1990-2018
United States
Environmental Protection
M »Agency
EPA 430-R-20-002
Inventory of
U.S. Greenhouse Gas
Emissions and Sinks

-------
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 2018, inclusive, will be made available for the
final report published on April 13, 2020 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, work on emissions from fuel combustion was led by Vincent
Camobreco. Sarah Roberts and Justine Geidosch directed the work on mobile combustion and transportation.
Work on fugitive methane emissions from the Energy sector was directed by Melissa Weitz and Chris Sherry.
Calculations for the Waste sector were led by Rachel Schmeltz. Tom Wirth directed work on the Agriculture and
the Land Use, Land-Use Change, and Forestry chapters, with support from John Steller. Work on Industrial
Processes and Product Use (IPPU) CO2, Cm, and N2O emissions was directed by John Steller and Vincent
Camobreco, with support from Chris Sherry and Amanda Chiu. Work on 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.
Other EPA offices also contributed data, analysis, and technical review for this report. 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. 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 and Maryalice Locke) 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, Mike
Nichols, and John Coulston) for compiling the inventories for CO2, CH4, and N2O fluxes associated with forest land.
We thank the Department of Agriculture's Agricultural Research Service (Stephen Del Grosso) and the Natural
Resource Ecology Laboratory at Colorado State University (Stephen Ogle, Keith Paustian, Bill Parton, F. Jay Breidt,
Shannon Spencer, Kendrick Killian, Ram Gurung, Ernie Marx, Stephen Williams, Cody Alsaker, Guhan
Dheenadayalan Sivakami, Amy Swan, and Chris Dorich) for compiling the inventories for CH4 emissions, N2O
emissions, and CO2 fluxes associated with soils in croplands, grasslands, and settlements.

-------
We thank Silvestrum Climate Associates (Stephen Crooks, Lisa Schile Beers, Christine May), National Oceanic and
Atmospheric Administration (Nate Herold, Ariana Sutton-Grier, Meredith Muth), the Smithsonian Environmental
Research Center (J. Patrick Megonigal, Blanca Bernal, James Holmquist, Meng Lu) and Florida International
University (Tiffany Troxler) and members of the U.S. Coastal Wetland Carbon Working Group for compiling
inventories of land use change, soil carbon stocks and stock change, Cm emissions, and N2O 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, Tommy
Hendrickson, Rebecca Ferenchiak, Kasey Knoell, Sabrina Andrews, Rani Murali, Cara Blumenthal, Louise Bruning,
Emily Peterson, Lou Browning, Helena Caswell, Katie O'Malley, Howard Marano, Neha Vaingankar, Megha Kedia,
Grace Tamble, Logan Pfeiffer, Mary Francis McDaniel, Mollie Carroll, Tyler Brewer, Madeleine Pearce, Carolyn
Pugh, and Claire Trevisan for synthesizing this report and preparing many of the individual analyses.
We thank Eastern Research Group for their significant analytical support. Deborah Bartram, Kara Edquist, Tara
Stout, and Amie Aguiar support the development of emissions estimates for wastewater. Cortney Itle, Amie Aguiar,
Kara Edquist, Amber Allen, 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, Casey Pickering,
Marty Wolf, Colin Peirce, and Aylin Sertkaya develop estimates for Natural Gas and Petroleum Systems. Cortney
Itle, Gopi Manne, Tara Stout, and Stephen Treimel support the development of emission estimates for coal mine
methane.
Finally, we thank the following teams for their significant analytical support: RTI International (Kate Bronstein,
Meaghan McGrath, Jeff Coburn, Keith Weitz, Michael Laney, Carson Moss, David Randall, Gabrielle Raymond,
Jason Goldsmith, Karen Schaffner, Melissa Icenhour); Raven Ridge Resources (James Marshall and Raymond
Pilcher).

-------
Preface
The United States Environmental Protection Agency (EPA) prepares the official U.S. Inventory of Greenhouse Gas
Emissions and Sinks to comply with existing commitments under the United Nations Framework Convention on
Climate Change (UNFCCC). Under decision 3/CP.5 of the UNFCCC Conference of the 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, the EPA has instituted an annual public
review and comment process for this document. The availability of the draft document is announced via Federal
Register Notice and is posted on the EPA Greenhouse Gas Emissions web site. Copies are also emailed upon
request. The public comment period is generally limited to 30 days; however, comments received after the closure
of the public comment period are accepted and considered for the next edition of this annual report. Public review
of this report occurred from February 12 to March 13, 2020 and comments received are posted to the docket EPA-
HQ-OAR-2019-0706. Responses to comments are posted to EPA's website within 2-4 weeks following publication
of this report.
v

-------
Table of Contents
TABLE OF CONTENTS	VI
LIST OF TABLES, FIGURES, AND BOXES	IX
EXECUTIVE SUMMARY	ES-1
ES.l Background Information	ES-2
ES.2 RecentTrends 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-25
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	1-15
1.5	Key Categories	1-16
1.6	Quality Assurance and Quality Control (QA/QC)	1-21
1.7	Uncertainty Analysis of Emission Estimates	1-24
1.8	Completeness	1-27
1.9	Organization of Report	1-27
2.	TRENDS IN GREENHOUSE GAS EMISSIONS	2-1
2.1	Recent Trends in U.S. Greenhouse Gas Emissions and Sinks	2-1
2.2	Emissions by Economic Sector	2-25
2.3	Precursor Greenhouse Gas Emissions (CO, NOx, NMVOCs, and SO2)	2-36
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 1A5)	3-47
3.3	Incineration of Waste (CRF Source Category 1A5)	3-55
3.4	Coal Mining (CRF Source Category lBla)	3-59
3.5	Abandoned Underground Coal Mines (CRF Source Category lBla)	3-64
vi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
3.6	Petroleum Systems (CRF Source Category lB2a)	3-68
3.7	Natural Gas Systems (CRF Source Category lB2b)	3-84
3.8	Abandoned Oil and Gas Wells (CRF Source Categories lB2a and lB2b)	3-101
3.9	Energy Sources of Precursor Greenhouse Gas Emissions	3-106
3.10	International Bunker Fuels (CRF Source Category 1: Memo Items)	3-107
3.11	Wood Biomass and Biofuels Consumption (CRF Source Category 1A)	3-112
4.	INDUSTRIAL PROCESSES AND PRODUCT USE	4-1
4.1	Cement Production (CRF Source Category 2A1)	4-9
4.2	Lime Production (CRF Source Category 2A2)	4-13
4.3	Glass Production (CRF Source Category 2A3)	4-19
4.4	Other Process Uses of Carbonates (CRF Source Category 2A4)	4-22
4.5	Ammonia Production (CRF Source Category 2B1)	4-27
4.6	Urea Consumption for Non-Agricultural Purposes	4-32
4.7	Nitric Acid Production (CRF Source Category 2B2)	4-35
4.8	Adipic Acid Production (CRF Source Category 2B3)	4-40
4.9	Caprolactam, Glyoxal and Glyoxylic Acid Production (CRF Source Category 2B4)	4-44
4.10	Carbide Production and Consumption (CRF Source Category 2B5)	4-47
4.11	Titanium Dioxide Production (CRF Source Category 2B6)	4-51
4.12	Soda Ash Production (CRF Source Category 2B7)	4-54
4.13	Petrochemical Production (CRF Source Category 2B8)	4-57
4.14	HCFC-22 Production (CRF Source Category 2B9a)	4-64
4.15	Carbon Dioxide Consumption (CRF Source Category 2B10)	4-68
4.16	Phosphoric Acid Production (CRF Source Category 2B10)	4-72
4.17	Iron and Steel Production (CRF Source Category 2C1) and Metallurgical Coke Production	4-76
4.18	Ferroalloy Production (CRF Source Category 2C2)	4-86
4.19	Aluminum Production (CRF Source Category 2C3)	4-90
4.20	Magnesium Production and Processing (CRF Source Category 2C4)	4-96
4.21	Lead Production (CRF Source Category 2C5)	4-101
4.22	Zinc Production (CRF Source Category 2C6)	4-105
4.23	Electronics Industry (CRF Source Category 2E)	4-110
4.24	Substitution of Ozone Depleting Substances (CRF Source Category 2F)	4-125
4.25	Electrical Transmission and Distribution (CRF Source Category 2G1)	4-134
4.26	Nitrous Oxide from Product Uses (CRF Source Category 2G3)	4-143
4.27	Industrial Processes and Product Use Sources of Precursor Gases	4-146
5.	AGRICULTURE	5-1
vii

-------
5.1	Enteric Fermentation (CRF Source Category 3A)	5-3
5.2	Manure Management (CRF Source Category 3B)	5-10
5.3	Rice Cultivation (CRF Source Category 3C)	5-19
5.4	Agricultural Soil Management (CRF Source Category 3D)	5-26
5.5	Liming (CRF Source Category 3G)	5-46
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	6-9
6.2	Forest Land Remaining Forest Land (CRF Category 4A1)	6-24
6.3	Land Converted to Forest Land (CRF Source Category 4A2)	6-47
6.4	Cropland Remaining Cropland (CRF Category 4B1)	6-54
6.5	Land Converted to Cropland (CRF Category 4B2)	6-65
6.6	Grassland Remaining Grassland (CRF Category 4C1)	6-72
6.7	Land Converted to Grassland (CRF Category 4C2)	6-83
6.8	Wetlands Remaining Wetlands (CRF Category 4D1)	6-90
6.9	Land Converted to Wetlands (CRF Source Category 4D2)	6-111
6.10	Settlements Remaining Settlements (CRF Category 4E1)	6-115
6.11	Land Converted to Settlements (CRF Category 4E2)	6-135
6.12	Other Land Remaining Other Land (CRF Category 4F1)	6-142
6.13	Land Converted to Other Land (CRF Category 4F2)	6-143
7.	WASTE	7-1
7.1	Landfills (CRF Source Category 5A1)	7-4
7.2	Wastewater Treatment (CRF Source Category 5D)	7-18
7.3	Composting (CRF Source Category 5B1)	7-36
7.4	Waste Incineration (CRF Source Category 5C1)	7-39
7.5	Waste Sources of Precursor Greenhouse Gases	7-40
8.	OTHER	8-1
9.	RECALCULATIONS AND IMPROVEMENTS	9-1
10.	REFERENCES	10-1
viii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 Eq.)	ES-7
Table ES-3: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)	ES-13
Table ES-4: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2 Eq.)
	ES-19
Table ES-5: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and
Forestry (MMT C02 Eq.)	ES-23
Table ES-6: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)	ES-25
Table ES-7: U.S. Greenhouse Gas Emissions by Economic Sector with Electricity-Related Emissions Distributed
(MMT CO2 Eq.)	ES-26
Table ES-8: Recent Trends in Various U.S. Data (Index 1990 = 100)	ES-27
Table 1-1: Global Atmospheric Concentration, Rate of Concentration Change, and Atmospheric Lifetime of
Selected Greenhouse Gases	1-4
Table 1-2: Global Warming Potentials and Atmospheric Lifetimes (Years) Used in this Report	1-9
Table 1-3: Comparison of 100-Year GWP values	1-10
Table 1-4: Key Categories for the United States (1990 and 2018)	1-16
Table 1-5: Estimated Overall Inventory Quantitative Uncertainty for 1990 (MMT CO2 Eq. and Percent)	1-25
Table 1-6: Estimated Overall Inventory Quantitative Uncertainty for 2018 (MMT CO2 Eq. and Percent)	1-26
Table 1-7: IPCC Sector Descriptions	1-28
Table 1-8: List of Annexes	1-28
Table 2-1: RecentTrends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 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 CO2 Eq.)... 2-8
Table 2-4: Emissions from Energy (MMT CO2 Eq.)	2-11
Table 2-5: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)	2-13
Table 2-6: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)	2-17
Table 2-7: Emissions from Agriculture (MMT CO2 Eq.)	2-19
Table 2-8: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and Forestry
(MMT CO2 Eq.)	2-22
Table 2-9: Emissions from Waste (MMT CO2 Eq.)	2-24
Table 2-10: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq. and Percent of Total in
2018)	2-26
Table 2-11: Electric Power-Related Greenhouse Gas Emissions (MMT CO2 Eq.)	2-29
ix

-------
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 2018	2-30
Table 2-13: Transportation-Related Greenhouse Gas Emissions (MMT CO2 Eq.)	2-33
Table 2-14: Recent Trends in Various U.S. Data (Index 1990 = 100)	2-35
Table 2-15: Emissions of NOx, CO, NMVOCs, and S02 (kt)	2-37
Table 3-1: CO2, CFU, and N2O Emissions from Energy (MMT CO2 Eq.)	3-3
Table 3-2: CO2, CFU, and N2O Emissions from Energy (kt)	3-3
Table 3-3: CO2, CFU, and N2O Emissions from Fossil Fuel Combustion (MMT CO2 Eq.)	3-6
Table 3-4: CO2, CFU, and N2O Emissions from Fossil Fuel Combustion (kt)	3-6
Table 3-5: CO2 Emissions from Fossil Fuel Combustion by Fuel Type and Sector (MMT CO2 Eq.)	3-7
Table 3-6: Annual Change in CO2 Emissions and Total 2018 CO2 Emissions from Fossil Fuel Combustion for Selected
Fuels and Sectors (MMT CO2 Eq. and Percent)	3-8
Table 3-7: CO2, CFU, and N2O Emissions from Fossil Fuel Combustion by Sector (MMT CO2 Eq.)	3-12
Table 3-8: CO2, CFU, and N2O Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)	3-13
Table 3-9: CO2 Emissions from Stationary Fossil Fuel Combustion (MMT CO2 Eq.)	3-14
Table 3-10: CFU Emissions from Stationary Combustion (MMT CO2 Eq.)	3-14
Table 3-11: N2O Emissions from Stationary Combustion (MMT CO2 Eq.)	3-15
Table 3-12: Electric Power Generation by Fuel Type (Percent)	3-16
Table 3-13: CO2 Emissions from Fossil Fuel Combustion in Transportation End-Use Sector (MMT CO2 Eq.)	3-26
Table 3-14: CFU Emissions from Mobile Combustion (MMT CO2 Eq.)	3-28
Table 3-15: N2O Emissions from Mobile Combustion (MMT CO2 Eq.)	3-29
Table 3-16: Carbon Intensity from Direct Fossil Fuel Combustion by Sector (MMT CO2 Eq./QBtu)	3-34
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)	3-37
Table 3-18: Approach 2 Quantitative Uncertainty Estimates for CFU and N2O Emissions from Energy-Related
Stationary Combustion, Including Biomass (MMT CO2 Eq. and Percent)	3-42
Table 3-19: Approach 2 Quantitative Uncertainty Estimates for CFU and N2O Emissions from Mobile Sources (MMT
CO2 Eq. and Percent)	3-45
Table 3-20: CO2 Emissions from Non-Energy Use Fossil Fuel Consumption (MMT CO2 Eq. and Percent)	3-48
Table 3-21: Adjusted Consumption of Fossil Fuels for Non-Energy Uses (TBtu)	3-49
Table 3-22: 2018 Adjusted Non-Energy Use Fossil Fuel Consumption, Storage, and Emissions	3-49
Table 3-23: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Non-Energy Uses of Fossil Fuels
(MMT CO2 Eq. and Percent)	3-51
Table 3-24: Approach 2 Quantitative Uncertainty Estimates for Storage Factors of Non-Energy Uses of Fossil Fuels
(Percent)	3-51
Table 3-25: CO2, CFU, and N2O Emissions from the Incineration of Waste (MMT CO2 Eq.)	3-55
Table 3-26: CO2, CFU, and N2O Emissions from the Incineration of Waste (kt)	3-56
Table 3-27: Municipal Solid Waste Generation (Metric Tons) and Percent Combusted (BioCycle dataset)	3-57
x Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-28: Approach 2 Quantitative Uncertainty Estimates for CO2 and N2O from the Incineration of Waste (MMT
CO2 Eq. and Percent)	3-58
Table 3-29: Coal Production (kt)	3-59
Table 3-30: CFU Emissions from Coal Mining (MMT CO2 Eq.)	3-60
Table 3-31: CFU Emissions from Coal Mining (kt)	3-60
Table 3-32: Approach 2 Quantitative Uncertainty Estimates for CFU Emissions from Coal Mining (MMT CO2 Eq. and
Percent)	3-63
Table 3-33: CFU Emissions from Abandoned Coal Mines (MMT CO2 Eq.)	3-65
Table 3-34: CFU Emissions from Abandoned Coal Mines (kt)	3-65
Table 3-35: Number of Gassy Abandoned Mines Present in U.S. Basins in 2018, Grouped by Class According to
Post-Abandonment State	3-67
Table 3-36: Approach 2 Quantitative Uncertainty Estimates for CFU Emissions from Abandoned Underground Coal
Mines (MMT CO2 Eq. and Percent)	3-68
Table 3-37: CFU Emissions from Petroleum Systems (MMT CO2 Eq.)	3-70
Table 3-38: CFU Emissions from Petroleum Systems (kt CH4)	3-71
Table 3-39: CO2 Emissions from Petroleum Systems (MMT CO2)	3-71
Table 3-40: CO2 Emissions from Petroleum Systems (kt CO2)	3-71
Table 3-41: N2O Emissions from Petroleum Systems (metric tons CO2 Eq.)	3-71
Table 3-42: N2O Emissions from Petroleum Systems (metric tons N2O)	3-71
Table 3-43: Approach 2 Quantitative Uncertainty Estimates for CFU and CO2 Emissions from Petroleum Systems
(MMT CO2 Eq. and Percent)	3-74
Table 3-44: Recalculations of CO2 in Petroleum Systems (MMT CO2)	3-76
Table 3-45: Recalculations of CH4 in Petroleum Systems (MMT CO2 Eq.)	3-77
Table 3-46: HF Oil Well Completions National CO2 Emissions (kt CO2)	3-77
Table 3-47: Offshore Oil Production National CH4 Emissions (metric tons CH4)	3-79
Table 3-48: Offshore Oil Production National CO2 Emissions (metric tons CO2)	3-79
Table 3-49: HF Oil Well Workovers National CO2 Emissions (kt CO2)	3-80
Table 3-50: Pneumatic Controller National CH4 Emissions (Metric Tons CH4)	3-80
Table 3-51: Associated Gas Flaring National CO2 Emissions (kt CO2)	3-80
Table 3-52: Miscellaneous Production Flaring National CO2 Emissions (kt CO2)	3-81
Table 3-53: Producing Oil Well Count Data	3-81
Table 3-54: Refineries National CFU Emissions (metric tons CH4)	3-82
Table 3-55: Quantity of CO2 Captured and Extracted for EOR Operations (MMT CO2)	3-83
Table 3-56: Quantity of CO2 Captured and Extracted for EOR Operations (kt)	3-84
Table 3-57: CFU Emissions from Natural Gas Systems (MMT CO2 Eq.)a	3-86
Table 3-58: CFU Emissions from Natural Gas Systems (kt)a	3-87
Table 3-59: Non-combustion CO2 Emissions from Natural Gas Systems (MMT)	3-87
xi

-------
Table 3-60: Non-combustion CO2 Emissions from Natural Gas Systems (kt)	3-87
Table 3-61: N2O Emissions from Natural Gas Systems (metric tons CO2 Eq.)	3-88
Table 3-62: N2O Emissions from Natural Gas Systems (metric tons N2O)	3-88
Table 3-63: Approach 2 Quantitative Uncertainty Estimates for CFU and Non-combustion CO2 Emissions from
Natural Gas Systems (MMT CO2 Eq. and Percent)	3-90
Table 3-64: Recalculations of CO2 in Natural Gas Systems (MMT CO2)	3-92
Table 3-65: Recalculations of CH4 in Natural Gas Systems (MMT CO2 Eq.)	3-92
Table 3-66: Gathering Stations National CFU Emissions (metric tons CH4)	3-94
Table 3-67: Gathering Stations National CO2 Emissions (metric tons CO2)	3-95
Table 3-68: Offshore Gas Production National Emissions (metric tons CH4)	3-97
Table 3-69: Offshore Gas Production National Emissions (metric tons CO2)	3-97
Table 3-70: Production Segment Pneumatic Controller National Emissions (metric tons CH4)	3-97
Table 3-71: Liquids Unloading National Emissions (metric tons CH4)	3-98
Table 3-72: Production Segment Storage Tanks National Emissions (metric tons CO2)	3-98
Table 3-73: HF Gas Well Workovers National Emissions (metric tons CH4)	3-98
Table 3-74: Non-HF Gas Well Workovers National Emissions (metric tons CH4)	3-99
Table 3-75: Producing Gas Well Count Data	3-99
Table 3-76: AGR National CO2 Emissions (kt CO2)	3-99
Table 3-77: Processing Segment Flares National Emissions (metric tons CH4)	3-100
Table 3-78: Processing Segment Reciprocating Compressors National Emissions (metric tons CH4)	3-100
Table 3-79: Processing Segment Blowdowns/Venting National Emissions (metric tons CH4)	3-100
Table 3-80: CFU Emissions from Abandoned Oil and Gas Wells (MMT CO2 Eq.)	3-102
Table 3-81: CFU Emissions from Abandoned Oil and Gas Wells (kt)	3-102
Table 3-82: CO2 Emissions from Abandoned Oil and Gas Wells (MMT CO2)	3-102
Table 3-83: CO2 Emissions from Abandoned Oil and Gas Wells (kt)	3-102
Table 3-84: Abandoned Oil Wells Activity Data, CH4 and CO2 Emissions (metric tons)	3-103
Table 3-85: Abandoned Gas Wells Activity Data, CH4 and CO2 Emissions (metric tons)	3-104
Table 3-86: Approach 2 Quantitative Uncertainty Estimates for CFU and CO2 Emissions from Petroleum and Natural
Gas Systems (MMT CO2 Eq. and Percent)	3-105
Table 3-87: NOx, CO, and NMVOC Emissions from Energy-Related Activities (kt)	3-106
Table 3-88: CO2, CFU, and N2O Emissions from International Bunker Fuels (MMT CO2 Eq.)	3-108
Table 3-89: CO2, CFU, and N2O Emissions from International Bunker Fuels (kt)	3-108
Table 3-90: Aviation Jet Fuel Consumption for International Transport (Million Gallons)	3-110
Table 3-91: Marine Fuel Consumption for International Transport (Million Gallons)	3-110
Table 3-92: CO2 Emissions from Wood Consumption by End-Use Sector (MMT CO2 Eq.)	3-112
Table 3-93: CO2 Emissions from Wood Consumption by End-Use Sector (kt)	3-113
xii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-94
Table 3-95
Table 3-96
Table 3-97
Table 3-98
Table 3-99
CO2 Emissions from Ethanol Consumption (MMT CO2 Eq.)	3-113
CO2 Emissions from Ethanol Consumption (kt)	3-113
CO2 Emissions from Biodiesel Consumption (MMT CO2 Eq.)	3-114
CO2 Emissions from Biodiesel Consumption (kt)	3-114
Woody Biomass Consumption by Sector (Trillion Btu)	3-115
Ethanol Consumption by Sector (Trillion Btu)	3-115
Table 3-100: Biodiesel Consumption by Sector (Trillion Btu)	3-115
Table 4-1: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)	4-3
Table 4-2: Emissions from Industrial Processes and Product Use (kt)	4-4
Table 4-3: CO2 Emissions from Cement Production (MMT CO2 Eq. and kt)	4-9
Table 4-4: Clinker Production (kt)	4-11
Table 4-5: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Cement Production (MMT CO2
Eq. and Percent)	4-12
Table 4-6: CO2 Emissions from Lime Production (MMT CO2 Eq. and kt)	4-14
Table 4-7: Gross, Recovered, and Net CO2 Emissions from Lime Production (kt)	4-14
Table 4-8: High-Calcium- and Dolomitic-Quicklime, High-Calcium- and Dolomitic-Hydrated, and Dead-Burned-
Dolomite Lime Production (kt)	4-15
Table 4-9: Adjusted Lime Production (kt)	4-16
Table 4-10: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lime Production (MMT CO2 Eq.
and Percent)	4-17
Table 4-11: CO2 Emissions from Glass Production (MMT CO2 Eq. and kt)	4-20
Table 4-12: Limestone, Dolomite, and Soda Ash Consumption Used in Glass Production (kt)	4-20
Table 4-13: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Glass Production (MMT CO2 Eq.
and Percent)	4-21
Table 4-14: CO2 Emissions from Other Process Uses of Carbonates (MMT CO2 Eq.)	4-23
Table 4-15: CO2 Emissions from Other Process Uses of Carbonates (kt)	4-23
Table 4-16: Limestone and Dolomite Consumption (kt)	4-25
Table 4-17: Soda Ash Consumption Not Associated with Glass Manufacturing (kt)	4-25
Table 4-18: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Other Process Uses of
Carbonates (MMT CO2 Eq. and Percent)	4-26
Table 4-19: CO2 Emissions from Ammonia Production (MMT CO2 Eq.)	4-28
Table 4-20: CO2 Emissions from Ammonia Production (kt)	4-28
Table 4-21: Ammonia Production, Recovered CO2 Consumed for Urea Production, and Urea Production (kt).... 4-29
Table 4-22: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Ammonia Production (MMT
CO2 Eq. and Percent)	4-30
Table 4-23: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (MMT CO2 Eq.)	4-32
Table 4-24: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (kt)	4-32
xiii

-------
Table 4-25: Urea Production, Urea Applied as Fertilizer, Urea Imports, and Urea Exports (kt)	4-34
Table 4-26: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Urea Consumption for Non-
Agricultural Purposes (MMT CO2 Eq. and Percent)	4-34
Table 4-27: N2O Emissions from Nitric Acid Production (MMT CO2 Eq. and kt N2O)	4-36
Table 4-28: Nitric Acid Production (kt)	4-38
Table 4-29: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Nitric Acid Production (MMT
CO2 Eq. and Percent)	4-39
Table 4-30: N2O Emissions from Adipic Acid Production (MMT CO2 Eq. and kt N2O)	4-40
Table 4-31: Adipic Acid Production (kt)	4-42
Table 4-32: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Adipic Acid Production (MMT
CO2 Eq. and Percent)	4-43
Table 4-33: N2O Emissions from Caprolactam Production (MMT CO2 Eq. and kt N2O)	4-45
Table 4-34: Caprolactam Production (kt)	4-46
Table 4-35: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Caprolactam, Glyoxal and
Glyoxylic Acid Production (MMT CO2 Eq. and Percent)	4-47
Table 4-36: CO2 and Cm Emissions from Silicon Carbide Production and Consumption (MMT CO2 Eq.)	4-48
Table 4-37: CO2 and Cm Emissions from Silicon Carbide Production and Consumption (kt)	4-48
Table 4-38: Production and Consumption of Silicon Carbide (Metric Tons)	4-49
Table 4-39: Approach 2 Quantitative Uncertainty Estimates for Cm and CO2 Emissions from Silicon Carbide
Production and Consumption (MMT CO2 Eq. and Percent)	4-50
Table 4-40: CO2 Emissions from Titanium Dioxide (MMT CO2 Eq. and kt)	4-51
Table 4-41: Titanium Dioxide Production (kt)	4-52
Table 4-42: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Titanium Dioxide Production
(MMT CO2 Eq. and Percent)	4-53
Table 4-43: CO2 Emissions from Soda Ash Production (MMT CO2 Eq. and kt CO2)	4-55
Table 4-44: Soda Ash Production (kt)	4-56
Table 4-45: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Soda Ash Production (MMT
CO2 Eq. and Percent)	4-56
Table 4-46: CO2 and Cm Emissions from Petrochemical Production (MMT CO2 Eq.)	4-59
Table 4-47: CO2 and Cm Emissions from Petrochemical Production (kt)	4-59
Table 4-48: Production of Selected Petrochemicals (kt)	4-61
Table 4-49: Approach 2 Quantitative Uncertainty Estimates for Cm Emissions from Petrochemical Production and
CO2 Emissions from Petrochemical Production (MMT CO2 Eq. and Percent)	4-62
Table 4-50: HFC-23 Emissions from HCFC-22 Production (MMT C02 Eq. and kt HFC-23)	4-65
Table 4-51: HCFC-22 Production (kt)	4-66
Table 4-52: Approach 2 Quantitative Uncertainty Estimates for HFC-23 Emissions from HCFC-22 Production (MMT
CO2 Eq. and Percent)	4-67
Table 4-53: CO2 Emissions from CO2 Consumption (MMT CO2 Eq. and kt)	4-68
xiv Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-54: CO2 Production (kt CO2) and the Percent Used for Non-EOR Applications	4-70
Table 4-55: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from CO2 Consumption (MMT CO2
Eq. and Percent)	4-71
Table 4-56: CO2 Emissions from Phosphoric Acid Production (MMT CO2 Eq. and kt)	4-72
Table 4-57: Phosphate Rock Domestic Consumption, Exports, and Imports (kt)	4-73
Table 4-58: Chemical Composition of Phosphate Rock (Percent by Weight)	4-74
Table 4-59: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Phosphoric Acid Production
(MMT CO2 Eq. and Percent)	4-75
Table 4-60:	CO2 Emissions from Metallurgical Coke Production (MMT CO2 Eq.)	4-77
Table 4-61:	CO2 Emissions from Metallurgical Coke Production (kt)	4-77
Table 4-62:	CO2 Emissions from Iron and Steel Production (MMT CO2 Eq.)	4-78
Table 4-63:	CO2 Emissions from Iron and Steel Production (kt)	4-78
Table 4-64:	CFU Emissions from Iron and Steel Production (MMT CO2 Eq.)	4-78
Table 4-65:	CFU Emissions from Iron and Steel Production (kt)	4-78
Table 4-66:	Material Carbon Contents for Metallurgical Coke Production	4-80
Table 4-67: Production and Consumption Data for the Calculation of CO2 Emissions from Metallurgical Coke
Production (Thousand Metric Tons)	4-80
Table 4-68: Production and Consumption Data for the Calculation of CO2 Emissions from Metallurgical Coke
Production (Million ft3)	4-80
Table 4-69: Material Carbon Contents for Iron and Steel Production	4-81
Table 4-70: CFU Emission Factors for Sinter and Pig Iron Production	4-82
Table 4-71: CO2 Emission Factors for Sinter Production, Direct Reduced Iron Production and Pellet Production . 4-82
Table 4-72: Production and Consumption Data for the Calculation of CO2 and CFU Emissions from Iron and Steel
Production (Thousand Metric Tons)	4-83
Table 4-73: Production and Consumption Data for the Calculation of CO2 Emissions from Iron and Steel Production
(Million ft3 unless otherwise specified)	4-83
Table 4-74: Approach 2 Quantitative Uncertainty Estimates for CO2 and CFU Emissions from Iron and Steel
Production and Metallurgical Coke Production (MMT CO2 Eq. and Percent)	4-85
Table 4-75: CO2 and CFU Emissions from Ferroalloy Production (MMT CO2 Eq.)	4-87
Table 4-76: CO2 and CFU Emissions from Ferroalloy Production (kt)	4-87
Table 4-77: Production of Ferroalloys (Metric Tons)	4-88
Table 4-78: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Ferroalloy Production (MMT
CO2 Eq. and Percent)	4-89
Table 4-79
Table 4-80
Table 4-81
Table 4-82
CO2 Emissions from Aluminum Production (MMT CO2 Eq. and kt)	4-91
PFC Emissions from Aluminum Production (MMT CO2 Eq.)	4-91
PFC Emissions from Aluminum Production (kt)	4-92
Production of Primary Aluminum (kt)	4-95
xv

-------
Table 4-83: Approach 2 Quantitative Uncertainty Estimates for CO2 and PFC Emissions from Aluminum Production
(MMT CO2 Eq. and Percent)	4-95
Table 4-84: SF6, HFC- 134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and Processing (MMT CO2
Eq.)	4-96
Table 4-85: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and Processing (kt)	4-97
Table 4-86: SF6 Emission Factors (kg SF6 per metric ton of magnesium)	4-99
Table 4-87: Approach 2 Quantitative Uncertainty Estimates for SF6, HFC-134a and CO2 Emissions from Magnesium
Production and Processing (MMT CO2 Eq. and Percent)	4-100
Table 4-88: CO2 Emissions from Lead Production (MMT CO2 Eq. and kt)	4-102
Table 4-89: Lead Production (Metric Tons)	4-103
Table 4-90: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lead Production (MMT CO2 Eq.
and Percent)	4-104
Table 4-91: CO2 Emissions from Zinc Production (MMT CO2 Eq. and kt)	4-106
Table 4-92: Zinc Production (Metric Tons)	4-106
Table 4-93: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Zinc Production (MMT CO2 Eq.
and Percent)	4-109
Table 4-94: PFC, HFC, SF6, NF3, and N2O Emissions from Electronics Manufacture (MMT CO2 Eq.)	4-112
Table 4-95: PFC, HFC, SF6, NF3, and N2O Emissions from Electronics Manufacture (metric tons)	4-113
Table 4-96: F-HTF Emissions from Electronics Manufacture by Compound Group (metric tons)	4-113
Table 4-97: F-GHGa Emissions from PV and MEMS manufacturing (MMT CO2 Eq.)	4-113
Table 4-98: Approach 2 Quantitative Uncertainty Estimates for HFC, PFC, SF6, NF3 and N2O Emissions from
Semiconductor Manufacture (MMT CO2 Eq. and Percent)3	4-123
Table 4-99: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.)	4-125
Table 4-100: Emissions of HFCs and PFCs from ODS Substitution (Metric Tons)	4-126
Table 4-101: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.) by Sector	4-126
Table 4-102: Approach 2 Quantitative Uncertainty Estimates for HFC and PFC Emissions from ODS Substitutes
(MMT CO2 Eq. and Percent)	4-129
Table 4-103: U.S. HFC Supply (MMT CO2 Eq.)	4-131
Table 4-104: Averaged U.S. HFC Demand (MMT C02 Eq.)	4-133
Table 4-105: SF6 Emissions from Electric Power Systems and Electrical Equipment Manufacturers (MMT CO2 Eq.)
	4-135
Table 4-106: SF6 Emissions from Electric Power Systems and Electrical Equipment Manufacturers (kt)	4-135
Table 4-107: Transmission Mile Coverage (Percent) and Regression Coefficients (kg per mile)	4-139
Table 4-108: Approach 2 Quantitative Uncertainty Estimates for SF6 Emissions from Electrical Transmission and
Distribution (MMT CO2 Eq. and Percent)	4-141
Table 4-109: N2O Production (kt)	4-143
Table 4-110: N2O Emissions from N2O Product Usage (MMT CO2 Eq. and kt)	4-143
Table 4-111: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from N2O Product Usage (MMT
CO2 Eq. and Percent)	4-145
xvi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-112: N0X, CO, and NMVOC Emissions from Industrial Processes and Product Use (kt)	4-146
Table 5-1: Emissions from Agriculture (MMT CO2 Eq.)	5-2
Table 5-2: Emissions from Agriculture (kt)	5-2
Table 5-3: Cm Emissions from Enteric Fermentation (MMT CO2 Eq.)	5-4
Table 5-4: Cm Emissions from Enteric Fermentation (kt)	5-4
Table 5-5: Cattle Sub-Population Categories for 2018 Population Estimates	5-7
Table 5-6: Approach 2 Quantitative Uncertainty Estimates for CFU Emissions from Enteric Fermentation (MMT CO2
Eq. and Percent)	5-8
Table 5-7: CFU and N2O Emissions from Manure Management (MMT CO2 Eq.)	5-12
Table 5-8: CFU and N2O Emissions from Manure Management (kt)	5-13
Table 5-9: Approach 2 Quantitative Uncertainty Estimates for CFU and N2O (Direct and Indirect) Emissions from
Manure Management (MMT CO2 Eq. and Percent)	5-16
Table 5-10: IPCC (2006) Implied Emission Factor Default Values Compared with Calculated Values for CH4 from
Manure Management (kg/head/year)	5-18
Table 5-11: CFU Emissions from Rice Cultivation (MMT CO2 Eq.)	5-20
Table 5-12: CFU Emissions from Rice Cultivation (kt)	5-21
Table 5-13: Rice Area Harvested (1,000 Hectares)	5-23
Table 5-14: Average Ratooned Area as Percent of Primary Growth Area (Percent)	5-24
Table 5-15: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Rice Cultivation (MMT CO2 Eq.
and Percent)	5-25
Table 5-16
Table 5-17
Table 5-18
Table 5-19
Table 5-20
(MMT CO2
Table 5-21
Table 5-22
Table 5-23
Table 5-24
Percent)...
N2O Emissions from Agricultural Soils (MMT CO2 Eq.)	
N2O Emissions from Agricultural Soils (kt)	
Direct N2O Emissions from Agricultural Soils by Land Use Type and N Input Type (MMT CO2 Eq.)...
Indirect N2O Emissions from Agricultural Soils (MMT CO2 Eq.)	
Quantitative Uncertainty Estimates of N2O Emissions from Agricultural Soil Management in 2018
Eq. and Percent)	
Emissions from Liming (MMT CO2 Eq.)	
Emissions from Liming (MMT C)	
Applied Minerals (MMT)	
Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Liming (MMT CO2 Eq. and
Table 5-25
Table 5-26
Table 5-27
Table 5-28
Percent)...
CO2 Emissions from Urea Fertilization (MMT CO2 Eq.)	
CO2 Emissions from Urea Fertilization (MMT C)	
Applied Urea (MMT)	
Quantitative Uncertainty Estimates for CO2 Emissions from Urea Fertilization (MMT CO2 Eq. and
Table 5-29: CH4 and N2O Emissions from Field Burning of Agricultural Residues (MMT CO2 Eq.
Table 5-30: CH4, N2O, CO, and NOx Emissions from Field Burning of Agricultural Residues (kt).
. 5-29
. 5-29
. 5-29
. 5-30
. 5-44
. 5-46
. 5-46
. 5-47
. 5-48
. 5-49
. 5-49
. 5-49
. 5-50
. 5-51
. 5-52
xvii

-------
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	5-56
Table 5-34: Greenhouse Gas Emission Ratios and Conversion Factors	5-57
Table 5-35: Approach 2 Quantitative Uncertainty Estimates for CFU and N2O Emissions from Field Burning of
Agricultural Residues (MMT CO2 Eq. and Percent)	5-58
Table 6-1
Table 6-2
Table 6-3
Table 6-4
Table 6-5
Table 6-6
Net CO2 Flux from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.)	6-2
Emissions from Land Use, Land-Use Change, and Forestry by Gas (MMT CO2 Eq.)	6-4
Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.).... 6-5
Emissions and Removals from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.)	6-6
Emissions and Removals from Land Use, Land-Use Change, and Forestry (kt)	6-7
Managed and Unmanaged Land Area by Land-Use Categories for All 50 States (Thousands of Hectares)
	6-10
Table 6-7: Land Use and Land-Use Change for the U.S. Managed Land Base for All 50 States (Thousands of
Hectares)	6-11
Table 6-8: Data Sources Used to Determine Land Use and Land Area for the Conterminous United States, Hawaii,
and Alaska	6-17
Table 6-9: Total Land Area (Hectares) by Land-Use Category for U.S. Territories	6-23
Table 6-10: Net CO2 Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and Harvested Wood
Pools (MMT CO2 Eq.)	6-28
Table 6-11: Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and Harvested Wood
Pools (MMT C)	6-29
Table 6-12: Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and Harvested Wood Pools
(MMT C)	6-30
Table 6-13: Estimates of CO2 (MMT per Year) Emissions from Forest Fires in the Conterminous 48 States and
Alaska3	6-32
Table 6-14: Quantitative Uncertainty Estimates for Net CO2 Flux from Forest Land Remaining Forest Land: Changes
in Forest C Stocks (MMT CO2 Eq. and Percent)	6-36
Table 6-15: Recalculations of Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C)	6-37
Table 6-16: Recalculations of Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C)	6-37
Table 6-17: Non-C02 Emissions from Forest Fires (MMT CO2 Eq.)a	6-39
Table 6-18: Non-C02 Emissions from Forest Fires (kt)a	6-39
Table 6-19: Quantitative Uncertainty Estimates of Non-C02 Emissions from Forest Fires (MMT CO2 Eq. and
Percent)3	6-40
Table 6-20: N2O Fluxes from Soils in Forest Land Remaining Forest Land and Land Converted to Forest Land (MMT
CO2 Eq. and kt N20)	6-41
Table 6-21: 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)	6-43
xviii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-22: Non-CC>2 Emissions from Drained Organic Forest Soilsa b (MMT CO2 Eq.)	6-44
Table 6-23: Non-CC>2 Emissions from Drained Organic Forest Soilsa b (kt)	6-44
Table 6-24: States identified as having Drained Organic Soils, Area of Forest on Drained Organic Soils, and
Sampling Error	6-45
Table 6-25: Quantitative Uncertainty Estimates for Non-C02 Emissions on Drained Organic Forest Soils (MMT CO2
Eq. and Percent)3	6-46
Table 6-26: Net CO2 Flux from Forest C Pools in Land Converted to Forest Land by Land Use Change Category (MMT
CO2 Eq.)	6-47
Table 6-27: Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use Change Category (MMT
C)	6-48
Table 6-28: Quantitative Uncertainty Estimates for Forest C Pool Stock Changes (MMT CO2 Eq. per Year) in 2018
from Land Converted to Forest Land by Land Use Change	6-51
Table 6-29: Recalculations of the Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use
Change Category (MMT C)	6-53
Table 6-30: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT CO2 Eq.)	6-55
Table 6-31: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT C)	6-55
Table 6-32: Approach 2 Quantitative Uncertainty Estimates for Soil C Stock Changes occurring within Cropland
Remaining Cropland (MMT CO2 Eq. and Percent)	6-63
Table 6-33: Area of Managed Land in Cropland Remaining Cropland that is not included in the current Inventory
(Thousand Hectares)	6-65
Table 6-34: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Land Converted to
Cropland by Land Use Change Category (MMT CO2 Eq.)	6-66
Table 6-35: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Land Converted to
Cropland (MMT C)	6-67
Table 6-36: 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)	6-70
Table 6-37: Area of Managed Land in Land Converted to Cropland that is not included in the current Inventory
(Thousand Hectares)	6-71
Table 6-38: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Grassland Remaining
Grassland (MMT C02 Eq.)	6-73
Table 6-39: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Grassland Remaining
Grassland (MMT C)	6-73
Table 6-40: Approach 2 Quantitative Uncertainty Estimates for C Stock Changes Occurring Within Grassland
Remaining Grassland (MMT CO2 Eq. and Percent)	6-78
Table 6-41: Area of Managed Land in Grassland Remaining Grassland in Alaska that is not included in the current
Inventory (Thousand Hectares)	6-79
Table 6-42: CH4 and N2O Emissions from Biomass Burning in Grassland (MMT CO2 Eq.)	6-80
Table 6-43: CH4, N2O, CO, and NOx Emissions from Biomass Burning in Grassland (kt)	6-81
Table 6-44: Thousands of Grassland Hectares Burned Annually	6-81
Table 6-45: Uncertainty Estimates for Non-C02 Greenhouse Gas Emissions from Biomass Burning in Grassland
(MMT CO2 Eq. and Percent)	6-82
xix

-------
Table 6-46: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Grassland (MMT CO2 Eq.)	6-84
Table 6-47: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Grassland (MMT C)	6-84
Table 6-48: 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)	6-88
Table 6-49: Area of Managed Land in Land Converted to Grassland in Alaska that is not included in the current
Inventory (Thousand Hectares)	6-90
Table 6-50: Emissions from Peatlands Remaining Peatlands (MMT CO2 Eq.)	6-92
Table 6-51: Emissions from Peatlands Remaining Peatlands (kt)	6-92
Table 6-52: Peat Production of Lower 48 States (kt)	6-94
Table 6-53: Peat Production of Alaska (Thousand Cubic Meters)	6-94
Table 6-54: Peat Production Area of Lower 48 States (hectares)	6-95
Table 6-55: Peat Production Area of Alaska (hectares)	6-95
Table 6-56: Peat Production (hectares)	6-95
Table 6-57: Approach 2 Quantitative Uncertainty Estimates for CO2, CH4, and N2O Emissions from Peatlands
Remaining Peatlands (MMT CO2 Eq. and Percent)	6-97
Table 6-58: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands
(MMT CO2 Eq.)	6-99
Table 6-59: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands
(MMT C)	6-100
Table 6-60: CFU Emissions from Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands (MMT CO2
Eq. and kt CH4)	6-100
Table 6-61: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes and CH4 Emissions occurring
within Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent)	6-101
Table 6-62: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to Unvegetated Open Water
Coastal Wetlands (MMT C02 Eq.)	6-103
Table 6-63: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to Unvegetated Open Water
Coastal Wetlands (MMT C)	6-103
Table 6-64: Approach 1 Quantitative Uncertainty Estimates for CO2 Flux Occurring within Vegetated Coastal
Wetlands Converted to Unvegetated Open Water Coastal Wetlands (MMT CO2 Eq. and Percent)	6-105
Table 6-65: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands Converted to
Vegetated Coastal Wetlands (MMT CO2 Eq.)	6-106
Table 6-66: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands Converted to
Vegetated Coastal Wetlands (MMT C)	6-107
Table 6-67: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes Occurring within Unvegetated
Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent)	6-109
Table 6-68: N2O Emissions from Aquaculture in Coastal Wetlands (MMT CO2 Eq. and kt N2O)	6-110
Table 6-69: Approach 1 Quantitative Uncertainty Estimates for N2O Emissions for Aquaculture Production in
Coastal Wetlands (MMT CO2 Eq. and Percent)	6-111
Table 6-70: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq.) 6-112
xx Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-71: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal Wetlands (MMT C)	6-112
Table 6-72: Cm Emissions from Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq. and kt CH4).... 6-112
Table 6-73: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes occurring within Land Converted
to Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent)	6-114
Table 6-74: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements (MMT CO2 Eq.)	6-116
Table 6-75: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements (MMT C)	6-116
Table 6-76: Thousands of Hectares of Drained Organic Soils in Settlements Remaining Settlements	6-116
Table 6-77: Uncertainty Estimates for CO2 Emissions from Drained Organic Soils in Settlements Remaining
Settlements (MMT CO2 Eq. and Percent)	6-117
Table 6-78: Area of Managed Land in Settlements Remaining Settlements that is not included in the current
Inventory (Thousand Hectares)	6-118
Table 6-79: Net Flux from Settlement Trees in Settlements Remaining Settlements (MMT CO2 Eq. and MMT C)a
	6-119
Table 6-80: 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)	6-121
Table 6-81: 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 (2018)	6-123
Table 6-82: Approach 2 Quantitative Uncertainty Estimates for Net CO2 Flux from Changes in C Stocks in
Settlement Trees (MMT CO2 Eq. and Percent)	6-125
Table 6-83: N2O Emissions from Soils in Settlements Remaining Settlements (MMT CO2 Eq. and kt N2O)	6-126
Table 6-84: Quantitative Uncertainty Estimates of N2O Emissions from Soils in Settlements Remaining Settlements
(MMT CO2 Eq. and Percent)	6-128
Table 6-85: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills (MMT CO2 Eq.)	6-130
Table 6-86: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills (MMT C)	6-130
Table 6-87: Moisture Contents, C Storage Factors (Proportions of Initial C Sequestered), Initial C Contents, and
Decay Rates for Yard Trimmings and Food Scraps in Landfills	6-133
Table 6-88: C Stocks in Yard Trimmings and Food Scraps in Landfills (MMT C)	6-133
Table 6-89: Approach 2 Quantitative Uncertainty Estimates for CO2 Flux from Yard Trimmings and Food Scraps in
Landfills (MMT CO2 Eq. and Percent)	6-134
Table 6-90: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Settlements (MMT CO2 Eq.)	6-136
Table 6-91: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Settlements (MMT C)	6-137
Table 6-92: 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)	6-139
Table 6-93: Area of Managed Land in Settlements Remaining Settlements that is not included in the current
Inventory (Thousand Hectares)	6-142
Table 7-1: Emissions from Waste (MMT CO2 Eq.)	7-2
Table 7-2: Emissions from Waste (kt)	7-2
xx i

-------
Table 7-3
Table 7-4
Table 7-5
Percent).
Table 7-6
Table 7-7
Table 7-8
Table 7-9
Cm Emissions from Landfills (MMT CO2 Eq.)	7-6
CH4 Emissions from Landfills (kt)	7-7
Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Landfills (MMT CO2 Eq. and
	7-13
Materials Discarded3 in the Municipal Waste Stream by Waste Type from 1990 to 2017 (Percent)15.. 7-17
CH4 and N2O Emissions from Domestic and Industrial Wastewater Treatment (MMT CO2 Eq.)	7-19
CH4 and N2O Emissions from Domestic and Industrial Wastewater Treatment (kt)	7-20
U.S. Population (Millions) and Domestic Wastewater BODs Produced (kt)	7-22
Table 7-10: Domestic Wastewater CH4 Emissions from Septic and Centralized Systems (2018, MMT CO2 Eq. and
Percent)	7-23
Table 7-11: Industrial Wastewater CH4 Emissions by Sector (2018, MMT CO2 Eq. and Percent)	7-24
Table 7-12: U.S. Pulp and Paper, Meat, Poultry, Vegetables, Fruits and Juices, Ethanol, Breweries, and Petroleum
Refining Production (MMT)	7-24
Table 7-13: Variables Used to Calculate Percent Wastewater Treated Anaerobically by Industry (Percent)	7-25
Table 7-14: Wastewater Flow (m3/ton) and BOD Production (g/L) for U.S. Vegetables, Fruits, and Juices Production
	7-27
Table 7-15: Wastewater Treatment Distribution for Breweries	7-30
Table 7-16: U.S. Population (Millions), Population Served by Biological Denitrification (Millions), Fraction of
Population Served by Wastewater Treatment (percent), Available Protein (kg/person-year), Protein Consumed
(kg/person-year), and Nitrogen Removed with Sludge (kt-N/year)	7-33
Table 7-17: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Wastewater Treatment (MMT
CO2 Eq. and Percent)	7-33
Table 7-18: CH4 and N2O Emissions from Composting (MMT CO2 Eq.)	7-37
Table 7-19: CH4 and N2O Emissions from Composting (kt)	7-37
Table 7-20: U.S. Waste Composted (kt)	7-38
Table 7-21: Tier 1 Quantitative Uncertainty Estimates for Emissions from Composting (MMT CO2 Eq. and Percent)
	7-38
Table 7-22: Emissions of NOx, CO, and NMVOC from Waste (kt)	7-40
Table 9-1: Revisions to U.S. Greenhouse Gas Emissions (MMT CO2 Eq.)	9-4
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.)	9-6
Figure ES-1
Figure ES-2
Figure ES-3
Figure ES-4
Figure ES-5
Figure ES-6
Gross U.S. Greenhouse Gas Emissions by Gas (MMT CO2 Eq.) 	ES-5
Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the Previous Year 	ES-5
Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990 (1990=0, MMT CO2 Eq.) .ES-6
2018 U.S. Greenhouse Gas Emissions by Gas (Percentages based on MMT CO2 Eq.) 	ES-10
2018 Sources of CO2 Emissions (MMT CO2 Eq.) 	ES-11
2018 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT CO2 Eq.)	ES-12
xxii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure ES-7: 2018 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion (MMT CO2 Eq.) 	ES-13
Figure ES-8: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.) 	ES-15
Figure ES-9: 2018 Sources of CFU Emissions (MMT CO2 Eq.) 	ES-16
Figure ES-10: 2018 Sources of N2O Emissions (MMT CO2 Eq.)	ES-17
Figure ES-11: 2018 Sources of HFCs, PFCs, SF6, and NF3 Emissions (MMT CO2 Eq.) 	ES-18
Figure ES-12: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2 Eq.) 	ES-19
Figure ES-13: 2018 U.S. Energy Consumption by Energy Source (Percent) 	ES-21
Figure ES-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)	ES-25
Figure ES-15: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed to Economic Sectors
(MMT CO2 Eq.) 	ES-27
Figure ES-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic Product (GDP)	ES-28
Figure ES-17: 2018 Key Categories (MMT C02 Eq.)a	ES-29
Figure 1-1: National Inventory Arrangements Diagram Inventory Process Inventory Process	1-12
Figure 1-2: U.S. QA/QC Plan Summary 	1-23
Figure 2-1: Gross 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, MMT CO2 Eq.).... 2-3
Figure 2-4: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2 Eq.) 	2-8
Figure 2-5: 2018 Energy Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 	2-10
Figure 2-6: 2018 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT CO2 Eq.)	2-14
Figure 2-7: 2018 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion (MMT CO2 Eq.) 	2-14
Figure 2-8: Electric Power Generation (Billion kWh) and Emissions (MMT C02 Eq.) 	2-15
Figure 2-9: 2018 Industrial Processes and Product Use Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 	2-17
Figure 2-10: 2018 Agriculture Chapter Greenhouse Gas Sources (MMT CO2 Eq.)	2-19
Figure 2-11: 2018 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.) 	2-21
Figure 2-12: 2018 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 	2-24
Figure 2-13: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)	2-25
Figure 2-14: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed to Economic Sectors
(MMT CO2 Eq.) 	2-30
Figure 2-15: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic Product 	2-36
Figure 3-1: 2018 Energy Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 	3-2
Figure 3-2: 2018 U.S. Fossil Carbon Flows (MMT C02 Eq.) 	3-2
Figure 3-3: 2018 U.S. Energy Use by Energy Source (Percent) 	3-9
Figure 3-4: Annual U.S. Energy Use (Quadrillion Btu)	3-9
Figure 3-5: 2018 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT CO2 Eq.) 	3-10
xxiii

-------
Figure 3-6: Annual Deviations from Normal Heating Degree Days for the United States (1950-2018, Index Normal =
100)	3-11
Figure 3-7: Annual Deviations from Normal Cooling Degree Days for the United States (1950-2018, Index Normal =
100)	3-11
Figure 3-8: Fuels Used in Electric Power Generation (TBtu) and Total Electric Power Sector CO2 Emissions	3-17
Figure 3-9: Electric Power Retail Sales by End-Use Sector (Billion kWh)	3-18
Figure 3-10: Industrial Production Indices (Index 2012=100)	3-19
Figure 3-11: Fuels Used in Residential and Commercial Sectors (TBtu), Heating and Cooling Degree Days, and Total
Sector CO2 Emissions	3-21
Figure 3-12: Fuels Used in Transportation Sector (TBtu), Onroad VMT, and Total Sector CO2 Emissions	3-23
Figure 3-13: Sales-Weighted Fuel Economy of New Passenger Cars and Light-Duty Trucks, 1990-2018
(miles/gallon)	3-25
Figure 3-14: Sales of New Passenger Cars and Light-Duty Trucks, 1990-2018 (Percent)	3-25
Figure 3-15: Mobile Source CH4 and N2O Emissions (MMT CO2 Eq.)	3-28
Figure 3-16: U.S. Energy Consumption and Energy-Related CO2 Emissions Per Capita and Per Dollar GDP	3-35
Figure 4-1
Figure 4-2
Figure 5-1
Figure 5-2
Figure 5-3
Figure 5-4
2018 Industrial Processes and Product Use Chapter Greenhouse Gas Sources (MMT CO2 Eq.)	4-2
U.S. HFC Consumption (MMT C02 Eq.)	4-132
2018 Agriculture Chapter Greenhouse Gas Emission Sources (MMT CO2 Eq.)	5-1
Annual CH4 Emissions from Rice Cultivation, 2015 (MT CO2 Eq./Year)	5-22
Sources and Pathways of N that Result in N2O Emissions from Agricultural Soil Management	5-28
Crops, 2015 Annual Direct N2O Emissions Estimated Using the Tier 3 DayCent Model (MT CO2
Eq./ha/year)	5-31
Figure 5-5: Grasslands, 2015 Annual Direct N2O Emissions Estimated Using the Tier 3 DayCent Model (MT CO2
Eq./ha/year)	5-32
Figure 5-6: Crops, 2015 Annual Indirect N2O Emissions from Volatilization Using the Tier 3 DayCent Model (MT CO2
Eq./ha/year)	5-33
Figure 5-7: Grasslands, 2015 Annual Indirect N2O Emissions from Volatilization Using the Tier 3 DayCent Model
(MTCCh Eq./ha/year)	5-33
Figure 5-8: Crops, 2015 Annual Indirect N2O Emissions from Leaching and Runoff Using the Tier 3 DayCent Model
(MTCCh Eq./ha/year)	5-34
Figure 5-9: Grasslands, 2015 Annual Indirect N2O Emissions from Leaching and Runoff Using the Tier 3 DayCent
Model (MT C02 Eq./ha/year)	5-34
Figure 6-1: 2018 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.)	6-5
Figure 6-2: Percent of Total Land Area for Each State in the General Land-Use Categories for 2018	6-13
Figure 6-3: Changes in Forest Area by Region for Forest Land Remaining Forest Land in the conterminous United
States and Alaska (1990-2018, Million Hectares)	6-27
Figure 6-4: Estimated Net Annual Changes in C Stocks for All C Pools in Forest Land Remaining Forest Land in the
Conterminous U.S. and Alaska (1990-2018, MMT C per Year)	6-31
Figure 6-5: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural Management within States,
2015, Cropland Remaining Cropland	6-56
xxiv Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-6: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural Management within States,
2015, Cropland Remaining Cropland	6-57
Figure 6-7: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural Management within States,
2015, Grassland Remaining Grassland	6-74
Figure 6-8: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural Management within States,
2015, Grassland Remaining Grassland	6-75
Figure 7-1: 2018 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.)	7-1
Figure 7-2: Methodologies Used Across the Time Series to Compile the U.S. Inventory of Emission Estimates for
MSW Landfills	7-8
Figure 7-3: Management of Municipal Solid Waste in the United States, 2017	7-16
Figure 7-4: MSW Management Trends from 1990 to 2017	7-16
Figure 7-5: Percent of Degradable Materials Diverted from Landfills from 1990 to 2017 (Percent)	7-18
Boxes
Box ES-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals, including
Relationship to EPA's Greenhouse Gas Reporting Program	1
Box ES-2: Improvements and Recalculations Relative to the Previous Inventory	6
Box ES-3: Recent Trends in Various U.S. Greenhouse Gas Emissions-Related Data	27
Box ES-4: Use of Ambient Measurements Systems for Validation of Emission Inventories	30
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: IPCC Reference Approach	1-24
Box 2-1: Methodology for Aggregating Emissions by Economic Sector	2-27
Box 2-2: Recent Trends in Various U.S. Greenhouse Gas Emissions-Related Data	2-35
Box 2-3: Sources and Effects of Sulfur Dioxide	2-38
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 CO2 Emissions from Fossil Fuel Combustion Trends	3-10
Box 3-3: Uses of Greenhouse Gas Reporting Program Data and Improvements in Reporting Emissions from
Industrial Sector Fossil Fuel Combustion	3-20
Box 3-4: Carbon Intensity of U.S. Energy Consumption	3-34
Box 3-5: Reporting of Lubricants, Waxes, and Asphalt and Road Oil Product Use in Energy Sector	3-54
Box 3-6: Carbon Dioxide Transport, Injection, and Geological Storage	3-82
Box 4-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals	4-6
Box 4-2: Industrial Process and Product Use Data from EPA's Greenhouse Gas Reporting Program	4-8
Box 5-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals	5-3
Box 5-2: Surrogate Data Method	5-24
xxv

-------
Box 5-3: Tier 1 vs. Tier 3 Approach for Estimating N2O 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	5-47
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	6-9
Box 6-2: Preliminary Estimates of Land Use in U.S. Territories	6-23
Box 6-3: CO2 Emissions from Forest Fires	6-31
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	6-60
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	7-4
Box 7-3: Nationwide Municipal Solid Waste Data Sources	7-11
Box 7-4: Overview of U.S. Solid Waste Management Trends	7-16
xxvi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 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
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..."3 The United States views this report as an opportunity
to fulfill these commitments.
This chapter summarizes the latest information on U.S. anthropogenic greenhouse gas emission trends from 1990
through 2018. To ensure that the U.S. emissions inventory is comparable to those of other UNFCCC Parties, the
estimates presented here were calculated using methodologies consistent with those recommended in the 2006
Intergovernmental Panel on Climate Change (IPCC) Guidelines for National Greenhouse Gas Inventories (IPCC
2006). The structure of this report is consistent with the UNFCCC guidelines for inventory reporting, as discussed in
Box ES-1.4
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 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
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). Subsequent
decisions by the Conference of the Parties elaborated the role of Annex I Parties in preparing national inventories. See
.
4	See .
Executive Summary ES-1

-------
sink categories and calculated using internationally-accepted methods provided by the 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 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.
f PA aho roileas s;rt r nhou^e eriiis'ion*; dava from ini'iiviciuri! Maimer ~nd •?up[i!ier-» of; erTiun fossil fuel? jnci
incnisti'ial gd-;e< through it? Greenhouse Gas Reporting Program (GHGRP).1' The GHGRP applies to direct
greenhouse gas emitters, fossil fuel suppliers, industrial gas suppliers, and facilities that inject carbon dioxide
(CO ) 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 CO 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. GHG 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).
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 (CO2), methane
(CH4), nitrous oxide (N2O), and several fluorine-containing halogenated substances. Although CO2, CH4, and N2O
occur naturally in the atmosphere, human activities have changed their atmospheric concentrations. From the pre-
industrial era (i.e., ending about 1750) to 2018, concentrations of these greenhouse gases have increased globally
by 46,165, and 23 percent, respectively (IPCC 2013; NOAA/ESRL 2019a, 2019b, 2019c). This annual report
5	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).
6	See  and .
ES-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).7 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 CO2 (IPCC 2013). Therefore
GWP-weighted emissions are provided in million metric tons of CO2 equivalent (MMT CO2 Eq.).8,9 Estimates for all
gases in this Executive Summary are presented in units of MMT CO2 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).10 All estimates are provided throughout the report in both CO2 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-125
3,500
HFC-134a
1,430
HFC-143a
4,470
HFC-227ea
3,220
HFC-236fa
9,810
cf4
7,390
c2f6
12,200
C3Fs
8,830
c-C4Fs
10,300
sf6
22,800
nf3
17,200
Other Fluorinated Gases
See Annex 6
7	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.
8	Carbon comprises 12/44 of carbon dioxide by weight.
9	One million metric ton is equal to 1012 grams or one teragram.
10	See .
Executive Summary ES-3

-------
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 2018, total gross U.S. greenhouse gas emissions were 6,676.6 million metric tons of carbon dioxide equivalent
(MMT CO2 Eq).11 Total U.S. emissions have increased by 3.7 percent from 1990 to 2018, down from a high of 15.2
percent above 1990 levels in 2007. Emissions increased from 2017 to 2018 by 2.9 percent (188.4 MMT CO2 Eq.).
Net emissions (including sinks) were 5,903 MMT CO2 Eq. Overall, net emissions increased 3.1 percent from 2017 to
2018 and decreased 10.2 percent from 2005 levels as shown in Table ES-2. The decline reflects many long-term
trends, including population, economic growth, energy market trends, technological changes including energy
efficiency, and energy fuel choices. Between 2017 and 2018, the increase in total greenhouse gas emissions was
largely driven by an increase in CO2 emissions from fossil fuel combustion. The increase in CO2 emissions from
fossil fuel combustion was a result of multiple factors, including increased energy use from greater heating and
cooling needs due to a colder winter and hotter summer in 2018 compared to 2017.
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 a detailed summary of
gross U.S. greenhouse gas emissions and sinks for 1990 through 2018. Note, 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.
11 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: 1990-2018

-------
Figure ES-1: Gross U.S. Greenhouse Gas Emissions by Gas (MMT CO2 Eq.)
HFCs, PFCs, SFe and NF3
Nitrous Oxide
I Net Emissions (Including Sinks)
9,000
8,000
7,000 m
6,000
O 5,000
4,000
3,000
2,000
1,000
Oi-HrNro^j-LnvorvooCT*
CT1 CT»	 01 CT*
 CT>
01	cti a> a> cn
psirsjr>jr>jfMr>jr>jr>jfNfMrMrMrMrMrMrsi
Executive Summary ES-5

-------
Figure ES-3: Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990
(1990=0, MMTCOz Eq.)
1,200
-i-HfNin^rmiDrvooCTiOT-HrMroTLn^DrvcoCTiOi-irNin^-muDrvco
OlOt<7>(T>0)O,iOlOl0^CT\CT>a»OOOOOOOOOOOOOOOOOOO
i-ii-ii-ii-ii-ii-ii-iiH*Hr>jfNrNjrNjp>jr>jrMfMrMrMPsifNrMrMrMfMfMrsifM
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. These improvements are implemented
consistently across the previous Inventory's time series (i.e., 1990 to 2017) 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
MMTCOz Eq.
•	Agricultural Soil Management (N2O)
•	Forest Land Remaining Forest Land: Changes in Forest Carbon Stocks (CO2)
•	Land Converted to Grassland: Changes in all Ecosystem Carbon Stocks (CO2)
•	Grassland Remaining Grassland: Changes in Mineral and Organic Carbon Stocks (CO2)
•	Natural Gas Systems (CH4)
•	Grassland Remaining Grassland: Changes in Mineral and Organic Carbon Stocks (CO2)
•	Land Converted to Cropland: Changes in all Ecosystem Carbon Stocks (CO2)
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 Agriculture chapter (Chapter 5), LULUCF chapter
(Chapter 6)). 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: 1990-2018

-------
Table ES-2: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
2018
CO?
5,128.3
6,131.9
5,561.7
5,412.4
5,292.3
5,253.6
5,424.9
Fossil Fuel Combustion
4,740.0
5,740.7
5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Transportation
1,469.1
1,856.1
1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
Electric Power
1,820.0
2,400.0
2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
Industrial
857.0
850.1
812.9
801.3
801.4
805.0
833.2
Residential
338.2
357.9
346.8
317.8
293.1
293.8
337.3
Commercial
228.2
226.9
232.8
245.4
232.3
232.8
246.5
U.S. Territories
27.6
49.7
41.4
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.5
139.7
120.0
127.0
113.7
123.1
134.6
Iron and Steel Production &







Metallurgical Coke Production
104.7
70.1
58.2
47.9
43.6
40.6
42.6
Cement Production
33.5
46.2
39.4
39.9
39.4
40.3
40.3
Petroleum Systems
9.6
12.2
30.5
32.6
23.0
24.5
36.8
Natural Gas Systems
32.2
25.3
29.6
29.3
29.9
30.4
35.0
Petrochemical Production
21.6
27.4
26.3
28.1
28.3
28.9
29.4
Ammonia Production
13.0
9.2
9.4
10.6
10.8
13.2
13.5
Lime Production
11.7
14.6
14.2
13.3
12.6
12.8
13.2
Incineration of Waste
8.0
12.5
10.4
10.8
10.9
11.1
11.1
Other Process Uses of Carbonates
6.3
7.6
13.0
12.2
10.5
9.9
10.0
Urea Fertilization
2.0
3.1
3.9
4.1
4.0
4.5
4.6
Carbon Dioxide Consumption
1.5
1.4
4.5
4.5
4.5
4.5
4.5
Urea Consumption for Non-







Agricultural Purposes
3.8
3.7
1.8
4.6
5.1
3.8
3.6
Liming
4.7
4.3
3.6
3.7
3.1
3.1
3.1
Ferroalloy Production
2.2
1.4
1.9
2.0
1.8
2.0
2.1
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.8
1.7
Titanium Dioxide Production
1.2
1.8
1.7
1.6
1.7
1.7
1.5
Aluminum Production
6.8
4.1
2.8
2.8
1.3
1.2
1.5
Glass Production
1.5
1.9
1.3
1.3
1.2
1.3
1.3
Zinc Production
0.6
1.0
1.0
0.9
0.9
1.0
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
1.0
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
323.2
317.7
317.2
322.2
328.9
International Bunker Fuelsb
103.5
113.1
103.4
110.9
116.6
120.1
122.1
CH4c
774.4
679.6
639.0
638.5
624.2
630.3
634.5
Enteric Fermentation
164.2
168.9
164.2
166.5
171.8
175.4
177.6
Natural Gas Systems
183.3
158.1
141.1
141.9
135.8
139.3
140.0
Landfills
179.6
131.3
112.6
111.3
108.0
107.7
110.6
Manure Management
37.1
51.6
54.3
57.9
59.6
59.9
61.7
Coal Mining
96.5
64.1
64.6
61.2
53.8
54.8
sin
Petroleum Systems
46.1
38.8
43.5
40.5
39.0
38.7
36.2
Wastewater T reatment
15.3
15.4
14.3
14.6
14.4
14.1
14.2
Executive Summary ES-7

-------
Rice Cultivation
16.0
18.0
15.4
16.2
13.5
12.8
13.3
Stationary Combustion
8.6
7.8
8.9
8.5
7.9
7.8
8.6
Abandoned Oil and Gas Wells
6.6
7.0
7.1
7.1
7.2
7.1
7.0
Abandoned Underground Coal Mines
7.2
6.6
6.3
6.4
6.7
6.4
6.2
Mobile Combustion
12.9
9.6
4.1
3.6
3.4
3.3
3.1
Composting
0.4
1.9
2.1
2.1
2.3
2.4
2.5
Field Burning of Agricultural Residues
0.3
0.4
0.4
0.4
0.4
0.4
0.4
Petrochemical Production
0.2
0.1
0.1
0.2
0.2
0.3
0.3
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
434.6
432.6
449.3
443.8
426.1
421.3
434.5
Agricultural Soil Management
315.9
313.0
349.2
348.1
329.8
327.4
338.2
Stationary Combustion
25.1
34.3
33.0
30.5
30.0
28.6
28.4
Manure Management
14.0
16.4
17.3
17.5
18.1
18.7
19.4
Mobile Combustion
42.0
37.3
19.7
18.3
17.4
16.3
15.2
AdipicAcid Production
15.2
7.1
5.4
4.3
7.0
7.4
10.3
Nitric Acid Production
12.1
11.3
10.9
11.6
10.1
9.3
9.3
Wastewater T reatment
3.4
4.4
4.8
4.8
4.9
5.0
5.0
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
1.9
2.0
2.2
2.2
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
2.0
1.9
1.7
1.5
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
0.2
0.3
0.3
Field Burning of Agricultural Residues
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Petroleum Systems
+
+
+
+
+
+
0.1
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
0.9
1.0
1.0
1.1
1.1
HFCs
46.5
128.7
166.3
170.5
170.5
172.5
171.6
Substitution of Ozone Depleting







Substancesd
0.2
108.4
160.9
165.8
167.3
166.9
167.8
HCFC-22 Production
46.1
20.0
5.0
4.3
2.8
5.2
3.3
Electronics Industry
0.2
0.2
0.3
0.3
0.3
0.4
0.4
Magnesium Production and







Processing
0.0
0.0
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.6
5.1
4.3
4.0
4.6
Electronics Industry
2.8
3.2
3.1
3.0
2.9
2.9
3.0
Aluminum Production
21.5
3.4
2.5
2.0
1.4
1.0
1.6
Substitution of Ozone Depleting







Substances
0.0
+
+
+
+
+
0.1
sf6
28.8
11.8
6.5
5.5
6.1
5.9
5.9
Electrical Transmission and







Distribution
23.2
8.4
4.8
3.8
4.1
4.1
4.1
Magnesium Production and







Processing
5.2
2.7
0.9
1.0
1.1
1.1
1.1
Electronics Industry
0.5
0.7
0.7
0.7
0.8
0.7
0.8
nf3
+
0.5
0.5
0.6
0.6
0.6
0.6
ES-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

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







nf3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total Emissions
6,437.0
7,391.8
6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
LULUCF Emissionsc
7.4
16.3
16.6
27.4
12.8
26.1
26.1
LULUCF CH4 Emissions
4.4
8.8
9.5
16.1
7.3
15.2
15.2
LULUCF N20 Emissions
3.0
7.5
7.0
11.2
5.5
10.8
10.9
LULUCF Carbon Stock Change8
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
LULUCF Sector Net Total'
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
Net Emissions (Sources and Sinks)
5,583.6
6,577.1
6,106.0
5,900.8
5,735.1
5,724.3
5,903.2
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 from Peatlands Remaining Peatlands; CH4 and N20 emissions reported for Non-C02 Emissions from
Forest Fires, Non-C02 Emissions from 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.
dSmall amounts of PFC emissions also result from this source.
e 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 CH4 and N20 emissions to the atmosphere plus net C stock changes.
Figure ES-4 illustrates the relative contribution of the direct greenhouse gases to total U.S. emissions in 2018,
weighted by global warming potential. The primary greenhouse gas emitted by human activities in the United
States was CO2, representing approximately 81.3 percent of total greenhouse gas emissions. The largest source of
CO2, and of overall greenhouse gas emissions, was fossil fuel combustion. Methane emissions (CH4) account for
nearly 10 percent of emissions and have decreased by 7 percent since 2005 and 18.1 percent since 1990. The
major sources of methane include enteric fermentation associated with domestic livestock, natural gas systems,
and decomposition of wastes in landfills. Agricultural soil management, stationary fuel combustion, manure
management, and mobile sources of fuel combustion were the major sources of N2O 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 (SFs) emissions. The electronics industry is the only
source of nitrogen trifluoride (NF3) emissions.
Executive Summary ES-9

-------
Figure ES-4: 2018 U.S. Greenhouse Gas Emissions by Gas (Percentages based on MMT CO2
Eq.)
2.7%
HFCs, PFCs, SFe and NF3 Subtotal
6.5%
NzO
9.5%
CH«
81.3%
CO2
Overall, from 1990 to 2018, total emissions of CO2 increased by 296.6 MMT CO2 Eq. (5.8 percent), while total
emissions of CH4 decreased by 140.0 MMT CO2 Eq. (18.1 percent) and emissions of N2O have remained constant
despite fluctuations throughout the time series. During the same period, aggregate weighted emissions of
hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SFs), and nitrogen trifluoride (NF3) rose
by 83.1 MMT CO2 Eq. (83.4 percent). From 1990 to 2018, HFCs increased by 125.0 MMT CO2 Eq. (268.8 percent),
PFCs decreased by 19.6 MMT CO2 Eq. (80.9 percent), SF6 decreased by 22.9 MMT CO2 Eq. (79.4 percent), and NF3
increased by 0.6 MMT CO2 Eq. (1,211.9 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 12.0
percent of total emissions in 2018. 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
CO2 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.12
Since the Industrial Revolution (i.e., about 1750), global atmospheric concentrations of CO2 have risen
approximately 46 percent (IPCC 2013; NOAA/ESRL 2019a), principally due to the combustion of fossil fuels for
12 The term "flux" is used to describe the net emissions of greenhouse gases accounting for both the emissions of C02 to and
the removals of C02 from the atmosphere. Removal of C02 from the atmosphere is also referred to as "carbon sequestration."
ES-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
energy. Globally, approximately 32,840 MMT of CO2 were added to the atmosphere through the combustion of
fossil fuels in 2017, of which the United States accounted for approximately 15 percent.13
Within the United States, fossil fuel combustion accounted for 92.8 percent of CO2 emissions in 2018. There are 25
additional sources of CO2 emissions included in the Inventory (see Figure ES-5). Although not illustrated in the
Figure ES-5, changes in land use and forestry practices can also lead to net CO2 emissions (e.g., through conversion
of forest land to agricultural or urban use) or to a net sink for CO2 (e.g., through net additions to forest biomass).
Figure ES-5: 2018 Sources of CO2 Emissions (MMT CO2 Eq.)
Fossil Fuel Combustion
Non-Energy Use of Fuels
Iron and Steel Prod. & Metallurgical Coke Prod.
Cement Production
Petroleum Systems
Natural Gas Systems
Petrochemical Production
Ammonia Production
Lime Production
Incineration of Waste
Other Process Uses of Carbonates
Urea Fertilization
Carbon Dioxide Consumption
Urea Consumption for Non-Agricultural Purposes
Liming
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Aluminum Production
Glass Production
Zinc Production
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
Abandoned Oil and Gas Wells
Magnesium Production and Processing
0	25	50	75	100	125	150
MMT COz Eq.
As the largest source of U.S. greenhouse gas emissions, CO2 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
2018, CO2 emissions from fossil fuel combustion increased from 4,740.0 MMT CO2 Eq. to 5,031.8 MMT CO2 Eq., a
6.2 percent total increase over the twenty-nine-year period. Conversely, CO2 emissions from fossil fuel combustion
decreased by 708.9 MMT CO2 Eq. from 2005 levels, a decrease of approximately 12.3 percent between 2005 and
2018. From 2017 to 2018, these emissions increased by 139.6 MMT CO2 Eq. (2.9 percent).
Historically, changes in emissions from fossil fuel combustion have been the driving factor affecting U.S. emission
trends. Changes in CO2 emissions from fossil fuel combustion are influenced by many long-term and short-term
factors. Important drivers influencing emissions levels include: (1) changes in demand for energy; and (2) a general
decline in the carbon 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
CO2 as a Portion of All
Emissions
13 Global C02 emissions from fossil fuel combustion were taken from International Energy Agency C02 Emissions from Fossil
Fuels Combustion Overview  (IEA 2019). The publication
has not yet been updated to include 2018 data.
Executive Summary ES-11

-------
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 contributing to CO2 emissions from fossil fuel combustion 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 a lack of specific consumption data for the individual end-use sectors within U.S. Territories. Fossil
fuel combustion for electric power also includes emissions of less than 0.5 MMT CO2 Eq. from geothermal-based
generation.
Figure ES-6: 2018 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT
COz Eq.)
2,500
2,000
iS" 1,500
IN
0
u
I-
1	1,000
500
0
U.S. Territories Commercial	Residential	Industrial	Electric Power Transportation
Figure ES-7 and Table ES-3 summarize CO2 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. 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.
Relative Contribution by Fuel Type
<0.05%
I Petroleum
¦	Coal
Natural Gas
¦	Geothermal
ES-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure ES-7: 2018 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion (MMT CO2
Eq.)
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

2014
2015
2016
2017
2018
Transportation
1,472.1

1,860.8

1,718.2
1,729.5
1,769.5
1,791.6
1,825.4
Combustion
1,469.1

1,856.1

1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
Electricity
3.0

4.7

4.4
4.3
4.2
4.3
4.7
Industrial
1,543.4

1,586.4

1,405.9
1,350.8
1,319.0
1,309.4
1,320.4
Combustion
857.0

850.1

812.9
801.3
801.4
805.0
833.2
Electricity
686.4

736.3

593.0
549.5
517.6
504.4
487.2
Residential
931.0

1,213.9

1,080.9
1,001.6
946.6
910.9
986.7
Combustion
338.2

357.9

346.8
317.8
293.1
293.8
337.3
Electricity
592.7

856.0

734.1
683.8
653.5
617.1
649.4
Commercial
765.9

1,029.9

938.5
908.5
866.0
839.0
858.0
Combustion
228.2

226.9

232.8
245.4
232.3
232.8
246.5
Electricity
537.7

803.0

705.6
663.0
633.6
606.2
611.5
U.S. Territories3
27.6

49.7

41.4
41.4
41.4
41.4
41.4
Total
4,740.0

5,740.7

5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Electric Power
1,820.0

2,400.0

2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
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
Transportation End-Use Sector. Transportation activities accounted for 36.3 percent of U.S. CO2 emissions from
fossil fuel combustion in 2018. The largest sources of transportation CO2 emissions in 2018 were passenger cars
(41.2 percent); freight trucks (23.2percent); light-duty trucks, which include sport utility vehicles, pickup trucks,
and minivans (17.4 percent); commercial aircraft (6.9 percent); pipelines (2.6 percent); other aircraft (2.4 percent);
rail (2.3 percent); and ships and boats (2.2 percent). Annex 3.2 presents the total emissions from all transportation
and mobile sources, including CO2, Cm, N2O, and HFCs.
In terms of the overall trend, from 1990 to 2018, total transportation CO2 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 46.1 percent from 1990 to 2018,14 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 CO2 emissions since 1990, improvements in
average new vehicle fuel economy since 2005 has slowed the rate of increase of CO2 emissions. Petroleum-based
products supplied 97.8 percent of the energy consumed for transportation, with 57.1 percent being related to
gasoline consumption in automobiles and other highway vehicles. Diesel fuel for freight trucks and jet fuel for
aircraft, accounted for 24.1 and 13.0 percent, respectively. The remaining 3.6 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 CO2 emissions, resulting both directly from the combustion of fossil fuels and
indirectly from the generation of electricity that is used by industry, accounted for 26 percent of CO2 emissions
from fossil fuel combustion in 2018. Approximately 63 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 14.4 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 20 and
17 percent, respectively, of CO2 emissions from fossil fuel combustion in 2018. The residential and commercial
sectors relied heavily on electricity for meeting energy demands, with 66 and 71 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 increased by 6 percent since 1990. Total direct and indirect
emissions from the commercial sector have increased by 12 percent since 1990.
Electric Power. The United States relies on electricity to meet a significant portion of its energy demands.
Electricity generators used 32 percent of U.S. energy from fossil fuels and emitted 35 percent of the CO2 from fossil
fuel combustion in 2018. The type of energy source used to generate electricity is the main factor influencing
emissions.15 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 2018.16 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 2018.17
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 twenty-nine-year period to represent 34 percent of electric power sector generation in 2018.
14	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.
15	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.
16	See Table 6.2 Coal Consumption by Sector of EIA (2019a).
17	Values represent electricity net generation from the electric power sector. See Table 7.2b Electricity Net Generation: Electric
Power Sector of EIA (2019a).
ES-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Across the time series, changes in electricity demand and the carbon intensity of fuels used for electric power also
have a significant impact on CO2 emissions. While CO2 emissions from the electric power sector have decreased by
approximately 3.7 percent since 1990, the carbon intensity of the electric power sector, in terms of CO2 Eq. per
QBtu input, has significantly decreased-by 13 percent-during that same timeframe. This decoupling of the level of
electric power generation and the resulting CO2 emissions is shown in Figure ES-8.
Figure ES-8: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.)
4,500
Ł>4,000
i
§ 3,500
S 3,000
0
E 2,500
CD
c
a>
u 2,000
g
Ł 1,500
u
1	1,000
LLI
500
0
Other significant CO2 trends included the following:
•	Carbon dioxide emissions from non-energy use of fossil fuels increased by 15.0 MMT CO2 Eq. (12.6
percent) from 1990 through 2018. Emissions from non-energy uses of fossil fuels were 134.6 MMT CO2
Eq. in 2018, which constituted 2.5 percent of total national CO2 emissions, approximately the same
proportion as in 1990.
•	Carbon dioxide emissions from iron and steel production and metallurgical coke production have
decreased by 62.1 MMT CO2 Eq. (59.3 percent) from 1990 through 2018, due to restructuring of the
industry, technological improvements, and increased scrap steel utilization.
•	Total C stock change (i.e., net CO2 removals) in the LULUCF sector decreased by approximately 7.1 percent
between 1990 and 2018. 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 CO2 at trapping heat in the atmosphere-by a factor of 25 based
on the IPCC Fourth Assessment Report estimate (IPCC 2007). Over the last two hundred and fifty years, the
concentration of CFU in the atmosphere increased by 165 percent (IPCC 2013; NOAA/ESRL 2019b). The main
anthropogenic sources of CFU include enteric fermentation from domestic livestock, natural gas systems, landfills,
domestic livestock manure management, coal mining, and petroleum systems (see Figure ES-9).
Nuclear Generation (Billion kWh)
Renewable Generation (Billion kWh)
Petroleum Generation (Billion kWh)
Coal Generation (Billion kWh)
Natural Gas Generation (Billion kWh)
I Total Emissions (MMT COz Eq.) [Right Axis]
3,500
3,000
2,500 Ł
u
2,000 z
(/>
c
1,500 $
E
LLI
1,000 S
o
500
Oi-irMm^-LnvorvcooN
CT»  Ci
CJ1 01 CT* O"*	O"* 0> 0) 0> Ct
OtHrMro^-Lnvor^cocriO^rMn^-Ln^rvoo
OOOOOOOOOOiHiHiHiHiHiHiHtHtH
0000000000000000000
ojfNrsiiNojfNjfNrJiNfMfMrMfMrMrsirMrMrMrM
Executive Summary ES-15

-------
Figure ES-9: 2018 Sources of ChU Emissions (MMT CO2 Eq.)
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
Ferroalloy Production
Carbide Production and Consumption
Iron and Steel Production & Metallurgical Coke Production
Incineration of Waste
< 0.5
< 0.5
< 0.5
< 0.5
< 0.5
< 0.5
CH4 as a Portion of All
Emissions
CO2
ChU
N2O
HFCs, PFCs, SFs and NFs
80 100
MMT CO2 Eq.
180
Note: LULUCF emissions 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 2018,
enteric fermentation Cm emissions were 177.6 MMT CO2 Eq. (28.0 percent of total Cm emissions), which
represents an increase of 13.4 MMT CO2 Eq. (8.2 percent) since 1990. This increase in emissions from
1990 to 2018 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 2018 with 140.0 MMT CO2 Eq. of Cm emitted into the atmosphere. Those emissions have
decreased by 43.4 MMT CO2 Eq. (23.7 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 (110.6 MMT
CO2 Eq.), accounting for 17.4 percent of total Cm emissions in 2018. From 1990 to 2018, Cm emissions
from landfills decreased by 69.0 MMT CO2 Eq. (40.0 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)
discarded in MSW landfills over the time series.18 While the amount of landfill gas collected and
18 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.
ES-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
combusted continues to increase, the rate of increase in collection and combustion no longer exceeds the
rate of additional Cm generation from the amount of organic MSW landfilled as the U.S. population
grows.
Nitrous Oxide Emissions
Nitrous oxide (N2O) 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 N2O
emissions are much lower than CO2 emissions, N2O is nearly 300 times more powerful than CO2 at trapping heat in
the atmosphere (IPCC 2007). Since 1750, the global atmospheric concentration of N2O has risen by approximately
23 percent (IPCC 2013; NOAA/ESRL 2019c). The main anthropogenic activities producing N2O in the United States
are agricultural soil management, stationary fuel combustion, manure management, fuel combustion in motor
vehicles, and adipic acid production (see Figure ES-10).
Figure ES-10: 2018 Sources of N2O Emissions (MMT CO2 Eq.)
Agricultural Soil Management
Stationary Combustion
Manure Management
Mobile Combustion
Adipic Acid Production
Nitric Acid Production
Wastewater Treatment
N2O from Product Uses
Composting
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Incineration of Waste
Electronics Industry < °-5
Field Burning of Agricultural Residues < 0,5
Petroleum Systems < 0,5
Natural Gas Systems < 0,5
MMT CO2 Eq.
NzO as a Portion of All
Emissions
< 0.5
CO2
Cl-U
N2O
HFCs, PFCs, SFe and NFa
Note: LULUCF emissions 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 N2O include the following:
•	Agricultural soils accounted for approximately 77.8 percent of N2O emissions and 5.1 percent of total
greenhouse gas emissions in the United States in 2018. Estimated emissions from this source in 2018
were 338.2 MMT CO2 Eq. Annual N2O emissions from agricultural soils fluctuated between 1990 and 2018,
although overall emissions were 7.0 percent higher in 2018 than in 1990. Year-to-year fluctuations are
largely a reflection of annual variation in weather patterns, synthetic fertilizer use, and crop production.
•	Nitrous oxide emissions from stationary combustion increased 3.3 MMT CO2 Eq. (13.1 percent) from 1990
to 2018. Nitrous oxide emissions from this source increased primarily as a result of an increase in the
number of coal fluidized bed boilers in the electric power sector.
•	Nitrous oxide emissions from mobile combustion decreased by 26.8 MMT CO2 Eq. (63.7 percent) from
1990 to 2018, primarily as a result of N2O national emission control standards and emission control
technologies for on-road vehicles.
Executive Summary ES-17

