington, D.C.
AISI (2006 through 2017) Personal communication, Mausami Desai, U.S. Environmental Protection Agency, and
American Iron and Steel Institute, December 2017.
AISI (2008) Personal communication, Mausami Desai, U.S. Environmental Protection Agency, and Bruce Steiner,
Technical Consultant with the American Iron and Steel Institute, October 2008.
10-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Carroll (2016) Personal communication, Mausami Desai, U.S. Environmental Protection Agency, and Colin P.
Carroll, Director of Environment, Health and Safety, American Iron and Steel Institute, December 2016.
Carroll (2017) Personal communication, John Steller, U.S. Environmental Protection Agency, and Colin P. Carroll,
Director of Environment, Health and Safety, American Iron and Steel Institute, November 2017.
DOE (2000) Energy and Environmental Profile of the U.S. Iron and Steel Industry. Office of Industrial Technologies,
U.S. Department of Energy. August 2000. DOE/EE-0229.EIA.
EIA (1998 through 2019) Quarterly Coal Report: October-December, Energy Information Administration, U.S.
Department of Energy. Washington, D.C.
EIA (2020) Natural Gas Annual 2019. Energy Information Administration, U.S. Department of Energy. Washington,
D.C. September 2020.
EIA (2017c) Monthly Energy Review, December 2017, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2015/12).EIA (1992) Coal and lignite production. EIA State Energy Data
Report 1992, Energy Information Administration, U.S. Department of Energy, Washington, D.C.
EPA (2010) Carbon Content Coefficients Developed for EPA's Mandatory Reporting Rule. Office of Air and
Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC/UNEP/OECD/IEA (1995) "Volume 3: Greenhouse Gas Inventory Reference Manual. Table 2-2." IPCC Guidelines
for National Greenhouse Gas Inventories. Intergovernmental Panel on Climate Change, United Nations
Environment Programme, Organization for Economic Co-Operation and Development, International Energy
Agency. IPCC WG1 Technical Support Unit, United Kingdom.
Tuck (2020) Personal communication. Christopher Tuck, Commodity Specialist, U.S. Geological Survey and Amanda
Chiu, U.S. Environmental Protection Agency. November 16, 2020.
USGS (2017) 2017 USGS Minerals Yearbook- Iron and Steel. U.S. Geological Survey, Reston, VA.
USGS (1991 through 2017) USGS Minerals Yearbook - Iron and Steel Scrap. U.S. Geological Survey, Reston, VA.
Ferroalloy Production
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Onder, H., and E.A. Bagdoyan (1993) Everything You've Always Wanted to Know about Petroleum Coke. Allis
Mineral Systems.
United States Geological Survey (USGS) (2020) 2016 Minerals Yearbook: Ferroalloys (Advanced Release). U.S.
Geological Survey, Reston, VA. January 2020.
USGS (2019) Mineral Industry Surveys: Silicon in May 2019. U.S. Geological Survey, Reston, VA. August 2019.
USGS (2018a) 2015 Minerals Yearbook: Ferroalloys. U.S. Geological Survey, Reston, VA. May 2018.
USGS (2018b) Mineral Industry Surveys: Silicon in July 2018. U.S. Geological Survey, Reston, VA. September 2018.
USGS (2017) Mineral Industry Surveys: Silicon in April 2017. U.S. Geological Survey, Reston, VA. June 2017.
United States Geological Survey (USGS) (2016a) 2014 Minerals Yearbook: Ferroalloys. U.S. Geological Survey,
Reston, VA. October 2016.
References 10-35

-------
USGS (2016b) Mineral Industry Surveys: Silicon in December 2016. U.S. Geological Survey, Reston, VA. December
2016.
USGS (2015a) 2012 Minerals Yearbook: Ferroalloys. U.S. Geological Survey, Reston, VA. April 2015.
USGS (2015b) Mineral Industry Surveys: Silicon in June 2015. U.S. Geological Survey, Reston, VA. September 2015.
USGS (2014) Mineral Industry Surveys: Silicon in September 2014. U.S. Geological Survey, Reston, VA. December
2014.
USGS (1996 through 2013) Minerals Yearbook: Silicon. U.S. Geological Survey, Reston, VA.
Aluminum Production
EPA (2020) Greenhouse Gas Reporting Program (GHGRP). Envirofacts, Subpart: F Aluminum Production. Available
online at: .
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
IPPC (2019) 2019 Refinement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National
Greenhouse Gas Inventories Programme, The Intergovernmental Panel on Climate Change. [Calvo Buendia, E.,
Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize, S., Osako, A., Pyrozhenko, Y., Shermanau, P. and
Federici, S. (eds.)]. Hayama, Kanagawa, Japan.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
USAA (2020) U.S. Primary Aluminum Production: Report for August 2020. U.S. Aluminum Association, Washington,
D.C. September 2020.
USAA (2019) U.S. Primary Aluminum Production: Report for August 2019. U.S. Aluminum Association, Washington,
D.C. September 2019.
USAA (2018) U.S. Primary Aluminum Production: Report for August 2018. U.S. Aluminum Association, Washington,
D.C. September 2018.
USAA (2017) U.S. Primary Aluminum Production: Report for September 2017. U.S. Aluminum Association,
Washington, D.C. October 2017.
USAA (2016a) U.S. Primary Aluminum Production: Report for February 2016. U.S. Aluminum Association,
Washington, D.C. March 2016.
USAA (2016b) U.S. Primary Aluminum Production: Report for August 2016. U.S. Aluminum Association,
Washington, D.C. August 2016.
USAA (2015) U.S. Primary Aluminum Production: Report for June 2015. U.S. Aluminum Association, Washington,
D.C.July 2015.
USAA (2014) U.S. Primary Aluminum Production 2013. U.S. Aluminum Association, Washington, D.C. October 2014.
USAA (2013) U.S. Primary Aluminum Production 2012. U.S. Aluminum Association, Washington, D.C. January 2013.
USAA (2012) U.S. Primary Aluminum Production 2011. U.S. Aluminum Association, Washington, D.C. January 2012.
USAA (2011) U.S. Primary Aluminum Production 2010. U.S. Aluminum Association, Washington, D.C.
USAA (2010) U.S. Primary Aluminum Production 2009. U.S. Aluminum Association, Washington, D.C.
USAA (2008, 2009) U.S. Primary Aluminum Production. U.S. Aluminum Association, Washington, D.C.
USAA (2004, 2005, 2006) Primary Aluminum Statistics. U.S. Aluminum Association, Washington, D.C.
10-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
USGS (2019) 2017 Mineral Yearbook: Aluminum. U.S. Geological Survey, Reston, VA.
USGS (2020) 2019 Mineral Commodity Summaries: Aluminum. U.S. Geological Survey, Reston, VA.
USGS (2007) 2006 Mineral Yearbook: Aluminum. U.S. Geological Survey, Reston, VA.
USGS (1995,1998, 2000, 2001, 2002) Minerals Yearbook: Aluminum Annual Report. U.S. Geological Survey, Reston,
VA.
Magnesium Production and Processing
ARB (2015) "Magnesium casters successfully retool for a cleaner future." California Air Resources Board News
Release. Release # 15-07. February 5, 2015. Accessed October 2017. Available online at:
.
Bartos S., C. Laush, J. Scharfenberg, and R. Kantamaneni (2007) "Reducing greenhouse gas emissions from
magnesium die castingJournal of Cleaner Production, 15: 979-987, March.
EPA (2020) Envirofacts. Greenhouse Gas Reporting Program (GHGRP), Subpart T: Magnesium Production and
Processing. Available online at: . Accessed on October 2020.
Gjestland, H. and D. Magers (1996) "Practical Usage of Sulphur [Sulfur] Hexafluoride for Melt Protection in the
Magnesium Die Casting Industry." #13,1996 Annual Conference Proceedings, International Magnesium
Association. Ube City, Japan.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
RAND (2002) RAND Environmental Science and Policy Center, "Production and Distribution of SF6 by End-Use
Applications" Katie D. Smythe. International Conference on SF6 and the Environment: Emission Reduction
Strategies. San Diego, CA. November 21-22, 2002.
USGS (2020, 2017, 2016, 2015, 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005a, 2003, 2002)
Minerals Yearbook: Magnesium Annual Report. U.S. Geological Survey, Reston, VA. Available online at:
.
USGS (2010b) Mineral Commodity Summaries: Magnesium Metal. U.S. Geological Survey, Reston, VA. Available
online at: .
USGS (2005b) Personal Communication between Deborah Kramer of the USGS and Jeremy Scharfenberg of ICF
Consulting.
Lead Production
Dutrizac, J.E., V. Ramachandran, and J.A. Gonzalez (2000) Lead-Zinc 2000. The Minerals, Metals, and Materials
Society.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Morris, D., F.R. Steward, and P. Evans (1983) Energy Efficiency of a Lead Smelter. Energy 8(5):337-349.
Sjardin, M. (2003) C02 Emission Factors for Non-Energy Use in the Non-Ferrous Metal, Ferroalloys and Inorganics
Industry. Copernicus Institute. Utrecht, the Netherlands.
Ullman (1997) Ullman's Encyclopedia of Industrial Chemistry: Fifth Edition. Volume A5. John Wiley and Sons.
References 10-37

-------
United States Geological Survey (USGS) (2020) 2020 Mineral Commodity Summary, Lead. U.S. Geological Survey,
Reston, VA. February 2020. Available online at: .
USGS (2019) 2019 Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. February 2019.
USGS (2018) 2018 Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2018.
USGS (2017) 2017 Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2017.
USGS (2016) 2016 Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2016.
USGS (2015) 2015 Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2015.
USGS (2014) Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. February 2014.
USGS (1995 through 2013) Minerals Yearbook: Lead Annual Report. U.S. Geological Survey, Reston, VA.
Zinc Production
AZR (2021) Summary of Company History. Available online at < https://azr.com/our-history/>. Accessed on March
16, 2021.
AZR (2020) Personal communication. Erica Livingston, Environmental Affairs Manager, American Zinc Recycling
Corp. and Amanda Chiu, U.S. Environmental Protection Agency. October 29, 2020.
Horsehead Corp. (2016) Form 10-k, Annual Report for the Fiscal Year Ended December 31, 2015. Available online
at: . Submitted
on January 25, 2017.
Horsehead Corp. (2015) Form 10-k, Annual Report for the Fiscal Year Ended December 31, 2014. Available online
at: .
Submitted on March 2, 2015.
Horsehead Corp. (2014) Form 10-k, Annual Report for the Fiscal Year Ended December 31, 2013. Available online
at: .
Submitted on March 13, 2014.
Horsehead Corp. (2013) Form 10-k, Annual Report for the Fiscal Year Ended December 31, 2012. Available online
at: . Submitted March 18, 2013.
Horsehead Corp. (2012a) Form 10-k, Annual Report for the Fiscal Year Ended December 31, 2011. Available online
at: . Submitted on
March 9, 2012.
Horsehead Corp. (2012b) Horsehead's New Zinc Plant and its Impact on the Zinc Oxide Business. February 22, 2012.
Available online at: . Accessed on September
10, 2015.
Horsehead Corp. (2011) 10-k Annual Report for the Fiscal Year Ended December 31, 2010. Available online at:
. Submitted on March 16, 2011.
Horsehead Corp. (2010a) 10-k Annual Report for the Fiscal Year Ended December 31, 2009. Available online at:
. Submitted on March 16, 2010.
Horsehead Corp. (2010b) Horsehead Holding Corp. Provides Update on Operations at its Monaco, PA Plant. July 28,
2010. Available online at: .
Horsehead Corp (2009) 10-k Annual Report for the Fiscal Year Ended December 31, 2008. Available online at: <
https://www.sec.gov/Archives/edgar/data/1385544/000095015209002674/l35087ael0vk.htm>. Submitted on
March 16, 2009.
10-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Horsehead Corp (2008) 10-k Annual Report for the Fiscal Year Ended December 31, 2007. Available online at:
. Submitted on March 31, 2008.
Horsehead Corp (2007) Registration Statement (General Form) S-l. Available online at . Submitted on April 13, 2007.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Nyrstar (2017) 2016 Clarksville Fact Sheet. Available online at:
. Accessed
on September 27, 2017.
Nyrstar (2016) 2015 Clarksville Fact Sheet.
PIZO (2017) Available online at . Accessed on January 12, 2017.
PIZO (2014) Available online at . Accessed on December 9, 2014.
PIZO (2012) Available online at . Accessed on October 10, 2012.
PIZO (2021) Personal communication. Thomas Rheaume, Arkansas Department of Environment and Environment
and Amanda Chiu, U.S. Environmental Protection Agency. February 16, 2021.
Recycling Today (2020) "AZR to restart for zinc recycling plant in North Carolina." March 6, 2020.
.
Accessed October 10, 2020.
Steel Dust Recycling (SDR) (2021) Personal communication. Jeremy Whitten, EHS Manager, Steel Dust Recycling
LLC and Amanda Chiu, U.S. Environmental Protection Agency. January 8, 2021.
SDR (2018) Personal communication. Jeremy Whitten, EHS Manager, Steel Dust Recycling LLC and John Steller, U.S.
Environmental Protection Agency. October 25, 2018.
SDR (2017) Personal communication. Jeremy Whitten, EHS Manager, Steel Dust Recycling LLC and John Steller, U.S.
Environmental Protection Agency. January 26, 2017.
SDR (2015) Personal communication. Jeremy Whitten, EHS Manager, Steel Dust Recycling LLC and Gopi Manne,
Eastern Research Group, Inc. September 22, 2015.
SDR (2014) Personal communication. Art Rowland, Plant Manager, Steel Dust Recycling LLC and Gopi Manne,
Eastern Research Group, Inc. December 9, 2014.
SDR (2013) Available online at . Accessed on October 29, 2013.
SDR (2012) Personal communication. Art Rowland, Plant Manager, Steel Dust Recycling LLC and Gopi Manne,
Eastern Research Group, Inc. October 5, 2012.
Sjardin (2003) C02 Emission Factors for Non-Energy Use in the Non-Ferrous Metal, Ferroalloys and Inorganics
Industry. Copernicus Institute. Utrecht, the Netherlands.
United States Geological Survey (USGS) (2020) 2020 Mineral Commodity Summary: Zinc. U.S. Geological Survey,
Reston, VA. February 2020. Available online at: .
USGS (2019) 2019 Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2019
USGS (2018) 2018 Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2018.
USGS (2017) 2017 Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2017.
USGS (2016) 2016 Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2016.
USGS (2015) 2015 Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2015.
References 10-39

-------
USGS (1995 through 2014) Minerals Yearbook: Zinc Annual Report. U.S. Geological Survey, Reston, VA.
Viklund-White (2000) The use ofLCAfor the environmental evaluation of the recycling of galvanized steel. ISIJ
International, Vol. 40. No. 3, pp 292-299.
Electronics Industry
Burton, C.S., and R. Beizaie (2001) "EPA's PFC Emissions Model (PEVM) v. 2.14: Description and Documentation"
prepared for Office of Global Programs, U. S. Environmental Protection Agency, Washington, DC. November 2001.
Citigroup Smith Barney (2005) Global Supply/Demand Model for Semiconductors. March 2005.
DisplaySearch. 2010. DisplaySearch Q4'09 Quarterly FPD Supply/Demand and Capital Spending Report.
DisplaySearch, LLC.
Doering, R. and Nishi, Y (2000) "Handbook of Semiconductor Manufacturing Technology", Marcel Dekker, New
York, USA, 2000.
EPA (2006) Uses and Emissions of Liquid PFC Heat Transfer Fluids from the Electronics Sector. U.S. Environmental
Protection Agency, Washington, DC. EPA-430-R-06-901.
EPA Greenhouse Gas Reporting Program (GHGRP) Envirofacts. Subpart I: Electronics Manufacture. Available online
at: .
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC (2019) 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National
Greenhouse Gas Inventories Programme, The Intergovernmental Panel on Climate Change. Calvo Buendia, E.,
Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize, S., Osako, A., Pyrozhenko, Y., Shermanau, P. and
Federici, S. (eds). Published: IPCC, Switzerland.
ISMI (2009) Analysis of Nitrous Oxide Survey Data. Walter Worth. June 8, 2009. Available online at:
.
ITRS (2007, 2008, 2011, 2013) International Technology Roadmap for Semiconductors: 2006 Update, January 2007;
International Technology Roadmap for Semiconductors: 2007 Edition, January 2008; International Technology
Roadmap for Semiconductors: 2011, January 2012; Update, International Technology Roadmap for
Semiconductors: 2013 Edition, Available online at: . These
and earlier editions and updates are available online at: . Information about the number of
interconnect layers for years 1990-2010 is contained in Burton and Beizaie, 2001. PEVM is updated using new
editions and updates of the ITRS, which are published annually. SEMI - Semiconductor Equipment and Materials
Industry (2017) World Fab Forecast, August 2018 Edition.
Platzer, Michaela D. (2015) U.S. Solar Photovoltaic Manufacturing: Industry Trends, Global Competition, Federal
Support. Congressional Research Service. January 27, 2015. .
SEMI - Semiconductor Equipment and Materials Industry (2018) World Fab Forecast, June 2018 Edition.
SEMI - Semiconductor Equipment and Materials Industry (2016) World Fab Forecast, May 2017 Edition.
SEMI - Semiconductor Equipment and Materials Industry (2013) World Fab Forecast, May 2013 Edition.
SEMI - Semiconductor Equipment and Materials Industry (2012) World Fab Forecast, August 2012 Edition.
Semiconductor Industry Association (SIA) (2009-2011) STATS: SICAS Capacity and Utilization Rates Q1-Q4 2008, Ql-
Q4 2009, Q1-Q4 2010. Available online at:
.
10-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
United States Census Bureau (USCB) (2011, 2012, 2015, 2016, 2017, 2018, 2019) Historical Data: Quarterly Survey
of Plant Capacity Utilization. Available online at: .
VLSI Research, Inc. (2012) Worldwide Silicon Demand. August 2012.
Substitution of Ozone Depleting Substances
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
EPA (2021a) Suppliers of Industrial GHGs and Products Containing GHGs. Greenhouse Gas Reporting Program.
Available online at: .
EPA (2021b). Summary of Updates to the Unitary Air Conditioning End-Use in the Vintaging Model. Prepared for
U.S. EPA's Stratospheric Protection Division by ICF under EPA Contract Number EP-BPA-16-H-0021.
EPA (2020a). Summary of Research and Proposed Assumptions for Adding a New Walk-in Cooler Panel Foam End-
Use to the Vintaging Model. Prepared for U.S. EPA's Stratospheric Protection Division by ICF under EPA Contract
Number EP-BPA-16-H-0021. July 29, 2020.
EPA (2020b). Summary of Research and Proposed Assumptions for Adding a New Display Case Insulation Foam
End-Use to the Vintaging Model. Prepared for U.S. EPA's Stratospheric Protection Division by ICF under EPA
Contract Number EP-BPA-16-H-0021. August 21, 2020.
EPA (2020c). Summary of Research and Proposed Assumptions for Adding Road Transport Insulation Foam and
Intermodal Container Insulation Foam End-Uses to the Vintaging Model. Prepared for U.S. EPA's Stratospheric
Protection Division by ICF under EPA Contract Number EP-BPA-16-H-0021. August 26, 2020.
EPA (2020d). Summary of Updates to the Ice Makers End-Use in the Vintaging Model. Prepared for U.S. EPA's
Stratospheric Protection Division by ICF under EPA Contract No. EP-BPA-16-H-0021. September 23, 2020.
EPA (2020e). Summary of Research and Proposed Assumptions for Adding a New Refrigerated Food Processing and
Dispensing Equipment Insulation Foam End-Use to the Vintaging Model. Prepared for U.S. EPA's Stratospheric
Protection Division by ICF under EPA Contract Number EP-BPA-16-H-0021. August 12, 2020.
EPA (2020f). Summary of Research and Proposed Assumptions for Adding a New Ice Machine Insulation Foam End-
Use to the Vintaging Model. Prepared for U.S. EPA's Stratospheric Protection Division by ICF under EPA Contract
Number EP-BPA-16-H-0021. August 12, 2020.
EPA (2020g). Summary of Research and Proposed Assumptions for Adding a New Stand-alone Equipment
Insulation Foam End-Use to the Vintaging Model. Prepared for U.S. EPA's Stratospheric Protection Division by ICF
under EPA Contract Number EP-BPA-16-H-0021. August 12, 2020.
EPA (2020h). Summary of Research and Proposed Assumptions for Adding a New Vending Machine Insulation
Foam End-Use to the Vintaging Model. Prepared for U.S. EPA's Stratospheric Protection Division by ICF under EPA
Contract Number EP-BPA-16-H-0021. August 12, 2020.
EPA (2020i). Observed Trends for HFC-227ea Emissions in the United States and HFC-227ea and HFC-134a
Emissions from the MDI Aerosols End-Use in EPA's Vintaging Model. Prepared for U.S. EPA's Stratospheric
Protection Division by ICF under EPA Contract Number EP-BPA-16-H-0021. June 24, 2020.
EPA (2020j). Summary of Research and Proposed Updates to the PU and PIR Boardstock End-Use in the Vintaging
Model. Prepared for U.S. EPA's Stratospheric Protection Division by ICF under EPA Contract Number EP-BPA-16-H-
0021. November 18, 2020.
EPA (2020k). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018. U.S. Environmental Protection
Agency, April 2020. Available at: .
References 10-41

-------
EPA (2018) EPA's Vintaging Model of ODS Substitutes: A Summary of the 2017 Peer Review. Office of Air and
Radiation. Document Number EPA-400-F-18-001. Available online at:
.
EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
Electrical Transmission and Distribution
IPCC (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. S. Solomon, D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.). Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
IPCC (1996) Climate Change 1995: The Science of Climate Change. Intergovernmental Panel on Climate Change, J.T.
Houghton, L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg, and K. Maskell (eds.). Cambridge University
Press. Cambridge, United Kingdom.
Levin et al. (2010) 'The Global SF6Source Inferred from Long-term High Precision Atmospheric Measurements and
its Comparison with Emission Inventories." Atmospheric Chemistry and Physics, 10: 2655-2662.
O'Connell, P., F. Heil, J. Henriot, G. Mauthe, H. Morrison, L. Neimeyer, M. Pittroff, R. Probst, J.P. Tailebois (2002)
SF6 in the Electric Industry, Status 2000, CIGRE. February 2002.
RAND (2004) 'Trends in SF6 Sales and End-Use Applications: 1961-2003," Katie D. Smythe. International Conference
on SFs and the Environment: Emission Reduction Strategies. RAND Environmental Science and Policy Center,
Scottsdale, AZ. December 1-3, 2004.
UDI (2017) 2017 UDI Directory of Electric Power Producers and Distributors, 125th Edition, Platts.
UDI (2013) 2013 UDI Directory of Electric Power Producers and Distributors,121st Edition, Platts.
UDI (2010) 2010 UDI Directory of Electric Power Producers and Distributors, 118th Edition, Platts.
UDI (2007) 2007 UDI Directory of Electric Power Producers and Distributors, 115th Edition, Platts.
UDI (2004) 2004 UDI Directory of Electric Power Producers and Distributors, 112th Edition, Platts.
UDI (2001) 2001 UDI Directory of Electric Power Producers and Distributors, 109th Edition, Platts.
UNFCCC (2014) Report of the Conference of the Parties on its nineteenth session, held in Warsaw from 11 to 23
November 2013. United Nations Framework Convention on Climate Change, Warsaw. (FCCC/CP/2013/10/Add.3).
January 31, 2014. Available online at: .
Nitrous Oxide from Product Use
CGA (2003) "CGA Nitrous Oxide Abuse Hotline: CGA/NWSA Nitrous Oxide Fact Sheet." Compressed Gas
Association. November 3, 2003.
CGA (2002) "CGA/NWSA Nitrous Oxide Fact Sheet." Compressed Gas Association. March 25, 2002.
Heydorn, B. (1997) "Nitrous Oxide—North America." Chemical Economics Handbook, SRI Consulting. May 1997.
IPCC (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. [S. Solomon, D. Qin, M. Manning, Z. Chen,
10-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Ottinger (2021) Personal communication. Deborah Ottinger, U.S. Environmental Protection Agency and Amanda
Chiu, U.S. Environmental Protection Agency. January 7, 2021.
Tupman, M. (2003) Personal communication. Martin Tupman, Airgas Nitrous Oxide and Daniel Lieberman, ICF
International. August 8, 2003.
Industrial Processes and Product Use Sources of Precursor
Greenhouse Gases
EPA (2020) "Criteria pollutants National Tier 1 for 1970 - 2019." National Emissions Inventory (NEI) Air Pollutant
Emissions Trends Data. Office of Air Quality Planning and Standards, April 2020. Available online at:
.
EPA (2003) Email correspondence containing preliminary ambient air pollutant data. Office of Air Pollution and the
Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. December 22, 2003.
EPA (1997) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC. October 1997.
Agriculture
Enteric Fermentation
Archibeque, S. (2011) Personal Communication. Shawn Archibeque, Colorado State University, Fort Collins,
Colorado and staff at ICF International.
Crutzen, P.J., I. Aselmann, and W. Seiler (1986) Methane Production by Domestic Animals, Wild Ruminants, Other
Herbivores, Fauna, and Humans. Tellus, 38B:271-284.
Donovan, K. (1999) Personal Communication. Kacey Donovan, University of California at Davis and staff at ICF
International.
Doren, P.E., J. F. Baker, C. R. Long and T. C. Cartwright (1989) Estimating Parameters of Growth Curves of Bulls, J
Animal Science 67:1432-1445.
Enns, M. (2008) Personal Communication. Dr. Mark Enns, Colorado State University and staff at ICF International.
EPA (2002) Quality Assurance/Quality Control and Uncertainty Management Plan for the U.S. Greenhouse Gas
Inventory: Procedures Manual for Quality Assurance/Quality Control and Uncertainty Analysis, U.S. Greenhouse
Gas Inventory Program, U.S. Environmental Protection Agency, Office of Atmospheric Programs, EPA430-R-02-
007B, June 2002.
ERG (2016) Development of Methane Conversion Rate Scaling Factor and Diet-Related Inputs to the Cattle Enteric
Fermentation Model for Dairy Cows, Dairy Heifers, and Feedlot Animals. ERG, Lexington, MA. December 2016.
Galyean and Gleghorn (2001) Summary of the 2000 Texas Tech University Consulting Nutritionist Survey. Texas
Tech University. Available online at .
June 2009.
References 10-43

-------
Holstein Association (2010) History of the Holstein Breed (website). Available online at:
. Accessed September 2010.
ICF (2006) Cattle Enteric Fermentation Model: Model Documentation. Prepared by ICF International for the
Environmental Protection Agency. June 2006.
ICF (2003) Uncertainty Analysis of 2001 Inventory Estimates of Methane Emissions from Livestock Enteric
Fermentation in the U.S. Memorandum from ICF International to the Environmental Protection Agency. May 2003.
IPCC (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. S. Solomon, D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.). Cambridge University Press. Cambridge, United Kingdom
996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
IPCC (2019) 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The
Intergovernmental Panel on Climate Change. Calvo Buendia, E., Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M.,
Ngarize, S., Osako, A., Pyrozhenko, Y., Shermanau, P. and Federici, S. (eds). Hayama, Kanagawa, Japan.Johnson, D.
(2002) Personal Communication. Don Johnson, Colorado State University, Fort Collins, and ICF International.
Johnson, D. (1999) Personal Communication. Don Johnson, Colorado State University, Fort Collins, and David
Conneely, ICF International.
Kebreab E., K. A. Johnson, S. L. Archibeque, D. Pape, and T. Wirth (2008) Model for estimating enteric methane
emissions from United States dairy and feedlot cattle. J. Anim. Sci. 86: 2738-2748.
Lippke, H., T. D. Forbes, and W. C. Ellis. (2000) Effect of supplements on growth and forage intake by stocker steers
grazing wheat pasture. J. Anim. Sci. 78:1625-1635.
National Bison Association (1999) Total Bison Population—1999. Report provided during personal email
communication with Dave Carter, Executive Director, National Bison Association, July 19, 2011.
Pinchak, W.E., D. R. Tolleson, M. McCloy, L. J. Hunt, R. J. Gill, R. J. Ansley, and S. J. Bevers (2004) Morbidity effects
on productivity and profitability of stocker cattle grazing in the southern plains. J. Anim. Sci. 82:2773-2779.
Platter, W. J., J. D. Tatum, K. E. Belk, J. A. Scanga, and G. C. Smith (2003) Effects of repetitive use of hormonal
implants on beef carcass quality, tenderness, and consumer ratings of beef palatability. J. Anim. Sci. 81:984-996.
Preston, R.L (2010) What's The Feed Composition Value of That Cattle Feed? Beef Magazine, March 1, 2010.
Available at: .
Skogerboe, T. L., L. Thompson, J. M. Cunningham, A. C. Brake, V. K. Karle (2000) The effectiveness of a single dose
of doramectin pour-on in the control of gastrointestinal nematodes in yearling stocker cattle. Vet. Parasitology
87:173-181.
Soliva, C.R. (2006) Report to the attention of IPCC about the data set and calculation method used to estimate
methane formation from enteric fermentation of agricultural livestock population and manure management in
Swiss agriculture. On behalf of the Federal Office for the Environment (FOEN), Berne, Switzerland.
U.S. Department of Agriculture (USDA) (2017) Quick Stats: Agricultural Statistics Database. National Agriculture
Statistics Service, U.S. Department of Agriculture. Washington, D.C. Available online at
. Accessed June 1, 2017.
USDA (2019) Quick Stats: Agricultural Statistics Database. National Agriculture Statistics Service, U.S. Department
of Agriculture. Washington, D.C. Available online at . Accessed August 1, 2016.
USDA (2012) Census of Agriculture: 2012 Census Report. United States Department of Agriculture. Available online
at: .
10-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
USDA (2007) Census of Agriculture: 2007 Census Report. United States Department of Agriculture. Available online
at: .
USDA (2002) Census of Agriculture: 2002 Census Report. United States Department of Agriculture. Available online
at: .
USDA (1997) Census of Agriculture: 1997 Census Report. United States Department of Agriculture. Available online
at: . Accessed July 18, 2011.
USDA (1996) Beef Cow/Calf Health and Productivity Audit (CHAPA): Forage Analyses from Cow/Calf Herds in 18
States. National Agriculture Statistics Service, U.S. Department of Agriculture. Washington, D.C. Available online at
. March 1996.
USDA (1992) Census of Agriculture: 1992 Census Report. United States Department of Agriculture. Available online
at: . Accessed July 18, 2011.
USDA:APHIS:VS (2010) Beef 2007-08, Part V: Reference of Beef Cow-calf Management Practices in the United
States, 2007-08. USDA-APHIS-VS, CEAH. Fort Collins, CO.
USDA:APHIS:VS (2002) Reference of 2002 Dairy Management Practices. USDA-APHIS-VS, CEAH. Fort Collins, CO.
Available online at .
USDA:APHIS:VS (1998) Beef'97, Parts l-IV. USDA-APHIS-VS, CEAH. Fort Collins, CO. Available online at
.
USDA:APHIS:VS (1996) Reference of 1996 Dairy Management Practices. USDA-APHIS-VS, CEAH. Fort Collins, CO.
Available online at .
USDA:APHIS:VS (1994) Beef Cow/Calf Health and Productivity Audit. USDA-APHIS-VS, CEAH. Fort Collins, CO.
Available online at .
USDA:APHIS:VS (1993) Beef Cow/Calf Health and Productivity Audit. USDA-APHIS-VS, CEAH. Fort Collins, CO.
August 1993. Available online at .
Vasconcelos and Galyean (2007) Nutritional recommendations of feedlot consulting nutritionists: The 2007 Texas
Tech University Study. J. Anim. Sci. 85:2772-2781.
Manure Management
A ASAE (1998) ASAE Standards 1998, 45th Edition. American Society of Agricultural Engineers. St. Joseph, Ml.
Bryant, M.P., V.H. Varel, R.A. Frobish, and H.R. Isaacson (1976) In H.G. Schlegel (ed.)]; Seminar on Microbial Energy
Conversion. E. Goltz KG. Gottingen, Germany.
Bush, E. (1998) Personal communication with Eric Bush, Centers for Epidemiology and Animal Health, U.S.
Department of Agriculture regarding National Animal Health Monitoring System's (NAHMS) Swine '95 Study.
EPA (2019) AgSTAR Anaerobic Digester Database. Available online at: . Accessed July 2019.
EPA (2008) Climate Leaders Greenhouse Gas Inventory Protocol Offset Project Methodology for Project Type
Managing Manure with Biogas Recovery Systems. Available online at:
.
EPA (2005) National Emission Inventory—Ammonia Emissions from Animal Agricultural Operations, Revised Draft
Report. U.S. Environmental Protection Agency. Washington, D.C. April 22, 2005. Available online at:
.
Accessed August 2007.
References 10-45

-------
EPA (2002a) Development Document for the Final Revisions to the National Pollutant Discharge Elimination System
(NPDES) Regulation and the Effluent Guidelines for Concentrated Animal Feeding Operations (CAFOS). U.S.
Environmental Protection Agency. EPA-821-R-03-001. December 2002.
EPA (2002b) Cost Methodology for the Final Revisions to the National Pollutant Discharge Elimination System
Regulation and the Effluent Guidelines for Concentrated Animal Feeding Operations. U.S. Environmental
Protection Agency. EPA-821-R-03-004. December 2002.
EPA (1992) Global Methane Emissions from Livestock and Poultry Manure, Office of Air and Radiation, U.S.
Environmental Protection Agency. February 1992.
ERG (2019) "Incorporation of USDA 2016 ARMS Dairy Data into the Manure Management Greenhouse Gas
Inventory." Memorandum to USDA OCE and EPA from ERG, December 2019.
ERG (2018) "Incorporation of USDA 2009 ARMS Swine Data into the Manure Management Greenhouse Gas
Inventory." Memorandum to USDA OCE and EPA from ERG, November 2018.
ERG (2010a) 'Typical Animal Mass Values for Inventory Swine Categories." Memorandum to EPA from ERG. July 19,
2010.
ERG (2010b) Telecon with William Boyd of USDA NRCS and Cortney Itle of ERG Concerning Updated VS and Nex
Rates. August 8, 2010.
ERG (2010c) "Updating Current Inventory Manure Characteristics new USDA Agricultural Waste Management Field
Handbook Values." Memorandum to EPA from ERG. August 13, 2010.
ERG (2008) "Methodology for Improving Methane Emissions Estimates and Emission Reductions from Anaerobic
Digestion System for the 1990-2007 Greenhouse Gas Inventory for Manure Management." Memorandum to EPA
from ERG. August 18, 2008.
ERG (2003a) "Methodology for Estimating Uncertainty for Manure Management Greenhouse Gas Inventory."
Contract No. GS-10F-0036, Task Order 005. Memorandum to EPA from ERG, Lexington, MA. September 26, 2003.
ERG (2003b) "Changes to Beef Calves and Beef Cows Typical Animal Mass in the Manure Management Greenhouse
Gas Inventory." Memorandum to EPA from ERG, October 7, 2003.
ERG (2001) Summary of development of MDP Factor for methane conversion factor calculations. ERG, Lexington,
MA. September 2001.
ERG (2000a) Calculations: Percent Distribution of Manure for Waste Management Systems. ERG, Lexington, MA.
August 2000.
ERG (2000b) Discussion of Methodology for Estimating Animal Waste Characteristics (Summary of Bo Literature
Review). ERG, Lexington, MA. June 2000.
Groffman, P.M., R. Brumme, K. Butterbach-Bahl, K.E. Dobbie, A.R. Mosier, D. Ojima, H. Papen, W.J. Parton, K.A.
Smith, and C. Wagner-Riddle (2000) "Evaluating annual nitrous oxide fluxes at the ecosystem scale." Global
Biogeochemical Cycles, 14(4):1061-1070.
Hashimoto, A.G. (1984) "Methane from Swine Manure: Effect of Temperature and Influent Substrate Composition
on Kinetic Parameter (k)." Agricultural Wastes, 9:299-308.
Hashimoto, A.G., V.H. Varel, and Y.R. Chen (1981) "Ultimate Methane Yield from Beef Cattle Manure; Effect of
Temperature, Ration Constituents, Antibiotics and Manure Age." Agricultural Wastes, 3:241-256.
Hill, D.T. (1984) "Methane Productivity of the Major Animal Types." Transactions of the ASAE, 27(2):530-540.
Hill, D.T. (1982) "Design of Digestion Systems for Maximum Methane Production." Transactions of the ASAE,
25(l):226-230.
IPCC (2019) 2019 Refinement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National
Greenhouse Gas Inventories Programme, The Intergovernmental Panel on Climate Change. [CalvoBuendia, E.,
10-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize S., Osako, A., Pyrozhenko, Y., Shermanau, P. and
Federici, S. (eds)]. Switzerland.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Morris, G.R. (1976) Anaerobic Fermentation of Animal Wastes: A Kinetic and Empirical Design Fermentation. M.S.
Thesis. Cornell University.
National Bison Association (1999) Total Bison Population—1999. Report provided during personal email
communication with Dave Carter, Executive Director, National Bison Association July 19, 2011.
Ott, S.L. (2000) Dairy '96 Study. Stephen L. Ott, Animal and Plant Health Inspection Service, U.S. Department of
Agriculture. June 19, 2000.
Robertson, G. P. and P. M. Groffman (2015). Nitrogen transformations. Soil Microbiology, Ecology, and
Biochemistry, pages 421-446. Academic Press, Burlington, Massachusetts, USA.
Safley, L.M., Jr. (2000) Personal Communication. Deb Bartram, ERG and L.M. Safley, President, Agri-Waste
Technology. June and October 2000.
Sweeten, J. (2000) Personal Communication. John Sweeten, Texas A&M University and Indra Mitra, ERG. June
2000.
UEP (1999) Voluntary Survey Results—Estimated Percentage Participation/Activity. Caged Layer Environmental
Management Practices, Industry data submissions for EPA profile development, United Egg Producers and National
Chicken Council. Received from John Thorne, Capitolink. June 2000.
United Nations Framework Convention on Climate Change (UNFCCC) (2017) Definitions. Available online at:
. Accessed on December 8, 2017.
USDA (2020) Quick Stats: Agricultural Statistics Database. National Agriculture Statistics Service, U.S. Department
of Agriculture. Washington, D.C. Available online at: .
USDA (2019a) Quick Stats: Agricultural Statistics Database. National Agriculture Statistics Service, U.S. Department
of Agriculture. Washington, D.C. Available online at: .
USDA (2019b) Chicken and Eggs 2018 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2019. Available online at:
.
USDA (2019c) Poultry - Production and Value 2018 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2019. Available online at:
.
USDA (2019d) 1987,1992,1997, 2002, 2007, 2012, and 2017 Census of Agriculture. National Agriculture Statistics
Service, U.S. Department of Agriculture. Washington, D.C. Available online at:
. May 2019.
USDA (2018a) Chicken and Eggs 2017 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2018. Available online at:
.
USDA (2018b) Poultry - Production and Value 2017 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2018. Available online at:
.
USDA (2017a) Chicken and Eggs 2016 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2017. Available online at:
.
References 10-47

-------
USDA (2017b) Poultry - Production and Value 2016 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2017. Available online at:
.
USDA (2016a) Chicken and Eggs 2015 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2016. Available online at:
.
USDA (2016b) Poultry - Production and Value 2015 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2016. Available online at:
.
USDA (2015a) Chicken and Eggs 2014 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2015. Available online at:
.
USDA (2015b) Poultry - Production and Value 2014 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2015. Available online at:
.
USDA (2014a) Chicken and Eggs 2013 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2014. Available online at:
.
USDA (2014b) Poultry - Production and Value 2013 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2014. Available online at:
.
USDA (2013a) Chicken and Eggs 2012 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2013. Available online at:
.
USDA (2013b) Poultry - Production and Value 2012 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2013. Available online at:
.
USDA (2012a) Chicken and Eggs 2011 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2012. Available online at:
.
USDA (2012b) Poultry - Production and Value 2011 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2012. Available online at:
.
USDA (2011a) Chicken and Eggs 2010 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2011. Available online at:
.
USDA (2011b) Poultry - Production and Value 2010 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2011. Available online at:
.
USDA (2010a) Chicken and Eggs 2009 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2010. Available online at:
.
USDA (2010b) Poultry - Production and Value 2009 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2010. Available online at:
.
10-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
USDA (2009a) Chicken and Eggs 2008 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2009. Available online at:
.
USDA (2009b) Poultry - Production and Value 2008 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2009. Available online at:
.
USDA (2009c) Chicken and Eggs - Final Estimates 2003-2007. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. March 2009. Available online at:
.
USDA (2009d) Poultry Production and Value—Final Estimates 2003-2007. National Agriculture Statistics Service,
U.S. Department of Agriculture. Washington, D.C. May 2009. Available online at:
.
USDA (2008) Agricultural Waste Management Field Handbook, National Engineering Handbook (NEH), Part 651.
Natural Resources Conservation Service, U.S. Department of Agriculture.
USDA (2004a) Chicken and Eggs—Final Estimates 1998-2003. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2004. Available online at:
.
USDA (2004b) Poultry Production and Value—Final Estimates 1998-2002. National Agriculture Statistics Service,
U.S. Department of Agriculture. Washington, D.C. April 2004. Available online at:
.
USDA (1999) Poultry Production and Value—Final Estimates 1994-97. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. March 1999. Available online at:
.
USDA (1998) Chicken and Eggs—Final Estimates 1994-97. National Agriculture Statistics Service, U.S. Department
of Agriculture. Washington, D.C. December 1998. Available online at:
.
USDA (1996) Agricultural Waste Management Field Handbook, National Engineering Handbook (NEH), Part 651.
Natural Resources Conservation Service, U.S. Department of Agriculture. July 1996.
USDA (1995a) Poultry Production and Value—Final Estimates 1988-1993. National Agriculture Statistics Service,
U.S. Department of Agriculture. Washington, D.C. March 1995. Available online at:
.
USDA (1995b) Chicken and Eggs—Final Estimates 1988-1993. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. December 1995. Available online at:
.
USDA (1994) Sheep and Goats—Final Estimates 1989-1993. National Agriculture Statistics Service, U.S. Department
of Agriculture. Washington, D.C. January 31,1994. Available online at:
.
USDA APHIS (2003) Sheep 2001, Part I: Reference of Sheep Management in the United States, 2001 and Part IV:
Baseline Reference of 2001 Sheep Feedlot Health and Management. USDA-APHIS-VS. Fort Collins, CO. #N356.0702.
Available online at.
USDA APHIS (2000) Layers '99—Part II: References of 1999 Table Egg Layer Management in the U.S. USDA-APHIS-
VS. Fort Collins, CO. Available online at
.
References 10-49

-------
USDA APHIS (1996) Swine '95: Grower/Finisher Part II: Reference of 1995 U.S. Grower/Finisher Health &
Management Practices. USDA-APHIS-VS. Fort Collins, CO. Available online at:
.
Rice Cultivation
Baicich, P. (2013) The Birds and Rice Connection. Bird Watcher's Digest. Available online at:
.
Brockwell, P.J., and R.A. Davis (2016) Introduction to time series and forecasting. Springer.
Cantens, G. (2004 through 2005) Personal Communication. Janet Lewis, Assistant to Gaston Cantens, Vice
President of Corporate Relations, Florida Crystals Company and ICF International.
Cheng, K., S.M. Ogle, W.J. Parton, G. Pan. (2014) "Simulating greenhouse gas mitigation potentials for Chinese
croplands using the DAYCENT ecosystem model." Global Change Biology 20:948-962.
Cheng, K., S.M. Ogle, W.J. Parton and G. Pan. (2013) "Predicting methanogenesis from rice paddies using the
DAYCENT ecosystem model." Ecological Modelling 261-262:19-31.
Del Grosso, S.J., S.M. Ogle, W.J. Parton, and F.J. Breidt (2010) "Estimating Uncertainty in N20 Emissions from U.S.
Cropland Soils." Global Biogeochemical Cycles, 24, GB1009, doi:10.1029/2009GB003544.
Deren, C. (2002) Personal Communication and Dr. Chris Deren, Everglades Research and Education Centre at the
University of Florida and Caren Mintz, ICF International. August 15, 2002.
Fitzgerald, G.J., K. M. Scow & J. E. Hill (2000) "Fallow Season Straw and Rice Management Effects on Methane
Emissions in California Rice." Global biogeochemical cycles, 14 (3), 767-776.
Fleskes, J. P., Perry, W.M., Petrik, K.L., Spell, R., and Reid, F. (2005) Change in area of winter-flood and dry rice in
the northern Central Valley of California determined by satellite imagery. California Fish and Game, 91: 207-215.
Gonzalez, R. (2007 through 2014) Email correspondence. Rene Gonzalez, Plant Manager, Sem-Chi Rice Company
and ICF International.
Hardke, J.T. (2015) Trends in Arkansas rice production, 2014. B.R. Wells Arkansas Rice Research Studies 2014.
Norman, R.J. and Moldenhauer, K.A.K. (Eds.). Research Series 626, Arkansas Agricultural Experiment Station,
University of Arkansas.
Hardke, J. (2014) Personal Communication. Dr. Jarrod Hardke, Rice Extension Agronomist at the University of
Arkansas Rice Research and Extension Center and Kirsten Jaglo, ICF International. September 11, 2014.
Hardke, J. (2013) Email correspondence. Dr. Jarrod Hardke, Rice Extension Agronomist at the University of
Arkansas Rice Research and Extension Center and Cassandra Snow, ICF International. July 15, 2013.
Hardke, J.T., and Wilson, C.E. Jr., (2014) Trends in Arkansas rice production, 2013. B.R. Wells Arkansas Rice
Research Studies 2013. Norman, R.J., and Moldenhauer, K.A.K., (Eds.). Research Series 617, Arkansas Agricultural
Experiment Station, University of Arkansas.
Hardke, J.T., and Wilson, C.E. Jr., (2013) Trends in Arkansas rice production. B.R. Wells Arkansas Rice Research
Studies 2012. Norman, R.J., and Moldenhauer, K.A.K., (Eds.). Research Series 609, Arkansas Agricultural Experiment
Station, University of Arkansas.
Hollier, C. A. (ed), (1999) Louisiana rice production handbook. Louisiana State University Agricultural Center. LCES
Publication Number 2321. 116 pp.
Holzapfel-Pschorn, A., R. Conrad, and W. Seiler (1985) "Production, Oxidation, and Emissions of Methane in Rice
Paddies." FEMS Microbiology Ecology, 31:343-351.
10-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Kirstein, A. (2003 through 2004, 2006) Personal Communication. Arthur Kirstein, Coordinator, Agricultural
Economic Development Program, Palm Beach County Cooperative Extension Service, FLand ICF International.
Klosterboer, A. (1997,1999 through 2003) Personal Communication. Arlen Klosterboer, retired Extension
Agronomist, Texas A&M University and ICF International. July 7, 2003.
Lindau, C.W. and P.K. Bollich (1993) "Methane Emissions from Louisiana First and Ratoon Crop Rice." So/7 Science,
156:42-48.
Linquist, B.A., M.A. Adviento-Borbe, C.M. Pittelkow, C.v. Kessel, et al. (2012) Fertilizer management practices and
greenhouse gas emissions from rice systems: A quantitative review and analysis. Field Crops Research, 135:10-21.
Linscombe, S. (1999, 2001 through 2014) Email correspondence. Steve Linscombe, Professor with the Rice
Research Station at Louisiana State University Agriculture Center and ICF International.
LSU, (2015) Louisiana ratoon crop and conservation: Ratoon & Conservation Tillage Estimates. Louisiana State
University, College of Agriculture AgCenter. Online at: www.lsuagcenter.com.
Miller, M.R., Garr, J.D., and Coates, P.S., (2010) Changes in the status of harvested rice fields in the Sacramento
Valley, California: Implications for wintering waterfowl. Wetlands, 30: 939-947.
Neue, H.U., R. Wassmann, H.K. Kludze, W. Bujun, and R.S. Lantin (1997) "Factors and processes controlling
methane emissions from rice fields." Nutrient Cycling in Agroecosystems 49: 111-117.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams and K. Paustian. (2007) "An empirically based approach for estimating
uncertainty associated with modeling carbon sequestration in soils." Ecological Modelling 205:453-463.
Ogle, S.M., S. Spencer, M. Hartman, L. Buendia, L. Stevens, D. du Toit, J. Witi (2016) "Developing national baseline
GHG emissions and analyzing mitigation potentials for agriculture and forestry using an advanced national GHG
inventory software system." In Advances in Agricultural Systems Modeling 6, Synthesis and Modeling of
Greenhouse Gas Emissions and Carbon Storage in Agricultural and Forestry Systems to Guide Mitigation and
Adaptation, S. Del Grosso, LR. Ahuja and W.J. Parton (eds.), American Society of Agriculture, Crop Society of
America and Soil Science Society of America, pp. 129-148.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel: Description
and Testing". Glob. Planet. Chang. 19: 35-48.
Parton, W.J., D.S. Schimel, C.V. Cole, D.S. Ojima (1987) "Analysis of factors controlling soil organic matter levels in
Great Plains grasslands." Soil Science Society of America Journal 51:1173-1179.
Sass, R. L. (2001) CH4 Emissions from Rice Agriculture. Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories. 399-417. Available online at: .
Sass, R.L., F.M. Fisher, P.A. Harcombe, and F.T. Turner (1990) "Methane Production and Emissions in a Texas Rice
Field." Global Biogeochemical Cycles, 4:47-68.
Sass, R.L., F.M. Fisher, S.T. Lewis, M.F. Jund, and F.T. Turner. (1994) "Methane emissions from rice fields: effect of
soil texture." Global Biogeochemical Cycles 8:135-140.
Schueneman, T. (1997,1999 through 2001) Personal Communication. Tom Schueneman, Agricultural Extension
Agent, Palm Beach County, FLand ICF International.
Slaton, N. (1999 through 2001) Personal Communication. Nathan Slaton, Extension Agronomist—Rice, University
of Arkansas Division of Agriculture Cooperative Extension Service and ICF International.
Stansel, J. (2004 through 2005) Email correspondence. Dr. Jim Stansel, Resident Director and Professor Emeritus,
Texas A&M University Agricultural Research and Extension Center and ICF International.
References 10-51