-------
HFC, PFC, SF6, 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 (SFs) 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, SFs, and NF3 are potent greenhouse gases. In addition to having very high global warming potentials,
SFs 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).
Figure ES-11: 2018 Sources of HFCs, PFCs, SFe, and NF3 Emissions (MMT CO2 Eq.)
Substitution of Ozone Depleting Substances
Electronics Industry
Electrical Transmission and Distribution
HCFC-22 Production
Aluminum Production
Magnesium Production and Processing
HFCs, PFCs, SF6, and NF3 as a Portion
of All Emissions
HFCs, PFCs, SFe and NFs
168
8 10 12
MMT CO2 Eq.
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 the substitution of ODS (e.g.,
chlorofluorocarbons [CFCs]) have been consistently increasing, from small amounts in 1990 to 167.9 MMT
CO2 Eq. in 2018. This increase was in large part the result of efforts to phase out CFCs and other ODS in
the United States. In the short term, this trend is expected to continue, and will likely continue over the
next decade as hydrochlorofluorocarbons (HCFCs), which are in use as interim substitutes in many
applications, are themselves phased out.
•	Emissions from HCFC-22 production were 3.3 MMT CO2 Eq. in 2018, a 93 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 34.9
percent from 1990 to 2018, reflecting the competing influences of industrial growth and the adoption of
emission reduction technologies. Within that time span, emissions peaked at 9.0 MMT CO2 Eq. in 1999,
the initial year of EPA's PFC Reduction/Climate Partnership for the Semiconductor Industry, but have since
declined to 4.8 MMT CO2 Eq. in 2018 (a 46.8 percent decrease relative to 1999).
ES-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
• Sulfur hexafluoride emissions from electric power transmission and distribution systems decreased by
82.4 percent (19.1 MMT CO2 Eq.) from 1990 to 2018. 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
In accordance with the UNFCCC decision to set the 2006IPCC Guidelines for National Greenhouse Gas Inventories
(IPCC 2006) as the standard for Annex I countries at the Nineteenth Conference of the Parties (UNFCCC 2014),
Figure ES-12 and Table ES-4 aggregate emissions and sinks by the sectors defined by those guidelines. Over the
twenty-nine-year period of 1990 to 2018, total emissions from the Energy, Industrial Processes and Product Use,
and Agriculture sectors grew by 209.1 MMT CO2 Eq. (3.9 percent), 30.9 MMT CO2 Eq. (9.0 percent), and 64.1 MMT
CO2 Eq. (11.6 percent), respectively. Emissions from the Waste sector decreased by 64.6 MMT CO2 Eq. (32.4
percent). Over the same period, total C sequestration in the LULUCF sector decreased by 61.1 MMT CO2 (7.1
percent decrease in total C sequestration), and CH4 and N2O emissions from the LULUCF sector increased by 18.7
MMT CO2 Eq. (254.2 percent).
Figure ES-12: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2
Eq.)
o
u
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
-1,000
Industrial Processes and Product Use
Waste
LULUCF (emissions)
Agriculture
Energy
Land Use, Land-Use Change and Forestry (LULUCF) (removals)
0\ O"*	0~v CJ"> CTv CX* O
C7\ C7\ CT* Cv CT* CTv	CT* CX* f~*>
(N(\lN(NlNrM(NrM(N(\rMlNrvJ(NfM(NlN(N
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

2014
2015
2016
2017
2018
Energy
5,338.1

6,294.4

5,704.0
5,550.1
5,421.6
5,383.8
5,547.2
Fossil Fuel Combustion
4,740.0

5,740.7

5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Natural Gas Systems
215.5

183.4

170.7
171.2
165.7
169.6
174.9
Non-Energy Use of Fuels
119.5

139.7

120.0
127.0
113.7
123.1
134.6
Petroleum Systems
55.7

51.0

74.0
73.2
62.0
63.2
73.1
Coal Mining
96.5

64.1

64.6
61.2
53.8
54.8
52.7
Executive Summary ES-19

-------
Stationary Combustion
33.7
42.1
41.8
39.0
38.0
36.4
37.0
Mobile Combustion
55.0
46.9
23.9
22.0
20.8
19.6
18.4
Incineration of Waste
8.4
12.9
10.7
11.1
11.2
11.4
11.4
Abandoned Oil and Gas Wells
6.6
7.0
7.1
7.2
7.2
7.1
7.0
Abandoned Underground Coal Mines
7.2
6.6
6.3
6.4
6.7
6.4
6.2
Industrial Processes and Product Use
345.6
366.8
380.8
377.1
370.4
370.7
376.5
Substitution of Ozone Depleting







Substances
0.2
108.5
161.0
165.8
167.3
166.9
167.9
Iron and Steel Production &







Metallurgical Coke Production
104.8
70.1
58.2
48.0
43.6
40.6
42.6
Cement Production
33.5
46.2
39.4
39.9
39.4
40.3
40.3
Petrochemical Production
21.8
27.5
26.4
28.2
28.6
29.2
29.7
Ammonia Production
13.0
9.2
9.4
10.6
10.8
13.2
13.5
Lime Production
11.7
14.6
14.2
13.3
12.6
12.8
13.2
Adipic Acid Production
15.2
7.1
5.4
4.3
7.0
7.4
10.3
Other Process Uses of Carbonates
6.3
7.6
13.0
12.2
10.5
9.9
10.0
Nitric Acid Production
12.1
11.3
10.9
11.6
10.1
9.3
9.3
Electronics Industry
3.6
4.8
4.9
5.0
5.0
4.9
5.1
Carbon Dioxide Consumption
1.5
1.4
4.5
4.5
4.5
4.5
4.5
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Electrical Transmission and







Distribution
23.2
8.4
4.8
3.8
4.1
4.1
4.1
Urea Consumption for Non-







Agricultural Purposes
3.8
3.7
1.8
4.6
5.1
3.8
3.6
HCFC-22 Production
46.1
20.0
5.0
4.3
2.8
5.2
3.3
Aluminum Production
28.3
7.6
5.4
4.8
2.7
2.3
3.0
Ferroalloy Production
2.2
1.4
1.9
2.0
1.8
2.0
2.1
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.8
1.7
Titanium Dioxide Production
1.2
1.8
1.7
1.6
1.7
1.7
1.5
Caprolactam, Glyoxal, and Glyoxylic







Acid Production
1.7
2.1
2.0
1.9
1.7
1.5
1.4
Glass Production
1.5
1.9
1.3
1.3
1.2
1.3
1.3
Magnesium Production and







Processing
5.2
2.7
1.0
1.1
1.2
1.2
1.2
Zinc Production
0.6
1.0
1.0
0.9
0.9
1.0
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
1.0
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
554.4
575.9
608.6
614.6
600.5
602.3
618.5
Agricultural Soil Management
315.9
313.0
349.2
348.1
329.8
327.4
338.2
Enteric Fermentation
164.2
168.9
164.2
166.5
171.8
175.4
177.6
Manure Management
51.1
67.9
71.6
75.4
77.7
78.5
81.1
Rice Cultivation
16.0
18.0
15.4
16.2
13.5
12.8
13.3
Urea Fertilization
2.0
3.1
3.9
4.1
4.0
4.5
4.6
Liming
4.7
4.3
3.6
3.7
3.1
3.1
3.1
Field Burning of Agricultural Residues
0.5
0.6
0.6
0.6
0.6
0.6
0.6
Waste
199.0
154.7
135.6
134.7
131.6
131.4
134.4
Landfills
179.6
131.3
112.6
111.3
108.0
107.7
110.6
Wastewater Treatment
18.7
19.8
19.1
19.3
19.2
19.1
19.2
Composting
0.7
3.5
4.0
4.0
4.3
4.6
4.7
Total Emissions3
6,437.0
7,391.8
6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
Land Use, Land-Use Change, and







Forestry
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
Forest land
(841.7)
(780.0)
(719.5)
(765.9)
(762.3)
(739.0)
(754.5)
Cropland
30.9
24.8
44.4
44.4
32.7
33.3
38.7
Grassland
2.6
(28.9)
(4.3)
(8.9)
(14.6)
(13.4)
(12.8)
Wetlands
(0.5)
(2.0)
(0.6)
(0.7)
(0.7)
(0.7)
(0.7)
Settlements
(44.7)
(28.5)
(43.0)
(44.5)
(44.1)
(44.3)
(44.2)
ES-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Net Emission (Sources and Sinks)b	5,583.6 6,577.1 6,106.0 5,900.8 5,735.1 5,724.3 5,903.2
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 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. CO2 emissions for
the period of 1990 through 2018.
In 2018, 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).
Energy-related activities are also responsible for CH4 and N2O emissions (40 percent and 10 percent of total U.S.
emissions of each gas, respectively). Overall, emission sources in the Energy chapter account for a combined 83.1
percent of total U.S. greenhouse gas emissions in 2018.
Figure ES-13: 2018 U.S. Energy Consumption by Energy Source (Percent)
Nuclear Electric Power
8.3%
Renewable Energy
11.3%
Petroleum
36.5%
Coal
13.1%
Natural Gas
30.8%
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 many non-energy-related industrial processes, which involve the
chemical or physical transformation of raw materials and can release waste gases such as CO2, Cm, N2O, and
fluorinated gases (e.g., HFC-23). These processes include iron and steel production and metallurgical coke
production, cement production, lime production, other process uses of carbonates (e.g., flux stone, flue gas
desulfurization, and glass manufacturing), ammonia production and urea consumption, petrochemical production,
aluminum production, HCFC-22 production, soda ash production and use, titanium dioxide production, ferroalloy
production, glass production, zinc production, phosphoric acid production, lead production, silicon carbide
production and consumption, nitric acid production, adipic acid production, and caprolactam production.
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).
Executive Summary ES-21

-------
These industries include electronics industry, electric power transmission and distribution, and magnesium metal
production and processing. In addition, N2O is used in and emitted by electronics industry and anesthetic and
aerosol applications, and CO2 is consumed and emitted through various end-use applications.
IPPU activities are responsible for 3.1, 0.1, and 5.9 percent of total U.S. CO2, CFU, and N2O 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 5.6 percent of U.S. greenhouse gas emissions in 2018.
Agriculture
The Agriculture chapter contains information on anthropogenic emissions from agricultural activities (except fuel
combustion, which is addressed in the Energy chapter, and some agricultural CO2, CFU, and N2O 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.
In 2018, agricultural activities were responsible for emissions of 618.5 MMT CO2 Eq., or 9.3 percent of total U.S.
greenhouse gas emissions. Methane, N2O, and CO2 were the primary greenhouse gases emitted by agricultural
activities. Methane emissions from enteric fermentation and manure management represented approximately
28.0 percent and 9.7 percent of total CH4 emissions from anthropogenic activities, respectively, in 2018.
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. N2O emissions in 2018,
accounting for 77.8 percent of total N2O 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 CO2 emissions from
anthropogenic activities.
Land Use, Land-Use Change, and Forestry
The LULUCF chapter contains emissions of CH4 and N2O, and emissions and removals of CO2 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.19 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 CO2 (C sequestration) in the United States.
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, 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 and landfilled yard trimmings
and food scraps, is a result of net tree growth and increased urban forest size, as well as long-term accumulation of
yard trimmings and food scraps carbon in landfills.
The LULUCF sector in 2018 resulted in a net increase in C stocks (i.e., net CO2 removals) of 799.6 MMT CO2 Eq.
(Table ES-5).20 This represents an offset of 12.0 percent of total (i.e., gross) greenhouse gas emissions in 2018.
Emissions of CH4 and N2O from LULUCF activities in 2018 were 26.1 MMT CO2 Eq. and represent 0.4 percent of
19	See .
20	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.
ES-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
total greenhouse gas emissions.21 Between 1990 and 2018, total C sequestration in the LULUCF sector decreased
by 7.1 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 CO2 emissions from Land Converted to Settlements.
Forest fires were the largest source of CH4 emissions from LULUCF in 2018, totaling 11.3 MMT CO2 Eq. (452 kt of
CH4). Coastal Wetlands Remaining Coastal Wetlands resulted in CH4 emissions of 3.6 MMT CO2 Eq. (144 kt of CH4).
Grassland fires resulted in CH4 emissions of 0.3 MMT CO2 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 CO2 Eq. each.
Forest fires were also the largest source of N2O emissions from LULUCF in 2018, totaling 7.5 MMT CO2 Eq. (25 kt of
N2O). Nitrous oxide emissions from fertilizer application to settlement soils in 2018 totaled to 2.4 MMT CO2 Eq. (8
kt of N2O). Additionally, the application of synthetic fertilizers to forest soils in 2018 resulted in N2O emissions of
0.5 MMT CO2 Eq. (2 kt of N2O). Grassland fires resulted in N2O emissions of 0.3 MMT CO2 Eq. (1 kt of N2O). Coastal
Wetlands Remaining Coastal Wetlands and Drained Organic Soils resulted in N2O emissions of 0.1 MMT CO2 Eq.
each (less than 0.5 kt of N2O). Peatlands Remaining Peatlands resulted in N2O emissions of less than 0.05 MMT CO2
Eq.
Carbon dioxide removals from C stock changes are presented in Table ES-5 along with CH4 and N2O 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.)
Gas/Land-Use Category
1990
2005
2014
2015
2016
2017
2018
Carbon Stock Change3
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
Forest Land Remaining Forest Land
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
Land Converted to Forest Land
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Land Converted to Cropland
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Grassland Remaining Grassland
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Land Converted to Grassland
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.4)
(4.4)
(4.4)
(4.4)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(109.6)
(116.6)
(126.6)
(126.8)
(125.7)
(125.9)
(125.9)
Land Converted to Settlements
62.9
85.0
81.4
80.1
79.4
79.3
79.3
ch4
4.4
8.8
9.5
16.1
7.3
15.2
15.2
Forest Land Remaining Forest Land:







Forest Firesb
0.9
5.0
5.6
12.2
3.4
11.3
11.3
Wetlands Remaining Wetlands: Coastal







Wetlands Remaining Coastal Wetlands
3.4
3.5
3.6
3.6
3.6
3.6
3.6
Grassland Remaining Grassland:







Grassland Firesc
0.1
0.3
0.4
0.3
0.3
0.3
0.3
Land Converted to Wetlands: Land







Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:







Drained Organic Soilsd
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
n2o
3.0
7.5
7.0
11.2
5.5
10.8
10.9
Forest Land Remaining Forest Land:







Forest Firesb
0.6
3.3
3.7
8.1
2.2
7.5
7.5
Settlements Remaining Settlements:







21 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 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
0.4
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 Emissions8
7.4
16.3
16.6
27.4
12.8
26.1
26.1
LULUCF Carbon Stock Change3
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
LULUCF Sector NetTotalh
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
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 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
c Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grassland.
d Estimates include emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
e Estimates include emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted to
Settlements.
f Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
s 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 CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
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 110.6 MMT CO2 Eq. and accounting for 82.2 percent of total
greenhouse gas emissions from waste management activities, and 17.4 percent of total U.S. CH4 emissions.22
Additionally, wastewater treatment generates emissions of 19.2 MMT CO2 Eq. and accounts for 14.3 percent of
total Waste sector greenhouse gas emissions, 2.2 percent of U.S. CH4 emissions, and 1.2 percent of U.S. N2O
emissions. Emissions of CH4 and N2O from composting are also accounted for in this chapter, generating emissions
of 2.5 MMT CO2 Eq. and 2.2 MMT CO2 Eq., respectively. Overall, emission sources accounted for in the Waste
chapter generated 134.4 MMT CChEq., or 2.0 percent of total U.S. greenhouse gas emissions in 2018.
22 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.
ES-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. While it is
important to use this characterization for consistency with UNFCCC reporting guidelines and to promote
comparability across countries, 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 2018, and Table ES-6 summarizes
emissions from each of these economic sectors.
Figure ES-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)
Electric Power Industry (Purple)
2,500
2,000
Transportation (Green)
1,500
Industry
2; 1,000
Agriculture
Commercial (Orange'
500
Residential (Blue)
ro
1—1 7—I
LT)
1—1
yo
1—1
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

2014
2015
2016
2017
2018
Transportation
1,527.1

1,973.4

1,791.6
1,800.2
1,835.6
1,852.3
1,882.6
Electric Power Industry
1,875.6

2,455.9

2,089.1
1,949.2
1,856.8
1,778.4
1,798.9
Industry
1,628.7

1,501.7

1,438.8
1,429.8
1,388.8
1,411.5
1,470.7
Agriculture
599.0

627.5

654.9
656.0
641.0
642.4
658.6
Commercial
428.7

405.1

429.4
442.5
427.0
426.8
443.3
Residential
344.7

370.1

378.6
352.0
328.3
330.2
375.9
U.S. Territories
33.3

58.0

46.6
46.6
46.6
46.6
46.6
Total Emissions
6,437.0

7,391.8

6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
LULUCF Sector Net Total3
(853.4)

(814.7)

(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
Net Emissions (Sources and Sinks)
5,583.6

6,577.1

6,106.0
5,900.8
5,735.1
5,724.3
5,903.2
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.
Executive Summary ES-25

-------
a The LULUCF Sector Net Total is the net sum of all CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
Using this categorization, emissions from transportation activities, in aggregate, accounted for the largest portion
(28.2 percent) of total U.S. greenhouse gas emissions in 2018. Electric power accounted for the second largest
portion (26.9 percent) of U.S. greenhouse gas emissions in 2018, while emissions from industry accounted for the
third largest portion (22.0 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 22.8 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 9.9 percent of U.S. emissions; unlike other economic sectors, agricultural sector
emissions were dominated by N2O 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 CO2 sequestration is
assigned to the LULUCF sector rather than the agriculture economic sector. The commercial and residential sectors
accounted for 6.6 percent and 5.6 percent of emissions, respectively, and U.S. Territories accounted for 0.7
percent of emissions; emissions from these sectors primarily consisted of CO2 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 2019a and Duffield 2006).23 These source categories include
CO2 from fossil fuel combustion and the use of limestone and dolomite for flue gas desulfurization, CO2 and N2O
from incineration of waste, CH4 and N2O 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 (28.9 percent and 28.3 percent,
respectively) in 2018. The commercial and residential sectors contributed the next largest shares of total U.S.
greenhouse gas emissions in 2018 (16.0 and 15.6 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, CO2
accounts for more than 80.6 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 2018.
Table ES-7: U.S. Greenhouse Gas Emissions by Economic Sector with Electricity-Related
Emissions Distributed (MMT CO2 Eq.)
Economic Sectors
1990
2005
2014
2015
2016
2017
2018
Industry
2,301.0
2,216.8
2,002.6
1,952.1
1,881.0
1,890.7
1,931.0
Transportation
1,530.2
1,978.3
1,796.2
1,804.6
1,839.9
1,856.7
1,887.4
Commercial
982.8
1,226.8
1,153.0
1,122.5
1,077.4
1,049.2
1,070.9
Residential
955.6
1,246.0
1,131.4
1,053.3
999.1
963.9
1,042.4
Agriculture
634.0
665.8
699.2
697.2
680.1
681.1
698.3
U.S. Territories
33.3
58.0
46.6
46.6
46.6
46.6
46.6
23 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.
ES-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Total Emissions 6,437.0

7,391.8

6,829.0 6,676.4 6,524.1 6,488.2 6,676.6
LULUCF Sector Net Total3 (853.4)

(814.7)

(723.0) (775.5) (788.9) (763.9) (773.5)
Net Emissions (Sources and Sinks) 5,583.6
6,577.1

6,106.0 5,900.8 5,735.1 5,724.3 5,903.2
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 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 (MMT CO2 Eq.)
2,500
Industry
2,000
Transportation
1,500
Commercial (Orange)
Z 1,000
Residential (Blue)
Agriculture
500
\Ł> r*s 00 o» o 1—1 rN
o o o o 1—1 1 th
0000000
m 't m 0 n co
1—I 1—I 1—I H H I
OOOOOO
Note: Emissions and removals from Land Use, Land Use Change, and Forestry are excluded from figure above. Excludes U.S.
Territories.
Box ES-3: Recent 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 0.2 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 0.7 percent 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)



Avg. Annual
Avg. Annual
Variable
1990
2005
2014 2015 2016 2017 2018 Growth Rate
Growth Rate
Executive Summary ES-27

-------







Since 1990a
Since 2005a
Greenhouse Gas Emissions'5
100
115
106
104
101
101
104
0.2%
-0.7%
Energy Usec
100
118
117
116
116
116
120
0.7%
0.1%
GDPd
100
159
181
186
189
193
199
2.5%
1.7%
Population6
100
118
127
128
129
130
131
1.0%
0.8%
a Average annual growth rate.
b GWP-weighted values.
c Energy content-weighted values (EIA 2019a).
d GDP in chained 2009 dollars (BEA 2020).
e 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
140
Population
Energy Use
120
in
x 100
.
ES-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 decreasing order of magnitude.
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: 2018 Key Categories (MMT CO2 Eq.)a
CO2 Emissions from Mobile Combustion: Road
CO2 Emissions from Stationary Combustion - Coal - Electricity Generation
Net Carbon Stock Change from Forest Land Remaining Forest Landt
CO2 Emissions from Stationary Combustion - Gas - Electricity Generation
CO2 Emissions from Stationary Combustion - Gas - Industrial
Direct N2O Emissions from Agricultural Soil Management
CO2 Emissions from Stationary Combustion - Gas - Residential
CO2 Emissions from Stationary Combustion - Oil - Industrial
CO2 Emissions from Stationary Combustion - Gas - Commercial
CO2 Emissions from Mobile Combustion: Aviation
CFI4 Emissions from Enteric Fermentation: Cattle
CH4 Emissions from Natural Gas Systems
CO2 Emissions from Non-Energy Use of Fuels
Emissions from Substitutes of ODS: Refrigeration and Air Conditioning
Net Carbon Stock Change from Settlements Remaining Settlements^
CFI4 Emissions from Landfills
Net Carbon Stock Change from Land Converted to Forest Land*1
Net Carbon Stock Change from Land Converted to Settlements^
CO2 Emissions from Stationary Combustion - Oil - Residential
Net Carbon Stock Change from Land Converted to Cropland^
CO2 Emissions from Stationary Combustion - Oil - Commercial
Fugitive Emissions from Coal Mining
Indirect N2O Emissions from Applied Nitrogen
CO2 Emissions from Mobile Combustion: Other
CO2 Emissions from Stationary Combustion - Coal - Industrial
CO2 Emissions from Iron and Steel Production & Metallurgical Coke Production
CO2 Emissions from Cement Production
CO2 Emissions from Mobile Combustion: Railways
CH4 Emissions from Petroleum Systems
CO2 Emissions from Mobile Combustion: Marine
CO2 Emissions from Petroleum Systems
CH4 Emissions from Manure Management: Cattle
CO2 Emissions from Natural Gas Systems
CO2 Emissions from Stationary Combustion - Oil - U.S. Territories
CO2 Emissions from Petrochemical Production
CH4 Emissions from Manure Management: Other Livestock
Net Carbon Stock Change from Land Converted to Grassland^
CO2 Emissions from Stationay Combustion - Oil - Electricity Generation
Net Carbon Stock Change from Cropland Remaining Cropland^
Net Carbon Stock Change from Grassland Remaining Grassland
CH4 Emissions from Abandoned Oil and Gas Wells
CH4 Emissions from Stationary Combustion - Residential
0	400 800 1,200
Emissions (MMT CO2 Eq.)
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 2006 IPCC Guidelines. The QA
process includes expert and public reviews for both the Inventory estimates and the Inventory report.
Key Categories as a Portion of All
Emissions
¦	Key Categories
¦	Key Categories LULUCF
¦	Other Categories
Executive Summary ES-29

-------
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.25 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.26 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.27 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 prioritize future work and improve overall Inventory quality. Some of the current
estimates, such as those for CO2 emissions from energy-related combustion activities, are considered to have low
uncertainties. This is because the amount of CO2 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, a lack of data increases the uncertainty or
systematic error associated with the estimates presented. 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 provided in accordance with UNFCCC reporting guidelines, a
qualitative discussion of uncertainty is presented for each source and sink category identifying specific factors
affecting the uncertainty surrounding the estimates.
26 See .
ES-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018. A summary of these estimates is provided in 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 molecular 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
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..."4 The United States views this report as an opportunity
to fulfill these commitments 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,
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
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). Subsequent
decisions by the Conference of the Parties elaborated the role of Annex I Parties in preparing national inventories. See
.
Introduction 1-1

-------
Greenhouse Gas Inventories at its Twenty-Fifth Session (Mauritius, April 2006). The 2006IPCC 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 at
the Nineteenth Conference of the Parties (Warsaw, November 11-23, 2013). This report presents information in
accordance with these guidelines.
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 and U.S. Territories.5 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 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 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 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
(CO2) underground for sequestration or other reasons and requires reporting by 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 CO2 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. GHG 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 2006 IPCC
Guidelines (e.g., higher tier methods). GHGRP data also allow EPA to disaggregate national inventory estimates
5	U.S. Territories include American Samoa, Guam, Puerto Rico, U.S. Virgin Islands, Wake Island, and other U.S. Pacific Islands.
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 .
1-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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, 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).
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,8 the U.S. Global Change Research Program (USGCRP),9 and the
National Academies of Sciences, Engineering, and Medicine (NAS).10
Greenhouse Gases
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
8	See .
9	See .
10	See .
Introduction 1-3

-------
effect is primarily a function of the concentration of water vapor, carbon dioxide (CO2), methane (CH4), nitrous
oxide (N2O), 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, CO2, CH4, N2O, and ozone (O3). 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.11 Some other fluorine-containing halogenated substances—
hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SFs), 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 (NO2), sulfur dioxide (SO2), and tropospheric (ground level) ozone (O3).
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
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, CFU, and N2O 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
CO?
ch4
n2o
sf6
cf4
Pre-industrial atmospheric concentration
280 ppm
0.700 ppm
0.270 ppm
Oppt
40 ppt
Atmospheric concentration
409 ppma
1.857 ppmb
0.331 ppmc
9.6 pptd
79 ppt0
Rate of concentration change
2.3 ppm/yrf
7 ppb/yrf'g
0.8 ppb/yrf
0.27 ppt/yrf
0.7 ppt/yrf
Atmospheric lifetime (years)
See footnote11
12.4'
121'
3,200
50,000
11 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: 1990-2018

-------
a The atmospheric C02 concentration is the 2018 annual average at the Mauna Loa, HI station (NOAA/ESRL 2019a). The
concentration in 2018 at Mauna Loa was 409 ppm. The global atmospheric C02 concentration, computed using an average of
sampling sites across the world, was 407 ppm in 2018.
b The values presented are global 2018 annual average mole fractions (NOAA/ESRL 2019b).
c The values presented are global 2018 annual average mole fractions (NOAA/ESRL 2019c).
d The values presented are global 2018 annual average mole fractions (NOAA/ESRL 2019d).
e 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 2018 (NOAA/ESRL 2019a).
The rate of concentration change for N20, SF6, and CF4 is the average rate of change between 2005 and 2011 (IPCC 2013).
s 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.
' 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 2018 and
has fluctuated between 1.5 to 3.0 ppm per year over this period (NOAA/ESRL 2019a).
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 (H2O). 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
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 (CO2). 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 CO2. Atmospheric CO2 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 409 ppmv in 2018, a 46 percent
increase (IPCC 2013; NOAA/ESRL 2019a).1213 The IPCC definitively states that "the increase of CO2... 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 CO2 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 CO2. 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
12	The pre-industrial period is considered as the time preceding the year 1750 (IPCC 2013).
13	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).
Introduction 1-5

-------
caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings
together," of which CO2 is the most important (IPCC 2013).
Methane (CHa). 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 Cm
have increased by about 165 percent since 1750, from a pre-industrial value of about 700 ppb to 1,857 ppb in
201814 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 Cm 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 CO2. Minor removal processes also include reaction with chlorine in the marine boundary
layer, a soil sink, and stratospheric reactions. Increasing emissions of Cm 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 (N2O). Anthropogenic sources of N2O 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 N2O has increased by
23 percent since 1750, from a pre-industrial value of about 270 ppb to 331 ppb in 2018,15 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).
Ozone (O3). Ozone is present in both the upper stratosphere,16 where it shields the Earth from harmful levels of
ultraviolet radiation, and at lower concentrations in the troposphere,17 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 CO2, 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).
14	This value is the global 2018 annual average mole fraction (NOAA/ESRL 2019b).
15	This value is the global 2018 annual average (NOAA/ESRL 2019c).
16	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.
17	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.
1-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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,18 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). An amendment to the Montreal Protocol
was adopted in 2016 which includes obligations for Parties to phase down the production and consumption of
HFCs.
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,
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 CO2. 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 NO2) 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.19 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 N2O. 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
18	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.
19	NOx emissions injected higher in the stratosphere, primarily from fuel combustion emissions from high altitude supersonic
aircraft, can lead to stratospheric ozone depletion.
Introduction 1-7

-------
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 carbonaceous20 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 GHG forcing" (IPCC 2013).21 Although because they remain in the atmosphere
for only days to weeks, their concentrations respond rapidly to changes in emissions.22 Not all aerosols have a
cooling effect. Current research suggests that another constituent of aerosols, black carbon, has a 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 CO2 (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 CO2, and therefore GWP-weighted emissions are measured in million metric tons of CO2 equivalent
(MMT CO2 Eq.).23 The relationship between kilotons (kt) of a gas and MMT CO2 Eq. can be expressed as follows:
MMT CO2 Eq. = Million metric tons of CO2 equivalent
20	Carbonaceous aerosols are aerosols that are comprised mainly of organic substances and forms of black carbon (or soot)
(IPCC 2013).
21	The IPCC (2013) defines high confidence as an indication of strong scientific evidence and agreement in this statement.
22	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).
23	Carbon comprises 12/44ths of carbon dioxide by weight.
Global Warming Potentials
( MMT \
Eq. = (kt of gas) x (GWP) x oqq J
MMT CO-
where,
1-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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...24
Greenhouse gases with relatively long atmospheric lifetimes (e.g., CO2, CFU, N2O, 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., SO2 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.
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-125
29
3,500
HFC-134a
14
1,430
HFC-143a
52
4,470
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
24 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).
Introduction 1-9

-------
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 CO2 radiative forcing and an improved
CO2 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
guidelines for national inventories.25 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-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
25 See .
1-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
SF6 23,900 22,800 23,500 26,087
NFs NA 17,200 16,100 17,885
1,100 700 3,287
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.
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 consistency of cross-cutting
issues in the Inventory.
Several other government agencies contribute to the collection and analysis of the underlying activity data used in
the Inventory calculations, 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). Formal and informal relationships
exist between EPA and other U.S. agencies that 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 military fuel consumption and bunker fuels. Informal relationships also exist with other U.S.
agencies to provide activity data for use in EPA's emission calculations. These 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.
Introduction 1-11

-------
Figure 1-1: National Inventory Arrangements Diagram Inventory Process Inventory Process
United States Greenhouse Gas Inventory Institutional Arrangements
1. Data Collection
Energy Data Sources
Agriculture and
LULUCF Data Sources
Industrial Processes
and Product Use Data
Sources
Waste Data Sources
2. Emissions
Calculations
U.S. Environmental
Protection Agency
Other U.S.
Government Agencies
(USF5, NOAA,
DOD, USGS, 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
Inventory Data Sources by Source and Sink Category.
Energy
Energy Information
Administration	Management
U.S. Department of Commerce Alaska Department of Natural
Agriculture
EPA Office of Land and EmergencyEPA Greenhouse Gas Reporting
- Bureau of the Census
U.S. Department of Defense -
Defense Logistics Agency
Federal Highway
Administration
EPA Acid Rain Program
EPA Office of Transportation
and Air Quality MOVES Model
Resources
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 Greenhouse Gas Reporting EPA Office of Air and Radiation
Program (GHGRP)
U.S. Department of Labor -
Mine Safety and Health
Administration
American Association of
Railroads
U.S. Department of Agriculture
(USDA) National Agricultural
Statistics Service and Agricultural
Research Service
USDA U.S. Forest Service Forest
Inventory and Analysis Program
American Public Transportation USDA Natural Resource
Association	Conservation Service (NRCS)
U.S. Department of Homeland USDA Economic Research Service
Security	(ERS)
U.S. Department of Energy and USDA Farm Service Agency (FSA)
its National Laboratories
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
Program (GHGRP)
American Chemistry Council
(ACC)
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
Waste
EPA Greenhouse Gas
Reporting Program (GHGRP)
EPA Office of Land and
Emergency Management
Data from research studies,
trade publications, and
industry associations
1-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 and ensuring
consistency and quality throughout the NIR and CRF tables. Emission calculations 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
"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 a new
methodology, gather the most appropriate activity data and emission factors (or in some cases direct emission
measurements) for the entire time series, and conduct a special category-specific review process involving relevant
experts from industry, government, and universities (see Box ES-2 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 regulations53 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.54 In the Inventory, EPA is publishing only data values that meet
53	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 .
54	Federal Register Notice on "Greenhouse Gas Reporting Program: Publication of Aggregated Greenhouse Gas Data." See pp.
79 and 110 of notice at ..
Introduction 1-13

-------
the GHGRP aggregation criteria.55 Specific uses of aggregated facility-level 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. Electronic copies of each year's
summary data, which contains 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.
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 and
quality 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.
55 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: 1990-2018

-------
QA/QC and Uncertainty
QA/QC and uncertainty analyses are guided by the QA/QC and uncertainty coordinators, who help maintain the
QA/QC plan and the overall uncertainty analysis procedures in coordination with the Inventory coordinator (see
sections on QA/QC and Uncertainty, below). These coordinators work 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 1.6 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.
Expert, Public, and UNFCCC Reviews
The compilation of the inventory is subject to 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
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.56
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 National Inventory Report and the accompanying Common Reporting Format (CRF) tables for
electronic reporting. EPA, as the National Inventory focal point and 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.57
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
56	See .
57	See < https://www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-and-sinks >.
Introduction 1-15

-------
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 2006IPCC 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. 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., Agricultural Soil Management).
1.5 Key Categories
The 2006 IPCC 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."58 This analysis can
identify source and sink categories that diverge from the overall trend in national emissions.
The 2006 IPCC Guidelines (IPCC 2006) define quantitative methods to identify key categories both in terms of
absolute level and trend, along with consideration of uncertainty. The first method, Approach 1, was implemented
to identify the key categories for the United States without considering uncertainty in its calculations. 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 assessment. This
analysis differs from Approach 1 by including each source category's uncertainty assessments (or proxies) in its
calculations and was also performed twice to include or exclude LULUCF categories.
In addition to conducting Approach 1 and 2 level and trend assessments, 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 update 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 2018)


Approach 1
Approach 2 (includes uncertainty)
2018


Level
Trend
Level
Trend
Level
Trend
Level
Trend
Emissions


Without
Without
With
With
Without
Without
With
With
(MMT
CRF Source Category
Gas
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
LULUCF
C02 Eq.)
Energy
l.A.3.b C02 Emissions










from Mobile
C02
•
•
•
•
•
•
•
•
1,521.9
Combustion: Road










See Chapter 4 Volume 1, "Methodological Choice and Identification of Key Categories" in IPCC (2006). See .
1-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
CRF Source Category
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
l.A.l C02 Emissions
from Stationary
Combustion - Coal -
Electricity Generation
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
1.A.2 C02 Emissions
from Stationary
Combustion - Coal -
Industrial
l.A.3.e C02 Emissions
from Mobile
Combustion: Other
1.A.3.C C02 Emissions
from Mobile
Combustion: Railways
1.B.2 C02 Emissions
from Petroleum
Systems
l.A.3.d C02 Emissions
from Mobile
Combustion: Marine
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
C02
Introduction 1-17

-------
CRF Source Category
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
1.B.2 C02 Emissions
from Natural Gas
Systems
1.A.5 C02 Emissions
from Stationary
Combustion - Oil -
U.S. Territories
l.A.l C02 Emissions
from Stationary
Combustion - Oil -
Electricity Generation
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
1.B.2 CH4 Emissions
from Natural Gas
Systems
l.B.l Fugitive
Emissions from Coal
Mining
1.B.2 CH4 Emissions
from Petroleum
Systems
1.B.2 CH4 Emissions
from Abandoned Oil
and Gas Wells
l.A.4.b CH4 Emissions
from Stationary
Combustion -
Residential
l.A.3.e CH4 Emissions
from Mobile
Combustion: Other
l.A.l N20 Emissions
from Stationary
Combustion - Coal -
Electricity Generation
l.A.3.b N20 Emissions
from Mobile
Combustion: Road
l.A.l N20 Emissions
from Stationary
Combustion - Gas -
Electricity Generation
C02
C02
C02
C02
C02
C02
CH4
ch4
ch4
ch4
ch4
ch4
n2o
n2o
n2o
1-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
CRF Source Category
Gas
Approach 1
Approach 2 (includes uncertainty)
2018
Emissions
(MMT
C02 Eq.)
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
1.A.2 N20 Emissions
from Stationary
Combustion -
Industrial
N20

•
2.6
Industrial Processes and Product Use
2.C.1 C02 Emissions










from Iron and Steel










Production &
C02
•
•
•
•
•
•
•
•
42.6
Metallurgical Coke










Production










2.A.1 C02 Emissions










from Cement
C02
•

•





40.3
Production










2.B.8CO2 Emissions










from Petrochemical
C02
•
•
•
•




29.4
Production










2.G SF6 Emissions










from Electrical
Transmission and
sf6
•
•
•
•

•

•
4.1
Distribution










2.B.9 HFC-23










Emissions from HCFC-
HFCs
•
•
•
•

•

•
3.3
22 Production










2.C.3 PFC Emissions










from Aluminum
PFCs
•
•

•




1.6
Production










2.F.1 Emissions from










Substitutes for Ozone
HFCs,
PFCs









Depleting Substances:
•
•
•
•
•
•
•
•
128.9
Refrigeration and Air









Conditioning










2.F.4 Emissions from










Substitutes for Ozone
Depleting Substances:
HFCs,
PFCs

•

•

•

•
19.2
Aerosols










2.F.2 Emissions from










Substitutes for Ozone
Depleting Substances:
HFCs,
PFCs

•

•




15.1
Foam Blowing Agents










2.F.3 Emissions from










Substitutes for Ozone
Depleting Substances:
HFCs,
PFCs





•


2.6
Fire Protection










2.F.5 Emissions from










Substitutes for Ozone
Depleting Substances:
HFCs,
PFCs





•


2.0
Solvents










Agriculture
3.G C02 Emissions
from Liming
C02

•
3.1
Introduction 1-19

-------


Approach 1
Approach 2 (includes uncertainty)
2018


Level Trend Level Trend
Level Trend Level Trend
Emissions


Without Without With With
Without Without With With
(MMT
CRF Source Category
Gas
LULUCF LULUCF LULUCF LULUCF
LULUCF LULUCF LULUCF LULUCF
C02 Eq.)
3.A.1 CH4 Emissions




from Enteric
ch4
• • • •
• •
171.7
Fermentation: Cattle




3.B.1 CH4 Emissions




from Manure
ch4
• • • •
• • •
35.7
Management: Cattle




3.D.1 Direct N20




Emissions from
Agricultural Soil
n2o
• •
• •
285.7
Management




3.D.2 Indirect N20




Emissions from
n2o
• • • •
• • • •
52.5
Applied Nitrogen




3.B.4 CH4 Emissions




from Manure
Management: Other
ch4
• •

26.0
Livestock




3.C CH4 Emissions
from Rice Cultivation
ch4

• •
13.3
Waste
5.A CH4 Emissions
from Landfills
ch4
• • • •
• • • •
110.6
5.D N20 Emissions




from Wastewater
n2o

•
5.0
Treatment




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




Emissions from Land
Converted to
co2
• •
• •
79.3
Settlements




4.B.2 Net C02




Emissions from Land
Converted to
co2
•
•
55.3
Cropland




4.C.1 Net C02




Emissions from
Grassland Remaining
co2

• •
11.2
Grassland




4.B.1 Net C02




Emissions from
Cropland Remaining
co2
• •
• •
(16.6)
Cropland




4.C.2 Net C02




Emissions from Land
Converted to
co2
• •
• •
(24.6)
Grassland




4.A.2 Net C02




Emissions from Land
Converted to Forest
co2
•
•
(110.6)
Land





1-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-
-2018


-------
CRF Source Category
Gas
Approach 1
Approach 2 (includes uncertainty)
2018
Emissions
(MMT
C02 Eq.)
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
4.E.1 Net C02
Emissions from
Settlements
Remaining
Settlements
4.A.1 Net C02
Emissions from Forest
Land Remaining
Forest Land
4.A.1 CH4 Emissions
from Forest Fires
4.A.1 N20 Emissions
from Forest Fires
C02
• •
• •
(125.9)
(663.2)
11.3
co2
ch4
n2o
• •
•
• •
•

7.5
Subtotal Without LULUCF
6,506.0
Total Emissions Without LULUCF
Percent of Total Without LULUCF
6,676.6
97%
Subtotal With LULUCF
5,673.6
Total Emissions With LULUCF
5,903.2
Percent of Total With LULUCF
96%
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)/or 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
Introduction 1-21

-------
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 2006IPCC 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.
1-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

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

Data

Data

Calculating

Gathering

^Documentation

k Emissions

• Obtain data in
1
'• Contact reports
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
4-J
V)
>.
o Protect cells

working/external

checkers for:
lu
• Develop

spreadsheets

o Input ranges
c
<
automatic

• Document

o Calculations
>.
L
checkers for:

assumptions

o Emission
O
+-»
o Outliers,

• Complete QA/QC

aggregation
c
CD
negative

checklists

o Trend and IEF
>
C
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
+->
IS)
automatic

docket for new

in
tU
checkers

citations

data/consistency
C
<
• Identify

• Review

with IPCC
u
spreadsheet

documentation

methodology
a
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
Introduction 1-23

-------
Box 1-3: IPCC Reference Approach
The UNFCCC reporting guidelines require countries to complete a "top-down" reference approach for
estimating CO2 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 plans have been developed. These plans 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 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; in some cases, however, estimates are based on
approximate methodologies, assumptions, and best available data. As new information becomes available, the
United States continues to improve and revise its emission estimates. Uncertainty estimates are an essential
element of a complete and transparent emissions inventory. Uncertainty information is not intended to dispute
the validity of the Inventory estimates, but to help prioritize efforts to improve the accuracy of future Inventories
and guide future decisions on methodological choice. While the U.S. Inventory calculates its emission estimates
with the highest possible accuracy, uncertainties are associated to a varying degree with the development of
emission estimates for any inventory. For some of the current estimates, such as CO2 emissions from energy-
related combustion activities, the impact of uncertainties on overall emission estimates is believed to be relatively
small. For some other limited categories of emissions, uncertainties could have a larger impact on the estimates
presented (i.e., cropland soil carbon). The UNFCCC reporting guidelines follow the recommendation in the 2006
1-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
IPCC Guidelines (IPCC 2006) and require that countries provide single point estimates for each gas and emission or
removal source category. Within the discussion of each emission source, specific factors affecting the uncertainty
associated with the estimates are discussed.
Additional research in the following areas could help reduce uncertainty in the U.S. Inventory:
•	Incorporating excluded emission and sink categories. Quantitative estimates for some of the sources and
sinks of greenhouse gas emissions are not available at this time. In particular, emissions from some land-
use activities (e.g., emissions and removals from U.S. Territories) and industrial processes are not included
in the inventory either because data are incomplete or because methodologies do not exist for estimating
emissions from these source categories. See Annex 5 of this report for a discussion of the sources of
greenhouse gas emissions and sinks excluded from this report.
•	Improving the accuracy of emission factors. Further research is needed in some cases to improve the
accuracy of emission factors used to calculate emissions from a variety of sources. For example, the
accuracy of current emission factors applied to Cm and N2O emissions from manure management are
uncertain, and research is underway to improve these emission factors per planned improvements noted
in Section 5.2 on Manure Management.
•	Collecting detailed activity data. Although methodologies exist for estimating emissions for some sources,
problems arise in obtaining activity data at a level of detail where more technology or process-specific
emission factors can be applied.
The overall uncertainty estimate for total U.S. greenhouse gas emissions was developed using the IPCC Approach 2
uncertainty estimation methodology. The IPCC provides good practice guidance on two approaches—Approach 1
and Approach 2—to estimating uncertainty for individual source categories. Approach 2 uncertainty analysis,
employing the Monte Carlo Stochastic Simulation technique, was applied wherever data and resources permitted.
See Annex 7 of this report for further details on the U.S. process for estimating uncertainty associated with the
emission estimates Consistent with good practices in the 2006 IPCC Guidelines (IPCC 2006), over a multi-year
timeframe, the United States expects to continue to improve the uncertainty estimates presented in this report,
prioritizing key categories.
Estimates of quantitative uncertainty for the total U.S. greenhouse gas emissions in 1990 (base year) and 2018 are
shown below in Table 1-5 and Table 1-6, respectively. The overall uncertainty surrounding the Total Net GHG
Emissions is estimated to be -6 to +8 percent in 1990 and -4 to +6 percent in and 2018. When the LULUCFsector is
excluded from the analysis the uncertainty is estimated to be -2 to +4 percent in 1990 and -2 to +5 percent in 2018.
Table 1-5: Estimated Overall Inventory Quantitative Uncertainty for 1990 (MMT CO2 Eq. and
Percent)
1990 Emission	Standard
Estimate Uncertainty Range Relative to Emission Estimate3	Mean'5 Deviation'5
Gas (MMTCOz
Eq.) (MMT C02 Eq.) (%)	(MMT C02 Eq.)


Lower
Upper
Lower
Upper




Boundc
Boundc
Bound
Bound


co2
5,128.3
5,017.9
5,339.6
-2%
4%
5,178.4
82.9
CH4d
774.4
720.6
868.7
-7%
12%
793.2
37.6
N2Od
434.6
344.3
549.3
-21%
26%
431.7
51.9
PFC, HFC, SF6, and NF3d
131.6
126.6
140.6
-4%
7%
133.4
3.6
Total
6,468.9
6,334.2
6,743.5
-2%
4%
6,536.7
106.0
LULUCF Emissions6
7.4
5.8
8.8
-22%
19%
7.2
0.8
LULUCF Total Net Fluxf
(860.7)
(1,178.4)
(527.7)
37%
-39%
(853.8)
164.4
LULUCF Sector Total6
(853.4)
(1,171.5)
(520.7)
37%
-39%
(846.6)
164.4
Net Emissions (Sources and Sinks)
5,615.6
5,305.2
6,075.2
-6%
8%
5,690.1
197.0
Introduction 1-25

-------
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 1990. The base year for uncertainty is 1995 for Substitution of Ozone
Depleting Substances.
e LULUCF emissions include the CH4 and N20 emissions reported for Non-C02 Emissions from Forest Fires, Emissions from
Drained Organic Soils, N20 Fluxes from Forest Soils, Non-C02 Emissions from Grassland Fires, N20 Fluxes from Settlement Soils,
Coastal Wetlands Remaining Coastal Wetlands, Peatlands Remaining Peatlands, and CH4 Emissions from Land Converted to
Coastal Wetlands.
f Net C02 flux 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, Changes in Organic Soils Carbon Stocks, Changes in Urban Tree Carbon Stocks, Changes in Yard Trimmings and Food
Scrap Carbon Stocks in Landfills, Land Converted to Settlements, Wetlands Remaining Wetlands, and Land Converted to
Wetlands.
s The LULUCF Sector Total is the net sum of all emissions (i.e., sources) of greenhouse gases to the atmosphere plus removals
of C02 (i.e., sinks or negative emissions) from the atmosphere.
Table 1-6: Estimated Overall Inventory Quantitative Uncertainty for 2018 (MMT CO2 Eq. and
Percent)

2018 Emission





Standard

Estimate
Uncertainty Range Relative to Emission Estimate3
Mean'5
Deviationb
Gas
(MMT C02







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

(MMT C02 Eq.)


Lower
Upper
Lower
Upper




Boundc
Boundc
Bound
Bound


C02
5,424.9
5,303.2
5,661.3
-2%
4%
5,479.0
90.1
CH4d
634.5
603.2
723.8
-5%
14%
662.1
30.9
N2Od
434.5
335.0
552.1
-23%
27%
429.1
54.8
PFC, HFC, SF6, and NF3d
182.7
181.2
200.7
-1%
10%
190.7
5.1
Total
6,676.6
6,550.7
6,985.0
-2%
5%
6,760.9
110.4
LULUCF Emissions6
26.1
22.3
32.8
-14%
25%
27.5
2.7
LULUCF Total Net Fluxf
(799.6)
(1,061.7)
(597.1)
33%
-25%
(829.6)
118.3
LULUCF Sector Totals
(773.5)
(1,034.5)
(569.3)
34%
-26%
(802.1)
118.3
Net Emissions (Sources and Sinks)
5,903.1
5,642.0
6,284.3
-4%
6%
5,958.8
163.1
+ 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 2018.
e LULUCF emissions include the CH4 and N20 emissions reported for Non-C02 Emissions from Forest Fires, Emissions from
Drained Organic Soils, N20 Fluxes from Forest Soils, Non-C02 Emissions from Grassland Fires, N20 Fluxes from Settlement Soils,
Coastal Wetlands Remaining Coastal Wetlands, Peatlands Remaining Peatlands, and CH4 Emissions from Land Converted to
1-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Coastal Wetlands.
f Net C02 flux 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, Changes in Organic Soils Carbon Stocks, Changes in Urban Tree Carbon Stocks, Changes in Yard Trimmings and Food
Scrap Carbon Stocks in Landfills, Land Converted to Settlements, Wetlands Remaining Wetlands, and Land Converted to
Wetlands.
g The LULUCF Sector Total is the net sum of all emissions (i.e., sources) of greenhouse gases to the atmosphere plus removals of
C02 (i.e., sinks or negative emissions) from the atmosphere.
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.
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 2018. 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
unavailable to develop an estimate and/or the categories were determined to be insignificant59 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 and seeking
to find the data required to estimate related emissions and removals. As such improvements are implemented,
new emission and removal estimates are quantified and included in the Inventory, focusing on categories that are
significant. 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.
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 1-7. In addition,
chapters on Trends in Greenhouse Gas Emissions and Other information to be considered as part of the U.S.
Inventory submission are included.
59 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."
1.8 Completeness

1.9 Organization of Report
Introduction 1-27

-------
Table 1-7: 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 emission 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.
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 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 CO2 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 1-8.
Table 1-8: 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
1-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
8.3.	Assessment Factors
8.4.	Responses During the Review Process
ANNEX 9 Use of Greenhouse Gas Reporting Program (GHGRP) in Inventory	
Introduction 1-29

-------
2. Trends in Greenhouse Gas Emissions
2.1 Recent Trends in U.S. Greenhouse Gas
Emissions and Sinks
In 2018, total gross U.S. greenhouse gas emissions were 6,676.6 million metric tons carbon dioxide equivalent
(MMT CO2 Eq).1 Total U.S. emissions have increased by 3.7 percent from 1990 to 2018, down from a high of 15.2
percent above 1990 levels in 2007. Emissions increased from 2017 to 2018 by 2.9 percent (188.4 MMT CO2 Eq.).
Net emissions (i.e., including sinks) were 5,903 MMT CO2 Eq. Overall, net emissions increased 3.1 percent from
2017 to 2018 and decreased 10.2 percent from 2005 levels as shown in Table 2-1. The decline reflects many long-
term trends, including population, economic growth, energy market trends, technological changes including
energy efficiency, and energy fuel choices. Between 2017 and 2018, the increase in total greenhouse gas emissions
was driven largely by an increase in CO2 emissions from fossil fuel combustion. The increase in CO2 emissions from
fossil fuel combustion was a result of multiple factors, including increased energy consumption from greater
heating and cooling needs due to a colder winter and hotter summer in 2018 compared to 2017.
Since 1990, U.S. emissions have increased at an average annual rate of 0.2 percent. 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: Gross U.S. Greenhouse Gas Emissions by Gas
9,000
8,000
HFCs, PFCs, SFe and NFs
Nitrous Oxide
I Methane
Carbon Dioxide
I Net Emissions (Including Sinks)
7,000
6,000
O 5,000
4,000
3.000
2,000
1,000
t ) iH T—I T—I	I	I	I iH T—I T—I
Figure 2-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the
Previous Year
1.8%
1.7% 1-3%
-0.5%
¦1.0%
-2.2%
-2.3%
-6.3%
fNfNfNfNjrMrNrsjrvjrsjrsjfNfNfNfNfNfNj
2-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 2-3: Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990
(1990=0, MMT COz Eq.)
Overall, from 1990 to 2018, total emissions of CO2 increased by 296.6 MMT CO2 Eq. (5.8 percent), while total
emissions of methane (CH4) decreased by 140.0MMT CO2 Eq. (18.1 percent), and total emissions of nitrous oxide
(N2O) remained constant despite fluctuations throughout the time series. During the same period, aggregate
weighted emissions of hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SFs), and nitrogen
trifluoride (NF3) rose by 83.1 MMT CO2 Eq. (83.4 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 12.0 percent (799.6 MMT CO2 Eq.) of total emissions in 2018.
Table 2-1 summarizes emissions and sinks from all U.S. anthropogenic sources in weighted units of MMT CO2 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

2014
2015
2016
2017
2018
C02
5,128.3
6,131.9

5,561.7
5,412.4
5,292.3
5,253.6
5,424.9
Fossil Fuel Combustion
4,740.0
5,740.7

5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Transportation
1,469.1
1,856.1

1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
Electric Power
1,820.0
2,400.0

2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
Industrial
857.0
850.1

812.9
801.3
801.4
805.0
833.2
Residential
338.2
357.9

346.8
317.8
293.1
293.8
337.3
Commercial
228.2
226.9

232.8
245.4
232.3
232.8
246.5
U.S. Territories
27.6
49.7

41.4
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.5
139.7

120.0
127.0
113.7
123.1
134.6
Iron and Steel Production &








Metallurgical Coke Production
104.7
70.1

58.2
47.9
43.6
40.6
42.6
Cement Production
33.5
46.2

39.4
39.9
39.4
40.3
40.3
Petroleum Systems
9.6
12.2

30.5
32.6
23.0
24.5
36.8
Natural Gas Systems
32.2
25.3

29.6
29.3
29.9
30.4
35.0
Petrochemical Production
21.6
27.4

26.3
28.1
28.3
28.9
29.4
Ammonia Production
13.0
9.2

9.4
10.6
10.8
13.2
13.5
Lime Production
11.7
14.6

14.2
13.3
12.6
12.8
13.2
Incineration of Waste
8.0
12.5

10.4
10.8
10.9
11.1
11.1
Other Process Uses of Carbonates
6.3
7.6

13.0
12.2
10.5
9.9
10.0
Urea Fertilization
2.0
3.1

3.9
4.1
4.0
4.5
4.6
Trends 2-3

-------
Carbon Dioxide Consumption
Urea Consumption for Non-
Agricultural Purposes
Liming
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Aluminum Production
Glass Production
Zinc Production
Phosphoric Acid Production
Lead Production
Carbide Production and
Consumption
Abandoned Oil and Gas Wells
Magnesium Production and
Processing
Wood Biomass, Ethanol, and
Biodiesel Consumptiona
International Bunker Fuelsb
CH4c
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
Ferroalloy Production
Carbide Production and
Consumption
Iron and Steel Production &
Metallurgical Coke Production
Incineration of Waste
International Bunker Fuelsb
N2Oc
Agricultural Soil Management
Stationary Combustion
Manure Management
Mobile Combustion
Adipic Acid Production
Nitric Acid Production
Wastewater Treatment
N20 from Product Uses
Composting
1.5
1.4
4.5
4.5
4.5
4.5
4.5
3.8
3.7
1.8
4.6
5.1
3.8
3.6
4.7
4.3
3.6
3.7
3.1
3.1
3.1
2.2
1.4
1.9
2.0
1.8
2.0
2.1
1.4
1.7
1.7
1.7
1.7
1.8
1.7
1.2
1.8
1.7
1.6
1.7
1.7
1.5
6.8
4.1
2.8
2.8
1.3
1.2
1.5
1.5
1.9
1.3
1.3
1.2
1.3
1.3
0.6
1.0
1.0
0.9
0.9
1.0
1.0
1.5
1.3
1.0
1.0
1.0
1.0
0.9
0.5
0.6
0.5
0.5
0.5
0.5
0.5
0.4
0.2
0.2
0.2
0.2
0.2
0.2
+
+
+
+
+
+
+
+
+
+
+
+
+
+
219.4
230.7
323.2
317.7
317.2
322.2
328.9
103.5
113.1
103.4
110.9
116.6
120.1
122.1
774.4
679.6
639.0
638.5
624.2
630.3
634.5
164.2
168.9
164.2
166.5
171.8
175.4
177.6
183.3
158.1
141.1
141.9
135.8
139.3
140.0
179.6
131.3
112.6
111.3
108.0
107.7
110.6
37.1
51.6
54.3
57.9
59.6
59.9
61.7
96.5
64.1
64.6
61.2
53.8
54.8
sin
46.1
38.8
43.5
40.5
39.0
38.7
36.2
15.3
15.4
14.3
14.6
14.4
14.1
14.2
16.0
18.0
15.4
16.2
13.5
12.8
13.3
8.6
7.8
8.9
8.5
7.9
7.8
8.6
6.6
7.0
7.1
7.1
7.2
7.1
7.0
7.2
6.6
6.3
6.4
6.7
6.4
6.2
12.9
9.6
4.1
3.6
3.4
3.3
3.1
0.4
1.9
2.1
2.1
2.3
2.4
2.5
0.3
0.4
0.4
0.4
0.4
0.4
0.4
0.2
0.1
0.1
0.2
0.2
0.3
0.3
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
0.2
0.1
0.1
0.1
0.1
0.1
0.1
434.6
432.6
449.3
443.8
426.1
421.3
434.5
315.9
313.0
349.2
348.1
329.8
327.4
338.2
25.1
34.3
33.0
30.5
30.0
28.6
28.4
14.0
16.4
17.3
17.5
18.1
18.7
19.4
42.0
37.3
19.7
18.3
17.4
16.3
15.2
15.2
7.1
5.4
4.3
7.0
7.4
10.3
12.1
11.3
10.9
11.6
10.1
9.3
9.3
3.4
4.4
4.8
4.8
4.9
5.0
5.0
4.2
4.2
4.2
4.2
4.2
4.2
4.2
0.3
1.7
1.9
1.9
2.0
2.2
2.2
2-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Caprolactam, Glyoxal, and Glyoxylic
Acid Production
1.7
2.1
2.0
1.9
1.7
1.5
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
0.2
0.3
0.3
Field Burning of Agricultural







Residues
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Petroleum Systems
+
+
+
+
+
+
0.1
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
0.9
1.0
1.0
1.1
1.1
HFCs
46.5
128.7
166.3
170.5
170.5
172.5
171.6
Substitution of Ozone Depleting







Substancesd
0.2
108.4
160.9
165.8
167.3
166.9
167.8
HCFC-22 Production
46.1
20.0
5.0
4.3
2.8
5.2
3.3
Electronics Industry
0.2
0.2
0.3
0.3
0.3
0.4
0.4
Magnesium Production and







Processing
0.0
0.0
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.6
5.1
4.3
4.0
4.6
Electronics Industry
2.8
3.2
3.1
3.0
2.9
2.9
3.0
Aluminum Production
21.5
3.4
2.5
2.0
1.4
1.0
1.6
Substitution of Ozone Depleting







Substancesd
0.0
+
+
+
+
+
0.1
sf6
28.8
11.8
6.5
5.5
6.1
5.9
5.9
Electrical Transmission and







Distribution
23.2
8.4
4.8
3.8
4.1
4.1
4.1
Magnesium Production and







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







and NF3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total Emissions
6,437.0
7,391.8
6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
LULUCF Emissionsc
7.4
16.3
16.6
27.4
12.8
26.1
26.1
LULUCF CH4 Emissions
4.4
8.8
9.5
16.1
7.3
15.2
15.2
LULUCF N20 Emissions
3.0
7.5
7.0
11.2
5.5
10.8
10.9
LULUCF Carbon Stock Change8
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
LULUCF Sector Net Total'
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
Net Emissions (Sources and Sinks)
5,583.6
6,577.1
6,106.0
5,900.8
5,735.1
5,724.3
5,903.2
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.
e 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 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
2014
2015
2016
2017
2018
co2
5,128,301
6,131,893
5,561,719
5,412,432
5,292,268
5,253,606
5,424,882
Fossil Fuel Combustion
4,740,006
5,740,660
5,184,776
5,031,762
4,942,421
4,892,234
5,031,813
Transportation
1,469,092
1,856,113
1,713,722
1,725,274
1,765,307
1,787,274
1,820,656
Electric Power
1,819,951
2,399,974
2,037,148
1,900,624
1,808,863
1,732,025
1,752,849
Industrial
857,009
850,072
812,899
801,260
801,422
805,006
833,207
Residential
338,209
357,934
346,811
317,798
293,148
293,818
337,251
Commercial
228,191
226,867
232,835
245,439
232,320
232,756
246,493
U.S. Territories
27,555
49,700
41,361
41,367
41,362
41,355
41,357
Non-Energy Use of Fuels
119,530
139,707
120,030
127,027
113,651
123,133
134,576
Iron and Steel Production &







Metallurgical Coke







Production
104,734
70,081
58,187
47,944
43,624
40,576
42,600
Cement Production
33,484
46,194
39,439
39,907
39,439
40,324
40,324
Petroleum Systems
9,630
12,163
30,536
32,644
22,980
24,472
36,814
Natural Gas Systems
32,174
25,291
29,620
29,334
29,862
30,365
34,972
Petrochemical Production
21,611
27,383
26,254
28,062
28,310
28,910
29,424
Ammonia Production
13,047
9,196
9,377
10,634
10,838
13,216
13,532
Lime Production
11,700
14,552
14,210
13,342
12,630
12,833
13,223
Incineration of Waste
7,951
12,469
10,435
10,756
10,919
11,111
11,113
Other Process Uses of







Carbonates
6,297
7,644
12,954
12,182
10,505
9,935
9,954
Urea Fertilization
2,011
3,150
3,923
4,082
4,041
4,514
4,598
Carbon Dioxide Consumption
1,472
1,375
4,471
4,471
4,471
4,471
4,471
Urea Consumption for Non-







Agricultural Purposes
3,784
3,653
1,807
4,578
5,132
3,769
3,628
Liming
4,667
4,349
3,609
3,737
3,081
3,080
3,147
Ferroalloy Production
2,152
1,392
1,914
1,960
1,796
1,975
2,063
Soda Ash Production
1,431
1,655
1,685
1,714
1,723
1,753
1,714
Titanium Dioxide Production
1,195
1,755
1,688
1,635
1,662
1,688
1,541
Aluminum Production
6,831
4,142
2,833
2,767
1,334
1,205
1,451
Glass Production
1,535
1,928
1,336
1,299
1,241
1,296
1,263
Zinc Production
632
1,030
956
933
925
1,009
1,009
Phosphoric Acid Production
1,529
1,342
1,037
999
998
1,028
940
Lead Production
516
553
459
473
500
513
513
Carbide Production and







Consumption
375
219
173
180
174
186
189
Abandoned Oil and Gas Wells
6
7
7
7
7
7
7
Magnesium Production and







Processing
1
3
2
3
3
3
1
Wood Biomass, Ethanol, and







Biodiesel Consumptiona
219,413
230,700
323,187
317,742
317,191
322,225
328,938
International Bunker Fuelsb
103,463
113,139
103,400
110,887
116,594
120,107
122,088
CH4c
30,976
27,182
25,560
25,539
24,970
25,212
25,378
Enteric Fermentation
6,566
6,755
6,567
6,660
6,874
7,016
7,103
Natural Gas Systems
7,332
6,324
5,643
5,674
5,433
5,570
5,598
Landfills
7,182
5,253
4,503
4,452
4,322
4,308
4,422
Manure Management
1,485
2,062
2,172
2,316
2,385
2,395
2,467
Coal Mining
3,860
2,565
2,583
2,449
2,154
2,191
2,109
2-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Petroleum Systems
1,844
1,553
1,739
1,622
1,559
1,548
1,449
Wastewater Treatment
614
618
573
583
575
566
569
Rice Cultivation
640
720
616
648
539
510
533
Stationary Combustion
344
313
355
340
318
312
346
Abandoned Oil and Gas Wells
263
278
284
286
289
282
281
Abandoned Underground







Coal Mines
288
264
253
256
268
257
247
Mobile Combustion
518
383
166
146
138
131
126
Composting
15
75
84
85
91
98
98
Field Burning of Agricultural







Residues
14
16
16
16
16
16
16
Petrochemical Production
9
3
5
7
10
10
12
Ferroalloy Production
1
+
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
3
4
4
4
4
N2Oc
1,458
1,452
1,508
1,489
1,430
1,414
1,458
Agricultural Soil Management
1,060
1,050
1,172
1,168
1,107
1,099
1,135
Stationary Combustion
84
115
111
102
101
96
95
Manure Management
47
55
58
59
61
63
65
Mobile Combustion
141
125
66
62
58
55
51
Adipic Acid Production
51
24
18
14
23
25
35
Nitric Acid Production
41
38
37
39
34
31
31
Wastewater Treatment
11
15
16
16
16
17
17
N20 from Product Uses
14
14
14
14
14
14
14
Composting
1
6
6
6
7
7
7
Caprolactam, Glyoxal, and







Glyoxylic Acid Production
6
7
7
6
6
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
3
4
4
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
0
+
+
+
+
+
+
sf6
1
1
+
+
+
+
+
Electrical Transmission and







Distribution
1
+
+
+
+
+
+
Trends 2-7

-------
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
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 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 twenty-nine-
year period of 1990 to 2018, total emissions from the Energy, Industrial Processes and Product Use, and
Agriculture sectors grew by 209.1 MMT CO2 Eq. (3.9 percent), 30.9 MMT CO2 Eq. (9.0 percent), and 64.1 MMT CO2
Eq. (11.6 percent), respectively. Emissions from the Waste sector decreased by 64.6 MMT CO2 Eq. (32.4 percent).
Over the same period, total C sequestration in the Land Use, Land-Use Change, and Forestry (LULUCF) sector
decreased by 61.1 MMT CO2 (7.1 percent decrease in total C sequestration), and emissions from the LULUCF sector
increased by 18.7 MMT CO2 Eq. (254.2 percent).
Figure 2-4: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2
Eq.)
7,000
6,000
5,000
w 4(000
Ol
o
^ 3,000
2,000
1,000
0
-1,000
Industrial Processes and Product Use
Waste
LULUCF (emissions)
Agriculture
Energy
CNJ
CT* CT\
CT* CTv
CT» 
-------
Fossil Fuel Combustion
Natural Gas Systems
Non-Energy Use of Fuels
Petroleum Systems
Coal Mining
Stationary Combustion
Mobile Combustion
Incineration of Waste
Abandoned Oil and Gas Wells
Abandoned Underground Coal Mines
Industrial Processes and Product Use
Substitution of Ozone Depleting
Substances
Iron and Steel Production &
Metallurgical Coke Production
Cement Production
Petrochemical Production
Ammonia Production
Lime Production
Adipic Acid Production
Other Process Uses of Carbonates
Nitric Acid Production
Electronics Industry
Carbon Dioxide Consumption
N20 from Product Uses
Electrical Transmission and
Distribution
Urea Consumption for Non-
Agricultural Purposes
HCFC-22 Production
Aluminum Production
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic
Acid Production
Glass Production
Magnesium Production and
Processing
Zinc Production
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
Agriculture
Agricultural Soil Management
Enteric Fermentation
Manure Management
Rice Cultivation
Urea Fertilization
Liming
Field Burning of Agricultural Residues
Waste
Landfills
Wastewater Treatment
Composting
4,740.0
5,740.7
5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
215.5
183.4
170.7
171.2
165.7
169.6
174.9
119.5
139.7
120.0
127.0
113.7
123.1
134.6
55.7
51.0
74.0
73.2
62.0
63.2
73.1
96.5
64.1
64.6
61.2
53.8
54.8
52.7
33.7
42.1
41.8
39.0
38.0
36.4
37.0
55.0
46.9
23.9
22.0
20.8
19.6
18.4
8.4
12.9
10.7
11.1
11.2
11.4
11.4
6.6
7.0
7.1
7.2
7.2
7.1
7.0
7.2
6.6
6.3
6.4
6.7
6.4
6.2
345.6
366.8
380.8
377.1
370.4
370.7
376.5
0.2
108.5
161.0
165.8
167.3
166.9
167.9
104.8
70.1
58.2
48.0
43.6
40.6
42.6
33.5
46.2
39.4
39.9
39.4
40.3
40.3
21.8
27.5
26.4
28.2
28.6
29.2
29.7
13.0
9.2
9.4
10.6
10.8
13.2
13.5
11.7
14.6
14.2
13.3
12.6
12.8
13.2
15.2
7.1
5.4
4.3
7.0
7.4
10.3
6.3
7.6
13.0
12.2
10.5
9.9
10.0
12.1
11.3
10.9
11.6
10.1
9.3
9.3
3.6
4.8
4.9
5.0
5.0
4.9
5.1
1.5
1.4
4.5
4.5
4.5
4.5
4.5
4.2
4.2
4.2
4.2
4.2
4.2
4.2
23.2
8.4
4.8
3.8
4.1
4.1
4.1
3.8
3.7
1.8
4.6
5.1
3.8
3.6
46.1
20.0
5.0
4.3
2.8
5.2
3.3
28.3
7.6
5.4
4.8
2.7
2.3
3.0
2.2
1.4
1.9
2.0
1.8
2.0
2.1
1.4
1.7
1.7
1.7
1.7
1.8
1.7
1.2
1.8
1.7
1.6
1.7
1.7
1.5
1.7
2.1
2.0
1.9
1.7
1.5
1.4
1.5
1.9
1.3
1.3
1.2
1.3
1.3
5.2
2.7
1.0
1.1
1.2
1.2
1.2
0.6
1.0
1.0
0.9
0.9
1.0
1.0
1.5
1.3
1.0
1.0
1.0
1.0
0.9
0.5
0.6
0.5
0.5
0.5
0.5
0.5
0.4
0.2
0.2
0.2
0.2
0.2
0.2
554.4
575.9
608.6
614.6
600.5
602.3
618.5
315.9
313.0
349.2
348.1
329.8
327.4
338.2
164.2
168.9
164.2
166.5
171.8
175.4
177.6
51.1
67.9
71.6
75.4
77.7
78.5
81.1
16.0
18.0
15.4
16.2
13.5
12.8
13.3
2.0
3.1
3.9
4.1
4.0
4.5
4.6
4.7
4.3
3.6
3.7
3.1
3.1
3.1
0.5
0.6
0.6
0.6
0.6
0.6
0.6
199.0
154.7
135.6
134.7
131.6
131.4
134.4
179.6
131.3
112.6
111.3
108.0
107.7
110.6
18.7
19.8
19.1
19.3
19.2
19.1
19.2
0.7
3.5
4.0
4.0
4.3
4.6
4.7
Trends 2-9

-------
Total Emissions3
6,437.0

7,391.8

6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
Land Use, Land-Use Change, and









Forestry
(853.4)

(814.7)

(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
Forest land
(841.7)

(780.0)

(719.5)
(765.9)
(762.3)
(739.0)
(754.5)
Cropland
30.9

24.8

44.4
44.4
32.7
33.3
38.7
Grassland
2.6

(28.9)

(4.3)
(8.9)
(14.6)
(13.4)
(12.8)
Wetlands
(0.5)

(2.0)

(0.6)
(0.7)
(0.7)
(0.7)
(0.7)
Settlements
(44.7)

(28.5)

(43.0)
(44.5)
(44.1)
(44.3)
(44.2)
Net Emission (Sources and Sinks)b
5,583.6

6,577.1

6,106.0
5,900.8
5,735.1
5,724.3
5,903.2
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 Net emissions with LULUCF.
Energy
Energy-related activities, primarily fossil fuel combustion, accounted for the vast majority of U.S. CO2 emissions for
the period of 1990 through 2018. Fossil fuel combustion is the largest source of energy-related emissions, with CO2
being the primary gas emitted (see Figure 2-5). Due to their relative importance, fossil fuel combustion-related CO2
emissions are considered in detail in the Energy chapter (see Energy chapter).
In 2018, approximately 80 percent of the energy consumed 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 CO2 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 N2O emissions (40 percent and 10 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.
Figure 2-5: 2018 Energy Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
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-CC>2 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 CO2 Eq.
2-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 2-4: Emissions from Energy (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
2018
CO?
4,909.3
5,930.3
5,375.4
5,231.5
5,119.8
5,081.3
5,249.3
Fossil Fuel Combustion
4,740.0
5,740.7
5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Transportation
1,469.1
1,856.1
1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
Electric Power Sector
1,820.0
2,400.0
2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
Industrial
857.0
850.1
812.9
801.3
801.4
805.0
833.2
Residential
338.2
357.9
346.8
317.8
293.1
293.8
337.3
Commercial
228.2
226.9
232.8
245.4
232.3
232.8
246.5
U.S. Territories
27.6
49.7
41.4
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.5
139.7
120.0
127.0
113.7
123.1
134.6
Petroleum Systems
9.6
12.2
30.5
32.6
23.0
24.5
36.8
Natural Gas Systems
32.2
25.3
29.6
29.3
29.9
30.4
35.0
Incineration of Waste
8.0
12.5
10.4
10.8
10.9
11.1
11.1
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Biomass-Wood"
215.2
206.9
233.8
224.7
216.3
221.4
229.1
International Bunker Fuelsb
103.5
113.1
103.4
110.9
116.6
120.1
122.1
Biofuels-Ethanola
4.2
22.9
76.1
78.9
81.2
82.1
81.9
Biofuels-Biodiesela
0.0
0.9
13.3
14.1
19.6
18.7
17.9
ch4
361.2
292.0
275.6
269.3
253.9
257.3
253.9
Natural Gas Systems
183.3
158.1
141.1
141.9
135.8
139.3
140.0
Coal Mining
96.5
64.1
64.6
61.2
53.8
54.8
sin
Petroleum Systems
46.1
38.8
43.5
40.5
39.0
38.7
36.2
Stationary Combustion
8.6
7.8
8.9
8.5
7.9
7.8
8.6
Abandoned Oil and Gas Wells
6.6
7.0
7.1
7.1
7.2
7.1
7.0
Abandoned Underground Coal
7.2
6.6
6.3
6.4
6.7
6.4
6.2
Mines







Mobile Combustion
12.9
9.6
4.1
3.6
3.4
3.3
3.1
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
n2o
67.6
72.1
53.1
49.2
47.8
45.2
44.0
Stationary Combustion
25.1
34.3
33.0
30.5
30.0
28.6
28.4
Mobile Combustion
42.0
37.3
19.7
18.3
17.4
16.3
15.2
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Petroleum Systems
+
+
+
+
+
+
0.1
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
0.9
1.0
1.0
1.1
1.1
Total
5,338.1
6,294.4
5,704.0
5,550.1
5,421.6
5,383.8
5,547.2
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, CO2 from fossil fuel combustion has accounted for
approximately 76 percent of GWP-weighted emissions across the time series Emissions from this source category
grew by 6.2 percent (291.8 MMT CO2 Eq.) from 1990 to 2018 and were responsible for most of the increase in
national emissions during this period. Conversely, CO2 emissions from fossil fuel combustion decreased by 708.8
MMT CO2 Eq. from 2005 and by 319.2 MMT CO2 Eq. from 2010, representing decreases of approximately 12.3
percent between 2005 and 2018 and 6.0 percent between 2010 and 2018. From 2017 to 2018, these emissions
Trends 2-11

-------
increased by 2.9 percent (139.6 MMT CO2 Eq.). Historically, changes in emissions from fossil fuel combustion have
been the main factor influencing U.S. emission trends.
Changes in CO2 emissions from fossil fuel combustion are affected by many long-term and short-term factors,
including population and economic growth, energy price fluctuations and market trends, technological changes,
energy fuel choices, and seasonal temperatures. 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 CO2 emissions also depend on the type of fuel consumed or energy used and its C intensity.
Producing a unit of heat or electricity using natural gas instead of coal, for example, reduces CO2 emissions
because of the lower C content of natural gas (see Table A-42 in Annex 2.1 for more detail on the C Content
Coefficient of different fossil fuels).
Trends in CO2 emissions from fossil fuel combustion over the past five years have been strongly influenced by the
electric power sector, which historically has accounted for the largest share of emissions from this source (see
Figure 2-6). The types of fuel consumed to produce electricity have changed in recent years, impacting emission
trends. Emissions increased 1.2 percent from 2017 to 2018 due to increasing electric power generation from
natural gas and renewables and decreasing generation from coal. Carbon dioxide emissions from coal consumption
for electric power generation decreased by 26.5 percent since 2014 and 42 percent since 2005, which can be
largely attributed to a shift to the use of less-CC>2-intensive natural gas to generate electricity and a rapid increase
in the use of renewable energy in the electric power sector in recent years. Electricity generation from renewable
sources increased by 32.6 percent from 2014 to 2018 and natural gas generation increased by 32.2 percent over
the same time period (see Table 3-12 for more detail on electricity generation by source). Total electric power
generation decreased by 1.5 percent from 2014 to 2017 but increased by 3.4 percent from 2017 to 2018. The
decrease in coal-powered electricity generation and increase in natural gas and renewable energy electricity
generation have contributed to a 14.0 percent decrease in overall CO2 emissions from electric power generation
from 2014 to 2018 and a 27 percent decrease from 2005 to 2018 (see Figure 2-8).
The trends in CO2 emissions from fossil fuel combustion over the past five years also follow changes in heating
degree days. Emissions from natural gas consumption in the residential and commercial sectors increased by 13.4
percent and 11.2 percent from 2017 to 2018, respectively. This trend can be largely attributed to a 11.8 percent
increase in heating degree days, which led to an increased demand for heating fuel in these sectors. Combined
residential and commercial sector CO2 emissions increased by 12.5 percent from 2017 to 2018.
Petroleum use is another major driver of CO2 emissions from fossil fuel combustion, particularly in the
transportation sector, which represents the largest source of CO2 emissions from fossil fuel combustion in 2018.
Emissions from petroleum consumption for transportation (including bunkers) have increased by 5.8 percent since
2014; this trend can be primarily attributed to a 7.1 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 CO2 emissions.
Overall, across all sectors, there was a 2.9 percent increase in total CO2 emissions from fossil fuel combustion from
2017 to 2018 and a 3.0 percent reduction since 2014. Carbon dioxide emissions from fossil fuel combustion,
separated by end-use sector, are presented in Table 2-5 and Figure 2-6 based on the underlying U.S. energy
consumer data collected by the U.S. Energy Information Administration (EIA). Figure 2-7 further describes direct
and indirect CO2 emissions from fossil fuel combustion, separated by end-use sector. Estimates of CO2 emissions
from fossil fuel combustion are calculated from these EIA "end-use sectors" based on total fuel consumption and
2-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
2014
2015
2016
2017
2018
Transportation
1,472.1
1,860.8
1,718.2
1,729.5
1,769.5
1,791.6
1,825.4
Combustion
1,469.1
1,856.1
1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
Electricity
3.0
4.7
4.4
4.3
4.2
4.3
4.7
Industrial
1,543.4
1,586.4
1,405.9
1,350.8
1,319.0
1,309.4
1,320.4
Combustion
857.0
850.1
812.9
801.3
801.4
805.0
833.2
Electricity
686.4
736.3
593.0
549.5
517.6
504.4
487.2
Residential
931.0
1,213.9
1,080.9
1,001.6
946.6
910.9
986.7
Combustion
338.2
357.9
346.8
317.8
293.1
293.8
337.3
Electricity
592.7
856.0
734.1
683.8
653.5
617.1
649.4
Commercial
765.9
1,029.9
938.5
908.5
866.0
839.0
858.0
Combustion
228.2
226.9
232.8
245.4
232.3
232.8
246.5
Electricity
537.7
803.0
705.6
663.0
633.6
606.2
611.5
U.S. Territories3
27.6
49.7
41.4
41.4
41.4
41.4
41.4
Total
4,740.0
5,740.7
5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Electric Power
1,820.0
2,400.0
2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
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.
Trends 2-13

-------
Figure 2-i
C02 Eq.)
6:
2018 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT
iiJ 1,500
Relative Contribution by Fuel Type
<0.05%
Petroleum
Coal
Natural Gas
Geothermal
U.S. Territories
Commercial
Residential
Industrial
Electric Power Transportation
Note on Figure 2-6: Fossil Fuel Combustion for electric power also includes emissions of less than 0.5 MMT C02 Eq. from
geothermal-based generation.
Figure 2-7: 2018 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion (MMT CO2
Eq.)
2,000 g Direct Fossil Fuel Combustion
I Indirect Fossil Fuel Combustion
U.S. Territories	Commercial	Residential	Industrial	Transportation
Electric power was the second largest emitter of CO2 in 2018 (surpassed by transportation); electric power
generators used 32 percent of U.S. energy from fossil fuels and emitted 35 percent of the CO2 from fossil fuel
combustion in 2018. Changes in electricity demand and the carbon intensity of fuels used for electric power
generation have a significant impact on CO2 emissions. Carbon dioxide emissions from the electric power sector
have decreased by approximately 3.7 percent since 1990, and the carbon intensity of the electric power sector, in
terms of CO2 Eq. per QBtu input, has significantly decreased by 13 percent during that same timeframe. This
decoupling of electric power generation and the resulting CO2 emissions is shown below in Figure 2-8.
2-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 2-\
8: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.)
Nuclear Generation (Billion kWh)
Renewable Generation (Billion kWh)
Petroleum Generation (Billion kWh)
Coal Generation (Billion kWh)
Natural Gas Generation (Billion kWh)
4,500
4,000
3,500
3,000
E 2,500
« 2,000
? 1,500
m 1-000
I Total Emissions (MMT CO2 Eq.) [Right Axis]
3,500
3,000
2,500 Łj
u
2,000 s
ui
c
1,500 ft
1,000
500
J3
.0
o^-HfMn^-Lnvoiv.GOc^Oi-ifMtn^-LrikOiv
cr>cncn
-------
•	Methane emissions from coal mining decreased by 43.8 MMT CO2 Eq. (45.4 percent) from 1990 through
2018, 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 CO2 Eq. (63.7 percent) from
1990 through 2018, primarily as a result of N2O national emission control standards and emission control
technologies for on-road vehicles.
•	Carbon dioxide emissions from non-energy uses of fossil fuels increased by 15.0 MMT CO2 Eq. (12.6
percent) from 1990 through 2018. Emissions from non-energy uses of fossil fuels were 134.6 MMT CO2
Eq. in 2018, which constituted 2.5 percent of total national CO2 emissions, approximately the same
proportion as in 1990.
•	Nitrous oxide emissions from stationary combustion increased by 3.3 MMT CO2 Eq. (13.1 percent) from
1990 through 2018. Nitrous oxide emissions from this source increased primarily as a result of an increase
in the number of coal fluidized bed boilers in the electric power sector.
•	Carbon dioxide emissions from incineration of waste (11.1 MMT CO2 Eq. in 2018) increased by 3.2 MMT
CO2 Eq. (39.8 percent) from 1990 through 2018, 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 CO2, CFU,
N2O, 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 NFsand
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, N2O is used in
and emitted by the electronics industry and anesthetic and aerosol applications. Figure 2-9 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 5.6 percent of U.S. greenhouse gas
emissions in 2018.
2-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 2-9: 2018 Industrial Processes and Product Use Chapter Greenhouse Gas Sources
(MMT COz Eq.)
Substitution of Ozone Depleting Substances
Iron and Steel Production & Metallurgical Coke Production
Cement Production
Petrochemical Production
Ammonia Production
Lime Production
Adipic Acid Production
Other Process Uses of Carbonates
Nitric Acid Production
Electronics Industry
Carbon Dioxide Consumption
N2O from Product Uses
Electrical Transmission and Distribution
Urea Consumption for Non-Agricultural Purposes
HCFC-22 Production
Aluminum Production
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Glass Production
Magnesium Production and Processing
Zinc Production
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
MMT COz Eq.
Table 2-6: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)
Gas/Source
1990

2005

2014
2015
2016
2017
2018
C02
212.3

194.1

178.8
173.1
165.3
164.7
167.8
Iron and Steel Production & Metallurgical Coke









Production
104.7

70.1

58.2
47.9
43.6
40.6
42.6
Iron and Steel Production
99.1

66.2

54.5
43.5
41.0
38.6
41.3
Metallurgical Coke Production
5.6

3.9

3.7
4.4
2.6
2.0
1.3
Cement Production
33.5

46.2

39.4
39.9
39.4
40.3
40.3
Petrochemical Production
21.6

27.4

26.3
28.1
28.3
28.9
29.4
Ammonia Production
13.0

9.2

9.4
10.6
10.8
13.2
13.5
Lime Production
11.7

14.6

14.2
13.3
12.6
12.8
13.2
Other Process Uses of Carbonates
6.3

7.6

13.0
12.2
10.5
9.9
10.0
Carbon Dioxide Consumption
1.5

1.4

4.5
4.5
4.5
4.5
4.5
Urea Consumption for Non-Agricultural









Purposes
3.8

3.7

1.8
4.6
5.1
3.8
3.6
Ferroalloy Production
2.2

1.4

1.9
2.0
1.8
2.0
2.1
Soda Ash Production
1.4

1.7

1.7
1.7
1.7
1.8
1.7
Titanium Dioxide Production
1.2

1.8

1.7
1.6
1.7
1.7
1.5
Aluminum Production
6.8

4.1

2.8
2.8
1.3
1.2
1.5
Glass Production
1.5

1.9

1.3
1.3
1.2
1.3
1.3
Zinc Production
0.6

1.0

1.0
0.9
0.9
1.0
1.0
Phosphoric Acid Production
1.5

1.3

1.0
1.0
1.0
1.0
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
+

+

+
+
+
+
+
ch4
0.3

0.1

0.2
0.2
0.3
0.3
0.3
Petrochemical Production
0.2

0.1

0.1
0.2
0.2
0.3
0.3
Ferroalloy Production
+

+

+
+
+
+
+
Industrial Processes and Product Use
as a Portion of All Emissions
5.6%
I Energy
¦ Agriculture
IPPU
Waste
Trends 2-17

-------
Carbide Production and Consumption
+
+
+
+
+
+
+
Iron and Steel Production & Metallurgical Coke