-------
TAMU (2015) Texas Rice Crop Survey. Texas A&M AgriUFE Research Center at Beaumont. Online at:
.
Texas Agricultural Experiment Station (2007 through 2014) Texas Rice Acreage by Variety. Agricultural Research
and Extension Center, Texas Agricultural Experiment Station, Texas A&M University System. Available online at:
.
Texas Agricultural Experiment Station (2006) 2005 - Texas Rice Crop Statistics Report. Agricultural Research and
Extension Center, Texas Agricultural Experiment Station, Texas A&M University System, p. 8. Available online at:
.
University of California Cooperative Extension (UCCE) (2015) Rice Production Manual. Revised (2015) UCCE, Davis,
in collaboration with the California Rice Research Board.
USDA (2005 through 2015) Crop Production Summary. National Agricultural Statistics Service, Agricultural Statistics
Board, U.S. Department of Agriculture, Washington, D.C. Available online at: .
USDA (2012) Summary of USDA-ARS Research on the Interrelationship of Genetic and Cultural Management
Factors That Impact Grain Arsenic Accumulation in Rice. News and Events. Agricultural Research Service, U.S.
Department of Agriculture, Washington, D.C. Available online at:
. September 2013.
USDA (2003) Field Crops, Final Estimates 1997-2002. Statistical Bulletin No. 982. National Agricultural Statistics
Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online at:
. September 2005.
USDA (1998) Field Crops Final Estimates 1992-1997. Statistical Bulletin Number 947 a. National Agricultural
Statistics Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online
at: . July 2001.
USDA (1994) Field Crops Final Estimates 1987-1992. Statistical Bulletin Number 896. National Agricultural Statistics
Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online at:
. July 2001.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory. Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
van Bodegom, P.M., R. Wassmann, T.M. Metra-Corton (2001) "A process based model for methane emission
predictions from flooded rice paddies." Global Biogeochemical Cycles 15: 247-263.
Wang, J.J., S.K. Dodla, S. Viator, M. Kongchum, S. Harrison, S. D. Mudi, S. Liu, Z. Tian (2013) Agriculture Field
Management Practices and Greenhouse Gas Emissions from Louisiana Soils. Louisiana Agriculture, Spring 2013: 8-
9. Available online at: .
Wassmann, R. H.U. Neue, R.S. Lantin, K. Makarim, N. Chareonsil5, LV. Buendia, and H. Rennenberg (2000a)
Characterization of methane emissions from rice fields in Asia II. Differences among irrigated, rainfed, and
deepwater rice." Nutrient Cycling in Agroecosystems, 58(l):13-22.
Wassmann, R., R.S. Lantin, H.U. Neue, LV. Buendia, et al. (2000b) "Characterization of Methane Emissions from
Rice Fields in Asia. III. Mitigation Options and Future Research Needs." Nutrient Cycling in Agroecosystems,
58(l):23-36.
Way, M.O., McCauley, G.M., Zhou, X.G., Wilson, L.T., and Morace, B. (Eds.), (2014) 2014 Texas Rice Production
Guidelines. Texas A&M AgriLIFE Research Center at Beaumont.
Wilson, C. (2002 through 2007, 2009 through 2012) Personal Communication. Dr. Chuck Wilson, Rice Specialist at
the University of Arkansas Cooperative Extension Service and ICF International.
10-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Wilson, C.E. Jr., and Branson, J.W., (2006) Trends in Arkansas rice production. B.R. Wells Arkansas Rice Research
Studies 2005. Norman, R.J., Meullenet, J.-F., and Moldenhauer, K.A.K., (Eds.). Research Series 540, Arkansas
Agricultural Experiment Station, University of Arkansas.
Wilson, C.E. Jr., and Branson, J.W., (2005) Trends in Arkansas rice production. B.R. Wells Arkansas Rice Research
Studies 2004. Norman, R.J., Meullenet, J.-F., and Moldenhauer, K.A.K., (Eds.). Research Series 529, Arkansas
Agricultural Experiment Station, University of Arkansas.
Wilson, C.E. Jr., Runsick, S.K., and Mazzanti, R., (2010) Trends in Arkansas rice production. B.R. Wells Arkansas Rice
Research Studies 2009. Norman, R.J., and Moldenhauer, K.A.K., (Eds.). Research Series 581, Arkansas Agricultural
Experiment Station, University of Arkansas.
Wilson, C.E. Jr., Runsick, S.K., Mazzanti, R., (2009) Trends in Arkansas rice production. B.R. Wells Arkansas Rice
Research Studies (2008) Norman, R.J., Meullenet, J.-F., and Moldenhauer, K.A.K., (Eds.). Research Series 571,
Arkansas Agricultural Experiment Station, University of Arkansas.
Wilson, C.E. Jr., and Runsick, S.K., (2008) Trends in Arkansas rice production. B.R. Wells Arkansas Rice Research
Studies 2007. Norman, R.J., Meullenet, J.-F., and Moldenhauer, K.A.K., (Eds.). Research Series 560, Arkansas
Agricultural Experiment Station, University of Arkansas.
Wilson, C.E. Jr., and Runsick, S.K., (2007) Trends in Arkansas rice production. B.R. Wells Arkansas Rice Research
Studies 2006. Norman, R.J., Meullenet, J.-F., and Moldenhauer, K.A.K., (Eds.). Research Series 550, Arkansas
Agricultural Experiment Station, University of Arkansas.
Yan, X., H. Akiyana, K. Yagi, and H. Akimoto (2009) "Global estimations of the inventory and mitigation potential of
methane emissions from rice cultivation conducted using the 2006 Intergovernmental Panel on Climate Change
Guidelines." Global Biogeochemical Cycles, 23, DOI: 0.1029/2008GB003299.
Young, M. (2013) Rice and Ducks. Ducks Unlimited, Memphis, TN. Available online at:
.
Agricultural Soil Management
AAPFCO (2008 through 2017) Commercial Fertilizers: 2008-2015. Association of American Plant Food Control
Officials. University of Missouri. Columbia, MO.
AAPFCO (1995 through 2000a, 2002 through 2007) Commercial Fertilizers: 1995-2007. Association of American
Plant Food Control Officials. University of Kentucky. Lexington, KY.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Cibrowski, P. (1996) Personal Communication. Peter Cibrowski, Minnesota Pollution Control Agency and Heike
Mainhardt, ICF Incorporated. July 29,1996.
Cheng, B., and D.M. Titterington (1994) "Neural networks: A review from a statistical perspective." Statistical
science 9: 2-30.
Claassen, R., M. Bowman, J. McFadden, D. Smith, and S. Wallander (2018) Tillage intensity and conservation
cropping in the United States, EIB 197. United States Department of Agriculture, Economic Research Service,
Washington, D.C.
CTIC (2004) 2004 Crop Residue Management Survey. Conservation Technology Information Center. Available at
.
Del Grosso, S.J., A.R. Mosier, W.J. Parton, and D.S. Ojima (2005) "DAYCENT Model Analysis of Past and
Contemporary Soil N20 and Net Greenhouse Gas Flux for Major Crops in the USA." Soil Tillage and Research, 83: 9-
24. doi: 10.1016/j.still.2005.02.007.
Del Grosso, S.J., S.M. Ogle, W.J. Parton, and F.J. Breidt (2010) "Estimating Uncertainty in N20 Emissions from U.S.
Cropland Soils." Global Biogeochemical Cycles, 24, GB1009, doi:10.1029/2009GB003544.
References 10-53

-------
Del Grosso, S.J., W.J. Parton, C.A. Keough, and M. Reyes-Fox. (2011) Special features of the DAYCENT modeling
package and additional procedures for parameterization, calibration, validation, and applications, in Methods of
Introducing System Models into Agricultural Research, L.R. Ahuja and Liwang Ma, editors, p. 155-176, American
Society of Agronomy, Crop Science Society of America, Soil Science Society of America, Madison, Wl. USA.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001) "Simulated
Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In Schaffer, M., L. Ma,
S. Hansen, (eds.). Modeling Carbon and Nitrogen Dynamics for Soil Management. CRC Press. Boca Raton, Florida.
303-332.
Del Grosso, S.J., T. Wirth, S.M. Ogle, W.J. Parton (2008) Estimating agricultural nitrous oxide emissions. EOS 89,
529-530.
Delgado, J.A., S.J. Del Grosso, and S.M. Ogle (2009) "15N isotopic crop residue cycling studies and modeling suggest
that IPCC methodologies to assess residue contributions to N20-N emissions should be reevaluated." Nutrient
Cycling in Agroecosystems, DOI 10.1007/sl0705-009-9300-9.
Edmonds, L, N. Gollehon, R.L. Kellogg, B. Kintzer, L Knight, C. Lander, J. Lemunyon, D. Meyer, D.C. Moffitt, and J.
Schaeffer (2003) "Costs Associated with Development and Implementation of Comprehensive Nutrient
Management Plans." Part 1. Nutrient Management, Land Treatment, Manure and Wastewater Handling and
Storage, and Recordkeeping. Natural Resource Conservation Service, U.S. Department of Agriculture.
EPA (2003) Clean Watersheds Needs Survey 2000—Report to Congress, U.S. Environmental Protection Agency.
Washington, D.C. Available online at: .
EPA (1999) Biosolids Generation, Use and Disposal in the United States. Office of Solid Waste, U.S. Environmental
Protection Agency. Available online at: .
EPA (1993) Federal Register. Part II. Standards for the Use and Disposal of Sewage Sludge; Final Rules. U.S.
Environmental Protection Agency, 40 CFR Parts 257, 403, and 503.
Firestone, M. K., and E.A. Davidson, Ed. (1989) Microbiological basis of NO and N20 production and consumption in
soil. Exchange of trace gases between terrestrial ecosystems and the atmosphere. New York, John Wiley & Sons.
ILENR (1993) Illinois Inventory of Greenhouse Gas Emissions and Sinks: 1990. Office of Research and Planning,
Illinois Department of Energy and Natural Resources. Springfield, IL.
IPCC (2013) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands. The
Intergovernmental Panel on Climate Change. [T, Hiraishi, T. Krug, K. Tanabe, N. Srivastava, B. Jamsranjav, M.
Fukuda and T. Troxler (eds.)]. Hayama, Kanagawa, Japan.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Little, R. (1988) "Missing-data adjustments in large surveys." Journal of Business and Economic Statistics 6: 287-
296.
McFarland, M.J. (2001) Biosolids Engineering, New York: McGraw-Hill, p. 2.12.
McGill, W.B., and C.V. Cole (1981) Comparative aspects of cycling of organic C, N, S and P through soil organic
matter. Geoderma 26:267-286.
Metherell, A.K., LA. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURYSoil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
NEBRA (2007) A National Biosolids Regulation, Quality, End Use & Disposal Survey. North East Biosolids and
Residuals Association, July 21, 2007.
10-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Noller, J. (1996) Personal Communication. John Noller, Missouri Department of Natural Resources and Heike
Mainhardt, ICF Incorporated. July 30,1996.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams and K. Paustian (2007) "Empirically-Based Uncertainty Associated with
Modeling Carbon Sequestration Rates in Soils." Ecological Modeling 205:453-463.
Oregon Department of Energy (1995) Report on Reducing Oregon's Greenhouse Gas Emissions: Appendix D
Inventory and Technical Discussion. Oregon Department of Energy. Salem, OR.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel: Description
and Testing." Glob. Planet. Chang. 19: 35-48.
Potter, C., S. Klooster, A. Huete, and V. Genovese (2007) Terrestrial carbon sinks for the United States predicted
from MODIS satellite data and ecosystem modeling. Earth Interactions 11, Article No. 13, DOI 10.1175/EI228.1.
Potter, C. S., J.T. Randerson, C.B. Fields, P.A. Matson, P.M. Vitousek, H.A. Mooney, and S.A. Klooster (1993)
'Terrestrial ecosystem production: a process model based on global satellite and surface data." Global
Biogeochemical Cycles 7:811-841.
PRISM Climate Group (2018) PRISM Climate Data, Oregon State University, ,
downloaded 18 July 2018.
Pukelsheim, F. (1994) 'The 3-Sigma-Rule." American Statistician 48:88-91.
Ruddy B.C., D.L. Lorenz, and D.K. Mueller (2006) County-level estimates of nutrient inputs to the land surface of
the conterminous United States, 1982-2001. Scientific Investigations Report 2006-5012. U.S Department of the
Interior.
Scheer, C., S.J. Del Grosso, W.J. Parton, D.W. Rowlings, P.R. Grace (2013) Modeling Nitrous Oxide Emissions from
Irrigated Agriculture: Testing DAYCENT with High Frequency Measurements, Ecological Applications, in press.
Available online at: .
Soil Survey Staff (2019) Gridded Soil Survey Geographic (gSSURGO) Database for the Conterminous United States.
United States Department of Agriculture, Natural Resources Conservation Service. Available online at
https://gdg.sc.egov.usda.gov/. April, 2019 (FY2019 official release).
Towery, D. (2001) Personal Communication. Dan Towery regarding adjustments to the CTIC (1998) tillage data to
reflect long-term trends, Conservation Technology Information Center, West Lafayette, IN, and Marlen Eve,
National Resource Ecology Laboratory, Fort Collins, CO. February 2001.
TVA (1991 through 1992a, 1993 through 1994) Commercial Fertilizers. Tennessee Valley Authority, Muscle Shoals,
AL.
USDA-ERS (2018) Agricultural Resource Management Survey (ARMS) Farm Financial and Crop Production Practices:
Tailored Reports. Available online at: .
USDA-ERS (1997) Cropping Practices Survey Data—1995. Economic Research Service, United States Department of
Agriculture. Available online at: .
USDA-NASS (2019) Quick Stats. National Agricultural Statistics Service, United States Department of Agriculture,
Washington, D.C. .
USDA-NASS (2017) 2017 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at .
USDA-NASS (2012) 2012 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at .
USDA-NASS (2004) Agricultural Chemical Usage: 2003 Field Crops Summary. Report AgChl(04)a. National
Agricultural Statistics Service, U.S. Department of Agriculture, Washington, D.C. Available online at:
.
References 10-55

-------
USDA-NASS (1999) Agricultural Chemical Usage: 1998 Field Crops Summary. Report AgCHl(99). National
Agricultural Statistics Service, U.S. Department of Agriculture, Washington, DC. Available online at:
.
USDA-NASS (1992) Agricultural Chemical Usage: 1991 Field Crops Summary. Report AgChl(92). National
Agricultural Statistics Service, U.S. Department of Agriculture, Washington, D.C. Available online at:
.
USDA-NRCS (2012) Assessment of the Effects of Conservation Practices on Cultivated Cropland in the Upper
Mississippi River Basin. U.S. Department of Agriculture, Natural Resources Conservation Service,
.
Tepordei, V.V. (2003b) Personal communication. Valentin Tepordei, U.S. Geological Survey and ICF Consulting,
August 18, 2003.
Tepordei, V.V. (1996) "Crushed Stone," In Minerals Yearbook 1994. U.S. Department of the Interior/Bureau of
Mines, Washington, D.C. Available online at:
. Accessed August 2000.
Tepordei, V.V. (1995) "Crushed Stone," In Minerals Yearbook 1993. U.S. Department of the Interior/Bureau of
Mines, Washington, D.C. pp. 1107-1147.
Tepordei, V. V. (1994) "Crushed Stone," In Minerals Yearbook 1992. U.S. Department of the Interior/Bureau of
Mines, Washington, D.C. pp. 1279-1303.
10-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
USGS (2020) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2020, U.S.
Geological Survey, Reston, VA. Available online at:
.
West, T.O., and A.C. McBride (2005) 'The contribution of agricultural lime to carbon dioxide emissions in the
United States: dissolution, transport, and net emissions," Agricultural Ecosystems & Environment 108:145-154.
West, T.O. (2008) Email correspondence. Tristram West, Environmental Sciences Division, Oak Ridge National
Laboratory, U.S. Department of Energy and Nikhil Nadkarni, ICF International on suitability of liming emission
factor for the entire United States. June 9, 2008.
Willett, J.C. (2020b) Personal communication. Jason Willett. Preliminary data tables from "Crushed Stone," In 2018
Minerals Yearbook. U.S. Department of the Interior/U.S. Geological Survey. Washington, D.C. December 01, 2020.
Willett, J.C. (2019) Personal communication. Jason Willett. Preliminary data tables from "Crushed Stone," In 2017
Minerals Yearbook. U.S. Department of the Interior/U.S. Geological Survey. Washington, D.C. September 10, 2019.
Willett, J.C. (2020a) "Crushed Stone," In Minerals Yearbook 2016. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed November 2020.
Willett, J.C. (2017) "Crushed Stone," In Minerals Yearbook 2015. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed November 2017.
Willett, J.C. (2016) "Crushed Stone," In Minerals Yearbook 2014. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed September 2016.
Willett, J.C. (2015) "Crushed Stone," In Minerals Yearbook 2013. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed September 2015.
Willett, J.C. (2014) "Crushed Stone," In Minerals Yearbook 2012. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed September 2014.
Willett, J.C. (2013a) "Crushed Stone," In Minerals Yearbook 2011. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed May 2013.
Willett, J.C. (2013b) Personal Communication. Jason Willet, U.S. Geological Survey and ICF International.
September 9, 2013.
Willett, J.C. (2011a) "Crushed Stone," In Minerals Yearbook 2009. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed August 2011.
Willett, J.C. (2011b) "Crushed Stone," In Minerals Yearbook 2010. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed September 2012.
Willett, J.C. (2010) "Crushed Stone," In Minerals Yearbook 2008. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed August 2010.
Willett, J.C. (2009) "Crushed Stone," In Minerals Yearbook 2007. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed August 2009.
References 10-57