Production
+
+
+
+
+
+
+
n2o
33.3
24.9
22.8
22.2
23.3
22.7
25.5
AdipicAcid Production
15.2
7.1
5.4
4.3
7.0
7.4
10.3
Nitric Acid Production
12.1
11.3
10.9
11.6
10.1
9.3
9.3
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
2.0
1.9
1.7
1.5
1.4
Electronics Industry
+
0.1
0.2
0.2
0.2
0.3
0.3
HFCs
46.5
128.7
166.3
170.5
170.5
172.5
171.6
Substitution of Ozone Depleting Substances3
0.2
108.4
160.9
165.8
167.3
166.9
167.8
HCFC-22 Production
46.1
20.0
5.0
4.3
2.8
5.2
3.3
Electronics Industry
0.2
0.2
0.3
0.3
0.3
0.4
0.4
Magnesium Production and Processing
0.0
0.0
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.6
5.1
4.3
4.0
4.6
Electronics Industry
2.8
3.2
3.1
3.0
2.9
2.9
3.0
Aluminum Production
21.5
3.4
2.5
2.0
1.4
1.0
1.6
Substitution of Ozone Depleting Substances
0.0
+
+
+
+
+
0.1
sf6
28.8
11.8
6.5
5.5
6.1
5.9
5.9
Electrical Transmission and Distribution
23.2
8.4
4.8
3.8
4.1
4.1
4.1
Magnesium Production and Processing
5.2
2.7
0.9
1.0
1.1
1.1
1.1
Electronics Industry
0.5
0.7
0.7
0.7
0.8
0.7
0.8
nf3
+
0.5
0.5
0.6
0.6
0.6
0.6
Electronics Industry
+
0.5
0.5
0.6
0.6
0.6
0.6
Unspecified Mix of HFCs, NF3, PFCs and SF6
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total
345.6
366.8
380.8
377.1
370.4
370.7
376.5
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 9.0 percent from 1990 to 2018. Significant trends in emissions
from IPPU source categories over the twenty-nine-year period from 1990 through 2018 included the following:
•	Hydrofluorocarbon and perfluorocarbon emissions resulting from the substitution of ODS (e.g.,
chlorofluorocarbons [CFCs]) have been increasing from small amounts in 1990 to 167.9 MMT CO2 Eq. in
2018. This increase was in large part the result of efforts to phase out CFCs and other ODS in the United
States. In the short term, this trend is expected to continue, and will likely continue over the next decade
as hydrochlorofluorocarbons (HCFCs), which are in use as interim substitutes in many applications, are
themselves phased-out.
•	Combined CO2 and CFU emissions from iron and steel production and metallurgical coke production
increased by 5.0 percent to 42.6 MMT CO2 Eq. from 2017 to 2018, but have declined overall by 62.1 MMT
CO2 Eq. (59.3 percent) from 1990 through 2018, 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 ammonia production (13.5 MMT CO2 Eq. in 2018) increased by 3.7 percent
(0.5 MMT CO2 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. Recent low prices for natural gas and increased demand for ammonia use in nitrogen
fertilizers has let to increases in ammonia production and emissions.
•	Carbon dioxide emissions from cement production increased by 20.4 percent (6.8 MMT CO2 Eq.) from
1990 through 2018. They rose from 1990 through 2006 and then fell until 2009 due to a decrease in
2-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
demand for construction materials during the economic recession. Since 2010, CO2 emissions from
cement production have risen 28.2 percent (8.9 MMT CO2 Eq.).
•	PFC emissions from aluminum production decreased by 92.6 percent (19.9 MMT CO2 Eq.) from 1990 to
2018, due to both industry emission reduction efforts and lower domestic aluminum production.
•	Nitrous oxide emissions from adipic acid production were 10.3 MMT CO2 Eq. in 2018, and have decreased
significantly (32.1 percent or 4.9 MMT CO2 Eq.) since 1990 due to both the widespread installation of
pollution control measures in the late 1990s and plant idling in the late 2000s.
Agriculture
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, N2O, and CO2 were the primary greenhouse gases emitted by agricultural activities.
In 2018, agricultural activities were responsible for emissions of 618.5 MMT CO2 Eq., or 9.3 percent of total U.S.
greenhouse gas emissions. Methane emissions from enteric fermentation and manure management represented
approximately 28.0 percent and 9.7 percent of total CH4 emissions from anthropogenic activities, respectively, in
2018. 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. N2O emissions in 2018,
accounting for 77.8 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 CO2 emissions from anthropogenic
activities. Figure 2-10 and Table 2-7 illustrate agricultural greenhouse gas emissions by source.
Figure 2-10: 2018 Agriculture Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
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 CO2 Eq.
Table 2-7: Emissions from Agriculture (MMT CO2 Eq.)
Gas/Source
1990

2005

2014
2015
2016
2017
2018
C02
6.7

7.5

7.5
7.8
7.1
7.6
7.7
Urea Fertilization
2.0

3.1

3.9
4.1
4.0
4.5
4.6
Liming
4.7

4.3

3.6
3.7
3.1
3.1
3.1
ch4
217.6

238.8

234.3
241.0
245.3
248.4
253.0
Enteric Fermentation
164.2

168.9

164.2
166.5
171.8
175.4
177.6
Manure Management
37.1

51.6

54.3
57.9
59.6
59.9
61.7
Rice Cultivation
16.0

18.0

15.4
16.2
13.5
12.8
13.3
Field Burning of Agricultural









Residues
0.3

0.4

0.4
0.4
0.4
0.4
0.4
Agriculture as a Portion of
All Emissions
9.3% 	
I Energy
¦ Agriculture
IPPU
Waste
Trends 2-19

-------
NzO 330.1 329.6	366.7	365.8	348.1	346.2	357.8
Agricultural Soil Management 315.9 313.0	349.2	348.1	329.8	327.4	338.2
Manure Management 14.0 16.4	17.3	17.5	18.1	18.7	19.4
Field Burning of Agricultural
Residues	02	02	02	02	02	02	0.2
Total 554.4 575.9	608.6	614.6	600.5	602.3	618.5
Note: Totals may not sum due to independent rounding.
Some significant trends in U.S. emissions from Agriculture source categories include the following:
•	Agricultural soils are the largest anthropogenic source of N2O emissions in the United States, accounting
for approximately 77.8 percent of N2O emissions in 2018 and 5.1 percent of total emissions in the United
States in 2018. Estimated emissions from this source in 2018 were 338.2 MMT CO2 Eq. Annual N2O
emissions from agricultural soils fluctuated between 1990 and 2018, although overall emissions were 22.2
MMT CO2 Eq. or 7.0 percent higher in 2018 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 2018,
enteric fermentation CFU emissions were 28.0 percent of total CFU emissions (177.6 MMT CO2 Eq.), which
represents an increase of 13.4 MMT CO2 Eq. (8.2 percent) since 1990. This increase in emissions from
1990 to 2018 in enteric fermentation generally follows the increasing trends in cattle populations. 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 2018, consistent with
an increase in beef cattle population over those same years.
•	Overall, emissions from manure management increased 58.7 percent between 1990 and 2018. This
encompassed an increase of 66.1 percent for CFU, from 37.1 MMT CO2 Eq. in 1990 to 61.7 MMT CO2 Eq. in
2018; and an increase of 39.0 percent for N2O, from 14.0 MMT CO2 Eq. in 1990 to 19.4 MMT CO2 Eq. in
2018. The majority of the increase observed in CH4 resulted from swine and dairy cattle manure, where
emissions increased 42.8 and 119.2 percent, respectively, from 1990 to 2018. From 2017 to 2018, there
was a 3.0 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 CO2 emissions reported in the Agriculture sector.
Estimated emissions from these sources were 3.1 and 4.6 MMT CO2 Eq., respectively. Liming emissions
increased by 2.2 percent relative to 2017 and decreased 1.5 MMT CO2 Eq. or 32.6 percent relative to
1990, while urea fertilization emissions increased by 1.9 percent relative to 2017 and 2.6 MMT CO2 Eq.
128.7 percent relative to 1990.
Land Use, LancHhc 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 CFU and N2O.
Overall, managed land is a net sink for CO2 (C sequestration) in the United States. The primary drivers of fluxes on
managed lands include, for example, forest management practices, tree planting in urban areas, the management
of agricultural soils, the landfilling of yard trimmings and food scraps, and activities that cause changes in C stocks
in coastal wetlands. 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 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.
2-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
The LULUCF sector in 2018 resulted in a net increase in C stocks (i.e., net CO2 removals) of 799.6 MMT CO2 Eq.
(Table 2-8).2 This represents an offset of approximately 11.9 percent of total (i.e., gross) greenhouse gas emissions
in 2018. Emissions of Cm and N2O from LULUCF activities in 2018 were 26.1 MMT CO2 Eq. and represent 0.4
percent of total greenhouse gas emissions.3 Between 1990 and 2018, total C sequestration in the LULUCF sector
decreased by 7.1 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 CO2 emissions from Land Converted to Settlements.
Forest fires were the largest source of CH4 emissions from LULUCF in 2018, totaling 11.3 MMT CO2 Eq. (452 kt of
CH4). Coastal Wetlands Remaining Coastal Wetlands resulted in CH4 emissions of 3.6 MMT CO2 Eq. (144 kt of CH4).
Grassland fires resulted in CH4 emissions of 0.3 MMT CO2 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 CO2 Eq. each.
Forest fires were also the largest source of N2O emissions from LULUCF in 2018, totaling 7.5 MMT CO2 Eq. (25 kt of
N2O). Nitrous oxide emissions from fertilizer application to settlement soils in 2018 totaled to 2.4 MMT CO2 Eq. (8
kt of N2O). Additionally, the application of synthetic fertilizers to forest soils in 2018 resulted in N2O emissions of
0.5 MMT CO2 Eq. (2 kt of N2O). Grassland fires resulted in N2O emissions of 0.3 MMT CO2 Eq. (1 kt of N2O). Coastal
Wetlands Remaining Coastal Wetlands and Drained Organic Soils resulted in N2O emissions of 0.1 MMT CO2 Eq.
each (less than 0.5 kt of N2O). Peatlands Remaining Peatlands resulted in N2O emissions of less than 0.05 MMT CO2
Eq.
Carbon dioxide removals from C stock changes are presented (green) in Figure 2-11 and Table 2-8 along with CH4
and N2O emissions (purple) for LULUCF source categories.
Figure 2-11: 2018 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.)
Forest Land Remaining Forest Land
Settlements Remaining Settlements
Land Converted to Forest Land
Land Converted to Grassland
Cropland Remaining Cropland
Wetlands Remaining Wetlands
Land Converted to Wetlands
Non-C02 Emissions from Peatlands Remaining Peatlands
CH4 Emissions from Land Converted to Coastal Wetlands
Non-C02 Emissions from Drained Organic Soils
N2O Emissions from Forest Soils
Non-C02 Emissions from Grassland Fires
N2O Emissions from Settlement Soils
Non-C02 Emissions from Coastal Wetlands Remaining Coastal Wetlands
Grassland Remaining Grassland
Non-C02 Emissions from Forest Fires
Land Converted to Cropland
Land Converted to Settlements
(300) (250) (200) (150) (100) (50) 0 50 100
MMT CO2 Eq.
(663.2)
I Non-C02 Emissions
Carbon Stock Change
l< 0.51
l< 0.51
l< 0.51
l< 0.5|
l< 0.51
L
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.
Trends 2-21

-------
Table 2-8: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-
Use Change, and Forestry (MMT CO2 Eq.)
Gas/Land-Use Category
1990
2005
2014
2015
2016
2017
2018
Carbon Stock Change3
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
Forest Land Remaining Forest Land
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
Land Converted to Forest Land
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Land Converted to Cropland
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Grassland Remaining Grassland
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Land Converted to Grassland
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.4)
(4.4)
(4.4)
(4.4)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(109.6)
(116.6)
(126.6)
(126.8)
(125.7)
(125.9)
(125.9)
Land Converted to Settlements
62.9
85.0
81.4
80.1
79.4
79.3
79.3
ch4
4.4
8.8
9.5
16.1
7.3
15.2
15.2
Forest Land Remaining Forest Land:







Forest Firesb
0.9
5.0
5.6
12.2
3.4
11.3
11.3
Wetlands Remaining Wetlands: Coastal







Wetlands Remaining Coastal Wetlands
3.4
3.5
3.6
3.6
3.6
3.6
3.6
Grassland Remaining Grassland:







Grassland Firesc
0.1
0.3
0.4
0.3
0.3
0.3
0.3
Land Converted to Wetlands: Land







Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:







Drained Organic Soilsd
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
n2o
3.0
7.5
7.0
11.2
5.5
10.8
10.9
Forest Land Remaining Forest Land:







Forest Firesb
0.6
3.3
3.7
8.1
2.2
7.5
7.5
Settlements Remaining Settlements:







Settlement Soilse
2.0
3.1
2.2
2.2
2.2
2.3
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
0.4
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 Emissions8
7.4
16.3
16.6
27.4
12.8
26.1
26.1
LULUCF Carbon Stock Change3
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
LULUCF Sector NetTotalh
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
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 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
c Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grassland.
d Estimates include emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted
to Forest Land.
e Estimates include emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted
to Settlements.
2-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
f Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted
to Forest Land.
g 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 CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
Other significant trends from 1990 to 2018 in emissions from LULUCF categories include:
•	Annual C sequestration by forest land (i.e., annual C stock accumulation in the five C pools and harvested
wood products for Forest Land Remaining Forest Land and Land Converted to Forest Land) has decreased
by approximately 8.2 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 14.9 percent over the period from
1990 to 2018. 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 26.1 percent from
1990 to 2018 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 2-12). In 2018,
landfills were the third-largest source of U.S. anthropogenic Cm emissions, generating 110.6 MMT CO2 Eq. and
accounting for 17.4 percent of total U.S. CFU emissions.4 Additionally, wastewater treatment generates emissions
of 19.2 MMT CO2 Eq. and accounts for 14.3 percent of waste emissions, 2.2 percent of U.S. CFU emissions, and 1.2
percent of U.S. N2O emissions. Emissions of CH4 and N2O from composting are also accounted for in this chapter,
generating emissions of 2.5 MMT CO2 Eq. and 2.2 MMT CO2 Eq., respectively. Overall, emission sources accounted
for in the Waste chapter generated 134.4 MMT CO2 Eq., or 2.0 percent of total U.S. greenhouse gas emissions in
2018. A summary of greenhouse gas emissions from the Waste chapter is presented in Table 2-9.
4 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 2-23

-------
Figure 2-12: 2018 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
Landfills
111
Wastewater
Treatment
Composting
Waste as a Portion of All Emissions
Energy
I Agriculture
IPPU
Waste
10
20
30
40
50 60
MMT CO2 Eq.
70
80
90
100
110
Table 2-9: Emissions from Waste (MMT CO2 Eq.)
Gas/Source
1990

2005

2014
2015
2016
2017
2018
ch4
195.3

148.6

129.0
128.0
124.7
124.3
127.2
Landfills
179.6

131.3

112.6
111.3
108.0
107.7
110.6
Wastewater Treatment
15.3

15.4

14.3
14.6
14.4
14.1
14.2
Composting
0.4

1.9

2.1
2.1
2.3
2.4
2.5
n2o
3.7

6.1

6.6
6.7
6.9
7.2
7.2
Wastewater Treatment
3.4

4.4

4.8
4.8
4.9
5.0
5.0
Composting
0.3

1.7

1.9
1.9
2.0
2.2
2.2
Total
199.0

154.7

135.6
134.7
131.6
131.4
134.4
Note: Totals may not sum due to independent rounding.
Some significant trends in U.S. emissions from waste source categories include the following:
•	From 1990 to 2018, net CH4 emissions from landfills decreased by 69.0 MMT CO2 Eq. (38.4 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.
•	Combined CH4 and N2O emissions from composting have generally increased approximately 3.9 MMT CO2
Eq. since 1990, from 0.7 MMT CO2 Eq. to 4.7 MMT CO2 Eq. in 2018, which represents more than a five-fold
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.
•	From 1990 to 2018, CFU and N2O emissions from wastewater treatment decreased by 1.1 MMT CO2 Eq.
(7.4 percent) and increased by 1.6 MMT CO2 Eq. (48.0 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.
2-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. While it is important to use this characterization for
consistency with United Nations Framework Convention on Climate Change (UNFCCC) reporting guidelines and to
promote comparability across countries, 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 (28.2 percent) of
total U.S. greenhouse gas emissions in 2018. Emissions from electric power accounted for the second largest
portion (26.9 percent), while emissions from industry accounted for the third largest portion (22.0 percent) of total
U.S. greenhouse gas emissions in 2018. 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 22.8 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 9.9 percent of emissions; unlike other economic sectors, agricultural sector
emissions were dominated by N2O emissions from agricultural soil management and CH4 emissions from enteric
fermentation, rather than CO2 from fossil fuel combustion. An increasing amount of carbon is stored in agricultural
soils each year, but this CO2 sequestration is assigned to the LULUCF sector rather than the agriculture economic
sector. The commercial and residential sectors accounted for roughly 6.6 percent and 5.6 percent of emissions,
respectively, and U.S. Territories accounted for 0.7 percent of emissions; emissions from these sectors primarily
consisted of CO2 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 2-13 shows the trend in emissions by sector from 1990 to 2018.
Figure 2-13: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)
Electric Power Industry (Purple)
2,500
2,000
Transportation (Green)
1,500
u
Industry
% 1,000
Agriculture
Commercial (Orange]
500
Residential (Blue)
in ID IV CO CTI o
0"i	Ql Ot CTt O
OI Ol Ot	O
CM ro
— o
o
o
o
LD
O
o
ID Pn OO CX»
— — o —
o
o
1—1
o
Cvl
ro
T—I
¦
-------
Table 2-10: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq. and
Percent of Total in 2018)
Sector/Source
1990
2005
2014
2015
2016
2017
2018
Percent3
Transportation
1,527.1
1,973.4
1,791.6
1,800.2
1,835.6
1,852.3
1,882.6
28.2%
C02 from Fossil Fuel Combustion
1,469.1
1,856.1
1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
27.3%
Substitution of Ozone Depleting








Substances
+
69.3
48.8
46.3
43.3
40.1
38.5
0.6%
Mobile Combustion
46.1
37.9
19.1
17.7
16.6
15.3
14.2
0.2%
Non-Energy Use of Fuels
11.8
10.2
10.0
11.0
10.4
9.6
9.3
0.1%
Electric Power Industry
1,875.6
2,455.9
2,089.1
1,949.2
1,856.8
1,778.4
1,798.9
26.9%
C02 from Fossil Fuel Combustion
1,820.0
2,400.0
2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
26.3%
Stationary Combustion
20.9
30.9
29.9
27.7
27.4
25.9
25.6
0.4%
Incineration of Waste
8.4
12.9
10.7
11.1
11.2
11.4
11.4
0.2%
Other Process Uses of Carbonates
3.1
3.8
6.5
6.1
5.3
5.0
5.0
0.1%
Electrical Transmission and








Distribution
23.2
8.4
4.8
3.8
4.1
4.1
4.1
0.1%
Industry
1,628.7
1,501.7
1,438.8
1,429.8
1,388.8
1,411.5
1,470.7
22.0%
C02 from Fossil Fuel Combustion
813.6
799.7
767.4
760.6
761.7
765.6
793.8
11.9%
Natural Gas Systems
215.5
183.4
170.7
171.2
165.7
169.6
174.9
2.6%
Non-Energy Use of Fuels
102.0
121.4
104.9
111.0
98.2
108.5
120.2
1.8%
Petroleum Systems
55.7
51.0
74.0
73.2
62.0
63.2
73.1
1.1%
Coal Mining
96.5
64.1
64.6
61.2
53.8
54.8
52.7
0.8%
Iron and Steel Production
104.8
70.1
58.2
48.0
43.6
40.6
42.6
0.6%
Cement Production
33.5
46.2
39.4
39.9
39.4
40.3
40.3
0.6%
Substitution of Ozone Depleting
Substances
+
9.8
27.0
29.8
32.1
33.9
35.3
0.5%
Petrochemical Production
21.8
27.5
26.4
28.2
28.6
29.2
29.7
0.4%
Ammonia Production
13.0
9.2
9.4
10.6
10.8
13.2
13.5
0.2%
Lime Production
11.7
14.6
14.2
13.3
12.6
12.8
13.2
0.2%
Adipic Acid Production
15.2
7.1
5.4
4.3
7.0
7.4
10.3
0.2%
Nitric Acid Production
12.1
11.3
10.9
11.6
10.1
9.3
9.3
0.1%
Abandoned Oil and Gas Wells
6.6
7.0
7.1
7.2
7.2
7.1
7.0
0.1%
Abandoned Underground Coal








Mines
7.2
6.6
6.3
6.4
6.7
6.4
6.2
0.1%
Electronics Industry
3.6
4.8
4.9
5.0
5.0
4.9
5.1
0.1%
Other Process Uses of Carbonates
3.1
3.8
6.5
6.1
5.3
5.0
5.0
0.1%
Carbon Dioxide Consumption
1.5
1.4
4.5
4.5
4.5
4.5
4.5
0.1%
N20 from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
0.1%
Stationary Combustion
4.8
4.6
4.2
4.2
4.1
4.2
4.2
0.1%
Urea Consumption for Non-








Agricultural Purposes
3.8
3.7
1.8
4.6
5.1
3.8
3.6
0.1%
Mobile Combustion
7.6
7.8
4.0
3.7
3.6
3.6
3.6
0.1%
HCFC-22 Production
46.1
20.0
5.0
4.3
2.8
5.2
3.3
+%
Aluminum Production
28.3
7.6
5.4
4.8
2.7
2.3
3.0
+%
Ferroalloy Production
2.2
1.4
1.9
2.0
1.8
2.0
2.1
+%
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.8
1.7
+%
Titanium Dioxide Production
1.2
1.8
1.7
1.6
1.7
1.7
1.5
+%
Caprolactam, Glyoxal, and








Glyoxylic Acid Production
1.7
2.1
2.0
1.9
1.7
1.5
1.4
+%
Glass Production
1.5
1.9
1.3
1.3
1.2
1.3
1.3
+%
Magnesium Production and








Processing
5.2
2.7
1.0
1.1
1.2
1.2
1.2
+%
2-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Zinc Production
0.6

1.0

1.0
0.9
0.9
1.0
1.0
+%
Phosphoric Acid Production
1.5

1.3

1.0
1.0
1.0
1.0
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
599.0

627.5

654.9
656.0
641.0
642.4
658.6
9.9%
N20 from Agricultural Soil










Management
315.9

313.0

349.2
348.1
329.8
327.4
338.2
5.1%
Enteric Fermentation
164.2

168.9

164.2
166.5
171.8
175.4
177.6
2.7%
Manure Management
51.1

67.9

71.6
75.4
77.7
78.5
81.1
1.2%
C02 from Fossil Fuel Combustion
43.4

50.4

45.5
40.7
39.7
39.4
39.4
0.6%
Rice Cultivation
16.0

18.0

15.4
16.2
13.5
12.8
13.3
0.2%
Urea Fertilization
2.0

3.1

3.9
4.1
4.0
4.5
4.6
0.1%
Liming
4.7

4.3

3.6
3.7
3.1
3.1
3.1
+%
Mobile Combustion
1.2

1.2

0.8
0.6
0.6
0.6
0.6
+%
Field Burning of Agricultural










Residues
0.5

0.6

0.6
0.6
0.6
0.6
0.6
+%
Stationary Combustion
0.1

+

0.1
0.1
0.1
0.1
0.1
+%
Commercial
428.7

405.1

429.4
442.5
427.0
426.8
443.3
6.6%
C02 from Fossil Fuel Combustion
228.2

226.9

232.8
245.4
232.3
232.8
246.5
3.7%
Landfills
179.6

131.3

112.6
111.3
108.0
107.7
110.6
1.7%
Substitution of Ozone Depleting










Substances
+

22.1

59.5
60.8
61.5
61.0
60.8
0.9%
Wastewater Treatment
15.3

15.4

14.3
14.6
14.4
14.1
14.2
0.2%
Human Sewage
3.4

4.4

4.8
4.8
4.9
5.0
5.0
0.1%
Composting
0.7

3.5

4.0
4.0
4.3
4.6
4.7
0.1%
Stationary Combustion
1.5

1.4

1.4
1.6
1.5
1.5
1.6
+%
Residential
344.7

370.1

378.6
352.0
328.3
330.2
375.9
5.6%
C02 from Fossil Fuel Combustion
338.2

357.9

346.8
317.8
293.1
293.8
337.3
5.1%
Substitution of Ozone Depleting










Substances
0.2

7.2

25.8
28.9
30.4
31.8
33.2
0.5%
Stationary Combustion
6.3

4.9

6.0
5.3
4.7
4.6
5.4
0.1%
U.S. Territories
33.3

58.0

46.6
46.6
46.6
46.6
46.6
0.7%
C02 from Fossil Fuel Combustion
27.6

49.7

41.4
41.4
41.4
41.4
41.4
0.6%
Non-Energy Use of Fuels
5.7

8.1

5.1
5.1
5.1
5.1
5.1
0.1%
Stationary Combustion
0.1

0.2

0.2
0.2
0.2
0.2
0.2
+%
Total Emissions
6,437.0

7,391.8

6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
100.0%
LULUCF Sector Net Totalb
(853.4)

(814.7)

(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
(11.6%)
Net Emissions (Sources and Sinks)
5,583.6

6,577.1

6,106.0
5,900.8
5,735.1
5,724.3
5,903.2
88.4%
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 2018.
b The LULUCF Sector Net Total is the net sum of all 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 CO2 emissions from the combustion of fossil fuels that are included
in the EIA electric power sector. Stationary combustion emissions of CH4 and N2O are also based on the EIA
electric power sector. Additional sources include CO2, Cm, and N2O from waste incineration, as the majority of
municipal solid waste is combusted in plants that produce electricity. The Electric Power economic sector also
Trends 2-27

-------
includes SF6 from Electrical Transmission and Distribution, and a portion of CO2 from Other Process Uses of
Carbonates (from pollution control equipment installed in electric power plants).
The Transportation economic sector includes CO2 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 N2O 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, CO2
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 CO2 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 N2O
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.
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 CO2 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 CO2 emissions from Non-Energy Uses of Fossil Fuels are assumed
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 CO2 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 2020). These supplementary data are subtracted from the industrial fuel use
reported by EIA to obtain agriculture fuel use. CO2 emissions from fossil fuel combustion, and CH4 and N2O
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
N2O emissions from Agricultural Soils, CFU from Enteric Fermentation, CH4 and N2O from Manure Management,
Cm from Rice Cultivation, CO2 emissions from Liming and Urea Application, and CH4 and N2O from Field Burning
of Agricultural Residues.
The Residential economic sector includes CO2 emissions from the combustion of fossil fuels that are included in
the EIA residential fuel-consuming sector. Stationary combustion emissions of CH4 and N2O 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 CO2 emissions from the combustion of fossil fuels that are included in
the EIA commercial fuel-consuming sector. Emissions of CH4 and N2O 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
2-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
direct CFU from Landfills, CFU and N2O from 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 26.9 percent of total U.S. greenhouse
gas emissions in 2018. Electric power-related emissions decreased by 4.1 percent since 1990 but increased by 1.2
percent from 2017 to 2018, primarily due to a significantly colder winter and a hotter summer in 2018 compared
to 2017, which increased the amount of energy required for heating and cooling. Between 2017 to 2018, the
consumption of natural gas and petroleum for electric power generation increased by 14.2 and 19.6 percent,
respectively, while the consumption of coal decreased by 4.5 percent, reflecting a continued shift from coal to
natural gas for electricity generation.
From 2017 to 2018, electricity sales to the residential and commercial end-use sectors increased by 6.6 percent
and 2.1 percent, respectively. Electricity sales to the industrial sector increased by approximately 1.8 percent.
Overall, from 2017 to 2018, the amount of electricity retail sales (in kWh) increased by 3.7 percent. Table 2-11
provides a detailed summary of emissions from electric power-related activities.
Table 2-11: Electric Power-Related Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Fuel Type or Source
1990

2005

2014
2015
2016
2017
2018
C02
1,831.0

2,416.3

2,054.1
1,917.5
1,825.0
1,748.1
1,768.9
Fossil Fuel Combustion
1,820.0

2,400.0

2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
Coal
1,546.5

1,982.8

1,568.6
1,351.4
1,242.0
1,207.1
1,152.9
Natural Gas
175.4

318.9

442.9
525.2
545.0
505.6
577.4
Petroleum
97.5

97.9

25.3
23.7
21.4
18.9
22.2
Geothermal
0.5

0.5

0.4
0.4
0.4
0.4
0.4
Incineration of Waste
8.0

12.5

10.4
10.8
10.9
11.1
11.1
Other Process Uses of









Carbonates
3.1

3.8

6.5
6.1
5.3
5.0
5.0
ch4
0.4

0.9

1.1
1.2
1.2
1.1
1.2
Stationary Sources3
0.4

0.9

1.1
1.2
1.2
1.1
1.2
Incineration of Waste
+

+

+
+
+
+
+
n2o
21.0

30.4

29.2
26.8
26.5
25.1
24.7
Stationary Sources3
20.5

30.1

28.9
26.5
26.2
24.8
24.4
Incineration of Waste
0.5

0.4

0.3
0.3
0.3
0.3
0.3
sf6
23.2

8.4

4.8
3.8
4.1
4.1
4.1
Electrical Transmission and









Distribution
23.2

8.4

4.8
3.8
4.1
4.1
4.1
Total
1,875.6

2,455.9

2,089.1
1,949.2
1,856.8
1,778.4
1,798.9
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
3 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 2019;
Duffield 2006). These source categories include CO2 from Fossil Fuel Combustion, CFU and N2O from Stationary
Combustion, Incineration of Waste, Other Process Uses of Carbonates, and SF6 from Electrical Transmission and
Trends 2-29

-------
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.5
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 (28.9 percent), followed closely by emissions
from transportation (28.3 percent). Emissions from the commercial and residential sectors also increase
substantially when emissions from electricity are included (16.0 and 15.6 percent, respectively). In all economic
end-use sectors except agriculture, CO2 accounts for more than 80.6 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 2-11 shows the trend in
these emissions by sector from 1990 to 2018.
Figure 2-14: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed
to Economic Sectors (MMT CO2 Eq.)
2,500
Industry
2,000
Transportation
1,500
Commercial (Orange)
Residential (Blue)
Agriculture
500
000
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 2018
Sector/Gas
1990

2005

2014
2015
2016
2017
2018
Percent3
Industry
2,301.0

2,216.8

2,002.6
1,952.1
1,881.0
1,890.7
1,931.0
28.9%
Direct Emissions
1,628.7

1,501.7

1,438.8
1,429.8
1,388.8
1,411.5
1,470.7
22.0%
C02
1,166.7

1,148.8

1,104.7
1,100.5
1,072.7
1,088.6
1,148.7
17.2%
ch4
348.1

282.4

266.5
260.9
246.2
249.8
245.7
3.7%
n2o
37.6

29.7

27.3
26.5
27.7
27.2
30.2
0.5%
HFCs, PFCs, SF6, and NF3
76.3

40.7

40.2
41.8
42.2
45.9
46.2
0.7%
Electricity-Related
672.3

715.2

563.8
522.4
492.2
479.2
460.3
6.9%
C02
656.4

703.6

554.3
513.8
483.8
471.0
452.6
6.8%
ch4
0.2

0.3

0.3
0.3
0.3
0.3
0.3
+%
5 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.
2-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
n2o
7.5
8.9
7.9
7.2
7.0
6.8
6.3
0.1%
sf6
8.3
2.4
1.3
1.0
1.1
1.1
1.0
+%
Transportation
1,530.2
1,978.3
1,796.2
1,804.6
1,839.9
1,856.7
1,887.4
28.3%
Direct Emissions
1,527.1
1,973.4
1,791.6
1,800.2
1,835.6
1,852.3
1,882.6
28.2%
C02
1,480.9
1,866.3
1,723.8
1,736.2
1,775.7
1,796.8
1,829.9
27.4%
ch4
5.9
3.0
1.7
1.6
1.5
1.5
1.4
+%
n2o
40.2
34.8
17.4
16.0
15.1
13.9
12.8
0.2%
HFCsb
+
69.3
48.8
46.3
43.3
40.1
38.5
0.6%
Electricity-Related
3.1
4.8
4.6
4.4
4.3
4.4
4.9
0.1%
C02
3.1
4.8
4.5
4.3
4.2
4.4
4.8
0.1%
ch4
+
+
+
+
+
+
+
+%
n2o
+
0.1
0.1
0.1
0.1
0.1
0.1
+%
sf6
+
+
+
+
+
+
+
+%
Commercial
982.8
1,226.8
1,153.0
1,122.5
1,077.4
1,049.2
1,070.9
16.0%
Direct Emissions
428.7
405.1
429.4
442.5
427.0
426.8
443.3
6.6%
C02
228.2
226.9
232.8
245.4
232.3
232.8
246.5
3.7%
ch4
196.4
149.7
130.1
129.2
125.9
125.5
128.5
1.9%
n2o
4.1
6.4
7.0
7.0
7.2
7.5
7.5
0.1%
HFCs
+
22.1
59.5
60.8
61.5
61.0
60.8
0.9%
Electricity-Related
554.2
821.7
723.6
680.0
650.4
622.4
627.5
9.4%
C02
541.0
808.4
711.5
668.9
639.3
611.8
617.1
9.2%
ch4
0.1
0.3
0.4
0.4
0.4
0.4
0.4
+%
n2o
6.2
10.2
10.1
9.4
9.3
00
00
8.6
0.1%
sf6
6.8
2.8
1.7
1.3
1.4
1.4
1.4
+%
Residential
955.6
1,246.0
1,131.4
1,053.3
999.1
963.9
1,042.4
15.6%
Direct Emissions
344.7
370.1
378.6
352.0
328.3
330.2
375.9
5.6%
C02
338.2
357.9
346.8
317.8
293.1
293.8
337.3
5.1%
ch4
5.2
4.1
5.0
4.5
3.9
3.8
4.5
0.1%
n2o
1.0
0.9
1.0
0.9
0.8
0.8
0.9
+%
HFCs
0.2
7.2
25.8
28.9
30.4
31.8
33.2
0.5%
Electricity-Related
610.9
875.9
752.8
701.3
670.8
633.6
666.5
10.0%
C02
596.4
861.8
740.2
689.9
659.3
622.8
655.4
9.8%
ch4
0.1
0.3
0.4
0.4
0.4
0.4
0.4
+%
n2o
6.8
10.9
10.5
9.7
9.6
8.9
9.2
0.1%
sf6
7.5
3.0
1.7
1.4
1.5
1.5
1.5
+%
Agriculture
634.0
665.8
699.2
697.2
680.1
681.1
698.3
10.5%
Direct Emissions
599.0
627.5
654.9
656.0
641.0
642.4
658.6
9.9%
C02
50.1
57.9
53.1
48.5
46.9
47.0
47.1
0.7%
ch4
218.3
239.4
234.5
241.1
245.5
248.6
253.1
3.8%
n2o
330.6
330.2
367.4
366.4
348.7
346.8
358.3
5.4%
Electricity-Related
35.1
38.3
44.3
41.2
39.1
38.7
39.7
0.6%
C02
34.2
37.7
43.6
40.6
38.4
38.1
39.1
0.6%
ch4
+
+
+
+
+
+
+
+%
n2o
0.4
0.5
0.6
0.6
0.6
0.5
0.5
+%
sf6
0.4
0.1
0.1
0.1
0.1
0.1
0.1
+%
U.S. Territories
33.3
58.0
46.6
46.6
46.6
46.6
46.6
0.7%
Total Emissions
6,437.0
7,391.8
6,829.0
6,676.4
6,524.1
6,488.2
6,676.6
100.0%
LULUCF Sector NetTotalc
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
(11.6%)
Net Emissions (Sources and








Sinks)
5,583.6
6,577.1
6,106.0
5,900.8
5,735.1
5,724.3
5,903.2
88.4%
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 2018.
b Includes primarily HFC-134a.
c The LULUCF Sector Net Total is the net sum of all CH4 and N20 emissions to the atmosphere plus net carbon stock changes.
Trends 2-31

-------
Industry
The industry end-use sector includes CO2 emissions from fossil fuel combustion from all manufacturing facilities, in
aggregate, and with the distribution of electricity-related emissions, accounts for 28.9 percent of U.S. greenhouse
gas emissions in 2018. 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 Cm emissions from petroleum and natural gas systems, fugitive CH4 emissions from coal mining,
byproduct CO2 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 16.1 percent. The decline has occurred both in direct
emissions and indirect emissions associated with electricity use. Structural changes within the U.S. economy that
led 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.
Transportation
When electricity-related emissions are distributed to economic end-use sectors, transportation activities
accounted for 28.3 percent of U.S. greenhouse gas emissions in 2018. The largest sources of transportation
greenhouse gas emissions in 2018 were passenger cars (41.2 percent); freight trucks (23.2 percent); light-duty
trucks, which include sport utility vehicles, pickup trucks, and minivans (17.4 percent); commercial aircraft (6.9
percent); pipelines (2.6 percent); other aircraft (2.4 percent); rail (2.3 percent); and ships and boats (2.2 percent).
These figures include direct CO2, Cm, and N2O 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 2018, 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 46.1 percent from 1990 to 2018, 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 2013,6 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 48 percent. Light-duty truck market share
was about 48 percent of new vehicles in model year 2018 (EPA 2019a).
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-121 in
Annex 3.
6 VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2017). 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.
2-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 from fossil fuel combustion, which increased by 24 percent from 1990 to
2018.7 This rise in CO2 emissions, combined with an increase in HFCs from close to zero emissions in 1990 to 38.5
MMT CO2 Eq. in 2018, led to an increase in overall greenhouse gas emissions from transportation activities of 23
percent.8
Table 2-13: Transportation-Related Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Vehicle
1990
2005
2014
2015
2016
2017
2018
Passenger Cars
639.6
693.1
760.3
760.2
770.6
767.3
777.5
C02
612.2
642.8
734.7
736.8
749.8
749.2
761.5
ch4
3.2
1.3
0.7
0.6
0.6
0.5
0.5
n2o
24.1
17.3
9.1
8.1
7.1
6.1
5.1
HFCs
0.0
31.7
15.8
14.7
13.2
11.4
10.4
Light-Duty Trucks
326.7
538.5
334.7
323.7
332.8
326.8
328.3
C02
312.2
490.7
305.9
297.2
308.7
305.0
308.0
ch4
1.7
0.8
0.3
0.3
0.2
0.2
0.2
n2o
12.8
13.6
3.9
3.2
2.9
2.4
2.0
HFCs
0.0
33.3
24.7
23.0
21.1
19.2
18.1
Medium- and Heavy-Duty







Trucks
230.3
400.1
402.5
410.1
414.2
427.6
437.9
C02
229.3
395.4
394.8
402.1
406.0
419.0
428.9
ch4
0.3
0.1
0.1
0.1
0.1
0.1
0.1
n2o
0.7
1.2
2.3
2.4
2.6
2.8
3.0
HFCs
0.0
3.4
5.3
5.5
5.5
5.7
5.9
Buses
8.5
12.2
19.0
19.4
19.0
20.4
21.9
C02
8.4
11.6
18.3
18.7
18.3
19.7
21.1
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
3.8
3.7
3.9
3.8
3.9
C02
1.7
1.6
3.8
3.7
3.8
3.7
3.8
ch4
+
+
+
+
+
+
+
n2o
+
+
+
+
+
+
+
Commercial Aircraft3
110.9
134.0
116.3
120.1
121.5
129.2
130.8
C02
109.9
132.7
115.2
119.0
120.4
128.0
129.6
ch4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
n2o
1.0
1.2
1.1
1.1
1.1
1.2
1.2
Other Aircraftb
78.3
59.7
35.0
40.4
47.5
45.6
44.7
C02
77.5
59.1
34.7
40.0
47.0
45.2
44.3
ch4
0.1
0.1
+
+
+
+
+
n2o
0.7
0.5
0.3
0.4
0.4
0.4
0.4
Ships and Boatsc
47.4
45.7
29.2
33.8
40.9
44.0
41.2
C02
46.3
44.2
26.2
30.5
37.1
39.9
36.8
ch4
0.6
0.5
0.3
0.3
0.3
0.3
0.3
n2o
0.6
0.6
0.3
0.4
0.5
0.5
0.5
HFCs
0.0
0.5
2.3
2.6
2.9
3.3
3.6
Rail
39.0
50.9
45.9
43.7
39.9
41.1
42.9
C02
38.5
50.3
45.2
43.0
39.3
40.5
42.3
ch4
0.1
0.1
0.1
0.1
0.1
0.1
0.1
7	See previous footnote.
8	See previous footnote.
Trends 2-33

-------
n2o
0.3
0.4
0.4
0.4
0.3
0.4
0.4
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
Pipelines8
36.0
32.4
39.4
38.5
39.2
41.3
49.2
C02
36.0
32.4
39.4
38.5
39.2
41.3
49.2
Lubricants
11.8
10.2
10.0
11.0
10.4
9.6
9.3
C02
11.8
10.2
10.0
11.0
10.4
9.6
9.3
Total Transportation
1,530.2
1,978.3
1,796.2
1,804.6
1,839.9
1,856.7
1,887.4
International Bunker FuelsS
54.8
44.7
28.7
31.6
35.0
34.6
32.5
Ethanol C029
4.1
21.6
74.0
74.2
76.9
77.7
78.6
Biodiesel C02g
0.0
0.9
13.3
14.1
19.6
18.7
17.9
+ 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).
e 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,
s 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.
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.
Commercial
The commercial end-use sector, with electricity-related emissions distributed, accounts for 16.0 percent of U.S.
greenhouse gas emissions in 2018 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 15.6 percent of U.S.
greenhouse gas emissions in 2018 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
2-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 10.5 percent of U.S. greenhouse gas emissions in 2018 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 2018, agricultural soil
management was the largest source of N2O emissions, and enteric fermentation was the largest source of CFU
emissions in the United States. This sector also includes small amounts of CO2 emissions from fossil fuel
combustion by motorized farm equipment such as tractors.
Box 2-2: Recent 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 0.2 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 2-15). 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 0.7 percent 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 2-14: Recent Trends in Various U.S. Data (Index 1990 = 100)
Variable
1990

2005

2014
2015
2016
2017
2018
Avg. Annual
Change
Since 1990a
Avg. Annual
Change
Since 2005a
Greenhouse Gas Emissions'5
100

115

106
104
101
101
104
0.2%
-0.7%
Energy Usec
100

118

117
116
116
116
120
0.7%
0.1%
GDPd
100

159

181
186
189
193
199
2.5%
1.7%
Population6
100

118

127
128
129
130
131
1.0%
0.8%
a Average annual growth rate.
b GWP-weighted values.
c Energy-content-weighted values (EIA 2019).
d GDP in chained 2009 dollars (BEA 2020).
e U.S. Census Bureau (2020).
Trends 2-35

-------
Figure 2-15: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic
Product
200
Real GDP
180
160
Population
140
Energy Use
120
U>
x 100
CD
-o
Emissions
Energy Use Per Capita
c
i—i
Emissions per Capita
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 UNFCCC9 request that information be provided on precursor greenhouse gases,
which include carbon monoxide (CO), nitrogen oxides (NOx), non-Cm volatile organic compounds (NMVOCs), and
sulfur dioxide (SO2). 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
SO2, 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
NO2) are created by lightning, fires, fossil fuel combustion, and in the stratosphere from N2O. 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, SO2
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
9 See .
2-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
indirect greenhouse gas formation into greenhouse gases is the interaction of CO with the hydroxyl radical—the
major atmospheric sink for Cm emissions—to form CO2. 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 SO2 (EPA 2019b),10
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
2014
2015
2016
2017
2018
NOx
21,738
17,338
10,797
10,286
9,572
9,293
8,892
Mobile Fossil Fuel Combustion
10,862
10,295
6,138
5,740
5,413
5,051
4,689
Stationary Fossil Fuel Combustion
10,023
5,858
3,313
3,036
2,876
2,757
2,719
Oil and Gas Activities
139
321
650
650
650
650
650
Industrial Processes and Product Use
592
572
414
414
414
414
414
Forest Fires
22
127
142
312
87
289
289
Waste Combustion
82
128
97
97
97
97
97
Grassland Fires
5
21
27
21
19
21
20
Agricultural Burning
12
14
14
13
13
13
13
Waste
+
2
2
2
2
2
2
CO
130,943
71,745
47,328
52,310
41,871
47,438
45,749
Mobile Fossil Fuel Combustion
119,360
58,615
34,135
33,159
30,786
29,112
27,438
Forest Fires
801
4,507
5,055
11,125
3,092
10,314
10,314
Stationary Fossil Fuel Combustion
5,000
4,648
3,686
3,686
3,686
3,686
3,686
Waste Combustion
978
1,403
1,776
1,776
1,776
1,776
1,776
Industrial Processes and Product Use
4,129
1,557
1,251
1,251
1,251
1,251
1,251
Oil and Gas Activities
302
318
637
637
637
637
637
Grassland Fires
84
358
442
356
324
345
331
Agricultural Burning
287
332
338
311
310
308
308
Waste
1
7
8
8
8
8
8
NMVOCs
20,930
13,154
11,130
10,965
10,718
10,513
10,307
Industrial Processes and Product Use
7,638
5,849
3,815
3,815
3,815
3,815
3,815
Mobile Fossil Fuel Combustion
10,932
5,724
3,754
3,589
3,342
3,137
2,931
Oil and Gas Activities
554
510
2,853
2,853
2,853
2,853
2,853
Stationary Fossil Fuel Combustion
912
716
497
497
497
497
497
Waste Combustion
222
241
143
143
143
143
143
Waste
673
114
68
68
68
68
68
Agricultural Burning
NA
NA
NA
NA
NA
NA
NA
S02
20,935
13,196
4,240
3,342
2,685
2,548
2,481
Stationary Fossil Fuel Combustion
18,407
11,541
3,532
2,635
1,978
1,841
1,774
Industrial Processes and Product Use
1,307
831
497
497
497
497
497
Oil and Gas Activities
390
180
94
94
94
94
94
Mobile Fossil Fuel Combustion
793
619
88
87
87
87
87
Waste Combustion
38
25
27
27
27
27
27
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.
10 NOx and CO emission estimates from Field Burning of Agricultural Residues were estimated separately, and therefore not
taken from EPA (2019b).
Trends 2-37

-------
NA (Not Available)
Source: (EPA 2019b) 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 (SO2) 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 SO2 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 SO2 emissions in the Clean Air Act.
Electric power is the largest anthropogenic source of SO2 emissions in the United States, accounting for 47.8
percent in 2018. 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.
2-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Energy-related activities were the primary sources of U.S. anthropogenic greenhouse gas emissions, accounting for
83.1 percent of total greenhouse gas emissions on a carbon dioxide (CO2) equivalent basis in 2018.1 This included
97, 40, and 10 percent of the nation's CO2, methane (CH4), and nitrous oxide (N2O) emissions, respectively. Energy-
related CO2 emissions alone constituted 78.6 percent of U.S. greenhouse gas emissions from all sources on a CO2-
equivalent basis, while the non-CC>2 emissions from energy-related activities represented a much smaller portion
of total national emissions (4.5 percent collectively).
Emissions from fossil fuel combustion comprise the vast majority of energy-related emissions, with CO2 being the
primary gas emitted (see Figure 3-1). Globally, approximately 32,840 million metric tons (MMT) of CO2 were added
to the atmosphere through the combustion of fossil fuels in 2017, of which the United States accounted for
approximately 15 percent.2 Due to their relative importance, fossil fuel combustion-related CO2 emissions are
considered separately and in more detail than other energy-related emissions (see Figure 3-2).
Fossil fuel combustion also emits CFU and N2O. Stationary combustion of fossil fuels was the second largest source
of N2O emissions in the United States and mobile fossil fuel combustion was the fourth 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 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 C02 Emissions from Fossil
Fuels Combustion Overview Available at:  (IEA 2019).
Energy 3-1

-------
Figure 3-1: 2018 Energy Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
CO2 Emissions from Fossil Fuel Combustion
Natural Gas Systems
Non-Energy Use of Fuels
Petroleum Systems
Coal Mining
Non-COz 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
waste
Energy
IPPU
Agriculture
83.1%
100	150	200
MMT CO2 Eq.
Figure 3-2: 2018 U.S. Fossil Carbon Flows (MMT CO2 Eq.)
Combustion
Emissions
Natural Gas
1,617
Combustion
Emissions 1,221
Atmospheric
Emissions
5,373
Domestic
Fossil Fuel
Production
5,179
Apparent
Consumption
5,523
Combustion
Emissions
2,327
Petroleum
1,717
Fossil Fuel
Energy
Imports
1:759_-
Petroleum
1,369- "
Fossil Fuel
Energy Exports
1,411
Non-Energy
Use Exports stock
Changes
* 81
International
Bunkers , . . . ,
121 Industrial
NEU Emissions 6
Natural Gas Emissions
1,612
NEU Emissions 13
Coal Emissions
1,209
NEU Emissions 116
Petroleum
Emissions
Non-Energy Balancing
Natural Gas Liquids,
Liquefied Refinery Gas,
& Other Liquids
353
Natural Gas 157
Coal 14
Other 220
Non-Energy Use
Carbon Sequestered
222
Non
Use
Use U.S.
Fossil Fuel Territories
•Energy Consumption
Imports U.S.
22 Territories
42
Item
72
Note: Totals may not sum due to independent rounding.
The "Balancing Item" above accounts for the statistical imbalances
and unknowns in the reported data sets combined here.
NEU = Non-Energy Use
Table 3-1 summarizes emissions from the Energy sector in units of MMT CO2 Eq., while unweighted gas emissions
in kilotons (kt) are provided in Table 3-2. Overall, emissions due to energy-related activities were 5,547.2 MMT CO2
Eq. in 2018,3 an increase of 3.9 percent since 1990 and an increase of 3.0 percent since 2017.
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.
3-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-1: CO2, ChU, and N2O Emissions from Energy (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
2018
CO?
4,909.3
5,930.3
5,375.4
5,231.5
5,119.8
5,081.3
5,249.3
Fossil Fuel Combustion
4,740.0
5,740.7
5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
Transportation
1,469.1
1,856.1
1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
Electric Power
1,820.0
2,400.0
2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
Industrial
857.0
850.1
812.9
801.3
801.4
805.0
833.2
Residential
338.2
357.9
346.8
317.8
293.1
293.8
337.3
Commercial
228.2
226.9
232.8
245.4
232.3
232.8
246.5
U.S. Territories
27.6
49.7
41.4
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.5
139.7
120.0
127.0
113.7
123.1
134.6
Petroleum Systems
9.6
12.2
30.5
32.6
23.0
24.5
36.8
Natural Gas Systems
32.2
25.3
29.6
29.3
29.9
30.4
35.0
Incineration of Waste
8.0
12.5
10.4
10.8
10.9
11.1
11.1
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Biomass-Wood"
215.2
206.9
233.8
224.7
216.3
221.4
229.1
International Bunker Fuelsb
103.5
113.1
103.4
110.9
116.6
120.1
122.1
Biofuels-Ethanola
4.2
22.9
76.1
78.9
81.2
82.1
81.9
Biofuels-Biodiesela
0.0
0.9
13.3
14.1
19.6
18.7
17.9
ch4
361.2
292.0
275.6
269.3
253.9
257.3
253.9
Natural Gas Systems
183.3
158.1
141.1
141.9
135.8
139.3
140.0
Coal Mining
96.5
64.1
64.6
61.2
53.8
54.8
52.7
Petroleum Systems
46.1
38.8
43.5
40.5
39.0
38.7
36.2
Stationary Combustion
8.6
7.8
8.9
8.5
7.9
7.8
8.6
Abandoned Oil and Gas Wells
6.6
7.0
7.1
7.1
7.2
7.1
7.0
Abandoned Underground







Coal Mines
7.2
6.6
6.3
6.4
6.7
6.4
6.2
Mobile Combustion
12.9
9.6
4.1
3.6
3.4
3.3
3.1
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
n2o
67.6
72.1
53.1
49.2
47.8
45.2
44.0
Stationary Combustion
25.1
34.3
33.0
30.5
30.0
28.6
28.4
Mobile Combustion
42.0
37.3
19.7
18.3
17.4
16.3
15.2
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Petroleum Systems
+
+
+
+
+
+
0.1
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
0.9
1.0
1.0
1.1
1.1
Total
5,338.1
6,294.4
5,704.0
5,550.1
5,421.6
5,383.8
5,547.2
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
2014
2015
2016
2017
2018
CO?
4,909,296
5,930,296
5,375,404
5,231,530
5,119,841
5,081,321
5,249,295
Fossil Fuel Combustion
4,740,006
5,740,660
5,184,776
5,031,762
4,942,421
4,892,234
5,031,813
Non-Energy Use of







Fuels
119,530
139,707
120,030
127,027
113,651
123,133
134,576
Petroleum Systems
9,630
12,163
30,536
32,644
22,980
24,472
36,814
Natural Gas Systems
32,174
25,291
29,620
29,334
29,862
30,365
34,972
Incineration of Waste
7,951
12,469
10,435
10,756
10,919
11,111
11,113
Energy 3-3

-------
Abandoned Oil and







Gas Wells
6
1
1
1
1
1
7
Biomass-Wood"
215,186
206,901
233,762
224,730
216,293
221,432
229,085
International Bunker







Fuelsb
103,463
113,139
103,400
110,887
116,594
120,107
122,088
Biofuels-Ethanola
4,227
22,943
76,075
78,934
81,250
82,088
81,917
Biofuels-Biodiesela
0
856
13,349
14,077
19,648
18,705
17,936
ch4
14,449
11,680
11,023
10,772
10,158
10,292
10,156
Natural Gas Systems
7,332
6,324
5,643
5,674
5,433
5,570
5,598
Coal Mining
3,860
2,565
2,583
2,449
2,154
2,191
2,109
Petroleum Systems
1,844
1,553
1,739
1,622
1,559
1,548
1,449
Stationary Combustion
344
313
355
340
318
312
346
Abandoned Oil and







Gas Wells
263
278
284
286
289
282
281
Abandoned







Underground Coal







Mines
288
264
253
256
268
257
247
Mobile Combustion
518
383
166
146
138
131
126
Incineration of Waste
+
+
+
+
+
+
+
International Bunker







Fuelsb
7
5
3
4
4
4
4
n2o
227
242
178
165
160
152
148
Stationary Combustion
84
115
111
102
101
96
95
Mobile Combustion
141
125
66
62
58
55
51
Incineration of Waste
2
1
1
1
1
1
1
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker







Fuelsb
3
3
3
3
3
4
4
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.
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 2017) to ensure that the trend
is accurate. Updates to CO2 emissions from Fossil Fuel Combustion in the Energy sector resulted in an average
decrease over the time series of about 6.6 MMT CO2 Eq. For more information on specific methodological updates,
please see the Recalculations for each category, in this chapter.
3-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 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 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
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
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 3.7 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
Section 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 CO2 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 2019) 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
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 .
Energy 3-5

-------
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 gases CO2, Cm, and N2O. Given that CO2 is the
primary gas emitted from fossil fuel combustion and represents the largest share of U.S. total emissions, CO2
emissions from fossil fuel combustion are discussed at the beginning of this section. Following that is a discussion
of emissions of all three gases from fossil fuel combustion presented by sectoral breakdowns. Methodologies for
estimating CO2 from fossil fuel combustion also differ from the estimation of CH4 and N2O 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 CO2, CH4, and N2O
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
2014
2015
2016
2017
2018
C02
4,740.0
5,740.7
5,184.8
5,031.8
4,942.4
4,892.2
5,031.8
ch4
21.5
17.4
13.0
12.1
11.4
11.1
11.8
n2o
67.1
71.6
52.7
48.9
47.4
44.9
43.6
Total
4,828.7
5,829.7
5,250.5
5,092.8
5,001.2
4,948.2
5,087.2
Note: Totals may not sum due to independent rounding.
Table 3-4: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion (kt)
Gas
C02
CH4
n2o
1990
4,740,006
862
225
2005
5,740,660
696
240
2014
2015
5,184,776 5,031,762
521	485
177	164
2016
4,942,421
455
159
2017
4,892,234
443
151
2018
5,031,813
471
146
CO2 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
2018, CO2 emissions from fossil fuel combustion increased by 2.9 percent relative to the previous year. The
increase in CO2 emissions from fossil fuel consumption was a result of a 4.1 percent increase in total energy use
and reflects a continued shift from coal to natural gas. Carbon dioxide emissions from natural gas consumption
increased by 160.3 MMT CO2 Eq. in 2018, an 11.0 percent increase from 2017, while CO2 emissions from coal
consumption decreased by 4.7 percent. The increase in natural gas consumption and emissions in 2018 is observed
across all sectors and is primarily driven by increased energy use from greater heating and cooling needs due to a
colder winter and hotter summer in 2018 (in comparison to 2017). In 2018, CO2 emissions from fossil fuel
combustion were 5,031.8 MMT CO2 Eq., or 6.2 percent above emissions in 1990 (see Table 3-5).6
6 An additional discussion of fossil fuel emission trends is presented in the Trends in U.S. Greenhouse Gas Emissions chapter.
3-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-5: CO2 Emissions from Fossil Fuel Combustion by Fuel Type and Sector (MMT CO2
Eq.)
Fuel/Sector
1990
2005
2014
2015
2016
2017
2018
Coal
1,717.3
2,111.2
1,652.4
1,424.7
1,307.5
1,267.5
1,208.5
Residential
3.0
0.8
NO
NO
NO
NO
NO
Commercial
12.0
9.3
3.8
3.0
2.3
2.0
1.8
Industrial
155.2
115.3
76.0
66.3
59.2
54.4
49.8
Transportation
NE
NE
NE
NE
NE
NE
NE
Electric Power
1,546.5
1,982.8
1,568.6
1,351.4
1,242.0
1,207.1
1,152.9
U.S. Territories
0.6
3.0
4.0
4.0
4.0
4.0
4.0
Natural Gas
999.7
1,167.0
1,420.0
1,460.2
1,471.8
1,451.4
1,611.6
Residential
237.8
262.2
277.7
252.7
238.4
241.5
273.7
Commercial
142.0
162.9
189.2
175.4
170.5
173.2
192.6
Industrial
408.5
388.6
467.1
464.4
474.8
485.8
514.8
Transportation
36.0
33.1
40.2
39.4
40.1
42.3
50.2
Electric Power
175.4
318.9
442.9
525.2
545.0
505.6
577.4
U.S. Territories
NO
1.3
3.0
3.0
3.0
3.0
3.0
Petroleum
2,022.4
2,462.1
2,111.9
2,146.5
2,162.7
2,172.9
2,211.3
Residential
97.4
94.9
69.1
65.1
54.8
52.3
63.5
Commercial
74.2
54.7
39.8
67.1
59.5
57.6
52.1
Industrial
293.3
346.2
269.9
270.5
267.4
264.8
268.6
Transportation
1,433.1
1,823.0
1,673.5
1,685.9
1,725.2
1,745.0
1,770.5
Electric Power
97.5
97.9
25.3
23.7
21.4
18.9
22.2
U.S. Territories
26.9
45.4
34.3
34.3
34.3
34.3
34.3
Geothermal3
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Total
4,740.0
5,740.7
5,184.8
5,031.8
4,942.5
4,892.3
5,031.8
Note: Totals may not sum due to independent rounding.
NE (Not Estimated)
NO (Not Occurring)
a Although not technically a fossil fuel, geothermal energy-related C02 emissions are included for reporting purposes.
Trends in CO2 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, 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
7 Based on national aggregate carbon content of all coal, natural gas, and petroleum fuels combusted in the United States.
Energy 3-7

-------
Table 3-6 shows annual changes in emissions during the last five years for coal, petroleum, and natural gas in
selected sectors.
Table 3-6: Annual Change in CO2 Emissions and Total 2018 CO2 Emissions from Fossil Fuel
Combustion for Selected Fuels and Sectors (MMT CO2 Eq. and Percent)
Sector
Fuel Type
2014 to 2015
2015 to 2016
2016 to 2017
2017 to 2018
Total 2018
Electric Power
Coal
-217.2
-13.8%
-109.4
-8.1%
-34.9
-2.8%
-54.2
-4.5%
1,152.9
Electric Power
Natural Gas
82.3
18.6%
19.8
3.8%
-39.4
-7.2%
71.7
14.2%
577.4
Electric Power
Petroleum
-1.6
-6.4%
-2.2
-9.3%
-2.5
-11.8%
3.3
17.4%
22.2
Transportation
Petroleum
12.4
0.7%
39.4
2.3%
19.8
1.1%
25.5
1.5%
1,770.5
Residential
Natural Gas
-24.9
-9.0%
-14.3
-5.7%
3.1
1.3%
32.3
13.4%
273.7
Commercial
Natural Gas
-13.8
-7.3%
-4.9
-2.8%
2.6
1.6%
19.4
11.2%
192.6
Industrial
Natural Gas
-2.6
-0.6%
10.4
2.2%
11.0
2.3%
29.0
6.0%
514.8
All Sectors3
All Fuels3
-153.0
-3.0%
-89.3
-1.8%
-50.2
-1.0%
139.6
2.9%
5,031.8
a Includes sector and fuel combinations not shown in this table.
As shown in Table 3-6, recent trends in CO2 emissions from fossil fuel combustion show a 3.0 percent decrease
from 2014 to 2015, then a 1.8 percent decrease from 2015 to 2016, a 1.0 percent decrease from 2016 to 2017, and
a 2.9 percent increase from 2017 to 2018. These changes contributed to an overall 3.0 percent decrease in CO2
emissions from fossil fuel combustion from 2014 to 2018.
Trends in CO2 emissions from fossil fuel combustion over the past five years have been largely driven by the
electric power sector, which historically has accounted for the largest portion of these emissions. The types of
fuels consumed to produce electricity have changed in recent years. Total electric power generation decreased by
1.5 percent from 2014 to 2017 but increased by 3.4 percent from 2017 to 2018. Emissions increased from 2017 to
2018 due to increasing electric power generation from natural gas and petroleum. Carbon dioxide emissions from
coal consumption for electric power generation decreased by 26.5 percent since 2014, which can be largely
attributed to a shift to the use of less-CC>2-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 CO2 emissions from fossil fuel combustion over the past five years also follow changes in heating
degree days. Emissions from natural gas consumption in the residential and commercial sectors increased by 13.4
percent and 11.2 percent from 2017 to 2018, respectively. This trend can be largely attributed to a 12 percent
increase in heating degree days, 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. 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
demands related to heating and cooling needs (EIA 2018; EIA 2019b).
Petroleum use in the transportation sector is another major driver of emissions, representing the largest source of
CO2 emissions from fossil fuel combustion in 2018. Emissions from petroleum consumption for transportation have
increased by 5.8 percent since 2014 and are primarily attributed to a 7.1 percent increase in vehicle miles traveled
(VMT) over the same time period.
In the United States, 80 percent of the energy used in 2018 was produced through the combustion of fossil fuels
such as petroleum, natural gas, and coal (see Figure 3-3 and Figure 3-4). Specifically, petroleum supplied the
largest share of domestic energy demands, accounting for 36 percent of total U.S. energy used in 2018. Natural gas
and coal followed in order of energy demand importance, accounting for approximately 31 percent and 13 percent
of total U.S. energy used, respectively. Petroleum was consumed primarily in the transportation end-use sector
and the vast majority of coal was used in the electric power sector. Natural gas was broadly consumed in all end-
use sectors except transportation (see Figure 3-5) (EIA 2019a). The remaining portion of energy used in 2018 was
3-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2019a).8
Figure 3-3: 2018 U.S. Energy Use by Energy Source (Percent)
Nuclear Electric Power
8.3%
Renewable Energy
11.3%
Petroleum
36.5%
Coal
13.1%
Natural Gas
30.8%
Figure 3-4: Annual U.S. Energy Use (Quadrillion Btu)
120
m
O"
100
80
Q.
Ł
o
u
>-
o>
60
40
20
Total Energy
Fossil Fuels
Renewable & Nuclear
T-iT-irNrMr\jr\jr\jrvjr\jrNr>jrMrMrMrMr\irMCMr\ir\i
8 Renewable energy, as defined in ElA's energy statistics, includes the following energy sources: hydroelectric power,
geothermal energy, biofuels, solar energy, and wind energy.
Energy 3-9

-------
Figure 3-5: 2018 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT
COz Eq.)
Relative Contribution by Fuel Type
<0.05%
Petroleum
Coal
Natural Gas
I Geothermal
41
U.S. Territories Commercial
Residential
1,753
1,821
Industrial	Electric Power Transportation
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 CO2 and smaller amounts of other gases,
including CFU, carbon monoxide (CO), and non-methane volatile organic compounds (NMVOCs).9 These other C-
containing non-CC>2 gases are emitted as a byproduct of incomplete fuel combustion, but are, for the most part,
eventually oxidized to CO2 in the atmosphere. Therefore, it is assumed all of the C in fossil fuels used to produce
energy is eventually converted to atmospheric CO2.
Box 3-2: Weather and Non-Fossil Energy Effects on CO2 Emissions from Fossil Fuel Combustion Trends
The United States in 2018 experienced a significantly colder winter overall compared to 2017, as heating degree
days increased 11.8 percent. Colder winter conditions compared to 2017 impacted the amount of energy
required for heating. In 2018 heating degree days in the United States were 5.7 percent below normal (see
Figure 3-6). Cooling degree days increased by 11.1 percent compared to 2017, which increased demand for air
conditioning in the residential and commercial sector. Hotter summer conditions compared to 2017 impacted
the amount of energy required for cooling, and 2018 cooling degree days in the United States were 29.2 percent
above normal (see Figure 3-7) (EIA 2019a). 10 The combination of colder winter and hotter summer conditions
led to residential and commercial energy consumption increases of 14.8 and 5.9 percent, respectively.
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).
3-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 3-6: Annual Deviations from Normal Heating Degree Days for the United States
(1950-2018, Index Normal = 100)
30 Normal
(4,538 Heating Degree Days)
c -10 99% Confidence
-20
Note: Climatological normal data are highlighted in dark red. Statistical confidence interval for "normal" climatology period of 1981
2q through 2010.
O (M	VD GO O fM ^vDOOOfM^-VDOOOfM^-vDCOOfM^- vDCOOfM^-vOCOOfM^-vDOO
ir)Lnir)LnLnkDvoiŁ>vDiorN.rvrv.rvrvoooococooo(nCTicncricr»ooooo^'^i-^-i'-iT-H
tJlO*OlfflOlOlOlOlO*tJlO>tJlCJlOlOlO^Ol0lOlOltJlOlCJlOlOlOOOOOOOOOO
l—i l—i i—i t—i i—i i—i i—i i—i i—i i—i i—i h i-H i—i l—i i—i i—i iHi iHi i—i iHIt—i i—ii iHi iHi rsj r\J rsj rsJ r\i rsi rsi rsi r\i rsi
Figure 3-7: Annual Deviations from Normal Cooling Degree Days for the United States
(1950-2018, Index Normal = 100)
40
Normal
(1,228 cooling degree days)
30
20
99% Confidence
10
0
-10
-30
Note: Climatological normal data are highlighted dark blue. Statistical confidence interval for "normal" climatology period of 1981
through 2010.
-40
O fM
m Ln
CO O fM
m kd kd
VD
VD CO O fM
VD VD hs rv
«D 00 O fM
N IN, 00 00
00
VD 00 O fM
oo oo cn a*
vO 00
cxi cn
o
o
fM
O
o
00
o
ID 00
i—i i—l
Energy 3-11

-------
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 2018 remained high at 93 percent. In
2018, 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), such that, they have become relatively
important electricity sources. Between 1990 and 2018, renewable energy generation (in kWh) from solar and wind
energy have increased from 0.1 percent in 1990 to 8 percent of total electricity generation in 2018, which helped
drive the decrease in the carbon intensity of the electricity supply in the United States.
Fossil Fuel Combustion Emissions by Sector
In addition to the CO2 emitted from fossil fuel combustion, Cm and N2O are emitted from stationary and mobile
combustion as well. Table 3-7 provides an overview of the CO2, Cm, and N2O emissions from fossil fuel combustion
by sector.
Table 3-7: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion by Sector (MMT CO2
Eq.)
End-Use Sector
1990

2005

2014
2015
2016
2017
2018
Transportation
1,524.0

1,903.0

1,737.6
1,747.3
1,786.1
1,806.8
1,839.0
C02
1,469.1

1,856.1

1,713.7
1,725.3
1,765.3
1,787.3
1,820.7
ch4
12.9

9.6

4.1
3.6
3.4
3.3
3.1
n2o
42.0

37.3

19.7
18.3
17.4
16.3
15.2
Electric Power
1,840.9

2,430.9

2,067.1
1,928.3
1,836.2
1,757.9
1,778.5
C02
1,820.0

2,400.0

2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
ch4
0.4

0.9

1.1
1.2
1.2
1.1
1.2
n2o
20.5

30.1

28.9
26.5
26.2
24.8
24.4
Industrial
861.9

854.7

817.2
805.6
805.6
809.3
837.5
C02
857.0

850.1

812.9
801.3
801.4
805.0
833.2
ch4
1.8

1.7

1.6
1.6
1.6
1.6
1.6
n2o
3.1

2.9

2.7
2.6
2.6
2.6
2.6
Residential
344.5

362.9

352.8
323.1
297.9
298.4
342.7
C02
338.2

357.9

346.8
317.8
293.1
293.8
337.3
ch4
5.2

4.1

5.0
4.5
3.9
3.8
4.5
n2o
1.0

0.9

1.0
0.9
0.8
0.8
0.9
Commercial
229.7

228.3

234.3
247.0
233.9
234.3
248.1
C02
228.2

226.9

232.8
245.4
232.3
232.8
246.5
ch4
1.1

1.1

1.1
1.2
1.2
1.2
1.2
n2o
0.4

0.3

0.3
0.4
0.3
0.3
0.3
U.S. Territories3
27.7

49.9

41.5
41.5
41.5
41.5
41.5
Total
4,828.7

5,829.7

5,250.5
5,092.8
5,001.2
4,948.2
5,087.2
Notes: 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.
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
(2019c).
3-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Other than CO2, gases emitted from stationary combustion include the greenhouse gases CH4 and N2O and
greenhouse gas precursors nitrogen oxides (NOx), CO, and NMVOCs.12 Methane and N2O emissions from stationary
combustion sources depend upon fuel characteristics, size, and vintage, 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 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, N2O, and greenhouse gas precursors including NOx, CO, and
NMVOCs. As with stationary combustion, N2O 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 idle, low speed, and cold start conditions. Methane and NMVOC emissions from motor vehicles are a
function of the Cm 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 were defined: transportation, industrial, residential, and
commercial. In the table 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 N2O from transportation.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 3-8.
Table 3-8: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion by End-Use Sector
(MMT COz Eq.)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation
1,527.1
1,907.7
1,742.0
1,751.5
1,790.3
1,811.1
1,843.8
C02
1,472.1
1,860.8
1,718.2
1,729.5
1,769.5
1,791.6
1,825.4
ch4
12.9
9.6
4.1
3.6
3.4
3.3
3.1
n2o
42.0
37.3
19.7
18.3
17.4
16.3
15.2
Industrial
1,556.2
1,600.5
1,418.9
1,363.1
1,331.0
1,321.2
1,331.8
C02
1,543.4
1,586.4
1,405.9
1,350.8
1,319.0
1,309.4
1,320.4
ch4
2.0
2.0
1.9
1.9
1.9
1.9
2.0
n2o
10.8
12.2
11.1
10.3
10.1
9.9
9.4
Residential
944.1
1,229.9
1,097.7
1,016.9
961.2
924.7
1,001.6
C02
931.0
1,213.9
1,080.9
1,001.6
946.6
910.9
986.7
ch4
5.4
4.4
5.4
4.9
4.3
4.2
5.0
n2o
7.7
11.6
11.4
10.5
10.3
9.6
10.0
Commercial
773.6
1,041.6
950.3
919.7
877.1
849.6
868.5
C02
765.9
1,029.9
938.5
908.5
866.0
839.0
858.0
ch4
1.2
1.4
1.5
1.6
1.6
1.6
1.6
12	Sulfur dioxide (S02) emissions from stationary combustion are addressed in Annex 6.3.
13	Separate calculations were performed for transportation-related CH4 and N20. The methodology used to calculate these
emissions are 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.
Energy 3-13

-------
N20	6.5	10.4	10.3	9.6	9.5	9.0	8.9
U.S. Territories3	27.7	49.9	41.5 41.5 41.5 41.5 41.5
Total	4,828.7	5,829.7	5,250.5 5,092.8 5,001.2 4,948.2 5,087.2
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.
a U.S. Territories are not apportioned by sector, and emissions are total greenhouse gas emissions from all
fuel combustion sources.
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 3-9 presents CO2 emissions from
fossil fuel combustion by stationary sources. The CO2 emitted is closely linked to the type of fuel being combusted
in each sector (see Methodology section of CO2 from Fossil Fuel Combustion). Other than CO2, gases emitted from
stationary combustion include the greenhouse gases CH4 and N2O. Table 3-10 and Table 3-11 present Cm and N2O
emissions from the combustion of fuels in stationary sources. The Cm and N2O 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 Cm and N2O from Stationary Combustion). Table 3-7
presents the corresponding direct CO2, Cm, and N2O emissions from all sources of fuel combustion, without
allocating emissions from electricity use to the end-use sectors.
Table 3-9: CO2 Emissions from Stationary Fossil Fuel Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2014
2015
2016
2017
2018
Electric Power
1,820.0
2,400.0
2,037.1
1,900.6
1,808.9
1,732.0
1,752.8
Coal
1,546.5
1,982.8
1,568.6
1,351.4
1,242.0
1,207.1
1,152.9
Natural Gas
175.4
318.9
442.9
525.2
545.0
505.6
577.4
Fuel Oil
97.5
97.9
25.3
23.7
21.4
18.9
22.2
Geothermal
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Industrial
857.0
850.1
812.9
801.3
801.4
805.0
833.2
Coal
155.2
115.3
76.0
66.3
59.2
54.4
49.8
Natural Gas
408.5
388.6
467.0
464.4
474.8
485.8
514.8
Fuel Oil
293.3
346.2
269.9
270.5
267.4
264.8
268.6
Commercial
228.2
226.9
232.8
245.4
232.3
232.8
246.5
Coal
12.0
9.3
3.8
3.0
2.3
2.0
1.8
Natural Gas
142.0
162.9
189.2
175.4
170.5
173.2
192.6
Fuel Oil
74.2
54.7
39.8
67.1
59.5
57.6
52.1
Residential
338.2
357.9
346.8
317.8
293.1
293.8
337.3
Coal
3.0
0.8
NO
NO
NO
NO
NO
Natural Gas
237.8
262.2
277.7
252.7
238.4
241.5
273.7
Fuel Oil
97.4
94.9
69.1
65.1
54.8
52.3
63.5
U.S. Territories
27.6
49.7
41.4
41.4
41.4
41.4
41.4
Coal
0.6
3.0
4.0
4.0
4.0
4.0
4.0
Natural Gas
NO
1.3
3.0
3.0
3.0
3.0
3.0
Fuel Oil
26.9
45.4
34.3
34.3
34.3
34.3
34.3
Total
3,270.9
3,884.5
3,471.1
3,306.5
3,177.1
3,105.0
3,211.2
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
Table 3-10: ChU Emissions from Stationary Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2014
2015
2016
2017
2018
Electric Power
0.4
0.9
1.1
1.2
1.2
1.1
1.2
Coal
0.3
0.4
0.3
0.3
0.2
0.2
0.2
Fuel Oil
+
+
+
+
+
+
+
3-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Natural gas
0.1
0.5
0.8
0.9
0.9
0.9
1.0
Wood
+
+
+
+
+
+
+
Industrial
1.8
1.7
1.6
1.6
1.6
1.6
1.6
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
1.1
1.0
1.1
1.1
Commercial
1.1
1.1
1.1
1.2
1.2
1.2
1.2
Coal
+
+
+
+
+
+
+
Fuel Oil
0.3
0.2
0.1
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
0.5
0.6
0.6
0.6
0.6
Residential
5.2
4.1
5.0
4.5
3.9
3.8
4.5
Coal
0.2
0.1
NO
NO
NO
NO
NO
Fuel Oil
0.3
0.3
0.3
0.2
0.2
0.2
0.2
Natural Gas
0.5
0.6
0.6
0.6
0.5
0.5
0.6
Wood
4.1
3.1
4.1
3.7
3.2
3.1
3.7
U.S. Territories
+
0.1
0.1
0.1
0.1
0.1
0.1
Coal
+
+
+
+
+
+
+
Fuel Oil
+
0.1
0.1
0.1
0.1
0.1
0.1
Natural Gas
NO
+
+
+
+
+
+
Wood
NO
NO
NO
NO
NO
NO
NO
Total
8.6
7.8
8.9
8.5
7.9
7.8
8.6
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
NO (Not Occurring)
Table 3-11: N2O Emissions from Stationary Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2014
2015
2016
2017
2018
Electric Power
20.5
30.1
28.9
26.5
26.2
24.8
24.4
Coal
20.1
28.0
25.7
22.8
22.4
21.2
20.3
Fuel Oil
0.1
0.1
+
+
+
+
+
Natural Gas
0.3
1.9
3.1
3.7
3.8
3.6
4.1
Wood
+
+
+
+
+
+
+
Industrial
3.1
2.9
2.7
2.6
2.6
2.6
2.6
Coal
0.7
0.5
0.4
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
1.7
1.7
1.7
Commercial
0.4
0.3
0.3
0.4
0.3
0.3
0.3
Coal
0.1
+
+
+
+
+
+
Fuel Oil
0.2
0.1
0.1
0.2
0.2
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
1.0
0.9
0.8
0.8
0.9
Coal
+
+
0.0
0.0
0.0
0.0
NO
Fuel Oil
0.2
0.2
0.2
0.2
0.1
0.1
0.2
Natural Gas
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wood
0.7
0.5
0.7
0.6
0.5
0.5
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
0.1
0.1
0.1
0.1
Natural Gas
NO
+
+
+
+
+
+
Energy 3-15

-------
Wood
NO
NO
NO NO NO	NO NO
Total	25.1	343	33.0 30.5 30.0 28.6 28.4
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
NO (Not Occurring)
Electric Power Sector
The process of generating electricity is the largest stationary source of CO2 emissions in the United States,
representing 32 percent of total CO2 emissions from all CO2 emissions sources across the United States. Methane
and N2O accounted for a small portion of total greenhouse gas emissions from electric power, representing 0.1
percent and 1.4 percent, respectively. Electric power also accounted for 34.8 percent of CO2 emissions from fossil
fuel combustion in 2018. Methane and N2O from electric power represented 10.3 and 55.9 percent of total CH4
and N2O emissions from fossil fuel combustion in 2018, 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 GHG emissions from the electric power sector have decreased by 3.4 percent since 1990. The carbon
intensity of the electric power sector, in terms of CO2 Eq. per QBtu, has significantly decreased - by 13 percent -
during that same timeframe with the majority of the emissions and carbon intensity decreases occurring in the
past decade. This decoupling of electric power generation and the resulting CO2 emissions is shown below in
Figure 3-8. 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 28 percent in
2018.16 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 29-year period to represent 34 percent of electric power sector generation in 2018
(see Table 3-12). Natural gas has a much lower carbon content than coal, which has led to lower emissions as
natural gas replaces coal-powered electricity generation. In 2018, natural gas had a carbon content of 0.0049 kg
C/kWh (14.43 MMT C/QBtu) while coal had a carbon content of 0.0089 kg C/kWh (26.09 MMT C/QBtu).
Table 3-12: Electric Power Generation by Fuel Type (Percent)
Fuel Type
1990
2005

2014
2015
2016
2017
2018
Coal
54.1%
51.1%

39.9%
34.2%
31.4%
30.9%
28.4%
Natural Gas
10.7%
17.5%

26.3%
31.6%
32.7%
30.9%
34.1%
Nuclear
19.9%
20.0%

20.3%
20.4%
20.6%
20.8%
20.1%
Renewables
11.3%
8.3%

12.8%
13.0%
14.7%
16.8%
16.7%
Petroleum
4.1%
3.0%

0.7%
0.7%
0.6%
0.5%
0.6%
Other Gases3
+%
0.1%

0.1%
0.1%
0.1%
0.1%
0.1%
Net Electricity Generation
(Billion kWh)b
2,905
3,902
3,936
3,917
3,917
3,877
4,009
+ Does not exceed 0.05 percent.
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 2019a).
3-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
In 2018, CO2 emissions from the electric power sector increased by 1.2 percent relative to 2017. This increase in
CO2 emissions was a result of an increase in fossil fuels consumed to produce electricity in the electric power
sector. Consumption of coal for electric power decreased by 4.5 percent while consumption of natural gas
increased 14.2 percent from 2017 to 2018. 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 2017 to 2018 (see Table 3-12). The decrease in coal-powered electricity generation and increase in
renewable energy electricity generation contributed to a decoupling of emissions trends from electric power
generation trends over the recent time series (see Figure 3-8).
Decreases in natural gas costs and the associated increase in natural gas generation, particularly between 2005
and 2018, 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 47 percent while the cost of
coal (in $/MMBtu) increased by 78 percent (EIA 2019a). Also, between 1990 and 2018, renewable energy
generation (in kWh) from wind and solar energy have increased from 0.1 percent of total generation in 1990 to 8
percent in 2018, 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 42 percent, from 2,713 billion kWh in 1990
to 3,860 billion kWh in 2018.
Figure 3-8: Fuels Used in Electric Power Generation (TBtu) and Total Electric Power Sector
CO2 Emissions
50,000
40,000
30,000
>.
Is 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
60
40
20
0
OrHfNm*riniDNcooi
O")	CT! 0"t CTi O^t	CT> 	
-------
Figure 3-9: Electric Power Retail Sales by End-Use Sector (Billion kWh)
1,600
Residential
1,400
I
Commercial
1 1,200
m
Industrial
1,000
800
In 2018, electricity sales to the residential and commercial end-use sectors, as presented in Figure 3-9, increased
by 6.6 percent and 2.1 percent relative to 2017, respectively. Electricity sales to the industrial sector in 2018
increased approximately 1.8 percent relative to 2017. Overall, in 2018, the amount of electricity retail sales (in
kWh) increased by 3.7 percent relative to 2017.
Industrial Sector
Industrial sector CO2, CFU, and N2O, emissions accounted for 17,14, and 6 percent of CO2, CFU, and N2O, emissions
from fossil fuel combustion, respectively in 2018. Carbon dioxide, CH4, and N2O 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 vast majority of the energy use (EIA 2019a; 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
impacts on heating 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 2017 to 2018, total industrial production and manufacturing output increased by 3.9 percent (FRB 2019).
Over this period, output increased across production indices for Food, Petroleum Refineries, Chemicals, Primary
Metals, and Nonmetallic Mineral Products, and decreased slightly for Paper (see Figure 3-10). In 2018, CO2, CFU,
and N2O emissions from fossil fuel combustion and electricity use within the industrial end-use sector totaled
1,331.8 MMT CO2 Eq., a 0.8 percent increase from 2017 emissions.
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 2017 to
2018, 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
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.
3-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
manufacturing; pulp, paper and print; food processing, beverages and tobacco; minerals manufacturing; and
agriculture-forest-fisheries.18
Figure 3-10: Industrial Production Indices (Index 2012=100)
140
Total Industrial excluding Computers, Communications Equipment, and Semiconductors
120
100
Total Industrial
80
60
140
Paper
120
100
Food
80
60
140
Stone, Clay, and Glass Products
120
100
Chemicals
80
60
140
Primary Metals
120
100
80
60
Petroleum Refineries
Ot-irMra^-Lrj^rxooa»OiHrMr^^-irj^rN.coCT»o^-irvir^^ir)vorNCX3
CT»CT»CT»CT»0*01CT>CT>^CT»00000000001-Ii-Ii—li—li—It—It—It—IiH
a->CTiCTia->ooooooooooooooooooo
t—It—l»—li—li—li—li—li—li—iTHfNrMfMfNrsirslfNirMrMrM (M fN fN f\l fN f\l CM IN fM
Despite the growth in industrial output (69 percent) and the overall U.S. economy (99 percent) from 1990 to 2018,
CO2 emissions from fossil fuel combustion in the industrial sector decreased by 2.8 percent over the same time
series. 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.
18 Further details on industrial sector combustion emissions are provided by EPA's GHGRP. See
.
Energy 3-19

-------
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 2018 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
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, etc.) 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 2018 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 increased 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. In the later part of the time series, energy use and emissions
begin to decouple due to decarbonization of the electric power sector (see Figure 3-11).
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 .
20	See .
3-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 3-11: Fuels Used in Residential and Commercial Sectors (TBtu), Heating and Cooling
Degree Days, and Total Sector CO2 Emissions
25,000
20,000
3
4-J
CD
b 15,000
a)
u>
z>
>¦
ai
Ł 10,000
LLI
5,000
0
In 2018 the residential and commercial sectors accounted for 7 and 5 percent of CO2 emissions from fossil fuel
combustion, respectively; 38 and 10 percent of CFU emissions from fossil fuel combustion, respectively; and 2 and
1 percent of N2O 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.
Coal consumption was a minor component of energy use for the commercial sector and did not contribute to any
energy use in the residential sector. In 2018, total emissions (CO2, CH4, and N2O) from fossil fuel combustion and
electricity use within the residential and commercial end-use sectors were 1.001.6 MMT CO2 Eq. and 868.5 MMT
CO2 Eq., respectively. Total CO2, CFU, and N2O emissions from combined fossil fuel combustion and electricity use
within the residential and commercial end-use sectors increased by 8.3 and 2.2 percent from 2017 to 2018,
respectively. This trend can be largely be attributed to a 12 percent increase in heating degree days, which led to
an increased demand for heating fuel in these sectors.
In 2018, combustion emissions from natural gas consumption represented 81 and 78 percent of the direct fossil
fuel CO2 emissions from the residential and commercial sectors, respectively. Carbon dioxide emissions from
natural gas combustion in the residential and commercial sectors in 2018 increased by 13.4 percent and 11.2
percent from 2017 to 2018, 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 CO2 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, CO2, CFU, and N2O emissions are not
presented for U.S. Territories in the tables above by sector, though the emissions will include some 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
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
CTl
>
X
O)
¦o
c
fN ro in io N 00
Q CTt CD CTi CT* CT") CT"i
CTl ffi Ol Cl CTl Ol Ol
0-i-i
-------
Table 3-8. Table 3-7 presents direct CO2, Cm, and N2O 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 1,843.8 MMT CO2 Eq. in 2018,
which represented 35 percent of CO2 emissions, 27 percent of CFU emissions, and 35 percent of N2O emissions
from fossil fuel combustion, respectively.21 Fuel purchased in the United States for international aircraft and
marine travel accounted for an additional 123.3 MMT CO2 Eq. in 2018; 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 2018, transportation emissions from fossil fuel combustion rose by 21 percent due, in large part, to
increased demand for travel (see Figure 3-12). The number of vehicle miles traveled (VMT) by light-duty motor
vehicles (passenger cars and light-duty trucks) increased 46 percent from 1990 to 2018,22 as a result of a
confluence of factors including population growth, economic growth, urban sprawl, and periods of low fuel prices.
From 2017 to 2018, CO2 emissions from the transportation end-use sector increased by 1.9 percent. The increase
in emissions is attributed to an increase in on-road and non-road fuel use, particularly by passenger cars, medium-
and heavy-duty trucks, and pipelines.
Commercial aircraft emissions increased between 2017 and 2018, but have decreased 7 percent since 2007 (FAA
2019).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 CO2 from fossil fuel combustion, which increased by 24 percent from 1990 to
2018. Annex 3.2 presents the total emissions from all transportation and mobile sources, including CO2, Cm, N2O
and HFCs.
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.
22	VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2018). 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 1990 and 2018
would likely have been even higher.
23	Commercial aircraft, as modeled in FAA's AEDT (FAA 2019), consists of passenger aircraft, cargo, and other chartered flights.
3-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 3-12: Fuels Used in Transportation Sector (TBtu), Onroad VMT, and Total Sector CO2
Emissions
40,000
35,000
30,000
25,000
20,000
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]
Sector CO2 Emissions (Index vs. 1990) [Right Axis]
200
180
160
140
120
100
80
60
40
20
>
6
"O
o^-HrMro^j-Lnusr-sOOO^
Oi en Oi cn cti	cn c ot o~>
cm m	m vo r*N 00
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.
Source: Information on fuel consumption was obtained from EIA (2019a).
Transportation Fossil Fuel Combustion CO2 Emissions
Domestic transportation CO2 emissions increased by 24 percent (353.3 MMT CO2) between 1990 and 2018, an
annualized increase of 0.8 percent. Among domestic transportation sources in 2018, light-duty vehicles (including
passenger cars and light-duty trucks) represented 59 percent of CO2 emissions from fossil fuel combustion,
medium- and heavy-duty trucks and buses 25 percent, commercial aircraft 7 percent, and other sources 12
percent. See Table 3-13 for a detailed breakdown of transportation CO2 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 2018, while biodiesel
consumption has increased from 0.01 billion gallons in 2001 to 1.9 billion gallons in 2018. For additional
information, see Section 3.11 on biofuel consumption at the end of this chapter and Table A-98 in Annex 3.2.
Carbon dioxide emissions from passenger cars and light-duty trucks totaled 1,069.5 MMT CO2 in 2018. This is an
increase of 16 percent (145.0 MMT CO2) from 1990 due, in large part, to increased demand for travel as fleet-wide
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 .
Energy 3-23

-------
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 2018). Carbon dioxide emissions from
passenger cars and light-duty trucks peaked at 1,151.5 MMT CO2 in 2004, and since then have declined about 7
percent. The decline in new light-duty vehicle fuel economy between 1990 and 2004 (Figure 3-13) is 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 20 13,25 then grew at a faster rate until 2016 (2.6 percent from 2014 to 2015, and 2.5 percent
from 2015 to 2016). Since 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 the light-duty truck share
decreased to about 33 percent in 2009 and has since varied from year to year between 36 and 48 percent. Light-
duty truck share is about 48 percent of new vehicles in model year 2018 (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 CO2 emissions increased by 87 percent from 1990 to 2018. This increase was largely
due to a substantial growth in medium- and heavy-duty truck VMT, which increased by 113 percent between 1990
and 2018.26 Carbon dioxide from the domestic operation of commercial aircraft increased by 18 percent (19.7
MMT CO2) from 1990 to 2018.27 Across all categories of aviation, excluding international bunkers, CO2 emissions
decreased by 7 percent (13.5 MMT CO2) between 1990 and 20 18.28 This includes a 66 percent (23.2 MMT CO2)
decrease in CO2 emissions from domestic military operations.
Transportation sources also produce CH4 and N2O; these emissions are included in Table 3-14 and Table 3-15 and
in the Cm and N2O from Mobile Combustion section. Annex 3.2 presents total emissions from all transportation
and mobile sources, including CO2, Cm, N2O, and HFCs.
25	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.
26	While FHWA data shows consistent growth in medium- and heavy-duty truck VMT over the 1990 to 2018 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 2018 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.
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.
3-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 3-13: Sales-Weighted Fuel Economy of New Passenger Cars and Light-Duty Trucks,
1990-2018 (miles/gallon)
r\i ro -si- m	r*s oo
1—1 T—I T-H T—I 1—1	1—1 1—|
o o o o o o o
Source: EPA (2019b).
Figure 3-14: Sales of New Passenger Cars and Light-Duty Trucks, 1990-2018 (Percent)
100%
90%
80%
a! 70%
Passenger Cars
60%
50%
40%
% Light-Duty Trucks
20%
10%
0%
Source: EPA (2019b).
Energy 3-25

-------
Table 3-13: CO2 Emissions from Fossil Fuel Combustion in Transportation End-Use Sector
(MMT COz Eq.)
Fuel/Vehicle Type
1990
2005
2014a
2015a
2016a
2017a
2018a
Gasolineb
958.9
1,152.7
1,077.4
1,070.0
1,095.3
1,091.7
1,107.7
Passenger Cars
604.3
638.6
730.2
732.0
744.9
744.2
756.0
Light-Duty Trucks
300.6
464.6
292.2
283.5
294.6
290.9
293.7
Medium- and Heavy-Duty





41.2
42.3
Trucks0
37.7
33.9
39.8
39.3
40.4


Buses
0.3
0.4
0.9
0.9
0.9
1.0
1.0
Motorcycles
1.7
1.6
3.8
3.7
3.8
3.7
3.8
Recreational Boatsd
14.3
13.8
10.6
10.6
10.7
10.7
10.8
Distillate Fuel Oil (Diesel)b
262.9
457.5
439.9
452.2
449.2
463.2
474.5
Passenger Cars
7.9
4.2
4.1
4.2
4.2
4.3
4.3
Light-Duty Trucks
11.5
25.8
13.6
13.7
13.9
13.9
14.0
Medium- and Heavy-Duty





377.5
386.2
Trucks0
190.5
360.2
354.7
362.4
365.2


Buses
8.0
10.6
16.6
16.9
16.5
17.7
19.0
Rail
35.5
45.5
41.2
39.3
35.9
37.1
38.9
Recreational Boatsd
2.7
2.8
2.5
2.6
2.7
2.7
2.8
Ships and Non-Recreational





10.0
9.3
Boatse
6.8
8.4
7.3
13.0
10.8


International Bunker Fuels1
11.7
9.4
6.1
8.4
8.7
9.0
9.9
Jet Fuel
184.2
189.3
148.4
157.6
166.0
171.8
172.3
Commercial Aircraft5
109.9
132.7
115.2
119.0
120.4
128.0
129.6
Military Aircraft
35.0
19.4
14.0
13.5
12.3
12.2
11.8
General Aviation Aircraft
39.4
37.3
19.2
25.1
33.4
31.5
30.9
International Bunker Fuels1
38.0
60.1
69.6
71.9
74.1
77.7
80.8
International Bunker Fuels from





74.5
77.7
Commercial Aviation
30.0
55.6
66.3
68.6
70.8


Aviation Gasoline
3.1
2.4
1.5
1.5
1.4
1.4
1.5
General Aviation Aircraft
3.1
2.4
1.5
1.5
1.4
1.4
1.5
Residual Fuel Oil
22.6
19.3
5.8
4.2
12.9
16.5
13.9
Ships and Boatse
22.6
19.3
5.8
4.2
12.9
16.5
13.9
International Bunker Fuels1
53.7
43.6
27.7
30.6
33.8
33.4
31.4
Natural GasJ
36.0
33.1
40.2
39.4
40.1
42.3
50.2
Passenger Cars
+
+
+
+
+
+
+
Light-Duty Trucks
+
+
+
+
+
+
+
Medium- and Heavy-Duty





+
+
Trucks
+
+
+
+
+


Buses
+
0.6
0.8
0.9
0.8
0.9
0.9
Pipeline11
36.0
32.4
39.4
38.5
39.2
41.3
49.2
LPG J
1.4
1.7
0.4
0.4
0.4
0.4
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





0.3
0.3
Trucks0
1.1
1.3
0.3
0.3
0.3


Buses
0.1
0.1
0.1
+
0.1
0.1
0.1
Electricity1
3.0
4.7
4.4
4.3
4.2
4.3
4.7
Passenger Cars
0
0
0.4
0.5
0.6
0.8
1.2
Light-Duty Trucks
0
0
+
+
0.1
0.1
0.2
Buses
0
0
+
+
+
+
+
Rail
3.0
4.7
4.0
3.7
3.5
3.4
3.4
Totalk	1,472.1	1,860.8	1,718.2 1,729.5 1,769.5 1,791.6 1,825.4
3-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Total (Including Bunkers)'
1,575.6
1,974.0
1,821.6
1,840.4
1,886.1
1,911.7
1,947.5
Biofuels-Ethanol'
4.1
21.6
74.0
74.2
76.9
77.7
78.6
Biofuels-Biodiesel'
+
0.9
13.3
14.1
19.6
18.7
17.9
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 2018 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 2018). 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 2017). TEDB data for 2018 has not been published yet,
therefore 2017 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 2018.
e 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,
s 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
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 2018 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 2018 time period.
Mobile Fossil Fuel Combustion CH4 and N2O Emissions
Mobile combustion includes emissions of Cm and N2O 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
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-
Energy 3-27

-------
both transportation and mobile sources. Table 3-14 and Table 3-15 provide mobile fossil fuel CFU and N2O emission
estimates in MMT CO2 Eq.31
Mobile combustion was responsible for a small portion of national CH4 emissions (0.5 percent) and was the fourth
largest source of national N2O emissions (3.5 percent). From 1990 to 2018, mobile source CFU emissions declined
by 76 percent, to 3.1 MMT CO2 Eq. (125 kt CH4), due largely to control technologies employed in on-road vehicles
since the mid-1990s to reduce CO, NOx, NMVOC, and CFU emissions. Mobile source emissions of N2O decreased by
64 percent, to 15.2 MMT CO2 Eq. (51 kt N2O). Earlier generation control technologies initially resulted in higher
N2O emissions, causing a 30 percent increase in N2O emissions from mobile sources between 1990 and 1997.
Improvements in later-generation emission control technologies have reduced N2O emissions, resulting in a 72
percent decrease in mobile source N2O emissions from 1997 to 2018 (Figure 3-15). Overall, CFU and N2O 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).
Figure 3-15: Mobile Source ChU and N2O Emissions (MMT CO2 Eq.)
Table 3-14: ChU Emissions from Mobile Combustion (MMT CO2 Eq.)
Fuel Type/Vehicle Type3
1990

2005

2014
2015
2016
2017
2018
Gasoline On-Roadb
5.2

2.2

1.1
1.0
0.9
0.8
0.7
Passenger Cars
3.2

1.3

0.7
0.6
0.6
0.5
0.5
Light-Duty Trucks
1.7

0.8

0.3
0.2
0.2
0.2
0.2
Medium- and Heavy-Duty









Trucks and Buses
0.3

0.1

0.1
0.1
0.0
0.0
0.0
Motorcycles
+

+

+
+
+
+
+
Diesel On-Roadb
+

+

0.1
0.1
0.1
0.1
0.1
Passenger Cars
+

+

+
+
+
+
+
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 2018.
3-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Light-Duty Trucks
+
+
+
+
+
+
+
Medium- and Heavy-Duty
+
+
+
+
0.1
0.1
0.1
Trucks and Buses







Alternative Fuel On-Road
+
0.2
0.2
0.2
0.2
0.2
0.2
Non-Roadc
7.7
7.2
2.8
2.4
2.3
2.2
2.1
Ships and Boats
0.6
0.5
0.3
0.3
0.3
0.3
0.3
Rail
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Aircraft
0.1
0.1
+
+
+
+
+
Agricultural Equipment
0.6
0.6
0.2
0.1
0.1
0.1
0.1
Construction/Mining
0.9
1.0
0.6
0.5
0.4
0.4
0.4
Equipment6







Other'
5.5
4.9
1.6
1.5
1.4
1.3
1.3
Total
12.9
9.6
4.1
3.6
3.4
3.3
3.1
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 2018 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 (FHWA 1996 through 2018). These mileage 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). TEDB data for 2018 has
not been published yet, therefore 2017 data are used as a proxy.
c Rail emissions do not include emissions from electric powered locomotives. Class II and Class III diesel
consumption data for 2014-2018 is estimated by applying the historical average fuel usage per carload factor to
the annual number of carloads. Intercity rail diesel consumption data for 2017 and 2018 is not available yet,
therefore 2016 data are used as a proxy. Commuter rail diesel consumption data for 2018 is not available yet,
therefore 2017 data are used as a proxy.
d Includes equipment, such as tractors and combines, as well as fuel consumption from trucks that are used off-
road in agriculture.
e 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
2014
2015
2016
2017
2018
Gasoline On-Roadb
37.5
31.8
13.3
11.6
10.2
8.7
7.3
Passenger Cars
24.1
17.3
9.0
8.0
7.0
6.0
5.1
Light-Duty Trucks
12.8
13.6
3.8
3.1
2.7
2.3
1.9
Medium- and Heavy-Duty







Trucks and Buses
0.5
0.9
0.5
0.4
0.4
0.3
0.3
Motorcycles
+
+
+
+
+
+
+
Diesel On-Roadb
0.2
0.3
1.9
2.2
2.4
2.6
2.9
Passenger Cars
+
+
+
+
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
1.8
2.0
2.2
2.4
2.7
Alternative Fuel On-Road
+
+
0.1
0.1
0.2
0.2
0.2
Non-Road
4.4
5.2
4.4
4.4
4.6
4.8
4.9
Ships and Boats
0.6
0.6
0.3
0.4
0.5
0.5
0.5
Energy 3-29

-------
Railc
0.3
0.3
0.3
0.3
0.3
0.3
0.3
Aircraft
1.7
1.8
1.4
1.5
1.5
1.6
1.6
Agricultural Equipment
0.5
0.6
0.6
0.6
0.6
0.5
0.5
Construction/Mining
S






Equipment6
0.6
1.0
0.8
0.8
0.8
0.9
0.9
Other'
0.6
0.9
1.0
1.0
1.0
1.0
1.0
Total	42.0	37.3	19.7	18.3	17.4	16.3	15.2
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 2018 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
(FHWA 1996 through 2018). These mileage 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). TEDB data for 2018 has not been
published yet, therefore 2016 data are used as a proxy.
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. Intercity rail diesel consumption data for 2017 and 2018 is not available yet, therefore 2016
data are used as a proxy. Commuter rail diesel consumption data for 2018 is not available yet, therefore 2017 data
are used as a proxy.
d Includes equipment, such as tractors and combines, as well as fuel consumption from trucks that are used off-road
in agriculture.
e 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.
C02 from Fossil Fuel Combustion
Methodology
CO2 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 2019a). EIA data includes 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 2017).33
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 41.4 MMT C02 Eq. in 2018.
3-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 data sets 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 2018), Coffeyville (2012), U.S. Census
Bureau (2001 through 2011), EIA (2020a, 2019a, 2019d), USAA (2008 through 2018), USGS (1991 through
2015a), (USGS 2018b), USGS (2014 through 2019b), USGS (2014 through 2017), USGS (1995 through
2013), USGS (1995, 1998, 2000, 2001, 2002, 2007), USGS (2019), USGS (1991 through 2015c), USGS (1991
through 2017), USGS (2014 through 2019a), USGS (1996 through 2013), USGS (1991 through 2015b),
USGS (2020), USGS (1991 through 2015c).36
3.	Adjust for biofuels, conversion of fossil fuels, and exports ofCC>2. Fossil fuel consumption estimates are
adjusted downward to exclude (1) fuels with biogenic origins, (2) fuels created from other fossil fuels, and
(3) exports of CO2. 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 Synthetic natural gas is created
from industrial coal and is currently included in EIA statistics for coal. Therefore, synthetic natural gas is
subtracted from coal consumption statistics.38 Since October 2000, the Dakota Gasification Plant has been
exporting CO2 to Canada by pipeline. Since this CO2 is not emitted to the atmosphere in the United States,
the associated fossil fuel burned to create the exported CO2 is subtracted from coal consumption
statistics. The associated fossil fuel is the total fossil fuel burned at the plant with the CO2 capture system
multiplied by the fraction of the plant's total site-generated CO2 that is recovered by the capture system.
To make these adjustments, additional data were collected from EIA (2019a), data for synthetic natural
gas were collected from EIA (2019d), and data for CO2 exports were collected from the Eastman
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.
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 (2019a) are already adjusted downward to account for biogas in natural gas.
38	These adjustments are explained in greater detail in Annex 2.1.
Energy 3-31

-------
Gasification Services Company (2011), Dakota Gasification Company (2006), Fitzpatrick (2002), Erickson
(2003), EIA (2008), and DOE (2012).
4.	Adjust Sectoral Allocation of Distillate Fuel Oil and Motor Gasoline. EPA had 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
5.	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 (2019a).
6.	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
international transport activities were calculated separately following the same procedures used for
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 2019)
supplied data on military jet fuel and marine fuel use. Commercial jet fuel use was obtained from FAA
(2019); 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 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.
7.	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 CO2. The
Carbon content coefficients used by the United States were obtained from ElA's Emissions of Greenhouse
Gases in the United States 2008 (EIA 2009a), and an EPA analysis of Carbon content coefficients
developed for the GHGRP (EPA 2010). A discussion of the methodology used to develop the Carbon
content coefficients are presented in Annexes 2.1 and 2.2.
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 2018).
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.
3-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
8.	Estimate CO2 Emissions. Total CO2 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 CO2 to C (44/12) to obtain total CO2 emitted from fossil fuel
combustion in million metric tons (MMT).
9.	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 (2019a) 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 2018); for each vehicle category, the
percent gasoline, diesel, and other (e.g., CNG, LPG) fuel consumption are estimated using data from
DOE (1993 through 2017). 43'44
•	For non-road vehicles, activity data were obtained from AAR (2008 through 2018), APTA (2007
through 2017), APTA (2006), BEA (2020), Benson (2002 through 2004), DLA Energy (2019), DOC (1991
through 2019), DOE (1993 through 2017), DOT (1991 through 2018), EIA (2009a), EIA (2019a), EIA
(2019f), EIA (1991 through 2018), EPA (2018),45 and Gaffney (2007).
•	For jet fuel used by aircraft, CO2 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.
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 2015. 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 (2019a). 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.
45	In 2014, EPA incorporated the NONROAD2008 model into MOVES2014. The current Inventory uses the NONROAD
component of MOVES2014b for years 1999 through 2018.
Energy 3-33

-------
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 CO2 Eq./QBtu for natural gas to upwards of 95 MMT CO2 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 CO2 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

2014
2015
2016
2017
2018
Residential3
57.4

56.6

55.4
55.5
55.1
55.0
55.2
Commercial3
59.6

57.7

55.7
57.2
56.8
56.6
56.0
Industrial3
64.5

64.5

61.5
61.2
60.8
60.5
60.2
Transportation3
71.1

71.4

71.5
71.5
71.5
71.5
71.4
Electric Powerb
87.3

85.8

81.2
78.1
76.8
77.3
75.5
U.S. Territories0
73.0

73.5

72.3
72.2
72.2
72.2
72.2
All Sectors0
73.0

73.5

70.8
69.7
69.2
69.2
68.3
Note: Excludes non-energy fuel use emissions and consumption.
3 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.
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 2018, was approximately 8.7 percent below levels in 1990
(see Figure 3-16). To differentiate these estimates from those of Table 3-16, the carbon intensity trend shown in
Figure 3-16 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 CO2
emissions per dollar of gross domestic product (GDP) have both declined since 1990 (BEA 2018).
3-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 3-16: 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)
s
s
"U
c
I—I
Energy Consumption/$GDP
o
o
UD
O
o
rv
o
o
00
o
o
cm m
00
o
o
o
fO
o
o
LO
O
o
o
1—1
m
1—1
1—1
o
o
T—I
1—1
o
1—1
o
1—1
o
o
o
Carbon intensity estimates were developed using nuclear and renewable energy data from EIA (2019a), EPA
(2010), and fossil fuel consumption data as discussed above and presented in Annex 2.1.
Uncertainty and Time-Series Consistency
For estimates of CO2 from fossil fuel combustion, the amount of CO2 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 CO2
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 CO2 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. Even with recent updates to carbon factors for natural gas and coal, the uncertainty estimates are not
impacted.
Although 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 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 problems in collecting
accurate energy statistics as firms in these industries have undergone significant restructuring.
To calculate the total CO2 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 CO2
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 CO2 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.
Energy 3-35

-------
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 CO2 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 CO2 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 CO2 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
Carlo sampling.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-17. Fossil fuel
combustion CO2 emissions in 2018 were estimated to be between 4,919.7 and 5,255.7 MMT CO2 Eq. at a 95
percent confidence level. This indicates a range of 2 percent below to 4 percent above the 2018 emission estimate
of 5,031.8 MMTCO2 Eq.
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.
3-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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)
2018 Emission Estimate Uncertainty Range Relative to Emission Estimate3
Fuel/Sector	(MMT CP2 Eq.)	(MMT CP2 Eq.)	(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Coalb
1,208.5
1,165.5
1,322.3
-4%
9%
Residential
NO
NE
NE
NE
NE
Commercial
1.8
1.7
2.1
-5%
15%
Industrial
49.8
47.4
57.6
-5%
16%
Transportation
NE
NE
NE
NE
NE
Electric Power
1,152.9
1,107.9
1,263.2
-4%
10%
U.S. Territories
4.0
3.5
4.8
-12%
19%
Natural Gasb
1,611.6
1,592.7
1,685.3
-1%
5%
Residential
273.7
266.1
292.7
-3%
7%
Commercial
192.6
187.1
206.1
-3%
7%
Industrial
514.8
499.1
551.6
-3%
7%
Transportation
50.2
48.7
53.7
-3%
7%
Electric Power
577.4
560.7
606.8
-3%
5%
U.S. Territories
3.0
2.6
3.5
-12%
17%
Petroleumb
2,211.3
2,075.5
2,338.3
-6%
6%
Residential
63.5
60.0
66.9
-6%
5%
Commercial
52.1
49.4
54.6
-5%
5%
Industrial
268.6
210.0
321.1
-22%
20%
Transportation
1,770.5
1,654.4
1,884.6
-7%
6%
Electric Power
22.2
21.2
23.9
-5%
8%
U.S. Territories
34.3
31.7
38.1
-8%
11%
Total (excluding Geothermal)b
5,031.4
4,919.2
5,255.2
-2%
4%
Geothermal
0.4
NE
NE
NE
NE
Total (including Geothermal)b'c
5,031.8
4,919.7
5,255.7
-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 2018. 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 CO2 emissions from any liquid
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 CO2 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 CO2 emissions from fossil fuel
Energy 3-37

-------
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
CO2 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 NEU source category emissions are
included in the reference approach (see Annex 4 for more details).
Recalculations Discussion
The Energy Information Administration (EIA 2019a) updated energy consumption statistics across the time series
relative to the previous Inventory.49 As a result of updated LPG and fuel ethanol heat contents, EIA updated LPG
consumption in the residential, commercial, industrial, and transportation sectors across the time series. EIA also
revised sector allocations for propane and total LPGs for 2010 through 2017, and for distillate fuel oil in 2017,
which impacted petroleum consumption by sector. EIA revised assumptions for the percentage of fossil fuels
consumed for non-combustion use which impacted the nonfuel sequestration statistics, particularly for petroleum
coke and residual fuel across the time series relative to the previous Inventory.
EIA also revised 2017 natural gas consumption in all sectors, 2017 kerosene consumption in the residential and
commercial sectors, 2009 and 2017 motor gasoline consumption in the commercial, industrial, and transportation
sectors, 1995 and 1997 through 2000 asphalt and road oil consumption in the industrial sector, 2017 residual fuel
oil and lubricants in the industrial and transportation sectors, 2017 petroleum coke consumption in the industrial
sector, 2009 through 2017 distillate fuel oil consumption in the transportation sector, and pentanes plus
consumption in the industrial sector across the time series.
To align with ElA's methodology for calculating industrial and commercial motor gasoline consumption, fuel
ethanol adjustments to motor gasoline consumption for the period 1990 through 1992 were corrected. To align
with ElA's methodology for calculating the amount of biofuel added to diesel fuel, both biodiesel and other
renewable diesel fuel were considered; EIA (2019a) data were used for 2009 forward. To improve the time series
consistency of distillate fuel oil consumption, data from ElA's Fuel Oil and Kerosene Sales Report (EIA 1991 through
2019) were used across the time series. Previously, distillate fuel oil consumption for the period 1990 through
2002 were obtained from ElA's State Energy Data System (SEDS) (EIA 1990-2002) and 2003 data were provided by
EIA (2003).
Revisions to LPG, lubricants, kerosene, jet fuel, distillate fuel, asphalt and road oil, residual fuel oil, petroleum coke,
pentanes plus, and motor gasoline consumption resulted in an average annual decrease of 6.6 MMT CO2 Eq. (0.3
percent) in CO2 emissions from petroleum. Revisions to natural gas consumption resulted in an increase of 1.0
MMT CO2 Eq. (0.1 percent) in CO2 emissions from natural gas in 2017. Overall, these changes resulted in an
49 Final estimates presented in this Inventory utilize energy statistics from ElA's Monthly Energy Review released in November
2019 (EIA 2019a). At the time of publication of this Inventory report there were no changes to energy statistics reported in later
iterations of the Monthly Energy Review.
3-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
average annual decrease of 6.6 MMT CO2 Eq. (0.1 percent) in CO2 emissions from fossil fuel combustion for the
period 1990 through 2017, relative to the previous Inventory.
As discussed in the Recalculations section of Chapter 4.5 - Ammonia Production, the carbon factors used to
determine the amount of natural gas used for ammonia feedstock were updated to be consistent with the factors
used in the fossil fuel combustion estimates. This update resulted in an annual average change to the amount of
natural gas subtracted from total natural gas consumption for energy use calculations of 0.08 percent over the
1990 to 2017 time period.
Planned Improvements
To reduce uncertainty of CO2 from fossil fuel combustion estimates for U.S. Territories, efforts will be made to
improve the quality of the U.S. Territories data, including through work with EIA and other agencies. This
improvement is part of an ongoing analysis and efforts to continually improve the CO2 from fossil fuel combustion
estimates. In addition, further expert elicitation may be conducted to better quantify the total uncertainty
associated with emissions from this source.
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, 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.50 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 CO2 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 2014. Additional work will look at
CO2 emissions from biomass to ensure they 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 GHGRP, the latest guidance from the IPCC on the use of facility-level data in national inventories will
continue to be relied upon.51
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.
In 2016, FHWA changed its methods for estimating the share of gasoline used in on-road and non-road
applications, creating a time-series inconsistency in the current Inventory between 2015 and previous years.52 EPA
50	See .
51	See .
52	The previous and new FHWA methodologies for estimating non-road gasoline are described in Off-Highway and Public-Use
Gasoline Consumption Estimation Models Used in the Federal Highway Administration, Publication Number FHWA-PL-17-012.
.
Energy 3-39

-------
has implemented an approach to address this inconsistency. EPA also tested an alternative approach that uses
MOVES on-road fuel consumption output to define the percentage of the FHWA consumption totals (from MF-21)
that are attributable to on-highway transportation sources, and applying this percentage to the EIA total, thereby
defining gasoline consumption from on-highway transportation sources (such that the remainder would be defined
as consumption by the industrial and commercial sectors). Results from this testing revealed differences between
fuel consumption calculated by MOVES and fuel consumption data from FHWA. Given this inconsistency, no
changes have been made to the methodology for estimating motor gasoline consumption for non-road mobile
sources.
EPA is also evaluating the methods used to adjust for conversion of fuels and exports of CO2. EPA is exploring the
approach used to account for CO2 transport, injection, and geologic storage, as part of this there may be changes
made to accounting for CO2 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.
EPA is currently evaluating proposed revisions to gasoline carbon factors used in this report. The current Inventory
continues to use NIPER (1990 through 2009) data to determine gasoline composition. NIPER has ceased to exist
and the current carbon factors have not been updated since 2010 (for the 1990-2008 Inventory Report). New data
and methods are available to estimate gasoline carbon factors over the time series. EPA has started reviewing data
and approaches and plans to update the gasoline carbon factors in a future report.
CH4and N20 from Stationary Combustion
Methodology
Methane and N2O 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 CFU and N2O emission estimates, consumption data for each fuel were
obtained from ElA's Monthly Energy Review (EIA 2019). 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 2017).53 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 (2018) and FHWA (1996 through 2018). 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 3.3). 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 2006IPCC
53 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.
3-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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 2020). Total fuel
consumption in the electric power sector from EIA (2019) was apportioned to each combustion technology type
and fuel combination using a ratio of fuel consumption by technology type derived from EPA (2020) data. The
combustion technology and fuel use data by facility obtained from EPA (2020) were only available from 1996 to
2018, so the consumption estimates from 1990 to 1995 were estimated by applying the 1996 consumption ratio by
combustion technology type from EPA (2020) to the total EIA (2019) 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.54
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 Cm and N2O
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 N2O
stationary source inventory estimation models with the model for CO2 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 CO2
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 N2O emission factors, based on the SAIC/EIA (2001) report.55 For these variables, the uncertainty
ranges were assigned to the input variables based on the data reported in SAIC/EIA (2001).56 However, the CFU
54	Several of the U.S. Tier 2 emission factors were used in IPCC (2006) as Tier 1 emission factors. See Table A-92 in Annex 3.1 for
emission factors by technology type and fuel type for the electric power sector.
55	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.
56	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 3-41

-------
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 Cm emissions in 2018 (including biomass) were estimated to be between 5.6 and 19.9 MMT CO2 Eq. at
a 95 percent confidence level. This indicates a range of 35 percent below to 130 percent above the 2018 emission
estimate of 8.7 MMT CO2 Eq.57 Stationary combustion N2O emissions in 2018 (including biomass) were estimated
to be between 20.7 and 42.8 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 27 percent
below to 51 percent above the 2018 emission estimate of 28.4 MMT CO2 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
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Stationary Combustion
ch4
8.6
5.6
19.9
-35% +130%
Stationary Combustion
n2o
28.4
20.7
42.8
-27% +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 N2O are greater than those associated with
estimates of CO2 from fossil fuel combustion, which mainly rely on the carbon content of the fuel combusted.
Uncertainties in both Cm and N2O 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 2018 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 Cm and N2O
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-CC>2 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
Cm, N2O, 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 N2O emissions from stationary sources (excluding CO2) across the entire time series were revised due
to revised data from EIA (2019) and EPA (2020) relative to the previous Inventory. Most notably, EIA (2019)
updated fuel oil consumption statistics in the residential, commercial, and industrial sectors across the time series
as a result of updated LPG and fuel ethanol heat contents; revised sectoral allocations for propane and total LPG
from 2010 to 2017 and for distillate fuel oil in 2017; and revised 2017 natural gas consumption statistics in all
57 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.
3-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
sectors. EPA (2020) revised coal, fuel oil, natural gas, and wood consumption statistics for 2017 in the electric
power sector. The historical data changes and methodology updates resulted in an average annual decrease of less
than 0.01 MMT CO2 Eq. (0.06 percent) in CH4 emissions, and an average annual decrease of 0.01 MMT CO2 Eq.
(0.04 percent) in N2O emissions for the 1990 through 2017 period.
Planned Improvements
Several items are being evaluated to improve the CH4 and N2O 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 since 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 Cm and N2O emissions from biogas can be included in future
inventories. EIA (2019) 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 Cm and N2O 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 are described in the subsections that follow. A complete discussion of the methodology used
to estimate Cm and N2O emissions from mobile combustion and the emission factors used in the calculations is
provided in Annex 3.2.
On-Road Vehicles
Estimates of Cm and N2O emissions from gasoline and diesel on-road vehicles are based on VMT and emission
factors 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 by vehicle and fuel type.58
Cm and N2O 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 N2O emissions factors for
older (model year 2003 and earlier) on-road gasoline vehicles were developed by ICF (2004). These factors were
derived from EPA, California Air Resources Board (CARB) and Environment Canada laboratory test results of
different vehicle and control technology types. The EPA, CARB and Environment Canada tests were designed
following the Federal Test Procedure (FTP), which 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 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 segment 2 were used to define running
58 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 3-43

-------
emissions and are subtracted from the total FTP emissions to determine start emissions. These were then
recombined 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 CO2, CO, HC, NOx, and PM from vehicles under
various conditions, to approximate average driving characteristics.59 Diesel on-road vehicle emission factors were
developed by ICF (2006a). CH4 and N2O emissions factors for newer (starting at model year 2007) on-road diesel
vehicles (those using aftertreatment) were calculated from annual vehicle certification data compiled by EPA.
Cm and N2O emission factors for AFVs were developed based on the 2018 GREET model. 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); 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 CH4 emission factors
based upon an incorrect CH4-to-THC ratio for diesel vehicles with aftertreatment technology.
Annual VMT data for 1990 through 2018 were obtained from the Federal Highway Administration's (FHWA)
Highway Performance Monitoring System database as reported in Highway Statistics (FHWA 1996 through 2018).60
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 2017)
and information on total motor vehicle fuel consumption by fuel type from FHWA (1996 through 2018). VMT for
AFVs were estimated based on Browning (2017 and 2018a). The age distributions of the U.S. vehicle fleet were
obtained from EPA (2018a, 2000), and the average annual age-specific vehicle mileage accumulation of U.S.
vehicles were obtained from EPA (2018a).
Control technology and standards data for on-road vehicles were obtained from EPA's Office of Transportation and
Air Quality (EPA 2018a, 2019c, 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
To estimate CH4 and N2O 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 N2O and CH4 per kilogram of fuel
consumed).61 Activity data were obtained from AAR (2008 through 2018), APTA (2007 through 2018), Raillnc (2014
through 2018), APTA (2006), BEA (1991 through 2015), Benson (2002 through 2004),, DLA Energy (2019), DOC
(1991 through 2019), DOE (1993 through 2017), DOT (1991 through 2018), EIA (2002, 2007, 2019a), EIA (2019f),
59	Additional information regarding the MOBILE model can be found online at .
60	The source of VMT 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 2018 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.
61	The consumption of international bunker fuels is not included in these activity data, but is estimated separately under the
International Bunker Fuels source category.
3-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
EIA (1991 through 2018), EPA (2018a), Esser (2003 through 2004), FAA (2019), FHWA (1996 through 2018),62
Gaffney (2007), and Whorton (2006 through 2014). Emission factors for non-road modes were taken from IPCC
(2006) and Browning (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 2018 estimates of CFU and N2O 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 and (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 since 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. However, a much higher level of uncertainty is associated with CH4 and N2O emission factors due to limited
emission test data, and because, unlike CO2 emissions, the emission pathways of CH4 and N2O are highly complex.
Mobile combustion CH4 emissions from all mobile sources in 2018 were estimated to be between 2.9 and 4.0 MMT
CO2 Eq. at a 95 percent confidence level. This indicates a range of 8 percent below to 27 percent above the
corresponding 2018 emission estimate of 3.1 MMT CO2 Eq. Also at a 95 percent confidence level, mobile
combustion N2O emissions from mobile sources in 2018 were estimated to be between 14.0 and 17.4 MMT CO2
Eq., indicating a range of 8 percent below to 14 percent above the corresponding 2018 emission estimate of 15.2
MMT CO2 Eq.
Table 3-19: Approach 2 Quantitative Uncertainty Estimates for ChU and N2O Emissions from
Mobile Sources (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate9
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Mobile Sources
ch4
3.1
2.9
4.0
-8% +27%
Mobile Sources
n2o
15.2
14.0
17.4
-8% +14%
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 analysis. 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 CFU and N2O please refer to Annex 7 - Uncertainty. As discussed in
Annex 5, data are unavailable to include estimates of CFU and N2O emissions from any liquid fuel used in pipeline
62 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 off-road trucks and equipment. For construction and
commercial/industrial gasoline estimates, the 2014 and older MF-24 volumes represented off-road trucks only; therefore, the
MOVES gasoline volumes for construction and commercial/industrial 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 3-45

-------
transport or some biomass used in transportation sources, but those emissions are assumed to 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 Cm and N2O emissions factors for on-road gasoline and diesel vehicles. Previously, these
factors were based on a regression analysis done by EPA for N2O and the ratio of NMOG emission standards for
Cm. In this year's Inventory, these emission factors for newer gasoline and diesel vehicles are based on annual
certification data compiled by EPA.
In prior Inventories, Class II and Class III rail fuel consumption data was provided by the American Short Line and
Regional Railroad Association (ASLRRA). Since ASLRRA no longer tracks and reports fuel consumption data of these
rail lines, it is now estimated for years 2014 onwards using carload data reported by Railinc (2014 through 2018).
The collective result of these changes was a net increase in CH4 emissions and a decrease in N2O emissions from
mobile combustion relative to the previous Inventory. Methane emissions increased by 0.5 percent. Nitrous oxide
emissions decreased by 1.1 percent.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018 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 N2O for the time periods 1990 to 2006 and 2007 to 2018. Details on
the emission trends and methodological inconsistencies through time are described in the Methodology section,
above.
Planned Improvements
While the data used for this report represent the most accurate information available, several areas have been
identified that could potentially be improved in the near term given available resources.
•	Determine new methane and nitrous oxide emission factors for non-road equipment using annual
certification data compiled by EPA.
•	In previous Inventories, EPA identified the need to evaluate and potentially update EPA's method for
estimating motor gasoline consumption for non-road mobile sources, in order to improve accuracy and
create a more consistent time series. As discussed in the Methodology section above and in Annex 3.2,
CH4 and N2O estimates for gasoline-powered non-road sources in this Inventory are based on a variety of
inputs, including FHWA Highway Statistics Table MF-24. In 2016, FHWA changed its methods for
estimating the share of gasoline used in on-road and non-road applications.63 These method changes
63 The previous and new FHWA methodologies for estimating non-road gasoline are described in Off-Highway and Public-Use
Gasoline Consumption Estimation Models Used in the Federal Highway Administration, Publication Number FHWA-PL-17-012.
.
3-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
created a time-series inconsistency in the current Inventory between 2015 and previous years in CFU and
N2O estimates for agricultural, construction, commercial, and industrial non-road mobile sources. EPA has
implemented an approach to address this inconsistency. EPA also tested an alternative approach that
uses MOVES on-road fuel consumption output to define the percentage of the FHWA consumption totals
(from MF-21) that are attributable to on-highway transportation sources, and applying this percentage to
the EIA total, thereby defining gasoline consumption from on-highway transportation sources (such that
the remainder would be defined as consumption by the industrial and commercial sectors). Results from
this testing revealed differences between fuel consumption calculated by MOVES and fuel consumption
data from FHWA. Given this inconsistency, no changes have been made to the methodology for
estimating motor gasoline consumption for non-road mobile sources.
•	Update emissions 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. 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 continues to
be investigated. Additionally, the feasibility of including data from a broader range of domestic and
international sources for domestic bunker fuels, including data from studies such as the Third IMO GHG
Study 2014, continues to be explored.
3.2 Carbon Emitted from Non-Energy Uses of
Fossil Fuels (CRF Source Category 1A5)
In addition to being combusted for energy, fossil fuels are also consumed for non-energy uses in the United States.
The fuels used for these purposes are diverse, including natural gas, liquefied petroleum gases (LPG), 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 a portion of non-energy uses of fossil fuels 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).
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
and across all uses, about 62 percent of the total C consumed for non-energy purposes was stored in products, 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 CO2 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. In addition, there is some overlap between fossil fuels consumed for non-energy uses and
the fossil-derived CO2 emissions accounted for in the IPPU chapter, especially for fuels used as reducing agents. To
Energy 3-47

-------
avoid double counting, the "raw" non-energy fuel consumption data reported by EIA are modified to account for
these overlaps. There are also net exports of petrochemicals 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 2018 from the non-energy uses of fossil fuels were 134.6 MMT CO2
Eq., which constituted approximately 2 percent of overall fossil fuel emissions. In 2018, the consumption of fuels
for non-energy uses (after the adjustments described above) was 5,264.7 TBtu (see Table 3-21). A portion of the C
in the 5,264.7 TBtu of fuels was stored (221.7 MMT CO2 Eq.), while the remaining portion was emitted (134.6 MMT
CO2 Eq.). Non-energy use emissions increased 9.3 percent from 2017 to 2018 mainly due to increases in coking coal
and petrochemical feedstock use, both of which are driven by changes in economic activity and changes in the
industrial sector, 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
2014
2015
2016
2017
2018
Potential Emissions
312.1
377.5
325.1
340.5
329.9
341.2
356.3
C Stored
192.5
237.8
205.1
213.5
216.2
218.0
221.7
Emissions as a % of Potential
38%
37%
37%
37%
34%
36%
38%
Emissions
119.5
139.7
120.0
127.0
113.7
123.1
134.6
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 (2019) (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.64,65 Consumption of natural
gas, LPG, 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 intermediary chemicals.
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, LPG,
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
64	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.
65	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 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 and the non-energy use estimates are roughly 20 percent of the emissions captured under IPPU. 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.
3-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
•	For industrial coking coal and distillate fuel oil, storage factors were taken from IPCC (2006), which in turn
draws 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
2014
2015
2016
2017
2018
Industry
4,215.8
5,110.7
4,602.9
4,764.6
4,634.2
4,799.5
5,049.6
Industrial Coking Coal
NO
80.4
48.8
121.8
88.6
111.8
124.7
Industrial Other Coal
8.2
11.9
10.3
10.3
10.3
10.3
10.3
Natural Gas to Chemical Plants
281.6
260.9
323.5
321.9
308.9
307.6
304.7
Asphalt & Road Oil
1,170.2
1,323.2
792.6
831.7
853.4
849.2
792.8
LPG
1,120.5
1,610.0
2,109.8
2,157.5
2,119.0
2,187.7
2,485.5
Lubricants
186.3
160.2
130.7
142.1
135.1
124.9
121.2
Pentanes Plus
117.6
95.5
43.5
78.4
53.1
81.5
104.8
Naphtha (<401 °F)
326.3
679.5
435.2
417.8
396.9
411.1
418.3
Other Oil (>401 °F)
662.1
499.5
236.2
216.8
204.0
241.8
217.7
Still Gas
36.7
67.7
164.5
162.2
166.1
163.8
166.9
Petroleum Coke
27.2
105.2
NO
NO
NO
NO
NO
Special Naphtha
100.9
60.9
104.5
97.0
88.7
94.9
86.5
Distillate Fuel Oil
7.0
11.7
5.8
5.8
5.8
5.8
5.8
Waxes
33.3
31.4
14.8
12.4
12.8
10.2
12.4
Miscellaneous Products
137.8
112.8
182.7
188.9
191.3
198.8
198.0
Transportation
176.0
151.3
149.4
162.8
154.4
142.0
137.8
Lubricants
176.0
151.3
149.4
162.8
154.4
142.0
137.8
U.S. Territories
85.6
123.2
77.3
77.3
77.3
77.3
77.3
Lubricants
0.7
4.6
1.0
1.0
1.0
1.0
1.0
Other Petroleum (Misc. Prod.)
84.9
118.6
76.2
76.2
76.2
76.2
76.2
Total
4,477.4
5,385.2
4,829.6
5,004.7
4,865.8
5,018.8
5,264.7
NO (Not Occurring).
Table 3-22: 2018 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







(MMT CO?
Sector/Fuel Type
(TBtu)
(MMT C/QBtu)
(MMT C)

(MMT C)
(MMT C)
Eq.)
Industry
5,049.6
NA
92.8
NA
60.1
32.8
120.2
Industrial Coking Coal
124.7
31.00
3.9
0.10
0.4
3.5
12.8
Industrial Other Coal
10.3
26.08
0.3
0.65
0.2
0.1
0.3
Natural Gas to







Chemical Plants
304.7
14.47
4.4
0.65
2.9
1.5
5.6
Asphalt & Road Oil
792.8
20.55
16.3
1.00
16.2
0.1
0.3
LPG
2,485.5
17.06
42.4
0.65
27.7
14.7
53.9
Lubricants
121.2
20.20
2.4
0.09
0.2
2.2
8.2
Energy 3-49

-------
Pentanes Plus
104.8
19.10
2.0
0.65
1.3
0.7
2.5
Naphtha (<401° F)
418.3
18.55
7.8
0.65
5.1
2.7
9.9
Other Oil (>401° F)
217.7
20.17
4.4
0.65
2.9
1.5
5.6
Still Gas
166.9
17.51
2.9
0.65
1.9
1.0
3.7
Petroleum Coke
+
27.85
+
0.30
+
+
+
Special Naphtha
86.5
19.74
1.7
0.65
1.1
0.6
2.2
Distillate Fuel Oil
5.8
20.17
0.1
0.50
0.1
0.1
0.2
Waxes
12.4
19.80
0.2
0.58
0.1
0.1
0.4
Miscellaneous







Products
198.0
20.31
4.0
0.00
+
4.0
14.7
Transportation
137.8
NA
2.8
NA
0.3
2.5
9.3
Lubricants
137.8
20.20
2.8
0.09
0.3
2.5
9.3
U.S. Territories
77.3
NA
1.5
NA
0.2
1.4
5.1
Lubricants
1.0
20.20
+
0.09
+
+
0.1
Other Petroleum







(Misc. Prod.)
76.2
20.00
1.5
0.10
0.2
1.4
5.0
Total
5,264.7

97.2

60.5
36.7
134.6
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 TBtu, MMT C, MMT C02 Eq.
NA (Not Applicable)
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, 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 2019a), Toxics Release
Inventory, 1998 (EPA 2000b), Biennial Reporting System (EPA 2000a, 2009), Resource Conservation and Recovery
Act Information System (EPA 2013b, 2015, 2016b, 2018b), 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); 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); 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, 2019b); 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, 2019b); and the Guide to the Business of Chemistry (ACC
2019a). Specific data sources are listed in full detail in Annex 2.3.
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
3-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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, LPG, pentanes plus, naphthas, other oils, still gas, special naphthas, and other industrial coal), (2)
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 2018 was estimated to be between
96.8 and 188.8 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 28 percent below to 40
percent above the 2018 emission estimate of 134.6 MMT CO2 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)
2018 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
83.7
53.9
142.3
-36%
+70%
Asphalt
C02
0.3
0.1
0.6
-58%
+118%
Lubricants
C02
17.5
14.4
20.3
-18%
+16%
Waxes
C02
0.4
0.3
0.7
-24%
+80%
Other
C02
32.7
18.8
35.6
-43%
+9%
Total
C02
134.6
96.8
188.8
-28%
40%
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)
2018 Storage Factor Uncertainty Range Relative to Emission Estimate3
(%)	(%)	(%, Relative)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Feedstocks
C02
65.3%
51.9%
71.9%
-21%
+10%
Asphalt
C02
99.6%
99.1%
99.8%
-0.5%
+0.2%
Lubricants
C02
9.2%
3.9%
17.5%
-57%
+90%
Waxes
C02
57.8%
47.6%
67.6%
-18%
+17%
Other
C02
6.3%
6.0%
42.8%
-4%
+582%
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-
Energy 3-51

-------
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
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 2018 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 2017 totals as well as their
trends across the time series.
Petrochemical input data reported by EIA will continue to be investigated in an attempt to address an input/output
discrepancy in the NEU model. Prior to 2001, the C balance inputs exceeded outputs, then starting in 2001 through
2009, outputs exceeded inputs. Inputs exceeded outputs in 2010, 2011, and 2013 through 2018, but outputs
exceeded inputs in 2012. A portion of this discrepancy has been reduced and two strategies have been developed
to address the remaining portion (see the Planned Improvements section, below).
Recalculations Discussion
Previously proxied data for five chemicals and fibers (polyester fiber, polyolefin fiber, nylon fiber, acetic acid, and
maleic anhydride) were updated using the Guide to the Business of Chemistry, 2019 for 1990 through 2017 values.
Overall, these changes resulted in an average annual decrease of less than 0.01 MMT CO2 Eq. (less than 0.01
percent) in carbon emissions from non-energy uses of fossil fuels for the period 1990 through 2017, relative to the
previous Inventory.
Planned Improvements
There are several future improvements planned:
•	Analyzing the fuel and feedstock data from EPA's GHGRP Subpart X (Petrochemical Production) to better
disaggregate CO2 emissions in NEU model and CO2 process emissions from petrochemical production.
•	More accurate accounting of C in petrochemical feedstocks. EPA has worked with EIA to determine the
3-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. 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
(miscellaneous products). A better understanding of these trends will be pursued to identify any
mischaracterized or misreported fuel consumption for non-energy uses. For example, "miscellaneous
products" category 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 from petroleum
and natural gas processing, and potentially also C black feedstock could be reported in this category.
Recovered sulfur would not be reported in the NEU calculation or elsewhere in the Inventory.
•	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;
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.
Energy 3-53

-------
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.66 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
Emitted from Non-Energy Uses of Fossil Fuels source category (CRF Source Category 1A5).67
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. 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 Section 3.2, 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.68 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. 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. These artificial adjustments would also result in the C emissions for lubricants,
waxes, and asphalt and road oil being reported under IPPU, while the C storage for lubricants, waxes, and
asphalt and road oil would be reported under Energy. 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, 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 IPPU chapter, as
they were consumed during non-energy related industrial activity. Emissions from 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.
66	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).
67	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.
68	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.
3-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018a; 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 CO2. According to the 2006IPCC Guidelines,
when the CO2 emitted is of fossil origin, it is counted as a net anthropogenic emission of CO2 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-
plastics, synthetic rubber, synthetic fibers, and carbon black in scrap tires—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 2018 from this source so the data were proxied to the 2011 estimate. Carbon
dioxide emissions from incineration of waste increased 40 percent since 1990, to an estimated 11.1 MMT CO2
(11,113 kt) in 2018, 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 N2O emissions (De Soete 1993; IPCC 2006). Methane emissions from
the incineration of waste were estimated to be less than 0.05 MMT CO2 Eq. (less than 0.5 kt CH4) in 2018 and have
decreased by 32 percent since 1990. Nitrous oxide emissions from the incineration of waste were estimated to be
0.3 MMT CO2 Eq. (1 kt N2O) in 2018 and have decreased by 32 percent since 1990.
Table 3-25: CO2, ChU, and N2O Emissions from the Incineration of Waste (MMT CO2 Eq.)
Gas/Waste Product
1990
2005
2014
2015
2016
2017
2018
C02
8.0
12.5
10.4
10.8
10.9
11.1
11.1
Plastics
5.6
6.9
5.9
6.2
6.2
6.4
6.4
Synthetic Rubber in Tires
0.3
1.6
1.2
1.1
1.2
1.2
1.2
Carbon Black in Tires
0.4
2.0
1.4
1.4
1.4
1.4
1.4
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
1.3
1.4
1.4
1.4
ch4
+
+
+
+
+
+
+
n2o
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Total
8.4
12.9
10.7
11.1
11.2
11.4
11.4
+ Does not exceed 0.05 MMT C02 Eq.
Energy 3-55

-------
Table 3-26: CO2, ChU, and N2O Emissions from the Incineration of Waste (kt)
Gas/Waste Product
1990
2005
2014
2015
2016
2017
2018
CO?
7,951
12,469
10,435
10,756
10,919
11,111
11,113
Plastics
5,588
6,919
5,928
6,184
6,227
6,388
6,388
Synthetic Rubber in Tires
308
1,599
1,154
1,149
1,160
1,171
1,171
Carbon Black in Tires
385
1,958
1,406
1,401
1,415
1,430
1,430
Synthetic Rubber in MSW
854
766
692
703
717
731
731
Synthetic Fibers
816
1,227
1,255
1,319
1,399
1,392
1,394
ch4
+
+
+
+
+
+
+
n2o
2
1
1
1
1
1
1
+ Does not exceed 0.5 kt.
fviet had ©logy
Emissions of CO2 from the incineration of waste include CO2 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 CO2 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 CO2 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) and detailed unpublished backup data for some years not shown in the reports
(Schneider 2007). For 2012 through 2018 data on total waste incinerated were assumed to equal to the 2011 value
from Shin (2014) for 2012 through 2018. For synthetic rubber and carbon black in scrap tires, information was
obtained biannually from U.S. Scrap Tire Management Summary for 2005 through 2018 data (RMA 2018). 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
2001 is the average of 1998 and 2002 values; and C content for 2002 to date is based on the 2002 value. Carbon
content for synthetic fibers was calculated from a weighted average of production statistics from 1990 to date.
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 CO2
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 CFUand N2O. These emissions were calculated as
a function of the total estimated mass of waste incinerated and emission factors. As noted above, CFU and N2O
emissions are a function of total waste incinerated in each year; for 1990 through 2008, these data were derived
3-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018, 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 N2O 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)
Year
Waste Discarded
Waste Incinerated
Incinerated (% of
Discards)
1990
235,733,657
30,632,057
13.0%
2005
259,559,787
25,973,520
10.0%
2014
273,116,704a
20,756,870
7.6%
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%
a Assumed equal to 2011 value.
Source: van Haaren et al. (2010).
Uncertainty and Time-Seri insistency
An Approach 2 Monte Carlo analysis was performed to determine the level of uncertainty surrounding the
estimates of CO2 emissions and N2O emissions from the incineration of waste (given the very low emissions for
Cm, 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
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 N2O 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
CO2 emissions in 2018 were estimated to be between 8.2 and 14.4 MMT CO2 Eq. at a 95 percent confidence level.
This indicates a range of 26 percent below to 29 percent above the 2018 emission estimate of 11.1 MMT CO2 Eq.
Also at a 95 percent confidence level, waste incineration N2O emissions in 2018 were estimated to be between 0.2
Energy 3-57

-------
and 1.3 MMT CO2 Eq. This indicates a range of 51 percent below to 328 percent above the 2018 emission estimate
of 0.3 MMT CO2 Eq. Differences observed in comparison to last year were due to a reevaluation and refinement of
assumptions on scrap tire weights of light and heavy-duty tires.
Table 3-28: Approach 2 Quantitative Uncertainty Estimates for CO2 and N2O from the
Incineration of Waste (MMT CO2 Eq. and Percent)


2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
Source
Gas
(MMT CO? Eq.)
(MMT CO?
Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Incineration of Waste
C02
11.1
8.2
14.4
-26%
29%
Incineration of Waste
N20
0.3
0.2
1.3
-51%
328%
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
EPA revised the percent of tires disposed for light duty tires and commercial tires in 2009 and 2013 to reflect
updated data. For 2009, EPA used data from the Rubber Manufacturers Association's (RMA) U.S. Scrap Tire
Management Summary 2005-2009 (RMA 2013), and RMA's 2013 U.S. Scrap Tire Management Summary (RMA
2014)	for 2013. These updates impacted CO2 emissions from synthetic rubber in tires and synthetic rubber in
MSW.
EPA also updated the total generation and recovery data for plastics, synthetic rubber, and synthetic fibers in MSW
for years 2016 and 2017. In the previous Inventory report, emissions were being proxied from 2015 values. EPA
used data from EPA's Advancing Sustainable Materials Management: Facts and Figures 2016 and 2017, Assessing
Trends in Material Generation, Recycling and Disposal in the United States (EPA 2019). The updates to MSW
discarded impacted CO2 emissions for those materials in 2016 and 2017.
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 previously came 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 GHGRP data were more complete (i.e., included more facilities), but did not contain data for all inventory
years (1990 through 2016). The EIA data can be used to supplement years not available in the GHGRP dataset. In
addition, the GHGRP data do not include specific waste components outside of an assumed biogenic and fossil
component, which is necessary for CO2 emission calculations. Data from EPA's GHGRP on fossil CO2 emissions can
be used to benchmark results for other waste components in the Inventory.
3-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Additional improvements will focus on investigating new methods and sources for CO2 emission estimates and
investigating new data sources for MSW incinerated values (i.e., tonnage) for estimating CO2 and non-CC>2 (CH4,
N2O) emissions.
Proposed improvements to the current CO2 emissions estimation methodology include opportunities for either
incorporating total CO2 emissions from existing waste incineration datasets (i.e., EIA and GHGRP data that provide
CO2 emission estimates) or updating emission factors (i.e., MSW carbon content) and continuing to use the Facts
and Figures disposal data for fossil-based products. Further research is required to compare the emission factors
(i.e., MSW carbon content, heating values) used across waste incineration CO2 emissions approaches, including the
current Inventory, EIA, and EPA's GHGRP. In addition, the currently used BioCycle percent combusted assumption
could be updated with Facts and Figures product tonnage combusted data.
Non-CC>2 improvements will focus on research of potential data sources for updating emission factors. EPA is also
researching potential data sources for incinerated MSW tonnage that can be used for future inventory years
instead of applying an incineration rate to generated MSW tonnage. EPA is analyzing the Facts and Figures non-tire
MSW combusted tonnage and previously compiled EIA and GHGRP tonnage data to compare organic and non-
organic components of these MSW tonnage data where available.
Additional improvements will be conducted to improve the transparency in the current reporting of waste
incineration. Currently, hazardous industrial waste incineration is included within the overall calculations for the
Carbon Emitted from Non-Energy Uses of Fossil Fuels source category. Waste incineration activities that do not
include energy recovery will be examined. Synthetic fibers within scrap tires are not included in this analysis and
will be explored for future Inventories. The C content of fibers within scrap tires will be used to calculate the
associated incineration emissions. Updated fiber content data from the Fiber Economics Bureau will also be
explored.
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 2018, 236 underground coal mines
and 430 surface mines were operating in the United States (EIA 2019). In recent years the total number of active
coal mines in the United States has declined. In 2018, the United States was the third-largest coal producer in the
world (686 MMT), after China (3,550 MMT) and India (771 MMT) (IEA 2019).
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
334,398
789
691,448
1,398
1,025,846
2014
345
321,783
613
583,974
958
905,757
2015
305
278,342
529
534,092
834
812,435
2016
251
228,707
439
431,285
690
659,991
2017
237
247,779
434
454,303
671
702,082
2018
236
249,802
430
435,521
666
685,324
Underground mines liberate CH4from 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
Energy 3-59

-------
the surface or boreholes drilled inside the mine that remove large, often highly concentrated volumes of Cm
before, during, or after mining. Some mines recover and use Cm generated from ventilation and degasification
systems, thereby reducing emissions to the atmosphere.
Surface coal mines liberate Cm 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
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, Cm is released during post-mining activities, as the coal is processed, transported, and stored for use.
Total Cm emissions in 2018 were estimated to be 2,109.3 kt (52.7 MMT CO2 Eq.), a decline of approximately 45
percent since 1990 (see Table 3-30 and Table 3-31). In 2018, underground mines accounted for approximately 74
percent of total emissions, surface mines accounted for 13 percent, and post-mining activities accounted for 13
percent. In 2018, total CH4 emissions from coal mining decreased by approximately 4 percent relative to the
previous year. This decrease was due to a modest decrease in coal production and an increase in CH4 recovered
and used. The amount of CH4 recovered and used in 2018 increased by approximately eleven percent compared to
2017 levels. This increase is primarily attributed to an increase in reported CH4 recovery and use at three mines.
Table 3-30: ChU Emissions from Coal Mining (MMT CO2 Eq.)
Activity
1990
2005
2014
2015
2016
2017
2018
Underground (UG) Mining
74.2
42.0
46.1
44.9
40.7
40.7
38.9
Liberated
80.8
59.7
63.0
61.2
57.0
57.6
57.7
Recovered & Used
(6.6)
(17.7)
(17.0)
(16.4)
(16.4)
(17.0)
(18.8)
Surface Mining
10.8
11.9
9.6
8.7
6.8
7.2
7.0
Post-Mining (UG)
9.2
7.6
6.7
5.8
4.8
5.3
5.3
Post-Mining (Surface)
2.3
2.6
2.1
1.9
1.5
1.6
1.5
Total
96.5
64.1
64.6
61.2
53.8
54.8
52.7
ble 3-31: ChU Emissions from Coal Mining (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
Underground (UG) Mining
2,968
1,682
1,844
1,796
1,629
1,626
1,556
Liberated
3,231
2,385
2,523
2,450
2,283
2,306
2,308
Recovered & Used
(263)
(704)
(679)
(654)
(654)
(679)
(752)
Surface Mining
430
475
386
347
273
290
280
Post-Mining (UG)
368
306
270
231
193
213
212
Post-Mining (Surface)
93
103
84
75
59
63
61
Total
3,860
2,565
2,583
2,449
2,154
2,191
2,109
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 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
3-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 Cm, thereby reducing emissions to the atmosphere. Total Cm emitted from underground mines
equals the Cm liberated from ventilation systems, plus the Cm liberated from degasification systems, minus the
Cm 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)69 (Subpart FF, "Underground Coal Mines"), data provided by the U.S. Mine Safety and Health
Administration (MSHA) (MSHA 2019), 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 CO2 Eq.)—have been required to
report to EPA's GHGRP (EPA 20 19).70 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.71
Since 2013, ventilation CH4 emission estimates have been calculated based on both GHGRP data submitted by
underground mines, and on quarterly measurement data obtained directly from MSHA for the remaining mines.
The quarterly measurements are used to determine the average daily CH4 emission rate for the reporting year
quarter. 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.
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. Eighteen
mines used degasification systems in 2018, and the CH4 removed through these systems was reported to EPA's
GHGRP under Subpart FF (EPA 2019). Based on the weekly measurements reported to EPA's GHGRP, degasification
data summaries for each mine were added to estimate the CH4 liberated from degasification systems. Eleven of
the 18 mines with degasification systems had operational CH4 recovery and use projects (see step 1.3 below), and
EPA's GHGRP reports show the remaining seven mines vented CH4from degasification systems to the
atmosphere.72
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
69	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).
70	Underground coal mines report to EPA under Subpart FF of the GHGRP (40 CFR Part 98). In 2018, 76 underground coal mines
reported to the program.
71	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.
72	Several of the mines venting CH4from 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 3-61

-------
exclusively to estimate Cm liberated from degasification systems at 14 of the 18 mines that used degasification
systems in 2018.
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.73 EPA's GHGRP does not require
gas production from virgin coal seams (coalbed methane) to be reported by coal mines under Subpart FF.74 Most
pre-mining wells drilled from the surface are considered coalbed methane wells prior to mine-through and
associated CFU 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 four mines with degasification systems that include pre-mining
wells that were mined through in 2018. For these mines, GHGRP data were supplemented with historical data from
state gas well production databases (GSA 2019; DMME 2019; WVGES 2019), 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 2019).
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 2018. Eleven of these projects involved degasification
systems, one did not use any degasification system, 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 was 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
were estimated using the following methods:
•	EPA's GHGRP data was exclusively used to estimate the CH4 recovered and used from seven of the 11
mines that deployed degasification systems in 2018. Based on weekly measurements, the GHGRP
degasification destruction data summaries for each mine were 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 four mines that
deployed degasification systems in 2018 (DMME 2019, GSA 2019). These four mines intersected pre-
mining wells in 2018. Supplemental information was 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 came 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 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
73 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.
3-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Annual Coal Report (EIA 2019) was multiplied by basin-specific Cm contents (EPA 1996, 2005) and a 150 percent
emission factor (to account for Cmfrom over- and under-burden) to estimate Cm emissions (King 1994, Saghafi
2013). For post-mining activities, basin-specific coal production was multiplied by basin-specific gas contents and a
mid-range 32.5 percent emission factor for Cm 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 daily emission
rate for the quarter. Additionally, the measurement equipment used can be expected to have resulted in an
average of 10 percent overestimation of annual Cm emissions (Mutmansky & Wang 2000). Equipment
measurement uncertainty is applied to both GHGRP and MSHA 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.
Beginning in 2015, a small level of uncertainty was introduced by using estimated rather than measured values of
recovered CH4 from two of the mines with degasification systems. An increased level of uncertainty was applied to
these two sub-sources, but the change had little impact on the overall uncertainty.
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 2018 were estimated to be between 43.9 and 59.2 MMT CO2 Eq. at a 95 percent confidence level. This
indicates a range of 16.7 percent below to 12.3 percent above the 2018 emission estimate of 52.7 MMT CO2 Eq.
Table 3-32: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Coal
Mining (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Coal Mining
ch4
52.7
43.9 59.2
-16.7% +12.3%
a Range of emission estimates predicted by Monte Carlo stochastic simulation for a 95 percent confidence interval.
Energy 3-63

-------
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 greenhouse gas report. Additional QA/QC and
verification procedures occur for each GHGRP subpart.
Recalculations Discussion
No recalculations were performed for the 1990 through 2017 portion of the time series.
Planned Improvements
EPA intends to add methods for estimating fugitive CO2 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:
•	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.
3-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Annual gross abandoned mine Cm emissions ranged from 7.2 to 10.8 MMT CO2 Eq. from 1990 through 2018,
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 CO2 Eq.) due to the large number
of gassy mine75 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 2018 there was one gassy mine closure. Gross abandoned
mine emissions decreased slightly from 9.2 MMT CO2 Eq. in 2017 to 8.9 MMT CO2 Eq. in 2018 (see Table 3-33 and
Table 3-34). Gross emissions are reduced by Cm recovered and used at 45 mines, resulting in net emissions in
2018 of 6.2 MMT CO2 Eq.
Table 3-33: ChU Emissions from Abandoned Coal Mines (MMT CO2 Eq.)
Activity
1990
2005
2014
2015
2016
2017
2018
Abandoned Underground Mines
7.2
8.4
8.7
9.0
9.5
9.2
8.9
Recovered & Used
0.0
(1-8)
(2.4)
(2.6)
(2.8)
(2.7)
(2.7)
Total
7.2
6.6
6.3
6.4
6.7
6.4
6.2
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
2014
2015
2016
2017
2018
Abandoned Underground Mines
288
334 P;
350
359
380
367
355
Recovered & Used
0.0
(70) 0
(97)
(102)
(112)
(109)
(107)
Total
288
264
253
256
268
257
247
+ Does not exceed 0.5 kt.
Methodology
Estimating Cm 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 which 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
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 Cm 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
75 A mine is considered a "gassy" mine if it emits more than 100 thousand cubic feet of CH4 per day (100 mcfd).
Energy 3-65

-------
the pressures of interest (atmospheric to 30 psia). The Cm 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 + 6At)("1/6)
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
Di	=	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 Cm 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)

-------
Table 3-35: Number of Gassy Abandoned Mines Present in U.S. Basins in 2018, Grouped by
Class According to Post-Abandonment State
Basin
Sealed
Vented
Flooded
Total
Known
Unknown Total Mines
Central Appl.
41
26
52
119
148
267
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
150
55
100
305
228
533
Inputs to the decline equation require the average Cm 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 Cm 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 United States,
representing 78 percent of the emissions. State-specific, initial emission rates were used based on average coal
mine CFU emissions rates during the 1970s (EPA 2004).
Abandoned mine emission estimates are based on all closed mines known to have active mine Cm 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 2019). 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. CFU degasification amounts were added to the quantity of Cm vented to determine the total CFU
liberation rate for all mines that closed between 1992 and 2018. 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 2018, emission totals were downwardly adjusted to reflect Cm emissions avoided from those
abandoned mines with Cm recovery and use or destruction systems. The Inventory totals were not adjusted for
abandoned mine Cm emission reductions from 1990 through 1992, because no data was reported for abandoned
coal mine Cm recovery and use or destruction projects during that time.
Uncertainty and Time-Serii insistency
A quantitative uncertainty analysis was conducted to estimate the uncertainty surrounding the estimates of
emissions from abandoned underground coal mines. The uncertainty analysis described below 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) Cm flow capacity as expressed by permeability; and 3)
Energy 3-67

-------
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 Cm emissions in 2018 were estimated to be between 5.0 and 7.1 MMT CO2 Eq. at a 95 percent
confidence level. This indicates a range of 20 percent below to 15 percent above the 2018 emission estimate of 6.2
MMT CO2 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 after 1972. Emissions from mines closed prior to 1972 have the largest degree of
uncertainty because no mine-specific Cm 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
2018 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Abandoned Underground
Coal Mines
ch4
6.2
5.0 7.1
-20% +15%
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.
Recalculations Discussion
No recalculations were performed for the 1990 through 2017 portion of the time series.
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
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 crude oil production and refining operations. Note, CO2
emissions exclude all combustion emissions (e.g., engine combustion) except for flaring CO2 emissions. All
3-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
combustion CO2 emissions (except for flaring) are accounted for in the fossil fuel combustion chapter (see Section
3.1). Emissions of N2O from petroleum systems are primarily associated with flaring. Total greenhouse gas
emissions (CH4, CO2, and N2O) from petroleum systems in 2018 were 73.1 MMT CO2 Eq., an increase of 31 percent
from 1990, primarily due to increases in CO2 emissions. Since 2008, total emissions increased by 30 percent; and
since 2017, total emissions increased by 16 percent. Total CO2 emissions from petroleum systems in 2018 were
36.8 MMT CO2 (36,814 kt CO2), an increase of a factor of 2.8 from 1990. Since 2008, total CO2 emissions increased
by a factor of 1.7, and since 2017 CO2 emissions increased by 50 percent. Total Cm emissions from petroleum
systems in 2018 were 36.2 MMT CO2 Eq. (1,449 kt CH4), a decrease of 21 percent from 1990. Since 2008, total CH4
emissions decreased by 15 percent; and since 2017, Cm emissions decreased by 6 percent. Total N2O emissions
from petroleum systems in 2018 were 0.07 MMT CO2 Eq. (0.24 kt N2O), an increase of a factor of 3.2 from 1990.
Since 2008, total N2O emissions increased by a factor of 2.7; and since 2017, N2O emissions increased by a factor of
1.6. Since 1990, U.S. oil production has increased by 49 percent; from 2008 to 2018, production increased by a
factor of 1.2; and from 2017 to 2018, production increased by 18 percent.
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 2017) to
ensure that the trend is accurate. Recalculations in petroleum systems in this year's Inventory include:
•	Revised offshore oil production methodology
•	Revised emissions for delayed cokers in refineries, due to a methodological change in GHGRP reporting
for Subpart Y
•	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 accounts for approximately 1
percent of total CH4 emissions (including leaks, vents, and flaring) from petroleum systems in 2018. The
predominant sources of emissions from exploration are hydraulically fractured oil well completions and well
drilling. Other sources include well testing and well completions without hydraulic fracturing. Since 1990,
exploration CH4 emissions have decreased 88 percent, and while the number of hydraulically fractured wells
completed increased by a factor of 2.6, there were decreases in the fraction of such completions without reduced
emissions completions (RECs) or flaring (from 90 percent in 1990 to 1 percent in 2018). Emissions of CH4 from
exploration were highest in 2012, over 20 times higher than in 2018; and lowest in 2017. Emissions of Cm from
exploration increased 11 percent from 2017 to 2018, due to an increase in hydraulically fractured oil well
completions with flaring. Exploration accounts for 8 percent of total CO2 emissions (including leaks, vents, and
flaring) from petroleum systems in 2018. Emissions of CO2 from exploration in 2018 increased by a factor of 7.4
from 1990 levels, and 76 percent from 2017, due to the abovementioned increase in hydraulically fractured oil well
completions with flaring. Emissions of CO2 from exploration were highest in 2014, around 11 percent higher than
in 2018. Exploration accounts for 2 percent of total N2O emissions from petroleum systems in 2018. Emissions of
N2O from exploration in 2018 increased by a factor of 8.4 from 1990, and by a factor of 1.4 from 2017, due to the
abovementioned increase in hydraulically fractured oil well completions with flaring.
Production. Production accounts for approximately 96 percent of total CH4 emissions (including leaks, vents, and
flaring) from petroleum systems in 2018. The predominant sources of emissions from production field operations
are pneumatic controllers, offshore oil platforms, gas engines, chemical injection pumps, leaks from oil wellheads,
and oil tanks. These six sources together account for 91 percent of the CH4 emissions from production. Since 1990,
Cm emissions from production have decreased by 17 percent due to decreases in emissions from offshore
platforms, tanks, and pneumatic controllers. Overall, production segment methane emissions decreased by 7
percent from 2017 levels due primarily to a decrease in the number of intermittent bleed controllers as use of low
bleed controllers grew in 2018. Production emissions account for 82 percent of the total CO2 emissions (including
leaks, vents, and flaring) from petroleum systems in 2018. The principal sources of CO2 emissions are associated
gas flaring, oil tanks with flares, and miscellaneous production flaring. These three sources together account for 98
percent of the CO2 emissions from production. Since 1990, CO2 emissions from production have increased by a
factor of 4.0, due to increases in flaring emissions from associated gas flaring, tanks, and miscellaneous production
Energy 3-69

-------
flaring. Overall, production segment CO2 emissions increased by 58 percent from 2017 levels primarily due to an
increase in associated gas flaring in the Permian and Williston basins. Production emissions account for 83 percent
of the total N2O emissions from petroleum systems in 2018. The principal sources of N2O emissions are oil tanks
with flares, miscellaneous production flaring, and associated gas flaring. Since 1990, N2O emissions from
production have increased by a factor of 6.9; and since 2017, N2O emissions from production have increased by a
factor of 2.8, due primarily to increases in N2O from oil tanks with flares and miscellaneous production flaring.
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 Cm emissions from petroleum
systems. Emissions from tanks, marine loading, and truck loading operations account for 75 percent of CH4
emissions from crude oil transportation. Since 1990, CH4 emissions from transportation have increased by 29
percent. In 2018, CH4 emissions from transportation increased by 10 percent from 2017 levels. Crude oil
transportation activities account for less than 0.01 percent of total CO2 emissions from petroleum systems.
Emissions from tanks, marine loading, and truck loading operations account for 75 percent of CO2 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) Cm 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 Cm in all
refined products. Within refineries, flaring accounts for 38 percent of the CH4 emissions, while delayed cokers,
uncontrolled blowdowns, and process vents account for 18,17, and 9 percent, respectively. Fugitive Cm emissions
from refining of crude oil have increased by 14 percent since 1990, and decreased 7 percent from 2017; 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 10 percent of total fugitive (including
leaks, vents, and flaring) CO2 emissions from petroleum systems. Of the total fugitive CO2 emissions, almost all
(about 98 percent) of it comes from flaring.76 Refinery fugitive CO2 emissions increased by 14 percent from 1990 to
2018 and increased by less than 1 percent from the 2017 levels. Flaring occurring at crude oil refining processes
and systems accounts for 15 percent of total fugitive N2O emissions from petroleum systems. Refinery fugitive N2O
emissions increased by 16 percent from 1990 to 2018 and decreased by 2 percent from 2017 levels.
Table 3-37: ChU Emissions from Petroleum Systems (MMT CO2 Eq.)
Activity
1990
2005
; 2014
2015
2016
2017
2018
Exploration3
3.0
4.5
5.1
2.1
0.5
0.3
0.4
Production (Total)
42.4
33.4
37.5
37.4
37.5
37.3
34.9
Pneumatic Controllers
19.3
17.6
19.6
19.7
20.6
21.3
18.4
Offshore Production
9.3
6.5
5.7
5.5
5.1
5.1
5.1
Equipment Leaksb
2.2
2.2
2.7
2.7
2.6
2.6
2.5
Gas Engines
2.1
1.7
2.3
2.3
2.2
2.2
2.3
Chemical Injection Pumps
1.2
1.7
2.2
2.2
2.1
2.1
2.0
Tanks
5.4
1-5
1.6
1.7
2.5
1.5
1.4
Other Sources
2.6
2.1
3.3
3.3
2.3
2.6
3.2
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
0.8
Total
46.1
38.8
i 43.5
40.5
39.0
38.7
36.2
Note: Totals may not sum due to independent rounding.
a Exploration includes well drilling, testing, and completions.
b Includes leak emissions from wellheads, separators, heaters/treaters, and headers.
76 Petroleum Systems includes fugitive emissions f(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.
3-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-38: ChU Emissions from Petroleum Systems (kt ChU)
Activity
1990
2005
2014
2015
2016
2017
2018
Exploration3
121
181
202
84
19
13
15
Production (Total)
1,689
1,336
1,498
1,496
1,499
1,494
1,395
Pneumatic Controllers
772
704
783
789
823
851
735
Offshore Production
372
261
230
220
205
205
202
Equipment Leaks
88
87
109
108
104
102
101
Gas Engines
86
70
93
93
90
89
91
Chemical Injection Pumps
49
68
88
87
84
82
81
Tanks
217
60
63
68
101
61
57
Other Sources
105
86
131
131
92
103
127
Crude Oil Transportation
7
5
8
8
8
8
8
Refining
27
31
31
33
33
33
31
Total	1,844 1,553	1,739 1,622 1,559 1,548 1,449
Note: Totals may not sum due to independent rounding.
a Exploration includes well drilling, testing, and completions.
Table 3-39: CO2 Emissions from Petroleum Systems (MMT CO2)
Activity
1990
2005
2014
2015
2016
2017
2018
Exploration
0.3
0.3
3.1
2.2
1.2
1.6
2.8
Production
6.0
8.1
24.1
26.4
17.8
19.2
30.3
Transportation
+
+
+
+
+
+
+
Crude Refining
3.3
3.7
3.4
4.1
4.0
3.7
3.7
Total
9.6
12.2
30.5
32.6
23.0
24.5
36.8
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02.
Table 3-40: CO2 Emissions from Petroleum Systems (kt CO2)
Activity
1990
2005
2014
2015
2016
2017
2018
Exploration
330
348
3,060
2,221
1,233
1,566
2,761
Production
6,014
8,087
24,056
26,355
17,755
19,190
30,317
Transportation
0.9
0.7
1.2
1.2
1.1
1.1
1.2
Crude Refining
3,284
3,728
3,419
4,067
3,991
3,714
3,734
Total
9,630
12,163
30,536
32,644
22,980
24,472
36,814
Note: Totals may not sum due to independent rounding.
Table 3-41: N2O Emissions from Petroleum Systems (metric tons CO2 Eq.)
Activity
1990
2005
2014
2015
2016
2017
2018
Exploration
172
178
1,563
1,139
628
690
1,623
Production
7,483
8,173
18,464
20,329
15,341
15,466
58,809
Transportation
NE
NE
NE
NE
NE
NE
NE
Crude Refining
9,138
10,372
9,659
11,656
11,575
10,796
10,557
Total
16,793
18,723
29,686
33,124
27,544
26,951
70,988
Note: Totals may not sum due to independent rounding.
NE (Not Estimated)
Table 3-42: N2O Emissions from Petroleum Systems (metric tons N2O)
Activity	1990	2005	2014 2015 2016 2017 2018
Exploration	0.6	0.6	5.2	3.8	2.1	2.3	5.4
Production	25.1	27.4	62.0	68.2	51.5	51.9 197.3
Energy 3-71

-------
Transportation	NE	NE	NE	NE	NE	NE	NE
Crude Refining	307	34J5	32.4	39.1	38.8	36.2	35.4
Total	5^4	6Z8	99.6 111.2 92.4 90.4 238.2
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, and 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).
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. 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, and analysis of GHGRP data (EPA 2019).
Emission factors for hydraulically fractured (HF) oil well completions and workovers (in four control categories)
were developed using 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 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.77 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 2018, 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 CO2 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. Data
77 See .
3-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
from all other basins were combined, and activity and emission factors developed for the other basins as a single
group for each emission source.
For the exploration and production segments, in general, CO2 emissions for each source were estimated with
GHGRP data or by multiplying CO2 content factors by the corresponding CH4 data, as the CO2 content of gas relates
to the CH4 content of gas. Sources with CO2 emission estimates calculated using GHGRP data were 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, CO2 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 CO2 are generally estimated by multiplying the CH4 emission factors by a conversion factor, which is the
ratio of CO2 content and CH4 content in produced associated gas.
For the exploration and production segments, N2O emissions were estimated for flaring sources using GHGRP data.
Sources with N2O emissions in the exploration segment were well testing and HF completions with flaring. Sources
with N2O emissions in the production segment were associated gas flaring, tank flaring, miscellaneous production
flaring, and HF workovers with flaring.
For crude oil transportation, emission factors for CH4 were largely developed using data from EPA (1997), API
(1992), and EPA (1999). Emission factors for CO2 were estimated by multiplying the CH4 emission factors by a
conversion factor, which is the ratio of CO2 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, CO2, and N2O emissions for all major activities starting
with emissions that occurred in 2010. The reported total of CH4, CO2, and N2O 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 CO2
emission factors (i.e., sum of activity emissions/sum of refinery feed) and 2010 to 2017 data were used to derive
N2O emission factors, which 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.
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 Drillinglnfo data (Enverus Drillinglnfo 2019), Energy Information
Administration (EIA) reports, Methane Emissions from the Natural Gas Industry by the Gas Research Institute and
EPA (EPA/GRI1996), 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 2019).
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 were not yet available.
A complete list of references for emission factors and activity data by emission source is provided in Annex 3.5.
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,
Energy 3-73

-------
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).78
EPA used Microsoft Excel's @ RISK add-in tool to estimate the 95 percent confidence bound around methane
emissions from petroleum systems for the current Inventory, then applied the calculated bounds to both Cm and
CO2 emissions estimates. Uncertainty estimates for N2O were not developed given the minor contribution of N2O
to emission totals. For the analysis, EPA focused on the six highest methane-emitting sources for the year 2018,
which together emitted 75 percent of methane from petroleum systems in 2018, and extrapolated the estimated
uncertainty for the remaining sources. 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." 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 provide the 95 percent confidence bound within which actual emissions from this
source category are likely to fall for the year 2018, using the recommended IPCC methodology. The results of the
Approach 2 uncertainty analysis are summarized in Table 3-43. Petroleum systems CFU emissions in 2018 were
estimated to be between 25.0 and 48.4 MMT CO2 Eq., while CO2 emissions were estimated to be between 25.4
and 49.3 MMT CO2 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.
Table 3-43: Approach 2 Quantitative Uncertainty Estimates for ChU and CO2 Emissions from
Petroleum Systems (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)b
(MMT CO? Eq.)

(%)



Lower Upper
Lower
Upper



Bound Bound
Bound
Bound
Petroleum Systems
ch4
36.2
25.0 48.4
-31%
+34%
Petroleum Systems0
C02
36.8
25.4 49.2
-31%
+34%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for the year 2018 CH4 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.
c An uncertainty analysis for the petroleum systems C02 emissions was not performed. The relative uncertainty estimated
(expressed as a percent) from the CH4 uncertainty analysis was applied to the point estimate of petroleum systems C02
emissions.
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
78 See .
3-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
memos are cited in the Recalculations Discussion). For information on other sources, please see the Methodology
Discussion above and Annex 3.5.
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
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 a stakeholder webinar on greenhouse gas data for oil and gas in September of 2019, and a
workshop in November of 2019. 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. A key challenge in using these types of studies
to assess Inventory results is having a relevant basis for comparison (i.e., 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.80 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.81
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 2018 GHGRP data indicates that 2 refineries stopped reporting in 2018
(i.e., 2017 is the last reported year). One of them permanently shutdown towards the end of 2017 and the other
one did not report in 2018 due to a merger. Based on this assessment, cessation of reporting does not impact the
79	See .
80	See .
81	See .
Energy 3-75

-------
completeness of data for 2018 refinery emissions and therefore no adjustment has been made to these estimates
for the Inventory.
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 2019, EPA released a draft memorandum that discussed changes under consideration and requested
stakeholder feedback on those changes. EPA then created an updated version of the memorandum to document
the methodology implemented into the current Inventory.82 The EPA memorandum Inventory of U.S. Greenhouse
Gas Emissions and Sinks 1990-2018: Update for Offshore Production Emissions (Offshore Production memo) is cited
in the Recalculations Discussion below.
EPA thoroughly evaluated relevant information available and made updates to production and refinery segment
methodologies for the Inventory, specifically: using updated BOEM, GHGRP, and other data to calculate emissions
and activity factors for offshore oil production, and revisiting emissions data for delayed coking in refineries to be
consistent with changes to Subpart Y. In addition, certain sources did not undergo methodological updates, but
Cm and/or CO2 emissions changed by greater than 0.05 MMT CO2 Eq., comparing the previous estimate for 2017
to the current (recalculated) estimate for 2017 (the emissions changes were mostly due to GHGRP data submission
revisions); these sources are discussed below and include hydraulically fractured oil well completions and
workovers, associated gas flaring, miscellaneous production flaring, and pneumatic controllers.
The combined impact of revisions to 2017 petroleum systems CH4 emission estimates, compared to the previous
Inventory, is an increase from 37.7 to 38.7 MMT CO2 Eq. (1.0 MMT CO2 Eq., or 3 percent). The recalculations
resulted in an average increase in CH4 emission estimates across the 1990 through 2017 time series, compared to
the previous Inventory, of 3.5 MMT CO2 Eq., or 9 percent, with the largest increase being in the estimate for 1996
(5.2 MMT CO2 Eq. or 14 percent) due to the recalculations for offshore oil production.
The combined impact of revisions to 2017 petroleum systems CO2 emission estimates, compared to the previous
Inventory, is an increase from 23.3 to 24.5 MMT CO2 (1.1 MMT CO2, or 5 percent). The recalculations resulted in an
average increase in emission estimates across the 1990 through 2017 time series, compared to the previous
Inventory, of 0.8 MMT CO2 Eq., or 6 percent with the largest changes being for 2017 (1.1 MMT CO2 or 5 percent)
due to the recalculations for offshore oil production.
The combined impact of revisions to 2017 petroleum systems N2O emission estimates, compared to the previous
Inventory, is an increase of 0.003 MMT CO2, Eq. or 11 percent. The recalculations resulted in an average increase in
emission estimates across the 1990 through 2017 time series, compared to the previous Inventory, of 0.003 MMT
CO2 Eq., or 19 percent.
In Table 3-44 and Table 3-45 below are categories in Petroleum Systems with updated methodologies or with
recalculations resulting in a change of greater than 0.05 MMT CO2 Eq., comparing the previous estimate for 2017
to the current (recalculated) estimate for 2017. For more information, please see the Recalculations Discussion
below.
Table 3-44: Recalculations of CO2 in Petroleum Systems (MMT CO2)
Previous Estimate	Current Estimate	Current Estimate
Year 2017,	Year 2017,	Year 2018,
2019 Inventory	2020 Inventory	2020 Inventory
1.7	1.6	2.8
1.6	1.5	2.7
18.0	19.2	30.3
Exploration
HF Oil Well Completions
Production
82 Stakeholder materials including draft and final memoranda for the current (i.e., 1990 to 2018) Inventory are available at
.
3-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Offshore Oil Production
+
0.5
0.5
Associated Gas Venting & Flaring
10.5
10.9
19.0
Miscellaneous Flaring
2.6
3.1
4.2
HF Oil Well Workovers
0.3
0.2
0.1
Transportation
+
+
+
Refining
3.7
3.7
3.7
Petroleum Systems Total
23.3
24.5
36.8
+ Does not exceed 0.05 MMT C02.



Table 3-45: Recalculations of ChU
in Petroleum Systems (MMT CO2 Eq.)