-------
Willett, J.C. (2007a) "Crushed Stone," In Minerals Yearbook 2005. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed August 2007.
Willett, J.C. (2007b) "Crushed Stone," In Minerals Yearbook 2006. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed August 2008.
Urea Fertilization
AAPFCO (2008 through 2018) Commercial Fertilizers. Association of American Plant Food Control Officials.
University of Missouri. Columbia, MO.
AAPFCO (1995 through 2000a, 2002 through 2007) Commercial Fertilizers. Association of American Plant Food
Control Officials. University of Kentucky. Lexington, KY.
AAPFCO (2000b) 1999-2000 Commercial Fertilizers Data, ASCII files. Available from David Terry, Secretary, AAPFCO.
EPA (2000) Preliminary Data Summary: Airport Deicing Operations (Revised). EPA-821-R-00-016. August 2000.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Itle, C. (2009) Email correspondence. Cortney Itle, ERG and Tom Wirth, U.S. Environmental Protection Agency on
the amount of urea used in aircraft deicing. January 7, 2009.
TVA (1991 through 1994) Commercial Fertilizers. Tennessee Valley Authority, Muscle Shoals, AL.
TVA (1992b) Fertilizer Summary Data 1992. Tennessee Valley Authority, Muscle Shoals, AL
Field Burning of Agricultural Residues
Akintoye, H.A., Agbeyi, E.O., and Olaniyan, A.B. (2005) "The effects of live mulches on tomato (Lycopersicon
esculentum) yield under tropical conditions." Journal of Sustainable Agriculture 26: 27-37.
Bange, M.P., Milroy, S.P., and Thongbai, P. (2004) "Growth and yield of cotton in response to waterlogging." Field
Crops Research 88:129-142.
Beyaert, R.P. (1996) The effect of cropping and tillage management on the dynamics of soil organic matter. PhD
Thesis. University of Guelph.
Bouquet, D.J., and Breitenbeck, G.A. (2000) "Nitrogen rate effect on partitioning of nitrogen and dry matter by
cotton." Crop Science 40: 1685-1693.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.Cantens, G.
(2004 through 2005) Personal Communication. Janet Lewis, Assistant to Gaston Cantens, Vice President of
Corporate Relations, Florida Crystals Company and ICF International.
Brouder, S.M., and Cassman, K.G (1990) "Root development of two cotton cultivars in relation to potassium uptake
and plant growth in a vermiculitic soil." Field Crops Res. 23:187-203.
Costa, L.D., and Gianquinto, G. (2002) "Water stress and watertable depth influence yield, water use efficiency,
and nitrogen recovery in bell pepper: lysimeter studies." Aust. J. Agric. Res. 53: 201-210.
Crafts-Brandner, S.J., Collins, M., Sutton, T.G., and Burton, H.R. (1994) "Effect of leaf maleic hydrazide
concentration on yield and dry matter partitioning in burley tobacco (Nicotiana tabacum L.)." Field Crops Research
37: 121-128.
10-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
De Pinheiro Henriques, A.R., and Marcelis, LF.M. (2000) "Regulation of growth at steady-state nitrogen nutrition in
lettuce (Lactuca sativa L): Interactive effects of nitrogen and irradiance." Annals of Botany 86: 1073-1080.0.
Diaz-Perez, J.C., Silvoy, J., Phatak,S.C., Ruberson, J., and Morse, R. (2008) Effect of winter cover crops and co-till on
the yield of organically-grown bell pepper (Capsicum annum L.). Acta Hort. 767:243-247.
Dua, K.L, and Sharma, V.K. (1976) "Dry matter production and energy contents often varieties of sugarcane at
Muzaffarnagar (Western Uttar Pradesh)." Tropical Ecology 17: 45-49.
Fritschi, F.B., Roberts, B.A., Travis, R.L., Rains, D.W., and Hutmacher, R.B. (2003) "Seasonal nitrogen concentration,
uptake, and partitioning pattern of irrigated Acala and Pima cotton as influenced by nitrogen fertility level." Crop
Science 44:516-527.
Gerik, T.J., K.L Faver, P.M. Thaxton, and K.M. El-Zik. (1996) "Late season water stress in cotton: I. Plant growth,
water use, and yield." Crop Science 36: 914-921.
Gibberd, M.R., McKay, A.G., Calder, T.C., and Turner, N.C. (2003) "Limitations to carrot (Daucus carota L.)
productivity when grown with reduced rates of frequent irrigation on a free-draining, sandy soil." Australian
Journal of Agricultural Research 54: 499-506.
Giglio, L., I. Csiszar, and C.O. Justice (2006) "Global distribution and seasonality of active fires as observed with the
Terra and Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) sensors" J. Geophys. Res. Ill, G02016,
doi:10.1029/2005JG000142.
Halevy, J. (1976) "Growth rate and nutrient uptake of two cotton cultivars grown under irrigation." Agronomy
Journal 68: 701-705.
Halvorson, A.D., Follett, R.F., Bartolo, M.E., and Schweissing, F.C. (2002) "Nitrogen fertilizer use efficiency of
furrow-irrigated onion and corn." Agronomy Journal 94: 442-449.
Heitholt, J.J., Pettigrew, W.T., and Meredith, W.R. (1992) "Light interception and lint yield of narrow-row cotton."
Crop Science 32: 728-733.
Hollifield, C.D., Silvertooth, J.C., and Moser, H. (2000) "Comparison of obsolete and modern cotton cultivars for
irrigated production in Arizona." 2000 Arizona Cotton Report, University of Arizona College of Agriculture,
.
Hopkinson, J.M. (1967) "Effects of night temperature on the growth of Nicotiana tabacum." Australian Journal of
Experimental Agriculture and Animal Husbandry 7: 78-82.
Huett, D.O., and Dettman, E.B. (1991) Effect of nitrogen on growth, quality and nutrient uptake of cabbages grown
in sand culture. Australian Journal of Experimental Agriculture 29: 875-81.
Huett, D.O., and Dettman, B. (1989) "Nitrogen response surface models of zucchini squash, head lettuce and
potato." Plant and Soil 134: 243-254.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC/UNEP/OECD/IEA (1997) Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories.
Intergovernmental Panel on Climate Change, United Nations Environment Programme, Organization for Economic
Co-Operation and Development, International Energy Agency, Paris, France.
Jacobs, J.L, Ward, G.N., and Kearney, G. (2004) "Effects of irrigation strategies and nitrogen fertilizer on turnip dry
matter yield, water use efficiency, nutritive characteristics and mineral content in western Victoria." Australian
Journal of Experimental Agriculture 44: 13-26.
Jacobs, J.L, Ward, G.N., McDowell, A.M., and Kearney, G. (2002) "Effect of seedbed cultivation techniques, variety,
soil type and sowing time, on brassica dry matter yields, water use efficiency and crop nutritive characteristics in
western Victoria." Australian Journal of Experimental Agriculture 42: 945-952.
References 10-59

-------
Jacobs, J.L, Ward, G.N., McDowell, A.M., and Kearney, G.A. (2001) "A survey on the effect of establishment
techniques, crop management, moisture availability and soil type on turnip dry matter yields and nutritive
characteristics in western Victoria." Australian Journal of Experimental Agriculture 41: 743-751.
Kage, H., Alt, C., and Stutzel, H. (2003) "Aspects of nitrogen use efficiency of cauliflower II. Productivity and
nitrogen partitioning as influenced by N supply." Journal of Agricultural Science 141: 17-29.
Kumar, A., Singh, D.P., and Singh, P. (1994) "Influence of water stress on photosynthesis, transpiration, water-use
efficiency and yield of Brassica juncea L." Field Crops Research 37: 95-101.
LANDFIRE (2008) Existing Vegetation Type Layer, LANDFIRE 1.1.0, U.S. Department of the Interior, Geological
Survey. Accessed 28 October 2010 at .
MacLeod, L.B., Gupta, U.C., and Cutcliffe, J.A. {1971) "Effect of N, P, and K on root yield and nutrient levels in the
leaves and roots of rutabagas grown in a greenhouse." Plant and Soil 35: 281-288.
Mahrani, A., and Aharonov, B. (1964) "Rate of nitrogen absorption and dry matter production by upland cotton
grown under irrigation." Israel J. Agric. Res. 14: 3-9.
Marcussi, F.F.N., Boas, R.L.V., de Godoy, L.J.G., and Goto, R. (2004) "Macronutrient accumulation and partitioning
in fertigated sweet pepper plants." Sci. Agric. (Piracicaba, Braz.) 61: 62-68.
McCarty, J.L. (2011) "Remote Sensing-Based Estimates of Annual and Seasonal Emissions from Crop Residue
Burning in the Contiguous United States." Journal of the Air & Waste Management Association, 61:1, 22-34, DOI:
10.3155/1047-3289.61.1.22.
McCarty, J.L. (2010) Agricultural Residue Burning in the Contiguous United States by Crop Type and State.
Geographic Information Systems (GIS) Data provided to the EPA Climate Change Division by George Pouliot,
Atmospheric Modeling and Analysis Division, EPA. Dr. McCarty's research was supported by the NRI Air Quality
Program of the Cooperative State Research, Education, and Extension Service, USDA, under Agreement No.
20063511216669 and the NASA Earth System Science Fellowship.
McCarty, J.L. (2009) Seasonal and Interannual Variability of Emissions from Crop Residue Burning in the Contiguous
United States. Dissertation. University of Maryland, College Park.
McPharlin, I.R., Aylmore, P.M., and Jeffery, R.C. (1992) "Response of carrots (Daucus carota L.) to applied
phosphorus and phosphorus leaching on a Karrakatta sand, under two irrigation regimes." Australian Journal of
Experimental Agriculture 32:225-232.
Mondino, M.H., Peterlin, O.A., and Garay, F. (2004) "Response of late-planted cotton to the application of growth
regulator (chlorocholine chloride, CYCOCEL75)." Expl Agric. 40: 381-387.
Moustakas, N.K., and Ntzanis, H. (2005) "Dry matter accumulation and nutrient uptake in flue-cured tobacco
(Nicotiana tabacum L.)." Field Crops Research 94:1-13.
Peach, L, Benjamin, L.R., and Mead, A. (2000) "Effects on the growth of carrots (Daucus carota L), cabbage
(Brassica oleracea var. capitata L.) and onion (Allium cepa L.) of restricting the ability of the plants to intercept
resources." Journal of Experimental Botany 51: 605-615.
Pettigrew, W.T., and Meredith, W.R., Jr. (1997) "Dry matter production, nutrient uptake, and growth of cotton as
affected by potassium fertilization." J. Plant Nutr. 20:531-548.
Pettigrew, W.T., Meredith, W.R., Jr., and Young, L.D. (2005) "Potassium fertilization effects on cotton lint yield,
yield components, and reniform nematode populations." Agronomy Journal 97: 1245-1251.
PRISM Climate Group (2015) PRISM Climate Data. Oregon State University. July 24, 2015. Available online at:
.
Reid, J.B., and English, J.M. (2000) "Potential yield in carrots (Daucus carota L.): Theory, test, and an application."
Annals of Botany 85: 593-605.
10-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Sadras, V.O., and Wilson, L.J. (1997) "Growth analysis of cotton crops infested with spider mites: II. Partitioning of
dry matter." Crop Science 37: 492-497.
Scholberg, J., McNeal, B.L, Jones, J.W., Boote, K.J., Stanley, C.D., and Obreza, T.A. (2000a) "Growth and canopy
characteristics of field-grown tomato." Agronomy Journal 92: 152-159.
Scholberg, J., McNeal, B.L, Boote, K.J., Jones, J.W., Locasio, S.J., and Olson, S.M. (2000b) "Nitrogen stress effects on
growth and nitrogen accumulation by field-grown tomato." Agronomy Journal 92:159-167.
Singels, A. and Bezuidenhout, C.N. (2002) "A new method of simulating dry matter partitioning in the Canegro
sugarcane model." Field Crops Research 78: 151 - 164.
Sitompul, S.M., Hairiah, K., Cadisch, G., and Van Noordwuk, M. (2000) "Dynamics of density fractions of macro-
organic matter after forest conversion to sugarcane and woodlots, accounted for in a modified Century model."
Netherlands Journal of Agricultural Science 48: 61-73.
Stirling, G.R., Blair, B.L, Whittle, P.J.L, and Garside, A.L (1999) "Lesion nematode (Pratylenchus zeae) is a
component of the yield decline complex of sugarcane." In: Magarey, R.C. (Ed.), Proceedings of the First
Australasian Soilborne Disease Symposium. Bureau of Sugar Experiment Stations, Brisbane, pp. 15-16.
Tan, D.K.Y., Wearing, A.H., Rickert, K.G., and Birch, C.J. (1999) "Broccoli yield and quality can be determined by
cultivar and temperature but not photoperiod in south-east Queensland." Australian Journal of Experimental
Agriculture 39: 901-909.
Tadesse, T., Nichols, M.A., and Fisher, K.J., 1999. Nutrient conductivity effects on sweet pepper plants grown using
a nutrient film technique. 1. Yield and fruit quality. New Zealand Journal of Crop and Horticultural Science, 27: 229-
237.
Torbert, H.A., and Reeves, D.W. (1994) "Fertilizer nitrogen requirements for cotton production as affected by
tillage and traffic." Soil Sci. Soc. Am. J. 58:1416-1423.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory, Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
USDA (2019) Quick Stats: U.S. & All States Data; Crops; Production and Area Harvested; 1990 - 2018. National
Agricultural Statistics Service, U.S. Department of Agriculture. Washington, D.C. U.S. Department of Agriculture,
National Agricultural Statistics Service. Washington, D.C., Available online at: .
Valantin, M., Gary, C., Vaissiere, B.E., and Frossard, J.S. (1999) "Effect of fruit load on partitioning of dry matter and
energy in cantaloupe (Cucumis melo L.)." Annals of Botany 84:173-181.
Wallach, D., Marani, A., and Kletter, E. (1978) 'The relation of cotton crop growth and development to final yield."
Field Crops Research 1: 283-294.
Wells, R., and Meredith, W.R., Jr. (1984) "Comparative growth of obsolete and modern cultivars. I. Vegetative dry
matter partitioning." Crop Science 24: 858-872.4.
Wiedenfels, R.P. (2000) "Effects of irrigation and N fertilizer application on sugarcane yield and quality." Field Crops
Research 43: 101-108.
Land Use, Land-Use Change, and Forestry
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
References 10-61

-------
UNFCCC (2014) Report of the Conference of the Parties on its nineteenth session, held in Warsaw from 11 to 23
November 2013. United Nations Framework Convention on Climate Change, Warsaw. (FCCC/CP/2013/10/Add.3).
January 31, 2014. Available online at: .
Representation of the U.S. Land Base
Alaska Department of Natural Resources (2006) Alaska Infrastructure 1:63,360. Available online at:
.
Alaska Interagency Fire Management Council (1998) Alaska Interagency Wildland Fire Management Plan. Available
online at: .
Alaska Oil and Gas Conservation Commission (2009) Oil and Gas Information System. Available online at:
.
EIA (2011) Coal Production and Preparation Report Shapefile. Available online at:
.
ESRI (2008) ESRI Data & Maps. Redlands, CA: Environmental Systems Research Institute. [CD-ROM],
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and J. Wickham. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Homer, C., J. Dewitz, J. Fry, M. Coan, N. Hossain, C. Larson, N. Herold, A. McKerrow, J.N. VanDriel and J. Wickham.
(2007) Completion of the 2001 National Land Cover Database for the Conterminous United States,
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v.
81, no. 5, p. 345-354.
IPCC (2014) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.). Published: IPCC,
Switzerland.
IPCC (2010) Revisiting the use of managed land as a proxy for estimating national anthropogenic emissions and
removals. [Eggleston HS, Srivastava N, Tanabe K, Baasansuren J, (eds.)]. Institute for Global Environmental Studies,
Intergovernmental Panel on Climate Change, Hayama, Kanagawa, Japan.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Jin, S., L. Yang, P. Danielson, C. Homer, J. Fry, and G. Xian. (2013) A comprehensive change detection method for
updating the National Land Cover Database to circa 2011. Remote Sensing of Environment, 132:159-175.
NOAA Coastal Change Analysis Program (C-CAP) Regional Land Cover Database. Data collected 1995-present
Charleston, SC: National Oceanic and Atmospheric Administration (NOAA) Coastal Services Center. Data accessed
at: .
Nusser, S.M. and J.J. Goebel (1997) 'The national resources inventory: a long-term multi-resource monitoring
programme." Environmental and Ecological Statistics 4:181-204.
Ogle, S.M., G. Domke, W.A. Zurz, M.T. Rocha, T. Huffman, A. Swan, J.E. Smith, C. Woodall, T. Krug (2018)
Delineating managed land for reporting greenhouse gas emissions and removals to the United Nations Framework
Convention on Climate Change. Carbon Balance and Management 13:9.
Smith, W.B., P.D. Miles, C.H. Perry, and S.A. Pugh (2009) Forest Resources of the United States, 2007. Gen. Tech.
Rep. WO-78. U.S. Department of Agriculture Forest Service, Washington, D.C.
10-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
U.S. Census Bureau (2010) Topologically Integrated Geographic Encoding and Referencing (TIGER) system
shapefiles. U.S. Census Bureau, Washington, D.C. Available online at: .
U.S. Department of Agriculture (2015) County Data - Livestock, 1990-2014. U.S. Department of Agriculture,
National Agriculture Statistics Service, Washington, D.C.
U.S. Department of Agriculture, Forest Service. Timber Product Output (TPO) Reports. Knoxville, TN: U.S.
Department of Agriculture Forest Service, Southern Research Station. 2012. . Accessed November 2017.
U.S. Department of Interior (2005) Federal Lands of the United States. National Atlas of the United States, U.S.
Department of the Interior, Washington D.C. Available online at:
.
United States Geological Survey (USGS), Gap Analysis Program (2012) Protected Areas Database of the United
States (PADUS), version 1.3 Combined Feature Class. November 2012.
USGS (2012) Alaska Resource Data File. Available online at: .
USGS (2005) Active Mines and Mineral Processing Plants in the United States in 2003. U.S. Geological Survey,
Reston, VA.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L, Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies. ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Forest Land Remaining Forest Land: Changes in Forest Carbon
Stocks
AF&PA (2006a and earlier) Statistical roundup. (Monthly). Washington, D.C. American Forest & Paper Association.
AF&PA (2006b and earlier) Statistics of paper, paperboard and wood pulp. Washington, D.C. American Forest &
Paper Association.
Amichev, B.Y. and J.M. Galbraith (2004) "A Revised Methodology for Estimation of Forest Soil Carbon from Spatial
Soils and Forest Inventory Data Sets." Environmental Management 33(Suppl. 1):S74-S86.
Bechtold, W.A.; Patterson, P.L (2005) The enhanced forest inventory and analysis program—national sampling
design and estimation procedures. Gen. Tech. Rep. SRS-80. Asheville, NC: U.S. Department of Agriculture Forest
Service, Southern Research Station. 85 p.
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In R.N.
Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Coulston, J.W., Wear, D.N., and Vose, J.M. (2015) Complex forest dynamics indicate potential for slowing carbon
accumulation in the southeastern United States. Scientific Reports. 5: 8002.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in standing
dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon Balance and
Management. 6:14.
Domke, G.M., Woodall, C.W., Smith, J.E., Westfall, J.A., McRoberts, R.E. (2012) Consequences of alternative tree-
level biomass estimation procedures on U.S. forest carbon stock estimates. Forest Ecology and Management. 270:
108-116.
Domke, G.M., Woodall, C.W., Walters, B.F., Smith, J.E. (2013) From models to measurements: comparing down
dead wood carbon stock estimates in the U.S. forest inventory. PLoS ONE 8(3): e59949.
References 10-63

-------
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter
carbon stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., Nave, L., Swanston, C. (2017) Toward inventory-based
estimates of soil organic carbon in forests of the United States. Ecological Applications. 27(4), 1223-1235.
EPA (2006) Municipal solid waste in the United States: 2005 Facts and figures. Office of Solid Waste, U.S.
Environmental Protection Agency. Washington, D.C. (5306P) EPA530-R-06-011. Available online at:
.
Frayer, W.E., and G.M. Furnival (1999) "Forest Survey Sampling Designs: A History." Journal of Forestry 97(12): 4-
10.
Freed, R. (2004) Open-dump and Landfill timeline spreadsheet (unpublished). ICF International. Washington, D.C.
Hair, D. (1958) "Historical forestry statistics of the United States." Statistical Bull. 228. U.S. Department of
Agriculture Forest Service, Washington, D.C.
Hair. D. and A.H. Ulrich (1963) The Demand and price situation for forest products - 1963. U.S. Department of
Agriculture Forest Service, Misc Publication No. 953. Washington, D.C.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
Howard, J. L. and Liang, S. (2019). U.S. timber production, trade, consumption, and price statistics 1965 to 2017.
Res. Pap. FPL-RP-701. Madison, Wl: USDA, Forest Service, Forest Products Laboratory.
Howard, J. L. and Jones, K.C. (2016) U.S. timber production, trade, consumption, and price statistics 1965 to 2013.
Res. Pap. FPL-RP-679. Madison, Wl: USDA, Forest Service, Forest Products Laboratory.
Howard, J. L. (2007) U.S. timber production, trade, consumption, and price statistics 1965 to 2005. Res. Pap. FPL-
RP-637. Madison, Wl: USDA, Forest Service, Forest Products Laboratory.
Howard, J. L. (2003) U.S. timber production, trade, consumption, and price statistics 1965 to 2002. Res. Pap. FPL-
RP-615. Madison, Wl: USDA, Forest Service, Forest Products Laboratory. Available online at:
.
IPCC (2014) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
[Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M., and Troxler, T.G. (eds.)]. Switzerland.
IPCC (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 [Solomon, S., D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom
and New York, NY, USA, 996 pp.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Jenkins, J.C., D.C. Chojnacky, L.S. Heath, and R.A. Birdsey (2003) "National-scale biomass estimators for United
States tree species." Forest Science 49(l):12-35.
Jandl, R., Rodeghiero, M., Martinez, C., Cotrufo, M. F., Bampa, F., van Wesemael, B., Harrison, R.B., Guerrini, I.A.,
deB Richter Jr., D., Rustad, L, Lorenz, K., Chabbi, A., Miglietta, F. (2014) Current status, uncertainty and future
needs in soil organic carbon monitoring. Science of the Total Environment, 468, 376-383.
Johnson, K. Domke, G.M., Russell, M.B., Walters, B.F., Horn, J., Peduzzi, A., Birdsey, R., Dolan, K., Huang, W. (2017).
Estimating aboveground live understory vegetation carbon in the United States. Environmental Research Letters.
10-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Nelson, M.D., Riitters, K.H., Coulston, J.W., Domke, G.M., Greenfield, E.J., Langner, L.L, Nowak, D.J., O'Dea, C.B.,
Oswalt, S.N., Reeves, M.C. and Wear, D.N., 2020. Defining the United States land base: a technical document
supporting the USDA Forest Service 2020 RPA assessment. Gen. Tech. Rep. NRS-191., 191, pp.1-70.
Ogle, S.M., Woodall, C.W., Swan, A., Smith, J.E., Wirth. T. In preparation. Determining the Managed Land Base for
Delineating Carbon Sources and Sinks in the United States. Environmental Science and Policy.
O'Neill, K.P., Amacher, M.C., Perry, C.H. (2005) Soils as an indicator of forest health: a guide to the collection,
analysis, and interpretation of soil indicator data in the Forest Inventory and Analysis program. Gen. Tech. Rep. NC-
258. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Research Station. 53 p.
Oswalt, S.N., Smith, W.B., Miles, P.D. and Pugh, S.A., 2019. Forest resources of the United States, 2017: Atechnical
document supporting the Forest Service 2020 RPA Assessment. Gen. Tech. Rep. WO-97. Washington, DC: U.S.
Department of Agriculture, Forest Service, Washington Office., 97.
Perry, C.H., C.W. Woodall, and M. Schoeneberger (2005) Inventorying trees in agricultural landscapes: towards an
accounting of "working trees". In: "Moving Agroforestry into the Mainstream." Proc. 9th N. Am. Agroforestry
Conf., Brooks, K.N. and Folliott, P.F. (eds.). 12-15 June 2005, Rochester, MN [CD-ROM], Dept. of Forest Resources,
Univ. Minnesota, St. Paul, MN, 12 p. Available online at: . (verified 23 Sept
2006).
Russell, M.B.; D'Amato, A.W.; Schulz, B.K.; Woodall, C.W.; Domke, G.M.; Bradford, J.B. (2014) Quantifying
understory vegetation in the U.S. Lake States: a proposed framework to inform regional forest carbon stocks.
Forestry. 87: 629-638.
Russell, M.B.; Domke, G.M.; Woodall, C.W.; D'Amato, A.W. (2015) Comparisons of allometric and climate-derived
estimates of tree coarse root carbon in forests of the United States. Carbon Balance and Management. 10: 20.
Skog, K.E. (2008) Sequestration of carbon in harvested wood products for the United States. Forest Products
Journal 58:56-72.
Smith, J.E.; Heath, L.S.; Skog, K.E.; Birdsey, R.A. (2006) Methods for calculating forest ecosystem and harvested
carbon with standard estimates for forest types of the United States. Gen. Tech. Rep. NE-343. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 216 p.
Smith, W. B., P. D. Miles, C. H. Perry, and S. A. Pugh (2009) Forest Resources of the United States, 2007. General
Technical Report WO-78, U.S. Department of Agriculture Forest Service, Washington Office.
Smith, J.E., L.S. Heath, and M.C. Nichols (2010) U.S. Forest Carbon Calculation Tool User's Guide: Forestland Carbon
Stocks and Net Annual Stock Change. General Technical Report NRS-13 revised, U.S. Department of Agriculture
Forest Service, Northern Research Station, 34 p.
Steer, Henry B. (1948) Lumber production in the United States. Misc. Pub. 669, U.S. Department of Agriculture
Forest Service. Washington, D.C.
Ulrich, Alice (1985) U.S. Timber Production, Trade, Consumption, and Price Statistics 1950-1985. Misc. Pub. 1453,
U.S. Department of Agriculture Forest Service. Washington, D.C.
Ulrich, A.H. (1989) U.S. Timber Production, Trade, Consumption, and Price Statistics, 1950-1987. USDA
Miscellaneous Publication No. 1471, U.S. Department of Agriculture Forest Service. Washington, D.C., 77.
United Nations Framework Convention on Climate Change (2013) Report on the individual review of the inventory
submission of the United States of America submitted in 2012. FCCC/ARR/2012/USA. 42 p.
USDA Forest Service (2020a) Forest Inventory and Analysis National Program: Program Features. U.S. Department
of Agriculture Forest Service. Washington, D.C. Available online at: .
Accessed 10 October 2020.
USDA Forest Service. (2020b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at:
. Accessed on 10 October 2020.
References 10-65