Previous Estimate
Current Estimate
Current Estimate

Year 2017,
Year 2017,
Year 2018,

2019 Inventory
2020 Inventory
2020 Inventory
Exploration
0.4
0.3
0.4
Production
36.4
37.4
34.9
Pneumatic Controllers
20.9
21.3
18.4
Offshore Oil Production
4.7
5.1
5.1
Transportation
0.2
0.2
0.2
Refining
0.7
0.8
0.8
Delayed Cokers
+
0.1
0.1
Petroleum Systems Total
37.7
38.7
36.2
+ Does not exceed 0.05 MMT C02 Eq.
Exploration
HF Oil Well Completions (Recalculation with Updated Data)
HF oil well completion CO2 emissions increased by an average of 9 percent across the time series and decreased by
6 percent in 2017, compared the to the previous Inventory. The CO2 emissions changes are due to GHGRP data
submission revisions. The recalculation of the EF for non-REC with flaring HF oil well completions had the largest
impact on times series emissions. Compared to the previous Inventory, the EF for non-REC with flaring increased
by 13 percent for all years of the time series except 2017; in 2017 it decreased by 6 percent.
Table 3-46: HF Oil Well Completions National CO2 Emissions (kt CO2)
Source		1990	2005	2014 2015 2016 2017 2018
HF Completions: Non-REC with
Venting
HF Completions: Non-REC with
Flaring
HF Completions: REC with
Venting
HF Completions: REC with
Flaring
Total Emissions
Previous Estimate
+ Does not exceed 0.05 kt C02.
NA (Not Applicable)
Production
Offshore Oil Production (Methodological Update)
2.5
89
0.0
0.0
92
81
4.0
139
0.0
0.0
143
127
4.0
690
0.2
2,107
2,801
2.719
1.4
446
0.2
1,518
1,966
1.913
0.2
252
0.1
940
1,192
1.162
0.2
360
0.1
1,168
1,529
1.619
552
0.1
2,178
2,730
NA
Energy 3-77

-------
EPA updated the offshore production methodology to estimate emissions for all offshore producing regions and to
use activity data sources that provide a full time series of data. The previous Inventory only estimated emissions
for offshore facilities in federal waters of the Gulf of Mexico (GOM); these facilities are under Bureau of Ocean
Energy Management (BOEM) jurisdiction and BOEM estimates their greenhouse gas emissions triennially via the
Gulfwide Emissions Inventory (GEI). The previous Inventory also relied on activity data sources that were no longer
updated, and surrogate activity data from 2008 and 2010 had been used to estimate emissions in more recent
years. The updated Inventory methodology now includes emissions estimates for offshore facilities in federal and
state waters of the GOM and offshore facilities in the Pacific and off the coast of Alaska.
The updated Inventory methodology for each region is presented here. EPA calculated vent and leak EFs for
offshore facilities in GOM federal waters for major complexes and minor complexes using BOEM GEI emissions
data from the 2005, 2008, 2011, 2014, and 2017 GEIs. Vent and leak EFs were calculated for 10 emission sources
(cold vents, equipment leaks, pneumatic pumps, losses from flashing, pneumatic controllers, combustion, glycol
hydrators, storage tanks, mud degassing, minor surrogates, and amine gas sweetening units) and paired with
active offshore complex counts over the time series. EPA calculated GOM federal waters flaring emissions using
flaring volumes reported in Oil and Gas Operations Reports (OGOR), Part B (OGOR-B). OGOR-B flaring volumes are
available over the time series but assumptions were necessary to assign the volumes to offshore gas production
versus offshore oil production for 1990 to 2010. The previous Inventory allocated all GOM federal waters flaring
emissions to offshore gas production facilities. EPA calculated production based EFs for offshore facilities in GOM
state waters using the resulting GOM federal waters emissions and oil production in each year. EPA also calculated
production based EFs for offshore facilities in the Pacific and Alaska regions, though the EFs for these regions were
derived from GHGRP data. EPA multiplied the production based EFs by the region-specific offshore production (i.e.,
GOM state waters production, Pacific production, and Alaska production) in a given year. The Offshore Production
memo provides details for the methodology update that was implemented into the Inventory.
Due to this recalculation, annual offshore oil production Cm emission estimates increased in the current Inventory
for 1990 to 2017 by an average of 67 percent, compared to the previous Inventory. The impacts varied across the
time series with estimates in 1990 through 2009 increasing by an average of84 percent and estimates in 2010
through 2017 increasing by an average of 25 percent. The increase in offshore oil production Cm emission
estimates over the time series are due in part to the inclusion of emissions from facilities located in GOM state
waters and the Pacific and Alaska regions. The increase in offshore oil production Cm emission estimates for 1990
to 2009 also resulted from an increase in calculated emissions for GOM federal waters due to differences in EFs
and activity data between the current and previous Inventory. The current Inventory applied EFs calculated from
2008 GEI data for this time period, whereas the previous Inventory applied EFs calculated from 2011 GEI data for
this time period and the 2008 GEI Cm emissions are higher. There are more offshore oil facilities in the current
Inventory compared to the previous Inventory. The current and previous Inventories have a different activity basis
(i.e., offshore complexes versus offshore structures), but a much higher percentage of offshore facilities in the
current Inventory are classified as oil rather than gas (an average of 66 percent oil facilities for 1990 through 2009)
compared to the previous Inventory (an average of 41 percent oil facilities over the same time period).
For comparison, total offshore production (for oil and gas combined) Cm emissions for facilities in GOM federal
waters are provided here for years 2011, 2014, and 2017 from the GEI, previous Inventory, and current Inventory.
For offshore facilities in GOM federal waters in year 2011, GEI Cm emissions equaled 246 kt, previous Inventory
Cm emissions equaled 338 kt, and current Inventory CH4 emissions equal 278 kt. For offshore facilities in GOM
federal waters in year 2014, GEI CH4 emissions equaled 205 kt, previous Inventory CH4 emissions equaled 338 kt,
and current Inventory CFU emissions equal 225 kt. For offshore facilities in GOM federal waters in year 2017, GEI
Cm emissions equaled 170 kt, previous Inventory CH4 emissions equaled 338 kt, and current Inventory CH4
emissions equal 206 kt.
Annual offshore oil production CO2 emission estimates increased in the current Inventory for 1990 to 2017 by a
factor of 72 on average, compared to the previous Inventory. This change is largely because all GOM federal
waters flaring emissions in the previous Inventory were allocated to offshore gas production, whereas the current
Inventory estimates GOM federal waters flaring emissions for both offshore gas and oil production, and a
significant portion of the CO2 is from offshore oil production. In addition, the Alaska region (which was not
3-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
previously included) is a significant contributor to CO2 emissions, due to flaring, and accounts for the highest
fraction of CO2 emissions from 1990 through 2007 in the current Inventory.
EPA received feedback on this update through its September 2019 memo and through the public review draft of
the Inventory. Two stakeholders supported the update to activity data. A stakeholder suggested clarifications on
the calculation of emission factors, and noted upcoming data that may be used to assess offshore emission factors.
A stakeholder suggested clarification on the development of activity counts and supported considering a different
approach which would use source-specific emission factors. As noted above, the emissions estimates were
calculated using complex-level factors for offshore operations in GOM federal waters, and using production-based
emission factors for offshore operations in state waters. An estimate of emissions source-level emissions was
developed using the fraction of emissions in each category in the GOM federal waters data set, applied to GOM
federal and state water total emission estimates, and using the fraction of emissions in each category in GHGRP for
Pacific and Alaska offshore production, and applied to the total estimates for Pacific and Alaska offshore
production. The emission source-level estimates are available in the annex. The stakeholder noted that the use of
emission factors calculated from data from the from the GHGRP reporting population (those emitting over the
GHGRP threshold), applied to all Pacific and Alaska offshore production could skew regional emission estimates.
The stakeholder also supported the use of GEI data as opposed to OGOR-B data to calculate emissions from flaring.
The emissions estimates were calculated using OGOR-B. GEI data is currently available for the years 2005, 2008,
2011, 2014, and 2017. The OGOR-B dataset can be used to calculate flaring emissions for the full 1990 to 2018
time series.
The recalculation also results in a change in the trend, in methane in particular where the 1990 to 2017 trend in
this Inventory is a decrease of 45 percent, versus a decrease of 11 percent in the previous Inventory. A stakeholder
provided several factors supporting this decreasing trend: more stringent limitations imposed by BSEE (Bureau of
Safety and Environmental Enforcement) related to venting and flaring, increased utilization of VRU equipment, and
replacement of older platforms with newer ones that include state of the art technology.
Table 3-47: Offshore Oil Production National ChU Emissions (metric tons ChU)
Source
1990
2005
2014
2015 2016

2017

2018
GOM Federal Waters
303,520
219,422
203,201
197,233 189,145
186,806
186,138
GOM State Waters
24,302
2,860
2,381
1,979 1,655

1,222

1,130
Pacific Waters
22,610
17,660
13,790
10,308 5,008

5,052

5,163
Alaska State Waters
21,936
21,192
10,516
10,703 9,680
12,164

9,834
Total Emissions
372,368
261,133
229,888
220,223 205,488
205,243
202,265
Previous Estimate
210,938
185,023
187,604
187,604 187,604
187,604

NA
NA (Not Applicable)








ible 3-48: Offshore Oil Production National CO2 Emissions (metric tons CO2)



Source
1990
2005
2014
2015 2016

2017
2018
GOM Federal Waters
188,356
147,743
. 313,103
368,773 373,468
379,413
414,023
GOM State Waters
15,081
1,926
3,669
3,700 3,269

2,482
2,514
Pacific Waters
70,319
54,925
42,889
32,060 11,052
13,440
8,688
Alaska State Waters
357,965
345,809
171,607
174,652 122,554
119,963
122,362
Total Emissions
631,721
550,402
531,267
579,185 510,342
515,299
547,587
Previous Estimatea
9,604
8,283
8,340
8,340 8,340

8,340
NA
NA (Not Applicable)
a Includes only C02 from leaks and vents.
HF Oil Well Workovers (Recalculation with Updated Data)
HF oil well workover CO2 emissions increased by an average of 8 percent across the time series, and decreased by
30 percent in 2017, compared the to the previous Inventory. The CO2 emissions changes are due to GHGRP data
submission revisions, which resulted in a recalculation of emission factors and activity data. HF oil well workover
CO2 time series emissions were most impacted by the recalculation of the EF for non-REC HF oil well workovers
Energy 3-79

-------
with flaring, which increased by 13 percent for 1990 to 2016 (compared to the previous Inventory). The
recalculation of activity data for REC HF oil well workovers with flaring had the largest impact on year 2017
emissions, with a smaller fraction of the population using REC with flaring.
Table 3-49: HF Oil Well Workovers National CO2 Emissions (kt CO2)
Source
1990
2005
2014
2015
2016
2017
2018
HF Workovers: Non-REC with







Venting
0.7
0.8
0.4
0.2
0.1
0.0
+
HF Workovers: REC with Venting
0.0
0.0
+
+
0.1
+
+
HF Workovers: Non-REC with







Flaring
25.1
28.3
36.6
35.2
32.2
18.2
3.6
HF Workovers: REC with Flaring
0.0
0.0
133.1
157.8
175.6
160.8
89.3
Total Emissions
25.8
29.1
170.1
193.3
207.8
179.0
92.9
Previous Estimate
22.9
25.8
168.3
192.1
207.4
257.5
NA
+ Does not exceed 0.05 kt C02.
NA (Not Applicable)
Pneumatic Controllers (Recalculation with Updated Data)
Pneumatic controller CH4 emission estimates increased by an average of less than 1 percent across the 1990 to
2017 time series, compared to the previous Inventory, due to GHGRP submission revisions and a small increase in
well counts throughout the time series due to updated Drilling Info data.
Table 3-50: Pneumatic Controller National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Pneumatic Controllers: High







Bleed
722,968
420,444
88,574
78,213
82,555
52,608
39,088
Pneumatic Controllers: Low







Bleed
49,343
44,058
28,772
25,461
17,517
19,651
30,628
Pneumatic Controllers: Int Bleed
0.0
239,899
665,830
685,810
722,917
778,365
665,108
Total Emissions
772,311
704,401
783,176
789,484
822,989
850,624
734,824
Previous Estimate
773,655
700,990
776,512
785,704
818,169
836,804
NA
NA (Not Applicable)
Associated Gas Flaring (Recalculation with Updated Data)
Associated gas flaring CO2 emission estimates increased by an average of 2 percent across the time series in the
current Inventory, compared to the previous Inventory. This change was due to GHGRP submission revisions. The
changes in CO2 emissions for 2017 (the year with the largest change) were mainly driven by the Williston and
Permian Basin data.
Table 3-51: Associated Gas Flaring National CO2 Emissions (kt CO2)
Source
1990
2005
2014
2015
2016
2017
2018
220 - Gulf Coast Basin (LA, TX)
234
127
631
673
404
740
686
360 - Anadarko Basin
108
65
230
238
2
57
37
395 - Williston Basin
966
1,239
7,799
8,412
5,838
6,530
10,132
430- Permian Basin
2,983
2,046
3,869
4,443
2,246
3,148
7,249
"Other" Basins
925
499
520
544
326
414
876
Total Emissions
5,217
3,977
13,050
14,311
8,815
10,889
18,980
220 - Gulf Coast Basin (LA, TX)
233
126
631
673
350
688
NA
360 - Anadarko Basin
106
65
222
239
2
55
NA
395 - Williston Basin
925
1,186
7,466
8,052
5,662
6,451
NA
430 - Permian Basin
2,982
2,048
3,869
4,447
2,247
2,897
NA
"Other" Basins
927
499
523
544
325
416
NA
3-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Previous Estimate	5,172	3,925	12,711 13,955 8,587 10,506	NA
NA (Not Applicable)
Miscellaneous Production Flaring (Recalculation with Updated Data)
Miscellaneous production flaring CO2 emission estimates increased by 17 percent in 2017 and increased by less
than 1 percent for other years of the time series, compared to the previous Inventory. The 2017 increase was
primarily due to recalculations of CO2 from flaring in the Permian and Williston basins, where GHGRP resubmission
revisions showed higher CO2 emissions from flaring, by 65 and 20 percent, respectively.
Table 3-52: Miscellaneous Production Flaring National CO2 Emissions (kt CO2)
Source
1990
2005
2014
2015
2016
2017
2018
220 - Gulf Coast Basin (LA, TX)
0
106
893
997
497
526
687
395 - Williston Basin
0
73
782
882
315
531
1,653
430- Permian Basin
0
215
687
825
794
1,424
1,183
Other Basins
0
407
796
870
592
585
703
Total Emissions
0
801
3,159
3,573
2,198
3,066
4,226
220 - Gulf Coast Basin (LA, TX)
0
107
901
1,005
496
523
NA
395 - Williston Basin
0
73
776
875
309
321
NA
430 - Permian Basin
0
215
686
824
794
1,185
NA
Other Basins
0
406
794
867
601
601
NA
Previous Total Estimate
0
800
3,157
3,571
2,201
2,631
NA
NA (Not Applicable)
Well Counts (Recalculation with Updated Data)
For total national well counts, EPA has used a more recent version of the Drillinglnfo dataset (Enverus Drillinglnfo
2019) to update well counts data in the Inventory. While this is not a significant recalculation (the update results in
an average increase of less than 1 percent), the well count dataset is a key input to the Inventory, and results are
highlighted here.
Table 3-53: Producing Oil Well Count Data
Oil Well Count
1990
2005
2014
2015
2016
2017
2018
Number of Oil Wells
Previous Estimate
562,356
564,090
482,887
480,482
i 610,121
605,259
600,519
597,635
580,917
577,515
570,331
566,726
564,186
NA
NA (Not Applicable)
In December 2019, EIA released an updated time series of national oil and gas well counts (covering 2000 through
2018). EIA estimates 982,371 total producing wells for year 2018. EPA's total well count for this year is 969,212.
EPA's well counts in recent time series years are generally 1 percent lower than ElA's. ElA's well counts include side
tracks, 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 and a higher fraction
of gas wells than EPA.
Refining
Refinery Cm emissions increased by an average of 12 percent across the time series, compared to the previous
Inventory, due to a recalculation of delayed coker emissions. The Subpart Y calculation methodology for delayed
cokers was updated for reporting year 2018 to use more accurate methods to quantify emissions for delayed
cokers. The update to the calculation methodology resulted in higher reported emissions from delayed cokers in
Energy 3-81

-------
2018 compared to previous years of reporting. The update did not impact all facilities in Subpart Y as some
facilities had already been reporting using the more accurate methods. For time-series consistency across 1990 to
2018 in the Inventory, emission estimates were updated for 1990 through 2017 using a ratio of reported emissions
for 2018 to 2017, comparing facilities that used different methods for those years. A stakeholder supported this
approach to updating estimates for delayed coker emissions.
Table 3-54: Refineries National ChU Emissions (metric tons ChU)
Source
1990

2005

2014
2015
2016
2017
2018
Delayed Cokers
Other Refining Sources
Total Refinery Emissions
3,873
23,299
27,172

4,395
26,445
30,841

5,506
25,089
30,595
5,447
27,294
32,742
5,787
27,245
33,032
5,142
27,992
33,134
5,435
25,503
30,938
Previous Delayed Cokers
Previous Other Refining Sources
Previous Total Refinery Estimate
1,146
23,294
24,440

1,301
26,440
27,740

1,057
24,979
26,036
931
27,271
28,202
960
27,171
28,131
1,029
27,305
28,333
NA
NA
NA
NA (Not Applicable)
Planned Improvements
Offshore Production
EPA updated the offshore production methodology for the Inventory, incorporating data from BOEM and GHGRP.
Detailed information and considerations for various approaches considered for the methodology update were
provided in a memorandum and discussed at a stakeholder workshop and webinar. Through the stakeholder
process and the public review period, stakeholders provided feedback on additional approaches or data sets that
could be used. In future inventories, EPA will consider alternate approaches or data sources, such as additional use
of BOEM data or data from upcoming studies. Stakeholders identified upcoming studies of offshore oil and gas
platform emissions that will include evaluation of different inventory estimates and methods.
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 malfunction and control efficiency 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
3-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
commercial and non-EOR industrial applications. This CO2 is produced from both naturally-occurring CO2
reservoirs and from industrial sources such as natural gas processing plants and ammonia plants. In the
Inventory, emissions from naturally-produced CO2 are estimated based on the specific application.
In the Inventory, CChthat 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.
For EOR CO2, as noted in the 2006IPCC Guidelines, "At the Tier 1 or 2 methodology levels [EOR CO2 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 CO2 include CO2 that was originally
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 CO2 from EOR and CO2 occurring in the
produced natural gas. Therefore, EOR CO2 emitted through those pathways is included in CO2 estimates in 1B2.
IPCC includes methodological guidance to estimate emissions from the capture, transport, injection, and
geological storage of CO2. 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 CO2 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 CO2 emission estimates for geologic
storage.
In the United States, facilities that produce CO2 for various end-use applications (including capture facilities such
as acid gas removal plants and ammonia plants), importers of CO2, exporters of CO2, facilities that conduct
geologic sequestration of CO2, and facilities that inject CO2 underground, are required to report greenhouse gas
data annually to EPA through its GHGRP. Facilities conducting geologic sequestration of CO2 are required to
develop and implement an EPA-approved site-specific monitoring, reporting and verification plan, and to report
the amount of CO2 sequestered using a mass balance approach.
GHGRP data relevant for this inventory estimate consists of national-level annual quantities of CO2 captured and
extracted for EOR applications for 2010 to 2018. However, for 2015 through 2018, data from EPA's GHGRP
(Subpart PP) were held constant from 2014 levels, for data confidentiality reasons. EPA will continue to evaluate
the availability of additional GHGRP data and other opportunities for improving the estimates. Several facilities
are reporting under Subpart RR (Geologic Sequestration of Carbon Dioxide). In 2016, one facility reported 3.1
MMT of CO2 sequestered in subsurface geological formations and 9,818 metric tons of CO2 emitted from
equipment leaks. In 2017, three facilities reported 9.1 MMT of CO2 sequestered in subsurface geological
formations, and 9,577 metric tons of CO2 emitted from equipment leaks. In 2018, five facilities reported 16.7
MMT of CO2 sequestered in subsurface geological formations and 11,023 metric tons of CO2 emitted from
equipment leaks.
The amount of CO2 captured and extracted from natural and industrial sites for EOR applications in 2018 is 59.3
MMT CO2 Eq. (59,318 kt) (see Table 3-55 and Table 3-56). The quantity of CO2 captured and extracted is noted
here for information purposes only; CO2 captured and extracted from industrial and commercial processes is
assumed to be emitted and included in emissions totals from those processes.
Table 3-55: Quantity of CO2 Captured and Extracted for EOR Operations (MMT CO2)
Stage
1990
2005
2014
2015
2016
2017
2018
Capture Facilities
4.8
6.5
13.1
13.1
13.1
13.1
13.1
Extraction Facilities
20.8
28.3
46.2
46.2
46.2
46.2
46.2
Total
25.6
34.7
59.3
59.3
59.3
59.3
59.3
Note: Totals may not sum due to independent rounding.
Energy 3-83

-------
Table 3-56: Quantity of CO2 Captured and Extracted for EOR Operations (kt)
Stage
1990
2005
2014
2015
2016
2017
2018
Capture Facilities
4,832
6,475
13,093
13,093
13,093
13,093
13,093
Extraction Facilities
20,811
28,267
46,225
46,225
46,225
46,225
46,225
Total
25,643
34,742
59,318
59,318
59,318
59,318
59,318
Note: Totals may not sum due to
independent rounding.





3.7 Natural Gas Systems (CRF Source Category
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, CO2, and
N2O) from natural gas systems in 2018 were 174.9 MMT CO2 Eq., a decrease of 19 percent from 1990, primarily
due to decreases in CFU emissions, and an increase of 3 percent from 2017, primarily due to increases in CO2
emissions. From 2008, emissions decreased by 6 percent, primarily due to decreases in CH4 emissions. National
total dry gas production in the United States increased by 71 percent from 1990 to 2018, and by 12 percent from
2017 to 2018, and by 52 percent from 2008 to 2018. Of the overall greenhouse gas emissions (174.9 MMT CO2 Eq.),
80 percent are CFU emissions (140.0 MMT CO2 Eq.), 20 percent are CO2 emissions (35.0 MMT), and less than 0.01
percent are N2O emissions (0.01 MMT CO2 Eq.).
Overall, natural gas systems emitted 140.0 MMT CO2 Eq. (5,598 kt CH4) of CFU in 2018, a 24 percent decrease
compared to 1990 emissions, and less than 1 percent increase compared to 2017 emissions (see Table 3-57 and
Table 3-58). There was a total of 35.0 MMT CO2 Eq. (34,972 kt) of non-combustion CO2 in 2018, an 9 percent
increase compared to 1990 emissions, and a 15 percent increase compared to 2017 levels. The 2018 N2O emissions
were estimated to be 0.01 MMT CO2 Eq. (0.03 kt N2O), a 116 percent increase compared to 1990 emissions.
The 1990 to 2018 trend is not consistent across segments or gases. Overall, the 1990 to 2018 decrease in CH4
emissions is due primarily to the decrease in emissions from the following segments: distribution (73 percent
decrease), transmission and storage (41 percent decrease), processing (43 percent decrease), and exploration (72
percent decrease). Over the same time period, the production segment saw increased CH4 emissions of 41 percent
(with onshore production emissions increasing 30 percent, offshore production emissions decreasing 80 percent,
and gathering and boosting [G&B] emissions increasing 91 percent). The 1990 to 2018 increase in CO2 emissions is
primarily due to increase in CO2 emissions in the production segment, where emissions from flaring have increased
overtime.
Methane and CO2 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 N2O from flaring activities are included in
the Inventory, with most of the emissions occurring in the processing and production segments. Note, CO2
emissions exclude all combustion emissions (e.g., engine combustion) except for flaring CO2 emissions. All
combustion CO2 emissions (except for flaring) are accounted for in Section 3.1- CO2 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 2017) to
ensure that the trend is accurate. Recalculations in natural gas systems in this year's Inventory include:
3-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
•	Updated methodology for G&B stations to use data from GHGRP, Zimmerle et al. 2019, and other sources.
•	Updated methodology for offshore gas production to use data from BOEM, GHGRP, and other sources.
•	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, CO2, and N2O emissions are discussed.
Exploration. Exploration includes well drilling, testing, and completions. Emissions from exploration account for 1
percent of CH4 emissions and 1 percent of CO2 emissions from natural gas systems in 2018. Well completions
account for approximately 97 percent of CH4 emissions from the exploration segment in 2018, with the rest
resulting from well testing and drilling. Flaring emissions account for most of the CO2 emissions. Methane
emissions from exploration decreased by 72 percent from 1990 to 2018, with the largest decreases coming from
hydraulically fractured gas well completions without reduced emissions completions (RECs). Methane emissions
decreased 10 percent from 2017 to 2018 due to decreases in emissions from hydraulically fractured well
completions. Methane emissions were highest from 2006 to 2008. Carbon dioxide emissions from exploration
increased by 1 percent from 1990 to 2018, and decreased 10 percent from 2017 to 2018 due to decreases in
flaring. Carbon dioxide emissions were highest from 2006 to 2008. Nitrous oxide emissions increased 80 percent
from 1990 to 2018, and increased 53 percent from 2017 to 2018.
Production (including gathering and boosting). In the production stage, 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) account for 58 percent of CH4 emissions and 27
percent of CO2 emissions from natural gas systems in 2018. Emissions from gathering and boosting and pneumatic
controllers in onshore production, account for most of the production segment CH4 emissions in 2018. Within
gathering and boosting, the largest sources are compressor exhaust slip, compressor venting and leaks, and
pneumatic controllers. Flaring emissions account for most of the CO2 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 41 percent from 1990 to 2018, 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 decreased 2 percent from 2017 to 2018 due to decreases in the number of high
bleed and intermittent bleed controllers. Methane emissions were highest in 2008-2013. Carbon dioxide emissions
from production increased approximately by a factor of 3 from 1990 to 2018 due to increases in emissions at flare
stacks in gathering and boosting and miscellaneous onshore production flaring, and increased 47 percent from
2017	to 2018 due primarily to increases in emissions from flare stacks in gathering and boosting and flaring at
tanks. Carbon dioxide emissions were highest in 2018. Nitrous oxide emissions increased 35 percent from 1990 to
2018	and increased 36 percent from 2017 to 2018. The increase in N2O emissions from 1990 to 2018 and from
2017 to 2018 is primarily due to increase in emissions from flare stacks at gathering and boosting.
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 CO2
emissions come from acid gas removal (AGR) units, which are designed to remove CO2 from natural gas. Processing
plants account for 9 percent of CH4 emissions and 70 percent of CO2 emissions from natural gas systems. Methane
emissions from processing decreased by 43 percent from 1990 to 2018 as emissions from compressors (leaks and
Energy 3-85

-------
venting) and equipment leaks decreased; and increased 6 percent from 2017 to 2018 due to increased emissions
from gas engines and blowdowns/venting. Carbon dioxide emissions from processing decreased by 14 percent
from 1990 to 2018, due to a decrease in AGR emissions, and increased 7 percent from 2017 to 2018 due to
increased emissions from flaring. Nitrous oxide emissions increased 29 percent from 2017 to 2018.
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 Cm emissions from these compressor stations and venting from pneumatic
controllers account for most of the emissions from this stage. Uncombusted compressor engine exhaust and
pipeline venting are also sources of Cm 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 Cm emissions from storage. Emissions from
liquified natural gas (LNG) stations and terminals are also calculated under the transmission and storage segment.
Methane emissions from the transmission and storage segment account for approximately 24 percent of emissions
from natural gas systems, while CO2 emissions from transmission and storage account for 1 percent of the CO2
emissions from natural gas systems. CH4 emissions from this source decreased by 41 percent from 1990 to 2018
due to reduced compressor station emissions (including emissions from compressors and leaks), and increased 5
percent from 2017 to 2018 due to increased emissions from transmission compressor exhaust and increased
emissions from reciprocating transmission compressors. CO2 emissions from transmission and storage have
increased by a factor of 2.7 from 1990 to 2018, due to increased emissions from LNG export terminals, and
decreased by less than 1 percent from 2017 to 2018. The quantity of LNG exported from the U.S. increased by a
factor of 21 from 1990 to 2018, and by 53 percent from 2017 to 2018. LNG emissions are about 1 percent of CH4
and 61 percent of CO2 emissions from transmission and storage in year 2018. Nitrous oxide emissions from
transmission and storage decreased by 24 percent from 1990 to 2018 and decreased 58 percent from 2017 to
2018.
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 1,305,781 miles of distribution mains in 2018, an increase of nearly 361,624 miles since 1990
(PHMSA 2019). Distribution system emissions, which account for 8 percent of CH4 emissions from natural gas
systems and less than 1 percent of CO2 emissions, resulting 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
CO2 emissions from this stage, as have station upgrades at metering and regulating (M&R) stations. Distribution
system Cm emissions in 2018 were 73 percent lower than 1990 levels and less than 1 percent lower than 2017
emissions. Distribution system CO2 emissions in 2018 were 73 percent lower than 1990 levels and less than 1
percent lower than 2017 emissions. Annual CO2 emission from this segment are less than 0.1 MMT CO2 Eq. across
the time series.
Total Cm emissions for the five major stages of natural gas systems are shown in MMT CO2 Eq. (Table 3-57) and kt
(Table 3-58). 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 6.8 MMT CO2 Eq. CH4 are subtracted from production segment
emissions and 6.7 MMT CO2 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 are
available in Annex 3.6. See Methodology for Estimating CH4 and CO2 Emissions from Natural Gas Systems.
Table 3-57: ChU Emissions from Natural Gas Systems (MMT CO2 Eq.)a
Stage
1990
2005
2014
2015
2016
2017
2018
Exploration b
4.0
10.3
1.0
1.0
0.7
1.2
1.1
Production
57.2
76.9
84.6
83.7
81.8
82.3
80.9
Onshore Production
34.9
51.4
49.2
46.9
45.1
45.5
45.3
Gathering and Boosting0
18.2
23.7
34.6
36.1
35.9
36.1
34.8
3-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Offshore Production
4.1
1.8
0.8
0.6
0.8
0.7
0.8
Processing
21.3
11.6
11.0
11.0
11.2
11.5
12.2
Transmission and Storage
57.2
36.1
32.3
34.1
30.1
32.3
33.9
Distribution
43.5
23.3
12.2
12.0
12.0
11.9
11.8
Total
183.3
158.1
141.1
141.9
135.8
139.3
140.0
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 3-58: ChU Emissions from Natural Gas Systems (kt)a
Stage
1990
2005
2014
2015
2016
2017
2018
Exploration b
162
411
39
41
27
49
44
Production
2,289
3,076
3,385
3,347
3,273
3,2911
3,238
Onshore Production
1,396
2,057
1,968
1,877
1,805
1,820
1,814
Gathering and Boosting0
729
946
1,386
1,445
1,435
1,443
1,391
Offshore Production
165
73
31
24
33
28
33
Processing
853
463
440
440
448
461
488
Transmission and Storage
2,228
1,442
1,292
1,365
1,205
1,294
1,355
Distribution
1,741
932
487
481
480
476
473
Total
7,332
6,324
5,643
5,674
5,433
5,570
5,598
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 3-59: Non-combustion CO2 Emissions from Natural Gas Systems (MMT)
Stage
1990
2005
2014
2015
2016
2017
2018
Exploration
0.4
1.6
0.8
0.3
0.2
0.5
0.4
Production
3.2
4.5
7.5
7.7
7.4
6.5
9.6
Processing
28.3
18.9
21.1
21.1
21.9
22.9
24.5
Transmission and Storage
0.2
0.2
0.2
0.2
0.3
0.5
0.5
Distribution
0.1
+
+
+
+
+
+
Total
32.2
25.3
29.6
29.3
29.9
30.4
35.0
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 3-60: Non-combustion CO2 Emissions from Natural Gas Systems (kt)
Stage
1990
2005
2014
2015
2016
2017
2018
Exploration
408
1,648
843
282
190
456
410
Production
3,197
4,548
7,464
7,740
7,450
6,505
9,591
Processing
28,338
18,893
21,075
21,075
21,908
22,896
24,465
Transmission and Storage
180
174
223
223
300
493
491
Distribution
51
27
14
14
14
14
14
Total
32,174
25,291
29,620
29,334
29,862
30,365
34,972
Note: Totals may not sum due to independent rounding.
Energy 3-87

-------
Table 3-61: N2O Emissions from Natural Gas Systems (metric tons CO2 Eq.)
Stage
1990
2005
2014
2015
2016
2017
2018
Exploration
241
442
514
3,204
111
285
436
Production
4,295
5,696
8,987
9,809
8,871
4,282
5,808
Processing
NO
3,347
5,764
5,764
3,794
3,042
3,922
Transmission and Storage
256
307
341
343
361
459
195
Distribution
NO
NO
NO
NO
NO
NO
NO
Total
4,792
9,791
15,606
19,120
13,136
8,068
10,361
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
Table 3-62: N2O Emissions from Natural Gas Systems (metric tons N2O)
Stage
1990
2005
2014
2015
2016
2017
2018
Exploration
0.8
1.5
1.7
10.8
0.4
1.0
1.5
Production
14.4
19.1
30.2
32.9
29.8
14.4
19.5
Processing
NO
11.2
19.3
19.3
12.7
10.2
13.2
Transmission and Storage
0.9
1.0
1.1
1.2
1.2
1.5
0.7
Distribution
NO
NO
NO
NO
NO
NO
NO
Total
16.1
32.9
52.4
64.2
44.1
27.1
34.8
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
Methodology
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, CO2, and N2O 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.
Emission Factors. Key references for emission factors for CH4 and CO2 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
2019), and others.
The GRI/EPA study developed over 80 CH4 emission factors to characterize emissions from the various components
within the operating stages 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 CFU 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 Cm composition of natural gas for each NEMS region.
GHGRP Subpart W data were used to develop CH4, CO2, and N2O 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
3-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 Cm emission factors include Zimmerle et al. (2015) for transmission and storage
station leaks and compressors, 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 CO2 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 2001GTI publication were used
to adapt the CH4 emission factors into related CO2 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 CO2 emission factors.
Flaring N2O 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, CO2, and N2O
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; Enverus Drillinglnfo, Inc. (Enverus
Drillinglnfo 2019); BOEM; Federal Energy Regulatory Commission (FERC); EIA; the Natural Gas STAR Program
annual data; Oil and Gas Journal; PHMSA; the Wyoming Conservation Commission; and the Alabama State Oil and
Gas Board.
For a few sources, recent direct activity data are not available. For these sources, either 2017 data were used as a
proxy for 2018 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 initial stakeholder feedback on
updates under consideration for the Inventory. Stakeholder feedback is noted below in Uncertainty and Time-
Series Consistency, Recalculations Discussion, and Planned Improvements.
Uncertainty and Time-Seri insistency
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
Energy 3-89

-------
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).83 EPA used Microsoft Excels @RISK
add-in tool to estimate the 95 percent confidence bound around CFU emissions from natural gas systems for the
current Inventory, then applied the calculated bounds to both Cm and CO2 emissions estimates. For the analysis,
EPA focused on the 13 highest-emitting sources for the year 2018, which together emitted 83 percent of methane
from natural gas systems in 2018, and extrapolated the estimated uncertainty for the remaining sources. 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 2018, using the IPCC methodology. The results of the Approach 2
uncertainty analysis are summarized in Table 3-63. Natural gas systems CH4 emissions in 2018 were estimated to
be between 118.2 and 159.6 MMT CO2 Eq. at a 95 percent confidence level. Natural gas systems CO2 emissions in
2018 were estimated to be between 29.5 and 39.9 MMT CO2 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.
Table 3-63: Approach 2 Quantitative Uncertainty Estimates for CH4 and Non-combustion CO2
Emissions from Natural Gas Systems (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)b
(MMT CO? Eq.)
(%)



Lower Upper
Lower Upper



Boundb Bound'5
Boundb Bound'5
Natural Gas Systems
ch4
140.0
118.2 159.6
-15% +14%
Natural Gas Systems0
C02
35.0
29.5 39.9
-15% +14%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for the year 2018 CH4 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 3-57 and Table 3-58.
c An uncertainty analysis for the C02 emissions was not performed. The relative uncertainty estimated (expressed as a
percent) from the CH4 uncertainty analysis was applied to the point estimate of C02 emissions.
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.
83 See < https://www.epa.gov/ghgemissions/natural-gas-and-petroleum-systems>.
3-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.84
As in previous years, EPA conducted early engagement and communication with stakeholders on updates prior to
public review. EPA held a stakeholder webinar in September of 2019 and a stakeholder workshop on greenhouse
gas data for oil and gas in November of 2019. 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
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. A key challenge in using these types of studies
to assess Inventory results is having a relevant basis for comparison (i.e., 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° x 0.1° spatial resolution, monthly temporal
resolution, and detailed scale-dependent error characterization.85 The gridded methane inventory is designed to
be consistent with the 2016 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2014 estimates for the
year 2012, which presents national totals.86
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 2019, EPA released draft memoranda that discussed changes under consideration, and
requested stakeholder feedback on those changes. EPA then created an updated version of the memoranda to
document the methodology implemented in the current Inventory.87 Memoranda cited in the Recalculations
Discussion below are: Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2018: Updates for Natural Gas
84	See .
85	See .
86	See .
87	Stakeholder materials including draft and final memoranda for the current (i.e., 1990-2018) Inventory are available at
.
Energy 3-91

-------
Gathering & Boosting Station Emissions (G&B Station memo) and Inventory of U.S. Greenhouse Gas Emissions and
Sinks 1990-2018: Updates for Offshore Production Emissions (Offshore Production memo).
EPA thoroughly evaluated relevant information available and made several updates to the Inventory, including:
using GHGRP, BOEM, and other data to calculate emissions from offshore production; and using GHGRP and
Zimmerle et al. 2019 study data to calculate gathering and boosting station emissions. In addition, certain sources
did not undergo methodological updates, but Cm and/or CO2 emissions changed by greater than 0.05 MMT CO2
Eq., comparing the previous estimate for 2017 to the current (recalculated) estimate for 2017 (the emissions
changes were mostly due to GHGRP data submission revisions). These sources are discussed below and include:
hydraulically fractured (HF) gas well completions; production segment pneumatic controllers; liquids unloading;
production segment storage tanks; HF and non-HF gas well workovers; and acid gas removal (AGR) vents, flares,
reciprocating compressors, and blowdowns at gas processing plants.
The combined impact of revisions to 2017 natural gas sector CH4 emissions, compared to the previous Inventory, is
a decrease from 165.6 to 139.3 MMT CO2 Eq. (26.3 MMT CO2 Eq., or 16 percent). The recalculations resulted in an
average decrease in CH4 emission estimates across the 1990 through 2017 time series, compared to the previous
Inventory, of 14.2 MMT CO2 Eq., or 8 percent.
The combined impact of revisions to 2017 natural gas sector CO2 emissions, compared to the previous Inventory, is
an increase from 26.3 MMT to 30.4 MMT, or 15 percent. The recalculations resulted in an average increase in
emission estimates across the 1990 through 2017 time series, compared to the previous Inventory, of 2.9 MMT
CO2 Eq., or 12 percent.
The combined impact of revisions to 2017 natural gas sector N2O emissions, compared to the previous Inventory, is
an increase from 4.7 kt CO2 Eq. to 8.1 kt CO2 Eq., or 70 percent. The recalculations resulted in an average increase
in emission estimates across the 1990 through 2017 time series, compared to the previous Inventory, of a factor of
2.5.
In Table 3-64 and Table 3-65 below are categories in Natural Gas Systems with recalculations resulting in a change
of greater than 0.05 MMT CO2 Eq., comparing the previous estimate for 2017 to the current (recalculated)
estimate for 2017. For more information, please see the Recalculations Discussion below.
Table 3-64: Recalculations of CO2 in Natural Gas Systems (MMT CO2)
Stage and Emission Source
Previous Estimate
Year 2017,
2019 Inventory
Current Estimate Year
2017,
2020 Inventory
Current Estimate
Year 2018,
2020 Inventory
Exploration
0.5
0.5
0.4
Production
2.8
6.5
9.6
Gathering Stations
0.2
4.3
7.0
Offshore Gas Production
0.4
+
+
Tanks
0.6
0.5
0.8
Processing
22.5
22.9
24.5
AGR Vents
16.7
17.2
17.5
Transmission and Storage
0.5
0.5
0.5
Distribution
+
+
+
Total
26.3
30.4
35.0
+ Does not exceed 0.05 MMT C02.
Table 3-65: Recalculations of CH4 in Natural Gas Systems (MMT CO2 Eq.)
Previous Estimate	Current Estimate Year	Current Estimate
Year 2017,	2017,	Year 2018,
Stage and Emission Source 2019 Inventory	2020 Inventory	2020 Inventory
Exploration 1.2	1.2	1.1
Production 108.4	82.3	80.9
3-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
G&B Stations
55.5
32.0
31.4
Offshore Gas Production
3.8
0.7
0.8
Non-HF Workovers
+
0.1
+
Pneumatic Controllers
26.4
26.6
25.4
Liquids Unloading
2.9
3.2
4.4
HF Workovers
0.8
0.8
0.6
Processing
11.7
11.5
12.2
Reciprocating Compressors
1.7
1.6
1.6
Flares
0.5
0.6
0.7
Blowdowns/Venting
0.9
0.7
1.1
Transmission and Storage
32.4
32.3
33.9
Distribution
11.9
11.9
11.8
Total
165.6
139.3
140.0
+ Does not exceed 0.05 MMT C02 Eq.
Exploration
There were no methodological updates to the exploration segment, but there were recalculations due to updated
data (e.g., GHGRP data for REC HF Completions with venting) that resulted in an average decrease in calculated
emissions over the time series from this segment of 0.3 MMT CO2 Eq. Cm (or 4 percent) and less than 0.05 MMT
CO2 (or 5 percent).
Production
Gathering and Boosting (G&B) Stations (Methodological Update)
EPA updated the G&B station methodology to use data from a Zimmerle et al. 2019 study. Zimmerle et al.
conducted Cm measurements at G&B stations, calculated CH4 EFs for certain equipment (compressors, tanks,
dehydrators, acid gas removal units, separators, and yard piping), and developed an approach to estimate national
activity data for G&B stations. EPA applied data from Zimmerle et al. and incorporated Subpart W data (for both
Cm and CO2) across the time series for the final methodology implemented in the Inventory. EPA did not retain
data from the previous methodology. EPA also applied the national average ratio of compressors per station and
the national-level scaling factor, both based on year 2017 data, from the Zimmerle et al. study. The G&B emissions
accounted for in the Inventory largely align with the G&B activities reported under Subpart W, because Subpart W
activity data were used to determine the national-level scaling factor. The G&B Station Memo provides details on
the methodology implemented into the final Inventory.
G&B station CH4 emission estimates decreased by an average of 36 percent in the current Inventory for the 1990
to 2017 time series, compared to the previous Inventory. The decrease in the CH4 emission estimate is due to
differences in the data between the current Inventory and previous Inventory. Calculated G&B station CH4
emission estimates decreased by an average of 36 percent in the current Inventory for each year in the 1990 to
2017 time series, compared to the previous Inventory. The decrease in the CH4 emission estimate is due to
differences in the input data between the current Inventory and the prior Inventory. The prior Inventory used data
from a Marchese et al. 2015 study to calculate CH4 emissions.88
Data were previously unavailable to quantify the largest sources of CO2 from G&B stations. By incorporating recent
Subpart W data on CO2 from flaring and acid gas removal units (previously not included in the Inventory), the
estimate of G&B station CO2 emission increased by a factor of 22 (from an average of 0.2 MMT CO2 to an average
of 3.5 MMT CO2) in the current Inventory for the 1990 to 2017 time series, compared to the previous Inventory.
88 Marchese, A. J. et al., Methane Emissions from United States Natural Gas Gathering and Processing. Environmental Science &
Technology, 49, 10718-10727. 2015.
Energy 3-93

-------
Feedback from three stakeholder comment letters supported the update to gathering and boosting. Of these
stakeholder comments, one also specifically supported the use of the Zimmerle et al. approach to developing the
national-level scaling factor to account for GHGRP non-reporters, and another suggested that the scaling factor
and national average ratio of compressors per station be updated annually in future Inventories if data are
available to do so.
One stakeholder comment letter did not support the update. The comment letter noted discrepancies found
between site-level and component-level emissions data in recent studies (citing work primarily focusing on the
onshore production segment. For comparison with an alternative national-level gathering and boosting estimate,
the letter references an estimate in Alvarez et al., which relied primarily on the Marchese et al. study (previous
Inventory data source), and the application of an adjustment factor of 10 percent. The comment letter
recommended retaining the previous (Marchese et al.) data source. In their paper, Zimmerle et al. discussed
differences between the Zimmerle et al. study (current data source) and the Marchese et al. Study (previous data
source). The differences noted in Zimmerle et al. are: (1) the Zimmerle et al. study uses an updated and likely more
representative mix of stations in terms of throughput and complexity, (2) the Zimmerle et al. study accessed
component level activity and emissions data from the GHGRP, which were not available at the time of the
Marchese et al. study, and which represented data from a large set of operators for the entire U.S., (3) the two
studies utilized different measurement methods, and (4) there may have been operational improvements to G&B
stations and/or construction of new lower-emitting stations during the intervening years between studies due to
increased attention to Cm emissions across the natural gas value chain.
The stakeholder comment letter that did not support the update to gathering and boosting also expressed concern
about the potential omission of "super-emitters." The Zimmerle et al. study detected a number of large emitters.
For example, the study noted that "For most leaker factors, 50% or more of emissions are due to the largest 5% of
emitters." The set of emission factors developed in the Zimmerle et al. study which were used to calculate
emissions in the GHG Inventory include estimates for all emissions detected in the field campaign, including
estimates for large emitters, and the study notes that these "Large emitter emissions have substantial impact on
major equipment emission factors, adding 70% - 83% to the impacted major equipment factors."
The stakeholder comment letter that did not support the update to gathering and boosting also sought additional
information justifying the use of the Zimmerle et al. (measurements conducted in 2017) and GHGRP (data available
starting in 2016) data across the time series as opposed to using data from Marchese et al. (measurements from
2013 and 2014) for previous years. EPA considered this approach but did not implement it in the Inventory due to
incongruencies between the studies noted in the previous paragraph. If the Marchese et al. study in emissions and
activity data were used for early years of the time series (e.g., 1990-2014) and the Zimmerle et al. and GHGRP data
were used in more recent years (e.g. 2016-2017), there would be a large decrease in emissions over a short period
of time due to this transition. Some fraction of the decrease would likely be attributable to improvements in
technologies and industry practices. However, as noted above there are other differences between the studies
such as study representativeness and the difference between the two is likely not entirely due to changes in
technologies (or any other single factor). For this reason, EPA did not implement an approach that uses data from
both of the studies in different parts of the time series.
Table 3-66: Gathering Stations National ChU Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Compressors
126,757
161,098
243,532
255,491
253,209
271,238
278,874
Tanks
135,802
172,593
260,910
273,723
271,278
205,261
180,945
Station Blowdowns
20,560
26,130
39,501
41,441
41,071
63,823
62,020
Dehydrator Vents -







Large units
29,975
38,096
57,590
60,419
59,879
51,668
48,401
Dehydrator Vents -







Small units
306
389
588
617
612
708
575
High-bleed Pneumatic







Devices
16,698
21,222
32,081
33,656
33,356
32,654
23,666
Intermittent Bleed







Pneumatic Devices
79,110
100,543
151,991
159,455
158,031
173,628
156,662
3-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Low-Bleed Pneumatic
Devices
2,835
3,603
5,446
5,714
5,663
6,344
5,722
Flare Stacks
5,300
6,736
10,183
10,683
10,588
9,394
13,935
AGRU
47
60
90
94
94
91
88
Pneumatic Pumps
15,844
20,137
30,441
31,936
31,651
23,391
24,878
Gas Engines
169,766
215,760
326,164
342,182
339,126
363,534
373,753
Dehydrator Leaks
851
1,081
1,634
1,715
1,699
1,852
1,882
Yard Piping
37,206
47,286
71,482
74,992
74,323
76,709
85,115
Separators
559
710
1,073
1,126
1,116
1,152
1,278
Desiccant Dehydrators
8
11
16
17
17
38
4
Total Emissions
641,624
815,454
1,232,724
1,293,262
1,281,711
1,281,484
1,257,799
Previous Estimate
1,051,775
1,217,024
2,063,775
2,163,417
2,143,324
2,218,773
NA
NA (Not Applicable)
Table 3-67: Gathering Stations National CO2 Emissions (metric tons CO2)
Source
1990
2005
2014
2015
2016
2017
2018
Compressors
15,277
19,416
29,351
30,793
30,517
32,690
33,611
Tanks
420,699
534,676
808,271
847,965
840,391
633,931
1,294,821
Station Blowdowns
1,587
2,017
3,049
3,199
3,170
4,923
9,572
Dehydrator Vents -







Large units
369,890
470,102
710,654
745,554
738,894
763,329
796,516
Dehydrator Vents-







Small units
332
422
638
669
663
1,266
4,860
High-bleed Pneumatic







Devices
1,143
1,452
2,195
2,303
2,282
2,120
1,714
Intermittent Bleed







Pneumatic Devices
5,240
6,659
10,067
10,561
10,467
13,172
13,066
Low-Bleed Pneumatic







Devices
213
271
409
429
425
399
410
Flare Stacks
1,354,751
1,721,783
2,602,824
2,730,646
2,706,255
2,300,171
4,205,760
AGRU
246,880
313,765
474,319
497,612
493,167
527,835
643,969
Pneumatic Pumps
963
1,224
1,850
1,941
1,924
1,683
1,679
Dehydrator Leaks
103
130
197
207
205
223
227
Yard Piping
4,484
5,699
8,615
9,038
8,958
9,245
10,258
Separators
67
86
129
136
135
139
154
Desiccant Dehydrators
+
+
+
+
+
+
+
Total Emissions
2,421,629
3,077,701
4,652,569
4,881,053
4,837,454
4,291,126
7,016,615
Previous Estimate
93,791
143,218
221,279
233,320
232,491
239,459
NA
NA (Not Applicable)
+ Less than 0.5 metric tons
Offshore Gas Production (Methodological Update)
EPA updated the offshore production methodology to estimate emissions for all offshore producing regions and to
use activity data sources that provide a full time series of data. The previous Inventory only estimated emissions
for offshore facilities in federal waters of the Gulf of Mexico (GOM); these facilities are under Bureau of Ocean
Energy Management (BOEM) jurisdiction and BOEM estimates their greenhouse gas emissions triennially via the
Gulfwide Emissions Inventory (GEI). The previous Inventory also relied on activity data sources that were no longer
updated, and surrogate activity data from 2008 and 2010 had been used to estimate emissions in more recent
years. The updated Inventory methodology now includes emissions estimates for offshore facilities in federal and
state waters of the GOM and offshore facilities off the coast of Alaska.
The updated Inventory methodology for each region is presented here. EPA calculated vent and leak EFs for
offshore facilities in GOM federal waters for major complexes and minor complexes using BOEM GEI emissions
data from the 2005, 2008, 2011, 2014, and 2017 GEIs. Vent and leak EFs were calculated for 11 emission sources
(cold vents, fugitives, pneumatic pumps, losses from flashing, pneumatic controllers, combustion, glycol
Energy 3-95

-------
dehydrators, storage tanks, mud degassing, minor surrogates, and amine gas sweetening units). These EFs were
paired with active offshore complex counts over the time series. EPA calculated GOM federal waters flaring
emissions using flaring volumes reported in Oil and Gas Operations Reports (OGOR), Part B (OGOR-B). OGOR-B
flaring volumes are available over the time series but assumptions were necessary to assign the volumes to
offshore gas production versus offshore oil production for 1990 to 2010. The previous Inventory allocated all GOM
federal waters flaring emissions to offshore gas production facilities. EPA calculated production based EFs for
offshore facilities in GOM state waters using the resulting GOM federal waters emissions and gas production in
each year. EPA also calculated production based EFs for offshore facilities in the Alaska region, and the EFs for
these regions were derived from GHGRP data. EPA multiplied the production based EFs by the region-specific
offshore production (i.e., GOM state waters production, and Alaska production) in a given year. The Offshore
Production memo provides details for the methodology update under consideration and that was implemented in
the Inventory.
Due to this recalculation, annual offshore gas production CFU emission estimates decreased in the current
Inventory for 1990 to 2017 by an average of 14 percent, compared to the previous Inventory. The impacts varied
across the time series with estimates in earlier years of the time series increasing (e.g., by an average of 19 percent
from 1990 to 2002) and estimates in more recent years of the time series decreasing (e.g., by an average of 73
percent from 2010 to 2017). The increase in offshore gas production Cm emission estimates from 1990 to 2002 is
due to the inclusion of emissions from facilities located in GOM state waters and the Alaska region. Examining the
same 1990 through 2002 period, there is not a significant difference between offshore gas production Cm
emission estimates in GOM federal waters between the current Inventory and previous Inventory, with an average
increase of only 4 percent.
The noticeable decrease in offshore gas production Cm emission estimates over the 2010 to 2017 time period is
due to a decrease in GOM federal waters emission estimates. The main factor that leads to a decrease in the
estimate of offshore gas production Cm emissions for GOM federal waters facilities is the use of updated activity
data. Activity data in the previous Inventory were last available for 2010, and the 2010 counts are applied as
surrogate to all following years. The updated methodology for the current Inventory uses a continuously updated
BOEM data source, and it shows a noticeable decrease in offshore facilities starting in 2008 that is not captured in
the previous Inventory's data.
For comparison, total offshore production (for oil and gas combined) Cm emissions for facilities in GOM federal
waters are provided here for years 2011, 2014, and 2017 from the GEI, previous Inventory, and current Inventory.
For offshore facilities in GOM federal waters in year 2011, GEI Cm emissions equaled 246 kt, previous Inventory
Cm emissions equaled 338 kt, and current Inventory CH4 emissions equal 278 kt. For offshore facilities in GOM
federal waters in year 2014, GEI CFU emissions equaled 205 kt, previous Inventory CH4 emissions equaled 338 kt,
and current Inventory CFU emissions equal 225 kt. For offshore facilities in GOM federal waters in year 2017, GEI
Cm emissions equaled 170 kt, previous Inventory CH4 emissions equaled 338 kt, and current Inventory CH4
emissions equal 206 kt.
Annual offshore gas production CO2 emission estimates decreased in the current Inventory for 1990 to 2017 by an
average of 71 percent, compared to the previous Inventory. This change is largely because all GOM federal waters
flaring emissions in the previous Inventory were allocated to offshore gas production, whereas the current
Inventory estimates GOM federal waters flaring emissions for both offshore gas and oil production, and a
significant portion of the CO2 is from offshore oil production.
EPA received feedback on this update through its September 2019 memo and through the public review draft of
the Inventory. Two stakeholders supported the update to activity data. A stakeholder suggested clarifications on
the calculation of emission factors, and noted upcoming data that may be used to assess offshore emission factors.
A stakeholder suggested clarification on the development of activity counts and supported considering a different
approach which would use source-specific emission factors. As noted above, the emissions estimates were
calculated using complex-level factors for offshore operations in GOM federal waters, and using production-based
emission factors for offshore operations in state waters. An estimate of emissions source-level emissions was
developed using the fraction of emissions in each category in the GOM federal waters data set, applied to GOM
federal and state water total emission estimates, and using the fraction of emissions in each category in GHGRP for
3-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Alaska offshore production, and applied to the total estimates for Alaska offshore production. The emission
source-level estimates are available in the annex. The stakeholder noted that the use of emission factors
calculated from data from the from the GHGRP reporting population (those emitting over the GHGRP threshold),
applied to all Alaska offshore production could skew regional emission estimates. The stakeholder also supported
the use of GEI data as opposed to OGOR-B data to calculate emissions from flaring. The emissions estimates were
calculated using OGOR-B. GEI data is currently available for the years 2005, 2008, 2011, 2014, and 2017. The
OGOR-B dataset can be used to calculate flaring emissions for the full 1990 to 2018 time series.
The recalculation results in a change in the trend, in methane in particular where the 1990 to 2017 trend in this
Inventory is a decrease of 83 percent, versus an increase of 7 percent in the previous Inventory. The stakeholder
provided several factors supporting this decreasing trend: more stringent limitations imposed by BSEE (Bureau of
Safety and Environmental Enforcement) related to venting and flaring, increased utilization of VRU equipment, and
replacement of older platforms with newer ones that include state of the art technology.
Table 3-68: Offshore Gas Production National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
GOM Federal Waters
153,457
60,823
21,847
17,890
23,752
19,563
20,151
GOM State Waters
9,296
10,790
9,110
5,836
8,241
7,995
12,373
Alaska State Waters
1,892
1,498
453
329
591
501
618
Total Emissions
164,645
73,111
31,410
24,055
32,585
28,060
33,141
Previous Estimate
140,949
173,459
150,565
150,565
150,565
150,565
NA
NA (Not Applicable)







ible 3-69: Offshore Gas Production National Emissions (metric tons CO2)


Source
1990
2005
2014
2015
2016
2017
2018
GOM Federal Waters
47,315
36,319
50,740
36,180
30,086
24,564
24,233
GOM State Waters
2,866
6,443
21,159
11,802
10,439
10,039
14,880
Alaska State Waters
19,825
15,695
4,745
3,448
2,563
3,483
1,877
Total Emissions
70,006
58,457
76,644
51,430
43,088
38,085
40,989
Previous Estimate
232,959
183,731
367,861
370,479
371,788
372,116
NA
NA (Not Applicable)
Pneumatic Controllers (Recalculation with Updated Data)
Pneumatic controller Cm emission estimates increased in the current Inventory by an average of 0.3 percent
across the time series, compared to the previous Inventory due to GHGRP submission revisions and Enverus
Drillinglnfo data revisions.
Table 3-70: Production Segment Pneumatic Controller National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Low Bleed
NO
23,168
32,486
31,784
31,790
34,639
33,867
High Bleed
296,948
463,604
130,339
101,509
104,353
108,294
87,372
Intermittent Bleed
193,647
536,998
931,781
939,438
900,993
919,154
895,118
Total Emissions
490,594
1,023,770
1,094,606
1,072,732
1,037,136
1,062,086
1,016,357
Previous Estimate
492,254
1,016,763
1,089,339
1,075,601
1,064,069
1,057,303
NA
NO (Not Occurring)
NA (Not Applicable)
Liquids Unloading (Recalculation with Updated Data)
Liquids unloading Cm emission estimates increased for 2017 by 11 percent in the current Inventory, compared to
the previous Inventory. Compared to the previous Inventory, on average across the time series, liquids unloading
Cm emission estimates increased by less than 0.1 percent. These changes were due to GHGRP submission
revisions.
Energy 3-97

-------
Table 3-71: Liquids Unloading National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Unloading with Plunger Lifts
NO
125,582
80,402
62,836
59,787
58,617
78,069
Unloading without Plunger Lifts
371,391
247,032
129,520
97,225
67,876
71,173
99,229
Total Emissions
371,391
372,614
209,921
160,061
127,663
129,790
177,298
Previous Estimated Emissions
372,325
373,442
210,784
160,706
130,778
117,379
NA
NO (Not Occurring)
NA (Not Applicable)
Tanks (Recalculation with Updated Data)
Production tank CO2 emission estimates decreased for 2017 by 10 percent in the current Inventory, compared to
the previous Inventory. Compared to the previous Inventory, on average across the time series, tank CO2 emission
estimates decreased by 1 percent. These changes were due to GHGRP submission revisions.
Table 3-72: Production Segment Storage Tanks National Emissions (metric tons CO2)
Source
1990
2005
2014
2015
2016
2017
2018
Large Tanks w/Flares
287,644
363,030
1,028,597
1,039,129
1,080,439
500,450
783,932
Large Tanks w/VRU
NO
760
2,782
2,811
2,434
44
58
Large Tanks w/o Control
164,501
88,897
153,447
155,018
902
219
7,235
Small Tanks w/Flares
NO
7,839
28,710
29,004
28,894
20,816
44,530
Small Tanks w/o Flares
5,638
4,300
9,850
9,950
12,388
4,090
8,943
Malfunctioning Separator
6
6
15
15
11
468
224
Dump Valves


Total Emissions
457,788
464,831
1,223,400
1,235,927
1,125,067
526,086
844,923
Previous Estimate
459,592
466,429
1,227,366
1,239,933
1,128,990
585,339
NA
NO (Not Occurring)
NA (Not Applicable)
HF Gas Well Workovers (Recalculation with Updated Data)
Recalculations of HF gas well workover Cm emissions resulted in an average decrease of 4 percent across the 1990
to 2017 time series when comparing the current Inventory to the previous Inventory. These changes were due to
GHGRP submission revisions.
Table 3-73: HF Gas Well Workovers National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
HF Workovers - Non-REC with
Venting
HF Workovers - Non-REC with
Flaring
25,774
60,903
24,642
1,752
7,530
10,263
2,393
365
953
460
80
72
509
799
HF Workovers - REC with Venting
NO
576
569
8,685
6,312
17,005
21,181
HF Workovers - REC with Flaring
NO
4
25
1,658
1,240
3,708
50
Total Emissions
26,139
62,437
25,695
12,175
15,155
31,485
24,422
Previous Estimate
26,188
67,717
26,608
13,161
15,551
33,711
NA
NO (Not Occurring)
NA (Not Applicable)
Non-HF Gas Well Workovers (Recalculation with Updated Data)
Recalculations of non-HF gas well workover emissions resulted in a 484 percent increase in 2017 CH4 estimates and
an average increase of 4 percent across the 1990 to 2016 time series when comparing the current Inventory to the
previous Inventory. The large increase for HF gas well workover emissions in 2017 results from GHGRP submission
revisions.
3-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-74: Non-HF Gas Well Workovers National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Non-HF Workovers - vented
484
667
443
532
539
3,484
342
Non-HF Workovers - flared
0
21
2
25
1
0
0
Total Emissions
484
688
444
557
540
3,484
343
Previous Estimate
486
634
427
537
523
597
NA
NA (Not Applicable)
Well Counts (Recalculation with Updated Data)
For total national well counts, EPA has used a more recent version of the Enverus Drillinglnfo data set (Enverus
Drillinglnfo 2019) to update well counts data in the Inventory. While this is not a significant recalculation (increases
are less than 1 percent across the time series), is the well count dataset is a key input to the Inventory, and results
are highlighted here.
Table 3-75: Producing Gas Well Count Data
Activity
1990
2005
2014
2015
2016
2017
2018
Number of Gas Wells
Previous Estimate
193,232
193,718
;? 346,484
.. 346,862
422,701
424,308
419,692
420,418
419,346
419,005
412,601
411,450
405,026
NA
NA (Not Applicable)
In December 2019, EIA released an updated time series of national oil and gas well counts (covering 2000 through
2018). EIA estimates 982,371 total producing wells for year 2018. EPA's total well count for this year is 969,212.
EPA's well counts in recent time series years are generally 1 percent lower than ElA's. ElA's well counts include side
tracks, 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 and a higher fraction
of gas wells than EPA.
Processing
Acid Gas Removal (Recalculation with Updated Data)
Acid gas removal unit (AGR) CO2 emission estimates for 2016 and 2017 increased on average by 2 percent,
comparing the current Inventory to the previous Inventory, due to GHGRP submission revisions, where a higher
emission factor was calculated from the GHGRP data. The emission estimates were essentially unchanged across
the 1990 to 2015 time series, comparing the current Inventory to the previous Inventory, with an average increase
of 0.1 percent.
Table 3-76: AGR National CO2 Emissions (kt CO2)
Source
1990
2005
2014
2015
2016
2017
2018
Acid Gas Removal
Previous Estimate
28,282
28,282
15,339
15,320
14,979
14,946
14,979
14,946
16,679
16,481
17,182
16,728
17,451
NA
NA (Not Applicable)
Flares (Recalculation with Updated Data)
Processing segment flare Cm emission estimates decreased by 4 percent across the 2011 to 2017 time series in the
current Inventory. Prior to 2011, flare-specific Cm emissions were not estimated. Instead, plant-wide emissions
were calculated for years prior to 2011. Processing segment flare Cm emission estimates increased by
approximately 15 percent for 2017 in the current Inventory, compared to the previous Inventory. This increase in
Energy 3-99

-------
CH4 emission estimates for 2017 is due to GHGRP submission revisions, where a higher emission factor was
calculated from the GHGRP data.
Table 3-77: Processing Segment Flares National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Flares
NO
NE
19,509
19,509
19,988
24,277
26,146
Previous Estimate
NO
NE
21,171
21,171
21,049
21,049
NA
NO (Not Occurring)
NA (Not Applicable)
NE (Not estimated)
Reciprocating Compressors (Recalculation with Updated Data)
Reciprocating compressor CH4 emission estimates decreased by 1 percent on average for 2011 to 2017 in the
current Inventory and decreased by 5 percent for 2017 in the current Inventory, compared to the previous
Inventory. This decrease in the CH4 emission estimate for 2017 is due to GHGRP submission revisions, where a
lower EF (mt CHVreciprocating compressor) was calculated from the GHGRP data.
Table 3-78: Processing Segment Reciprocating Compressors National Emissions (metric tons
cm)
Source
1990
2005
2014
2015
2016
2017
2018
Reciprocating Compressors
Previous Estimate
324,939
324,939
NA
NA
67,982
68,408
67,982
68,408
63,682
63,351
64,955
68,178
62,574
NA
NA (Not Applicable)
Blowdowns/Venting (Recalculation with Updated Data)
Blowdowns and venting CH4 emission estimates decreased by 2 percent across the 1990 to 2017 time series in the
current Inventory and decreased by 20 percent for 2016 and 2017 in the current Inventory, compared to the
previous Inventory. This decrease in CH4 emissions for 2016 and 2017 is due to GHGRP submission revisions, where
a lower emission factor (CH4 from blowdowns/venting per plant) was calculated from the GHGRP data.
Table 3-79: Processing Segment Blowdowns/Venting National Emissions (metric tons ChU)
Source
1990
2005
2014
2015
2016
2017
2018
Blowdowns/Venting
Previous Estimate
59,507
59,507
34,234
34,264
34,890
34,943
34,890
34,943
28,447
36,428
29,061
36,266
45,499
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 0.04 MMT CO2 Eq. CH4 (or 0.7 percent) and less than 0.04 MMT CO2 (or 18 percent).
Distribution
There were no methodological updates to the distribution segment, and recalculations due to updated data
resulted in average increases in calculated CH4 and CO2 emissions over the time series of 0.01 percent.
Planned Improvements
EPA seeks stakeholder feedback on the improvements noted below for future Inventories.
3-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Gathering and Boosting Stations
EPA updated the G&B station methodology for the Inventory, incorporating the Zimmerle et al. 2019 study and
Subpart W data. Comments on the public review draft of the Inventory suggested continuing to confirm and
update the scaling factor applied to calculate national emissions. EPA plans to periodically reassess this factor. See
the G&B Station memo for details on the updates under consideration and specific requests for stakeholder
feedback.
Offshore Production
EPA updated the offshore production methodology for the Inventory, incorporating data from BOEM and GHGRP.
Detailed information and considerations for various approaches considered for the methodology update were
provided in a memorandum and discussed at a stakeholder workshop and webinar. Through the stakeholder
process and the public review period, stakeholders provided feedback on additional approaches or data sets that
could be used. In future inventories, EPA will consider alternate approaches or data sources, such as additional use
of BOEM data or data from upcoming studies. Stakeholders identified upcoming studies of offshore oil and gas
platform emissions that will include evaluation of different inventory estimates and methods.
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 malfunction and control efficiency data.
•	Activity data and emissions data for production facilities that do not report to GHGRP.
•	Natural gas leaks at point of use estimates.
•	Anomalous leak events, such as a 2018 well blowout in Ohio.
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:
•	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.
•	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 is around 3.2 million (with around 2.6 million abandoned oil wells and 0.6
million abandoned gas wells). Abandoned wells emit both Cm and CO2. Wells that are plugged have much lower
Energy 3-101

-------
average emissions than wells that are unplugged (less than 1 kg Cm per well per year, versus over 100 kg Cm per
well per year). Around a third 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 227 kt Cm and 5 kt CO2 in 2018. Emissions of both gases
decreased by 1 percent from 1990, while the total population of abandoned oil wells increased 27 percent.
Emissions of both gases decreased by less than 1 percent between 2017 and 2018 as a result of well plugging
activities.
Abandoned gas wells. Abandoned gas wells emitted 54 kt Cm and 2 kt CO2 in 2018. Emissions of both gases
increased by 50 percent from 1990, as the total population of abandoned gas wells increased 79 percent.
Emissions of both gases decreased by less than 1 percent between 2017 and 2018 as a result of well plugging
activities.
Table 3-80: ChU Emissions from Abandoned Oil and Gas Wells (MMT CO2 Eq.)
Activity
1990
2005
2014
2015
2016
2017
2018
Abandoned Oil Wells
5.7
5.9
5.8
5.8
5.8
5.7
5.7
Abandoned Gas Wells
0.9
1.1
1.3
1.3
1.4
1.4
1.4
Total
6.6
6.9
7.1
7.1
7.2
7.1
7.0
Note: Totals may not sum due to independent rounding.
Table 3-81: ChU Emissions from Abandoned Oil and Gas Wells (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
Abandoned Oil Wells
227
236
232
233
234
227
227
Abandoned Gas Wells
36
43
52
53
55
55
54
Total
263
278
284
286
289
282
281
Note: Totals may not sum due to independent rounding.
Table 3-82: CO2 Emissions from Abandoned Oil and Gas Wells (MMT CO2)
Activity 1990
2005
2014
2015
2016
2017
2018
Abandoned Oil Wells +
+
+
+
+
+
+
Abandoned Gas Wells +
+
+
+
+
+
+
Total +
+
+
+
+
+
+
+ Does not exceed 0.05 MMT C02.






able 3-83: CO2 Emissions from Abandoned Oil and Gas Wells (kt)



Activity 1990
2005
2014
2015
2016
2017
2018
Abandoned Oil Wells 5
5
5
5
5
5
5
Abandoned Gas Wells 2
2
2
2
2
2
2
Total 6
7
7
7
7
7
7
Note: Totals may not sum due to independent rounding.
Methodology
EPA developed abandoned well Cm emission factors using data from Kang et al. (2016) and Townsend-Small et al.
(2016). Plugged and unplugged abandoned well Cm 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
3-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 emission factors using the CFU emission factors and an assumed ratio of CO2-
to-Cm gas content, similar to the approach used to calculate CO2 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 CO2/MT CH4, which was derived through API TankCalc modeling runs. For
abandoned gas wells, EPA used the Natural Gas Systems default production segment CFU and CO2 gas content
values (GRI/EPA 1996, GTI 2001) to develop a ratio of 0.044 MT CO2/MT CH4.
The total population of abandoned wells over the time series was estimated using historical data and Drillinglnfo
data. For the most recent year of the Inventory time series (year 2018), the prior year total counts are used as
surrogate data, as the Drillinglnfo query approach for the most recent year would likely overestimate abandoned
well counts, because many wells might be spud and not reporting production—not because they are
dry/abandoned, but due to the time required for completion. The abandoned well population was then split into
plugged and unplugged wells by assuming that all abandoned wells were unplugged in 1950, using year-specific
Drilling info data to calculate the fraction of abandoned wells plugged in 2016 (31 percent) and 2017 and 2018 (34
percent in both years), 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.89
Abandoned Oil Wells
Table 3-84: Abandoned Oil Wells Activity Data, ChU and CO2 Emissions (metric tons)
Source
1990
2005
2014
2015
2016
2017
2018
Plugged abandoned oil wells
387,506
617,887
759,781
780,434
801,199
882,850
889,068
Unplugged abandoned oil







wells
1,688,445
1,789,493
1,784,161
1,792,458
1,800,130
1,750,802
1,744,585
Total Abandoned Oil Wells
2,075,950
2,407,380
2,543,943
2,572,893
2,601,329
2,633,652
2,633,652
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)
318
477
564
577
592
652
657
CH4from unplugged







abandoned oil wells (MT)
226,740
235,212
231,461
232,197
233,191
226,801
225,995
Total ChUfrom Abandoned







oil wells (MT)
227,058
235,688
232,025
232,773
233,782
227,453
226,652
Total C02 from Abandoned







oil wells (MT)
4,607
4,782
4,708
4,723
4,744
4,615
4,599
89 See .
Energy 3-103

-------
Abandoned Gas Wells
Table 3-85: Abandoned Gas Wells Activity Data, ChU and CO2 Emissions (metric tons)
Source
1990
2005
2014
2015
2016
2017
2018
Plugged abandoned gas wells
60,126
104,652
154,844
162,215
171,979
193,375
194,736
Unplugged abandoned gas







wells
261,982
303,089
363,614
372,566
386,402
383,486
382,124
Total Abandoned Gas Wells
322,108
407,741
518,458
534,781
558,381
576,861
576,861
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 (MT)
53
97
147
155
164
185
186
CH4from unplugged







abandoned gas wells (MT)
36,199
42,582
51,591
52,919
54,884
54,470
54,276
Total CH4 from abandoned







gas wells (MT)
36,253
42,679
51,738
53,074
55,048
54,654
54,462
Total C02 from abandoned







gas wells (MT)
1,589
1,870
2,268
2,326
2,413
2,395
2,387
Uncertainty and Time-Seri insistency
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 2018, then applied the calculated bounds to both CH4 and CO2 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 CFU 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 3-86 provide the 95 percent confidence bound within which actual emissions
from abandoned oil and gas wells are likely to fall for the year 2018, using the recommended IPCC methodology.
Abandoned oil well CFU emissions in 2018 were estimated to be between 1.0 and 18.1 MMT CO2 Eq., while
abandoned gas well CFU emissions were estimated to be between 0.2 and 4.3 MMT CO2 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.
3-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-86: Approach 2 Quantitative Uncertainty Estimates for ChU and CO2 Emissions from
Petroleum and Natural Gas Systems (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.)b
(MMT C02
Eq.)

(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Abandoned Oil Wells
ch4
5.7
1.0
18.1
-83%
+219%
Abandoned Gas Wells
ch4
1.4
0.2
4.3
-83%
+219%
Abandoned Oil Wells
co2
0.005
0.001
0.015
-83%
+219%
Abandoned Gas Wells
co2
0.002
0.0004
0.008
-83%
+219%
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 2018.
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
2018 by summing an estimate of total abandoned wells not included in recent databases, to an annual estimate of
abandoned wells in the Enverus Drillinglnfo data set (with year 2017 estimates used as surrogates for year 2018
data). 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 Drilling info data to calculate the fraction of
abandoned wells plugged in 2016 through 2018, 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.
j?h1 jtion 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 a
stakeholder webinar on greenhouse gas data for oil and gas in September of 2019, and a workshop in November of
2019.
Recalculations Discussion
The counts of national abandoned wells were recalculated across the time series to use the latest Drillinglnfo data,
which resulted in minor changes to the total abandoned well population and the allocation between petroleum
and natural gas systems. The minor changes resulted from changes to the year-specific data for 1990 to 2017 as
processed from Drillinglnfo, which led EPA to recalculate the estimate of historical wells not included in the
Drillinglnfo data set (which decreased from 1,108,648 to 1,075,849 historical wells not included in Drillinglnfo).
Compared with the previous Inventory, counts of abandoned oil and gas wells are on average 0.3 percent and 0.8
percent, respectively, higher over 1990 to 2017. The impact was largest in recent years, with abandoned oil and
gas well counts recalculated to be 1.4 percent and 3.1 percent, respectively, higher for 2017 comparing the
previous Inventory values to the current Inventory values; this change is primarily due to the use of year-specific
data for year 2017 (as the previous Inventory used year 2016 estimates as surrogate for year 2017 per the
established methodology described above).
Energy 3-105

-------
Planned Improvements
The abandoned wells source was added to the Inventory in 2018. 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.
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 UNFCCC90 request that information be provided on precursor
greenhouse gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-Cm volatile organic
compounds (NMVOCs), and sulfur dioxide (SO2). 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 SO2, 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. Total emissions of NOx, CO, and NMVOCs from energy-related activities from 1990 to 2018 are reported in
Table 3-87. Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and Annex 6.3.
Table 3-87: NOx, CO, and NMVOC Emissions from Energy-Related Activities (kt)
Gas/Activity
1990
2005
2014
2015
2016
2017
2018
NOx
21,106
16,602
10,198
9,523
9,037
8,555
8,154
Mobile Fossil Fuel Combustion
10,862
10,295
6,138
5,740
5,413
5,051
4,689
Stationary Fossil Fuel Combustion
10,023
5,858
3,313
3,036
2,876
2,757
2,719
Oil and Gas Activities
139
321
650
650
650
650
650
Waste Combustion
82
128
97
97
97
97
97
International Bunker Fuelsa
1,956
1,704
1,211
1,363
1,470
1,481
1,462
CO
125,640
64,985
40,234
39,258
36,885
35,211
33,537
Mobile Fossil Fuel Combustion
119,360
58,615
34,135
33,159
30,786
29,112
27,438
Stationary Fossil Fuel Combustion
5,000
4,648
3,686
3,686
3,686
3,686
3,686
Waste Combustion
978
1,403
1,776
1,776
1,776
1,776
1,776
Oil and Gas Activities
302
318
637
637
637
637
637
International Bunker Fuelsa
103
133
137
144
150
156
160
NMVOCs
12,620
7,191
7,247
7,082
6,835
6,629
6,423
Mobile Fossil Fuel Combustion
10,932
5,724
3,754
3,589
3,342
3,137
2,931
Oil and Gas Activities
554
510
2,853
2,853
2,853
2,853
2,853
Stationary Fossil Fuel Combustion
912
716
497
497
497
497
497
Waste Combustion
222
241
143
143
143
143
143
International Bunker Fuelsa
57
54
42
47
50
51
51
Note: Totals may not sum due to independent rounding.
a These values are presented for informational purposes only and are not included in totals.
90 See .
3-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Methodology
Emission estimates for 1990 through 2018 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2019), 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 2018. Details on the emission trends through time are described in more detail in the Methodology
section, above.
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.91 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).92
Two transport modes are addressed under the IPCC definition of international bunker fuels: aviation and marine.93
Greenhouse gases emitted from the combustion of international bunker fuels, like other fossil fuels, include CO2,
Cm and N2O for marine transport modes, and CO2 and N2O 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.
91	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).
92	Note that the definition of international bunker fuels used by the UNFCCC differs from that used by the International Civil
Aviation Organization.
93	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).
Energy 3-107

-------
The 2006IPCC 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.94
Emissions of CChfrom 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 Cm is emitted by modern engines (Anderson et al. 2011), and as a result, Cm
emissions from this category are reported as zero. In jet engines, N2O 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
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 2018 from the combustion of international bunker fuels from both
aviation and marine activities were 123.3 MMT CO2 Eq., or 18 percent above emissions in 1990 (see Table 3-88 and
Table 3-89). Emissions from international flights and international shipping voyages departing from the United
States have increased by 112.4 percent and decreased by 36.9 percent, respectively, since 1990. The majority of
these emissions were in the form of CO2; however, small amounts of Cm (from marine transport modes) and N2O
were also emitted.
Table 3-88: CO2, ChU, and N2O Emissions from International Bunker Fuels (MMT CO2 Eq.)
Gas/Mode
1990
2005
2014
2015
2016
2017
2018
CO?
103.5
113.1
103.4
110.9
116.6
120.1
122.1
Aviation
38.0
60.1
69.6
71.9
74.1
77.7
80.8
Commercial
30.0
55.6
66.3
68.6
70.8
74.5
77.7
Military
8.1
4.5
3.3
3.3
3.3
3.2
3.1
Marine
65.4
53.0
33.8
38.9
42.5
42.4
41.3
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
0.9
1.0
1.0
1.1
1.1
Aviation
0.4
0.6
0.7
0.7
0.7
0.7
0.8
Marine
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Total
104.5
114.2
104.4
112.0
117.7
121.3
123.3
Notes: Totals may not sum due to independent rounding. Includes aircraft cruise altitude emissions.
a CH4 emissions from aviation are estimated to be zero.
Table 3-89: CO2, ChU, and N2O Emissions from International Bunker Fuels (kt)
Gas/Mode	1990	2005	2014	2015	2016	2017	2018
94 Naphtha-type jet fuel was used in the past by the military in turbojet and turboprop aircraft engines.
3-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
co2
103,463
113,139
103,400
110,887
116,594
120,107
122,088
Aviation
38,034
60,125
69,609
71,942
74,059
77,696
80,788
Marine
65,429
53,014
33,791
38,946
42,535
42,412
41,300
ch4
7
5
3
4
4
4
4
Aviation3
0
0
0
0
0
0
0
Marine
7
5
3
4
4
4
4
n2o
3
3
3
3
3
4
4
Aviation
1
2
2
2
2
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 CO2 were estimated by applying C content and fraction oxidized factors to fuel consumption activity
data. This approach is analogous to that described under Section 3.1 - CO2 from Fossil Fuel Combustion. Carbon
content and fraction oxidized factors for jet fuel, distillate fuel oil, and residual fuel oil were taken directly from EIA
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 (2019) and USAF (1998), and heat content for jet fuel was taken from EIA (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 CFU and N2O 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 N2O emissions were
obtained from the 2006IPCC 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 N2O (IPCC 2006). For marine vessels
consuming either distillate diesel or residual fuel oil the following values (g/MJ), were employed: 0.32 for CFU and
0.08 for N2O. 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 2018 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 CO2 estimates for 1990 and 2000 through 2018 were obtained directly from FAA's AEDT
model (FAA 2019). The radar-informed method that was used to estimate CO2 emissions for commercial aircraft
for 1990 and 2000 through 2018 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 CO2
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
Energy 3-109

-------
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 Under Secretary 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 2019). 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 3-90 See Annex 3.8 for additional discussion of military data.
Table 3-90: Aviation Jet Fuel Consumption for International Transport (Million Gallons)
U.S. and Foreign Carriers
3,222
5,983
7,126
7,383
7,610
8,011
8,352
U.S. Military
862
462
339
341
333
326
315
Total
4,084
6,445
7,465
7,725
7,943
8,338
8,667
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 2019) for 1990 through 2001, 2007 through 2018, 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 (2019). 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 3-91.
Table 3-91: Marine Fuel Consumption for International Transport (Million Gallons)
Fuel Type
1990
2005
2014
2015
2016
2017
2018
Residual Fuel Oil
4,781
3,881
2,466
2,718
3,011
2,975
2,790
Distillate Diesel Fuel & Other
617
444
261
492
534
568
684
U.S. Military Naval Fuels
522
471
331
326
314
307
285
Total
5,920
4,796
3,058
3,536
3,858
3,850
3,759
Note: Totals may not sum due to independent rounding.
Uncertainty and lime-Serfi insistency
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.95 For example, smaller aircraft on shorter routes often carry sufficient
95 See uncertainty discussions under section 3.1 Carbon Dioxide Emissions from Fossil Fuel Combustion.
3-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 3-lllortland3-lll, 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 3-lllortland3-lll from ports with
low fuel costs.
Uncertainties exist with regard to the total fuel used by military aircraft and ships, and in the activity data on
military operations and training that were used to estimate percentages of total fuel use reported as bunker fuel
emissions. Total aircraft and ship fuel use estimates were developed from DoD records, which document fuel sold
to the Navy and Air Force from the Defense Logistics Agency. These data may slightly over or under estimate actual
total fuel use in aircraft and ships because each Service may have procured fuel from, and/or may have sold to,
traded with, and/or given fuel to other ships, aircraft, governments, or other entities. There are uncertainties in
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 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 2001.
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, department and military service 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 additional data
collection.
Although aggregate fuel consumption data have been used to estimate emissions from aviation, the recommended
method for estimating emissions of gases other than CO2 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 CO2.96
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 2018. Details on the emission trends through time are described in more detail in the Methodology
section, above.
96 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 3-111

-------
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
CO2, CH4, and N2O 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.
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 CO2 in addition to CH4 and N2O already covered in this chapter. In line
with the reporting requirements for inventories submitted under the UNFCCC, CO2 emissions from biomass
combustion have been estimated separately from fossil fuel CO2 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 I 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 CO2 emissions to U.S. totals
within the Land Use, Land-Use Change, and Forestry sector's approach.
Therefore, CO2 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 2018, total CO2 emissions from the burning of woody biomass in the industrial, residential, commercial, and
electric power sectors were approximately 229.1 MMT CO2 Eq. (229,085 kt) (see Table 3-92 and Table 3-93). As the
largest consumer of woody biomass, the industrial sector was responsible for 63.0 percent of the CO2 emissions
from this source. The residential sector was the second largest emitter, constituting 23.3 percent of the total, while
the commercial and electric power sectors accounted for the remainder.
Table 3-92: CO2 Emissions from Wood Consumption by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Industrial
135.3
136.3
140.3
138.5
138.3
144.5
144.3
Residential
59.8
44.3
59.7
52.9
46.2
44.6
53.3
Commercial
6.8
7.2
7.9
8.2
8.6
8.6
8.7
Electric Power
13.3
19.1
25.9
25.1
23.1
23.6
22.8
Total
215.2
206.9
233.8
224.7
216.3
221.4
229.1
Note: Totals may not sum due to independent rounding.
3-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-93: CO2 Emissions from Wood Consumption by End-Use Sector (kt)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Industrial
135,348
136,269
140,331
138,537
138,339
144,502
144,285
Residential
59,808
44,340
59,657
52,872
46,180
44,649
53,336
Commercial
6,779
7,218
7,867
8,176
8,635
8,634
8,669
Electric Power
13,252
19,074
25,908
25,146
23,140
23,647
22,795
Total
215,186
206,901
233,762
224,730
216,293
221,432
229,085
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 2018, the United States transportation sector consumed an estimated 1,148.2 trillion Btu of ethanol (95 percent
of total), and as a result, produced approximately 78.6 MMT CO2 Eq. (78,603 kt) (see Table 3-94 and Table 3-95) of
CO2 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 3-94: CO2 Emissions from Ethanol Consumption (MMT CO2 Eq.)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation3
4.1
21.6
74.0
74.2
76.9
77.7
78.6
Industrial
0.1
1.2
1.6
1.9
1.8
1.9
1.4
Commercial
0.1
0.2
0.4
2.8
2.6
2.5
1.9
Total
4.2
22.9
76.1
78.9
81.2
82.1
81.9
Note: Totals may not sum due to independent rounding.
a See Annex 3.2, Table A-98 for additional information on transportation consumption of these fuels.
Table 3-95: CO2 Emissions from Ethanol Consumption (kt)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation3
4,059
21,616
74,006
74,187
76,903
77,671
78,603
Industrial
105
1,176
1,647
1,931
1,789
1,868
1,401
Commercial
63
151
422
2,816
2,558
2,550
1,913
Total
4,227
22,943
76,075
78,934
81,250
82,088
81,917
Note: Totals may not sum due to independent rounding.
a See Annex 3.2, Table A-98 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 2019a). 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 2019b).
In 2018, the United States consumed an estimated 242.9 trillion Btu of biodiesel, and as a result, produced
approximately 17.9 MMT CO2 Eq. (17,936 kt) (see Table 3-96 and Table 3-97) of CO2 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 2019b). There was no
measured biodiesel consumption prior to 2001 EIA (2019a).
Energy 3-113

-------
Table 3-96: CO2 Emissions from Biodiesel Consumption (MMT CO2 Eq.)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation3
NO
0.9
13.3
14.1
19.6
18.7
17.9
Total
NO
0.9
13.3
14.1
19.6
18.7
17.9
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
a See Annex 3.2, Table A-98 for additional information on transportation consumption of these fuels.
Table 3-97: CO2 Emissions from Biodiesel Consumption (kt)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation3
NO
856
13,349
14,077
19,648
18,705
17,936
Total
NO
856
13,349
14,077
19,648
18,705
17,936
Note: Totals may not sum due to independent rounding.
NO (Not Occurring)
a See Annex 3.2, Table A-98 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 2019a) (see Table 3-98), 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 CO2 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 CO2.
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 0
Fossil Fuel Combustion. Total U.S. ethanol consumption from EIA (2019a) 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 3-99). 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 2019a) (see Table 3-100).97
97 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.
3-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 3-98: Woody Biomass Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Industrial
1,441.9
1,451.7
1,495.0
1,475.9
1,473.8
1,539.4
1,537.1
Residential
580.0
430.0
578.5
512.7
447.8
433.0
517.2
Commercial
65.7
70.0
76.3
79.3
83.7
83.7
84.1
Electric Power
128.5
185.0
251.3
243.9
224.4
229.3
221.1
Total
2,216.2
2,136.7
2,401.1
2,311.8
2,229.8
2,285.5
2,359.5
Note: Totals may not sum due to independent rounding.





ible 3-99: Ethanol
Consumption by Sector (Trillion Btu)



End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation
59.3
315.8
1,081.1
1,083.7
1,123.4
1,134.6
1,148.2
Industrial
1.5
17.2
24.1
28.2
26.1
27.2
20.5
Commercial
0.9
2.2
6.2
41.1
37.4
37.2
27.9
Total
61.7
335.1
1,111.3
1,153.1
1,186.9
1,199.1
1,196.6
Note: Totals may not sum due to independent rounding.
Table 3-100: Biodiesel Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2014
2015
2016
2017
2018
Transportation
NO
11.6
180.8
190.6
266.1
253.3
242.9
Total
NO
11.6
180.8
190.6
266.1
253.3
242.9
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 CO2. 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 2018. Details on the emission trends through time are described in more detail in the Methodology
section, above.
Recalculations Discussion
EIA (2019a) updated heat contents for fuel ethanol, which resulted in updated ethanol consumption statistics and
CO2 emissions from ethanol consumption increased by less than 0.01 percent in 2017 relative to the previous
report. EIA (2019a) also updated biodiesel consumption statistics for 2016 and CO2 emissions from biodiesel
consumption increased by less than 0.01 percent relative to the previous report.
Planned Improvements
Future research will look into 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.
Energy 3-115

-------
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.98 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 CO2 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 CO2 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 GHGRP, the latest guidance
from the IPCC on the use of facility-level data in national inventories will be relied upon."
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 and
the biogenic components of MSW. EPA will examine EIA data on biogas to see if it can be included in future
inventories. EIA (2019a) 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. 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 2019a), whereas non-CC>2 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 2020) to fuel consumption data from EIA (2019a). 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 N2O from Stationary Combustion). EPA will
continue to investigate this discrepancy in order to apply a consistent approach to both CO2 and non-CC>2 emission
calculations for woody biomass consumption in the electric power sector.
98	See .
99	See .
3-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. 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.
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 (CO2), methane (CH4), nitrous oxide (N2O), 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, lime production, other process uses of carbonates (e.g., flux
stone, flue gas desulfurization, and glass manufacturing), ammonia production and urea consumption,
petrochemical production, aluminum production, HCFC-22 production, soda ash production and use, titanium
dioxide production, ferroalloy production, glass production, zinc production, phosphoric acid production, lead
production, silicon carbide production and consumption, nitric acid production, adipic acid production, and
caprolactam production.
Greenhouse gases that are used in manufacturing processes or by end-consumers include man-made compounds
such as hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SFs), 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
accumulate in the atmosphere as long as emissions continue. 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 2018, IPPU generated emissions of 376.5 million metric tons of CO2 equivalent (MMT CO2 Eq.), or 5.6 percent of
total U.S. greenhouse gas emissions.1 Carbon dioxide emissions from all industrial processes were 167.8 MMT CO2
Eq. (167,841 kt CO2) in 2018, or 3.1 percent of total U.S. CO2 emissions. Methane emissions from industrial
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

-------
processes resulted in emissions of approximately 0.3 MMT CO2 Eq. (13 kt CH4) in 2018, which was less than 1
percent of U.S. CFU emissions. Nitrous oxide emissions from IPPU were 25.5 MMT CO2 Eq. (86 kt N2O) in 2018, or
5.9 percent of total U.S. N2O emissions. In 2018 combined emissions of HFCs, PFCs, SF6, and NF3 totaled 182.8
MMT CO2 Eq. Total emissions from IPPU in 2018 were 9.0 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: 2018 Industrial Processes and Product Use Chapter Greenhouse Gas Sources
(MMT COz Eq.)
Substitution of Ozone Depleting Substances
Iron and Steel Production & Metallurgical Coke Production
Cement Production
Petrochemical Production
Ammonia Production
Lime Production
Adipic Acid Production
Other Process Uses of Carbonates
Nitric Acid Production
Electronics Industry
Carbon Dioxide Consumption
N2O from Product Uses
Electrical Transmission and Distribution
Urea Consumption for Non-Agricultural Purposes
HCFC-22 Production
Aluminum Production
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Glass Production
Magnesium Production and Processing
Zinc Production
Phosphoric Acid Production
Lead Production
Carbide Production and Consumption
168
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
The increase in overall IPPU emissions since 1990 reflects a range of emission trends among the emission sources.
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).
Similarly, CO2 and CFU emissions from many chemical production sources have either decreased or not changed
significantly since 1990, with the exception of petrochemical production which has steadily increased. Emissions
from mineral sources have either increased (e.g., cement manufacturing) or not changed significantly (e.g., glass
and lime manufacturing) since 1990 but largely follow economic cycles. Hydrofluorocarbon emissions from the
substitution of ODS have increased drastically since 1990, 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 N2O 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 CO2 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
2 See .
4-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
generally to: mineral products, chemical production, metal production, and emissions from the uses of HFCs, PFCs,
SFs, 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 2017) to
ensure that the trend is accurate. This year's estimates of HFC emissions from use of Ozone Depleting Substances
Substitutes reflect updates to stock and emission estimates to align with a recent national market characterization.
In addition, a technical aerosol end-use was added to the aerosols sector, in order to capture a portion of the
market that was not adequately encompassed by the current non-MDI aerosol end-use (EPA 2019b). Within the
Fire Protection sector, a correction was made to the lifetime for streaming agents, which was changed from 18
years to 24 years. Carbon content factors were also updated for the Iron and Steel emissions calculations. Finally,
the methods to estimate the CO2 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 increased greenhouse gas
emissions an average of 11.8 MMT CO2 Eq. (1.3 percent) across the time series.
In addition to the methodological updates noted above, the Inventory includes new categories not included in the
previous Inventory that improve completeness of the national IPPU estimates. This year's IPPU estimates include
fluorinated greenhouse gases (HFCs, NF3, PFCs, and SFs) from the Electronics Industry from manufacturing micro-
electronic mechanical systems (MEMS) and photovoltaics (PV), and this update increases greenhouse gas
emissions an average of 0.01 MMT CO2 Eq. over the time series. For more information on specific methodological
updates, please see the Recalculations discussion within the respective source category section of this chapter.
Table 4-1: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
2018
CO?
212.3
194.1
178.8
173.1
165.3
164.7
167.8
Iron and Steel Production &







Metallurgical Coke Production
104.7
70.1
58.2
47.9
43.6
40.6
42.6
Iron and Steel Production
99.1
66.2
54.5
43.5
41.0
38.6
41.3
Metallurgical Coke Production
5.6
3.9
3.7
4.4
2.6
2.0
1.3
Cement Production
33.5
46.2
39.4
39.9
39.4
40.3
40.3
Petrochemical Production
21.6
27.4
26.3
28.1
28.3
28.9
29.4
Ammonia Production
13.0
9.2
9.4
10.6
10.8
13.2
13.5
Lime Production
11.7
14.6
14.2
13.3
12.6
12.8
13.2
Other Process Uses of Carbonates
6.3
7.6
13.0
12.2
10.5
9.9
10.0
Carbon Dioxide Consumption
1.5
1.4
4.5
4.5
4.5
4.5
4.5
Urea Consumption for Non-







Agricultural Purposes
3.8
3.7
1.8
4.6
5.1
3.8
3.6
Ferroalloy Production
2.2
1.4
1.9
2.0
1.8
2.0
2.1
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.8
1.7
Titanium Dioxide Production
1.2
1.8
1.7
1.6
1.7
1.7
1.5
Aluminum Production
6.8
4.1
2.8
2.8
1.3
1.2
1.5
Glass Production
1.5
1.9
1.3
1.3
1.2
1.3
1.3
Zinc Production
0.6
1.0
1.0
0.9
0.9
1.0
1.0
Phosphoric Acid Production
1.5
1.3
1.0
1.0
1.0
1.0
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
+
+
+
+
+
+
+
ch4
0.3
0.1
0.2
0.2
0.3
0.3
0.3
Petrochemical Production
0.2
0.1
0.1
0.2
0.2
0.3
0.3
Ferroalloy Production
+
+
+
+
+
+
+
Industrial Processes and Product Use 4-3

-------
LarPiae Production ana







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







Metallurgical Coke Production
+
+
+
+
+
+
+
n2o
33.3
24.9
22.8
22.2
23.3
22.7
25.5
AdipicAcid Production
15.2
7.1
5.4
4.3
7.0
7.4
10.3
Nitric Acid Production
12.1
11.3
10.9
11.6
10.1
9.3
9.3
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
2.0
1.9
1.7
1.5
1.4
Electronics Industry
+
0.1
0.2
0.2
0.2
0.3
0.3
HFCs
46.5
128.7
166.3
170.5
170.5
172.5
171.6
Substitution of Ozone Depleting







Substances3
0.2
108.4
160.9
165.8
167.3
166.9
167.8
HCFC-22 Production
46.1
20.0
5.0
4.3
2.8
5.2
3.3
Electronics Industry
0.2
0.2
0.3
0.3
0.3
0.4
0.4
Magnesium Production and







Processing
0.0
0.0
0.1
0.1
0.1
0.1
0.1
PFCs
24.3
6.7
5.6
5.1
4.3
4.0
4.6
Electronics Industry
2.8
3.2
3.1
3.0
2.9
2.9
3.0
Aluminum Production
21.5
3.4
2.5
2.0
1.4
1.0
1.6
Substitution of Ozone Depleting







Substances
0.0
+
+
+
+
+
0.1
sf6
28.8
11.8
6.5
5.5
6.1
5.9
5.9
Electrical Transmission and







Distribution
23.2
8.4
4.8
3.8
4.1
4.1
4.1
Magnesium Production and







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







and NF3
+
+
+
+
+
+
+
Electronics Industry
+
+
+
+
+
+
+
Total
345.6
366.8
380.8
377.1
370.4
370.7
376.5
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
2014
2015
2016
2017
2018
co2
212,326
194,098
178,783
173,083
165,304
164,691
167,841
Iron and Steel Production &







Metallurgical Coke Production
104,734
70,081
58,187
47,944
43,624
40,576
42,600
Iron and Steel Production
99,126
66,160
54,467
43,528
40,981
38,598
41,318
Metallurgical Coke Production
5,608
3,921
3,721
4,417
2,643
1,978
1,282
Cement Production
33,484
46,194
39,439
39,907
39,439
40,324
40,324
Petrochemical Production
21,611
27,383
26,254
28,062
28,310
28,910
29,424
Ammonia Production
13,047
9,196
9,377
10,634
10,838
13,216
13,532
Lime Production
11,700
14,552
14,210
13,342
12,630
12,833
13,223
Other Process Uses of Carbonates
6,297
7,644
12,954
12,182
10,505
9,935
9,954
Carbon Dioxide Consumption
1,472
1,375
4,471
4,471
4,471
4,471
4,471
Urea Consumption for Non-
Agricultural Purposes
Ferroalloy Production
3,784
2,152
3,653
1,392
1,807
1,914
4,578
1,960
5,132
1,796
3,769
1,975
3,628
2,063
4-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Soda Ash Production
1,431
1,655
1,685
1,714
1,723
1,753
1,714
Titanium Dioxide Production
1,195
1,755
1,688
1,635
1,662
1,688
1,541
Aluminum Production
6,831
4,142
2,833
2,767
1,334
1,205
1,451
Glass Production
1,535
1,928
1,336
1,299
1,241
1,296
1,263
Zinc Production
632
1,030
956
933
925
1,009
1,009
Phosphoric Acid Production
1,529
1,342
1,037
999
998
1,028
940
Lead Production
516
553
459
473
500
513
513
Carbide Production and







Consumption
375
219
173
180
174
186
189
Magnesium Production and







Processing
1
3
2
3
3
3
1
ch4
12
4
6
9
11
11
13
Petrochemical Production
9
3
5
7
10
10
12
Ferroalloy Production
1
+
1
1
1
1
1
Carbide Production and







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







Metallurgical Coke Production
1
1
+
+
+
+
+
n2o
112
84
77
74
78
76
86
Adipic Acid Production
51
24
18
14
23
25
35
Nitric Acid Production
41
38
37
39
34
31
31
N20 from Product Uses
14
14
14
14
14
14
14
Caprolactam, Glyoxal, and







Glyoxylic Acid Production
6
7
7
6
6
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). 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
Industrial Processes and Product Use 4-5

-------
(e.g., ceramics, non-metallurgical magnesium production, glyoxal and glyoxylic acid production, CFU 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). 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 IPPU, rather than Energy; however, due to national circumstances regarding
the allocation of energy statistics and carbon I balance data, the United States reports non-energy uses in the
Energy chapter of this Inventory. Reporting these non-energy use emissions under IPPU would involve making
artificial adjustments to the non-energy use C balance. These artificial adjustments would also result in the C
emissions for lubricants, waxes, and asphalt and road oil being reported under IPPU, while the C storage for
lubricants, waxes, and asphalt and road oil would be reported under Energy. To avoid presenting an incomplete C
balance, double-counting, and adopting a less transparent approach, the entire calculation of C storage and C
emissions is therefore conducted in the Non-Energy Uses of Fossil Fuels category calculation methodology and
reported under the Energy sector. For more information, see the Methodology section for CO2 from Fossil Fuel
Combustion and Section 3.2, 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 and zinc production) are reported in the IPPU chapter, unless otherwise noted due to
specific national circumstances. More information on the methodology to adjust for these emissions within the
Energy chapter is described in the Methodology section of CO2 from Fossil Fuel Combustion (3.1 Fossil Fuel
Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating Emissions of CO2 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). 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
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 was tailored to include specific procedures
4-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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, or Tier 1, QA/QC procedures 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
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,
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.
3 See .
Industrial Processes and Product Use 4-7

-------
Box 4-2: Industrial Process and Product Use Data from EPA's Greenhouse Gas Reporting Program
On October 30, 2009, the U.S. 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 GHGRP. The rule applies to direct greenhouse gas emitters, fossil fuel suppliers,
industrial gas suppliers, and facilities that inject CO2 underground for sequestration or other reasons and
requires reporting by sources or suppliers in 41 industrial categories ("Subparts"). 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 CO2 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.
EPA's GHGRP dataset and the data presented in this Inventory 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. 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
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 definitions for source
categories 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-level
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 sections6, EPA is continuing to analyze how facility-level
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).
See Annex 9 for more information on use of GHGRP data in the Inventory.
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.
4-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (CO2) 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 (CaCOs), 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 CO2 in a process known as
calcination or calcining. The quantity of CO2 emitted during clinker production is directly proportional to the lime
content of the clinker. During calcination, each mole of CaCC>3 heated in the clinker kiln forms one mole of CaO and
one mole of CO2. The CO2 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 there are
few process emissions of C02as a result of the reactions. The clinker is then rapidly cooled to maintain quality,
then very finely ground with a small amount of gypsum and potentially other materials (e.g., ground granulated
blast furnace slag, etc.), and used to make Portland and similar cements.7
Carbon dioxide emitted from the chemical process of cement production is the second largest source of industrial
CO2 emissions in the United States. Cement is produced in 34 states and Puerto Rico. Texas, California, Missouri,
Florida, and Alabama were the leading cement-producing states in 2018 and accounted for almost 50 percent of
total U.S. production (USGS 2019). Based on both GHGRP data (EPA 2018) and USGS reported data, clinker
production in 2018 remained at relatively flat levels compared to 2017. Cement sales remained relatively stagnant
in between 2017 to 2018 and imports of clinker for consumption decreased by approximately 25 percent over this
same period (USGS 2019). In 2018, U.S. clinker production totaled 77,500 kilotons (EPA 2018). The resulting CO2
emissions were estimated to be 40.3 MMT CO2 Eq. (40,324 kt) (see Table 4-3).
Table 4-3: CO2 Emissions from Cement Production (MMT CO2 Eq. and kt)
Year
MMTCOz Eq.
kt
1990
33.5
33,484
2005
46.2
46,194
2014
39.4
39,439
2015
39.9
39,907
2016
39.4
39,439
2017
40.3
40,324
2018
40.3
40,324
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
7 Approximately three percent of total clinker production is used to produce masonry cement, which is produced using
plasticizers (e.g., ground limestone, lime, etc.) and Portland cement (USGS 2011). Carbon dioxide emissions that result from the
production of lime used to create masonry cement are included in the Lime Manufacture source category.
Industrial Processes and Product Use 4-9

-------
increased by 20 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 associated decrease in demand for construction materials. Since 2010, emissions have increased by
roughly 28 percent due to increasing cement consumption. 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.
Methodology
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,8 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 CO2 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 CO2 per ton of clinker produced, which was
determined as follows:
EFciinker = 0.650 CaO X [(44.01 g/mole CO2) -v- (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 CChemissions.
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 CO2 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 CO2 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
to produce clinker but no information is currently available on the total amount of CKD produced annually.9
Because data are not currently available to derive a country-specific CKD correction factor, a default correction
factor of 1.02 (two percent) was used to account for CKD CO2 emissions, as recommended by the IPCC (IPCC
2006).10 Total cement production emissions were calculated by adding the emissions from clinker production to
the emissions assigned to CKD.
8 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 2017 and 2018, the percentage of facilities not using CEMS was 12 percent and
8	percent, respectively.
9	The USGS Minerals Yearbook: Cement notes that CKD values used for clinker production are likely underreported.
10	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)..."
4-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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). The data
were compiled by USGS (to the nearest ton) through questionnaires sent to domestic clinker and cement
manufacturing plants, including the facilities in Puerto Rico. Clinker production values in the current Inventory
report utilize GHGRP data for the years 2014 through 2017 (EPA 2018). 2017 GHGRP data are used as a proxy for
2018 as GHGRP data are not available for this report. 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
2014	75,800
2015	76,700
2016	75,800
2017	77,500
201	8	77,500	
Notes: Clinker production from 1990 through
2018 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 CO2 from CKD loss can range from 1.5 to 8 percent
depending upon plant specifications. Additionally, some amount of CO2 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 CO2 in the atmosphere to create calcium carbonate. This reaction
only occurs in roughly the outer 0.2 inches of the total thickness. Because the amount of CO2 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 CO2 emission factor for clinker production, and the
emission factor for additional CO2 emissions from CKD, 2018 CO2 emissions from cement production were
estimated to be between 37.8 and 42.8 MMT CO2 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 40.3 MMT CO2
Eq.
Industrial Processes and Product Use 4-11

-------
Table 4-5: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Cement
Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Cement Production
C02
40.3
37.8 42.8
-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 2018. 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 the 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.11 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 4-12ortland cement to report greenhouse gas emissions, including
CO2 process emissions from each kiln, CO2 combustion emissions from each kiln, CH4 and N2O combustion
emissions from each kiln, and CO2, CH4, and N2O 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.
As mentioned above, EPA compares GHGRP clinker production data to the USGS clinker production data. For the
year 2014, USGS and GHGRP clinker production data showed a difference of approximately 2 percent, while in
2015,	2016, and in 2017 that difference decreased to 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 CO2 Eq. in 2015,
2016,	and in 2017. The information collected by the USGS National Minerals Information Center surveys continue
to be an important data source.
Recalculations Discussion
No recalculations were performed for the 1990 through 2017 portion of the time series.
11 See GHGRP Verification Fact Sheet .
4-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 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.12 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 CO2 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 will be to identify 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 CO2
reabsorption rates via carbonation for various cement products. This information is not reported by facilities
subject to report to GHGRP.
4.2 Lime Production fCRF 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 (CO2) is
generated during the calcination stage, when limestone—mostly calcium carbonate (CaCOs)—is roasted at high
temperatures in a kiln to produce calcium oxide (CaO) and CO2. The CO2 is given off as a gas and is normally
emitted to the atmosphere.
CaCO3 —> CaO + C02
Some of the CO2 generated during the production process, however, is recovered at some facilities for use in sugar
refining and precipitated calcium carbonate (PCC) production.13 Emissions from fuels consumed for energy
purposes during the production of lime are included for 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,
37 percent; environmental uses, 31 percent; chemical and industrial uses, 22 percent; construction uses, 9
percent; and refractory dolomite, 1 percent (USGS 2018). The major uses are in steel making, flue gas
12	See IPCC Technical Bulletin on Use of Facility-Specific Data in National Greenhouse Gas Inventories .
13	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.
Industrial Processes and Product Use 4-13

-------
desulfurization 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 CO2 scrubber,
and there has been experimentation on the use of lime to capture CO2 from electric power plants.
Lime production in the United States—including Puerto Rico—was reported to be 18,100 kilotons in 2018 (USGS
2020). Lime production in 2018 increased by about 3 percent compared to 2017 levels, due primarily to an
increase in hydrated lime output (USGS 2019 and 2020). At year-end 2018, there were 74 operating primary lime
plants in the United States, including Puerto Rico.14 Principal lime producing states are Missouri, Alabama, Ohio,
Texas, and Kentucky (USGS 2019).
U.S. lime production resulted in estimated net CO2 emissions of 13.2 MMT CO2 Eq. (13,223 kt) (see Table 4-6 and
Table 4-7). The trends in CO2 emissions from lime production are directly proportional to trends in production,
which are described below.
Table 4-6: CO2 Emissions from Lime Production (MMT CO2 Eq. and kt)
Year MMT CP2 Eq.	kt
1990	11.7	11,700
2005	14.6	14,552
2014	14.2	14,210
2015	13.3	13,342
2016	12.6	12,630
2017	12.8	12,833
201	8	1^2	13,223
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
2014	14,715	505	14,210
2015	13,764	422	13,342
2016	13,000	370	12,630
2017	13,234	401	12,833
2018	13,624	401	13,223
Note: Totals may not sum due to independent rounding.
a 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 CO2 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:
14 In 2018, 74 operating primary lime facilities in the United States reported to the EPA Greenhouse Gas Reporting Program.
4-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
[(44.01 g/mole C02) 4- (56.08 g/mole CaO)] x (0.9500 CaO/lime) = 0.7455 g C02/g lime
For dolomitic lime:
[(88.02 g/mole C02) 4 (96.39 g/mole CaO)] x (0.9500 CaO/lime) = 0.8675 g C02/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 24.3 percent for high-calcium hydrated lime, and 27.2 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. 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 2018)
based on reported facility-level data for years 2010 through 2017. 2018 C02 captured for on-site process use is
proxied with the 2017 value due to GHGRP data availability at the time of this Inventory. 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 2017. 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, high-calcium- and dolomitic-quicklime, high-calcium- and dolomitic-hydrated, and
dead-burned dolomite) for 1990 through 2018 (see Table 4-8) were obtained from U.S. Geological Survey (USGS)
(USGS 2019) annual reports and are compiled by USGS to the nearest ton. The high-calcium quicklime and
dolomitic quicklime values were estimated using the ratio of the 2015 quicklime values to the 2018 total values.
The 2015 values for high-calcium hydrated, dolomitic hydrated, and dead-burned dolomite were used since there
is less fluctuation in their production from year to year. Natural hydraulic lime, which is produced from CaO and
hydraulic calcium silicates, is not manufactured in the United States (USGS 2018). 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 is presented in Table 4-9 (IPCC 2006). The CaO and CaO*MgO contents of
lime 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
2014
14,100
2,740
2,190
279
200
2015
13,100
2,550
2,150
279
200
2016
12,281
2,390
2,150
279
200
2017
12,532
2,439
2,150
279
200
Industrial Processes and Product Use 4-15

-------

High-Calcium
Dolomitic
High-Calcium
Dolomitic
Dead-Burned
Year
Quicklime
Quicklime
Hydrated
Hydrated
Dolomite
2018
12,950
2,521
2,150
279
200
Table 4-9: Adjusted Lime Production (kt)
Year
High-Calcium
Dolomitic
1990
12,466
2,800
2005
15,721
3,522
2014
15,699
3,135
2015
14,670
2,945
2016
13,850
2,786
2017
14,101
2,835
2018
14,520
2,916
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 CO2 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 CO2 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, CO2 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 CO2;
whereas most of the lime used in steel making reacts with impurities such as silica, sulfur, and aluminum
compounds. Quantifying the amount of CO2 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
CO2 are "reused" are required to quantify the amount of CO2 that is reabsorbed. Research conducted thus far has
not yielded the necessary information to quantify CO2 reabsorption rates.15 However, some additional information
on the amount of CO2 consumed on site at lime facilities has been obtained from EPA's GHGRP.
In some cases, lime is generated from calcium carbonate byproducts at pulp mills and water treatment plants.16
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 I is present from the wood. Kraft
mills recover the calcium carbonate "mud" after the causticizing operation and calcine it back into lime—thereby
15	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).
16	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 13 C2H2 + Ca(OH) 2], not calcium
carbonate [CaCOs]. Thus, the calcium hydroxide is heated in the kiln to simply expel the water [Ca(OH)2 + heat HCaO + H20] and
no C02 is released.
4-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
generating CO2—for reuse in the pulping process. Although this re-generation of lime could be considered a lime
manufacturing process, the CO2 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
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 CO2 emissions
for 2018 were estimated to be between 12.9 and 13.5 MMT CO2 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
13.2 MMT CO2 Eq.
Table 4-10: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lime
Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Lime Production
C02
13.2
12.9 13.5 -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 2018. 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 CO2
captured for onsite use applicable to lime manufacturing facilities can be found under Subpart S (Lime
Industrial Processes and Product Use 4-17

-------
Manufacturing) of the GHGRP regulation (40 CFR Part 98).17 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).18 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 years 2016 and 2017 based on updated quicklime production data from USGS.
The updates resulted in a decrease in CO2 emissions of about 2.4 percent for both 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 in particular, aggregated activity data on lime production by type. In addition, initial review of
data has identified that there are several facilities that 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 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.19
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. 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. However, 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. A next step to continue this planned
improvements is to identify the appropriate staff within NLA to work and review the remaining data needs,
including GHGRP data. At the time of this Inventory, due to limited resources and need for additional outreach and
information, this planned improvement is still in process and has not been incorporated into this current Inventory
report.
17	See .
18	See .
19	See .
4-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
(CO2) 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 CO2 emissions
during the glass melting process are limestone, dolomite, and soda ash. The main former in all types of glass is
silica (SiCh). 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,
Na2CC>3) and potash (potassium carbonate, K2O). 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
are limestone (CaCOs), dolomite (CaCOsMgCOs), alumina (AI2O3), magnesia (MgO), barium carbonate (BaCOs),
strontium carbonate (SrCOs), lithium carbonate (IJ2CO3), and zirconia (ZrCh) (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 cullet broker services.
The raw materials (primarily limestone, dolomite and soda ash) release CO2 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 CO2 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.20
In 2018, 713 kilotons of limestone and 2,280 kilotons of soda ash were consumed for glass production (USGS 2019;
USGS 2019a). Dolomite consumption data for glass manufacturing was reported to be zero for 2018. Use of
limestone and soda ash in glass production resulted in aggregate CO2 emissions of 1.3 MMT CO2 Eq. (1,263 kt) (see
Table 4-11). Overall, emissions have decreased 18 percent from 1990 through 2018.
Emissions in 2018 decreased approximately 3 percent from 2017 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 a corresponding decrease in 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).
20 Excerpt from Glass & Glass Product Manufacturing Industry Profile, First Research. Available online at:
.
Industrial Processes and Product Use 4-19

-------
Table 4-11: CO2 Emissions from Glass Production (MMT CO2 Eq. and kt)
Year
MMT C02 Eq.
kt
1990
1.5
1,535
2005
1.9
1,928
2014
2015
2016
2017
2018
1.3
1.3
1.2
1.3
1.3
1,336
1,299
1,241
1,296
1,263
Note: Totals may not sum due to
independent rounding.
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 CCh/metric ton carbonate): limestone, 0.43971; dolomite, 0.47732; and soda ash, 0.41492.
Consumption data for 1990 through 2018 of limestone, dolomite, and soda ash used for glass manufacturing were
obtained from the U.S. Geological Survey (USGS) Minerals Yearbook: Crushed Stone Annual Report (1995 through
2016a), 2017 and 2018 preliminary data from the USGS Crushed Stone Commodity Expert (Willett 2019a), the
USGS Minerals Yearbook: Soda Ash Annual Report (1995 through 2015) (USGS 1995 through 2015b), USGS Mineral
Industry Surveys for Soda Ash in December 2018 (USGS 2019) and the U.S. Bureau of Mines (1991 and 1993a),
which are reported to the nearest ton. 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 consumption by end use to the
1992 total limestone and dolomite consumption values.
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; or (2) the average
percent of total limestone or dolomite for the withheld end-use in the preceding and succeeding years.
A large quantity of limestone and dolomite reported to the USGS under the categories "unspecified-reported" and
"unspecified-estimated." A portion of this consumption is believed to be limestone or dolomite used for glass
manufacturing. The quantities listed under the "unspecified" categories were, therefore, allocated to glass
manufacturing according to the percent limestone or dolomite consumption for glass manufacturing end use for
that year.21 For 2018, the unspecified uses of both limestone and dolomite consumption were not available at the
time of publication, so 2017 values were used as a proxy for these values.
Based on the 2018 reported data, the estimated distribution of soda ash consumption for glass production
compared to total domestic soda ash consumption is 47 percent (USGS 1995 through 2015b, 2018, 2019).
Table 4-12: Limestone, Dolomite, and Soda Ash Consumption Used in Glass Production (kt)
Activity	1990 2005 2014 2015 2016 2017 2018
Limestone	430	920	765 699 455 720 720
21 This approach was recommended by USGS.
4-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Dolomite
Soda Ash
59
3,177
541
3,050
0
2,410
0
2,390
0
2,510
0
2,360
0
2,280
Total
3,666
4,511
3,175 3,089 2,965 3,080 3,000
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 2018, 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 2018, glass
production CO2 emissions were estimated to be between 1.2 and 1.3 MMT CO2 Eq. at the 95 percent confidence
level. This indicates a range of approximately 4 percent below and 5 percent above the emission estimate of 1.3
MMTCCh Eq.
Table 4-13: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Glass
Production (MMT CO2 Eq. and Percent)
Source
„ 2018 Emission Estimate
Gas
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTCOzEq.) (%)


Lower Upper
Bound Bound
Lower Upper
Bound Bound
Glass Production
C02 1.3
1.2 1.3
-4% +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 2018. 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).
Industrial Processes and Product Use 4-21

-------
Recalculations Discussion
For the current Inventory, 1990 through 2018, updated USGS data on limestone and dolomite consumption was
available for 2016 and 2017. The revised values used in the current Inventory resulted in updated emissions
estimates for the years 2016 (decrease of 0.6 percent) and 2017 (decrease of 1.4 percent).
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, there are other carbonates that are also consumed for glass
manufacturing, although in smaller quantities. EPA has initiated review of available activity data on carbonate
consumption by type in the glass industry from EPA's Greenhouse Gas Reporting Program (GHGRP) reported
annually since 2010, as well as USGS publications. This is a long-term planned improvement.
EPA has initiated review of EPA's GHGRP data to help understand the completeness of emission estimates and
facilitate category-specific QC per Volume 1 of the 2006IPCC Guidelines for the Glass Production source category.
EPA's 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 EPA's GHGRP, the latest guidance from the IPCC on the use of facility-level data in national
inventories will be relied upon.22 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 (CaCOs), dolomite (CaCOsMgCOs),23 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 CO2 as a byproduct.
CaCO3 —> CaO + C02
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 consumption
22	See .
23	Limestone and dolomite are collectively referred to as limestone by the industry, and intermediate varieties are seldom
distinguished.
4-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).
Limestone is 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 2020). Similarly, dolomite deposits are
also widespread throughout the world. 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 1995a through 2020). Internationally,
two types of soda ash are produced, natural and synthetic. In 2016, 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, with only two states producing natural soda ash: Wyoming and California (USGS 2016b).
Similar to glass production discussed above, emissions from soda ash production are reported under that category
(i.e., CRF Source Category 2B7).
In 2018,19,577 kt of limestone, 1,931 kt of dolomite, and 2,576 kt of soda ash were consumed for these emissive
applications, excluding glass manufacturing (Willett 2019, USGS 2019). Usage of limestone, dolomite and soda ash
resulted in aggregate CO2 emissions of 10.0 MMT CO2 Eq. (9,954 kt) (see Table 4-14 and Table 4-15). While 2018
emissions have increased under 1 percent compared to 2017, overall emissions have increased 58 percent from
1990 through 2018.
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	Usesb	Total
1990 2.6	1.4	0.1	1.4	0.8	6.3
2005 2.6	3.0	0.0	1.3	0.7	7.6
2014
2.9
7.1
0.0
1.1
1.8
13.0
2015
2.9
7.3
0.0
1.1
0.9
12.2
2016
2.5
5.9
0.0
1.1
1.1
10.5
2017
2.4
5.6
0.0
1.1
0.8
9.9
2018
2.4
5.6
0.0
1.1
0.8
10.0
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.
Table 4-15: CO2 Emissions from Other Process Uses of Carbonates (kt)





Other




Magnesium
Soda Ash
Miscellaneous

Year
Flux Stone
FGD
Production
Consumption3
Usesb
Total
1990
2,592
1,432
64
1,390
819
6,297
2005
2,649
2,973
0
1,305
718
7,644
2014
2,911
7,111
0
1,143
1,790
12,954
2015
2,901
7,335
0
1,075
871
12,182
2016
2,477
5,860
0
1,082
1,087
10,505
Industrial Processes and Product Use 4-23

-------
2017	2,444	5,598	0	1,058	835	9,935
201	8	2,443	5,606	0	1,069	836	9,954
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.
M
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 - limestone: 0.43971 metric ton CCh/metric ton carbonate, and dolomite: 0.47732 metric ton
CCh/metric ton carbonate.24 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 CO2 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 CO2 emissions ceased its
operations (USGS 1995b through 2012; USGS 2013).
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), preliminary data for 2017 and 2018 from USGS Crushed Stone Commodity Expert (Willett
2018a, 2018b, 2019), American Iron and Steel Institute limestone and dolomite consumption data (AISI 2018,
2019), and the U.S. Bureau of Mines (1991 and 1993a), which are reported to the nearest ton. For 2018, estimates
of the unspecified uses of both limestone and dolomite consumption were available at the time of publication,
however the specified uses were not available, so 2017 values were used as a proxy for these values. The
production capacity data for 1990 through 2018 of dolomitic magnesium metal also came from the USGS (1995b
through 2012; USGS 2013) 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.
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
24 2006 IPCC Guidelines, Volume 3: Chapter 2, Table 2.1.
4-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
"unspecified uses" was, therefore, allocated to all other reported end-uses according to each end-use's fraction of
total consumption in that year.25
Table 4-16: Limestone and Dolomite Consumption (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
Flux Stone
6,737
7,022
7,599
7,834
6,933
6,861
6,857
Limestone
5,804
3,165
4,243
4,590
4,995
4,963
4,926
Dolomite
933
3,857
3,356
3,244
1,938
1,899
1,931
FGD
3,258
6,761
16,171
16,680
13,327
12,732
12,749
Other Miscellaneous Uses
1,835
1,632
4,069
1,982
2,471
1,900
1,902
Total
11,830
15,415
27,839
26,496
22,731
21,493
21,508
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 CO2 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 CO2) are released for every metric ton of soda
ash consumed. The activity data for soda ash consumption for 1990 to 2018 (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). 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
2014
2015
2016
2017
2018
Soda Asha
3,351
3,144
2,754
2,592
2,608
2,550
2,576
Total
3,351
3,144
2,754
2,592
2,608
2,550
2,576
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
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. This year, EPA
reinitiated dialogue with 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
25 This approach was recommended by USGS, the data collection agency.
Industrial Processes and Product Use 4-25

-------
published by USGS. During this discussion, the expert confirmed that EPA's range of uncertainty was still
reasonable (Willett 2017a).
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 2018 were estimated to be between 8.9 and 11.4 MMT CO2 Eq.
at the 95 percent confidence level. This indicates a range of approximately 11 percent below and 14 percent above
the emission estimate of 9.4 MMT CO2 Eq.
Table 4-18: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Other
Process Uses of Carbonates (MMT CO2 Eq. and Percent)


2018 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


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



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Other Process Uses
of Carbonates
C02
10.0
8.9
11.4
-11% +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 2018. 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).
Recalculations Discussion
For the current Inventory, 1990 through 2018, updated USGS data on limestone and dolomite consumption was
available for 2016 and 2017. The revised values used in the current Inventory resulted in updated emissions
estimates for the years 2016 and 2017. Compared to the previous Inventory, 1990 through 2017, emissions in the
current Inventory for 2016 decreased by 4 percent (464 kt CO2 Eq.) and decreased by 2 percent (204 kt CO2 Eq.) for
2017.
4-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 are 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 (CO2) 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 CO2 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 CO2 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 CO2 produced by the process is captured and used to produce urea rather than
being emitted to the atmosphere. In 2018, there were 15 companies operating 34 ammonia producing facilities in
16 states. Approximately 50 percent of domestic ammonia production capacity is concentrated in the states of
Louisiana, Oklahoma, and Texas (USGS 2019).
There are five principal process steps in synthetic ammonia production from natural gas feedstock. The primary
reforming step converts methane (CH4) to CO2, 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 CO2 in this
step of the process. The secondary reforming step converts the remaining CH4 feedstock to CO and CO2. The CO in
the process gas from the secondary reforming step (representing approximately 15 percent of the process gas) is
converted to CO2 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 CO2 is included in a waste gas stream
with other process impurities and is absorbed by a scrubber solution. In regenerating the scrubber solution, CO2 is
released from the solution.
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
Industrial Processes and Product Use 4-27

-------
To produce synthetic ammonia from petroleum coke, the petroleum coke is gasified and converted to CO2 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 CO2 produced during the production of ammonia is emitted directly to the atmosphere. Some of the
ammonia and some of the CO2 produced by the synthetic ammonia process are used as raw materials in the
production of urea [COfNHhh], 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 CO2 emitted directly to the atmosphere from the synthetic ammonia production process is accounted for
in determining emissions from ammonia production. The CO2 that is captured during the ammonia production
process and used to produce urea does not contribute to the CO2 emission estimates for ammonia production
presented in this section. Instead, CO2 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 CO2 resulting
from agricultural applications of urea are accounted for in the Agriculture chapter. Previously, these emission
estimates from the agricultural application of urea were accounted for in the Cropland Remaining Cropland section
of the Land Use, Land Use Change, and Forestry chapter. Emissions of CO2 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 CO2 from ammonia production in 2018 were 13.5 MMT CO2 Eq. (13,532 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 increased by about 4 percent. Emissions in 2018 have increased by
approximately 2 percent from the 2017 levels. Agricultural demands continue to drive demand for nitrogen
fertilizers (USGS 2019).
Table 4-19: CO2 Emissions from Ammonia Production (MMT CO2 Eq.)
Source
1990
2005
2014
2015
2016
2017
2018
Ammonia Production
13.0
9.2
9.4
10.6
10.8
13.2
13.5
Total	13.0	9.2	9.4 10.6 10.8 13.2 13.5
Table 4-20: CO2 Emissions from Ammonia Production (kt)
Source
1990
2005
2014
2015
2016
2017
2018
Ammonia Production
13,047
9,196
9,377
10,634
10,838
13,216
13,532
Total
13,047
9,196
9,377
10,634
10,838
13,216
13,532
Methodology
"•J" m
For the U.S. Inventory, CO2 emissions from the production of synthetic ammonia from natural gas feedstock are
estimated using a country-specific approach modified from the 2006IPCCGuidelines (IPCC 2006) Tier 1 and 2
methods. In the country-specific approach, emissions are not based on total fuel requirement per the 2006 IPCC
Guidelines due to data disaggregation limitations of energy statistics provided by the Energy Information
Administration (EIA). A country-specific emission factor is developed and applied to national ammonia production
to estimate emissions. The method uses a CO2 emission factor published by the European Fertilizer Manufacturers
Association (EFMA) that is based on natural gas-based ammonia production technologies that are similar to those
employed in the United States. This CO2 emission factor of 1.2 metric tons CCh/metric ton NH3 (EFMA 2000a) is
applied to the percent of total annual domestic ammonia production from natural gas feedstock.
4-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
The emission factor of 1.2 metric ton CCh/metric ton NH3 for production of ammonia from natural gas feedstock
was taken from the EFMA Best Available Techniques publication, Production of Ammonia (EFMA 2000a). The EFMA
reported an emission factor range of 1.15 to 1.30 metric ton CCh/metric ton NH3, with 1.2 metric ton CCh/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 CO2.
Emissions of CChfrom ammonia production are then adjusted to account for the use of some of the CO2 produced
from ammonia production as a raw material in the production of urea. The CO2 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 CCh/urea factor of 44/60, assuming complete conversion of ammonia (NH3) and CO2 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 CO2 emission factor for production of ammonia from petroleum coke is based on plant-specific
data, wherein all carbon contained in the petroleum coke feedstock that is not used for urea production is
assumed to be emitted to the atmosphere as CO2 (Bark 2004). 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. The CO2 emission factor of 3.57 metric tons CCh/metric ton Nl-hfor
the petroleum coke feedstock process (Bark 2004) is applied to the percent of total annual domestic ammonia
production from petroleum coke feedstock. The implied CO2 emission factor for total ammonia production is
therefore a combination of the emissions factor 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 and 2015 and
2016 there were increases in the amount of ammonia produced from petroleum coke which caused increases in
the implied emission factor across those years.
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 CChfrom Fossil Fuel
Combustion. See the Planned Improvements section on improvements of reporting fuel and feedstock CO2
emissions utilizing EPA's GHGRP data to improve consistency with 2006 IPCC Guidelines.
The total ammonia production data for 2011 through 2018 were obtained from American Chemistry Council (ACC
2019). 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 2018) for
2012 through 2018. 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 in the current Inventory report utilize GHGRP data for the years 2011 through 2017 (EPA
2018). GHGRP urea production data for 2018 were not yet published and so 2017 data were used as a proxy.
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
Industrial Processes and Product Use 4-29

-------
1990
15,425
5,463
7,450
2005
10,143
3,865
5,270
2014
10,515
4,078
5,561
2015
11,765
4,312
5,880
2016
12,305
5,419
7,390
2017
14,070
5,419
7,390
2018
14,370
5,419
7,390
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 collocated 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 CO2 from ammonia production plants for purposes other than urea production (e.g., commercial sale,
etc.) has not been considered in estimating the CO2 emissions from ammonia production, as data concerning the
disposition of recovered CO2 are not available. Such recovery may or may not affect the overall estimate of CO2
emissions depending upon the end use to which the recovered CO2 is applied. Further research is required to
determine whether byproduct CO2 is being recovered from other ammonia production plants for application to
end uses that are not accounted for elsewhere. However, for reporting purposes, CO2 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 2018 were estimated to be between 12.9 and 14.1 MMT CO2 Eq. at the 95
percent confidence level. This indicates a range of approximately 4 percent below and 5 percent above the
emission estimate of 13.5 MMT CO2 Eq.
Table 4-22: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Ammonia Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Ammonia Production
C02
13.5
12.9 14.1
-4% +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 2018. Details on the emission trends through time are described in more detail in the Methodology
section, above.
4-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).26 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.27 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
No recalculations of ammonia emissions were performed for the 1990 through 2017 portion of the time series.
However, the carbon factors used to determine the amount of natural gas used for ammonia feedstock were
updated to be consistent with the factors used in the fossil fuel combustion estimates. This update did not have an
impact on process-related ammonia emissions presented here but did impact the amount of natural gas
subtracted from energy use as part of the CO2 Emissions from Fossil Fuel Combustion calculations (see Annex 2.3
for more information).
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 89 188),28 that include facility-
level ammonia production data and feedstock consumption. This data will first be reported by facilities in 2018 and
available post-verification to assess in early 2019 for use in future Inventories (e.g., 2021 Inventory report) if the
data meets GHGRP CBI aggregation criteria. 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-level 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.29 Specifically, the planned improvements
include assessing the anticipated new data to update the emission factors to include both fuel and feedstock CO2
emissions to improve consistency with 2006 IPCC Guidelines, in addition to reflecting CO2 capture and storage
26	See .
27	See .
28	See .
29	See .
Industrial Processes and Product Use 4-31

-------
practices (beyond use of CO2 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 so 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 and carbon dioxide (CO2) as raw materials. All urea produced in the United States
is assumed to be produced at ammonia production facilities where both ammonia and CO2 are generated. There
were 34 plants producing ammonia in the United States during 2018, with two additional plants sitting idle for the
entire year (USGS 2019b).
The chemical reaction that produces urea is:
2nh3+ C02 -> NH2COONH4 -> CO(NH2)2 +h2o
This section accounts for CO2 emissions associated with urea consumed exclusively for non-agricultural purposes.
Carbon dioxide emissions associated with urea consumed for fertilizer are accounted for in the Agriculture
chapter.
Urea is used as a nitrogenous fertilizer for agricultural applications and also in a variety of industrial applications.
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 CO2 from urea consumed for non-agricultural purposes in 2018 were estimated to be 3.6 MMT CO2
Eq. (3,628 kt), and are summarized in Table 4-23 and Table 4-24. Net CO2 emissions from urea consumption for
non-agricultural purposes in 2018 have decreased by approximately 4 percent from 1990. The significant decrease
in emissions during 2014 can be attributed to a decrease in the amount of urea imported by the United States
during that year. Similarly, 2017 also saw a decrease in the amount of urea imported to the United States as well
as a significant increase in urea exports.
Table 4-23: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (MMT CO2
Eq.)
Source
1990
2005
2014
2015
2016
2017
2018
Urea Consumption
3.8
3.7
1.8
4.6
5.1
3.8
3.6
Total	3.8	3.7	1.8 4.6 5.1 3.8 3.6
Table 4-24: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (kt)
Source
1990
2005
2014
2015
2016
2017
2018
Urea Consumption
3,784
: 3,653
1,807
4,578
5,132
3,769
3,628
Total
3,784
3,653
1,807
4,578
5,132
3,769
3,628
4-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Methodology
Emissions of CO2 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 CO2 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 CO2 during use, and consistent with the 2006IPCC Guidelines.
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) and is reported in Table 4-25, from the total domestic supply of urea. In previous Inventory reports, the
quantity of urea fertilizer applied to agricultural lands was obtained directly from the Cropland Remaining
Cropland section of the Land Use, Land Use Change, and Forestry chapter. The domestic supply of urea is
estimated based on the amount of urea produced plus the sum of net urea imports and exports. A factor of 0.733
tons of CO2 per ton of urea consumed is then applied to the resulting supply of urea for non-agricultural purposes
to estimate CO2 emissions from the amount of urea consumed for non-agricultural purposes. The 0.733 tons of CO2
per ton of urea emission factor is based on the stoichiometry of producing urea from ammonia and CO2. This
corresponds to a stoichiometric CCh/urea factor of 44/60, assuming complete conversion of NH3 and CChto urea
(IPCC 2006; EFMA 2000).
Urea production data for 1990 through 2008 were obtained from the Minerals Yearbook: Nitrogen (USGS 1994
through 2009a). Urea production data for 2009 through 2010 were obtained from the U.S. Census Bureau (2011).
The U.S. Census Bureau ceased collection of urea production statistics in 2011. Starting with the previous Inventory
(i.e., 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 2017 (EPA 2018).
For this public review draft of the current Inventory (i.e., 1990 through 2018), GHGRP data are not available and
urea production values for 2018 are proxied using 2017 values.
Urea import data for 2018 are not yet publicly available and so 2017 data have been used as proxy. Urea import
data for 2013 to 2017 were obtained from the 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 2018 are not yet publicly available and so 2017 data have been used as proxy. Urea export
data for 2013 to 2017 were obtained from the 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.
Industrial Processes and Product Use 4-33

-------
Table 4-25: Urea Production, Urea Applied as Fertilizer, Urea Imports, and Urea Exports (kt)
Year
Urea
Urea Applied
Urea
Urea
Production
as Fertilizer
Imports
Exports
1990
7,450
3,296
1,860
854
2005
5,270
4,779
5,026
536
2014
5,561
6,156
3,510
451
2015
5,880
6,447
7,190
380
2016
7,390
6,651
6,580
321
2017
7,390
6,888
5,510
872
2018
7,390
7,080
5,510
872
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 CO2 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 2018 were estimated to be
between 3.0 and 4.2 MMT CO2 Eq. at the 95 percent confidence level. This indicates a range of approximately 16
percent below and 16 percent above the emission estimate of 3.6 MMT CO2 Eq.
Table 4-26: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Urea
Consumption for Non-Agricultural Purposes (MMT CO2 Eq. and Percent)
Source Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Urea Consumption





for Non-Agricultural C02
3.6
3.0
4.2
-16%
+16%
Purposes





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 2018. Details on the emission trends through time are described in more detail in the Methodology
section, above.
4-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).30 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.31 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
This current Inventory (i.e., 1990 through 2018) has been updated to include more recent 2017 United States urea
imports and exports data. Utilizing updated values resulted in an approximately 24 percent decrease in 2017
emissions reported in the current Inventory (i.e., 1990 through 2018) compared to the year 2017 emissions from
the previous Inventory (i.e., 1990 through 2017). The previous Inventory relied on proxy data for imports and
exports for 2017, the updated data used in this Inventory resulted in lower imports and increased exports in 2017
which reduced consumption and emissions.
4.7 Nitric Acid Production (CRF Source
Category 2B2)
Nitrous oxide (N2O) is emitted during the production of nitric acid (HNO3), an inorganic compound used primarily
to make synthetic commercial fertilizers. It 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 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 2018, there were 32 active nitric acid production
plants, including one high-strength nitric acid production plant in the United States (EPA 2010; EPA 2018).
The basic process technology for producing nitric acid has not changed significantly over time. During this process,
N2O 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 (O2) in two stages. The overall reaction is:
4NH3 + 802 -> 4HNO:i + 4H2
30	See .
31	See .
Industrial Processes and Product Use 4-35

-------
Currently, the nitric acid industry controls emissions of NO and NO2 (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 N2O. However,
NSCR units 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 with NSCRs installed at
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 9.3 MMT CO2 Eq. (31 kt of N2O) in 2018 (see Table
4-27). Emissions from nitric acid production have decreased by 23 percent since 1990, while production has
increased by 8 percent over the same time period. Emissions have decreased by 35 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 MMT CP2 Eq. kt N2Q
1990	12.1	41
2005	11.3	38
2014	10.9	37
2015	11.6	39
2016	10.1	34
2017	9.3	31
2018	9.3	31
Methodology
Emissions of N2O were calculated using the estimation methods provided by the 2006IPCC Guidelines and a
country-specific method utilizing EPA's GHGRP. The 2006 IPCC Guidelines Tier 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 N2O emissions for 2010 through 2018.
2010 through 2018
Process N2O emissions and nitric acid production data were obtained directly from EPA's GHGRP for 2010 through
2018 by aggregating reported facility-level data (EPA 2018). 2017 values were used as proxy for 2018, as GHGRP
data for 2018 were not available at the time of this current draft. However, based on GHGRP FLIGHT data, the level
of emissions from nitric acid production in 2018 are consistent with 2017.
As of 2018, in the United States, all nitric acid facilities are required to report annual greenhouse gas emissions
data to EPA as per the requirements of the GHGRP. 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 and no Subpart V facilities have stopped reporting as a result of the provisions in 98.2(i)(l) or
98.2(i)(2). As of 2018, there were 32 facilities that reported to EPA, including the known single high-strength nitric
acid production facility in the United States (EPA 2018). All nitric acid (weak acid) facilities are required to calculate
process emissions using a site-specific emission factor developed through annual performance testing under
typical operating conditions or by directly measuring N2O emissions using monitoring equipment.32
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.
4-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
GHGRP nitric acid production data are utilized to develop weighted country-specific emission factors used to
calculate emissions estimates for the years 2010 to 2018. Based on aggregated nitric acid production data by
abatement type (i.e., with, without) provided by EPA's GHGRP, the percent of production values and associated
emissions of nitric acid with and without abatement technologies are calculated. These percentages are the basis
for developing the country-specific weighted emission factors which vary from year to year based on the amount
of nitric acid production with and without abatement technologies. To maintain consistency across the time series
and with the rounding approaches taken by other data sets, GHGRP nitric acid data are also rounded for
consistency.
1990 through 2009
Using GHGRP data for 2010,33 country-specific N2O emission factors were calculated for nitric acid production with
abatement and without abatement (i.e., controlled and uncontrolled emission factors), as previously stated. The
following 2010 emission factors were derived for production with abatement and without abatement: 3.3 kg
INhO/metric ton HNO3 produced at plants using abatement technologies (e.g., tertiary systems such as NSCR
systems) and 5.99 kg INhO/metric ton HNO3 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 N2O 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 also year over 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 N2O 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 N2O emission factors were used in conjunction with annual production to estimate
N2O emissions for 1990 through 2009, using the following equations:
El — Pi X EFWelgfr):eCl:l
EFweighted,i =	X EFc) + (%PUnc,i X EFunc)\
where,
Ei	= Annual N2O Emissions for year I (kg/yr)
Pi	= Annual nitric acid production for year I (metric tons HNO3)
EFweighted.i	= Weighted N2O emission factor for year I (kg INhO/metric ton HNO3)
%Pc,i	= Percent national production of HNO3 with N2O abatement technology (%)
EFc	= N2O emission factor, with abatement technology (kg INhO/metric ton HNO3)
%Punc,i	= Percent national production of HNO3 without N2O abatement technology (%)
EFunc	= N2O emission factor, without abatement technology (kg INhO/metric ton HNO3)
I	= year from 1990 through 2009
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 4-37

-------
•	For 2009: Weighted N2O emission factor = 5.46 kg N20/metric ton HNO3.
•	For 1990 through 2008: Weighted N2O emission factor = 5.66 kg N20/metric ton HNO3.
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
2014
7,660
2015
7,210
2016
7,810
2017
7,780
2018
7,780
Uncertainty and Time-Series Consistency
Uncertainty associated with the parameters used to estimate N2O 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.
At this time EPA does not estimate uncertainty of the aggregated facility-1 eve I information. As noted in the QA/QC
and verification section below, 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. The annual production reported by each nitric acid
facility under EPA's GHGRP and then aggregated to estimate national N2O 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 8.9 and 9.8 MMT CO2 Eq. at the 95 percent
confidence level. This indicates a range of approximately 5 percent below to 5 percent above the 2017 emissions
estimate of 9.3 MMT CO2 Eq.
4-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-29: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Nitric
Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Nitric Acid Production
N20
9.3
8.9 9.8
-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 2018.
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
Recalculations Discussion
No recalculations were performed for the 1990 through 2017 portion of the time series.
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. Longer-term, in 2020, EPA anticipates having information from GHGRP facilities on the
installation date of any N2O abatement equipment, per revisions finalized in December 2016 to EPA's GHGRP. This
information will enable more accurate estimation of N2O emissions from nitric acid production over the time
series.
34	See Subpart V monitoring and reporting regulation .
35	See GHGRP Verification Factsheet .
Industrial Processes and Product Use 4-39

-------
4.8 Adipic Acid Production (CRF Source
Category 2B3)
Adipic acid is produced through a two-stage process during which nitrous oxide (N2O) 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:
(iCH2)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 N2O 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 N2O abatement measures at adipic acid facilities (EPA 2019).
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 2018,
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).
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 9 percent over the period of 1990 through 2018, to
approximately 825,000 metric tons (ACC 2019). Nitrous oxide emissions from adipic acid production were
estimated to be 10.3 MMT CO2 Eq. (35 kt N2O) in 2018 (see Table 4-30). Over the period 1990 through 2018,
emissions have been reduced by 32 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 increased by
approximately 40 percent from GHGRP Reporting Year (RY) 2017 to RY2018 due to a significant change in
emissions from one facility. The facility confirmed that there was an increase in adipic acid production and a
decrease in the use of the N2O abatement device in RY2018, resulting in a large increase in greenhouse gas
emissions. As noted above, changes in control measures and abatement technologies at adipic 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 GHGRP.
Table 4-30: N2O Emissions from Adipic Acid Production (MMT CO2 Eq. and kt N2O)
Year MMT CP2 Eq. kt N2Q
1990	15.2	51
2005	7.1	24
2014	5.4	18
4-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
2015	4.3	14
2016	7.0	23
2017	7.4	25
2018	10.3	35
fviet had ©logy
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.
2010 through 2018
All emission estimates for 2010 through 2018 were obtained through analysis of GHGRP data (EPA 2010 through
2013; EPA 2014 through 2018; EPA 2019), 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 2018 (EPA 2010
through 2013; EPA 2014 through 2018; EPA 2019) and aggregated to national N2O 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 N2O emissions using monitoring equipment.36
1990 through 2009
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 N2O emissions were estimated using the following Tier 2 equation from the 2006
IPCC Guidelines:
Eaa = Qaa X EFaa X (1 - [DF X UF])
where,
Eaa —
N2O emissions from adipic acid production, metric tons
Qaa =
Quantity of adipic acid produced, metric tons
EFaa
Emission factor, metric ton INhO/metric ton adipic acid produced
DF
N2O destruction factor
UF
Abatement system utility factor
The adipic acid production is multiplied by an emission factor (i.e., N2O emitted per unit of adipic acid produced),
which has been estimated to be approximately 0.3 metric tons of N2O per metric ton of product (IPCC 2006). The
"N2O destruction factor" in the equation represents the percentage of N2O 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
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.
Industrial Processes and Product Use 4-41

-------
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 2019; 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 2018 were obtained from the American
Chemistry Council (ACC 2019).
Table 4-31: Adipic Acid Production (kt)
Year
kt
1990
755
2005
865
2014
1,025
2015
1,055
2016
860
2017
830
2018
825
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2018.
Uncertainty and Time Serii insistency
Uncertainty associated with N2O 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 Table 4-32. Nitrous oxide
emissions from adipic acid production for 2018 were estimated to be between 9.8 and 10.8 MMT CO2 Eq. at the 95
percent confidence level. These values indicate a range of approximately 5 percent below to 5 percent above the
2018 emission estimate of 10.3 MMT CO2 Eq.
4-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-32: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Adipic
Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)
(MMT CO? Eq.)
(%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Adipic Acid Production
N20
10.3
9.8 10.8
-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 2018.
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 2017 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 4-43

-------
4.9 Caprolactam, Glyoxal and Glyoxylic Acid
Production (CRF Source Category 2B4)
Caprolactam
Caprolactam (CsHnNO) is a colorless monomer produced for nylon-6 fibers and plastics, with a substantial
proportion of the fiber used in carpet manufacturing. Commercial processes for the manufacture of caprolactam
are based on either toluene or benzene. The production of caprolactam can give rise to significant emissions of
nitrous oxide (N2O).
During the production of caprolactam, emissions of N2O can occur from the ammonia oxidation step, emissions of
carbon dioxide (CO2) from the ammonium carbonate step, emissions of sulfur dioxide (SO2) from the ammonium
bisulfite step, and emissions of non-methane volatile organic compounds (NMVOCs). Emissions of CO2, SO2 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 (CsHioO). The classical route (Raschig process) and basic reaction
equations for production of caprolactam from cyclohexanone are (IPCC 2006):
Oxidation of NH3 to NO/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 ({NH2OH)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^NO. H2S04 (+4NH3 and H20) -> C^NO + 2(NH4)2S04
In 1999, there were four caprolactam production facilities in the United States. As of 2018, the United States had
three companies that produce caprolactam with a total of three caprolactam production facilities: AdvanSix in
Virginia (AdvanSix 2019), BASF in Texas (BASF 2019), and Fibrant LLC in Georgia (Fibrant 2019; TechSci n.d. 2017).
4-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Nitrous oxide emissions from caprolactam production in the United States were estimated to be 1.4 MMT CO2 Eq.
(5 kt N2O) in 2018 (see Table 4-33). National emissions from caprolactam production decreased by approximately
15 percent over the period of 1990 through 2018. Emissions in 2018 decreased by approximately 3 percent from
the 2017 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
7
6
6
5
5
2014	2.0
2015	1.9
2016	1.7
2017	1.5
2018	1.4
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 N2O is emitted in the process of oxidation of
acetaldehyde.
Glyoxal (ethanedial) (C2H2O2) is produced from oxidation of acetaldehyde (ethanal) (C2H4O) with concentrated
nitric acid (HNO3). 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.
Emissions of N2O 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 2018, as shown in this formula:
EN2o = EF x CP
where,
Enzo	= Annual N2O Emissions (kg)
EF	= N2O emission factor (default) (kg N20/metric ton caprolactam produced)
CP	= Caprolactam production (metric tons)
During the caprolactam production process, N2O 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 4-45

-------
amount of N2O emissions can be estimated based on the chemical reaction shown above. Based on this formula,
which is consistent with an IPCCTier 1 approach, approximately 111.1 metric tons of caprolactam are required to
generate one metric ton of N2O, resulting in an emission factor of 9.0 kg N2O 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 N2O 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 N2O emissions, such as those
employed at nitric acid plants. Thus, the 530 thousand metric tons (kt) of caprolactam produced in 2018 (ACC
2019) resulted in N2O emissions of approximately 1.4 MMT CO2 Eq. (5 kt).
The activity data for caprolactam production (see Table 4-34) from 1990 to 2018 were obtained from the American
Chemistry Council's Guide to the Business of Chemistry report (ACC 2019). 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
2014	755
2015	700
2016	640
2017	545
2018	530
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-Seri insistency
Estimation of emissions of N2O from caprolactam production can be treated as analogous to estimation of
emissions of N2O from nitric acid production. Both production processes involve an initial step of NH3 oxidation,
which is the source of N2O 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 2018 were estimated to be between 1.0
and 1.9 MMT CO2 Eq. at the 95 percent confidence level. These values indicate a range of approximately 32
percent below to 32 percent above the 2018 emission estimate of 1.4 MMT CO2 Eq.
4-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-35: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from
Caprolactam, Glyoxal and Glyoxylic Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Caprolactam Production
N20
1.4
1.0 1.9
-32% +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 2018. 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
Revisions to historical activity data were available from ACC and incorporated which resulted in changes to
emissions estimates for previous years in the time series. The updates resulted in a decrease of 0.2 MMT CO2 Eq. in
2015, decrease of 0.3 MMT CO2 Eq. in 2016, and an increase of 0.1 MMT CO2 Eq. in 2017.
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 comment period for the current Inventory
report, EPA continued 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 (CO2) and methane (CH4) are emitted from the production of silicon carbide (SiC), a material
produced 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 CO2, Cm, 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)
Industrial Processes and Product Use 4-47

-------
Carbon dioxide and Cm 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. However, as noted in Annex 5 to this report, Cm emissions from calcium
carbide production are not estimated as 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
other non-abrasive applications, primarily in iron and steel production (USGS 1991a through 2015). 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. However, 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 2016 in the United States (USGS 2018a).
Carbon dioxide emissions from SiC production and consumption in 2018 were 0.2 MMT CO2 Eq. (189 kt CO2) (see
Table 4-36 and Table 4-37). Approximately 49 percent of these emissions resulted from SiC production while the
remainder resulted from SiC consumption. Methane emissions from SiC production in 2018 were 0.01 MMT CO2
Eq. (0.4 kt CH4) (see Table 4-36 and Table 4-37). Emissions have not fluctuated greatly in recent years, but 2018
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
2014
2015
2016
2017
2018
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
1990
2005
2014
2015
2016
2017
2018
C02
375
219
173
180
174
186
189
ch4
1
+
+
+
+
+
+
+ Does not exceed 0.5 kt
Methodology
Emissions of CO2 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 ~ EFSCŁ02 * Qsc
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.
4-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
/I metric ton\
Esc,CH4 = EFSCjCH4 x Qsc x 1000 kg )
where,
Esc,C02
EFsc,C02
Esc,CH4
EFsc,CH4
CO2 emissions from production ofSiC, metric tons
Emission factor for production of SiC, metric ton CCh/metric ton SiC
Quantity of SiC produced, metric tons
Cm emissions from production of SiC, metric tons
Emission factor for production of SiC, kilogram CHVmetric ton SiC
Emission factors were taken from the 2006IPCC Guidelines:
•	2.62 metric tons CCh/metric ton SiC
•	11.6 kg CHVmetric ton SiC
Production data for metallurgical and other non-abrasive applications of silicon carbide is not available; therefore,
both CO2 and CH4 estimates for silicon carbide are based solely upon production data for silicon carbide for
industrial abrasive applications.
SiC industrial abrasives production data for 1990 through 2013 were obtained from the 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 were
obtained from the Mineral Industry Surveys, Manufactured Abrasives in the First Quarter 2019, Table 1, July 2019
(USGS 2019a). 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 silicon carbide
provided by the U.S. Census Bureau (2005 through 2019) (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 CO2 from silicon carbide consumption for metallurgical uses were calculated by multiplying the annual
utilization ofSiC for metallurgical uses (reported annually in the USGS Minerals Yearbook: Silicon) by the carbon
content ofSiC (31.5 percent), which was determined according to the molecular weight ratio ofSiC.
Emissions of CChfrom silicon carbide consumption for other non-abrasive uses were calculated by multiplying the
annual SiC consumption for non-abrasive uses by the carbon content ofSiC (31.5 percent). The annual SiC
consumption for non-abrasive uses was calculated by multiplying the annual SiC consumption (production plus net
imports) by the percent used in metallurgical and other non-abrasive uses (50 percent) (USGS 1991a through 2015)
and then subtracting the SiC consumption for metallurgical use.
The petroleum coke portion of the total CO2 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 CO2 from Fossil Fuel Combustion (Section 3.1) and Annex 2.1, Methodology for
Estimating Emissions of CO2 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
2014	35,000	140,733
Industrial Processes and Product Use 4-49

-------
2015
35,000
153,475
2016
35,000
142,104
2017
35,000
163,492
2018
35,000
168,531
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
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 Cm, 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 Cm 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 CO2 emissions from 2017 were estimated to be between 10 percent below and 9
percent above the emission estimate of 0.19 MMT CO2 Eq. at the 95 percent confidence level. Silicon carbide
production Cm emissions were estimated to be between 9 percent below and 9 percent above the emission
estimate of 0.01 MMT CO2 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
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Silicon Carbide Production
and Consumption
C02
0.19
0.17
0.21
-10%
+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 2018. Details on the emission trends through time are described in more detail in the Methodology
section above.
jfkl 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
No recalculations were performed for the 1990 through 2017 portion of the time series.
4-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
4.11 Titanium Dioxide Production (CRF Source
Category 2B6)
Titanium dioxide (TiCh) 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
(CO2). 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:
The sulfate process does not use petroleum coke or other forms of carbon as a raw material and does not emit
CO2.
The C in the first chemical reaction is provided by petroleum coke, which is oxidized in the presence of the chlorine
and FeTiC>3 (rutile ore) to form CO2. Since 2004, all TiCh 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 TiCh 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 2018, U.S. TiC>2 production totaled 1,200,000 metric tons
(USGS 2019). There were a total five plants producing TiC>2 in the United States in 2018.
Emissions of CO2 from titanium dioxide production in 2018 were estimated to be 1.5 MMT CO2 Eq. (1,541 kt CO2),
which represents an increase of 29 percent since 1990 (see Table 4-40). Compared to 2017, emissions from
titanium dioxide production decreased by 9 percent in 2018 due to a 95 percent decrease in production.
Table 4-40: CO2 Emissions from Titanium Dioxide (MMT CO2 Eq. and kt)
Year MMT CP2 Eq.	kt
1990	12	1,195
2005	1.8	1,755
2014	1.7	1,688
2015	1.6	1,635
2016	1.7	1,662
2017	1.7	1,688
2018	1.5	1,541
Emissions of CO2 from TiC>2 production were calculated by multiplying annual national TiC>2 production by chloride
process-specific emission factors using a Tier 1 approach provided in 2006IPCC Guidelines. The Tier 1 equation is
as follows:
2FeTi03 + 7CZ2 + 3C —> 2TiCl^ + 2FeCl^ + 3C02
2TiCl4 + 2(?2 ~~* 2Ti02 ~l~ 4CZ2
Methodology
Etd = EFtd X Qtd
where,
Etd
EFtd
Qtd
CO2 emissions from TiC>2 production, metric tons
Emission factor (chloride process), metric ton CCh/metric ton TiC>2
Quantity of Ti02 produced
Industrial Processes and Product Use 4-51

-------
The petroleum coke portion of the total CO2 process emissions from TiC>2 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 CChfrom Fossil Fuel Combustion (Section 3.1 Fossil Fuel Combustion) and Annex 2.1, Methodology for
Estimating Emissions of CO2 from Fossil Fuel Combustion.
Data were obtained for the total amount of TiC>2 produced each year. For years prior to 2004, it was assumed that
TiC>2 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 post-2004 production uses the chloride process (USGS 2005). The percentage of
production from the chloride process is estimated at 100 percent since 2004. An emission factor of 1.34 metric
tons CCh/metric ton TiC>2 was applied to the estimated chloride-process production (IPCC 2006). It was assumed
that all TiC>2 produced using the chloride process was produced using petroleum coke, although some TiC>2 may
have been produced with graphite or other carbon inputs.
The emission factor for the TiC>2 chloride process was taken from the 2006 IPCC Guidelines. Titanium dioxide
production data and the percentage of total TiC>2 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 2018 were obtained from the
Minerals Commodity Summary: Titanium and Titanium Dioxide (USGS 2020).40 Data on the percentage of total TiC>2
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
2014	1,260
2015	1,220
2016	1,240
2017	1,260
2018	1,150
Uncertainty and Time-Series Consistency
Each year, the USGS collects titanium industry data for titanium mineral and pigment production operations. If
TiC>2 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 TiC>2 pigment plants over
the time series.
Although some TiC>2 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 CChper unit of TiC>2 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
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.
4-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
agent used in each process rather than on the amount of TiC>2 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 TiCh 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 TiC>2 production capacity that
was attributed to the chloride process was multiplied by total TiCh production to estimate the amount of TiC>2
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 TiC>2 production, literature data
were used for petroleum coke composition. Certain grades of petroleum coke are manufactured specifically for
use in the TiC>2 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 CO2 emissions from 2018 were estimated to be between 1.3 and 1.7 MMT CO2 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 CO2 Eq.
Table 4-42: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Titanium
Dioxide Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(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 2018. 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
No recalculations were performed for the 1990 through 2017 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 4-53

-------
4.12 Soda Ash Production (CRF Source
Category 2B7)
Carbon dioxide (CO2) is generated as a byproduct of calcining trona ore to produce soda ash, and is eventually
emitted into the atmosphere. In addition, CO2 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 CO2 during trona-based production is based on the
following reaction:
2Na2C03 ¦ NaHC03 ¦ 2H20(Trona) -» 3Na2C03(Soda Ash) + 5H20 +C02
Soda ash (sodium carbonate, Na2COs) 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 2018c). 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 2019a). Only two states
produce natural soda ash: Wyoming and California. Of these two states, net emissions of COzfrom soda ash
production were only calculated for Wyoming, due to specifics regarding the production processes employed in
the state.42 Based on 2018 reported data, the estimated distribution of soda ash by end-use in 2018 (excluding
glass production) was chemical production, 56 percent; wholesale distributors (e.g., for use in agriculture, water
treatment, and grocery wholesale), 12 percent; soap and detergent manufacturing, 11 percent; other uses, 10
percent; flue gas desulfurization, 7 percent; pulp and paper production, 2 percent, and water treatment, 2 percent
(USGS 2019).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, which surpassed the
United States in soda ash production in 2003, is 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.
4-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
In 2018, CO2 emissions from the production of soda ash from trona ore were 1.7 MMT CO2 Eq. (1,714 kt CO2) (see
Table 4-43). Total emissions from soda ash production in 2018 decreased by approximately 2 percent from
emissions in 2017, and have increased by approximately 20 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 2018 since experiencing a decline in domestic and
export sales caused by adverse global economic conditions in 2009, although production dropped slightly in 2018
relative to the prior year.
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
2014	1.7	1,685
2015	1.7	1,714
2016	1.7	1,723
2017	1.8	1,753
2018	1.7	1,714
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 CO2, or an emission factor
of 0.0974 metric tons CO2 per metric ton of trona ore (IPCC 2006). Thus, the 17.6 million metric tons of trona ore
mined in 2018 for soda ash production (USGS 2019) resulted in CO2 emissions of approximately 1.7 MMT CO2 Eq.
(1,714 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 CO2 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).
The activity data for trona ore production (see Table 4-44) for 1990 through 2018 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, 2018b, 2019). 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.
However, 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 4-55

-------
Table 4-44: Soda Ash Production (kt)
Year
Production3
1990
14,700
2005
17,000
2014
17,300
2015
17,600
2016
17,700
2017
18,000
2018
17,600
a Soda ash produced from trona ore only.
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
2018b).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-45. Soda Ash Production
CO2 emissions for 2018 were estimated to be between 1.5 and 1.8 MMT CO2 Eq. at the 95 percent confidence
level. This indicates a range of approximately 9 percent below and 8 percent above the emission estimate of 1.7
MMTCCh Eq.
Table 4-45: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Soda Ash
Production (MMT CO2 Eq. and Percent)
Source

2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3

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



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Soda Ash Production
C02
1.7
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 2018.
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).
4-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Recalculations Discussion
No recalculations were performed for the 1990 through 2017 portion of the time series.
Planned Improvements
EPA plans to use GHGRP data for conducting category-specific QC of emission estimates consistent with both
Volume 1, Chapter 6 of the 2006IPCC 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 small amounts of carbon dioxide (CO2) 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 Cm 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 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 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 also produces byproduct CO2, carbon monoxide (CO), and water from
the direct oxidation of the propylene feedstock, and produces other hydrocarbons from side reactions in the
ammoxidation process.
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 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 CO2 and uncombusted CH4 emissions are released from thermal incinerators used as control devices,
process dryers, and equipment leaks. Carbon black is also produced in the United States by the thermal cracking of
46 See .
Industrial Processes and Product Use 4-57

-------
acetylene-containing feedstocks (i.e., acetylene black process), by the thermal cracking of other hydrocarbons (i.e.,
thermal black process), and by the open burning of carbon black feedstock (i.e., lamp black process); each of these
processes is used at only one U.S. plant (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:
C2H6 -» C2H4 + H2
Small amounts of Cm are also generated from the steam cracking process. In addition, CO2 and Cm 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 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 + i02 + 2HCI -» C2H4Cl2 + 2H20 (oxychlorination)
C2H4 + 302 -» 2C02 + 2H20 (direct oxidation of ethylene during oxychlorination)
In addition to the byproduct CO2 produced from the direction oxidation of the ethylene feedstock, CO2 and Cm
emissions are also generated from combustion units.
Ethylene oxide (C2H4O) 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 CO2 from
the direct oxidation of the ethylene feedstock is removed from the process vent stream using a recycled carbonate
solution, and the recovered CO2 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 CO2 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 CO2) 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 CO2 and CH4 from petrochemical production in 2018 were 29.4 MMT CO2 Eq. (29,424 kt CO2) and 0.3
MMT CO2 Eq. (12 kt CH4), respectively (see Table 4-46 and Table 4-47). Since 1990, total CO2emissions from
petrochemical production increased by 36 percent. Methane emissions from petrochemical (methanol and
acrylonitrile) production reached a low of 1.8 kt Cm in 2011, given declining methanol production; however, Cm
emissions have been increasing every year since 2011 and are now 38 percent greater than in 1990 (though still
less than the peak in 1997) due to a rebound in methanol production.
4-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-46: CO2 and ChU Emissions from Petrochemical Production (MMT CO2 Eq.)
Year
1990
2005
2014
2015
2016
2017
2018
C02
21.6
27.4
26.3
28.1
28.3
28.9
29.4
ch4
O
k>
0.1
0.1
0.2
0.2
0.3
0.3
Total
21.8
27.5
26.4
28.2
28.6
29.2
29.7
Note: Totals may not sum due to independent rounding.




ible 4-47:
CO2 and CH4
Emissions from Petrochemical Production (kt)

Year
1990
2005
2014
2015
2016
2017
2018
C02
21,611
27,383
26,254
28,062
28,310
28,910
29,424
ch4
9
3
5
7
10
10
12
Note: Totals may not sum due to independent rounding.
Methodology
Emissions of CO2 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 CO2
and Cm emissions from production of acrylonitrile and methanol,47 and a country-specific approach similar to the
IPCC Tier 2 method was used to estimate CO2 emissions from production of carbon black, ethylene oxide, ethylene,
and ethylene dichloride. The Tier 2 method for petrochemicals is a total feedstock C mass balance method used to
estimate total CO2 emissions, but 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
CO2. Further, the guideline states that while the total C mass balance method estimates total C emissions from the
process but does not directly provide an estimate of the amount of the total C emissions emitted as CO2, CH4, or
non-CH4 volatile organic compounds (NMVOCs). This method accounts for all the C as CO2, including CH4.
Note, a small subset of facilities reporting under EPA's GHGRP use Continuous Emission Monitoring Systems
(CEMS) to monitor CO2 emissions from process vents and/or stacks from stationary combustion units, these
facilities are required to also report CO2, CH4 and N2O emissions from combustion of process off-gas in flares. The
CO2 from flares are included in aggregated CO2 results. Preliminary analysis of aggregated annual reports shows
that flared CH4 and N2O emissions are less than 500 kt CO2 Eq./year. EPA's GHGRP is still reviewing this 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 2018
Carbon dioxide emissions and national production were aggregated directly from EPA's GHGRP dataset for 2010
through 2018 (EPA 2019). In 2018, data reported to the GHGRP included CO2 emissions of 3,400,000 metric tons
from carbon black production; 19,500,000 metric tons of CChfrom ethylene production; 480,000 metric tons of
CO2 from ethylene dichloride production; and 1,310,000 metric tons of CO2 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 CO2 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
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 4-59

-------
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 CO2 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 CO2. 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 CO2 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.
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 CChfrom Fossil Fuel Combustion (3.1 Fossil Fuel Combustion (IPCC
Source Category 1A)) and Annex 2.1, Methodology for Estimating Emissions of CO2 from Fossil Fuel Combustion.
1990 through 2009
Prior to 2010, for each of these 4 types of petrochemical processes, an average national CO2 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 CO2 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 CO2 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. 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 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 CO2 emission factors that were calculated from the GHGRP data are as follows:
•	2.59 metric tons CCh/metric ton carbon black produced
•	0.79 metric tons CCh/metric ton ethylene produced
•	0.040 metric tons CCh/metric ton ethylene dichloride produced
48 A few facilities producing ethylene dichloride and ethylene used C02 CEMS, those C02 emissions have been included in the
aggregated GHGRP emissions presented here. For ethylene production processes, nearly all process emissions are from the
combustion of process off-gas. Under EPA's GHGRP, Subpart X, ethylene facilities can report C02 emissions from burning of
process gases using the optional combustion methodology for ethylene production processes, which requires estimating
emissions based on fuel quantity and carbon contents of the fuel. This is consistent with the 2006 IPCC Guidelines (p. 3.57)
which recommends including combustion emissions from fuels obtained from feedstocks (e.g., off-gases) in petrochemical
production under in the IPPU sector. In 2014, for example, this methodology was used by more than 20 of the 65 reporting
facilities. In addition to C02, these facilities are required to report emissions of CH4 and 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 so not significant enough to
prioritize for inclusion in the report at this time. Pending resources and significance, EPA may include these emissions in future
reports to enhance completeness.
4-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
•	0.46 metric tons CCh/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 and ethylene
dichloride for 1990 through 2009 were obtained from the American Chemistry Council's (ACC's) Guide to the
Business of Chemistry (ACC 2002, 2003, 2005 through 2011). Annual production data for ethylene oxide were
obtained from ACC's U.S. Chemical Industry Statistical Handbook for 2003 through 2009 (ACC 2014a) and from
ACC's Business of Chemistry for 1990 through 2002 (ACC 2014b).
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 1CO2 and Cm
emission factors to estimate emissions for 1990 through 2018. Emission factors used to estimate acrylonitrile
production emissions are as follows:
•	0.18 kg Cm/metric ton acrylonitrile produced
•	1.00 metric tons CCh/metric ton acrylonitrile produced
Annual acrylonitrile production data for 1990 through 2018 were obtained from ACC's Business of Chemistry (ACC
2019).
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 1CO2 and Cm emission
factors to estimate emissions for 1990 through 2018. Emission factors used to estimate methanol production
emissions are as follows:
•	2.3 kg CHVmetric ton methanol produced
•	0.67 metric tons CCh/metric ton methanol produced
Annual methanol production data for 1990 through 2018 were obtained from the ACC's Business of Chemistry (ACC
2019).
Table 4-48: Production of Selected Petrochemicals (kt)
Chemical
1990
2005
2014
2015
2016
2017
2018
Carbon Black
1,307
1,651
1,210
1,220
1,190
1,240
1,280
Ethylene
16,542
j 23,975
25,500
26,900
26,600
27,800
30,500
Ethylene Dichloride
6,283
; 11,260
11,300
11,300
11,700
12,400
12,500
Ethylene Oxide
2,429
3,220
3,160
3,240
3,270
3,350
3,280
Acrylonitrile
1,214
1,325
1,095
1,050
955
1,040
1,250
Methanol
3,750
1,225
2,105
3,065
4,250
4,295
5,200
As noted earlier in the introduction section of the Petrochemical Production chapter, 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 2006 IPCC Guidelines. According to the 2006 IPCC 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
Industrial Processes and Product Use 4-61

-------
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 2006IPCC 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 Cm and CO2 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 CO2 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 CO2 emissions from 2018 were estimated to be between 27.8 and 31.1 MMT CO2 Eq. at the 95 percent
confidence level. This indicates a range of approximately 6 percent below to 6 percent above the emission
estimate of 29.4 MMT CO2 Eq. Petrochemical production CH4 emissions from 2018 were estimated to be between
0.11 and 0.37 MMT CO2 Eq. at the 95 percent confidence level. This indicates a range of approximately 57 percent
below to 46 percent above the emission estimate of 0.3 MMT CO2 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
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3


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

(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Petrochemical
Production
C02
29.4
27.8
31.1
-6%
+6%
Petrochemical
Production
ch4
0.30
0.11
0.37
-57%
+46%
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 2018.
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
49 See .
4-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 emissions calculated using the
GHGRP data to the CO2 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 when the difference was 18 percent). 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 CO2, CH4, and
N2O 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, CO2 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 CO2.
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 CO2 emissions from the ethylene process units).
EPA continues to assess the GHGRP data for ways to better disaggregate the data and incorporate it into the
inventory.
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
As previously noted above, EPA's GHGRP data are used to develop CO2 emission factors for carbon black, ethylene,
ethylene dichloride, and ethylene oxide production. These factors are used with production data to estimate CO2
50	See .
51	See .
Industrial Processes and Product Use 4-63

-------
emissions from production of these petrochemicals in 1990 through 2009. In previous Inventories, average
emission factors were developed from all years of available GHGRP data. Based on a review of the
representativeness of GHGRP data for more recent years, the emission factor for the above-mentioned
petrochemical types in the current Inventory has been updated to reflect GHGRP data only from 2010 through
2013 as these years are more representative of operations from 1990 through 2009. This resulted in an average
annual increase in total petrochemical emissions of about 1 percent compared to the previous (i.e., 1990 to 2017)
Inventory.
The previous Inventory used proxy data for 2017 production and emissions values for carbon black, ethylene,
ethylene dichloride and ethylene oxide as GHGRP data for 2017 was not available. The 2017 data for production
and emissions from those sources has been updated with the GHGRP data for 2017 for this Inventory. It resulted in
a 2 percent increase in total petrochemical emissions for 2017 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 N2O 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 CO2 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 CO2 estimates of non-energy use of fuels in the energy sector and CO2 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
and the non-energy use estimates are roughly 20 percent of the emissions captured here. 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.
4.14 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
4-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (CHCb) and hydrogen fluoride (HF) in the presence of a catalyst,
SbCls. 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 (CHCbF), 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 2018. Emissions of HFC-23 from this activity in 2018 were
estimated to be 3.3 MMT CO2 Eq. (0.2 kt) (see Table 4-50). This quantity represents a 36 percent decrease from
2017 emissions and a 93 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 decrease from
2017 emissions was caused both by a decrease in the HFC-23 emission rate and by a decrease in HCFC-22
production. 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.
Table 4-50: HFC-23 Emissions from HCFC-22 Production (MMT COz Eq. and kt HFC-23)
Year
MMT CO? Eq.
kt HFC-23
1990
46.1
3
2005
20.0
1
2014
5.0
0.3
2015
4.3
0.3
2016
2.8
0.2
2017
5.2
0.3
2018
3.3
0.2
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 2018 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
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 4-65

-------
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 2018
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.
Table 4-51: HCFC-22 Production (kt)
Year
kt
1990
139
2005
156
/
2012
96
2013-2018
C
C(CBI)
Note: HCFC-22 production in 2013 through 2018
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 2018. 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 2018 (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
4-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 3.1 and 3.6 MMT CO2 Eq. at the 95 percent confidence
level. This indicates a range of approximately 7 percent below and 10 percent above the emission estimate of 3.3
MMT CO2 Eq.
Table 4-52: Approach 2 Quantitative Uncertainty Estimates for HFC-23 Emissions from
HCFC-22 Production (MMT CO2 Eq. and Percent)
Source
2018 Emission Estimate
aS (MMT C02 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)


Lower Upper
Bound Bound
Lower Upper
Bound Bound
HCFC-22 Production
HFC-23 3.3
3.1 3.6
-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 2018. See Methods discussion of this 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). 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.
Recalculations Discussion
A recent review of the time series of HFC-23 emissions from HCFC-22 production found small errors for the values
for 2014 and 2017. For these two years, HFC-23 emissions from a facility that does not produce HCFC-22 had been
inadvertently included in the total, leading to overestimates by 750 and 85 metric tons of CO2 Eq., respectively.
The revised time series excludes the emissions from this facility for all years.
53 EPA (2015). Greenhouse Gas Reporting Program Report Verification. Available online at
.
Industrial Processes and Product Use 4-67