-------
USDA Forest Service. (2020c) Forest Inventory and Analysis National Program, FIA library: Field Guides, Methods
and Procedures. U.S. Department of Agriculture Forest Service. Washington, D.C. Available online at:
. Accessed on 10 October 2020.
USDA Forest Service (2020d) Forest Inventory and Analysis National Program, FIA library: Database
Documentation. U.S. Department of Agriculture, Forest Service, Washington Office. Available online at:
. Accessed on 10 October 2020.
U.S. Census Bureau (1976) Historical Statistics of the United States, Colonial Times to 1970, Vol. 1. Washington,
D.C.
Wear, D.N., Coulston, J.W. (2015) From sink to source: Regional variation in U.S. forest carbon futures. Scientific
Reports. 5: 16518.
Westfall, J.A., Woodall, C.W., Hatfield, M.A. (2013) A statistical power analysis of woody carbon flux from forest
inventory data. Climatic Change. 118: 919-931.
Woodall, C.W., Coulston, J.W., Domke, G.M., Walters, B.F., Wear, D.N., Smith, J.E., Anderson, H.-E., Clough, B.J.,
Cohen, W.B., Griffith, D.M., Hagan, S.C., Hanou, I.S.; Nichols, M.C., Perry, C.H., Russell, M.B., Westfall, J.A., Wilson,
B.T. (2015a) The U.S. Forest Carbon Accounting Framework: Stocks and Stock change 1990-2016. Gen. Tech. Rep.
NRS-154. Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 49 pp.
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols (2011a) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Woodall, C.W., Amacher, M.C., Bechtold, W.A., Coulston, J.W., Jovan, S., Perry, C.H., Randolph, K.C., Schulz, B.K.,
Smith, G.C., Tkacz, B., Will-Wolf, S. (2011b) "Status and future of the forest health indicators program of the United
States." Environmental Monitoring and Assessment. 177: 419-436.
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., Walters, B.F., Oswalt, S.N., Domke, G.M., Toney, C., Gray, A.N. (2013) Biomass and carbon
attributes of downed woody materials in forests of the United States. Forest Ecology and Management 305: 48-59.
Woodall, C.W., Walters, B.F., Coulston, J.W., D'Amato, A.W., Domke, G.M., Russell, M.B., Sowers, P.A. (2015b)
Monitoring network confirms land use change is a substantial component of the forest carbon sink in the eastern
United States. Scientific Reports. 5: 17028.
Zhu, Zhiliang, and McGuire, A.D., eds., (2016) Baseline and projected future carbon storage and greenhouse-gas
fluxes in ecosystems of Alaska: U.S. Geological Survey Professional Paper 1826,196 p., Available online at:
.
Forest Land Remaining Forest Lami ^n-C02 Emissions from
Forest Fires
Eidenshink, J., Schwind, B., Brewer, K., Zhu, Z.L., Quayle, B. and Howard, S., (2007). A project for monitoring trends
in burn severity. Fire ecology, 3(1), pp.3-21.
Homer, C., Dewitz, J., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N., Wickham, J. and Megown, K.,
(2015). Completion of the 2011 National Land Cover Database for the conterminous United States-representing a
decade of land cover change information. Photogrammetric Engineering & Remote Sensing, 81(5), pp.345-354.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
10-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
MTBS Data Summaries. 2018. MTBS Data Access: Fire Level Geospatial Data. (2018, August - last revised). MTBS
Project (USDA Forest Service/U.S. Geological Survey). Available online: .
Ogle, S. M., G. M. Domke, W. A. Kurz, M. T. Rocha, T. Huffman, A. Swan, J. E. Smith, C. W. Woodall, and T. Krug.
(2018). Delineating managed land for reporting national greenhouse gas emissions and removals to the United
Nations framework convention on climate change. Carbon Balance and Management 13:9.
Ruefenacht, B., Finco, M.V., Nelson, M.D., Czaplewski, R., Helmer, E.H., Blackard, J.A., Holden, G.R., Lister, A.J.,
Salajanu, D., Weyermann, D. and Winterberger, K., (2008). Conterminous US and Alaska forest type mapping using
forest inventory and analysis data. Photogrammetric Engineering & Remote Sensing, 74(11), pp.1379-1388.
Smith, J. E., L. S. Heath, and C. M. Hoover. (2013). Carbon factors and models for forest carbon estimates for the
2005-2011 National Greenhouse Gas Inventories of the United States. For. Ecology and Management 307:7-19.
USDA Forest Service (2020b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 11 October 2020.
USDA Forest Service (2020d) Forest Inventory and Analysis National Program, FIA library: Database
Documentation. U.S. Department of Agriculture, Forest Service, Washington Office. Available online at:
. Accessed on 10 October 2020.
Forest Land Remaining Forest Land: N20 Emissions from Soils
Albaugh, T.J., Allen, H.L, Fox, T.R. (2007) Historical Patterns of Forest Fertilization in the Southeastern United
States from 1969 to 2004. Southern Journal of Applied Forestry, 31,129-137(9).
Binkley, D. (2004) Email correspondence regarding the 95 percent confidence interval for area estimates of
southern pine plantations receiving N fertilizer (±20%) and the rate applied for areas receiving N fertilizer (100 to
200 pounds/acre). Dan Binkley, Department of Forest, Rangeland, and Watershed Stewardship, Colorado State
University and Stephen Del Grosso, Natural Resource Ecology Laboratory, Colorado State University. September
19, 2004.
Binkley, D., R. Carter, and H.L. Allen (1995) Nitrogen Fertilization Practices in Forestry. In: Nitrogen Fertilization in
the Environment, P.E. Bacon (ed.), Marcel Decker, Inc., New York.
Briggs, D. (2007) Management Practices on Pacific Northwest West-Side Industrial Forest Lands, 1991-2005: With
Projections to 2010. Stand Management Cooperative, SMC Working Paper Number 6, College of Forest Resources,
University of Washington, Seattle.
Fox, T.R., H. L. Allen, T.J. Albaugh, R. Rubilar, and C.A. Carlson (2007) Tree Nutrition and Forest Fertilization of Pine
Plantations in the Southern United States. Southern Journal of Applied Forestry, 31, 5-11.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
USDA Forest Service (2001) U.S. Forest Facts and Historical Trends. FS-696. U.S. Department of Agriculture Forest
Service, Washington, D.C. Available online at: .
Forest Land Remaining Forest Land: Drained Organic Soils
IPCC (2014) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands,
Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.). Published: IPCC,
Switzerland.
References 10-67

-------
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change, H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
STATSG02 (2016) Soil Survey Staff, Natural Resources Conservation Service, United States Department of
Agriculture. U.S. General Soil Map (STATSG02). Available online at .
Accessed 10 November 2016.
USDA Forest Service (2020b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, DC; 2015. Available online at . Accessed 10 October 2020.
Land Converted to Forest Land
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In R.N.
Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Brockwell, Peter J., and Richard A. Davis. Introduction to time series and forecasting. Springer, 2016.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in standing
dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon Balance and
Management. 6:14.Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., Nave, L, Swanston, C. (2017) Toward
inventory-based estimates of soil organic carbon in forests of the United States. Ecological Applications. 27(4),
1223-1235.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter
carbon stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Domke, G.M., Woodall, C.W., Walters, B.F., Smith, J.E. (2013) From models to measurements: comparing down
dead wood carbon stock estimates in the U.S. forest inventory. PLoS ONE 8(3): e59949.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Jenkins, J.C., D.C. Chojnacky, L.S. Heath, and R.A. Birdsey (2003) "National-scale biomass estimators for United
States tree species." Forest Science 49(l):12-35.0gle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003)
"Uncertainty in estimating land use and management impacts on soil organic carbon storage for U.S.
agroecosystems between 1982 and 1997." Global Change Biology 9:1521-1542.
Ogle, S.M., F.J. Breidt, and K. Paustian. (2006) "Bias and variance in model results due to spatial scaling of
measurements for parameterization in regional assessments." Global Change Biology 12:516-523.
Smith, J.E.; Heath, L.S.; Skog, K.E.; Birdsey, R.A. (2006) Methods for calculating forest ecosystem and harvested
carbon with standard estimates for forest types of the United States. Gen. Tech. Rep. NE-343. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 216 p.
USDA Forest Service (2020b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 10 October 2020.
USDA Forest Service (2020c) Forest Inventory and Analysis National Program, FIA library: Field Guides, Methods
and Procedures. U.S. Department of Agriculture Forest Service. Washington, D.C. Available online at:
. Accessed on 10 October 2020.
10-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory, Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
USDA-NRCS (1997) "National Soil Survey Laboratory Characterization Data," Digital Data, Natural Resources
Conservation Service, U.S. Department of Agriculture. Lincoln, NE.
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols (2011a) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., Walters, B.F., Coulston, J.W., D'Amato, A.W., Domke, G.M., Russell, M.B., Sowers, P.A. (2015b)
Monitoring network confirms land use change is a substantial component of the forest carbon sink in the eastern
United States. Scientific Reports. 5: 17028.
Woodall, C.W., Walters, B.F., Oswalt, S.N., Domke, G.M., Toney, C., Gray, A.N. (2013) Biomass and carbon
attributes of downed woody materials in forests of the United States. Forest Ecology and Management 305: 48-59.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L, Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies. ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Cropland Remaining Cropland
Armentano, T. V., and E.S. Menges (1986). Patterns of change in the carbon balance of organic soil-wetlands of the
temperate zone. Journal of Ecology 74: 755-774.
Brady, N.C. and R.R. Weil (1999) The Nature and Properties of Soils. Prentice Hall. Upper Saddle River, NJ, 881.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Cheng, B., and D.M. Titterington (1994) "Neural networks: A review from a statistical perspective." Statistical
science 9: 2-30.
Claassen, R., M. Bowman, J. McFadden, D. Smith, and S. Wallander (2018) Tillage intensity and conservation
cropping in the United States, EIB 197. United States Department of Agriculture, Economic Research Service,
Washington, D.C.
Conant, R. T., K. Paustian, and E.T. Elliott (2001). "Grassland management and conversion into grassland: effects on
soil carbon." Ecological Applications 11: 343-355.
CTIC (2004) National Crop Residue Management Survey: 1989-2004. Conservation Technology Information Center,
Purdue University, Available online at: .
Daly, C., R.P. Neilson, and D.L Phillips (1994) "A Statistical-Topographic Model for Mapping Climatological
Precipitation Over Mountainous Terrain." Journal of Applied Meteorology 33:140-158.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001) "Simulated
Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In Modeling Carbon
and Nitrogen Dynamics for Soil Management, Schaffer, M., L. Ma, S. Hansen, (eds.). CRC Press, Boca Raton, Florida,
pp. 303-332.
Del Grosso, S.J., S.M. Ogle, W.J. Parton (2011) Soil organic matter cycling and greenhouse gas accounting
methodologies, Chapter 1, pp 3-13 DOI: 10.1021/bk-2011-1072.ch001. In: Understanding Greenhouse Gas
References 10-69

-------
Emissions from Agricultural Management, L Guo, A. Gunasekara, L McConnell (eds.). American Chemical Society,
Washington, D.C.
Edmonds, L, R. L Kellogg, B. Kintzer, L Knight, C. Lander, J. Lemunyon, D. Meyer, D.C. Moffitt, and J. Schaefer
(2003) "Costs associated with development and implementation of Comprehensive Nutrient Management Plans."
Part I—Nutrient management, land treatment, manure and wastewater handling and storage, and recordkeeping.
Natural Resources Conservation Service, U.S. Department of Agriculture. Available online at:
.
Euliss, N., and R. Gleason (2002) Personal communication regarding wetland restoration factor estimates and
restoration activity data. Ned Euliss and Robert Gleason of the U.S. Geological Survey, Jamestown, ND, to Stephen
Ogle of the National Resource Ecology Laboratory, Fort Collins, CO. August 2002.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) Completion of the 2006
National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Griscom, B. W., Adams, J., Ellis, P. W., Houghton, R. A., Lomax, G., Miteva, D. A., Schlesinger, W. H., Shoch, D., Siikamaki, J.
V., Smith, P., Woodbury, P., Zganjar, C., Blackman, A., Campari, J., Conant, R. T., Delgado, C., Elias, P., Gopalakrishna, T.,
Hamsik, M. R., Herrero, M., Kiesecker, J., Landis, E., Laestadius, L., Leavitt, S. M., Minnemeyer, S., Polasky, S., Potapov, P.,
Putz, F. E., Sanderman, J., Silvius, M., Wollenberg, E. & Fargione, J. (2017) "Natural climate solutions." Proceedings of the
National Academy of Sciences of the United States of America 114(44): 11645-11650.
Hijmans, R.J., S.E. Cameron, J.L Parra, P.G. Jones and A. Jarvis (2005) Very high resolution interpolated climate
surfaces for global land areas. International Journal of Climatology 25:1965-1978.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and
Wickham, J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v.
81, no. 5, p. 345-354.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC (2003) Good Practice Guidance for Land Use, Land-Use Change, and Forestry. The Intergovernmental Panel on
Climate Change, National Greenhouse Gas Inventories Programme, J. Penman, et al., eds. August 13, 2004.
Available online at: .
Lai, R., Kimble, J. M., Follett, R. F. & Cole, C. V. (1998) The potential of U.S. cropland to sequester carbon and
mitigate the greenhouse effect. Chelsea, Ml: Sleeping Bear Press, Inc.
Little, R. (1988) "Missing-data adjustments in large surveys." Journal of Business and Economic Statistics 6: 287-
296.
McGill, W.B., and C.V. Cole (1981) Comparative aspects of cycling of organic C, N, S and P through soil organic
matter. Geoderma 26:267-286.
Metherell, A.K., LA. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURY Soil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
Mesinger, F., G. DiMego, E. Kalnay, K. Mitchell, P. C. Shafran, W. Ebisuzaki, D. Jovic, J. Woollen, E. Rogers, E. H.
Berbery, M. B. Ek, Y. Fan, R. Grumbine, W. Higgins, H. Li, Y. Lin, G. Manikin, D. Parrish, and W. Shi (2006) North
American regional reanalysis. Bulletin of the American Meteorological Society 87:343-360.
10-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
NRCS (1999) Soil Taxonomy: A basic system of soil classification for making and interpreting soil surveys, 2nd
Edition. Agricultural Handbook Number 436, Natural Resources Conservation Service, U.S. Department of
Agriculture, Washington, D.C.
NRCS (1997) "National Soil Survey Laboratory Characterization Data," Digital Data, Natural Resources Conservation
Service, U.S. Department of Agriculture. Lincoln, NE.
NRCS (1981) Land Resource Regions and Major Land Resource Areas of the United States, USDA Agriculture
Handbook 296, United States Department of Agriculture, Natural Resources Conservation Service, National Soil
Survey Cente., Lincoln, NE, pp. 156.
Ogle, S. M., Alsaker, C., Baldock, J., Bernoux, M., Breidt, F. J., McConkey, B., Regina, K. & Vazquez-Amabile, G. G.
(2019) "Climate and Soil Characteristics Determine Where No-Till Management Can Store Carbon in Soils and
Mitigate Greenhouse Gas Emissions." Scientific Reports 9(1): 11665.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams, K. Killian, and K. Paustian (2010) "Scale and uncertainty in modeled
soil organic carbon stock changes for U.S. croplands using a process-based model." Global Change Biology 16:810-
820.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams and K. Paustian (2007) "Empirically-Based Uncertainty Associated with
Modeling Carbon Sequestration Rates in Soils." Ecological Modeling 205:453-463.
Ogle, S.M., F.J. Breidt, and K. Paustian (2006) "Bias and variance in model results due to spatial scaling of
measurements for parameterization in regional assessments." Global Change Biology 12:516-523.
Ogle, S. M., et al. (2005) "Agricultural management impacts on soil organic carbon storage under moist and dry
climatic conditions of temperate and tropical regions." Biogeochemistry 72: 87-121.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel: Description
and Testing". Glob. Planet. Chang. 19: 35-48.
Parton, W.J., D.S. Ojima, C.V. Cole, and D.S. Schimel (1994) "A General Model for Soil Organic Matter Dynamics:
Sensitivity to litter chemistry, texture and management," in Quantitative Modeling of Soil Forming Processes.
Special Publication 39, Soil Science Society of America, Madison, Wl, 147-167.
Parton, W.J., D.S. Schimel, C.V. Cole, D.S. Ojima (1987) "Analysis of factors controlling soil organic matter levels in
Great Plains grasslands." Soil Science Society of America Journal 51:1173-1179.
Parton, W.J., J.W.B. Stewart, C.V. Cole. (1988) "Dynamics of C, N, P, and S in grassland soils: a model."
Biogeochemistry 5:109-131.
Paustian, K., et al. (1997a). "Agricultural soils as a sink to mitigate C02 emissions." Soil Use and Management 13:
230-244.
Paustian, K., et al. (1997b) Management controls on soil carbon. In Soil organic matter in temperate
agroecosystems: long-term experiments in North America (Paul E.A., K. Paustian, and C.V. Cole, eds.). Boca Raton,
CRC Press, pp. 15-49.
Potter, C. S., J.T. Randerson, C.B. Fields, P.A. Matson, P.M. Vitousek, H.A. Mooney, and S.A. Klooster (1993)
'Terrestrial ecosystem production: a process model based on global satellite and surface data." Global
Biogeochemical Cycles 7:811-841.
Potter, C., S. Klooster, A. Huete, and V. Genovese (2007) Terrestrial carbon sinks for the United States predicted
from MODIS satellite data and ecosystem modeling. Earth Interactions 11, Article No. 13, DOI 10.1175/EI228.1.
PRISM Climate Group (2018) PRISM Climate Data, Oregon State University, ,
downloaded 18 July 2018.
References 10-71

-------
Pukelsheim, F. (1994) 'The 3-Sigma-Rule." American Statistician 48:88-91
Soil Survey Staff (2016) State Soil Geographic (STATSGO) Database for State. Natural Resources Conservation
Service, United States Department of Agriculture. Available online at:
.
Spencer, S., S.M. Ogle, F.J. Breidt, J. Goebel, and K. Paustian. (2011) "Designing a national soil carbon monitoring
network to support climate change policy: a case example for US agricultural lands." Greenhouse Gas Management
& Measurement 1: 167-178.
Towery, D. (2001) Personal Communication. Dan Towery regarding adjustments to the CTIC (1998) tillage data to
reflect long-term trends, Conservation Technology Information Center, West Lafayette, IN, and Marlen Eve,
National Resource Ecology Laboratory, Fort Collins, CO. February 2001.
USDA-ERS (2018) Agricultural Resource Management Survey (ARMS) Farm Financial and Crop Production Practices:
Tailored Reports. Available online at: .
USDA-ERS (1997) Cropping Practices Survey Data—1995. Economic Research Service, United States Department of
Agriculture. Available online at: .
USDA-FSA (2015) Conservation Reserve Program Monthly Summary-September 2015. U.S. Department of
Agriculture, Farm Service Agency, Washington, D.C. Available online at: .
USDA-NASS (2017) 2017 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at .
USDA-NASS (2012) 2012 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at .
USDA-NASS (2004) Agricultural Chemical Usage: 2003 Field Crops Summary. Report AgChl(04)a. National
Agricultural Statistics Service, U.S. Department of Agriculture, Washington, D.C. Available online at:
.
USDA-NASS (1999) Agricultural Chemical Usage: 1998 Field Crops Summary. Report AgCHl(99). National
Agricultural Statistics Service, U.S. Department of Agriculture, Washington, DC. Available online at:
.
USDA-NASS (1992) Agricultural Chemical Usage: 1991 Field Crops Summary. Report AgChl(92). National
Agricultural Statistics Service, U.S. Department of Agriculture, Washington, D.C. Available online at:
.
USDA-NRCS (2012) Assessment of the Effects of Conservation Practices on Cultivated Cropland in the Upper
Mississippi River Basin. U.S. Department of Agriculture, Natural Resources Conservation Service,
.
USDA-NRCS (2018a) Summary Report: 2015 National Resources Inventory. Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
USDA-NRCS (2018b) CEAP Cropland Farmer Surveys. USDA Natural Resources Conservation Service.
https://www.nrcs.usda.gov/wps/portal/nrcs/detail/national/technical/nra/ceap/na/?cid=nrcsl43_014163.
USDA-NRCS (2000) Digital Data and Summary Report: 1997 National Resources Inventory. Revised December 2000.
Resources Inventory Division, Natural Resources Conservation Service, United States Department of Agriculture,
Beltsville, MD.
Van Buuren, S. (2012) "Flexible imputation of missing data." Chapman & Hall/CRC, Boca Raton, FL.
10-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L., Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) "A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies." ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Zomer RJ, Trabucco A, Bossio DA, van Straaten O, Verchot LV (2008) Climate Change Mitigation: A Spatial Analysis
of Global Land Suitability for Clean Development Mechanism Afforestation and Reforestation. Agric. Ecosystems
and Envir. 126: 67-80.
Zomer RJ, Bossio DA, Trabucco A, Yuanjie L, Gupta DC & Singh VP (2007) Trees and Water: Smallholder
Agroforestry on Irrigated Lands in Northern India. Colombo, Sri Lanka: International Water Management Institute,
pp 45. (IWMI Research Report 122).
Land Converted to Cropland
Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001) "Simulated
Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In Modeling Carbon
and Nitrogen Dynamics for Soil Management, Schaffer, M., L. Ma, S. Hansen, (eds.). CRC Press, Boca Raton, Florida,
pp. 303-332.
Del Grosso, S.J., S.M. Ogle, W.J. Parton (2011) "Soil organic matter cycling and greenhouse gas accounting
methodologies." Chapter 1, pp 3-13 DOI: 10.1021/bk-2011-1072.ch001. In: Understanding Greenhouse Gas
Emissions from Agricultural Management (L. Guo, A. Gunasekara, L. McConnell. Eds.), American Chemical Society,
Washington, D.C.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001) "Simulated
Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In Schaffer, M., L. Ma,
S. Hansen, (eds.); Modeling Carbon and Nitrogen Dynamics for Soil Management. CRC Press. Boca Raton, Florida.
303-332.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) "Accounting for density reduction and structural loss in
standing dead trees: Implications for forest biomass and carbon stock estimates in the United States". Carbon
Balance and Management 6:14.
Domke, G.M., et al. (2013) "From models to measurements: comparing down dead wood carbon stock estimates in
the U.S. forest inventory." PLoS ONE 8(3): e59949.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) "A framework for estimating litter
carbon stocks in forests of the United States." Science of the Total Environment 557-558: 469-478.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) "Completion of the
2006 National Land Cover Database for the Conterminous United States." PE&RS, Vol. 77(9):858-864.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and Wickham,
J. (2007) "Completion of the 2001 National Land Cover Database for the Conterminous United States."
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) "Completion of the 2011 National Land Cover Database for the conterminous United States-
References 10-73

-------
Representing a decade of land cover change information." Photogrammetric Engineering and Remote Sensing 81:
345-354.
Houghton, R.A., et al. (1983) "Changes in the carbon content of terrestrial biota and soils between 1860 and 1980:
a net release of C02 to the atmosphere." Ecological Monographs 53: 235-262.
Houghton, R. A. and Nassikas, A. A. (2017) "Global and regional fluxes of carbon from land use and land cover
change 1850-2015." Global Biogeochemical Cycles 31(3): 456-472.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Jenkins, J.C., D.C. Chojnacky, L.S. Heath, and R.A. Birdsey (2003) "National-scale biomass estimators for United
States tree species." Forest Science 49(l):12-35.
Metherell, A.K., LA. Harding, C.V. Cole, and W.J. Parton (1993) CENTURY Soil Organic Matter Model Environment.
Agroecosystem version 4.0. Technical documentation, GPSRTech. Report No. 4, USDA/ARS, Ft. Collins, CO.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams, K. Killian, and K. Paustian (2010) "Scale and uncertainty in modeled
soil organic carbon stock changes for U.S. croplands using a process-based model." Global Change Biology 16:810-
820.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel: Description
and Testing". Glob. Planet. Chang. 19: 35-48.
Parton, W.J., D.S. Ojima, C.V. Cole, and D.S. Schimel (1994) "A General Model for Soil Organic Matter Dynamics:
Sensitivity to litter chemistry, texture and management," in Quantitative Modeling of Soil Forming Processes.
Special Publication 39, Soil Science Society of America, Madison, Wl, 147-167.
Parton, W.J., D.S. Schimel, C.V. Cole, D.S. Ojima (1987) "Analysis of factors controlling soil organic matter levels in
Great Plains grasslands." Soil Science Society of America Journal 51:1173-1179.
Parton, W.J., J.W.B. Stewart, C.V. Cole. (1988) "Dynamics of C, N, P, and S in grassland soils: a model."
Biogeochemistry 5:109-131.
PRISM Climate Group (2018) PRISM Climate Data, Oregon State University, ,
downloaded 18 July 2018.
Smith, J.E.; Heath, L.S.; Skog, K.E.; Birdsey, R.A. (2006) Methods for calculating forest ecosystem and harvested
carbon with standard estimates for forest types of the United States. Gen. Tech. Rep. NE-343. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 216 p.
Tubiello, F. N., et al. (2015) "The Contribution of Agriculture, Forestry and other Land Use activities to Global
Warming, 1990-2012." Global Change Biology 21:2655-2660.
USDA Forest Service (2020) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 10 October 2020.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory. Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
10-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols (2011) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L, Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) "A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies." ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Grassland Remaining Grassland: Soil Carbon Stock Changes
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Del Grosso, S.J., S.M. Ogle, W.J. Parton (2011) Soil organic matter cycling and greenhouse gas accounting
methodologies, Chapter 1, pp 3-13 DOI: 10.1021/bk-2011-1072.ch001. In: Understanding Greenhouse Gas
Emissions from Agricultural Management (L Guo, A. Gunasekara, L McConnell. Eds.), American Chemical Society,
Washington, D.C.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001) "Simulated
Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In Modeling Carbon
and Nitrogen Dynamics for Soil Management, Schaffer, M., L. Ma, S. Hansen, (eds.). CRC Press, Boca Raton, Florida,
pp. 303-332.
Edmonds, L, R. L Kellogg, B. Kintzer, L. Knight, C. Lander, J. Lemunyon, D. Meyer, D.C. Moffitt, and J. Schaefer
(2003) "Costs associated with development and implementation of Comprehensive Nutrient Management Plans."
Part I—Nutrient management, land treatment, manure and wastewater handling and storage, and recordkeeping.
Natural Resources Conservation Service, U.S. Department of Agriculture. Available online at:
.
EPA (1999) Biosolids Generation, Use and Disposal in the United States. Office of Solid Waste, U.S. Environmental
Protection Agency. Available online at: .
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and Wickham,
J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v. 81,
no. 5, p. 345-354.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Kellogg, R.L., C.H. Lander, D.C. Moffitt, and N. Gollehon (2000) Manure Nutrients Relative to the Capacity of
Cropland and Pastureland to Assimilate Nutrients: Spatial and Temporal Trends for the United States. U.S.
Department of Agriculture, Washington, D.C. Publication number nps00-0579.
Metherell, A.K., LA. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURY Soil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
NEBRA (2007) A National Biosolids Regulation, Quality, End Use & Disposal Survey. North East Biosolids and
Residuals Association. July 21, 2007.
References 10-75

-------
Nusser, S.M. and J.J. Goebel (1997) The national resources inventory: a long-term multi-resource monitoring
programme. Environmental and Ecological Statistics 4:181-204.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams, K. Killian, and K. Paustian (2010) "Scale and uncertainty in modeled
soil organic carbon stock changes for U.S. croplands using a process-based model." Global Change Biology 16:810-
820.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
Parton, W.J., D.S. Ojima, C.V. Cole, and D.S. Schimel (1994) "A General Model for Soil Organic Matter Dynamics:
Sensitivity to litter chemistry, texture and management," in Quantitative Modeling of Soil Forming Processes.
Special Publication 39, Soil Science Society of America, Madison, Wl, 147-167.
Parton, W.J., D.S. Schimel, C.V. Cole, D.S. Ojima (1987) "Analysis of factors controlling soil organic matter levels in
Great Plains grasslands." Soil Science Society of America Journal 51:1173-1179.
Parton, W.J., J.W.B. Stewart, C.V. Cole. (1988) "Dynamics of C, N, P, and S in grassland soils: a model."
Biogeochemistry 5:109-131.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel: Description
and Testing". Glob. Planet. Chang. 19: 35-48.PRISM Climate Group, Oregon State University,
, created 24 July 2015.
PRISM Climate Group (2018) PRISM Climate Data, Oregon State University, ,
downloaded 18 July 2018.
United States Bureau of Land Management (BLM) (2014) Rangeland Inventory, Monitoring, and Evaluation
Reports. Bureau of Land Management. U.S. Department of the Interior. Available online at:
.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory. Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
USDA Forest Service (2020) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 10 October 2020.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L, Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) "A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies." ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Grassland Remaining Grassland: PJon-C02 Emissions from
issiand Fires
Anderson, R.C. Evolution and origin of the Central Grassland of North America: climate, fire and mammalian
grazers. Journal of the Torrey Botanical Society 133: 626-647.
Andreae, M.O. and P. Merlet (2001) Emission of trace gases and aerosols from biomass burning. Global
Biogeochemical Cycles 15:955-966.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Chapin, F.S., S.F. Trainor, O. Huntington, A.L. Lovecraft, E. Zavaleta, D.C. Natcher, A.D. McGuire, J.L. Nelson, L. Ray,
M. Calef, N. Fresco, H. Huntington, T.S. Rupp, L. DeWilde, and R.L Naylor (2008) Increasing wildfires in Alaska's
Boreal Forest: Pathways to potential solutions of a wicked problem. Bioscience 58:531-540.
10-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Daubenmire, R. (1968) Ecology of fire in grasslands. Advances in Ecological Research 5:209-266.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and Wickham,
J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v. 81,
no. 5, p. 345-354.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Ogle, S.M., S. Spencer, M. Hartman, L. Buendia, L. Stevens, D. du Toit, J. Witi (2016) "Developing national baseline
GHG emissions and analyzing mitigation potentials for agriculture and forestry using an advanced national GHG
inventory software system." In Advances in Agricultural Systems Modeling 6, Synthesis and Modeling of
Greenhouse Gas Emissions and Carbon Storage in Agricultural and Forestry Systems to Guide Mitigation and
Adaptation, S. Del Grosso, LR. Ahuja and W.J. Parton (eds.), American Society of Agriculture, Crop Society of
America and Soil Science Society of America, pp. 129-148.
Nusser, S.M. and J.J. Goebel (1997) The national resources inventory: a long-term multi-resource monitoring
programme. Environmental and Ecological Statistics 4:181-204.
USDA-NRCS (2015) Summary Report: 2012 National Resources Inventory, Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa. Available
online at: .
Land Converted to Grassland
Asner, G.P., Archer, S., Hughes, R.F., Ansley, R.J. and Wessman, C.A. (2003) "Net changes in regional woody
vegetation cover and carbon storage in Texas drylands, 1937-1999." Global Change Biology 9(3): 316-335.
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In R.N.
Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Breshears, D.D., Knapp, A.K., Law, D.J., Smith, M.D., Twidwell, D. and Wonkka, C.L., 2016. Rangeland Responses to
Predicted Increases in Drought Extremity. Rangelands, 38(4), pp.191-196.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Del Grosso, S.J., S.M. Ogle, W.J. Parton. (2011) Soil organic matter cycling and greenhouse gas accounting
methodologies, Chapter 1, pp 3-13 DOI: 10.1021/bk-2011-1072.ch001. In: Understanding Greenhouse Gas
Emissions from Agricultural Management (L. Guo, A. Gunasekara, L. McConnell. Eds.), American Chemical Society,
Washington, D.C.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001) "Simulated
Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In Modeling Carbon
and Nitrogen Dynamics for Soil Management (Schaffer, M., L. Ma, S. Hansen, (eds.). CRC Press, Boca Raton, Florida,
pp. 303-332.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in standing
dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon Balance and
Management. 6:14.
References 10-77

-------
Domke, G.M., et al. 2013. From models to measurements: comparing down dead wood carbon stock estimates in
the U.S. forest inventory. PLoS ONE 8(3): e59949.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter
carbon stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Epstein, H.E., Gill, R.A., Paruelo, J.M., Lauenroth, W.K., Jia, G.J. and Burke, I.C., 2002. The relative abundance of
three plant functional types in temperate grasslands and shrublands of North and South America: effects of
projected climate change. Journal of Biogeography, 29(7), pp.875-888.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and Wickham,
J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v. 81,
no. 5, p. 345-354.
Houghton, R.A., et al. (1983) "Changes in the carbon content of terrestrial biota and soils between 1860 and 1980:
a net release of C02 to the atmosphere." Ecological Monographs 53: 235-262.
Houghton, R. A. and Nassikas, A. A. (2017) "Global and regional fluxes of carbon from land use and land cover
change 1850-2015." Global Biogeochemical Cycles 31(3): 456-472.
Huang, C.Y., Asner, G.P., Martin, R.E., Barger, N.N. and Neff, J.C. (2009) "Multiscale analysis of tree cover and
aboveground carbon stocks in pinyon-juniper woodlands." Ecological Applications 19(3): 668-681.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change, [H.S. Eggleston, L. Buendia, K. Miwa, T
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Jenkins, J.C., D.C. Chojnacky, L.S. Heath, and R.A. Birdsey (2003) "National-scale biomass estimators for United
States tree species." Forest Science 49(l):12-35.
Jurena, P.N. and Archer, S., (2003). Woody plant establishment and spatial heterogeneity in grasslands. Ecology,
84(4), pp.907-919.
Lenihan, J.M., Drapek, R., Bachelet, D. and Neilson, R.P., (2003). Climate change effects on vegetation distribution,
carbon, and fire in California. Ecological Applications, 13(6), pp.1667-1681.
Metherell, A.K., LA. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURY Soil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams, K. Killian, and K. Paustian (2010) "Scale and uncertainty in modeled
soil organic carbon stock changes for U.S. croplands using a process-based model." Global Change Biology 16:810-
820.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
10-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Parton, W.J., D.S. Ojima, C.V. Cole, and D.S. Schimel (1994) "A General Model for Soil Organic Matter Dynamics:
Sensitivity to litter chemistry, texture and management," in Quantitative Modeling of Soil Forming Processes.
Special Publication 39, Soil Science Society of America, Madison, Wl, 147-167.
Parton, W.J., D.S. Schimel, C.V. Cole, D.S. Ojima (1987) "Analysis of factors controlling soil organic matter levels in
Great Plains grasslands." Soil Science Society of America Journal 51:1173-1179.
Parton, W.J., J.W.B. Stewart, C.V. Cole (1988) "Dynamics of C, N, P, and S in grassland soils: a model."
Biogeochemistry 5:109-131.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel: Description
and Testing". Glob. Planet. Chang. 19: 35-48.
PRISM Climate Group (2018) PRISM Climate Data, Oregon State University, ,
downloaded 18 July 2018.
Scholes, R.J. and Archer, S.R., 1997. Tree-grass interactions in savannas 1. Annual review of Ecology and
Systematics, 28(1), pp.517-544.
Sims, P.L., Singh, J.S. and Lauenroth, W.K., 1978. The structure and function often western North American
grasslands: I. Abiotic and vegetational characteristics. The Journal of Ecology, pp.251-285.
Smith, J.E.; Heath, L.S.; Skog, K.E.; Birdsey, R.A. (2006) Methods for calculating forest ecosystem and harvested
carbon with standard estimates for forest types of the United States. Gen. Tech. Rep. NE-343. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 216 p.
Tarhouni, M., et al. (2016) Measurement of the aboveground biomass of some rangeland species using a digital
non-destructive technique. Botany Letters 163(3):281-287.
Tubiello, F. N., et al. (2015) "The Contribution of Agriculture, Forestry and other Land Use activities to Global
Warming, 1990-2012." Global Change Biology 21:2655-2660.
United States Bureau of Land Management (BLM) (2014) Rangeland Inventory, Monitoring, and Evaluation
Reports. Bureau of Land Management. U.S. Department of the Interior. Available online at:
.
USDA Forest Service (2019) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, DC; 2015. Available online at . Accessed 2 October 2019.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory. Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols. (2011) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L., Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) "A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies." ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
References 10-79

-------
Wetlands Remaining Wetlands: C02, CH4, and W20 Emissions
from Peatlands Remaining Peatlands
Apodaca, L. (2011) Email correspondence. Lori Apodaca, Peat Commodity Specialist, USGS and Emily Rowan, ICF
International. November.
Apodaca, L. (2008) E-mail correspondance. Lori Apodaca, Peat Commodity Specialist, USGS and Emily Rowan, ICF
International. October and November.
Cleary, J., N. Roulet and T.R. Moore (2005) "Greenhouse gas emissions from Canadian peat extraction, 1990-2000:
A life-cycle analysis." Ambio 34:456-461.
Division of Geological & Geophysical Surveys (DGGS), Alaska Department of Natural Resources (1997-2015)
Alaska's Mineral Industry Report (1997-2014). Alaska Department of Natural Resources, Fairbanks, AK. Available
online at .
IPCC (2013) 2013 Supplement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.). Published: IPCC,
Switzerland.
IPCC (2007) Climate Change 2007: Synthesis Report. Contribution of Working Groups I, II and III to the Fourth
Assessment Report (AR4) of the IPCC. The Intergovernmental Panel on Climate Change, R.K. Pachauri, A. Resinger
(eds.). Geneva, Switzerland.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
Szumigala, D.J. (2011) Phone conversation. Dr. David Szumigala, Division of Geological and Geophysical Surveys,
Alaska Department of Natural Resources and Emily Rowan, ICF International. January 18, 2011.
Szumigala, D.J. (2008) Phone conversation. Dr. David Szumigala, Division of Geological and Geophysical Surveys,
Alaska Department of Natural Resources and Emily Rowan, ICF International. October 17, 2008.
USGS (1991-2017) Minerals Yearbook: Peat (1994-2017). United States Geological Survey, Reston, VA. Available
online at .
USGS (2018) Minerals Yearbook: Peat - Tables-only release (2018). United States Geological Survey, Reston, VA.
Available online at .
USGS (2020) Mineral Commodity Summaries: Peat (2020). United States Geological Survey, Reston, VA. Available
online at .
Wetlands Remaining Coastal Wetlands: Emissions and
Removals from Coastal Wetlands Remaining Coastal Wetlands
Bianchi, T. S., Allison, M. A., Zhao, J., Li, X., Comeaux, R. S., Feagin, R. A., & Kulawardhana, R. W. (2013) Historical
reconstruction of mangrove expansion in the Gulf of Mexico: linking climate change with carbon sequestration in
coastal wetlands. Estuarine, Coastal and Shelf Science 119: 7-16.
Byrd, K. B., Ballanti, L. R., Thomas, N. M., Nguyen, D. K., Holmquist, J. R., Simard, M., Windham-Myers, L., Schile, L.
M., Parker, V. T.,... and Castaneda-Moya, E. (2017) Biomass/Remote Sensing dataset: 30m resolution tidal marsh
biomass samples and remote sensing data for six regions in the conterminous United States: U.S. Geological Survey
data release, .
Byrd, K. B., Ballanti, L., Thomas, N., Nguyen, D., Holmquist, J.R., Simard, M., and Windham-Myers, L. (2018) A
remote sensing-based model of tidal marsh aboveground carbon stocks for the conterminous United States. ISPRS
10-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Journal of Photogrammetry and Remote Sensing 139: 255-271.
Byrd, K. B., Ballanti, L., Thomas, N., Nguyen, D., Holmquist, J.R., Simard, M., and Windham-Myers, L. (2020)
Corrigendum to "A remote sensing-based model of tidal marsh aboveground carbon stocks for the conterminous
United States". ISPRS Journal of Photogrammetry and Remote Sensing 166: 63-67.
Callaway, J. C., Borgnis, E. L., Turner, R. E. & Milan, C. S. (2012a) Carbon sequestration and sediment accretion in
San Francisco Bay tidal wetlands. Estuaries and Coasts 35(5): 1163-1181.
Callaway, J. C., Borgnis, E. L., Turner, R. E., Milan, C. S., Goodfriend, W., & Richmond, S. (2012b) "Wetland Sediment
Accumulation at Corte Madera Marsh and Muzzi Marsh". San Francisco Bay Conservation and Development
Commission.
Church, T. M., Sommerfield, C. K., Velinsky, D. J., Point, D., Benoit, C., Amouroux, D. & Donard, O. F. X. (2006)
Marsh sediments as records of sedimentation, eutrophication and metal pollution in the urban Delaware Estuary.
Marine Chemistry 102(1-2): 72-95.
Couvillion, B. R., Barras, J. A., Steyer, G. D., Sleavin, W., Fischer, M., Beck, H., & Heckman, D. (2011) Land area
change in coastal Louisiana (1932 to 2010) (pp. 1-12). U.S. Department of the Interior, U.S. Geological Survey.
Couvillion, B. R., Fischer, M. R., Beck, H. J. and Sleavin, W. J. (2016) Spatial Configuration Trends in Coastal
Louisiana from 1986 to 2010. Wetlands 1-13.
Craft, C. B., & Richardson, C. J. (1998) Recent and long-term organic soil accretion and nutrient accumulation in the
Everglades. Soil Science Society of America Journal 62(3): 834-843.
Crooks, S., Findsen, J., Igusky, K., Orr, M. K. and Brew, D. (2009) Greenhouse Gas Mitigation Typology Issues Paper:
Tidal Wetlands Restoration. Report by PWA and SAIC to the California Climate Action Reserve.
Crooks, S., Rybczyk, J., O'Connell, K., Devier, D. L, Poppe, K., Emmett-Mattox, S. (2014) Coastal Blue Carbon
Opportunity Assessment for the Snohomish Estuary: The Climate Benefits of Estuary Restoration. Report by
Environmental Science Associates, Western Washington University, EarthCorps, and Restore America's Estuaries.
DeLaune, R. D., & White, J. R. (2012) Will coastal wetlands continue to sequester carbon in response to an increase
in global sea level?: A case study of the rapidly subsiding Mississippi river deltaic plain. Climatic Change, 110(1),
297-314.
Holmquist, J. R., Windham-Myers, L, Bliss, N., Crooks, S., Morris, J. T., Megonigal, J. P. & Woodrey, M. (2018)
Accuracy and Precision of Tidal Wetland Soil Carbon Mapping in the Conterminous United States. Scientific reports
8(1): 9478.
Hu, Z., Lee, J. W., Chandran, K., Kim, S. and Khanal, S. K. (2012) N20 Emissions from Aquaculture: A Review.
Environmental Science & Technology 46(12): 6470-6480.
Hussein, A. H., Rabenhorst, M. C. & Tucker, M. L. (2004) Modeling of carbon sequestration in coastal marsh soils.
Soil Science Society of America Journal 68(5): 1786-1795.
IPCC (2000) Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories.
Quantifying Uncertainties in Practice, Chapter 6. Penman, J., Kruger, D., Galbally, I., Hiraishi, T., Nyenzi, B.,
Emmanuel, S., Buendia, L, Hoppaus, R., Martinsen, T., Meijer, J., Miwa, K. and Tanabe, K. (eds). Institute of Global
Environmental Strategies (IGES), on behalf of the Intergovernmental Panel on Climate Change (IPCC): Hayama,
Japan.
IPCC (2003) Good Practice Guidance for Land Use, Land-Use Change and Forestry. LUCF Sector Good Practice
Guidance, Chapter 3. Penman, J., Gytarsky, M., Hiraishi, T., Krug, T., Kruger, D., Pipatti, R., Buendia, L, Miwa, K.,
Ngara, T., Tanabe, K. and Wagner, F. (eds). Institute of Global Environmental Strategies (IGES), on behalf of the
Intergovernmental Panel on Climate Change (IPCC): Hayama, Japan.
IPCC (2006) IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by the National Greenhouse Gas
Inventories Programme, Eggleston H.S., Buendia L, Miwa K., Ngara T. and Tanabe K. (eds). IGES, Japan.
References 10-81