-------
4.15 Carbon Dioxide Consumption (CRF Source
Category 2610}
Carbon dioxide (CO2) 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). CO2 used for EOR is injected underground to enable additional petroleum to be
produced. For the purposes of this analysis, CO2 used in commercial applications other than EOR is assumed to be
emitted to the atmosphere. Carbon dioxide used in EOR applications is discussed in the Energy chapter under
"Carbon Capture and Storage, including Enhanced Oil Recovery" and is not discussed in this section.
Carbon dioxide is produced from naturally-occurring CO2 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 CO2 as a component. Only CO2
produced from naturally occurring CO2 reservoirs, and as a byproduct from energy and industrial processes, and
used in industrial applications other than EOR is included in this analysis. Carbon dioxide captured from biogenic
sources (e.g., ethanol production plants) is not included in the Inventory. Carbon dioxide captured from crude oil
and gas production is used in EOR applications and is therefore reported in the Energy chapter.
Carbon dioxide is produced as a byproduct of crude oil and natural gas production. This CO2 is separated from the
crude oil and natural gas using gas processing equipment, and may be emitted directly to the atmosphere, or
captured and reinjected into underground formations, used for EOR, or sold for other commercial uses. A further
discussion of CO2 used in EOR is described in the Energy chapter in Box 3-6 titled "Carbon Dioxide Transport,
Injection, and Geological Storage."
In 2018, the amount of CO2 produced and captured for commercial applications and subsequently emitted to the
atmosphere was 4.5 MMT CO2 Eq. (4,471 kt) (see Table 4-53). This is consistent with 2014 through 2018 levels and
is an increase of approximately 204 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
2014	4.5	4,471
2015	4.5	4,471
2016	4.5	4,471
2017	4.5	4,471
2018	4.5	4,471
fviethiMlolGgy
Carbon dioxide emission estimates for 1990 through 2018 were based on the quantity of CO2 extracted and
transferred for industrial applications (i.e., non-EOR end-uses). Some of the CO2 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 CO2 production used in commercial applications other than EOR is eventually
released into the atmosphere.
4-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
2010 through 2018
For 2010 through 2018, 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 2019). However, for the years 2015 through 2018,
GHGRP Subpart PP values did not pass GHGRP confidential business information (CBI) criteria for data aggregation.
Facilities report CO2 extracted or produced from natural reservoirs and industrial sites, and CO2 captured from
energy and industrial processes and transferred to various end-use applications to EPA's GHGRP. This analysis
includes only reported CO2 transferred to food and beverage end-uses. EPA is continuing to analyze and assess
integration of CO2 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 CO2 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 CO2 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 2018 is much higher (ranging from 44 to
48) 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 CO2
transferred to end-use applications from naturally occurring CO2 reservoirs and excludes industrial sites.
As previously mentioned, data from EPA's GHGRP (Subpart PP) was unavailable for use for the years 2015 through
2018 due to data confidentiality reasons. As a result, the emissions estimates for 2015 through 2018 have been
held constant from 2014 levels to avoid disclosure of proprietary information. EPA continues to evaluate options
for utilizing GHGRP data to update these values for future Inventories. Additional information on evaluating
GHGRP Subpart PP data is included in the Planned Improvements section.
1990 through 2009
For 1990 through 2009, data from EPA's GHGRP are not available. For this time period, CO2 production data from
four naturally-occurring CO2 reservoirs were used to estimate annual CO2 emissions. These facilities were Jackson
Dome in Mississippi, Brave and West Bravo Domes in New Mexico, and McCallum Dome in Colorado. The facilities
in Mississippi and New Mexico produced CO2 for use in both EOR and in other commercial applications (e.g.,
chemical manufacturing, food production). The fourth facility in Colorado (McCallum Dome) produced CO2 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 CO2 production in units of MMCF CO2 per day for
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 (ARI 1990 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
54 See .
Industrial Processes and Product Use 4-69

-------
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
Total C02
% Non-

MS
NM
NMCOz
Dome, CO
Production from
EOR3

C02 Production
CO2 Production
Production
C02 Production
Extraction and


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





EOR)
Facilities (kt)

1990
1,344 (100%)
63 (1%)
+
65 (100%)
NA
NA
2005
1,254 (27%)
58 (1%)
+
63(100%)
NA
NA
2014
NA
NA
NA
NA
72,000b
6%
2015
NA
NA
NA
NA
72,000b
6%
2016
NA
NA
NA
NA
72,000b
6%
2017
NA
NA
NA
NA
72,000b
6%
2018
NA
NA
NA
NA
72,000b
6%
+ Does not exceed 0.5 percent.
NA (Not Available)
a Includes only food & beverage applications.
b For 2010 through 2018, the publicly available GHGRP data were aggregated at the national level. From 2010 through 2014,
those aggregated values based GHGRP CBI criteria. For 2015 through 2018, values were held constant with those from
2014. Facility-level data are not publicly available from EPA's GHGRP.
Uncertainty and Time-Seri insistency
There is uncertainty associated with the data reported through EPA's GHGRP. Specifically, there is uncertainty
associated with the amount of CO2 consumed for food and beverage applications given a threshold for reporting
under GHGRP applicable to those reporting under Subpart PP, in addition to the exclusion of the amount of CO2
transferred to all other end-use categories. This latter category might include CO2 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 CO2 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 CO2 emissions for 2018 were estimated to be between 4.2 and 4.7 MMT CO2 Eq. at the 95 percent
confidence level. This indicates a range of approximately 5 percent below to 5 percent above the emission
estimate of 4.5 MMT CO2 Eq.
55 See .
4-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-55: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from CO2
Consumption (MMT CO2 Eq. and Percent)
Source Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)


Lower Upper
Lower Upper


Bound Bound
Bound Bound
C02 Consumption C02
4.5
4.2 4.7
-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 2018.
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 CO2 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
No recalculations were performed for the 1990 through 2017 portion of the time series.
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 In addition, EPA is also investigating the possibility of utilizing only extraction facility Subpart
PP data, while also updating the values for 2015 through 2018.
These improvements, in addition to updating the time series when new data is available, are still in process and
will be incorporated into future Inventory reports. These are near- to medium-term improvements.
56	See .
57	See .
58	See .
Industrial Processes and Product Use 4-71

-------
4.16 Phosphoric Acid Production (CRF Source
Category 2610}
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 (CO2) 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, as well as in Idaho and Utah, and is used primarily as a raw material for wet-process phosphoric acid
production (USGS 2018). The composition of natural phosphate rock varies depending upon 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 also may contain organic carbon. The calcium carbonate component of the
phosphate rock is integral to the phosphate rock chemistry. Phosphate rock can also contain organic carbon that is
physically incorporated into the mined rock but is not an integral component of the phosphate rock chemistry.
The phosphoric acid production process involves chemical reaction of the calcium phosphate (CasfPCUh)
component of the phosphate rock with sulfuric acid (H2SO4) and recirculated phosphoric acid (H3PO4) (EFMA 2000).
However, the generation of CO2 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 2018 was an estimated 23.0 million metric tons (USGS 2019). Total
imports of phosphate rock to the United States in 2018 were estimated to be approximately 3.0 million metric tons
(USGS 2019). Between 2014 and 2017, most of the imported phosphate rock (68 percent) came from Peru, with 31
percent from Morocco and 1 percent from other sources (USGS 2019). All phosphate rock mining companies in the
U.S. are vertically integrated with fertilizer plants that produce phosphoric acid located near the mines. Some
additional phosphoric acid production facilities that used imported phosphate rock are located in Louisiana.
Over the 1990 to 2018 period, domestic phosphoric acid production has decreased by nearly 53 percent. Total CO2
emissions from phosphoric acid production were 0.9 MMT CO2 Eq. (940 kt CO2) in 2018 (see Table 4-56). Domestic
consumption of phosphate rock in 2018 was estimated to have decreased 9 percent relative to 2017 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
2014
1.0
1,037
2015
1.0
999
2016
1.0
998
2017
1.0
1,028
2018
0.9
940
Methodology
Carbon dioxide emissions from production of phosphoric acid from phosphate rock are estimated by multiplying
the average amount of inorganic carbon (expressed as CO2) contained in the natural phosphate rock as calcium
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:
4-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Epa Cpr X Qpr
where,
Epa =	CO2 emissions from phosphoric acid production, metric tons
Cpr =	Average amount of carbon (expressed as CO2) in natural phosphate rock, metric ton
CO2/ metric ton phosphate rock
Qpr =	Quantity of phosphate rock used to produce phosphoric acid
The CO2 emissions calculation methodology assumes that all of the inorganic C (calcium carbonate) content of the
phosphate rock reacts to produce CO2 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 CO2 and
that all of the organic C content remains in the phosphoric acid product. The United States uses a country-specific
methodology 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 2018, 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 2018, 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 2018 were obtained from USGS Minerals Yearbook: Phosphate Rock (USGS 1994 through 2015b),
and from USGS Minerals Commodity Summaries: Phosphate Rock (USGS 2016, 2017, 2018, 2019, 2020). From 2004
through 2018, the USGS reported no exports of phosphate rock from U.S. producers (USGS 2005 through 2015b).
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).
Carbonate content data for phosphate rock mined in Florida are used to calculate the CO2 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 are used to calculate CO2 emissions
from consumption of imported phosphate rock. The CO2 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).
Table 4-57: Phosphate Rock Domestic Consumption, Exports, and Imports (kt)
Location/Year
1990
2005
2014
2015
2016
2017
2018
U.S. Domestic Consumption
49,800
35,200
26,700
26,200
26,700
26,300
23,300
FLand NC
42,494
28,160
21,360
20,960
21,360
21,040
18,640
ID and UT
7,306
7,040
5,340
5,240
5,340
5,260
4,660
Exports—FL and NC
6,240
0
0
0
0
0
0
Imports
451
2,630
2,380
1,960
1,590
2,470
2,770
59 The 2006IPCC Guidelines do not provide a method for estimating process emissions (C02) from Phosphoric Acid Production.
Industrial Processes and Product Use 4-73

-------
Total U.S. Consumption
44,011
37,830
29,080 28,160 28,290 28,770 26,070
Table 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-Seri insistency
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 2018. 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 2018 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 2018 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 CO2 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 CO2 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 CO2. 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 CO2 in the elemental phosphorus production process. The calculation for
CO2 emissions assumes that phosphate rock consumption, for purposes other than phosphoric acid production,
results in CChemissions 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. 2018 phosphoric acid
production CO2 emissions were estimated to be between 0.8 and 1.2 MMT CO2 Eq. at the 95 percent confidence
4-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
level. This indicates a range of approximately 18 percent below and 20 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
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower Upper
Bound Bound
Lower Upper
Bound Bound
Phosphoric Acid Production
C02
0.9
.
Industrial Processes and Product Use 4-75

-------
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 (CO2)
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 CO2 generated from the iron and steel industry is a result of the
production of crude iron.
In addition to the production processes mentioned above, CO2 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
deducted and reported in the Energy chapter. In general, CO2 emissions are generated in these production
processes through the reduction and consumption of various carbon-containing inputs (e.g., ore, scrap, flux, coke
byproducts). In addition, fugitive CFU 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. These facilities have 21 active blast furnaces between them as of 2018. 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. In addition, there are 14 cokemaking facilities, of
which 3 facilities are co-located with integrated iron and steel facilities (ACCCI 2020). In the United States, four
states-Indiana, Ohio, Michigan, and Pennsylvania-account for roughly 51 percent of total raw steel production
(USGS 2019).
Total annual production of crude steel in the United States was fairly constant between 2000 and 2008 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 4.8 percent of world production in 2018
(AISI 2004 through 2018).
The majority of CO2 emissions from the iron and steel production process come from the use of 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.
4-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. However, the approaches and
emission estimates for both metallurgical coke production and iron and steel production 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 CO2 from metallurgical coke production in 2018 were 1.3 MMT CO2 Eq. (1,282 kt CO2) (see Table 4-60
and Table 4-61). Emissions decreased significantly in 2018 by 35 percent from 2017 levels and have decreased by
77 percent (4.3 MMT CO2 Eq.) since 1990. Coke production in 2018 was 34 percent lower than in 2000 and 50
percent below 1990.
Table 4-60: CO2 Emissions from Metallurgical Coke Production (MMT CO2 Eq.)
Gas
1990
2005
2014
2015
2016
2017
2018
C02
5.6
3.9
3.7
4.4
2.6
2.0
1.3
Total
5.6
3.9
3.7
4.4
2.6
2.0
1.3
Table 4-61: CO2 Emissions from Metallurgical Coke Production (kt)
Gas
1990
2005
2014
2015
2016
2017
2018
C02
5,608
3,921
3,721
4,417
2,643
1,978
1,282
Total
5,608
3,921
3,721
4,417
2,643
1,978
1,282
Iron and Steel Production
Emissions of CO2 and CH4 from iron and steel production in 2018 were 41.3 MMT CO2 Eq. (41,318 kt) and 0.0079
MMT CO2 Eq. (0.3 kt CH4), respectively (see Table 4-62 through Table 4-65), totaling approximately 41.3 MMT CO2
Eq. Emissions slightly increased in 2018 from 2017 but 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 2018, domestic production of pig iron increased by 7 percent from 2017 levels. Overall, domestic pig iron
production has declined since the 1990s. Pig iron production in 2018 was 50 percent lower than in 2000 and 52
percent below 1990. Carbon dioxide emissions from iron production have decreased by 79 percent since 1990.
Carbon dioxide emissions from steel production have decreased by 27 percent (2.2 MMT CO2 Eq.) since 1990,
while overall CO2 emissions from iron and steel production have declined by 58 percent (57.8 MMT CO2 Eq.) from
1990 to 2018.
Industrial Processes and Product Use 4-77

-------
Table 4-62: CO2 Emissions from Iron and Steel Production (MMT CO2 Eq.)
Source/Activity
Data
1990
2005
2014
2015
2016
2017
2018
Sinter Production
2.4
1.7
1.1
1.0
0.9
0.9
0.9
Iron Production
45.7
17.7
16.8
10.3
9.9
8.2
9.6
Pellet Production
1.8
1.5
1.1
1.0
0.9
0.9
0.9
Steel Production
8.0
9.4
7.5
6.9
6.9
6.2
5.8
Other Activities3
41.2
35.9
27.9
24.3
22.5
22.4
24.1
Total
99.1
66.2
54.5
43.5
41.0
38.6
41.3
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
2014
2015
2016
2017
2018
Sinter Production
2,448
1,663
1,104
1,016
877
869
937
Iron Production
45,704
17,664
16,848
10,333
9,930
8,239
9,583
Pellet Production
1,817
1,503
1,126
964
869
867
867
Steel Production
7,965
9,396
7,477
6,935
6,854
6,226
5,781
Other Activities3
41,193
35,934
27,911
24,280
22,451
22,396
24,149
Total
99,126
66,160
54,467
43,528
40,981
38,598
41,318
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
2014
2015
2016
2017
2018
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
2014
2015
2016
2017
2018
Sinter Production
0.9
0.6
0.4
0.3
0.3
0.3
0.3
Total
0.9
0.6
0.4
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.
The Tier 2 methodology equation is as follows:
Eco2 ~
^(<2a X Ca) - ^(<2fc X Cb)
44
12
4-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
where,
Eco2	=	Emissions from coke, pig iron, EAF steel, or BOF steel production, metric tons
a	=	Input material a
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 CO2 to C
The Tier 1 methodology equations are as follows:
Es,p = Qs x EFSJ)
Ed,C02 = Qd X EFd,C02
Ep,co2 = Qp x EFp C02
where,
Es,p	=	Emissions from sinter production process for pollutant p (CO2 or CH4), metric ton
Qs	=	Quantity of sinter produced, metric tons
EFs,p	=	Emission factor for pollutant p (CO2 or CH4), metric ton p/metric ton sinter
Ed,co2	=	Emissions from DRI production process for CO2, metric ton
Qd	=	Quantity of DRI produced, metric tons
EFd,co2	=	Emission factor for CO2, metric ton C02/metric ton DRI
QP	=	Quantity of pellets produced, metric tons
EFP,co2	=	Emission factor for CO2, metric ton C02/metric ton pellets produced
Metallurgical Coke Production
Coking coal is used to manufacture metallurgical coke that 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 4-79

-------
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 Cm emission factor for metallurgical coke production (i.e., 0.1 g
Cm per metric ton of coke production), it is not appropriate to use because CO2 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 CO2 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 2019) 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
2014
2015
2016
2017
2018
Metallurgical Coke Production







Coking Coal Consumption at Coke Plants
35,269
21,259
19,321
17,879
14,955
15,910
16,635
Coke Production at Coke Plants
25,054
15,167
13,748
12,479
10,755
11,746
12,525
Coal Breeze Production
2,645
1,594
1,449
1,341
1,122
1,193
1,248
Coal Tar Production
1,058
638
580
536
449
477
499
Table 4-68: Production and Consumption Data for the Calculation of CO2 Emissions from
Metallurgical Coke Production (Million ft3)
Source/Activity Data	1990 2005	2014 2015 2016 2017 2018
Metallurgical Coke Production
Coke Oven Gas Production	250,767
Natural Gas Consumption	599
114,213 102,899 84,336 74,807 74,997 80,750
2,996 3,039 2,338 2,077 2,103 2,275
4-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Blast Furnace Gas Consumption
24,602
4,460
4,346 4,185 3,741 3,683 4,022
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 CO2 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 CO2 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 CFU, 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 CO2 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 CO2 emissions; the
Industrial Processes and Product Use 4-81

-------
estimation of Cm emissions is precluded. A preliminary analysis of facility-level emissions reported during iron
production further supports this assumption and indicates that Cm emissions are below 500 kt CO2 Eq. and well
below 0.05 percent of total national emissions. The production of direct reduced iron also results in emissions of
Cm 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 CChfrom 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 CO2
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 CO2 from Fossil Fuel Combustion.
Sinter consumption and pellet consumption data for 1990 through 2018 were obtained from AISI's Annual
Statistical Report (AISI 2004 through 2019) 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 2016) and personal communication with the
USGS Iron and Steel Commodity Specialist (Fenton 2015 through 2019). 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 CO2 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 2019) 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 2019) and through personal communications with AISI (AISI 2006
through 2016 and AISI 2008). The factor for the quantity of EAF anode consumed per ton of EAF steel produced
4-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
was 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 2019) 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 2016). 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 2019) 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 2019). 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 2018). Heat contents for coke
oven gas and blast furnace gas were provided in Table 37 of AISI's Annual Statistical Report (AISI 2004 through
2019) 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
2014
2015
2016
2017
2018
Sinter Production







Sinter Production
12,239
8,315
5,521
5,079
4,385
4,347
4,687
Direct Reduced Iron







Production







Direct Reduced Iron







Production
516
1,303
2,113
2,722
C
C
C
Pellet Production







Pellet Production
60,563
50,096
37,538
32,146
28,967
28,916
28,916
Pig Iron Production







Coke Consumption
24,946
13,832
11,136
7,969
7,124
7,101
7,618
Pig Iron Production
49,669
37,222
29,375
25,436
22,293
22,395
24,058
Direct Injection Coal







Consumption
1,485
2,573
2,425
2,275
1,935
2,125
2,569
EAF Steel Production







EAF Anode and Charge







Carbon Consumption
67
1,127
1,062
1,072
1,120
1,127
1,133
Scrap Steel







Consumption
42,691
46,600
48,900
44,000
C
C
C
Flux Consumption
319
695
771
998
998
998
998
EAF Steel Production
33,511
52,194
55,174
49,451
52,589
55,825
58,904
BOF Steel Production







Pig Iron Consumption
47,307
34,400
23,800
20,300
C
C
C
Scrap Steel







Consumption
14,713
11,400
5,920
4,530
C
C
C
Flux Consumption
576
582
454
454
408
408
408
BOF Steel Production
43,973
42,705
33,000
29,396
25,888
25,788
27,704
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	2014 2015 2016 2017 2018
Pig Iron Production
Natural Gas Consumption	56,273	59,844	47,734 43,294 38,396 38,142 40,204
Industrial Processes and Product Use 4-83

-------
Fuel Oil Consumption
(thousand gallons)
163,397
16,170
16,674
9,326
6,124
4,352
3,365
Coke Oven Gas Consumption
22,033
16,557
16,896
13,921
12,404
12,459
13,337
Blast Furnace Gas







Production
1,439,380
1,299,980
1,000,536
874,670
811,005
808,499
871,860
EAF Steel Production







Natural Gas Consumption
15,905
19,985
9,622
8,751
3,915
8,105
8,556
BOF Steel Production







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







Coke Oven Gas Consumption
224,883
97,132
85,479
70,029
62,036
62,164
67,008
Blast Furnace Gas







Consumption
1,414,778
1,295,520
996,190
870,485
807,264
804,816
867,838
Uncertainty and Time-Series Consistency
The estimates of CO2 emissions from metallurgical coke production are based on 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 CO2 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 CO2 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 CO2 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
4-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2006IPCC 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
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 CO2 emissions from metallurgical coke production and iron and
steel production for 2018 were estimated to be between 35.2 and 50.4 MMT CO2 Eq. at the 95 percent confidence
level. This indicates a range of approximately 17 percent below and 18 percent above the emission estimate of
41.3 MMT CO2 Eq. Total CH4 emissions from metallurgical coke production and iron and steel production for 2018
were estimated to be between 0.006 and 0.009 MMT CO2 Eq. at the 95 percent confidence level. This indicates a
range of approximately 20 percent below and 20 percent above the emission estimate of 0.008 MMT CO2 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
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)
(MMT CO;
>Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Metallurgical Coke & Iron
and Steel Production
C02
41.3
35.2
50.4
-17%
+18%
Metallurgical Coke & Iron
and Steel Production
ch4
+
+
+
-20%
+20%
+ 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 2018.
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 2006 IPCC Guidelines, see the QA/QC and Verification Procedures section in the introduction of
the IPPU chapter.
Recalculations Discussion
The carbon balance calculations for metallurgical coke production for previous Inventories used a C content of 73
percent by weight for coking coal based on Table 4.3 of the 2006 IPCC Guidelines for National Greenhouse Gas
Inventories. Based on recommendations as part of the Inventory UNFCCC review this factor was updated to be
more consistent with factors used in the Energy calculations of the Inventory. For this Inventory report the C
content value for coking coal was updated to 75.4 percent carbon by weight based on data from the U.S. Energy
Information Administration (EIA). This change resulted in an annual average increase in emissions of 1.8 MMT CO2
Eq.
Industrial Processes and Product Use 4-85

-------
Planned Improvements
Future improvements involve improving activity data and emission factor sources for estimating CO2 and Cm
emissions 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
inventories will be relied upon.62 This is a medium-term improvement and EPA estimates that earliest this
improvement could be incorporated is the 2021 Inventory submission.
Additional improvements include accounting for emission estimates for the production of metallurgical coke to 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 medium-term improvement and is still in development; therefore, it is not included
in this current Inventory report and is not expected until the next (i.e., 2021) 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 (CO2) 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, CO2 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 CO2. A representative reaction equation for the
production of 50 percent ferrosilicon (FeSi) is given below:
Fe203 + 2Si02 + 7C —> 2FeSi + 7C0
62 See .
4-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
While most of the carbon contained in the process materials is released to the atmosphere as CO2, a percentage is
also released as Cm 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 2018,12 companies in the United States produce ferroalloys (USGS 2018a).
Emissions of CO2 from ferroalloy production in 2018 were 2.1 MMT CO2 Eq. (2,063 kt CO2) (see Table 4-75 and
Table 4-76), which is a 4 percent reduction since 1990. Emissions of CH4 from ferroalloy production in 2018 were
0.01 MMT CO2 Eq. (0.6 kt CH4), which is a 15 percent decrease since 1990.
Table 4-75: CO2 and ChU Emissions from Ferroalloy Production (MMT CO2 Eq.)
Gas
1990
2005
2014
2015
2016
2017
2018
C02
2.2
1.4
1.9
2.0
1.8
2.0
2.1
ch4
+
+ *
+
+
+
+
+
Total
2.2
1.4
1.9
2.0
1.8
2.0
2.1
+ Does not exceed 0.05 MMT C02 Eq.
Table 4-76: CO2 and ChU Emissions from Ferroalloy Production (kt)
Gas
1990
2005
2014
2015
2016
2017
2018
C02
ch4
2,152
1
1,392
+
1,914
1
1,960
1
1,796
1
1,975
1
2,063
1
+ Does not exceed 0.05 MMT C02 Eq.
IViethiMlGiGgf
Emissions of CO2 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 CO2 and Cm emissions are as follows:
Eco2 = YSMPi x EFi)
i
where,
Eco2 =	CO2 emissions, metric tons
MP, =	Production of ferroalloy type/', metric tons
EFi =	Generic emission factor for ferroalloy type /', metric tons CCh/metric ton specific
ferroalloy product
where,
ECHi =	X EF^
Ech4 =	Cm emissions, kg
MP, =	Production of ferroalloy type/', metric tons
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 4-87

-------
EFi =	Generic emission factor for ferroalloy type /', kg Cm/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 CO2 and CFU estimates:
•	Ferrosilicon, 25 to 55 percent Si and Miscellaneous Alloys, 32 to 65 percent Si - 2.5 metric tons
CCh/metric ton of alloy produced; 1.0 kg Cm/metric ton of alloy produced.
•	Ferrosilicon, 56 to 95 percent Si - 4.0 metric tons CCh/metric ton alloy produced; 1.0 kg Cm/metric ton of
alloy produced.
•	Silicon Metal - 5.0 metric tons CCh/metric ton metal produced; 1.2 kg Cm/metric ton metal produced.
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 CO2 from Fossil Fuel
Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating
Emissions of CO2 from Fossil Fuel Combustion.
Ferroalloy production data for 1990 through 2018 (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). 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 2018 (USGS 2013, 2014, 2015b, 2016b, 2017,
2018b, 2019).
Table 4-77: Production of Ferroalloys (Metric Tons)
Year
Ferrosilicon
Ferrosilicon
Silicon Metal
Misc. Alloys 32-

25%-55%
56%-95%

65%
1990
321,385
109,566
145,744
72,442
2005
123,000
86,100
148,000
NA
2014
176,161
155,436
170,404
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
4-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
2018	189,846	167,511	183,642	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
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.
Also, 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 CO2 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 will
reduce Cm 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 CO2 emissions from 2018 were estimated to be between 1.8 and 2.3 MMT CO2 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 CO2 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 CO2 Eq.
Table 4-78: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Ferroalloy Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTCOzEq.) (%)



Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Ferroalloy Production
Ferroalloy Production
C02
ch4
2.1
+
1.8
+
2.3
+
-12%
-12%
+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 2018. Details on the emission trends through time are described in more detail in the Methodology
section, above.
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 4-89

-------
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 2017 portion of the time series.
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
2019a). 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
(CO2) and two perfluorocarbons (PFCs): perfluoromethane (CF4) and perfluoroethane (C2F6).
Carbon dioxide is emitted during the aluminum smelting process when alumina (aluminum oxide, AI2O3) 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 (NasAIFs). 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 CO2.
Process emissions of CO2 from aluminum production were estimated to be 1.5 MMT CO2 Eq. (1,451 kt) in 2018 (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 CO2 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 CO2 from Fossil
65 See .
4-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Fuel Combustion source category of the Energy sector. Similarly, the coal tar pitch portion of these CO2 process
emissions is accounted for here.
Table 4-79: CO2 Emissions from Aluminum Production (MMT CO2 Eq. and kt)
Year
MMT CO? Eq.
kt
1990
6.8
6,831
2005
4.1
4,142
2014
2.8
2,833
2015
2.8
2,767
2016
1.3
1,334
2017
1.2
1,205
2018
1.5
1,451
In addition to CO2 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 "anode effects." These anode effects 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.
Since 1990, emissions of CF4 and C2F6 have declined by 94 percent and 88 percent, respectively, to 1.1 MMT CO2
Eq. of CF4 (0.21 kt) and 0.4 MMT CO2 Eq. of C2F6 (0.03 kt) in 2018, 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 78 percent, while the combined CF4 and C2F6
emission rate (per metric ton of aluminum produced) has been reduced by 67 percent. PFC emissions increased by
approximately 51 percent between 2017 and 2018 due to increases in both aluminum production and CF4
emissions per metric ton of aluminum produced. Increases in CF4 emissions per metric ton of aluminum may be
due to a combination of increased production, increased anode effect duration and/or frequency, and increases in
the smelter-specific slope coefficients at individual facilities. The decrease in the ratio of C2F6toCF4 emissions may
be due to combination of a decrease in the measured C2F6 to CF4 weight ratio at some facilities and a change in the
relative share of production at each facility.
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
2014
1.9
0.6
2.5
2015
1.5
0.5
2.0
2016
0.9
0.4
1.4
2017
0.7
0.4
1.0
2018
1.1
0.4
1.6
Note: Totals may not sum due to
independent rounding.
Industrial Processes and Product Use 4-91

-------
Table 4-81: PFC Emissions from Aluminum Production (kt)
Year
cf4
c2f6
1990
2.4
0.3
2005
0.4
+
2014
0.3
0.1
2015
0.2
+
2016
0.1
+
2017
0.1
+
2018
0.2
+
+ Does not exceed 0.05 kt.
In 2018, U.S. primary aluminum production totaled approximately 0.9 million metric tons, a 21 percent increase
from 2017 production levels (USAA 2019). In 2018, three companies managed production at seven operational
primary aluminum smelters. Two smelters that were idle at the end of 2017 were restarted and one other smelter
restarted production in 2018. One smelter remained on standby throughout 2018 (USGS 2019b). During 2018,
monthly U.S. primary aluminum production was higher for every month when compared to the corresponding
months in 2017 (USAA 2019, 2018).
For 2019, total production for the January to August period was approximately 0.8 million metric tons compared to
0.5 million metric tons for the same period in 2018, a 37.9 percent increase (USAA 2019). Based on the increase in
production, process CO2 and PFC emissions are likely to be higher in 2019 compared to 2018 if there are no
significant changes in process controls at operational facilities.
Process CO2 and PFC (i.e., CF4 and C2F6) emission estimates from primary aluminum production for 2010 through
2018 are available from EPA's GHGRP—Subpart F (Aluminum Production) (EPA 2019). Under EPA's GHGRP,
facilities began reporting primary aluminum production process emissions (for 2010) in 2011; as a result, GHGRP
data (for 2010 through 2018) 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, CO2 emissions from anode consumption during electrolysis in all
prebake and S0derberg cells, and all CO2 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 CO2 from
Fossil Fuel Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for
Estimating Emissions of CO2 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 .
4-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Process CO2 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
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 CO2 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, CO2 is also emitted during the anode baking process. These emissions can
account for approximately 10 percent of total process CO2 emissions from prebake smelters.
Depending on the availability of smelter-specific data, the CO2 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 CO2 emission factors. The first approach tracks the
consumption and carbon content of the anode, assuming that all C in the anode is converted to CO2. 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 CO2 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 CO2 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 CO2 emissions data or CO2 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), CO2 emission estimates were estimated using Tier 1 S0derberg and/or Prebake emission factors (metric ton
of CO2 per metric ton of aluminum produced) from IPCC (2006).
Industrial Processes and Product Use 4-93

-------
Process PFC Emissions from Anode Effects
Smelter-specific PFC emissions from aluminum production for 2010 through 2018 were reported to EPA under its
GHGRP. To estimate their PFC emissions 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. 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.
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 2018 were obtained via USAA (USAA 2019). 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 2018, national aluminum
production data were obtained from the USAA's Primary Aluminum Statistics (USAA 2004 through 2006, 2008
through 2019).
4-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 4-82: Production of Primary Aluminum (kt)
Year	kt_
1990 4,048
2005 2,478
2014	1,710
2015	1,587
2016	818
2017	741
2018	897
Uncertainty and Time-Series Consistency
Uncertainty was assigned to the CO2, CF4, and C2F6 emission values reported by each individual facility to EPA's
GHGRP. As previously mentioned, the methods for estimating emissions for EPA's GHGRP and this report are the
same, and follow the 2006IPCC Guidelines methodology. As a result, it was possible to assign uncertainty bounds
(and distributions) based on an analysis of the uncertainty associated with the facility-specific emissions estimated
for previous Inventory years. Uncertainty surrounding the reported CO2, CF4, and C2F6 emission values were
determined to have a normal distribution with uncertainty ranges of ±6, ±16, and ±20 percent, respectively. A
Monte Carlo analysis was applied to estimate the overall uncertainty of the CO2, 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 CO2 emissions were estimated to be between 1.42 and 1.49 MMT CO2 Eq. at the 95 percent
confidence level. This indicates a range of approximately 2 percent below to 2 percent above the emission
estimate of 1.45 MMT CO2 Eq. Also, production-related CF4 emissions were estimated to be between 1.06 and 1.24
MMT CO2 Eq. at the 95 percent confidence level. This indicates a range of approximately 8 percent below to 8
percent above the emission estimate of 1.15 MMT CO2 Eq. Finally, aluminum production-related C2F6 emissions
were estimated to be between 0.36 and 0.49 MMT CO2 Eq. at the 95 percent confidence level. This indicates a
range of approximately 15 percent below to 16 percent above the emission estimate of 0.43 MMT CO2 Eq.
Table 4-83: Approach 2 Quantitative Uncertainty Estimates for CO2 and PFC Emissions from
Aluminum Production (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)
(MMT CO?
Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Aluminum Production
C02
1.45
1.42
1.49
-2%
2%
Aluminum Production
cf4
1.15
1.06
1.24
-8%
8%
Aluminum Production
c2f6
0.43
0.36
0.49
-15%
16%
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 2018. 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
Industrial Processes and Product Use 4-95

-------
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 (2015).67 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
In a few instances GHGRP facilities revised their GHGRP reports due to previously identified reporting errors in
2017, resulting in a decrease in total emissions of PFCs.
4.20 Magnesium Production and Processing
(CRF Source Category 2C4)
The magnesium metal production and casting industry uses sulfur hexafluoride (SFs) 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
(CO2) 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 (SO2) 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 1.1 MMT CO2 Eq. (0.05 kt) of SF6, 0.1 MMT CO2 Eq. (0.1 kt) of HFC-134a, and
0.001 MMT CO2 Eq. (1.4 kt) of CO2 in 2018. This represents an increase of approximately 2 percent from total 2017
emissions (see Table 4-84) and an increase in SF6 emissions by 4 percent. The increase can be attributed to an
increase in die casting and permanent mold SF6 emissions between 2017 and 2018 as reported through the
GHGRP, including from two first-time reporters to the GHGRP. In 2018, total HFC-134a emissions decreased from
0.098 MMT CO2 Eq. to 0.090 MMT CO2 Eq., or a 9 percent decrease as compared to 2017 emissions. FK 5-1-12
emissions decreased from 2017 levels. The emissions of the carrier gas, CO2, decreased from 3.1 kt in 2017 to 1.4
kt in 2018, or 53 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
2014
2015
2016
2017
2018
sf6
5.2
2.7
0.9
1.0
1.1
1.1
1.1
HFC-134a
0.0
O
O
0.1
0.1
0.1
0.1
0.1
C02
+
+
+
+
+
+
+
FK 5-1-12°
0.0
0.0
+
+
+
+
+
Total
5.2
2.7
1.0
1.1
1.2
1.2
1.2
+ Does not exceed 0.05 MMT C02 Eq.
67 GHGRP Report Verification Factsheet. .
4-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
a Emissions of FK 5-1-12 are not included in totals.
Table 4-85: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and
Processing (kt)
Year
1990
2005
2014
2015
2016
2017
2018
sf6
0.2
0.1
+
+
+
+
+
HFC-134a
0.0
0.0
0.1
0.1
0.1
0.1
0.1
C02
1.4
2.9
2.3
2.6
2.7
3.1
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
SFs 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 also reported their HFC-134a and FK 5-1-12 usage, which is also
assumed to be equal to emissions. The last reporting year was 2010 under the Partnership. Emissions data for
2011 through 2018 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 CO2. 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 2018 (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
SFs 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 is reported in the RAND survey of major SF6 manufacturers, which
shows 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. These emission factors for the other processes (i.e., permanent mold, wrought, and
anode casting) were based on discussions with industry representatives.
The quantities of CO2 carrier gas used for each production type have been estimated using the 1999 estimated CO2
emissions data and the annual calculated rate of change of SF6 use in the 1990 through 1999 time period. For each
Industrial Processes and Product Use 4-97

-------
year and production type, the rate of change of SF6 use between the current year and the subsequent year was
first estimated. This rate of change is then applied to the CO2 emissions of the subsequent year to determine the
CO2 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. This was likely due
to a temporary decrease in production at many facilities between 2008 and 2010, where those facilities were
operating at production levels significantly less than full capacity.
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 CO2 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 CO2 as a carrier gas through the GHGRP. Using this approach helped ensure
4-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. For
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 2018
For 2011 through 2018, 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 SFs 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). 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 2018). USGS data for 2018 was not yet available at the time of the analysis, so the 2016 values were
held constant through 2018 as a proxy. Where data was submitted late or with errors for 2018 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 2018 is attributed to the uncertainties around SF6, HFC-
134a, and CO2 emission estimates. To estimate the uncertainty surrounding the estimated 2018 SF6 emissions from
magnesium production and processing, the uncertainties associated with three variables were estimated: (1)
emissions reported by magnesium producers and processors for 2018 through EPA's GHGRP, (2) emissions
estimated for magnesium producers and processors that reported via the Partnership in prior years but did not
report 2018 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
Industrial Processes and Product Use 4-99

-------
(per the 2006IPCC Guidelines). If facilities did not report emissions data during the current reporting year through
EPA's GHGRP, SFs 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 2017 and
2018.
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 1.11 and 1.28 MMT CO2 Eq.
at the 95 percent confidence level. This indicates a range of approximately 7 percent below to 7 percent above the
2018 emission estimate of 1.20 MMT CO2 Eq. The uncertainty estimates for 2018 are similar to the uncertainty
reported for 2017 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
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO? Eq.)
(MMT CO? Eq.)
(%)



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Magnesium
SF6, HFC-
1.20
1.11 1.28
-7% 7%
Production
134a, C02



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 2018. 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 20 15).68 Based on the results
of the verification process, EPA follows up with facilities to resolve mistakes that may have occurred. The post-
68 GHGRP Report Verification Factsheet. .
4-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
In a few instances GHGRP facilities revised their GHGRP reports due to previously identified reporting errors in
2016, resulting in a change of SF6 emissions for die casting and sand casting in 2016. The emission factors for die
casting shown in Table 4-86 were updated by holding activity data constant at 2012 levels between 2009 to 2012
based on additional information from USGS on activity data.
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 2006 IPCC 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 CO2 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 2018, lead was produced in the United States only using secondary production processes. Until 2014, both lead
production in the United States involved both primary and secondary processes—both of which emit carbon
dioxide (CO2) (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 CO2 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
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, CO2 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 2018, secondary lead production accounted for 100 percent of total
lead production. The lead-acid battery industry accounted for more than 85 percent of the reported U.S. lead
consumption in 2018 (USGS 2019).
Industrial Processes and Product Use 4-101

-------
In 2018, total secondary lead production in the United States was similar to that in 2017. A new secondary lead
refinery, located in Nevada, was completed and began production in 2016. The plant produces high-purity refined
lead for use in advanced lead-acid batteries using an electromechanical battery recycling technology system. 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. During the
first 10 months of 2018, however, 22.9 million spent SLI lead-acid batteries were exported, which was 44 percent
more than exports in 2017 (USGS 2019).
As in 2017, U.S. primary lead production remained at production levels of zero for 2018. This is due to the closure
of the only domestic primary lead smelter in 2013 (year-end), as stated previously. In 2018, U.S. secondary lead
production was similar to 2017 levels, and has increased by 24 percent since 1990 (USGS 1995 through 2019).
In 2018, U.S. lead production totaled 1,140,000 metric tons (USGS 2020). The resulting emissions of CO2 from 2018
lead production were estimated to be 0.5 MMT CO2 Eq. (513 kt) (see Table 4-88). The 2016 and 2017 CO2 values
were also updated and are summarized in Table 4-88 (USGS 2020).
At last reporting, the United States was the fourth largest mine producer of lead in the world, behind China,
Australia, and Peru accounting for approximately 6 percent of world production in 2018 (USGS 2019).
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
2014
0.5
459
2015
0.5
473
2016
0.5
500
2017
0.5
513
2018
0.5
513
After a steady increase in total emissions from 1995 to 2000, total emissions have gradually decreased since 2000
and are currently 1 percent lower than 1990 levels.
Methodology
The methods used to estimate emissions for lead production69 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:
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 CCh/metric ton lead product
EFs	=	Emission factor for secondary materials, metric tons CCh/metric ton lead product
69 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.
4-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
For primary lead production using direct smelting, Sjardin (2003) and the IPCC (2006) provide an emission factor of
0.25 metric tons CCh/metric ton lead. For secondary lead production, Sjardin (2003) and IPCC (2006) provide an
emission factor of 0.25 metric tons CCh/metric ton lead for direct smelting, as well as an emission factor of 0.2
metric tons CCh/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 CO2 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 CO2 from Fossil Fuel
Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating
Emissions of CO2 from Fossil Fuel Combustion.
The 1990 through 2018 activity data for primary and secondary lead production (see Table 4-89) were obtained
from the U.S. Geological Survey (USGS 1995 through 2020). The 2016 and 2017 lead production values were also
updated and are summarized in Table 4-89 (USGS 2020).
Table 4-89: Lead Production (Metric Tons)
Year
Primary
Secondary
1990
404,000
922,000
2005
143,000
1,150,000
2014
1,000
1,020,000
2015
0
1,050,000
2016
0
1,110,000
2017
0
1,140,000
2018
0
1,140,000
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 CO2 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 CO2
emissions in 2018 were estimated to be between 0.4 and 0.6 MMT CO2 Eq. at the 95 percent confidence level. This
indicates a range of approximately 15 percent below and 15 percent above the emission estimate of 0.5 MMT CO2
Eq.
Industrial Processes and Product Use 4-103

-------
Table 4-90: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lead
Production (MMT CO2 Eq. and Percent)
Source Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)


Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Lead Production C02
0.5
0.4 0.6
-15%
+15%
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 2018. 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 is used for QA purposes only.
Recalculations Discussion
For the current 1990 through 2018 Inventory, updated USGS data on lead were available. The revised production
values used in the current Inventory resulted in revised emissions estimates for the years 2016 and 2017.
Compared to the previous Inventory, emissions in the current Inventory for 2016 increased by approximately 13
percent (56 kt CO2 Eq.) and increased 1 percent (4 kt CO2 Eq.) for 2017.
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.70
70 See .
4-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
4.22 Zinc Production (CRF Source Cati—ry
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 (CO2) 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 CO2 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 CO2 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"), PIZO, and Steel Dust Recycling (SDR). For AZR,
EAF dust is recycled in Waelz kilns at their Calumet, IL; Palmerton, PA; Rockwood, TN; and Barnwell, SC facilities.
These Waelz kiln 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. The new Mooresboro facility uses a hydrometallurgical process (i.e., solvent extraction
with electrowinning technology) to produce zinc products. The current capacity of the new facility is 155,000 short
tons, with plans to expand to 170,000 short tons per year. 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).
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
Industrial Processes and Product Use 4-105

-------
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). Hydrometallurgical production processes are
assumed to be non-emissive since no carbon is used in these processes (Sjardin 2003).
PIZO and SDR recycle EAF dust into intermediate zinc products using Waelz kilns, and then sell the intermediate
products to companies who smelt it into refined products.
Emissions of CO2 from zinc production in 2018 were estimated to be 1.0 MMT CO2 Eq. (1,009 kt CO2) (see Table
4-91). All 2018 CO2 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 2018, emissions were estimated to be 60 percent higher than they were in
1990.
Table 4-91: CO2 Emissions from Zinc Production (MMT CO2 Eq. and kt)
Year
MMTCOz Eq.
kt
1990
0.6
632
2005
1.0
1,030
2014
1.0
956
2015
0.9
933
2016
0.9
925
2017
1.0
1,009
2018
1.0
1,009
In 2018, United States primary and secondary refined zinc production were estimated to total 116,000 metric tons
(USGS 2020) (see Table 4-92). Domestic zinc mine production increased slightly in 2018, owing to the addition of
production from a reopened mine in New York (USGS 2019). Refined zinc production decreased slightly owing to
maintenance outages at the Clarksville, TN, smelter (USGS 2019). Primary zinc production (primary slab zinc)
decreased by fourteen percent in 2018, while secondary zinc production in 2018 stayed the same relative to 2017.
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
2014
110,000
70,000
180,000
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
4-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
fvieth ado logy
The methods used to estimate non-energy CO2 emissions from zinc production71 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 EFdefaldt
where,
Eco2 =	CO2 emissions from zinc production, metric tons
Zn =	Quantity of zinc produced, metric tons
EFdefauit =	Default emission factor, metric tons CCh/metric ton zinc produced
The Tier 1 emission factors provided by IPCC for Waelz kiln-based secondary production were derived from 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
t-j ^ aelz KlL~h ~~	¦	¦	^	^	.	-—	.	.
metric tons zinc metric tons coke	metric tons C	metric tons zinc
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
EFuAp Dust — ! : ! 	 ~ X ¦ ; ¦ " X ¦
metric tons EAF Dust metric tons coke	metric tons C	metric tons EAF Dust
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). Total EAF dust consumed by AZR at their Waelz kilns was not available for 2018 so 2015
data was used as proxy. Consumption levels for 1990 through 2005 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). The EAF dust consumption values for
71 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.
Industrial Processes and Product Use 4-107

-------
each year were then multiplied by the 1.24 metric tons CCh/metric ton EAF dust consumed emission factor to
develop CO2 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 2017 (SDR 2012, 2014, 2015, and 2017). SDR data for 2018 was not available
at time of Public Review so 2017 data was used as a proxy. SDR's facility in Alabama underwent expansion in 2011
to include a second unit (operational since early- to mid-2012). SDR's facility has been operational since 2008.
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).
PIZO Technologies Worldwide LLC's facility in Arkansas has been operational since 2009. The amount of EAF dust
consumed by PIZO's facility for 2009 through 2018 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 2018 were estimated by applying the average annual capacity utilization rates for AZR and SDR
(Grupo PROMAX) to PIZO's annual capacity (Horsehead 2012, 2013, 2014, 2015, and 2016; SDR 2012, 2014 and
2017; PIZO 2012, 2014 and 2017). The 1.24 metric tons C02/metric ton EAF dust consumed emission factor was
then applied to PIZO's and SDR's estimated EAF dust consumption to develop CO2 emission estimates for those
Waelz kiln facilities.
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 was replaced by AZR's new facility in Mooresboro, NC. The
new facility uses hydrometallurgical process to produce refined zinc products. This process is assumed to be non-
emissive. Production levels for 1990 through 2004 were extrapolated using the percentage changes in annual
refined zinc production at secondary smelters in the United States as provided by USGS Minerals Yearbook: Zinc
(USGS 1995 through 2005). The 3.70 metric tons C02/metric ton zinc emission factor was then applied to the
Monaca facility's production levels to estimate CO2 emissions for the facility. The Waelz kiln production emission
factor was applied in this case rather than the EAF dust consumption emission factor since AZR's Monaca facility
did not consume EAF dust.
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 CO2 from Fossil Fuel
Combustion (3.1 Fossil Fuel Combustion (CRF Source Category 1A)) and Annex 2.1, Methodology for Estimating
Emissions of CO2 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 2018, only one facility produced primary
zinc. Primary zinc produced from this facility was subtracted from the USGS 2016 to 2018 total zinc production
statistic to estimate secondary zinc production for these years.
Uncertainty and Time-Seri insistency
The uncertainty associated with these estimates is two-fold, relating to activity data and emission factors used.
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
4-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
part of its financial reporting to the Securities and Exchange Commission (SEC), 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). Also, the EAF dust consumption for PIZO's facility for 2011 through 2016 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 and SDR's annual EAF dust consumption values (except SDR's EAF dust consumption for 2011
through 2017, which were obtained from SDR's recycling facility in Alabama).
Second, there is uncertainty associated with the emission factors used to estimate CO2 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 CO2
emissions from 2018 were estimated to be between 0.8 and 1.2 MMT CO2 Eq. at the 95 percent confidence level.
This indicates a range of approximately 16 percent below and 16 percent above the emission estimate of 1.0 MMT
CO2 Eq.
Table 4-93: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Zinc
Production (MMT CO2 Eq. and Percent)

2018 Emission



Source Gas
Estimate
Uncertainty Range Relative to Emission Estimate'
i

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



Lower Upper
Lower
Upper


Bound Bound
Bound
Bound
Zinc Production C02
1.0

-------
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.72 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 (N2O) 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 (HFC-23 or
CHF3), perfluoromethane (CF4), perfluoroethane (C2F6), nitrogen trifluoride (NF3), and sulfur hexafluoride (SFs),
although other fluorinated compounds such as perfluoropropane (C3F8) and perfluorocyclobutane (c-C4Fs) 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 N2O and a combination of other HFCs and unsaturated, low-GWP PFCs such as CsFs, 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 CO2 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
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
72 See .
4-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 10,000.73
For 2018, total GWP-weighted emissions of all fluorinated greenhouse gases and N2O from deposition, etching,
and chamber cleaning processes in the U.S. semiconductor industry were estimated to be 5.1 MMT CO2 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, C4F8O, CsFs, HFC-32, HFC-41, and HFC-134a. These
gases have been grouped as "Other F-GHGs". Emissions from all fluorinated greenhouse gases and N2O are
presented in Table 4-94 and Table 4-95 below for the years 1990, 2005, and the period 2014 to 2018. 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 GHGRP during years 2012 through 2018. 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)74 of semiconductor
products led to an increase in emissions of 153 percent between 1990 and 1999, when emissions peaked at 9.1
MMT CO2 Eq. Emissions began to decline after 1999, reaching a low point in 2009 before rebounding slightly and
plateauing at the current level, which represents a 45 percent decline from 1999 levels. Together, industrial
growth, adoption of emissions reduction technologies (including but not limited to abatement technologies), and
shifts in gas usages resulted in a net increase in emissions of approximately 41 percent between 1990 and 2018.
Total emissions from semiconductor manufacture in 2018 were similar to 2017 emissions, increasing by 3 percent.
The emissions reported by facilities manufacturing MEMS included emissions of C2F6, C3F8, C4F8, CF4, HFC-23, NF3,
and SFe,75 and were equivalent to only 0.08 percent to 0.40 percent of the total reported emissions from
semiconductor manufacturing in 2011 to 2018. These emissions ranged from 0.0001 to 0.0185 MMT CO2 Eq. from
1991 to 2018. Based upon information in the World Fab Forecast (WFF), it appears that some GHGRP reporters
73	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.
74	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.
75	Gases not reported by MEMS manufacturers to the GHGRP are currently listed as "NE" in the CFR. 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.
Industrial Processes and Product Use 4-111

-------
that manufacture both semiconductors and MEMS are reporting their emissions as only from semiconductor
manufacturing (GHGRP reporters must choose a single classification per fab). Some fabs that reported as
manufacturing MEMS in 2011 also later reported their emissions as emissions from manufacturing
semiconductors. Thus, the decrease in estimated emissions from MEMS manufacturing between 2011 and 2018
may be partially due to emissions from some fabs being included in the MEMS estimates in the earlier years of the
GHGRP but are now included under semiconductor manufacturing emissions. Emissions from non-reporters have
not been estimated.
Total GWP-weighted emissions from manufacturing of photovoltaic cells were estimated to range from 0.0018
MMT CO2 Eq. to 0.0247 MMT CO2 Eq. from 1998 to 2018 and were equivalent to between 0.02 percent to 0.50
percent of the total reported emissions from semiconductor 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, C4F8, and CHF3.76
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 CO2 Eq. and 1.1 MMT CO2
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 11 percent and 18 percent of total annual emissions (F-GHG, N2O and F-
HTFs) from semiconductor manufacturing.77 Table 4-96 shows F-HTF emissions in tons by compound group based
on reporting to EPA's GHGRP during years 2012 through 2018.78
Table 4-94: PFC, HFC, SFe, NF3, and N2O Emissions from Electronics Manufacture79 (MMT
COz Eq.)
Year
1990
2005
2014
2015
2016
2017
2018
cf4
0.8
1.1
1.5
1.5
1.5
1.6
1.7
c2f6
2.0
2.0
1.4
1.3
1.2
1.2
1.1
CsFs
+
0.1
0.1
0.1
0.1
0.1
0.1
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
0.3
0.4
0.4
sf6
0.5
0.7
0.7
0.7
0.8
0.7
0.8
nf3
+
0.5
0.5
0.6
0.6
0.6
0.6
Other F-GHGs
+
+
+
+
+
+
+
Total F-GHGs
3.6
4.6
4.6
4.7
4.7
4.6
4.8
76	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.
77	Emissions data for HTFs (in tons of gas) from the semiconductor industry from 2011 through 2018 were obtained from the
EPA GHGRP annual facility emissions reports.
78	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 .
79	An extremely small portion of emissions included in the totals for Semiconductor Manufacture are from the manufacturing
of MEMS and photovoltaic cells.
4-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
n2o80
+
0.1
0.2
0.2
0.2
0.3
0.3
HFC, PFC and SF6 F-HTFs
0.000
0.028
0.026
0.019
0.018
0.021
0.020
MEMS
0.000
0.013
0.007
0.006
0.005
0.006
0.008
PV
0.000
0.014
0.030
0.037
0.025
0.025
0.025
Total
3.6
4.8
4.9
5.0
5.0
4.9
5.1
Table 4-95: PFC, HFC, SFe, NF3, and N2O Emissions from Electronics Manufacture (metric
tons)
Year
1990
2005
2014
2015
2016
2017
2018
cf4
115
145
201
206
209
219
233
c2f6
160
161
114
108
98
95
91
CsFs
0
9
15
15
14
11
12
C4Fs
0
11
6
6
5
6
6
HFC-23
15
14
21
22
23
25
25
sf6
22
30
32
32
36
31
33
nf3
3
28
30
34
34
35
37
n2o
120
412
734
793
791
922
857
Total
435
811
1,153
1,216
1,210
1,344
1,294
Table 4-96: F-HTF Emissions from Electronics Manufacture by Compound Group (metric
tons)
Year
2012
2013
2014
2015
2016
2017
2018
HFCs
1.3
0.9
2.0
1.6
2.7
1.6
1.5
PFCs
1.1
0.4
0.2
0.3
0.3
0.2
0.4
sf6
0.5
0.4
0.9
0.6
0.5
0.7
0.6
HFEs
26.1
29.0
25.2
18.9
13.5
16.5
23.5
PFPMIEs
21.9
18.1
18.2
20.7
17.3
14.3
18.3
Perfluoalkylromorpholines
10.7
10.7
10.8
8.1
7.6
5.2
5.9
Perfluorotrialkylamines
45.6
29.5
49.3
43.7
38.6
37.6
42.5
Total F-HTFs
107.3
89.1
106.5
93.9
80.4
76.2
92.6
Table 4-97: F-GHGa Emissions from PV and MEMS manufacturing (MMT CO2 Eq.)
Year
1990
2005
2014
2015
2016
2017
2018
MEMS
0.0
0.013
0.007
0.006
0.005
0.006
0.008
PV
0.0
0.014
0.035
0.030
0.025
0.025
0.025
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.
Methodology
Emissions are based on data reported through Subpart I, Electronics Manufacture, of EPA's GHGRP, Partner-
reported emissions data received through EPA's PFC81 Reduction/Climate Partnership, EPA's PFC Emissions Vintage
Model (PEVM)—a model that estimates industry emissions from etching and chamber cleaning processes in the
80	Emissions of N20 from semiconductor manufacturing are reported in the CRF under 2H3.
81	In the context of the EPA Partnership and PEVM, PFC refers to perfluorocompounds, not perfluorocarbons.
Industrial Processes and Product Use 4-113

-------
absence of emission control strategies (Burton and Beizaie 2001),82 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 2018 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 2018. 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 2018. The methodology discussion below for these time
periods focuses on semiconductor emissions from etching, chamber cleaning, and uses of N2O. 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 2018. 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 CO2 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 calculating the ratio of manufacturing capacity of reporters to non-reporters and
then multiplying this ratio by the reported emissions, to calculate the total U.S. manufacturing emissions.
Manufacturing capacities in megawatts were drawn from a 2015 Congressional Research Service Report on U.S.
Solar Photovoltaic Manufacturing83 and self-reported capacity by the GHGRP reporter84 EPA estimated that during
the 2015 to 2017 period, 28 percent of emissions were reported through the GHGRP. These emissions are
estimated for the full time series by linearly scaling the total U.S. capacity between zero in 1997 to the total
capacity reported in the Congressional Research Service in 2012. Capacities were held constant for non-reporters
for 2012 to 2018. Emissions per MW from the GHGRP reporter in 2015 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, 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 2018. EPA estimates the emissions of F-HTFs from non-reporting facilities by
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
82	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.
83	Platzer, Michaela D. (2015) U.S. Solar Photovoltaic Manufacturing: Industry Trends, Global Competition, Federal Support.
Congressional Research Service. January 27, 2015. < https://fas.org/sgp/crs/misc/R42509.pdf>.
84	.
4-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).85 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),86 and (2) product type (discrete, memory or
logic).87 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. The emission factor is used to estimate world uncontrolled emissions using publicly-available data on
world silicon consumption.
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, C4F8, HFC-23, SF6, and NF3) used in semiconductor
manufacturing. Through reporting under Subpart I, data on other F-GHGs (C4F6, CsFs, HFC-32, HFC-41, HFC-134a)
85	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.
86	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).
87	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.
Industrial Processes and Product Use 4-115

-------
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 N2O emissions, it is assumed the proportion of N2O 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, CsFs, HFC-32, HFC-41, HFC-134a) were estimated using the
method described above for 1990 to 1994.
For this time period, the N2O 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 N2O 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 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.
emissions figure by the non-Partner share of U.S. total silicon capacity for each year as described above.88 Gas-
specific emissions from non-Partners were estimated using linear interpolation of gas-specific emission distribution
of 1999 (assumed same as total U.S. Industry in 1994) and 2011 (calculated from a subset of non-Partner facilities
from GHGRP reported emissions data). Annual updates to PEVM reflect published figures for actual silicon
88 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.
4-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).89,90,91
For this time period emissions of other F-GHGs (C4F6, CsFs, 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.92 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
figure was assumed (SIA 2009).
In addition, publicly-available actual 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
89	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 2-3 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.
90	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.
91	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.
92	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.
Industrial Processes and Product Use 4-117

-------
emissions for non-Partners were estimated using the same method as for 2000 through 2006.
For this time period emissions of other F-GHGs (CsFs, 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 CO2 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 GaAs technology in addition to Si technology.93 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 N2O emissions from CVD and other processes. The F-GHGs and N2O 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
estimated site-specific DRE,94 if a site-specific DRE was indicated), and the fab-wide DREs reported in
2014.95 To adjust emissions for facilities that abated emissions in 2011 through 2013, EPA first estimated
93	GaAs and Si technologies refer to the wafer on which devices are manufactured, which use the same PFCs but in different
ways.
94	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.
95	If abatement information was not available for 2014 or the reported incorrectly in 2014, data from 2015 or 2016 was
substituted.
4-118 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.96
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 N2O and estimates of manufacturing activity. The new emission factors (in units of mass of CO2
Eq./TMLA [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).97 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 or less 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, C4F8, CHF3, SF6 and NF3)98 were regressed
against the corresponding TMLA to estimate an aggregate F-GHG emissions factor (CO2 Eq./MSI TMLA), and
facility-reported N2O emissions were regressed against the corresponding TMLA to estimate a N2O emissions
factor (CO2 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.
Non-reporting fabs were then broken out into similar subpopulations by wafer size 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
96	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.
97	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 POU abatement. These fabs were therefore excluded from the regression analysis. (They are still
included in the national totals.)
98	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.
Industrial Processes and Product Use 4-119

-------
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 N2O 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 N2O for 2011, 2012, and 2015
through 2018, resulting in one set of proportions for F-GHGs and one set for N2O, 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 affected 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 2018
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 2018, 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 2018 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 CO2 Eq.-weighted controlled F-GHG and N2O emissions (emissions after the use of abatement) divided by
total fab CO2 Eq.-weighted uncontrolled F-GHG and N2O 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 N2O) 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 N2O), and wafer size (200 mm 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 N2O 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.
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 2b Method in the 2006IPCC Guidelines, this
method uses different emission and byproduct generation factors for different F-GHGs and process types, but it
goes beyond the Tier 2b Method by requiring use of updated factors for different wafer sizes (i.e., 300mm vs. 150
and 200mm) and CVD clean subtypes (in situ thermal, in situ thermal, 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 (40 CFR Part 98). For the years 2011 through 2013 reported emissions,
semiconductor manufacturers used older emission factors to estimate their F-GHG emissions (Federal Register /
4-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2006IPCCGuidelines. 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.nzo) + Non-Reporters' Estimated N2O Emissions
(Enr,N2o)
where Er and Enr denote totals for the indicated subcategories of emissions for F-GHG and N2O, 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
Facilities under Subpart I, docket EPA-HQ-OAR-2011-0028).99 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
99 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 4-121

-------
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 of 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
SFs. 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 Er,f-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 N2O consumed and the N2O 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 N2O consumed was assumed to be 20 percent. Consumption
of N2O for GHGRP reporting facilities was estimated by back-calculating from emissions reported and assuming no
abatement. The quantity of N2O 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 N2O
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
each of the GHGRP reporting, INhO-emitting facilities. The uncertainty for the total reported N2O emissions was
then estimated by combining the uncertainties of each of the 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
4-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 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.
For simplicity, the results of the Monte Carlo simulations on the bounds of the gas- and wafer size-specific
emissions as well as the TMLA and emission factors are assumed to be normally distributed and the uncertainty
bounds are assigned at 1.96 standard deviations around the estimated mean. The departures from normality were
observed to be small.
The final step in estimating the uncertainty in emissions of non-reporting facilities is convolving the distribution of
emission factors with the distribution of TMLA using Monte Carlo simulation.
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 N2O emissions from semiconductor
manufacturing were estimated to be between 4.7 and 5.3 MMT CO2 Eq. at a 95 percent confidence level. This
range represents 6 percent below to 6 percent above the 2018 emission estimate of 5.0 MMT CO2 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.
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


2018 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimateb


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



Lower
Upper
Lower Upper



Boundc
Boundc
Bound Bound
Semiconductor
HFC, PFC, SF6,
5.0
4.7
5.3
-6% 6%
Manufacture
NFs, and N20
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.
Industrial Processes and Product Use 4-123

-------
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, and 2018, which are included in the overall emissions estimates,
were based on an updated set of default emission factors. This may have affected the trend seen between 2013
and 2014 (a 24 percent increase), which reversed the trend seen between 2011 and 2013. As discussed in the
Planned Improvements section, EPA is planning to conduct analysis to determine how much of the 2013 to 2014
trend may be attributable to the updated factors and to improve time-series consistency.
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).100 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 2018 were updated to reflect updated emissions reporting in EPA's GHGRP, relative
to the previous Inventory. Additionally, non-reporter estimates were revised. EPA identified several facilities that
report to the GHGRP but were being categorized as non-reporters, causing an over-estimation of non-reporter
TMLA and consequently non-reporter emissions. Together these revisions resulted in an average change of 4
percent through the 2011 through 2018 timeseries.
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.
Emission factors for semiconductor processes have also been revised overtime. Recently, the 2011 to 2013
portion of the inventory was updated to reflect emission factors and DREs that were revised in 2013 to improve
times series consistency. However, the effects of these revisions have not yet been applied to the 2000 to 2010
portion of the time series.
100 GHGRP Report Verification Factsheet. .
4-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 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).
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 1990.101 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 4-100.102
Table 4-99: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.)
Gas
1990
2005
2014
2015
2016
2017
2018
HFC-23
0.0
+
+
+
+
+
+
HFC-32
0.0
0.3
3.4
3.9
4.6
5.3
6.0
HFC-125
+
9.0
40.0
43.4
47.0
50.0
53.3
HFC-134a
+
81.3
76.7
75.5
71.2
66.4
63.4
HFC-143a
+
9.4
26.9
27.6
28.3
28.0
27.7
HFC-236fa
0.0
1.2
1.4
1.3
1.3
1.2
1.2
cf4
0.0
+
+
+
+
+
0.1
Others3
0.2
7.3
12.5
13.9
15.0
15.8
16.2
Total
0.2
108.5
161.0
165.8
167.3
166.9
167.9
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.
101	[42 U.S.C § 7671, CAA Title VI],
102	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 4-125

-------
Table 4-100: Emissions of HFCs and PFCs from ODS Substitution (Metric Tons)
Gas
1990
2005
2014
2015
2016
2017
2018
HFC-23
0
1
2
2
2
2
2
HFC-32
0
397
5,001
5,841
6,799
7,799
8,821
HFC-125
+
2,583
11,439
12,403
13,416
14,291
15,243
HFC-134a
+
56,863
53,636
52,813
49,791
46,468
44,362
HFC-143a
+
2,096
6,011
6,183
6,326
6,272
6,198
HFC-236fa
0
118
145
134
129
124
118
cf4
0
2
5
6
6
6
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.103 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 167.9 MMT CO2 Eq. emitted in 2018. 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 2018. The
end-use sectors that contributed the most toward emissions of HFCs and PFCs as ODS substitutes in 2018 include
refrigeration and air-conditioning (128.9 MMT CO2 Eq., or approximately 77 percent), aerosols (19.2 MMT CO2 Eq.,
or approximately 11 percent), and foams (15.1 MMT CO2 Eq., or approximately 9 percent). Within the refrigeration
and air-conditioning end-use sector, large retail food was the highest emitting end-use (33.1 MMT CO2 Eq.),
followed by motor vehicle 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
2014
2015
2016
2017
2018
Refrigeration/Air Conditioning
+
89.7
122.5
124.8
126.5
126.8
128.9
Aerosols
0.2
11.9
22.6
23.5
22.1
20.7
19.2
Foams
+
4.1
11.8
13.4
14.5
15.0
15.1
Solvents
+
1.7
1.8
1.8
1.9
1.9
2.0
103 R-404A contains HFC-125, HFC-143a, and HFC-134a.
4-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Fire Protection	+	1.1	2.2	2.3	2.4	2.5	2.6
Total	0.2	108.5	161.0 165.8 167.3 166.9 167.9
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
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,104 R-404A, and R-507A.105 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 CO2 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 CO2 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.
104	R-410A contains HFC-32 and HFC-125.
105	R-507A, also called R-507, contains HFC-125 and HFC-143a.
Industrial Processes and Product Use 4-127

-------
Solvents
Chlorofluorocarbons, methyl chloroform (1,1,1-trichloroethane or TCA), and to a lesser extent carbon tetrachloride
(CCU) 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
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.
fviet had ©logy
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 68 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 68
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-
4-128 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. The most
significant sources of uncertainty for this source category include the total stock of refrigerant installed in
industrial process refrigeration and cold storage equipment, as well as the emission factor for refrigerant installed
in industrial process refrigeration and cold storage equipment.
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 166.5 and 185.4 MMT CO2 Eq.
at the 95 percent confidence level. This indicates a range of approximately 0.8 percent below to 10.5 percent
above the emission estimate of 167.9 MMT CO2 Eq.
Table 4-102: Approach 2 Quantitative Uncertainty Estimates for HFC and PFC Emissions
from ODS Substitutes (MMT CO2 Eq. and Percent)


2018 Emission


Source
Gases
Estimate
Uncertainty Range Relative to Emission Estimate3


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



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Substitution of Ozone
Depleting Substances
HFCs and
PFCs
167.9
166.5 185.4
-0.8% +10.5%
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 time-series consistency from 1990
through 2018. 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)106 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
106 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 4-129

-------
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
United States.107 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, including atmospheric measurements of HFC emissions for the United States and EPA's GHGRP,
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.108 It is assumed that the total demand
equals the amount supplied by either new production, chemical import, or quantities recovered (usually
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 nine 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
107	Chemical that is exported, transformed, or destroyed—unless otherwise imported back to the United States—will never be
emitted in the United States.
108	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.
4-130 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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 4-2 compare the published net supply of saturated HFCs (excluding HFC-23) in MMT CO2
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 2018 (U.S. EPA 2019a) and the chemical demand as calculated
by the Vintaging Model for the same time series. 2018 Subpart 00 GHGRP values are not yet publicly available and
are proxied using the last available estimate value, 2017, plus the average year-to-year change since the start of
EPA's GHGRP.
Table 4-103: U.S. HFC Supply (MMT COz Eq.)

2010
2011
2012
2013
2014
2015
2016
2017
2018
Reported Net Supply (GHGRP)
235
248
245
295
279
290
268
317
326
Industrial GHG Suppliers
235
241
227
278
254
264
240
285
292
HFCs in Products and Foams3
NA
7
18
17
25
26
28
32
34
Modeled Supply (Vintaging Model)
264
269
274
279
286
285
287
273
274
Percent Difference
12%
9%
12%
-5%
2%
-2%
7%
-14%
-16%
NA (Not Available)
a Importers and exporters of fluorinated gases in products were not required to report 2010 data.
Industrial Processes and Product Use 4-131

-------
Figure 4-2: U.S. HFC Consumption (MMT CO2 Eq.)
¦	Reported Imports in Products and Foams
¦	Modeled Consumption
¦	Reported Bulk Supply
2010 2011 2012 2013 2014 2015 2016 2017 2018
350 -i
As shown, the estimates from the Vintaging Model are higher than the GHGRP estimates by an average of 0.6
percent across the time series (i.e., 2010 through 2018). Potential reasons for the differences between the
reported and modeled data, include:
•	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 CO2 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 generally increases over time (although some year-to-
year variations exist) and the Vintaging Model estimates also increase (through 2016), 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 and 2018
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
4-132 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
than or equal to 25,000 metric tons of CO2 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.
• 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 (MMT CCh Eq.)

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

Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Reported Net
Supply (GHGRP)
242
247
270
287
285
279
293
322
Modeled Demand
(Vintaging Model)
266
272
277
283
285
286
280
274
Percent Difference
10%
10%
2%
-2%
0%
2%
-4%
-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
through 2016, and a slight lowering after that, actual consumption for specific chemicals or equipment
may vary over 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 OO (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 2018 emissions from
Industrial Processes and Product Use 4-133

-------
that non-modeled source (0.1 MMT CO2 Eq.) are much smaller than those for the ODS substitute sector
(167.9 MMT CO2 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:
•	Renaming the non-metered dose inhaler (non-MDI) aerosol end-use to consumer aerosol and updating
stock and emission estimates to align with a recent national market characterization.
•	Adding a technical aerosol end-use to the aerosols sector, in order to capture a portion of the market that
was not adequately encompassed by the former non-MDI aerosol end-use (EPA 2019b).
•	Correcting the lifetime for streaming agents, which was changed from 18 years to 24 years, within the Fire
Protection sector.
•	Renaming the polyurethane rigid spray foam end-use to high pressure two-component spray foam and
updating market size and foam blowing agent transition assumptions to align with stakeholder input and
market research (EPA 2020).
•	Adding a low pressure two-component spray foam end-use to the foams sector, in order to capture a
portion of the market that was not adequately encompassed by the former polyurethane rigid spray foam
end-use (EPA 2020).
Together, these updates increased greenhouse gas emissions on average by 3.3 percent between 1990 and 2017.
Planned Improvements
Future improvements to the Vintaging Model are planned for the Foam Blowing sector. Blowing agent transitions
and quantities for specific equipment types are under review for commercial refrigeration foam to determine if the
end-use can be disaggregated to align with refrigeration end-uses.
4.25 Electrical Transmission and Distribution
(CRF Source Category 2G1)
The largest use of sulfur hexafluoride (SFs), 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.
4-134 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 4.1 MMT CO2 Eq. (0.2 kt) in 2018. 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 2018. 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 33 percent from 2011 to
2018,109 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).
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	2Z8	03	23.2
2005	7.7	0.7	8.4
2014	4.4	0.4	4.8
2015	3.5	0.3	3.8
2016	3.8	0.3	4.1
2017	3.8	0.3	4.1
2018	3.7	0.3	4.1
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
2014	0.2
2015	0.2
109 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 GHG emissions below 15,000 metric tons of carbon dioxide
equivalent (mtC02 Eq.) for three consecutive years or below 25,000 mtC02 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 mtC02 Eq., which may occur for a variety of reasons, including changes in facility size
and changes in emission rates.
Industrial Processes and Product Use 4-135

-------
2016
2017
2018
0.2
0.2
0.2
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.110 (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 SFs) = SF6 purchased to refill existing equipment (kilograms) + nameplate capacity of retiring
equipment (kilograms)111
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 SFs 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
110	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.
111	Nameplate capacity is defined as the amount of SF6 within fully charged electrical equipment.
4-136 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
for each respective year by the estimated U.S. emissions of SF6 from electric power systems in 1999 (estimated to
be 13.6 MMT C02 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
which case sales will rise more quickly than emissions. This effect was accounted for by applying 3-year smoothing
to utility SFs 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 2018 Emissions from Electric Power Systems
Emissions from electric power systems from 1999 to 2018 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 2018, Partner utilities, which for inventory purposes are defined as utilities that
either currently are or previously have been part of the Partnership,112 represented 50 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 2018, approximately 1 percent of the total
emissions attributed to Partner utilities were reported through Partnership reports. Approximately 93 percent of
the total emissions attributed to Partner utilities were reported and verified through EPA's GHGRP. Partners
without verified 2018 data accounted for approximately 6 percent of the total emissions attributed to Partner
utilities.113
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 CO2 Eq.) for three consecutive years or below 25,000 MT CO2 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-
112	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.
113	Only data reported as of August 4, 2019 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 4-137

-------
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
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 24 percent of U.S. transmission miles and 23
percent of estimated U.S. emissions from electric power system in 2018.114
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.115 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,
114	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.
115	In the United States, SF6 is contained primarily in transmission equipment rated above 34.5 kV.
4-138 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
the emissions data from both groups were combined to develop regression equations for 2012. This was
repeated for 2013 through 2018 using Partner and GHGRP-Only Reporter data for each year.
o The 2018 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 70 percent of total U.S. transmission miles). The
regression equation for 2018 is:
Emissions (kg) = 0.221 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 2018.
Table 4-107: Transmission Mile Coverage (Percent) and Regression Coefficients (kg per

1999
2005
2014
2015
2016
2017
2018
Percentage of Miles Covered by Reporters
50%
50%
74%
73%
73%
74%
70%
Regression Coefficient3
0.71
0.35 /
0.23
0.19
0.21
0.24
0.22
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 2018 was calculated to be 0.6 percent, as
transmission miles increased by approximately 30,000 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 2018 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 2018 Emissions from Manufacture of Electrical Equipment
Three different methods were used to estimate 1990 to 2018 emissions from original electrical equipment
manufacturers (OEMs).
Industrial Processes and Product Use 4-139

-------
•	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
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 CO2 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
SFs 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 2018 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 5.2 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.116 Based on a
Monte Carlo analysis, the cumulative uncertainty of all GHGRP-Only reported data was estimated to be 8.8
percent.
There are two sources of uncertainty associated with the regression equations used to estimate emissions in 2016
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 3.5 and 4.7 MMT CO2 Eq. at the 95
116 Uncertainty is assumed to be higher for the GHGRP-Only category, because 2011 is the first year that those utilities have
reported to EPA.
4-140 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
percent confidence level. This indicates a range of approximately 13 percent below and 15 percent above the
emission estimate of 4.1 MMT CO2 Eq.
Table 4-108: Approach 2 Quantitative Uncertainty Estimates for SF6 Emissions from
Electrical Transmission and Distribution (MMT CO2 Eq. and Percent)


2018 Emission



Source
Gas
Estimate
Uncertainty Range Relative to 2018 Emission Estimate3


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




Lower Upper
Lower
Upper



Bound Bound
Bound
Bound
Electrical Transmission
and Distribution
sf6
4.1
3.5 4.7
-13%
+15%
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
SFs 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 2018. 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).117 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 2017.
117 GHGRP Report Verification Factsheet. .
Industrial Processes and Product Use 4-141

-------
•	GHGRP report resubmissions: Historical estimates for the period 2011 through 2017 were updated
relative to the previous report based on revisions to reported historical data in EPA's GHGRP.
•	Missing report gap-filling: Previously, only missing data from Partner utilities were gap-filled for, while
GHGRP-only utilities with missing data were considered non-reporters. Between 2011 and 2018, missing
data is interpolated between reporting years for all reporting utilities. Data is extrapolated for three years
if a reporting utility has stopped reporting using a utility specific transmission mile growth rate for 2011
through 2016 and an industry-wide growth rate for 2017 and 2018. See methodology section for more
information.
•	Nameplate capacity: The previous year's methodology determined the end of year nameplate capacity by
summing the Beginning of Year Nameplate Capacity and the Net Increase in Nameplate Capacity for the
GHGRP reporters, which aggregates a small portion of hermetically sealed equipment and high-voltage
equipment. Beginning in the 2017 reporting year, EPA's GHGRP required that reporters distinguish
between the nameplate capacity of non-hermetically sealed equipment from equipment that is
hermetically sealed. EPA now calculates the end of year nameplate capacity for 2010 to 2017 by using the
reported beginning of year nameplate capacity reported for the following year. For 2018, the last year in
the time series, the end of year nameplate was determined by using the reported beginning of year
nameplate and the net increase in non-hermetically sealed equipment. If, however, a facility stopped
reporting prior to 2017, the previous inventory's methodology (i.e., summing the Beginning of Year
Nameplate Capacity and the Net Increase in Nameplate Capacity) was used to determine the end of year
nameplate capacity with the net increase in nameplate capacity scaled down to adjust for the nameplate
capacity of hermetically sealed equipment. EPA calculated the adjustment factor by taking the net
increase in non-hermetically sealed equipment divided by the total net increase of both hermetically and
non-hermetically sealed equipment using data from the 2017 and 2018 reporting years.
•	Transmission miles: First, this inventory year's methodology interpolates between known years of UDI
facility-specific transmission mile data and calculates a growth rate year to year on these interpolated
values; whereas, previously, UDI transmission mile data growth was assumed to be the same for all
facilities for years where EPA did not have data and did not result in an accurate gap-filling methodology.
Estimates from 1990 through 1998 were updated as a result of recalculations made to some Partner
transmission mile growth rates which caused a recalculation to the 1999 U.S. emission estimate. As
discussed in the Methodology above, the 1990 to 1998 estimates are based, in part, on the emissions
estimated for this source category in 1999. Second, a correction was made to address an incorrect growth
rate being used for extrapolating for transmission miles for all utilities from last year's inventory.
As a result of the recalculations, SF6 emissions from electrical transmission and distribution decreased by 3.9
percent for 2017 relative to the previous report, and SF6 nameplate capacity decreased by 3.5 percent for 2017
relative to the previous report. On average, SF6 emission estimates for the entire time series decreased by
approximately 0.18 percent per year.
Planned Improvements
EPA plans to more closely examine transmission miles data by company provided by the UDI data sets, which are
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.
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.
4-142 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
4.26 Nitrous Oxide from Product Uses (CRF
Source Category 2G3)
Nitrous oxide (N2O) 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 N2O that is actually emitted depends upon the
specific product use or application.
There are a total of three N2O production facilities currently operating in the United States (Ottinger 2014). 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 N2O is as a
propellant in pressure and aerosol products, the largest application being pressure-packaged whipped cream.
Small quantities of N2O 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 N2O in 2018 was approximately 15 kt (see Table 4-109).
Table 4-109: N2O Production (kt)
Year kt
1990 16
2005 15
2014	15
2015	15
2016	15
2017	15
2018	15
Nitrous oxide emissions were 4.2 MMT CO2 Eq. (14 kt N2O) in 2018 (see Table 4-110). Production of N2O 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 N2O. The use
of N2O 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
2014
4.2
14
2015
4.2
14
2016
4.2
14
2017
4.2
14
2018
4.2
14
Industrial Processes and Product Use 4-143

-------
Methodology
Emissions from N2O product uses were estimated using the following equation:
Epu = X Sa x ERa)
a
where,
Sa
ER;
p
a
N2O emissions from product uses, metric tons
Total U.S. production of N2O, metric tons
specific application
Share of N2O usage by application a
Emission rate for application a, percent
The share of total quantity of N2O usage by end-use represents the share of national N2O produced that is used by
the specific subcategory (e.g., anesthesia, food processing). In 2018, the medical/dental industry used an
estimated 86.5 percent of total N2O produced, followed by food processing propellants at 6.5 percent. All other
categories combined used the remainder of the N2O produced. This subcategory breakdown has changed only
slightly over the past decade. For instance, the small share of N2O usage in the production of sodium azide has
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 N2O 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). The N2O was allocated across the following categories: medical applications, food
processing propellant, and sodium azide production (pre-1996). A usage emissions rate was then applied for each
sector to estimate the amount of N2O emitted.
Only the medical/dental and food propellant subcategories were estimated to release emissions into the
atmosphere, and therefore these subcategories were the only usage subcategories with emission rates. For the
medical/dental subcategory, due to the poor solubility of N2O in blood and other tissues, none of the N2O 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 N2O used as a propellant in
pressurized and aerosol food products, none of the N2O is reacted during the process and all of the N2O 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 N2O is consumed/reacted during the process, and therefore the emission rate
was considered to be zero percent (Tupman 2003).
The 1990 through 1992 N2O production data were obtained from SRI Consulting's Nitrous Oxide, North America
report (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 N2O 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 N2O emissions for years 1993 through 2001
(Tupman 2003). The 2002 and 2003 N2O 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 N2O production to range between 13.6 and 15.9 thousand
metric tons. Due to the unavailability of data, production estimates for years 2004 through 2018 were held
constant at the 2003 value.
The 1996 share of the total quantity of N2O used by each subcategory was obtained from SRI Consulting's Nitrous
Oxide, North America report (Heydorn 1997). The 1990 through 1995 share of total quantity of N2O used by each
subcategory was kept the same as the 1996 number provided by SRI Consulting. The 1997 through 2001 share of
4-144 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
total quantity of N2O usage by sector was obtained from communication with a N2O industry expert (Tupman
2003). The 2002 and 2003 share of total quantity of N2O usage by sector was obtained from CGA (2002, 2003). Due
to the unavailability of data, the share of total quantity of N2O usage data for years 2004 through 2018 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 report (Heydorn 1997), and confirmed by a N2O industry
expert (Tupman 2003). The emissions rate for all other subcategories was obtained from communication with a
N2O industry expert (Tupman 2003). The emissions rate for the medical/dental subcategory was obtained from the
2006IPCC Guidelines.
Uncertainty and Time-Series Consistency
The overall uncertainty associated with the 2018 N2O emission estimate from N2O product usage was calculated
using the 2006 IPCC Guidelines (2006) Approach 2 methodology. Uncertainty associated with the parameters used
to estimate N2O 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 N2O product usage were estimated to be between 3.2 and 5.2 MMT CO2 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 CO2 Eq.
Table 4-111: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from N2O
Product Usage (MMT CO2 Eq. and Percent)
Source
Gas
2018 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.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2018. 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 2006 IPCC 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 2017 portion of the time series.
Planned Improvements
EPA has 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 N2O 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.
Industrial Processes and Product Use 4-145

-------
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 N2O. Additionally, planned improvements include considering imports and exports of N2O for
product uses.
Finally, for future Inventories, EPA will examine data from EPA's GHGRP to improve the emission estimates for the
N2O 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 UNFCCC118 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 (SO2). 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 SO2, 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
"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 2018
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
2014
2015
2016
2017
2018
NOx
592
572
414
414
414
414
414
Industrial Processes


f




Other Industrial Processes3
343
437
300
300
300
300
300
Metals Processing
88
60
63
63
63
63
63
Chemical and Allied Product


*




Manufacturing
152
55
43
43
43
43
43
118 See .
4-146 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Storage and Transport
3
15
5
5
5
5
5
Miscellaneous15
5
2
2
2
2
2
2
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
O
4,129
1,557
1,251
1,251
1,251
1,251
1,251
Industrial Processes







Metals Processing
2,395
752
553
553
553
553
553
Other Industrial Processes3
487
484
530
530
530
530
530
Chemical and Allied Product







Manufacturing
1,073
189
117
117
117
117
117
Miscellaneous15
101
32
42
42
42
42
42
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
IMVOCs
7,638
5,849
3,815
3,815
3,815
3,815
3,815
Industrial Processes







Storage and Transport
1,352
1,308
613
613
613
613
613
Other Industrial Processes3
364
414
314
314
314
314
314
Chemical and Allied Product







Manufacturing
575
213
70
70
70
70
70
Metals Processing
111
45
26
26
26
26
26
Miscellaneous15
20
17
24
24
24
24
24
Product Uses







Surface Coating
2,289
1,578
1,134
1,134
1,134
1,134
1,134
Non-Industrial Processes0
1,724
1,446
1,039
1,039
1,039
1,039
1,039
Degreasing
675
280
202
202
202
202
202
Dry Cleaning
195
230
165
165
165
165
165
Graphic Arts
249
194
139
139
139
139
139
Other Industrial Processes3
85
88
63
63
63
63
63
Other
+
36
26
26
26
26
26
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt
NA (Not Available)
3 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 2018 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2019), and disaggregated based on EPA (2003). Data were
collected for emissions of CO, NOx, volatile organic compounds (VOCs), and SO2 from metals processing, chemical
Industrial Processes and Product Use 4-147

-------
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 2018. 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 2017 portion of the time series.
4-148 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
5. Agriculture
Agricultural activities contribute directly to emissions of greenhouse gases through a variety of processes. This
chapter provides an assessment of methane (Cm) and nitrous oxide (N2O) 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 (CO2) emissions from liming and urea fertilization (see Figure
5-1). Additional CO2, CFU and N2O fluxes from agriculture-related land-use and land-use conversion activities, such
as cultivation of cropland, grassland fires and conversion of forest land to cropland, are presented in the Land Use,
Land-Use Change, and Forestry (LULUCF) chapter. Carbon dioxide emissions from on-farm energy use are reported
in the Energy chapter.
Figure 5-1: 2018 Agriculture Chapter Greenhouse Gas Emission Sources (MMT CO2 Eq.)
Agricultural Soil Management
Enteric Fermentation
Agriculture as a Portion of
All Emissions
Manure Management
Rice Cultivation
Urea Fertilization
Energy
¦ Agriculture
IPPU
Waste
Field Burning of Agricultural Residues
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320
MMT CO2 Eq.
In 2018, the Agriculture sector was responsible for emissions of 618.5 MMT CO2 Eq.,1 or 9.3 percent of total U.S.
greenhouse gas emissions.2 Methane emissions from enteric fermentation and manure management represent
28.0 percent and 9.7 percent of total CFU 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 N2O by agricultural soil management through
activities such as fertilizer application and other agricultural practices that increased nitrogen availability in the soil
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.
2	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 are
not included.
Agriculture 5-1

-------
was the largest source of U.S. N2O emissions, accounting for 77.8 percent. Manure management and field burning
of agricultural residues were also small sources of N2O emissions. Urea fertilization and liming each accounted for
0.1 percent of total CO2 emissions from anthropogenic activities.
Table 5-1 and Table 5-2 present emission estimates for the Agriculture sector. Between 1990 and 2018, CO2 and
Cm emissions from agricultural activities increased by 16.0 percent and 16.2 percent, respectively, while N2O
emissions from agricultural activities fluctuated from year to year, but increased by 8.4 percent overall.
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 2017)
to ensure that the trend is accurate. This year's major updates include (1) Manure Management: updated waste
management system distribution data for dairy cows; and (2) Agricultural Soil Management: incorporating the
2015 National Resources Inventory along with new data on crop histories, and updating the DayCent soil process
model to extend soil depth from 20 to 30 centimeters and capture the effects of freeze thaw. In total, the
improvements made to the Agriculture sector in this Inventory increased greenhouse gas emissions by 60.2 MMT
CO2 Eq. (11 percent) in 2017. For more information on specific methodological updates, please see the
Recalculations discussions within the respective source category sections of this chapter.
Table 5-1: Emissions from Agriculture (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
2018
CO?
6.7
7.5
7.5
7.8
7.1
7.6
7.7
Urea Fertilization
2.0
3.1
3.9
4.1
4.0
4.5
4.6
Liming
4.7
4.3
3.6
3.7
3.1
3.1
3.1
ch4
217.6
238.8
234.3
241.0
245.3
248.4
253.0
Enteric Fermentation
164.2
168.9
164.2
166.5
171.8
175.4
177.6
Manure Management
37.1
51.6
54.3
57.9
59.6
59.9
61.7
Rice Cultivation
16.0
18.0
15.4
16.2
13.5
12.8
13.3
Field Burning of Agricultural Residues
0.3
0.4
0.4
0.4
0.4
0.4
0.4
n2o
330.1
329.6
366.7
365.8
348.1
346.2
357.8
Agricultural Soil Management
315.9
313.0
349.2
348.1
329.8
327.4
338.2
Manure Management
14.0
16.4
17.3
17.5
18.1
18.7
19.4
Field Burning of Agricultural Residues
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Total
554.4
575.9
608.6
614.6
600.5
602.3
618.5
Note: Totals may not sum due to independent
rounding.