-------
IPCC (2014) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.). Published: IPCC,
Switzerland.
Kearney, M. S. & Stevenson, J. C. (1991) Island land loss and marsh vertical accretion rate evidence for historical
sea-level changes in Chesapeake Bay. Journal of Coastal Research 7(2): 403-415.
Koster, D., Lichter, J., Lea, P. D., & Nurse, A. (2007) Historical eutrophication in a river-estuary complex in mid-
coast Maine. Ecological Applications 17(3): 765-778.
Lu, M & Megonigal, J. P. (2017) Final Report for RAE Baseline Assessment Project. Memo to Silvestrum Climate
Associates by Smithsonian Environmental Research Center, Maryland.
Lynch, J. C. (1989) Sedimentation and nutrient accumulation in mangrove ecosystems of the Gulf of Mexico. M.S.
thesis, Univ. of Southwestern Louisiana, Lafayette, LA.
Marchio, D. A., Savarese, M., Bovard, B., & Mitsch, W. J. (2016) Carbon sequestration and sedimentation in
mangrove swamps influenced by hydrogeomorphic conditions and urbanization in Southwest Florida. Forests 7:
116-135.
McCombs, J. W., Herold, N. D., Burkhalter, S. G. and Robinson C. J. (2016) Accuracy Assessment of NOAA Coastal
Change Analysis Program 2006-2010 Land Cover and Land Cover Change Data. Photogrammetric Engineering &
Remote Sensing. 82:711-718.
Merrill, J. Z. (1999) Tidal Freshwater Marshes as Nutrient Sinks: particulate Nutrient Burial and Denitrification.
Ph.D. Dissertation, University of Maryland, College Park, MD, 342 pp.
National Marine Fisheries Service (2020) Fisheries of the United States, 2017. U.S. Department of Commerce,
NOAA Current Fishery Statistics No. 2018.
National Oceanic and Atmospheric Administration, Office for Coastal Management (2020) Coastal Change Analysis
Program (C-CAP) Regional Land Cover. Charleston, SC: NOAA Office for Coastal Management. Accessed October
2020 at .
Noe, G. B., Hupp, C. R., Bernhardt, C. E., & Krauss, K. W. (2016) Contemporary deposition and long-term
accumulation of sediment and nutrients by tidal freshwater forested wetlands impacted by sea level rise. Estuaries
and Coasts 39(4): 1006-1019.
Orson, R. A., Simpson, R. L, & Good, R. E. (1990) Rates of sediment accumulation in a tidal freshwater marsh.
Journal of Sedimentary Research 60(6): 859-869.
Orson, R., Warren, R. & Niering, W. (1998) Interpreting sea level rise and rates of vertical marsh accretion in a
southern New England tidal salt marsh. Estuarine, Coastal and Shelf Science 47(4): 419-429.
Roman, C., Peck, J., Allen, J., King, J. & Appleby, P. (1997) Accretion of a New England (USA) salt marsh in response
to inlet migration, storms, and sea-level rise. Estuarine, Coastal and Shelf Science 45(6): 717-727.
Villa, J. A. & Mitsch W. J. (2015) Carbon sequestration in different wetland plant communities of Southwest Florida.
International Journal for Biodiversity Science, Ecosystems Services and Management 11: 17-28
Weston, N. B., Neubauer, S. C., Velinsky, D. J., & Vile, M. A. (2014) Net ecosystem carbon exchange and the
greenhouse gas balance of tidal marshes along an estuarine salinity gradient. Biogeochemistry 120: 163-189.
Land Converted to Wetlands
Bianchi, T. S., Allison, M. A., Zhao, J., Li, X., Comeaux, R. S., Feagin, R. A., & Kulawardhana, R. W. (2013) Historical
reconstruction of mangrove expansion in the Gulf of Mexico: linking climate change with carbon sequestration in
coastal wetlands. Estuarine, Coastal and Shelf Science 119: 7-16.
Byrd, K. B., Ballanti, L. R., Thomas, N. M., Nguyen, D. K., Holmquist, J. R., Simard, M., Windham-Myers, L., Schile, L.
10-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
M., Parker, V. T.,... and Castaneda-Moya, E. (2017) Biomass/Remote Sensing dataset: 30m resolution tidal marsh
biomass samples and remote sensing data for six regions in the conterminous United States: U.S. Geological Survey
data release, .
Byrd, K. B., Ballanti, L., Thomas, N., Nguyen, D., Holmquist, J.R., Simard, M., and Windham-Myers, L. (2018) A
remote sensing-based model of tidal marsh aboveground carbon stocks for the conterminous United States. ISPRS
Journal of Photogrammetry and Remote Sensing 139: 255-271.
Byrd, K. B., Ballanti, L., Thomas, N., Nguyen, D., Holmquist, J.R., Simard, M., and Windham-Myers, L. (2020)
Corrigendum to "A remote sensing-based model of tidal marsh aboveground carbon stocks for the conterminous
United States". ISPRS Journal of Photogrammetry and Remote Sensing 166: 63-67.
Callaway, J. C., Borgnis, E. L., Turner, R. E. & Milan, C. S. (2012a) Carbon sequestration and sediment accretion in
San Francisco Bay tidal wetlands. Estuaries and Coasts 35(5): 1163-1181.
Callaway, J. C., Borgnis, E. L., Turner, R. E., Milan, C. S., Goodfriend, W., & Richmond, S. (2012b). "Wetland
Sediment Accumulation at Corte Madera Marsh and Muzzi Marsh". San Francisco Bay Conservation and
Development Commission.
Church, T. M., Sommerfield, C. K., Velinsky, D. J., Point, D., Benoit, C., Amouroux, D. & Donard, O. F. X. (2006).
Marsh sediments as records of sedimentation, eutrophication and metal pollution in the urban Delaware Estuary.
Marine Chemistry 102(1-2): 72-95.
Craft, C. B., & Richardson, C. J. (1998). Recent and long-term organic soil accretion and nutrient accumulation in
the Everglades. Soil Science Society of America Journal 62(3): 834-843.
Crooks, S., Rybczyk, J., O'Connell, K., Devier, D.L., Poppe, K., Emmett-Mattox, S. (2014) Coastal Blue Carbon
Opportunity Assessment for the Snohomish Estuary: The Climate Benefits of Estuary Restoration. Report by
Environmental Science Associates, Western Washington University, EarthCorps, and Restore America's Estuaries.
Hussein, A. H., Rabenhorst, M. C. & Tucker, M. L. (2004) Modeling of carbon sequestration in coastal marsh soils.
Soil Science Society of America Journal 68(5): 1786-1795.
IPCC (2019). Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Volume 4:
Agriculture, Forestry, and Other Land Use. Calvo Buendia, E., Tanabe K., Kranjc, A., Baasansuren, J., Fukuda, M.,
Ngarize, S., Osako, A., Pyrozhenko, Y., Shermanau, P., & Federici, S. (eds). IPCC, Switzerland.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Prepared by the National Greenhouse
Gas Inventories Programme, H.S.Eggleston, L. Buendia, K. Miwa, T. Ngara & K. Tanabe (eds). IGES, Japan.
IPCC (2014) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.). Published: IPCC,
Switzerland.
IPCC (2003) Good Practice Guidance for Land Use, Land-Use Change and Forestry. LUCF Sector Good Practice
Guidance, Chapter 3. Penman, J., Gytarsky, M., Hiraishi, T., Krug, T., Kruger, D., Pipatti, R., Buendia, L, Miwa, K.,
Ngara, T., Tanabe, K. & F. Wagner (eds). Institute of Global Environmental Strategies (IGES), on behalf of the
Intergovernmental Panel on Climate Change (IPCC): Hayama, Japan.
IPCC (2000) Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories.
Quantifying Uncertainties in Practice, Chapter 6. Penman, J and Kruger, D and Galbally, I and Hiraishi, T and Nyenzi,
B and Emmanuel, S and Buendia, L and Hoppaus, R and Martinsen, T and Meijer, J and Miwa, K and Tanabe, K
(eds). Institute of Global Environmental Strategies (IGES), on behalf of the Intergovernmental Panel on Climate
Change (IPCC): Hayama, Japan.
Kearney, M. S. & Stevenson, J. C. (1991) Island land loss and marsh vertical accretion rate evidence for historical
sea-level changes in Chesapeake Bay. Journal of Coastal Research 7(2): 403-415.
Koster, D., Lichter, J., Lea, P. D., & Nurse, A. (2007). Historical eutrophication in a river-estuary complex in mid-
coast Maine. Ecological Applications 17(3): 765-778.
References 10-83

-------
Lu, M & Megonigal, J.P. (2017) Final Report for RAE Baseline Assessment Project. Memo to Silvestrum Climate
Associates by Smithsonian Environmental Research Center, Maryland.
Lynch, J. C., Sedimentation and nutrient accumulation in mangrove ecosystems of the Gulf of Mexico, M.S. thesis,
Univ. of Southwestern Louisiana, Lafayette, La., 1989.
Marchio, D.A., Savarese, M., Bovard, B., & Mitsch, W.J. (2016) Carbon sequestration and sedimentation in
mangrove swamps influenced by hydrogeomorphic conditions and urbanization in Southwest Florida. Forests 7:
116-135.
McCombs, J.W., Herold, N.D., Burkhalter, S.G. and Robinson C.J., (2016) Accuracy Assessment of NOAA Coastal
Change Analysis Program 2006-2010 Land Cover and Land Cover Change Data. Photogrammetric Engineering &
Remote Sensing. 82:711-718.
Merrill, J. Z. 1999. Tidal Freshwater Marshes as Nutrient Sinks: particulate Nutrient Burial and Denitrification. Ph.D.
Dissertation, University of Maryland, College Park, MD, 342pp.
National Oceanic and Atmospheric Administration, Office for Coastal Management (2020) Coastal Change Analysis
Program (C-CAP) Regional Land Cover. Charleston, SC: NOAA Office for Coastal Management. Accessed October
2020 at .
Noe, G. B., Hupp, C. R., Bernhardt, C. E., & Krauss, K. W. (2016) Contemporary deposition and long-term
accumulation of sediment and nutrients by tidal freshwater forested wetlands impacted by sea level rise. Estuaries
and Coasts 39(4): 1006-1019.
Orson, R. A., Simpson, R. L, & Good, R. E. (1990) Rates of sediment accumulation in a tidal freshwater marsh.
Journal of Sedimentary Research 60(6): 859-869.
Orson, R., Warren, R. & Niering, W. (1998) Interpreting sea level rise and rates of vertical marsh accretion in a
southern New England tidal salt marsh. Estuarine, Coastal and Shelf Science 47(4): 419-429.
Roman, C., Peck, J., Allen, J., King, J. & Appleby, P. (1997) Accretion of a New England (USA) salt marsh in response
to inlet migration, storms, and sea-level rise. Estuarine, Coastal and Shelf Science 45(6): 717-727.
Villa, J. A. & Mitsch W. J. (2015) "Carbon sequestration in different wetland plant communities of Southwest
Florida". International Journal for Biodiversity Science, Ecosystems Services and Management 11: 17-28.
Weston, N. B., Neubauer, S. C., Velinsky, D. J., & Vile, M. A. (2014) Net ecosystem carbon exchange and the
greenhouse gas balance of tidal marshes along an estuarine salinity gradient. Biogeochemistry 120: 163-189.
Settlements Remaining Settlements: Soil Carbon Stock
Changes
Armentano, T. V., and E.S. Menges (1986). Patterns of change in the carbon balance of organic soil-wetlands of the
temperate zone. Journal of Ecology 74: 755-774.
Brady, N.C. and R.R. Weil (1999) The Nature and Properties of Soils. Prentice Hall. Upper Saddle River, NJ, 881.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and J. Wickham. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS 77(9):858-864.
Homer, C., J. Dewitz, J. Fry, M. Coan, N. Hossain, C. Larson, N. Herold, A. McKerrow, J.N. VanDriel and J. Wickham.
(2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing 73(4): 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
10-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing
81(5) :345-354.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
NRCS (1999) Soil Taxonomy: A basic system of soil classification for making and interpreting soil surveys, 2nd
Edition. Agricultural Handbook Number 436, Natural Resources Conservation Service, U.S. Department of
Agriculture, Washington, D.C.
Nusser, S.M. and J.J. Goebel (1997) The national resources inventory: a long-term multi-resource monitoring
programme. Environmental and Ecological Statistics 4:181-204.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997. Global Change Biology
9:1521-1542.
Soil Survey Staff (2011) State Soil Geographic (STATSGO) Database for State. Natural Resources Conservation
Service, United States Department of Agriculture. Available online at:
.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory, Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L, Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies. ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Settlements Remaining Settlements: Changes in Carbon Stocks
in Settlement Trees
deVries, R.E. (1987) A Preliminary Investigation of the Growth and Longevity of Trees in Central Park. M.S. thesis,
Rutgers University, New Brunswick, NJ.
Fleming, LE. (1988) Growth Estimation of Street Trees in Central New Jersey. M.S. thesis, Rutgers University, New
Brunswick, NJ.
Frelich, L.E. (1992) Predicting Dimensional Relationships for Twin Cities Shade Trees. University of Minnesota,
Department of Forest Resources, St. Paul, MN, p. 33.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
MRLC (2013) National Land Cover Database 2001 (NLCD2001). Available online at:
. Accessed August 2013.
Nowak, D.J. (1986) Silvics of an Urban Tree Species: Norway maple (Acer platanoides L). M.S. thesis, College of
Environmental Science and Forestry, State University of New York, Syracuse, NY.
Nowak, D.J. (1994) Atmospheric carbon dioxide reduction by Chicago's urban forest. In: Chicago's Urban Forest
Ecosystem: Results of the Chicago Urban Forest Climate Project. E.G. McPherson, D.J. Nowak, and R.A. Rowntree
(eds.). General Technical Report NE-186. U.S. Department of Agriculture Forest Service, Radnor, PA. pp. 83-94.
Nowak, D.J. (2012) Contrasting natural regeneration and tree planting in 14 North American cities. Urban Forestry
and Urban Greening. 11: 374-382.
References 10-85

-------
Nowak, D.J. and D.E. Crane (2002) Carbon storage and sequestration by urban trees in the United States.
Environmental Pollution 116(3):381-389.
Nowak, D.J. and E. Greenfield (2010) Evaluating the National Land Cover Database tree canopy and impervious
cover estimates across the conterminous United States: A comparison with photo-interpreted estimates.
Environmental Management. 46: 378-390.
Nowak, D.J. and E.J. Greenfield (2018a) U.S. urban forest statistics, values and projections. Journal of Forestry.
116(2):164-177
Nowak, D.J. and E.J. Greenfield (2018b) Declining urban and community tree cover in the United States. Urban
Forestry and Urban Greening. 32:32-55.
Nowak, D.J., D.E. Crane, J.C. Stevens, and M. Ibarra (2002) Brooklyn's Urban Forest. General Technical Report NE-
290. U.S. Department of Agriculture Forest Service, Newtown Square, PA.
Nowak, D.J., R.E. Hoehn, D.E. Crane, J.C. Stevens, J.T. Walton, and J. Bond (2008) A ground-based method of
assessing urban forest structure and ecosystem services. Arboric. Urb. For. 34(6): 347-358.
Nowak, D.J., E.J. Greenfield, R.E. Hoehn, and E. Lapoint (2013) Carbon storage and sequestration by trees in urban
and community areas of the United States." Environmental Pollution 178: 229-236.
Nowak, D.J. A.R. Bodine, R.E. Hoehn, C.B. Edgar, D.R. Hartel, T.W. Lister, T.J. Brandeis (2016) Austin's Urban Forest,
2014. USDA Forest Service, Northern Research Station Resources Bulletin. NRS-100. Newtown Square, PA. 55 p.
Nowak, D.J. A.R. Bodine, R.E. Hoehn, C.B. Edgar, G. Riley, D.R. Hartel, K.J. Dooley, S.M. Stanton, M.A. Hatfield, T.J.
Brandeis, T.W. Lister (2017) Houston's Urban Forest, 2015. USDA Forest Service, Southern Research Station
Resources Bulletin. SRS-211. Newtown Square, PA. 91 p.
Smith, W.B. and S.R. Shifley (1984) Diameter Growth, Survival, and Volume Estimates for Trees in Indiana and
Illinois. Research Paper NC-257. North Central Forest Experiment Station, U.S. Department of Agriculture Forest
Service, St. Paul, MN.
U.S. Department of Interior (2018) National Land Cover Database 2011 (NLCD2011). Accessed online August 16,
2018. Available online at: .
Settlements Remaining Settlements: W20 Emissions from Soils
Brakebill, J.W. and Gronberg, J.M. (2017) County-Level Estimates of Nitrogen and Phosphorus from Commercial
Fertilizer for the Conterminous United States, 1987-2012. U.S. Geological Survey,
.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
Soil Survey Staff (2016) State Soil Geographic (STATSGO) Database for State. Natural Resources Conservation
Service, United States Department of Agriculture. Available online at:
.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory, Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
.
10-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Settlements Remaining Settlements: Changes in Yard
Trimmings and Food Scrap Carbon Stocks in Landfills
Barlaz, M.A. (2008) "Re: Corrections to Previously Published Carbon Storage Factors." Memorandum to Randall
Freed, ICF International. February 28, 2008.
Barlaz, M.A. (2005) "Decomposition of Leaves in Simulated Landfill." Letter report to Randall Freed, ICF Consulting.
June 29, 2005.
Barlaz, M.A. (1998) "Carbon Storage during Biodegradation of Municipal Solid Waste Components in Laboratory-
Scale Landfills." Global Biogeochemical Cycles 12:373-380.
De la Cruz, F.B. and M.A. Barlaz (2010) "Estimation of Waste Component Specific Landfill Decay Rates Using
Laboratory-Scale Decomposition Data" Environmental Science & Technology 44:4722- 4728.
Eleazer, W.E., W.S. Odle, Y. Wang, and M.A. Barlaz (1997) "Biodegradability of Municipal Solid Waste Components
in Laboratory-Scale Landfills." Environmental Science & Technology 31:911-917.
EPA (2019) Advancing Sustainable Materials Management: Facts and Figures 2017. U.S. Environmental Protection
Agency, Office of Solid Waste and Emergency Response, Washington, D.C. Available online at
.
EPA (2016) Municipal Solid Waste Generation, Recycling, and Disposal in the United States: Facts and Figures. U.S.
Environmental Protection Agency, Office of Solid Waste and Emergency Response, Washington, D.C. Available
online at .
EPA (1995) Compilation of Air Pollutant Emission Factors. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, NC. AP-42 Fifth Edition. Available online at
.
EPA (1991) Characterization of Municipal Solid Waste in the United States: 1990 Update. U.S. Environmental
Protection Agency, Office of Solid Waste and Emergency Response, Washington, D.C. EPA/530-SW-90-042.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
IPCC (2003) Good Practice Guidance for Land Use, Land-Use Change, and Forestry. The Intergovernmental Panel on
Climate Change, National Greenhouse Gas Inventories Programme, J. Penman et al. (eds.). Available online at
.
Oshins, C. and D. Block (2000) "Feedstock Composition at Composting Sites." Biocycle 41(9):31-34.
Tchobanoglous, G., H. Theisen, and S.A. Vigil (1993) Integrated Solid Waste Management, 1st edition. McGraw-Hill,
NY. Cited by Barlaz (1998) "Carbon Storage during Biodegradation of Municipal Solid Waste Components in
Laboratory-Scale Landfills." Global Biogeochemical Cycles 12:373-380.
Land Converted to Settlements
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In R.N.
Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer. Domke, G.M.,
Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter carbon stocks in
forests of the United States. Science of the Total Environment 557-558: 469-478.
References 10-87

-------
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in standing
dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon Balance and
Management. 6:14.
Domke, G.M., Woodall, C.W., Walters, B.F., Smith, J.E. (2013) From models to measurements: comparing down
dead wood carbon stock estimates in the U.S. forest inventory. PLoS ONE 8(3): e59949.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter
carbon stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and
Wickham, J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L, Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v.
81, no. 5, p. 345-354.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change, H.S. Eggleston, L. Buendia, K. Miwa, T
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
Jenkins, J.C., D.C. Chojnacky, L.S. Heath, and R.A. Birdsey (2003) "National-scale biomass estimators for United
States tree species." Forest Science 49(l):12-35.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
Ogle, S.M., F.J. Breidt, and K. Paustian (2006) "Bias and variance in model results due to spatial scaling of
measurements for parameterization in regional assessments." Global Change Biology 12:516-523.
Schimel, D.S. (1995) "Terrestrial ecosystems and the carbon cycle." Global Change Biology 1: 77-91.
Smith, J.E.; Heath, L.S.; Skog, K.E.; Birdsey, R.A. (2006) Methods for calculating forest ecosystem and harvested
carbon with standard estimates for forest types of the United States. Gen. Tech. Rep. NE-343. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 216 p.
Tubiello, F. N., et al. (2015). "The Contribution of Agriculture, Forestry and other Land Use activities to Global
Warming, 1990-2012." Global Change Biology 21:2655-2660.
USDA Forest Service (2020) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 10 October 2020.
USDA-NRCS (2018) Summary Report: 2015 National Resources Inventory, Natural Resources Conservation Service,
Washington, D.C., and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.

-------
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols. (2011) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Yang, L, Jin, S., Danielson, P., Homer, C., Gass, L, Bender, S. M., Case, A., Costello, C., Dewitz, J., Fry, J., Funk, M.,
Granneman, B., Liknes, G. C., Rigge, M. & Xian, G. (2018) A new generation of the United States National Land
Cover Database: Requirements, research priorities, design, and implementation strategies. ISPRS Journal of
Photogrammetry and Remote Sensing 146: 108-123.
Waste
Landfills
40 CFR Part 60, Subpart WWW (2005) Standards of Performance for Municipal Solid Waste Landfills, 60.750-
60.759, Code of Federal Regulations, Title 40. Available online at:
.
40 CFR Part 258, Subtitle D of RCRA (2012) Criteria for Municipal Solid Waste Landfills, 258.1—258.75, Code of
Federal Regulations, Title 40. Available online at: .
BioCycle (2010) "The State of Garbage in America" By L. Arsova, R. Van Haaren, N. Goldstein, S. Kaufman, and N.
Themelis. BioCycle. December 2010. Available online at: .
BioCycle (2006) "The State of Garbage in America" By N. Goldstein, S. Kaufman, N. Themelis, and J. Thompson Jr.
BioCycle. April 2006. Available online at: .
Bronstein, K., Coburn, J., and R. Schmeltz (2012) "Understanding the EPA's Inventory of U.S. Greenhouse Gas
Emissions and Sinks and Mandatory GHG Reporting Program for Landfills: Methodologies, Uncertainties,
Improvements and Deferrals." Prepared for the U.S. EPA International Emissions Inventory Conference, August
2012, Tampa, Florida. Available online at:
.
Business for Social Responsibility (BSR) (2014). Analysis of U.S. Food Waste Among Food Manufacturers, Retailers,
and Restaurants. Available online at: .
BSR (2013) Analysis of U.S. Food Waste Among Food Manufacturers, Retailers, and Restaurants. Available online
at: 
Czepiel, P., B. Mosher, P. Crill, and R. Harriss (1996) "Quantifying the Effect of Oxidation on Landfill Methane
Emissions." Journal of Geophysical Research, 101(Dll):16721-16730.Dou, Z.; Ferguson, J. D.; Galligan, D. T.; Kelly,
A. M.; Finn, S. T.; Giegengack, R. (2016) "Assessing U.S. food wastage and opportunities for reduction. Global Food
Security Volume 8, March 2016, Pages 19-26. https://doi.Org/10.1016/j.gfs.2016.02.001.
EIA (2007) Voluntary Greenhouse Gas Reports for EIA Form 1605B (Reporting Year 2006). Available online at:
.
EPA (2020a) Landfill Methane Outreach Program (LMOP). 2020 Landfill and Project Level Data. August 2020.
Available online at: .
References 10-89

-------
EPA (2020b) Greenhouse Gas Reporting Program (GHGRP). 2020 Amazon S3 Data. Subpart HH: Municipal Solid
Waste Landfills and Subpart TT: Industrial Waste Landfills. Accessed on October 1, 2020.
EPA (2020c) Wasted Food Measurement Methodology Scoping Memo. July 2020. Available online at <
https://www.epa.gov/sites/production/files/2020-
06/documents/food_measurement_methodology_scoping_memo-6-18-20.pdf>.
EPA (2020d) Advancing Sustainable Materials Management: Facts and Figures 2018. December 2020. Available
online at: .
EPA (2019a) Landfill Methane Outreach Program (LMOP). 2019 Landfill and Project Level Data. September 2019.
Available online at: .
EPA (2019b) Greenhouse Gas Reporting Program (GHGRP). 2019 Amazon S3 Data. Subpart HH: Municipal Solid
Waste Landfills and Subpart TT: Industrial Waste Landfills.
EPA (2019c) Advancing Sustainable Materials Management: Facts and Figures 2016 and 2017. November 2019.
Available online at: .
EPA (2018) Advancing Sustainable Materials Management: Facts and Figures 2015. July 2018. Available online at: <
https://www.epa.gov/sites/production/files/2018-
07/documents/smm_2015_tables_and_figures_07252018_fnl_508_0.pdf>.
EPA (2016a) Industrial and Construction and Demolition Landfills. Available online at:
https://www.epa.gov/landfills/industrial-and-construction-and-demolition-cd-landfills.
EPA (2016b) Advancing Sustainable Materials Management: Facts and Figures 2014. December 2016. Available
online at: .
EPA (2014) Advancing Sustainable Materials Management: Facts and Figures 2014. February 2014. Available online
at: .
EPA (2008) Compilation of Air Pollution Emission Factors, Publication AP-42, Draft Section 2.4 Municipal Solid
Waste Landfills. October 2008.
EPA (1993) Anthropogenic Methane Emissions in the United States, Estimates for 1990: Report to Congress, U.S.
Environmental Protection Agency, Office of Air and Radiation. Washington, D.C. EPA/430-R-93-003. April 1993.
EPA (1988) National Survey of Solid Waste (Municipal) Landfill Facilities, U.S. Environmental Protection Agency.
Washington, D.C. EPA/530-SW-88-011. September 1988.
EREF (The Environmental Research & Education Foundation) (2016). Municipal Solid Waste Management in the
United States: 2010 & 2013.
ERG (2020) Production Data Supplied by ERG for 1990-2018 for Pulp and Paper, Fruits and Vegetables, and Meat.
June 5, 2020.
Food Waste Reduction Alliance (FWRA) (2016). Analysis of U.S. Food Waste Among Food Manufacturers, Retailers,
and Restaurants. Available online at: .
Intergovernmental Panel on Climate Change (IPCC) (2019) 2019 Refinement to the 2006 IPCC Guidelines for
National Greenhouse Gas Inventories. Available online at .
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
10-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Mancinelli, R. and C. McKay (1985) "Methane-Oxidizing Bacteria in Sanitary Landfills." Proc. First Symposium on
Biotechnical Advances in Processing Municipal Wastes for Fuels and Chemicals, Minneapolis, MN, 437-450. August.
RTI (2018a) Methodological changes to the scale-up factor used to estimate emissions from municipal solid waste
landfills in the Inventory. Memorandum prepared by K. Bronstein and M. McGrath for R. Schmeltz (EPA). March 22,
2018.
RTI (2018b) Comparison of industrial waste data reported under Subpart TT and the Solid Waste chapter of the
GHG Inventory. Memorandum prepared by K. Bronstein, B. Jackson, and M. McGrath for R, Schmeltz (EPA).
October 12, 2018.
RTI (2017) Methodological changes to the methane emissions from municipal solid waste landfills as reflected in
the public review draft of the 1990-2015 Inventory. Memorandum prepared by K. Bronstein and M. McGrath for R.
Schmeltz (EPA). March 31, 2017.
RTI (2011) Updated Research on Methane Oxidation in Landfills. Memorandum prepared by K. Weitz (RTI) for R.
Schmeltz (EPA). January 14, 2011.
Waste Business Journal (WBJ) (2016) Directory of Waste Processing & Disposal Sites 2016.
WBJ (2010) Directory of Waste Processing & Disposal Sites 2010.
WTO (2017). "China's import ban on solid waste queried at import licensing meeting". World Trade Organization.
Published October 3, 2017. Available online at:
.
Wastewater Treatment and Discharge
AF&PA (2018) "2018 AF&PA Sustainability Report: Advancing U.S. Paper and Wood Products Industry Sustainability
Performance." American Forest & Paper Association. Available online at:  Accessed July 2019.
AF&PA (2016) "2016 AF&PA Sustainability Report: Advancing U.S. Paper and Wood Products Industry Sustainability
Performance." American Forest & Paper Association. Available online at:  Accessed May 2017.
AF&PA (2014) "2014 AF&PA Sustainability Report." American Forest & Paper Association. Available online at:
. Accessed
June 2017.
Beecher et al. (2007) "A National Biosolids Regulation, Quality, End Use & Disposal Survey, Preliminary Report."
Northeast Biosolids and Residuals Association, April 14, 2007.
Beer Institute (2011) Brewers Almanac. Available online at: .
Benyahia, F., M. Abdulkarim, A. Embaby, and M. Rao. (2006) Refinery Wastewater Treatment: A true Technological
Challenge. Presented at the Seventh Annual U.A.E. University Research Conference.
BIER (2017) Beverage Industry Environmental Roundtable. 2016 Trends and Observations. Available online at:
. Accessed April 2018.
Brewers Association (2020) Statistics: Number of Breweries. Available online at:
. Accessed May 2020.
Brewers Association (2017). 2016 Sustainability Benchmarking Update. Available online at:
. Accessed
April 2018.
References 10-91

-------
Brewers Association (2016a) 2015 Sustainability Benchmarking Report. Available online at:
. Accessed
March 2018.
Brewers Association (2016b) Wastewater Management Guidance Manual. Available online at:
.
Accessed September 2017.
Cabrera (2017) "Pulp Mill Wastewater: Characteristics and Treatment." Biological Wastewater Treatment and
Resource Recovery. InTech. pp. 119-139.
CAST (1995) Council for Agricultural Science and Technology. Waste Management and Utilization in Food
Production and Processing. U.S.A. October 1995. ISBN 1-887383-02-6. Available online at: .
Climate Action Reserve (CAR) (2011) Landfill Project Protocol V4.0, June 2011. Available online at:
.
Cooper (2018) Email correspondence. Geoff Cooper, Renewable Fuels Association to Kara Edquist, ERG. "Wet Mill
vs. Dry Mill Ethanol Production." May 18, 2018.
DOE (2013) U.S. Department of Energy Bioenergy Technologies Office. Biofuels Basics. Available online at:
. Accessed September 2013.
Donovan (1996) Siting an Ethanol Plant in the Northeast. C.T. Donovan Associates, Inc. Report presented to
Northeast Regional Biomass Program (NRBP). (April). Available online at: .
Accessed October 2006.
EIA (2020) Energy Information Administration. U.S. Refinery and Blender Net Production of Crude Oil and
Petroleum Products (Thousand Barrels). Available online at:
. Accessed May 2020.
EPA (2019) Preliminary Effluent Guidelines Program Plan 14. EPA-821-R-19-005. Office of Water, U.S.
Environmental Protection Agency. Washington, DC. October 2019. Available online at:
. Accessed
July 2020.
EPA (2013) U.S. Environmental Protection Agency. Report on the Performance of Secondary Treatment
Technology. EPA-821-R-13-001. Office of Water, U.S. Environmental Protection Agency. Washington, D.C. March
2013. Available online at: .
EPA (2012) U.S. Environmental Protection Agency. Clean Watersheds Needs Survey 2012 - Report to Congress. U.S.
Environmental Protection Agency, Office of Wastewater Management. Washington, D.C. Available online at:
. Accessed
February 2016.
EPA (2008) U.S. Environmental Protection Agency. Clean Watersheds Needs Survey 2008 - Report to Congress. U.S.
Environmental Protection Agency, Office of Wastewater Management. Washington, D.C. Available online at:
. Accessed December
2015.
EPA (2004a) U.S. Environmental Protection Agency. Clean Watersheds Needs Survey 2004 - Report to Congress.
U.S. Environmental Protection Agency, Office of Wastewater Management. Washington, D.C.
EPA (2004b) Technical Development Document for the Final Effluent Limitations Guidelines and Standards for the
Meat and Poultry Products Point Source Category (40 CFR 432). Office of Water. EPA-821-R-04-011, Washington
DC, July.
10-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
EPA (2002) U.S. Environmental Protection Agency. Development Document for the Proposed Effluent Limitations
Guidelines and Standards for the Meat and Poultry Products Industry Point Source Category (40 CFR 432). EPA-
821-B-01-007. Office of Water, U.S. Environmental Protection Agency. Washington, D.C. January 2002.
EPA (2000) U.S. Environmental Protection Agency. Clean Watersheds Needs Survey 2000 - Report to Congress.
Office of Wastewater Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
. Accessed July 2007.
EPA (1999) U.S. Environmental Protection Agency. Biosolids Generation, Use and Disposal in the United States.
Office of Solid Waste and Emergency Response, U.S. Environmental Protection Agency. Washington, D.C. EPA530-
R-99-009. September 1999.
EPA (1998) U.S. Environmental Protection Agency. "AP-42 Compilation of Air Pollutant Emission Factors." Chapter
2.4, Table 2.4-3, page 2.4-13. Available online at: .
EPA (1997a) U.S. Environmental Protection Agency. Estimates of Global Greenhouse Gas Emissions from Industrial
and Domestic Wastewater Treatment. EPA-600/R-97-091. Office of Policy, Planning, and Evaluation, U.S.
Environmental Protection Agency. Washington, D.C. September 1997.
EPA (1997b) U.S. Environmental Protection Agency. Supplemental Technical Development Document for Effluent
Guidelines and Standards (Subparts B & E). EPA-821-R-97-011. Office of Water, U.S. Environmental Protection
Agency. Washington, D.C. October 1997.
EPA (1996) U.S. Environmental Protection Agency. 1996 Clean Water Needs Survey Report to Congress.
Assessment of Needs for Publicly Owned Wastewater Treatment Facilities, Correction of Combined Sewer
Overflows, and Management of Storm Water and Nonpoint Source Pollution in the United States. Office of
Wastewater Management, U.S. Environmental Protection Agency. Washington, D.C.
EPA (1993a) U.S. Environmental Protection Agency, "Anthropogenic Methane Emissions in the U.S.: Estimates for
1990, Report to Congress." Office of Air and Radiation, Washington, DC. April 1993.
EPA (1993b) U.S. Environmental Protection Agency. Development Document for the Proposed Effluent Limitations
Guidelines and Standards for the Pulp, Paper and Paperboard Point Source Category. EPA-821-R-93-019. Office of
Water, U.S. Environmental Protection Agency. Washington, D.C. October 1993.
EPA (1993c) Standards for the Use and Disposal of Sewage Sludge. 40 CFR Part 503.
EPA (1992) U.S. Environmental Protection Agency. Clean Watersheds Needs Survey 1992 - Report to Congress.
Office of Wastewater Management, U.S. Environmental Protection Agency. Washington, D.C.
EPA (1982) U.S. Environmental Protection Agency. Development Document for Effluent Limitations Guidelines and
standards for the Petroleum Refining. United States Environmental Protection Agency, Office of Water. EPA-440/1-
82-014. Washington D.C. October 1982.
EPA (1975) U.S. Environmental Protection Agency. Development Document for Interim Final and Proposed Effluent
Limitations Guidelines and New Source Performance Standards for the Fruits, Vegetables, and Specialties Segment
of the Canned and Preserved Fruits and Vegetables Point Source Category. United States Environmental Protection
Agency, Office of Water. EPA-440/1-75-046. Washington D.C. October 1975.
EPA (1974) U.S. Environmental Protection Agency. Development Document for Effluent Limitations Guidelines and
New Source Performance Standards for the Apple, Citrus, and Potato Processing Segment of the Canned and
Preserved Fruits and Vegetables Point Source Category. Office of Water, U.S. Environmental Protection Agency,
Washington, D.C. EPA-440/l-74-027-a. March 1974.
ERG (2021) Revised Memorandum: Improvements to the 1990-2019 Wastewater Treatment and Discharge
Greenhouse Gas Inventory. March 2021.
ERG (2018a) Memorandum: Updates to Domestic Wastewater BOD Generation per Capita. August 2018.
ERG (2018b) Memorandum: Inclusion of Wastewater Treatment Emissions from Breweries. July 2018.
References 10-93

-------
ERG (2016) Revised Memorandum: Recommended Improvements to the 1990-2015 Wastewater Greenhouse Gas
Inventory. November 2016.
ERG (2013a) Memorandum: Revisions to Pulp and Paper Wastewater Inventory. October 2013.
ERG (2013b) Memorandum: Revisions to the Petroleum Refinery Wastewater Inventory. October 2013.
ERG (2008a) Memorandum: Planned Revisions of the Industrial Wastewater Inventory Emission Estimates for the
1990-2007 Inventory. 10 August 2008.
ERG (2008b) Memorandum: Estimation of Onsite Treatment at Industrial Facilities and Review of Wastewater
Characterization Data. 15 April 2008.
ERG (2006a) Memorandum: Recommended Improvements to EPA's Wastewater Inventory for Industrial
Wastewater. Prepared for Melissa Weitz, EPA. 11 August 2006.
ERG (2006b) Memorandum: Assessment of Greenhouse Gas Emissions from Wastewater Treatment of U.S. Ethanol
Production Wastewaters. Prepared for Melissa Weitz, EPA. 10 October 2006.
FAO (2020a) FAOSTAT-Forestry Database. Available online at:
. Accessed April 2020.
FAO (2020b) "Pulp and Paper Capacities Report." United States. From 1998 - 2003, 2000 - 2005, 2001 - 2006,
2002 - 2007, 2003 - 2008, 2010 - 2015, 2011 - 2016, 2012 - 2017, 2013 - 2018, 2014 - 2019, 2015 - 2020, 2016 -
2021, 2017 - 2022 reports. Available online at: . Accessed April
2020.
FAO (2020c) FAOSTAT-Food Balance Sheets. Available online at:
. Accessed May 2020.
Foley et al. (2015) N20 and CH4 Emission from Wastewater Collection and Treatment Systems: State of the Science
Report and Technical Report. GWRC Report Series. IWA Publishing, London, UK.
Great Lakes-Upper Mississippi River Board of State and Provincial Public Health and Environmental Managers.
(2004) Recommended Standards for Wastewater Facilities (Ten-State Standards).
Guisasola et al. (2008) Methane formation in sewer systems. Water Research 42(6-7): 1421-1430.
IPCC (2019) 2019 Refinement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National
Greenhouse Gas Inventories Programme, The Intergovernmental Panel on Climate Change. [CalvoBuendia, E.,
Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize S., Osako, A., Pyrozhenko, Y., Shermanau, P. and
Federici, S. (eds)]. Switzerland.
IPCC (2014) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
[Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.)]. Published:
IPCC, Switzerland.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Kenari et. al (2010). An Investigation on the Nitrogen Content of a Petroleum Refinery Wastewater and its Removal
by Biological Treatment. Journal of Environmental Health, Sciences, and Engineering. 7(1): 391-394.Leverenz, H.L.,
G. Tchobanoglous, and J.L. Darby (2010) "Evaluation of Greenhouse Gas Emissions from Septic Systems." Water
Environment Research Foundation. Alexandria, VA.
Malmberg, B. (2018) Draft Pulp and Paper Information for Revision of EPA Inventory of U.S. Greenhouse Gas
Emissions and Sinks, Waste Chapter. National Council for Air and Stream Improvement, Inc. Prepared for Rachel
Schmeltz, EPA. June 13, 2018.
McFarland (2001) Biosolids Engineering, New York: McGraw-Hill, p. 2.12.
10-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
Merrick (1998) Wastewater Treatment Options for the Biomass-to-Ethanol Process. Report presented to National
Renewable Energy Laboratory (NREL). Merrick & Company. Subcontract No. AXE-8-18020-01. October 22,1998.
Metcalf & Eddy, Inc. (2014) Wastewater Engineering: Treatment and Resource Recovery, 5th ed. McGraw Hill
Publishing.
Metcalf & Eddy, Inc. (2003) Wastewater Engineering: Treatment, Disposal and Reuse, 4th ed. McGraw Hill
Publishing.
Nemerow, N.L. and A. Dasgupta (1991) Industrial and Hazardous Waste Treatment. Van Nostrand Reinhold. NY.
ISBN 0-442-31934-7.
NRBP (2001) Northeast Regional Biomass Program. An Ethanol Production Guidebook for Northeast States.
Washington, D.C. (May 3). Available online at: . Accessed October 2006.
Rendleman, C.M. and Shapouri, H. (2007) New Technologies in Ethanol Production. USDA Agricultural Economic
Report Number 842.
RFA (2020a). Renewable Fuels Association. Annual U.S. Fuel Ethanol Production. Available online at:
. Accessed May 2020.
RFA (2020b). Renewable Fuels Association. Monthly Grain Use for U.S. Ethanol Production Report. Available online
at: . Accessed May 2020.
Ruocco (2006a) Email correspondence. Dr. Joe Ruocco, Phoenix Bio-Systems to Sarah Holman, ERG. "Capacity of
Bio-Methanators (Dry Milling)." October 6, 2006.
Ruocco (2006b) Email correspondence. Dr. Joe Ruocco, Phoenix Bio-Systems to Sarah Holman, ERG. "Capacity of
Bio-Methanators (Wet Milling)." October 16, 2006.
Short et al. (2017) Dissolved Methane in the Influent of Three Australian Wastewater Treatment Plants Fed by
Gravity Sewers. Sci Total Environ 599-600: 85-93.
Short et al. (2014) Municipal Gravity Sewers: an Unrecognised Source of Nitrous Oxide. Sci Total Environ 468-469:
211-218.
Stier, J. (2018) Personal communications between John Stier, Brewers Association Sustainability Mentor and Amie
Aguiar, ERG. Multiple dates.
Sullivan (SCS Engineers) (2010) The Importance of Landfill Gas Capture and Utilization in the U.S. Presented to
SWICS, April 6, 2010. Available online at:
.
Sullivan (SCS Engineers) (2007) Current MSW Industry Position and State of the Practice on Methane Destruction
Efficiency in Flares, Turbines, and Engines. Presented to Solid Waste Industry for Climate Solutions (SWICS). July
2007. Available online at:
.
TTB (2020) Alcohol and Tobacco Tax and Trade Bureau. Beer Statistics. Available online at:
. Accessed May 2020.
UNFCCC (2012) CDM Methodological tool, Project emissions from flaring (Version 02.0.0). EB 68 Report. Annex 15.
Available online at: .
U.S. Census Bureau (2020) International Database. Available online at:
. Accessed May 2020.
U.S. Census Bureau (2017) "American Housing Survey." Table 1A-4: Selected Equipment and Plumbing-All Housing
Units. From 1989, 1991,1993, 1995, 1997, 1999, 2001, 2003, 2005, 2007, and 2009 reports. Table C-04-AO
Plumbing, Water, and Sewage Disposal-All Occupied Units. From 2011, 2013, 2015, and 2017 reports. Available
online at . Accessed May 2020.
References 10-95

-------
U.S. Census Bureau (2013) "American Housing Survey." Table 1A-4: Selected Equipment and Plumbing-All Housing
Units. From 1989, 1991,1993, 1995, 1997, 1999, 2001, 2003, 2005, 2007, and 2009 reports. Table C-04-AO
Plumbing, Water, and Sewage Disposal-All Occupied Units. From 2011, and 2013 reports. Available online at
. Accessed May 2020.
USDA (2020a) U.S. Department of Agriculture. National Agricultural Statistics Service. Washington, D.C. Available
online at:  and
. Accessed May 2020.
USDA (2020b) U.S. Department of Agriculture. Economic Research Service. Nutrient Availability. Washington D.C.
Available online at: . Accessed May 2020.
USDA (2020c) U.S. Department of Agriculture. National Agricultural Statistics Service. Vegetables 2019 Summary.
Available online at: . Accessed April
2020.
U.S. Poultry (2006) Email correspondence. John Starkey, USPOULTRY to D. Bartram, ERG. 30 August 2006.
White and Johnson (2003) White, P.J. and Johnson, L.A. Editors. Corn: Chemistry and Technology. 2nd ed. AACC
Monograph Series. American Association of Cereal Chemists. St. Paul, MN.
World Bank (1999) Pollution Prevention and Abatement Handbook 1998, Toward Cleaner Production. The
International Bank for Reconstruction and Development/The WORLDBANK. 1818 H Street, N.W. Washington, DC.
20433, USA. ISBN 0-8213-3638-X.
Composting
BioCycle (2017) The State of Organics Recycling in the U.S. Prepared by Nora Goldstein. Available online at:

Cornell Composting (1996). Monitoring Compost Moisture. Cornell Waste Management Institute. Available online
at: .
Cornell Waste Management Institute (2007) The Science of Composting. Available online at:
.
EPA (2020) Advancing Sustainable Materials Management: Facts and Figures 2018. November 2020. Available
online at: .
EPA (2019) Advancing Sustainable Materials Management: 2016 and 2017 Tables and Figures. Office of Solid Waste
and Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2018) Advancing Sustainable Materials Management: 2015 Tables and Figures. Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
Harvard Law School and Center for EcoTechnology (CET) (2019) Bans and Beyond: Designing and Implementing
Organic Waste Bans and Mandatory Organics Recycling Laws. Prepared by Katie Sandson and Emily Broad Leib,
Harvard Law School Food Law and Policy Clinic, with input from Lorenzo Macaluso and Coryanne Mansell, Center
for EcoTechnology (CET). Available online at: .
10-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019

-------
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Volume 5: Waste, Chapter 4: Biological
Treatment of Solid Waste, Table 4.1. The National Greenhouse Gas Inventories Programme, The
Intergovernmental Panel on Climate Change, H.S. Eggleston, L Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.).
Hayama, Kanagawa, Japan. Available online at: .
Northeast Recycling Council (NERC) (2020) Disposal Bans & Mandatory Recycling in the United States. Available
online at: < https://nerc.org/documents/disposal_bans_mandatory_recycling_united_states.pdf>.
University of Maine (2016). Compost Report Interpretation Guide. Soil Testing Lab. Available online at:
.
U.S. Census Bureau (2019) Table 1. Annual Estimates of the Resident Population for the United States, Regions,
States, and Puerto Rico: April 1, 2010 to July 1, 2019 (NST-EST2019-01). Available online at:

U.S. Composting Council (2010) Yard Trimmings Bans: Impact and Support. Prepared by Stuart Buckner, Executive
Director, U.S., Composting Council. Available online at: .
Stand-Alone Anaerobi jestion
IPCC (2006). 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Volume 5: Waste, Chapter 4: Biological
Treatment of Solid Waste, Table 4.1. The National Greenhouse Gas Inventories Programme, The Intergovernmental
Panel on Climate Change, H.S. Eggleston, L Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.). Hayama, Kanagawa,
Japan. Available online at .
EPA (2020). Types of Anaerobic Digesters: Common Ways to Describe Digesters. Available online at
.
EPA (2019). Anaerobic Digestion Facilities Processing Food Waste in the United States in 2016: Survey Results.
September 2019 EPA/903/S-19/001. Available online at .
EPA (2018). Anaerobic Digestion Facilities Processing Food Waste in the United States in 2015: Survey Results. May
2018 EPA/903/S-18/001. Available online at .
EPA (2016). Frequently Asked Questions About Anaerobic Digestion. Available online at
.
EPA (1993). Anthropogenic Methane Emissions in the U.S.: Estimates for 1990, Report to Congress. Office of Air
and Radiation, Washington, DC. April 1993.
Water Environment Federation (WEF) (2012). What Every Operator Should Know about Anaerobic Digestion.
Available online at .
Waste Incineration
RTI (2009) Updated Hospital/Medical/lnfectious Waste Incinerator (HMIWI) Inventory Database. Memorandum
dated July 6, 2009. Available online at: .
References 10-97

-------
Waste Sources of Prec r Greenhouse Gas Emissions
EPA (2020) "Criteria pollutants National Tier 1 for 1970 - 2019." National Emissions Inventory (NEI) Air Pollutant
Emissions Trends Data. Office of Air Quality Planning and Standards, April 2020. Available online at:
.
EPA (2003) Email correspondence containing preliminary ambient air pollutant data. Office of Air Pollution and the
Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. December 22, 2003.
Recalculations and Improvements
ArSova, Ljupka, Rob van Haaren, Nora Goldstein, Scott M. Kaufman, and Nickolas J. Themelis (2008) "16th Annual
BioCycle Nationwide Survey: The State of Garbage in America" BioCycle, JG Press, Emmaus, PA. December.
EIA (2020) Monthly Energy Review, November 2020. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035 (2020/11).
EIA (2019) Personal communication between EIA and ICF on November 11,
2019.11/documents/2016_and_2017_facts_and_figures_data_tables_0.pdf>.
EPA (2019Motor Vehicle Emissions Simulator (MOVES). Office of Transportation and Air Quality, U.S.
Environmental Protection Agency. Available online at: .
EPA (2018a) Advancing Sustainable Materials Management: 2015 Data Tables. Office of Land and Emergency
Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2016) Advancing Sustainable Materials Management: 2014 Fact Sheet. Office of Land and Emergency
Management, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2015) Advancing Sustainable Materials Management: Facts and Figures 2013 - Assessing Trends in Material
Generation, Recycling and Disposal in the United States. Office of Solid Waste and Emergency Response, U.S.
Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2007, 2008, 2011, 2013, 2014) Municipal Solid Waste in the United States: Facts and Figures. Office of Solid
Waste and Emergency Response, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
ERG (2020) Improvements to the 1990-2018 Wastewater Treatment and Discharge Greenhouse Gas Inventory. July
2020.
GTI (2019) Classification of Methane Emissions from Industrial Meters, Vintage vs Modern Plastic Pipe, and Plastic-
lined Steel and Cast-Iron Pipe. June 2019. Gas Technology Institute and U.S. Department of Energy GTI Project
Number 22070. DOE project Number ED-FE0029061.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Ogle, S.M., F.J. Breidt, and K. Paustian. (2006) "Bias and variance in model results due to spatial scaling of
measurements for parameterization in regional assessments." Global Change Biology 12:516-523.
RMA (2018) "2017 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association, Washington, DC.
July 2018. Available online at: 
-------
Schneider, S. (2007) E-mail between Shelly Schneider of Franklin Associates (a division of ERG) and Sarah Shapiro of
ICF International, January 10, 2007.
STATSG02 (2016) Soil Survey Staff, Natural Resources Conservation Service, United States Department of
Agriculture. U.S. General Soil Map (STATSG02). Available online at .
Accessed 10 November 2016.
References 10-99

-------