ible 5-2: Emissions from Agriculture (kt)






Gas/Source
1990
2005
2014
2015
2016
2017
2018
CO?
6,678
7,499
7,532
7,819
7,122
7,594
7,745
Urea Fertilization
2,011
3,150
3,923
4,082
4,041
4,514
4,598
Liming
4,667
4,349
3,609
3,737
3,081
3,080
3,147
ch4
8,705
9,553
9,371
9,639
9,813
9,938
10,119
Enteric Fermentation
6,566
6,755
6,567
6,660
6,874
7,016
7,103
Manure Management
1,485
2,062
2,172
2,316
2,385
2,395
2,467
Rice Cultivation
640
720
616
648
539
510
533
Field Burning of Agricultural Residues
14
16
16
16
16
16
16
N20
1,108
1,106
1,231
1,227
1,168
1,162
1,201
Agricultural Soil Management
1,060
1,050
1,172
1,168
1,107
1,099
1,135
Manure Management
47
55
58
59
61
63
65
Field Burning of Agricultural Residues
1
1
1
1
1
1
1
Note: Totals may not sum due to independent rounding.
5-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 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 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
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 Cm as a byproduct, which can be exhaled or eructated by the animal. The
amount of Cm produced and emitted by an individual animal depends primarily upon the animal's digestive
system, and the amount and type of feed it consumes.
Ruminant animals (e.g., cattle, buffalo, sheep, goats, and camels) are the major emitters of CFU 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 Cm emissions per unit of body mass among all animal
types.
Non-ruminant animals (e.g., swine, horses, and mules and asses) also produce Cm emissions through enteric
fermentation, although this microbial fermentation occurs in the large intestine. These non-ruminants emit
significantly less CFU on a per-animal-mass basis than ruminants because the capacity of the large intestine to
produce CFU is lower.
In addition to the type of digestive system, an animal's feed quality and feed intake also affect Cm emissions. In
general, lower feed quality and/or higher feed intake leads to higher Cm 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 CFU
emissions in 2018 were 177.6 MMT CO2 Eq. (7,103 kt). Beef cattle remain the largest contributor of CH4 emissions
Agriculture 5-3

-------
from enteric fermentation, accounting for 72 percent in 2018. Emissions from dairy cattle in 2018 accounted for 25
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
2014
2015
2016
2017
2018
Beef Cattle
119.1
125.2
116.5
118.0
123.0
126.3
128.1
Dairy Cattle
39.4
37.6
42.0
42.6
43.0
43.3
43.6
Swine
2.0
2.3
2.4
2.6
2.6
2.7
2.8
Horses
1.0
1.7
1.5
1.4
1.4
1.3
1.2
Sheep
2.3
1.2
1.0
1.1
1.1
1.1
1.1
American Bison
0.1
0.4
0.4
0.4
0.4
0.4
0.4
Goats
0.3
0.4
0.3
0.3
0.3
0.3
0.3
Mules and Asses
+
0.1
0.1
0.1
0.1
0.1
0.1
Total
164.2
168.9
164.2
166.5
171.8
175.4
177.6
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

2014
2015
2016
2017
2018
Beef Cattle
4,763
5,007

4,660
4,722
4,919
5,052
5,125
Dairy Cattle
1,574
1,503

1,679
1,706
1,722
1,730
1,744
Swine
81
92

96
102
105
108
111
Horses
40
70

60
57
54
51
48
Sheep
91
49

42
42
42
42
42
American Bison
4
17

14
14
15
15
15
Goats
13
14

13
13
13
13
14
Mules and Asses
1
2

3
3
3
3
3
Total
6,566
6,755

6,567
6,660
6,874
7,016
7,103
Note: Totals may not sum due to independent rounding.
From 1990 to 2018, emissions from enteric fermentation have increased by 8.2 percent. Emissions have also
increased from 2017 to 2018 by 1.2 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 4.6 percent over the entire time series, the population has declined by 2.6
percent, and milk production increased 57 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 2018,
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
U.S.). Dairy cattle emissions have continued to trend upward since 2007, in line with dairy cattle population
3 Enteric fermentation emissions from camels and poultry are not estimated for this Inventory. See Annex 5 for more
information on sources and sinks of greenhouse gas emissions not included in this Inventory.
5-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
increases. Regarding trends in other animals, populations of sheep have steadily declined, with an overall decrease
of 54 percent since 1990. Horse populations are 22 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 2018. Swine populations have trended upward through most of the time series, increasing 37 percent
from 1990 to 2018. The population of American bison more than tripled over the 1990 to 2018 time period, while
the population of mules and asses increased by a factor of 5.
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,
account for the majority of enteric fermentation Cm 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.
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 Cm
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), and a simplified approach
was used to estimate 2018 enteric emissions from cattle.
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
Agriculture 5-5

-------
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
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 CFU 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 Cm 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 Cm 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 Inventory Methodology for Cattle
As noted above, a simplified approach for cattle enteric emissions was used in lieu of the CEFM for 2018. First,
2018 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 previous Inventory
year. Dairy cow, beef cow, and bull populations for 2018 were based on data directly from the USDA-NASS
QuickStats database (USDA 2019). Because the remaining CEFM cattle sub-population categories do not
4 Due to inconsistencies in the 2003 literature values, the 2002 values were used for 2003 as well.
5-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
correspond exactly to the remaining QuickStats cattle categories, 2018 populations for these categories were
estimated by extrapolating the 2017 populations based on percent changes from 2017 to 2018 in similar
QuickStats categories, consistent with Volume 1, Chapter 5 of the 2006IPCC 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.
Table 5-5: Cattle Sub-Population Categories for 2018 Population Estimates
CEFM Cattle Category	USDA-NASS QuickStats Cattle Category
Dairy Calves
Cattle, Calves

Dairy Cows
Cattle, Cows, Milk

Dairy Replacements 7-11 months
Cattle, Heifers, GE 500 lbs,
Milk Replacement
Dairy Replacements 12-23 months
Cattle, Heifers, GE 500 lbs,
Milk Replacement
Bulls
Cattle, Bulls, GE 500 lbs

Beef Calves
Cattle, Calves

Beef Cows
Cattle, Cows, Beef

Beef Replacements 7-11 months
Cattle, Heifers, GE 500 lbs,
Beef Replacement
Beef Replacements 12-23 months
Cattle, Heifers, GE 500 lbs,
Beef Replacement
Steer Stockers
Cattle, Steers, GE 500 lbs

Heifer Stockers
Cattle, Heifers, GE 500 lbs,
(Excl. Replacement)
Steer Feedlot
Cattle, On Feed

Heifer Feedlot
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 Cm emissions from livestock in the United States from 1990 through 2018.
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 2018 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 2018 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).
See Annex 3.10 for more detailed information on the methodology and data used to calculate Cm emissions from
enteric fermentation.
Agriculture 5-7

-------
Uncertainty and Time-Series Consistency
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 2018 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
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.
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 Cm emissions in 2018 were estimated to be between 158.1 and 209.6
MMT CO2 Eq. at a 95 percent confidence level, which indicates a range of 11 percent below to 18 percent above
the 2018 emission estimate of 177.6 MMT CO2 Eq. 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.
Table 5-6: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Enteric
Fermentation (MMT CO2 Eq. and Percent)


2018 Emission


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


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



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Enteric Fermentation
ch4
177.6
158.1 209.6
-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 2018 estimates.
c The overall uncertainty calculated in 2003, and applied to the 2018 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.
5-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018. Details on the emission trends and
methodologies through time are described in more detail in the Introduction and Methodology sections.
jfkl 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.
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.
Recalculations Discussion
No recalculations were performed for the 1990 to 2017 estimates. The 2018 estimates were developed using a
simplified approach, as noted earlier in the chapter.
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. Recent improvements efforts have yielded information that the 4 percent
value is still representative of U.S. milk fat for the year 2018, but EPA continues to investigate the
availability of data across the 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;
Agriculture 5-9

-------
•	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);
•	Investigation of methodologies and emission factors for including enteric fermentation emission
estimates from poultry;
•	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; and
•	Analysis and integration of a more representative spatial distribution of animal populations by state,
particularly for poultry animal populations.
EPA received comments during the Public Review period of the 1990 through 2017 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. In addition, EPA received
comments during the Public Review period of the current Inventory (i.e., 1990 through 2018) regarding the use of
a 100-year GWP. EPA is 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 to the potential improvements listed above, EPA will review the final 2019 Refinement to the 2006IPCC
Guidelines and incorporate any changes, as applicable, to update the current Inventory estimation data and
methodologies.
5.2 Manure Management (CRF Source
Category 3B)
The treatment, storage, and transportation of livestock manure can produce anthropogenic Cm and N2O
emissions. 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 or poultry manure are 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 CO2 and little or no CFU. Ambient temperature,
moisture, and manure storage or residency time affect the amount of CH4 produced because they influence the
5-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
growth of the bacteria responsible for CFU formation. For non-liquid-based manure systems, moist conditions
(which are a function of rainfall and humidity) can promote CFU production. Manure composition, which varies by
animal diet, growth rate, and animal type (particularly the different animal digestive systems), also affects the
amount of Cm produced. In general, the greater the energy content of the feed, the greater the potential for Cm
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, N2O emissions are produced through both direct and indirect pathways. Direct N2O emissions
are produced as part of the nitrogen (N) cycle through the nitrification and denitrification of the N in livestock dung
and urine.5 There are two pathways for indirect N2O 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 NO3") 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 N2O 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 N2O emissions to occur, the manure must first be handled aerobically
where organic N is mineralized or decomposed to NH4 which is then nitrified to NO3 (producing some N2O as a
byproduct) (nitrification). Next, the manure must be handled anaerobically where the nitrate is then denitrified to
N2O 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 N2O in the waste management system (WMS). Indirect N2O 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 CFU emissions from manure management in 2018 were 61.7 MMT CO2 Eq. (2,467 kt); in 1990,
emissions were 37.1 MMT CO2 Eq. (1,485 kt). This represents a 66 percent increase in emissions from 1990.
Emissions increased on average by 1.0 MMT CO2 Eq. (2.0 percent) annually over this period. The majority of this
increase is due to swine and dairy cow manure, where emissions increased 43 and 119 percent, respectively. From
2017 to 2018, there was a 3.0 percent increase in total CFU 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
5 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.
Agriculture 5-11

-------
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 Cm emissions than dry systems. This significant shift in both
the dairy cattle and swine industries was accounted for by incorporating state and WMS-specific Cm 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 2018, total N2O emissions from manure management were estimated to be 19.4 MMT CO2 Eq. (65 kt); in 1990,
emissions were 14.0 MMT CO2 Eq. (47 kt). These values include both direct and indirect N2O emissions from
manure management. Nitrous oxide emissions have increased since 1990. Small changes in N2O emissions from
individual animal groups exhibit the same trends as the animal group populations, with the overall net effect that
N2O emissions showed a 39 percent increase from 1990 to 2018 and a 4.2 percent increase from 2017 through
2018. 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 N2O emissions.
Table 5-7 and Table 5-8 provide estimates of CH4 and N2O emissions from manure management by animal
category.6
Table 5-7: ChU and N2O Emissions from Manure Management (MMT CO2 Eq.)
Gas/Animal Type
1990
2005
2014
2015
2016
2017
2018
CH4a
37.1
51.6
54.3
57.9
59.6
59.9
61.7
Dairy Cattle
14.7
24.3
29.7
30.8
31.5
31.8
32.3
Swine
15.5
20.3
18.0
20.2
21.1
21.0
22.2
Poultry
3.3
3.2
3.3
3.4
3.4
3.4
3.5
Beef Cattle
3.1
3.3
3.0
3.1
3.3
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
17.3
17.5
18.1
18.7
19.4
Beef Cattle
5.9
7.2
7.8
7.7
8.1
8.6
9.2
Dairy Cattle
5.3
5.5
5.8
6.0
6.1
6.1
6.1
Swine
1.2
1.6
1.7
1.8
1.9
2.0
2.0
Poultry
1.4
1.6
1.6
1.6
1.6
1.6
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
71.6
75.4
77.7
78.5
81.1
Notes: Emissions from manure deposited on pasture are included in the Agricultural Soils
Management sector. Totals may not sum due to independent rounding.
6 Manure management emissions from camels are not estimated for this Inventory. See Annex 5 for more information on
sources and sinks of greenhouse gas emissions not included in this Inventory.
5-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
+ 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.
cThere are no American bison N20 emissions from managed systems; American bison are
maintained entirely on pasture, range, and paddock.
Table 5-8: ChU and N2O Emissions from Manure Management (kt)
Gas/Animal Type
1990
2005
2014
2015
2016
2017
2018
CH4a
1,485
2,062
2,172
2,316
2,385
2,395
2,467
Dairy Cattle
589
970
1,190
1,233
1,259
1,270
1,292
Swine
622
812
719
808
846
840
888
Poultry
131
129
132
136
136
137
141
Beef Cattle
126
133
120
126
132
136
135
Horses
9
12
8
8
8
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
58
59
61
63
65
Beef Cattle
20
24
26
26
27
29
31
Dairy Cattle
18
18
20
20
20
20
21
Swine
4
5
6
6
6
7
7
Poultry
5
5
5
5
5
5
6
Sheep
+
1
1
1
1
1
1
Horses
+
+
+
+
+
+
+
Goats
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
American Bisonc
NA
NA
NA
NA
NA
NA
NA
Notes: Emissions from manure deposited on pasture 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 N2O emission estimates for each animal
type. This section presents a summary of the methodologies used to estimate CH4 and N2O emissions from manure
management. See Annex 3.11 for more detailed information on the methodology and data used to calculate CH4
and N2O emissions 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);
Agriculture 5-13

-------
•	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 (Bo) of the volatile solids (by animal type); and
•	Methane conversion factors (MCF), the extent to which the Cm 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).
5-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
•	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
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 Cm emissions (kg CFU per year) from
each WMS. The amount of VS (kg per year) were multiplied by the Bo (m3 CFU per kg VS), the MCF for that WMS
(percent), and the density of Cm (kg Cm per m3 Cm). The CFU emissions for each WMS, state, and animal type
were summed to determine the total U.S. Cm emissions.
Nitrous Oxide Calculation Methods
The following inputs were used in the calculation of direct and indirect manure management N2O 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 N2O emission factor (EFwms);
•	Indirect N2O emission factor for volatilization (EFvoiatiiization);
•	Indirect N2O emission factor for runoff and leaching (EFrunoff/ieach);
•	Fraction of N loss from volatilization of NH3 and NOx (Fracgas); and
•	Fraction of N loss from runoff and leaching (Fracmnoff/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 rates 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 rates were assumed to be the same as NOF bulls.7
•	All N2O emission factors (direct and indirect) were taken from IPCC (2006). These data are appropriate
because they were developed using U.S. data.
•	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). Fracmnoff/ieaching values were based on regional cattle runoff data from EPA's Office
of Water (EPA 2002b; see Annex 3.11).
To estimate N2O 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
7 The N20 emissions from N excreted (Nex) by 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.
Agriculture 5-15

-------
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 N2O emissions were calculated by multiplying the amount of N excreted (kg per year) in each WMS by the
N2O direct emission factor for that WMS (EFwms, in kg N2O-N per kg N) and the conversion factor of N2O-N to N2O.
These emissions were summed over state, animal, and WMS to determine the total direct N2O emissions (kg of
N2O per year).
Next, indirect N2O emissions from volatilization (kg N2O per year) were calculated by multiplying the amount of N
excreted (kg per year) in each WMS by the fraction of N lost through volatilization (Fractas) divided by 100, the
emission factor for volatilization (EFvoiatiiization, in kg N2O per kg N), and the conversion factor of N2O-N to N2O.
Indirect N2O emissions from runoff and leaching (kg N2O 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 (Fracmnoff/ieach)
divided by 100, and the emission factor for runoff and leaching (EFrunoff/ieach, in kg N2O per kg N), and the conversion
factor of N2O-N to N2O. The indirect N2O emissions from volatilization and runoff and leaching were summed to
determine the total indirect N2O emissions.
Following these steps, direct and indirect N2O emissions were summed to determine total N2O emissions (kg N2O
per year) for the years 1990 to 2018.
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 Cm and N2O 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 Cm and N2O 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 2018 emission estimates.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 5-9. Manure management
Cm emissions in 2018 were estimated to be between 50.6 and 74.0 MMT CO2 Eq. at a 95 percent confidence level,
which indicates a range of 18 percent below to 20 percent above the actual 2018 emission estimate of 61.7 MMT
CO2 Eq. At the 95 percent confidence level, N2O emissions were estimated to be between 16.3 and 24.1 MMT CO2
Eq. (or approximately 16 percent below and 24 percent above the actual 2018 emission estimate of 19.4 MMT CO2
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)


2018 Emission




Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


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




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Manure Management
ch4
61.7
50.6
74.0
-18%
+20%
Manure Management
n2o
19.4
16.3
24.1
-16%
+24%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
5-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018. 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 N2O emissions from managed systems and Cm 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, Bo, 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. 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
trend towards larger farm sizes; large farms are more likely to manage manure as a liquid and therefore produce
more Cm emissions.
Agriculture 5-17

-------
Table 5-10: IPCC (2006) Implied Emission Factor Default Values Compared with Calculated
Values for ChU from Manure Management (kg/head/year)
IPCC Default
CH4 Emission	Implied CH4 Emission Factors (kg/head/year)
Animal Type
Factors 	

(ke/head/vear)a
1990
2005
2014
2015
2016
2017
2018
Dairy Cattle
48-112
30.2
54.5
64.2
65.6
66.8
67.2
67.9
Beef Cattle
1-2
1.5
1.6
1.6
1.7
1.7
1.7
1.6
Swine
10-45
11.5
13.3
11.2
11.8
12.1
11.7
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
2.5
2.6
2.6
2.6
2.6
American Bison
NA
1.8
2.0
2.0
2.1
2.1
2.1
2.1
Mules and Asses
0.76-1.14
0.9
1.0
0.9
1.0
1.0
1.0
1.0
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 N2O were compared to the U.S. Inventory implied N2O
emission factors. Default N2O emission factors from the 2006 IPCC Guidelines were used to estimate N2O 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
The manure management emission estimates include the following recalculations relative to the previous
Inventory:
•	State animal populations were updated to reflect updated USDA NASS datasets, which resulted in
population changes for:
o	Poultry in 2017,
o	Market swine in 2013-2017,
o	Breeding swine in 2017, and
o	American bison, goats, horses, and mules and asses in 2013-2015 (USDA 2019a).
•	Incorporated 2017 USDA Census of Agriculture data which affected animal populations (bison, goats,
horses, and mules and asses), farm-level distribution data which affect WMS distributions for dairy cows
and swine, and county-level temperature data which affects MCFs. These updates affected methane and
nitrous oxide emissions for 2013 through 2017 (USDA 2019d).
•	WMS distribution data for dairy cows were updated with data from the 2016 USDA Agricultural Resource
Management Survey (ARMS) of dairy producers (ERG 2019).
•	Anaerobic digestion data were updated for swine, dairy cows, and poultry using data from EPA's AgSTAR
Program (EPA 2019).
These changes impacted total emission estimates for 1990 through 2017, overall decreasing annual estimations
from less than 1 percent to 5.1 percent across the time series. The most significant changes were to the dairy cow
emissions estimates, resulting primarily from the dairy cow WMS update. Total dairy cow annual estimations
decreased throughout the entire time series, but most significantly for 2008 through 2015 during which time they
decreased by over 10 percent.
5-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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:
•	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.
•	Updating the Bo data used in the Inventory, as data become available.
EPA notes 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 methodology for population distribution to states where USDA population data are withheld
due to disclosure concerns. EPA previously discussed these changes with the National Emissions Inventory
staff to potentially improve consistency across U.S. inventories.
•	Revising the anaerobic digestion estimates to estimate Cm 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 new swine WMS data included this WMS category).
•	Comparing Cm and N2O emission estimates with estimates from other models and more recent studies
and compare the results to the Inventory, such as USDA's Dairy Gas Emissions Model.
•	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 N2O
estimates.
•	EPA acknowledges IPCC's 2019 Refinement to 2006IPCC Guidelines for National Greenhouse Gas
Inventories will provide updated emission factors that may affect emissions estimates for manure
management. EPA will work to review these updates and incorporate changes as time and resources
allow.
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 CFU
production through a process known as methanogenesis. Approximately 60 to 90 percent of the CFU produced by
methanogenic bacteria in flooded rice fields is oxidized in the soil and converted to CO2 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 CFU 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 CFU emissions in rice cultivation, and improved
water management has the largest potential to mitigate emissions (Yan et al. 2009). Upland rice fields are not
Agriculture 5-19

-------
flooded, and therefore do not produce Cm, but large amounts of CFUcan 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 N2O 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 CFU transport to the atmosphere through the plant compared to shallow-flooded systems (Sass
2001).
Other management practices also influence CFU 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
amount of root exudates8 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 CFU emissions (Wassmann et al. 2000b).
Fertilization practices also influences CFU emissions, particularly the use of fertilizers with sulfate (Wassmann et al.
2000b; Linquist et al. 2012), which can reduce CFU 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 CFU 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 sometimes 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 2018 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 5-2). Most emissions occur in Arkansas, California, Louisiana Mississippi,
Missouri and Texas. In 2018, CFU emissions from rice cultivation were 13.3 MMT CO2 Eq. (533 kt). Annual emissions
fluctuate between 1990 and 2018, 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 2018 are 17 percent lower
than emissions in 1990.
Table 5-11: ChU Emissions from Rice Cultivation (MMT CO2 Eq.)
State
1990
2005
2014
2015
2016
2017
2018
Arkansas
5.4
7.9
5.7
6.4
NE
NE
NE
California
3.3
3.4
3.9
4.1
NE
NE
NE
Florida
+
+
+
+
NE
NE
NE
Illinois
+
+
+
+
NE
NE
NE
Kentucky
+
+
+
+
NE
NE
NE
Louisiana
2.6
2.8
3.2
2.6
NE
NE
NE
Minnesota
+
0.1
+
+
NE
NE
NE
8 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.
5-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Mississippi	1.1	1.4
Missouri	0.6	1.1
New York	+	+
South Carolina	+	+
Tennessee	+	+
Texas	3^	13
Total	16.0	18.0
0.8	1.0	NE	NE	NE
0.8	0.7	NE	NE	NE
+	+	NE	NE	NE
+	+	NE	NE	NE
+	+	NE	NE	NE
0.9	1A	NE	NE	NŁ
15.4	16.2	13.5	12.8	13.3
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
NE (Not Estimated). State-level emissions are not estimated for 2016 through 2018 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
2014
2015
2016
2017
2018
Arkansas
216
315
229
256
NE
NE
NE
California
131
135
155
166
NE
NE
NE
Florida
+
1
+
+
NE
NE
NE
Illinois
+
+
+
+
NE
NE
NE
Kentucky
+
+
+
+
NE
NE
NE
Louisiana
103
113
130
103
NE
NE
NE
Minnesota
1
2
+
+
NE
NE
NE
Mississippi
45
55
31
40
NE
NE
NE
Missouri
22
45
34
26
NE
NE
NE
New York
+
+
+
+
NE
NE
NE
South Carolina
+
+
+
+
NE
NE
NE
Tennessee
+
+
+
+
NE
NE
NE
Texas
122
54
37
57
NE
NE
NE
Total
640
720
616
648
539
510
533
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 2018 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.
Agriculture 5-21

-------
Figure 5-2: Annual CHU Emissions from Rice Cultivation, 2015 (MT CO2 Eq./Year)
MT C02 Eq. ha 1 yr
~ < 5
J 5 to 10
| 10 to 15
¦ 15 to 20
¦ >20
Note: Only national-scale emissions are estimated for 2016 through 2018 in this Inventory using the surrogate data method
described in the Methodology section; therefore, the fine-scale emission patterns in this map are based on the estimates for
2015.
Methodology
The methodoiogy used to estimate Cm emissions from rice cultivation is based on a combination of IPCC Tier 1 and
3 approaches. The Tier 3 method utilizes a process-based model (DayCent) 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 cm, as well as cm transport through the plant
and via ebullition (Cheng et al. 2013). The method simulates 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 CHU emissions, DayCent simulates soil C stock changes and N2O 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
5-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).9
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 CFU 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 Cm emission 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). 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
2014
2015
2016
2017
2018
Arkansas
600
784
700
679
NE
NE
NE
California
249
236
257
280
NE
NE
NE
Florida
0
4
0
0
NE
NE
NE
Illinois
0
0
0
0
NE
NE
NE
Kentucky
0
0
0
0
NE
NE
NE
Louisiana
381
402
375
368
NE
NE
NE
Minnesota
4
9
1
1
NE
NE
NE
Mississippi
123
138
92
98
NE
NE
NE
Missouri
48
94
93
62
NE
NE
NE
New York
1
0
0
0
NE
NE
NE
South Carolina
0
0
0
0
NE
NE
NE
Tennessee
0
1
0
0
NE
NE
NE
Texas
302
118
112
131
NE
NE
NE
Total
1,707
1,788
1,631
1,619
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 2018 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. 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. Ratooned crop area as a percent of primary crop
area is presented in Table 5-14.
9 See .
Agriculture 5-23

-------
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).
dTexas: 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 Cm 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
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 2018 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 the 1990 through 2015 emissions
data that were derived using the Tier 1 and 3 methods (Brockwell and Davis 2016). Surrogate data for this model
are based on rice commodity statistics from USDA-NASS.10 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
10 See .
5-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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., CH emissions), xp contains specific surrogate data depending on the
response variable, and e 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 2018.
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 CH
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 2018,
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 Cm emissions estimate using
simple error propagation (IPCC 2006). Additional details on the uncertainty methods are provided in Annex 3.12.
Rice cultivation CFU emissions in 2018 were estimated to be between 9.2 and 21.6 MMT CO2 Eq. at a 95 percent
confidence level, which indicates a range of 31 percent below to 62 percent above the 2018 emission estimate of
13.3 MMT CO2 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
2018 Emission
Estimate
Uncertainty Range Relative to Emission Estimate3


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




Lower
Upper
Lower Upper




Bound
Bound
Bound Bound
Rice Cultivation
Tier 3
ch4
10.8
6.9
14.8
-36% +36%
Rice Cultivation
Tier 1
ch4
2.5
1.3
3.7
-48% +48%
Rice Cultivation
Total
ch4
13.3
9.2
21.6
-31% +62%
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 2018. 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
Agriculture 5-25

-------
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. Two errors were found in the spreadsheets. First,
Cm emissions from rice cultivation were not included in the national totals due to an incorrect formula. Second,
the amount of residue returned to the field was estimated in units of C, but should be in units of dry matter. Both
errors were corrected.
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
The major improvements to the current Inventory were (1) incorporating new land use and crop histories from the
NRI survey; and (2) modeling SOC stock changes to 30 cm depth with the Tier 3 approach (previously modeled to
20 cm depth), which impacts the simulation of methanogenesis in DayCent. The surrogate data method was also
applied to re-estimate stock changes from 2016 to 2017. These changes resulted in an average increase in rice
cultivation CH4 emissions of 1.2 MMT CO2 Eq. from 1990 to 2017, which is an average of 9 percent larger compared
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 the timeline may be extended
if there are insufficient resources to fund this improvement.
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).11 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.12 Several agricultural activities increase mineral N availability in soils that lead to direct N2O
emissions at the site of a management activity (see Figure 5-3) (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
11	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 denitrification, which leaks from microbial cells into the soil
and then into the atmosphere. Nitrous oxide is also produced during nitrification, although by a less well-understood
mechanism (Nevison 2000).
12	Asymbiotic N fixation is the fixation of atmospheric N2 by bacteria living in soils that do not have a direct relationship with
plants.
5-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
and forages); and drainage of organic soils13 (i.e., Histosols) (IPCC 2006). Additionally, agricultural soil management
activities, including irrigation, drainage, tillage practices, cove crops, and fallowing of land, can influence N
mineralization from soil organic matter and levels of asymbiotic N fixation. Indirect emissions of N2O occur when N
is transported from a site and is subsequently converted to N2O; 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.14 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.
13	Drainage of organic soils in former wetlands enhances mineralization of N-rich organic matter, thereby increasing N20
emissions from these soils.
14	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 (HNO3), and NOx.
In addition, hydrological processes lead to leaching and runoff of NO3" 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 5-3: 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: 1990-2018

-------
Agricultural soils produce the majority of N2O emissions in the United States. Estimated emissions in 2018 are
338.2 MMT CO2 Eq. (1,135 kt) (see Table 5-16 and Table 5-17). Annual N2O emissions from agricultural soils are 7
percent greater in the 2018 compared to 1990, but emissions fluctuated between 1990 and 2018 due to inter-
annual variability largely associated with weather patterns, synthetic fertilizer use, and crop production. From
1990 to 2018, 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 N2O 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
2014
2015
2016
2017
2018
Direct
272.5
272.2
302.3
294.5
281.0
280.0
285.7
Cropland
185.9
184.1
207.6
200.2
191.6
191.3
196.0
Grassland
86.6
88.1
94.6
94.3
89.4
88.7
89.7
Indirect
43.4
40.8
47.0
53.6
48.8
47.4
52.5
Cropland
34.2
31.8
37.9
43.0
39.2
37.8
42.8
Grassland
9.2
9.1
9.1
10.6
9.6
9.6
9.7
Total
315.9
313.0
349.2
348.1
329.8
327.4
338.2
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
2014
2015
2016
2017
2018
Direct
915
914
1,014
988
943
939
959
Cropland
623.8
617.7
696.8
671.8
642.9
641.9
657.7
Grassland
290.7
295.8
317.5
316.4
300.0
297.5
300.9
Indirect
146
137
158
180
164
159
176
Cropland
114.8
106.6
127.1
144.2
131.5
126.9
143.5
Grassland
30.7
30.4
30.5
35.6
32.3
32.2
32.6
Total
1,060
1,050
1,172
1,168
1,107
1,099
1,135
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
2014
2015
2016
2017
2018
Cropland
185.8
184.0
207.6
200.2
191.6
191.3
196.0
Mineral Soils
182.1
180.3
204.2
196.8
188.2
187.9
192.6
Synthetic Fertilizer
63.1
64.0
70.5
64.8
60.8
60.5
61.8
Organic Amendment3
12.6
13.4
14.2
14.1
14.1
14.0
14.0
Residue Nb
39.3
39.6
42.4
39.0
37.7
37.7
38.7
Mineralization and
67.1
63.3
77.1
78.9
75.5
75.7
78.1
Asymbiotic Fixation







Drained Organic Soils
3.8
3.7
3.4
3.4
3.4
3.4
3.4
Grassland
86.7
88.2
94.6
94.3
89.4
88.7
89.7
Mineral Soils
84.2
85.8
92.2
91.8
86.9
86.2
87.2
Synthetic Fertilizer
+
+
+
+
+
+
+
PRP Manure
14.6
12.8
11.6
11.6
11.3
11.2
11.3
Managed Manurec
+
+
+
+
+
+
+
Biosolids (i.e., treated
0.2
0.5
0.6
0.6
0.6
0.6
0.6
Sewage Sludge)







Residue Nd
29.7
30.8
31.8
30.4
28.6
28.4
28.7
Mineralization and
39.5
41.7
48.2
49.2
46.3
45.9
46.5
Asymbiotic Fixation







Agriculture 5-29

-------
Drained Organic Soils	23	2.4	2.5 2.5 2.5 2.5 2.5
Total	272.5	272.2	302.3 294.5 281.0 280.0 285.7
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).
bCropland 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
2014
2015
2016
2017
2018
Cropland
34.2
31.8
37.9
43.0
39.2
37.8
42.8
Volatilization & Atm.
6.5
7.3
8.2
8.6
8.3
8.1
8.2
Deposition







Surface Leaching & Run-Off
27.7
24.4
29.7
34.4
30.9
29.7
34.6
Grassland
9.2
9.1
9.1
10.6
9.6
9.6
9.7
Volatilization & Atm.
3.6
3.6
3.6
3.5
3.4
3.4
3.4
Deposition







Surface Leaching & Run-Off
5.6
5.5
5.5
7.1
6.3
6.2
6.3
Total
43.4
40.8
47.0
53.6
48.8
47.4
52.5
Notes: Estimates after 2015 are based on a data splicing method (See Methodology section). Totals may not
sum due to independent rounding.
Figure 5-4 and Figure 5-5 show regional patterns for direct N2O emissions. Figure 5-6 and Figure 5-7 show indirect
N2O emissions from volatilization, and Figure 5-8 and Figure 5-9 show the indirect N2O emissions from leaching and
runoff in croplands and grasslands, respectively.
5-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 5-4: Crops, 2015 Annual Direct N2O Emissions Estimated Using the Tier 3 DayCent
Model (MT CO2 Eq./ha/year)
Note: Only national-scale emissions are estimated for 2016 to 2018 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Direct N2O 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 5-4). 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 highest from states in the Great Plains and western United States (see Figure
5-5) where a high proportion of the land is dominated by grasslands and used for cattle and sheep grazing.
However, there are relatively large emissions from local areas in the Southeast, particularly Kentucky, Florida 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 due to greater stocking rates of livestock per unit
of area, compared to other regions of the United States.
Agriculture 5-31

-------
Figure 5-5: Grasslands, 2015 Annual Direct N2O Emissions Estimated Using the Tier 3
DayCent Model (MT CO2 Eq./ha/year)
Note: Only national-scale emissions are estimated for 2016 to 2018 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Indirect N2O emissions from volatilization in croplands have a similar pattern as the direct N2O emissions with
higher emissions in the Midwestern Corn Belt, Lower Mississippi River Basin and Great Plains. Indirect N2O
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 N2O 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
additions, small areas of high emissions occur in portions of the Great Plains that have relatively large areas of
irrigated croplands that can have relatively high leaching rates of applied/mineralized N. Indirect N2O 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: 1990-2018

-------
Figure 5-6: Crops, 2015 Annual Indirect N2O Emissions from Volatilization Using the Tier 3
DayCent Model (MT CO2 Eq./ha/year)
Note: Only national-scale emissions are estimated for 2016 to 2018 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Figure 5-7: Grasslands, 2015 Annual Indirect N2O Emissions from Volatilization Using the
Tier 3 DayCent Model (MT CO2 Eq./ha/year)
Note: Only national-scale emissions are estimated for 2016 to 2018 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 5-8: Crops, 2015 Annual Indirect N2O Emissions from Leaching and Runoff Using the
Tier 3 DayCent Model (MT CO2 Eq./ha/year)
Note: Only riational-scale emissions are estimated for 2016 to 2018 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2015.
Figure 5-9: Grasslands, 2015 Annual Indirect N2O Emissions from Leaching and Runoff
Using the Tier 3 DayCent Model (MT CO2 Eq./ha/year)
Note: Only national-scale emissions are estimated for 2016 to 2018 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: 1990-2018

-------
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) applications, 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 N2O.
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 N2O emissions from asymbiotic fixation15 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 N2O Emissions
The methodology used to estimate direct N2O 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 biosolids
(i.e., treated sewage sludge) amendments (Del Grosso et al. 2010). The Tier 3 approach has been specifically
designed and tested to estimate N2O emissions in the United States, accounting for more of the environmental and
management influences on soil N2O emissions than the IPCC Tier 1 method (see Box 5-3 for further elaboration).
Moreover, the Tier 3 approach addresses direct N2O 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 N2O) 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,16 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 an average of 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 N2O 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).
15	N inputs from asymbiotic N fixation are not directly addressed in 2006 IPCC Guidelines, but are a component of the total
emissions from managed lands and are included in the Tier 3 approach developed for this source.
16	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 N2O 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 N2O 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 N2O
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 N2O emissions only during that year and cannot be stored in soils and contribute to N2O emissions in
subsequent years. This is a simplifying assumption that is likely to create bias in estimated N2O emissions for a
specific year. In contrast, the process-based model used 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 N2O during
subsequent years.
DayCent is used to estimate N2O 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 N2O emissions from other crops or rotations with other crops,17 such as
sugarcane, some vegetables, tobacco, 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 N2O
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 N2O emissions from 2016 to 2018 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,18 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
17	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.
18	See .
5-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
2018 without surrogate data. See Box 5-4 for more information about the splicing method. Emission estimates for
2016 to 2018 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
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 2018.
A critical issue when applying splicing methods is to account for the additional uncertainty introduced by
predicting emissions with related information 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 2018. 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
N2O 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 MYD13Q119 (Potter et al. 1993, 2007). The model simulates soil temperature, and water dynamics,
using daily weather data using a 4-kilometer gridded product developed by the PRISM Climate Group (2018), and
19 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 N2O 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
N2O emissions, but these are not model inputs.
5-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
The N2O 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-CC>2 greenhouse gas
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 N2O emissions from mineral soils are described in
Annex 3.12.
For the Tier 3 method, soil N2O emissions from 2016 to 2018 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 N2O production (nitrification and denitrification). It is not
possible to partition N2O emissions into each anthropogenic activity directly from model outputs due to the
complexity of the interactions (e.g., N2O 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 N2O 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 N2O, 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 N2O emissions with individual sources of N.
Tier 1 Approach for Mineral Cropland Soils
The IPCC (2006) Tier 1 methodology is used to estimate direct N2O 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 N2O emissions from N applications are based on mineral
Agriculture 5-39

-------
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
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.20
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 2017).21 The fertilizer sales for 2015 will be updated when data are
released. After subtracting the portion of fertilizer applied to crops and grasslands simulated by DayCent
(see Tier 3 Approach for Mineral Cropland Soils and Direct N2O 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
3.10). 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 N2O 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 to units of N using average organic fertilizer N content (TVA 1991 through 1994; AAPFCO 1995
through 2017). Commercial fertilizers do include some manure and biosolids (i.e., treated sewage sludge),
but the amounts are removed from the commercial fertilizer data to avoid double counting 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 N2O emissions using the Tier 1 method. Further elaboration
on the methodology and data used to estimate N2O emissions from mineral soils are described in Annex 3.12.
Soil N2O emissions from 2016 to 2018 for Tier 1 mineral soil emissions are estimated using a splicing method that is
described in Box 5-4. 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.
20	Commercial organic fertilizers include dried blood, tankage, compost, and other, but the dried manure and biosolids (i.e.,
treated sewage sludge) is removed from the dataset in order to avoid double counting with other datasets that are used for
manure N and biosolids.
21	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).
5-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Tier 1 Approach for Drainage of Organic Soils in Croplands and Grasslands
The IPCC (2006) Tier 1 method is used to estimate direct N2O 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. Further elaboration on the methodology and data used to estimate N2O emissions from organic soils are
described in Annex 3.12.
Tier 1 and 3 Approaches for Direct N2O Emissions from Grassland Soils
As with N2O 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 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, fertilization, or inter-seeding legumes. DayCent is
used to simulate N2O 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 based on the methods described in Manure Management
section (Section 5.2) and associated annex (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 input rates are used
in the DayCent simulations. DayCent simulations of non-federal grasslands accounted for approximately 77
percent of total PRP manure N in aggregate across the country.22 The remainder of the PRP manure N in each state
is assumed to be excreted on federal grasslands, and the N2O emissions are estimated using the IPCC (2006) Tier 1
method.
Biosolids (i.e., treated sewage sludge) are assumed to be applied on grasslands because of the heavy metal content
and other pollutants in human waste that limit its use as an amendment to croplands. Biosolids application 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 soil amendments are only available at the national scale, and it
22 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

-------
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 amendments on N2O emissions from grassland
soils, and consequently, emissions from biosolids are estimated using the IPCC (2006) Tier 1 method.
Soil N2O 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. N2O 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 N2O emissions
from mineral soils are described in Annex 3.12.
Soil N2O 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 2018 are
estimated using a splicing method as described in Box 5-4. As with croplands, estimates for 2016 to 2018 will be
recalculated in a future Inventory when new NRI data are released by USDA. Biosolids application data are
compiled through 2018 in this Inventory, and therefore soil N2O emissions and confidence intervals are estimated
using the Tier 1 method for all years in the time series without application of the splicing method.
Total Direct N2O 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 N2O emissions from agricultural soil management
(see Table 5-16 and Table 5-17).
Indirect N2O Emissions Associated with Nitrogen Management in Cropland and
Grasslands
Indirect N2O 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 N2O. 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 N2O. The second pathway occurs via leaching and runoff of soil N (primarily in
the form of NO3") 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 NO3" is subject to denitrification in water bodies, which leads to N2O
emissions. Regardless of the eventual location of the indirect N2O 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 N2O. 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 N2O emissions in the Tier 3 Approach for Mineral Cropland
Soils and Direct N2O 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: 1990-2018

-------
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 volatilization data generated from both DayCent and
Tier 1 methods to estimate indirect N2O emissions occurring due to re-deposition of the volatilized N (see Table
5-19). Further elaboration on the methodology and data used to estimate indirect N2O 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 N2O. 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, biosolids amendments 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 N2O 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 plus
irrigation does not exceed 80 percent of PET.
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 N2O emissions that occur in groundwater and waterways (see Table 5-19). Further
elaboration on the methodology and data used to estimate indirect N2O emissions are described in Annex 3.12.
Indirect soil N2O emissions from 2016 to 2018 are estimated using the splicing method that is described in Box 5-4.
As with the direct N2O 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 N2O 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 calculated with the IPCC (2006) Tier 1 method;
(4) the components of indirect emissions (N volatilized and leached or runoff) calculated with the IPCC (2006) Tier
1 method; and (5) indirect emissions estimated with the IPCC (2006) Tier 1 method. Uncertainty in direct
emissions, which account for the majority of N2O emissions from agricultural management, as well as the
components of indirect emissions calculated by DayCent are estimated with a Monte Carlo Analysis, addressing
uncertainties in model inputs and structure (i.e., algorithms and parameterization) (Del Grosso et al. 2010). For
2016 to 2018, 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 estimate confidence intervals for direct emissions
calculated 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 N2O emissions. Uncertainty in the splicing method is also included
in the error propagation for 2016 to 2018 (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 N2O emissions. The estimated emissions ranges from 31
percent below to 31 percent above the 2018 emission estimate of 285.7 MMT CO2 Eq. The combined uncertainty
for indirect soil N2O emissions ranges from 69 percent below to 151 percent above the 2018 estimate of 52.5 MMT
CO2 Eq.
Agriculture 5-43

-------
Table 5-20: Quantitative Uncertainty Estimates of N2O Emissions from Agricultural Soil
Management in 2018 (MMT CO2 Eq. and Percent)


2018 Emission



Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate


(MMT C02 Eq.)
(MMT CO
2 Eq.)
(%)



Lower
Upper
Lower Upper



Bound
Bound
Bound Bound
Direct Soil N20 Emissions
N20
285.7
197.5
373.8
-31% 31%
Indirect Soil N20 Emissions
n2o
52.5
16.1
132.0
-69% 151%
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 N2O emissions from managed croplands and
grasslands in Hawaii and Alaska. The Inventory currently includes the N2O 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 2018. 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 N2O emissions and NO3" 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 N2O emission and nitrate leaching for these sites reasonably well.
See Annex 3.12 for more detailed information about the comparisons.
The original statistical model developed from the comparisons to experimental data did not separate freeze-thaw
affected areas from areas that are not affected by freeze-thaw cycles. Freeze-thaw cycles lead to hot moments or
pulses in emissions that substantially increase annual emissions (Wagner-Riddle et al. 2017). The empirical model
estimated that emissions were too high at NRI sites with freeze-thaw effects because most of the experimental
sites are not influenced by freeze-thaw events, and this led to a reduction in emissions from freeze-thaw events.
Therefore, corrective actions were taken to include a freeze-thaw indicator variable in the statistical model to
address differences in the DayCent model prediction capability for experimental sites with and without freeze-
thaw events.
In addition, quality control uncovered an error in the DayCent simulations associated with no grazing on pastures
and rangelands during the recent historical period from 1980 to 2015. In the initial simulations, this led to a large
increase in N additions to soils from crop and grass residues. Corrective actions were taken to ensure grazing was
simulated on pastures and rangelands by the DayCent Model.
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
5-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
Several major improvements have been implemented in this Inventory leading to the need for recalculations,
including (1) development of a more detailed time series of management activity data by combining information in
an imputation analysis from USDA-NRCS CEAP survey, USDA-ERS ARMS data, CTIC data and USDA Census of
Agriculture data; (2) incorporating new land use and crop histories from the NRI survey; (3) incorporating new land
use data from the NLCD; (4) modeling SOC stock changes to 30 cm depth with the Tier 3 approach (previously
modeled to 20 cm depth), which influences the mineralization of N from soil organic matter decomposition; (5)
modeling the N cycle with freeze-thaw effects on soil N2O emissions; and (6) addressing the effect of cover crops
on greenhouse gas emissions and removals. Other improvements include better resolving the timing of tillage,
planting, fertilization and harvesting based on the USDA-NRCS CEAP survey and state level information on planting
and harvest dates; improving the timing of irrigation; and crop senescence using growing degree relationships. The
surrogate data method was also applied to re-estimate N2O emissions from 2016 to 2017. These changes resulted
in an average increase in emissions of 22 percent from 1990 to 2017 relative to 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.
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 N2O emissions. Experimental study sites will continue to be added for quantifying model structural
uncertainty. Studies that have continuous (daily) measurements of N2O (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 N2O 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 N2O 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
N2O 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.
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.
Agriculture 5-45

-------
5.5 Liming (CRF Source Category 3G)
Crushed limestone (CaCOs) and dolomite (CaMgfCOsh) 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 used in other
industrial process sectors (e.g., cement production, glass production, etc.) are accounted for within the IPPU
chapter. Emissions from liming of soils have fluctuated over the past 25 years in the United States, ranging from
3.1 MMT CO2 Eq. to 6.0 MMT CO2 Eq. In 2018, liming of soils in the United States resulted in emissions of 3.1 MMT
CO2 Eq. (0.9 MMT C), representing a 33 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
2014
2015
2016
2017
2018
Limestone
4.1
3.9
3.3
3.5
2.8
2.9
3.0
Dolomite
0.6
0.4
0.3
0.3
0.3
0.2
0.2
Total
4.7
4.3
3.6
3.7
3.1
3.1
3.1
Note: Totals may not sum due to independent rounding.
Table 5-22: Emissions from Liming (MMT C)
Source
1990
2005
2014
2015
2016
2017
2018
Limestone
1.1
1.1
0.9
0.9
0.8
0.8
0.8
Dolomite
0.2
0.1 /
0.1
0.1
0.1
0.1
0.1
Total
1.3
1.2
1.0
1.0
0.8
0.8
0.9
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 applied (see Table 5-23)
were multiplied by CO2 emission factors from West and McBride (2005). These 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).
The annual application rates of limestone and dolomite were derived from estimates and industry statistics
provided in the Minerals Yearbook and Mineral Industry Surveys (Tepordei 1993 through 2006; Willett 2007a,
2007b, 2009, 2010, 2011a, 2011b, 2013a, 2014, 2015, 2016, 2017, 2018; USGS 2008 through 2018). 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).
5-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CaCC>3 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 2018 U.S. emission estimate from liming of soils is 3.1 MMT CO2
Eq. using the U.S.-specific factors. In contrast, emissions would be estimated at 6.4 MMT CO2 Eq. using the IPCC
(2006) default emission factors.
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 2018 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 2018 data, 2017 fractions were applied
to a 2018 estimate of total crushed stone presented in the USGS Mineral Industry Surveys: Crushed Stone and Sand
and Gravel in the First Quarter of 2019 (USGS 2019).
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 ¦ 2014
2015
2016
2017
2018
Limestone
Dolomite
19.0 18.1 15.3
2.4 1.9 1.3
16.0
1.2
13.0
1.1
13.4
0.8
13.7
0.8
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
Agriculture 5-47

-------
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 CO2 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 2018 were estimated to be between -0.34 and 5.94
MMT CO2 Eq. at the 95 percent confidence level. This confidence interval represents a range of 111 percent below
to 88 percent above the 2018 emission estimate of 3.1 MMT CO2 Eq. Note that there is a small probability of a
negative emissions value leading to a net uptake of CO2 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).
Table 5-24: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Liming
(MMT CO2 Eq. and Percent)
Source
_ 2018 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 3.1
(0.34) 5.94
-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 2018. 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.
:alcu!ations Discussion
Adjustments were made in the current Inventory to improve the results. First, limestone and dolomite application
data for 2016 and 2017 were updated with the recently published data from USGS (2019), rather than
approximated by a ratio method for 2017. With this revision in the activity data, the emissions decreased by 3.9
and 3.2 percent for 2016 and 2017, respectively, relative to the previous Inventory estimates.
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 industrial production process. In the presence of water and urease enzymes, urea is converted
into ammonium (NhV), hydroxyl ion (OH), and bicarbonate (HCO3 ). The bicarbonate then evolves into CO2 and
water. Emissions from urea fertilization in the United States totaled 4.6 MMT CO2 Eq. (1.3 MMT C) in 2018 (Table
5-25 and Table 5-26). Carbon dioxide emissions have increased by 129 percent between 1990 and 2018 due to an
increasing amount of urea that is applied to soils. The variation in emissions across the time series is driven by
increasing amounts of fertilizer applied to soils. Carbon dioxide emissions associated with urea consumed for non-
agricultural purposes are accounted for in the IPPU chapter.
5-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 5-25: CO2 Emissions from Urea Fertilization (MMT CO2 Eq.)
Source	1990 2005	2014 2015 2016 2017 2018
Urea Fertilization	2.0	3.1	3.9 4.1 4.0 4.5	4.6
Table 5-26: CO2 Emissions from Urea Fertilization (MMT C)
Source	1990 i 2005	2014 2015 2016 2017 2018
Urea Fertilization	0.5	0.9	1.1 1.1 1.1 1.2	1.3
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 all CO2 fixed during the industrial process for urea production is
released after application. 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).23 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.
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, 2017, and 2018 fertilizer years (i.e., July 2015 through June 2016, July 2016
through June 2017 and July 2017 through June 2018) were not available for this Inventory. Therefore, urea
application in the 2016, 2017, and 2018 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 CO2 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 data using the method described above.
Table 5-27: Applied Urea (MMT)

1990
2005
2014
2015
2016
2017
2018
Urea Fertilizer3
3.3
4.8
6.2
6.4
6.7
6.9
7.1
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-Serii insistency
An Approach 2 Monte Carlo analysis was conducted as described by the IPCC (2006). The largest source of
uncertainty was the default emission factor, which assumes that 100 percent of the C in CO(NH2)2 applied to soils is
ultimately emitted into the environment as CO2. This factor does not incorporate the possibility that some of the C
23 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

-------
may be retained in the soil, and therefore the uncertainty range was set from 50 percent emissions to the
maximum emission value of 100 percent using a triangular distribution. In addition, urea consumption data also
have uncertainty that are represented as normal density distributions. Due to the highly skewed distribution of the
emissions from the Monte Carlo analysis, the estimated emissions are based on the mode of the posterior
distribution 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 2018 were estimated to be between 2.97 and 5.35
MMT CO2 Eq. at the 95 percent confidence level. This indicates a range of 35 percent below to 16 percent above
the 2018 emission estimate of 4.6 MMT CO2 Eq. (Table 5-28).
Table 5-28: Quantitative Uncertainty Estimates for CO2 Emissions from Urea Fertilization
(MMT CO2 Eq. and Percent)


Uncertainty Range Relative to Emission
Source Gas
2018 Emission Estimate
Estimate3

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


Lower Upper
Lower Upper


Bound Bound
Bound Bound
Urea Fertilization C02
4.6
2.97 5.35
-35% +16%
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. Urea for non-fertilizer use, such as
aircraft deicing, may be included in consumption totals, but the amount is likely very small. For example, research
on aircraft deicing practices based on a 1992 survey found a known annual usage of approximately 2,000 tons of
urea for deicing; this would constitute 0.06 percent of the 1992 consumption of urea (EPA 2000). Similarly, surveys
conducted from 2002 to 2005 indicate that total urea use for deicing at U.S. airports is estimated to be 3,740
metric tons per year, or less than 0.07 percent of the fertilizer total for 2007 (Itle 2009). In addition, there is
uncertainty surrounding the underlying assumptions behind the calculation that converts 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 corresponding increase or decrease in the value for the subsequent year.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018. 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 Urea Fertilization has been developed and implemented, consistent with the U.S.
Inventory QA/QC plan. No errors were found in the calculation. Based on the quality control review, it was not
clear if Urea Ammonium Nitrate (UAN) should also be included as a source of CO2 emissions. This will be further
investigated in a future Inventory.
Recalculations Discussion
Emissions estimates were derived directly from the Monte Carlo analysis in this Inventory. The mode was selected
due to the highly skewed distribution of emissions from the Monte Carlo analysis. The entire time series was
recalculated to use the mode of the distribution. This improvement in the calculation of emissions led to estimates
that averaged about 13 percent lower than the previous Inventory across the time series.
Planned Improvements
A key planned improvement is to investigate the composition of Urea Ammonium Nitrate (UAN), and determine if
UAN should be included in the estimation of Urea CO2 emissions.
5-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 emissions because the C released to the atmosphere as CO2 during burning is reabsorbed during the next
growing season by the crop. However, crop residue burning is a net source of CFU, N2O, 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, and wheat (McCarty 2009). In 2018, CH4 and N2O emissions from field
burning of agricultural residues were 0.4 MMT CO2 Eq. (16 kt) and 0.2 MMT CO2 Eq. (0.6 kt), respectively (Table
5-29 and Table 5-30). Annual emissions of CFU and N2O have increased from 1990 to 2018 by 15.7 percent and 15.2
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
2014
2015
2016
2017
2018
ch4
0.3
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
+
+
0.1
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
n2o
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Maize
+
+
0.1
+
+
+
+
Rice
+
+
+
+
+
+
+
Wheat
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
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 5-30: ChU, N2O, CO, and NOx Emissions from Field Burning of Agricultural Residues
(kt)
Gas/Crop Type
1990
2005
2014
2015
2016
2017
2018
ch4
14
16
16
16
16
16
16
Maize
2
3
5
5
5
5
5
Rice
3
3
3
2
2
2
2
Wheat
5
5
4
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
1
2
2
2
2
2
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
n2o
1
1
1
1
1
1
1
Maize
+
+
+
+
+
+
+
5-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
287
332
338
311
310
308
308
NOx
12
14
14
13
13
13
13
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt
Methodology
A U.S.-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 U.S.-specific approach to the IPCC (2006) default
approach, see Box 5-6) and a data splicing method with a linear extrapolation was applied to complete the
emissions time series from 2015 to 2018. In order to estimate the amounts of C and N released during burning, the
following equation is used:
C or N released = Ł for all crop types and states
AB
CAH x CP x RCR x DMF x BE x CE x (FC or FN)
where,
Area Burned (AB)
Crop Area Harvested (CAH)
Crop Production (CP)
Residue: Crop Ratio (RCR)
Dry Matter Fraction (DMF)
Fraction of C or N (FC or FN)
Burning Efficiency (BE)
Combustion Efficiency (CE)
= Total area of crop burned, by state
= Total area of crop harvested, by state
= Annual production of crop in kt, by state
= Amount of residue produced per unit of crop production
= Amount of dry matter per unit of biomass for a crop
= Amount of C or N per unit of dry matter for a crop
= The proportion of prefire fuel biomass consumed24
= The proportion of C or N released with respect to the total amount of C or N
available in the burned material, respectively
24 In IPCC/UNEP/OECD/IEA (1997), the equation for C or N released contains the variable 'fraction oxidized in burning.' This
variable is equivalent to (burning efficiency x combustion efficiency).
Agriculture 5-53

-------
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.25 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 CFU, CO, N2O, and NOx emissions from the Field Burning of Agricultural Residues:
CH4 and CO, or N2O and N0X = C or N Released x ER x CF
where,
Emissions Ratio (ER) = g CH4-C or CO-C/g C released, or g N2O-N or NOx-N/g N released
Conversion Factor (CF) = conversion, by molecular weight ratio, of CH4-C to C (16/12), or CO-C to C
(28/12), or N2O-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 and factors used in: (1) the current Inventory and (2) the default IPCC (2006)
approach was undertaken for the time series from 1990 through 2014 to determine the difference in overall
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:
Emissions (kt) = AB x (Mbx Cf) x Gef x 10"6
where,
Area Burned (AB) = Total area of crop burned (ha)
Mass Burned (Mb x Cf) = IPCC (2006) default carbon fractions with fuel biomass consumption US-
Specific Values using NASS Statistics26 (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 that utilizes default combustion factors and emission factors with mass
of fuel values derived from national datasets resulted in 27 percent lower emissions of CH4 and 49 percent
lower emissions of N2O compared to this Inventory. In summary, the IPCC/UNEP/OECD/IEA (1997) method is
25	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.
26	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.
5-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
considered more appropriate for U.S. conditions because it is more flexible for incorporating country-specific
data and emissions are estimated based on specific C and N content of the fuel, which is converted into CH , CO,
N O and NO., 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 analyzed for 2015 to 2018 so a
data splicing method is used to estimate emissions for that 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 have 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.
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 developed by 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 often, and avoid models that appear to
Agriculture 5-55

-------
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
0
0
0
0
Rice
8%
8%
4%
6%
Wheat
1%
2%
2%
1%
Barley
1%
0
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
0
0
0
0
Legume Hay
0
0
0
0
Peas
0
0
0
0
Sunflower
0
0
0
0
Tobacco
2%
2%
3%
3%
Vegetables
0
0
0
0
Chickpeas
0
1%
0
0
Dry Beans
1%
1%
0
0
Lentils
0
0
0
0
Peanuts
3%
3%
3%
3%
Soybeans
0
0
1%
1%
Potatoes
0
0
0
0
Sugarbeets
0
0
0
0
Additional parameters are needed to estimate the amount of burning, including residue: crop ratios, dry matter
fractions, carbon fractions, nitrogen fractions, burning efficiency 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 burning efficiency is assumed to be 93 percent, and the combustion efficiency is
assumed to be 88 percent, for all crop types (EPA 1994). 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).
Table 5-33: Parameters for Estimating Emissions from Field Burning of Agricultural Residues
Crop
Residue/Crop
Ratio
Dry
Matter
Fraction
Carbon
Fraction
Nitrogen
Fraction
Burning
Efficiency
(Fraction)
Combustion
Efficiency
(Fraction)
Maize
0.707
0.56
0.47
0.01
0.93
0.88
Rice
1.340
0.89
0.47
0.01
0.93
0.88
Wheat
1.725
0.89
0.47
0.01
0.93
0.88
Barley
1.181
0.89
0.47
0.01
0.93
0.88
Oats
1.374
0.89
0.47
0.01
0.93
0.88
Other Small Grains
1.777
0.88
0.47
0.01
0.93
0.88
5-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Sorghum
0.780
0.60
0.47
0.01
0.93
0„
Cotton
7.443
0.93
0.47
0.01
0.93
0„
Grass Hay
0.208
0.90
0.47
0.02
0.93
0„
Legume Hay
0.290
0.67
0.47
0.01
0.93
0„
Peas
1.677
0.91
0.47
0.01
0.93
0„
Sunflower
1.765
0.88
0.47
0.01
0.93
0„
Tobacco
0.300
0.87
0.47
0.01
0.93
0„
Vegetables
0.708
0.08
0.47
0.01
0.93
0„
Chickpeas
1.588
0.91
0.47
0.01
0.93
0„
Dry Beans
0.771
0.90
0.47
0.01
0.93
0„
Lentils
1.837
0.91
0.47
0.02
0.93
0„
Peanuts
1.600
0.94
0.47
0.02
0.93
0„
Soybeans
1.500
0.91
0.47
0.01
0.93
0„
Potatoes
0.379
0.25
0.47
0.02
0.93
0„
Sugarbeets
0.196
0.22
0.47
0.02
0.93
0„
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 et al. 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 valuesare 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 etal. 1992; Gibberd etal. 2003; Reid and English 2000; Peach etal. 2000; see IPCC Tubers for
R:S and N fraction.
Lettuce, cabbage: combines 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.
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).
Agriculture 5-57

-------
For this Inventory, new activity data on the burned areas have not been analyzed for 2015 to 2018. To complete
the emissions time series, a linear extrapolation of the trend is applied to estimate the emissions in the last four
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 Cm, N2O, CO and NOx for the last 4 years in the time series from 2015 to 2018 (Brockwell and Davis 2016). The
Tier 2 method described previously will be applied to recalculate the emissions for the last 4 years in the time
series (2015 to 2018) in a future Inventory.
Uncertainty and Time-Series Consistency
Emissions are estimated using a linear regression model with autoregressive moving-average (ARMA) errors for
2018. 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 2018 are between 0.33 and 0.46 MMT CO2 Eq. at a 95 percent confidence level. This
indicates a range of 16 percent below and 16 percent above the 2018 emission estimate of 0.4 MMT CO2 Eq.
Nitrous oxide emissions are between 0.14 and 0.20 MMT CO2 Eq., or approximately 19 percent below and 13
percent above the 2018 emission estimate of 0.2 MMT CO2 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)


2018 Emission




Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


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

(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Field Burning of Agricultural
Residues
ch4
0.4
0.33
0.46
-16%
+16%
Field Burning of Agricultural
Residues
n2o
0.2
0.14
0.20
-19%
+13%
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 2018. 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 field burning of agricultural residues was implemented with Tier 1 analyses,
consistent with the U.S. Inventory QA/QC plan outlined in Annex 8. Errors were identified in the assignment of
yields to grass hay, legume hay and other close grown crops for calculation of residue burned, and these errors
were documented and corrected in the analysis.
:alcu!ations Discussion
Methodological recalculations are associated with two improvements, a) incorporation of new survey data from
the USDA National Resources Inventory (USDA-NRCS 2018), and b) a revision to the logistical regression predicting
burned area in states that were not directly analyzed for fire occurrence based on remote sensing products (see
Methodology section). The logistical regression incorporated revised information on the timing of state legislation
to restrict burning of residues in agricultural fields. As a result of these two improvements, the emissions increased
5-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
on average across the time series by 178 percent and 189 percent for Cm and N2O, respectively. The absolute
increases in emissions are 0.2 MMT CO2 Eq. and 0.1 MMT CO2 Eq. for Cm and N2O, respectively.
Planned Improvements
The 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 the DayCent model, and burning events will be simulated directly within the 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 these sources, and better ensure
mass balance of C and N in the Inventory analysis.
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)
stocks (i.e., aboveground biomass, belowground biomass, dead wood, litter, and C stock changes from mineral and
organic soils), harvested wood pools, and non-carbon dioxide (non-CCh) 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 the non-CC>2 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-CC>2 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 (Cm) and nitrous oxide (N2O)
emissions from managed peatlands, as well as aboveground and soil C stock changes in all coastal wetlands, CH4
emissions from vegetated coastal wetlands, and N2O emissions from aquaculture in coastal wetlands. Estimates for
Land Converted to Wetlands include aboveground 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, N2O emissions from
nitrogen fertilizer additions to soils, and CChfluxes 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 net emissions of greenhouse gases accounting for both the emissions of C02 to and the
removals of C02 from the atmosphere. Removal 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.
The land use, land-use change, and forestry (LULUCF) sector in 2018 resulted in a net increase in C stocks (i.e., net
CO2 removals) of 799.6 MMT CO2 Eq. (218.1 MMT C).2 This represents an offset of approximately 12.0 percent of
total (i.e., gross) greenhouse gas emissions in 2018. Emissions of CFU and N2O from LULUCF activities in 2018 are
26.1 MMT CO2 Eq. and represent 0.4 percent of total greenhouse gas emissions.3
Total C sequestration in the LULUCF sector decreased by approximately 7.1 percent between 1990 and 2018. This
decrease was primarily due to a decline in the rate of net C accumulation in Forest Land and Cropland Remaining
Cropland, as well as an increase in emissions from Land Converted to Settlements,4 Specifically, there was a net C
accumulation in Settlements Remaining Settlements, which increased from 1990 to 2018, 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 2018 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. Emissions from Land Converted to Grassland decreased during this period. The C stock
change from LULUCF is summarized in Table 6-1.
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 2017) to
ensure that the trend is accurate. This year's major updates include (1) Representation of the U.S. Land Base:
incorporating revised land use data from the updated 2015 National Resource Inventory, National Forest Inventory
and National Land Cover Database; (2) Forest Lands: use of new data from the National Forest Inventory and
refined estimates in the Digital General Soil Map, and new data on harvested wood products from 2003-2017; and
(3) Cropland/Grassland: incorporating the 2015 National Resources Inventory along with new data on crop
histories, and updating the DayCent soil process model to extend soil depth from 20 to 30 centimeters and capture
the effects of freeze thaw. Together, these updates for 2017 increased total sequestration of CO2 by 60.3 MMT CO2
Eq. (8 percent) and increased total non-CC>2 emissions by 10.6 MMT CO2 Eq. (68 percent). For more information on
specific methodological updates, please see the Recalculations discussion within the respective source category
section of this chapter.
Table 6-1: Net CO2 Flux from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.)
Land-Use Category
1990
2005
2014
2015
2016
2017
2018
Forest Land Remaining Forest Land
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
Changes in Forest Carbon Stocks3
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
Land Converted to Forest Land
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Changes in Forest Carbon Stocksb
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Changes in Mineral and Organic Soil







Carbon Stocks
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
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 Peatiands, 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.
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.
Note: Totals may not sum due to independent rounding.
6-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Land Converted to Cropland
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Changes in all Ecosystem Carbon







Stocks0
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Grassland Remaining Grassland
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Changes in Mineral and Organic Soil







Carbon Stocks
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Land Converted to Grassland
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Changes in all Ecosystem Carbon







Stocks0
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.4)
(4.4)
(4.4)
(4.4)
Changes in Organic Soil Carbon Stocks







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







Carbon Stocks in Coastal Wetlands
(5.1)
(6.8)
(5.1)
(5.1)
(5.1)
(5.1)
(5.1)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Changes in Aboveground and Soil







Carbon Stocksd
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(109.6)
(116.6)
(126.6)
(126.8)
(125.7)
(125.9)
(125.9)
Changes in Organic Soil Carbon Stocks
11.3
12.2
15.1
15.7
16.0
16.0
15.9
Changes in Settlement Tree Carbon







Stocks
(96.4)
(117.4)
(129.4)
(130.4)
(129.8)
(129.8)
(129.8)
Changes in Yard Trimmings and Food







Scrap Carbon Stocks in Landfills
(24.5)
(11.4)
(12.3)
(12.1)
(11.9)
(12.1)
(12.0)
Land Converted to Settlements
62.9
85.0
81.4
80.1
79.4
79.3
79.3
Changes in all Ecosystem Carbon







Stocks0
62.9
85.0
81.4
80.1
79.4
79.3
79.3
LULUCF Carbon Stock Change
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
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 Includes the net changes to carbon stocks stored in all forest ecosystem pools (excludes drained organic soils which
are included in the flux from Forest Land Remaining Forest Land because it is not possible to separate the activity data
at this time).
c 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.
d Includes aboveground and soil carbon stock changes for land converted to vegetated coastal wetlands.
Emissions of CFU from LULUCF activities are shown in Table 6-2. Forest fires were the largest source of CFU
emissions from LULUCF in 2018, totaling 11.3 MMT CO2 Eq. (452 kt of CH4). Coastal Wetlands Remaining Coastal
Wetlands resulted in CH4 emissions of 3.6 MMT CO2 Eq. (144 kt of CH4). Grassland fires resulted in CFU emissions of
0.3 MMT CO2 Eq. (12 kt of CH4). Land Converted to Wetlands, Drained Organic Soils on forest lands, and Peatlands
Remaining Peatlands resulted in CFU emissions of less than 0.05 MMT CO2 Eq. each.
For N2O emissions, forest fires were also the largest source from LULUCF in 2018, totaling 7.5 MMT CO2 Eq. (25 kt
of N2O). Nitrous oxide emissions from fertilizer application to settlement soils in 2018 totaled to 2.4 MMT CO2 Eq.
(8 kt of N2O). This represents an increase of 20.1 percent since 1990. Additionally, the application of synthetic
fertilizers to forest soils in 2018 resulted in N2O emissions of 0.5 MMT CO2 Eq. (2 kt of N2O). 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 N2O emissions of 0.3 MMT CO2 Eq. (1 kt of
N2O). Coastal Wetlands Remaining Coastal Wetlands and drained organic soils on forest lands resulted in N2O
emissions of 0.1 MMT CO2 Eq. each (less than 0.5 kt of N2O), and Peatlands Remaining Peatlands resulted in N2O
emissions of less than 0.05 MMT CO2 Eq.
Emissions and removals from LULUCF are summarized in Figure 6-1 and Table 6-3 by land-use and category, and
Table 6-4 and Table 6-5 by gas in MMT CO2 Eq. and kt, respectively.
Land Use, Land-Use Change, and Forestry 6-3

-------
Table 6-2: Emissions from Land Use, Land-Use Change, and Forestry by Gas (MMT CO2 Eq.)
Gas/Land-Use Sub-Category
1990
2005
2014
2015
2016
2017
2018
ch4
4.4
8.8
9.5
16.1
7.3
15.2
15.2
Forest Land Remaining Forest Land:







Forest Fires3
0.9
5.0
5.6
12.2
3.4
11.3
11.3
Wetlands Remaining Wetlands: Coastal







Wetlands Remaining Coastal Wetlands
3.4
3.5
3.6
3.6
3.6
3.6
3.6
Grassland Remaining Grassland:







Grassland Firesb
0.1
0.3
0.4
0.3
0.3
0.3
0.3
Land Converted to Wetlands: Land







Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:







Drained Organic Soilsc
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
n2o
3.0
7.5
7.0
11.2
5.5
10.8
10.9
Forest Land Remaining Forest Land:







Forest Fires3
0.6
3.3
3.7
8.1
2.2
7.5
7.5
Settlements Remaining Settlements:







Settlement Soilsd
2.0
3.1
2.2
2.2
2.2
2.3
2.4
Forest Land Remaining Forest Land:







Forest Soilse
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Grassland Remaining Grassland:







Grassland Firesb
0.1
0.3
0.4
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 Soilsc
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
LULUCF Emissions
7.4
16.3
16.6
27.4
12.8
26.1
26.1
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
a Estimates include emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
b Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grass/and.
c Estimates include emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
d Estimates include emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted to
Settlements.
e Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
6-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-1: 2018 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.)
Forest Land Remaining Forest Land )
Settlements Remaining Settlements
Land Converted to Forest Land
Land Converted to Grassland
Cropland Remaining Cropland
Wetlands Remaining Wetlands
Land Converted to Wetlands
Non-C02 Emissions from Peatlands Remaining Peatlands
ChU Emissions from Land Converted to Coastal Wetlands
Non-C02 Emissions from Drained Organic Soils
N2O Emissions from Forest Soils
Non-C02 Emissions from Grassland Fires
N2O Emissions from Settlement Soils
Non-C02 Emissions from Coastal Wetlands Remaining Coastal Wetlands
Grassland Remaining Grassland
Non-C02 Emissions from Forest Fires
Land Converted to Cropland
Land Converted to Settlements
(663.2)
Non-C02 Emissions
Carbon Stock Change
l< 0.51
l< 0.51
l< 0.51
l< 0.51
l< 0.51
(300) (250) (200) (150) (100) (50)
MMT CO2 Eq.
50 100
Note: Parentheses indicate net sequestration.
Table 6-3: Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and
Forestry (MMT CO2 Eq.)
Land-Use Category
1990
2005
2014
2015
2016
2017
2018
Forest Land Remaining Forest Land
(732.2)
(669.8)
(609.0)
(655.3)
(651.7)
(628.4)
(643.9)
Changes in Forest Carbon Stocks3
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
Non-C02 Emissions from Forest Firesb
1.5
8.2
9.2
20.3
5.6
18.8
18.8
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
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Changes in Forest Carbon Stockse
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Changes in Mineral and Organic Soil







Carbon Stocks
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Land Converted to Cropland
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Changes in all Ecosystem Carbon Stocks'
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Grassland Remaining Grassland
9.3
11.4
20.6
14.3
10.2
11.5
11.8
Changes in Mineral and Organic Soil







Carbon Stocks
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Non-C02 Emissions from Grassland Fires5
0.2
0.7
0.8
0.7
0.6
0.6
0.6
Land Converted to Grassland
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Changes in all Ecosystem Carbon Stocks'
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Wetlands Remaining Wetlands
(0.5)
(2.0)
(0.6)
(0.6)
(0.7)
(0.7)
(0.7)
Changes in Organic Soil Carbon Stocks in







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







Stocks in Coastal Wetlands
(5.1)
(6.8)
(5.1)
(5.1)
(5.1)
(5.1)
(5.1)
CH4 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
3.4
3.5
3.6
3.6
3.6
3.6
3.6
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 Use, Land-Use Change, and Forestry 6-5

-------
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Changes in Aboveground and Soil Carbon







Stocks
(+)
<+>
(+)
(+)
(+)
(+)
(+)
CH4 Emissions from Land Converted to







Coastal Wetlands
+
+
+
+
+
+
+
Settlements Remaining Settlements
(107.6)
(113.5)
(124.3)
(124.6)
(123.5)
(123.6)
(123.5)
Changes in Organic Soil Carbon Stocks
11.3
12.2
15.1
15.7
16.0
16.0
15.9
Changes in Settlement Tree Carbon







Stocks
(96.4)
(117.4)
(129.4)
(130.4)
(129.8)
(129.8)
(129.8)
Changes in Yard Trimming and Food







Scrap Carbon Stocks in Landfills
(24.5)
(11.4)
(12.3)
(12.1)
(11.9)
(12.1)
(12.0)
N20 Emissions from Settlement Soilsh
2.0
3-1
2.2
2.2
2.2
2.3
2.4
Land Converted to Settlements
62.9
85.0
81.4
80.1
79.4
79.3
79.3
Changes in all Ecosystem Carbon Stocks'
62.9
85.0
81.4
80.1
79.4
79.3
79.3
LULUCF Emissions'
7.4
16.3
16.6
27.4
12.8
26.1
26.1
LULUCF Carbon Stock Change1'
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
LULUCF Sector Net Totalk
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
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 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
e 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.
g Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grass/and.
h Estimates include 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.
' 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 CH4 and N20 emissions to the atmosphere plus net carbon stock changes
in units of MMT C02 eq.
Table 6-4: Emissions and Removals from Land Use, Land-Use Change, and Forestry (MMT
COz Eq.)
Gas/Land-Use Category
1990
2005
2014
2015
2016
2017
2018
Carbon Stock Change3
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
Forest Land Remaining Forest Land
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
Land Converted to Forest Land
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
Cropland Remaining Cropland
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Land Converted to Cropland
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Grassland Remaining Grassland
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Land Converted to Grassland
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.4)
(4.4)
(4.4)
(4.4)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(109.6)
(116.6)
(126.6)
(126.8)
(125.7)
(125.9)
(125.9)
Land Converted to Settlements
62.9
85.0
81.4
80.1
79.4
79.3
79.3
ch4
4.4
8.8
9.5
16.1
7.3
15.2
15.2
Forest Land Remaining Forest Land:







Forest Firesb
0.9
5.0
5.6
12.2
3.4
11.3
11.3
Wetlands Remaining Wetlands: Coastal
3.4
3.5
3.6
3.6
3.6
3.6
3.6
6-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Wetlands Remaining Coastal Wetlands
Grassland Remaining Grassland:







Grassland Firesc
0.1
0.3
0.4
0.3
0.3
0.3
0.3
Land Converted to Wetlands: Land







Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:







Drained Organic Soilsd
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:







Peatlands Remaining Peatlands
+
+
+
+
+
+
+
n2o
3.0
7.5
7.0
11.2
5.5
10.8
10.9
Forest Land Remaining Forest Land:







Forest Firesb
0.6
3.3
3.7
8.1
2.2
7.5
7.5
Settlements Remaining Settlements:







Settlement Soilse
2.0
3.1
2.2
2.2
2.2
2.3
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
0.4
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 Emissions8
7.4
16.3
16.6
27.4
12.8
26.1
26.1
LULUCF Carbon Stock Change3
(860.7)
(831.0)
(739.6)
(802.9)
(801.7)
(790.0)
(799.6)
LULUCF Sector Net Totalh
(853.4)
(814.7)
(723.0)
(775.5)
(788.9)
(763.9)
(773.5)
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 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
c Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grassland.
d Estimates include emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
e Estimates include emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted to
Settlements.
f Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
s 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 CH4 and N20 emissions to the atmosphere plus net carbon stock changes
in units of MMT C02 Eq.
Table 6-5: Emissions and Removals from Land Use, Land-Use Change, and Forestry (kt)
Gas/Land-Use Category
1990
2005
2014
2015
2016
2017
2018
Carbon Stock Change (C02)a
(860,747)
(830,952)
(739,565)
(802,929)
(801,734)
(790,019)
(799,622)
Forest Land Remaining Forest







Land
(733,893)
(678,611)
(618,785)
(676,144)
(657,899)
(647,721)
(663,247)
Land Converted to Forest Land
(109,423)
(110,220)
(110,475)
(110,557)
(110,572)
(110,576)
(110,579)
Cropland Remaining Cropland
(23,176)
(29,002)
(12,247)
(12,826)
(22,730)
(22,292)
(16,602)
Land Converted to Cropland
54,092
53,816
56,652
57,197
55,454
55,629
55,333
Grassland Remaining







Grassland
9,132
10,705
19,738
13,610
9,590
10,911
11,230
Land Converted to Grassland
(6,686)
(40,309)
(24,878)
(23,164)
(24,761)
(24,908)
(24,613)
Land Use, Land-Use Change, and Forestry 6-7

-------
Wetlands Remaining Wetlands
Land Converted to Wetlands
Settlements Remaining
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
Soilse
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
(4,049)
(44)
(109,567)
62,867
176
35
137
+
10
(5,689)
(32)
(116,642)
85,032
352
198
140
13
+
25
11
10
2
1
(4,328) (4,358) (4,389) (4,398) (4,445)
(44)	(44)	(44)	(44)	(44)
(126,550) (126,789) (125,734) (125,929) (125,926)
81,351
382
222
143
16
1
1
+
24
12
80,145
645
489
143
13
1
1
+
38
27
79,350
292
136
144
11
1
1
+
18
79,310
610
452
144
12
1
1
+
36
25
79,271
610
452
144
12
1
1
+
37
25
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 emissions from fires on both Forest Land Remaining Forest Land and Land Converted to Forest Land.
c Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grassland.
d Estimates include emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
e Estimates include emissions from N fertilizer additions on both Settlements Remaining Settlements and Land Converted to
Settlements.
f Estimates include emissions from N fertilizer additions on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
6-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 to the
2006 Guidelines for National Greenhouse Gas Inventories: Wetlands. 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 6-6), (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 6-7), and (3) account for greenhouse gas fluxes on all managed
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 6-7) (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
5	See .
6	NRI data are available at .
Land Use, Land-Use Change, and Forestry 6-9

-------
Characteristics Consortium (MRLC) National Land Cover Dataset (NLCD).8 For this Inventory, NRI data have been
extended through 2015 for the conterminous United States and Hawaii (non-federal lands), NLCD data have been
extended through 2016 for the conterminous United States and new FIA data cover the entire time series of land
use data in the conterminous United States and Alaska.
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 46 million hectares is unmanaged,
which has not changed much over the time series of the Inventory (Table 6-7). In 2018, 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
(less than 0.01 percent decrease 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 6-7). 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 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.
Dominant land uses vary by region, largely due to climate patterns, soil types, geology, proximity to coastal
regions, and historical settlement patterns (Figure 6-2). 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 6-6: Managed and Unmanaged Land Area by Land-Use Categories for All 50 States
(Thousands of Hectares)
Land Use Categories
1990
2005
2014
2015
2016a
2017a
2018a
Managed Lands
886,515
886,513
886,513
886,513
886,513
886,513
886,513
Forest
281,621
281,681
281,903
281,945
281,796
281,652
281,546
Croplands
174,471
165,727
162,543
161,929
161,933
161,933
161,933
Grasslands
336,840
337,621
336,437
336,529
336,657
336,781
336,863
Settlements
33,446
40,469
44,367
44,799
44,795
44,797
44,797
Wetlands
38,422
39,017
39,048
39,076
39,089
39,108
39,132
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 this discrepancy in a future Inventory.
12	These "managed area" discrepancies also occur in the Common Reporting Format (CRF) tables submitted to the UNFCCC.
6-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Other	21,715	21,997	22,215 22,236 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	290,111 290,153 290,004 289,860 289,754
Croplands	174,471	165,727	162,543 161,929 161,933 161,933 161,933
Grasslands	362,370	363,583	363,045 363,138 363,266 363,389 363,471
Settlements	33,446	40,469	44,367 44,799 44,795 44,797 44,797
Wetlands	42,589	43,183	43,213 43,241 43,254 43,273 43,297
	Other	32,457	32,725	32,917 32,937 32,944 32,944 32,944
a The land use data for 2017 to 2018 were only partially updated based on new Forest Inventory and Analysis (FIA) data and
land use data for 2016 were partially updated with data from National Land Cover Dataset (NLCD) and FIA. In addition, there
were no new data incorporated for Alaska. New activity data for the National Resources Inventory (NRI) and NLCD will be
incorporated in a future Inventory to update 2016-2018 and 2017-2018, respectively.
Table 6-7: 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
2014
2015
2016b
2017b
2018b
Total Forest Land
281,621
281,681
281,903
281,945
281,796
281,652
281,546
FF
280,393
280,207
280,438
280,528
280,529
280,380
280,274
CF
169
167
143
139
134
135
135
GF
919
1,162
1,171
1,125
989
992
992
WF
77
28
26
25
25
25
25
SF
12
24
26
27
26
26
26
OF
50
93
99
100
93
93
93
Total Cropland
174,471
165,727
162,543
161,929
161,933
161,933
161,933
CC
162,163
150,304
149,492
148,880
148,885
148,884
148,884
FC
182
86
61
58
58
58
58
GC
11,738
14,820
12,616
12,609
12,609
12,609
12,609
WC
118
178
103
104
104
104
104
SC
75
100
92
99
99
99
99
OC
195
239
178
179
179
179
179
Total Grassland
336,840
337,621
336,437
336,529
336,657
336,781
336,863
GG
327,446
315,161
316,242
316,287
316,408
316,502
316,622
FG
593
560
546
547
553
583
545
CG
8,237
17,523
16,229
16,600
16,600
16,600
16,600
WG
176
542
327
308
308
308
308
SG
43
509
386
346
346
346
346
OG
345
3,328
2,707
2,442
2,442
2,442
2,442
Total Wetlands
38,422
39,017
39,048
39,076
39,089
39,108
39,132
WW
37,860
37,035
37,433
37,602
37,616
37,634
37,658
FW
83
59
57
54
54
54
54
CW
132
566
477
440
440
440
440
GW
297
1,187
928
836
836
836
836
SW
0
38
30
25
25
25
25
OW
50
133
123
118
118
118
118
Total Settlements
33,446
40,469
44,367
44,799
44,795
44,797
44,797
SS
30,585
31,522
37,281
38,210
38,210
38,210
38,210
FS
310
549
574
544
539
541
541
Land Use, Land-Use Change, and Forestry 6-11

-------
cs
1,237
3,602
2,662
2,452
2,452
2,452
2,452
GS
1,255
4,499
3,586
3,352
3,352
3,352
3,352
WS
4
61
51
46
46
46
46
OS
54
235
214
197
197
197
197
Total Other Land
21,715
21,997
22,215
22,236
22,243
22,243
22,243
OO
20,953
18,231
18,734
19,000
19,007
19,007
19,007
FO
41
70
94
90
90
90
90
CO
301
590
677
678
678
678
678
GO
391
2,965
2,564
2,331
2,331
2,331
2,331
WO
26
121
127
121
121
121
121
SO
2
20
18
16
16
16
16
Grand Total
886,515
886,513
886,513
886,513
886,513
886,513
886,513
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).
bThe land use data for 2017 to 2018 were only partially updated based on new Forest Inventory and Analysis
(FIA) data and land use data for 2016 were partially updated with data from National Land Cover Dataset
(NLCD) and FIA. In addition, there were no new data incorporated for Alaska. New activity data for the
National Resources Inventory (NRI) and NLCD will be incorporated in a future Inventory to update 2016-2018
and 2017-2018, respectively.
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.
6-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-2: Percent of Total Land Area for Each State in the General Land-Use Categories for
2018
Forest Lands	Croplands
Hawaii
Grasslands
Wetlands
11 - 30
31 - 50
T
~ 11 - 30
31 - 50
Alaska
Hawaii
Alaska
Settlements
Other Lands
Land Use, Land-Use Change, and Forestry 6-13

-------
fvieth ado logy
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. Therefore,
unless wetlands are managed for 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.
6-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
indirectly by human actions such as atmospheric deposition of chemical species produced in industry or
CO2 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 hay or 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.
Land Use, Land-Use Change, and Forestry 6-15

-------
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 grasses, 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 of 0.25 acres (0.1 ha) or
more 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-CC>2 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, D*>hrription and Apphranori 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 6-8). 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, N2O, 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.
20	2006 IPCC Guidelines do not include provisions to separate desert and tundra as land-use categories.
6-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 6-8: Data Sources Used to Determine Land Use and Land Area for the Conterminous
United States, Hawaii, and Alaska


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 2018 for this Inventory, but the time series will be
updated when new data are released.
Land Use, Land-Use Change, and Forestry 6-17

-------
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-219 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.
6-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Managed Land Designation
Lands are designated as managed in the United States based on the definition provided earlier in this section. In
order to apply the definition in an analysis of managed land, the following criteria are used:
•	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 Grassland is 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 land base. The resulting
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.
Land Use, Land-Use Change, and Forestry 6-19

-------
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 6-7) using definitions developed to meet national circumstances, while adhering to IPCC
guidelines (2006).25 In practice, the land was initially classified into a variety of 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 differences in Forest Land definitions and the resulting discrepancies in areas among the land use and
land-use change categories. There are three steps in this process. The first step involves adjustments for Land
Converted to Forest Land (Grassland, Cropland, Settlements, Other Lands, and Wetlands), followed by adjustments
in Forest Land converted to another land use (i.e., Grassland, Cropland, Settlements, Other Lands, and Wetlands),
and finally adjustments to 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 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.
6-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 6-8). 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-CC>2 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.
Land Use, Land-Use Change, and Forestry 6-21

-------
•	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 acre (4.05 ha)
threshold and are Grassland, they will be classified as such 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. The Census has about 46 million more hectares of land in the United States land base 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
6-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
Major updates were made in this Inventory associated with the release of new land use data. The land
representation data were recalculated from the previous Inventory with the following datasets: a) updated FIA
data from 1990 to 2018 for the conterminous United States and Alaska, b) updated NRI data from 1990 to 2015 for
the conterminous United States and Hawaii, and c) updated NLCD data for the conterminous United States from
2001 through 2016. With recalculations, managed Forest Land increased by an average of 1.3 percent across the
time series from 1990 to 2017 according to the new FIA data. According to the new NRI and NLCD data, as well as
harmonization of these data with the new FIA data (See section "Approach for Combining Data Sources"),
Cropland, Grassland, and Other Land decreased by an average of 0.1 percent, 0.6 percent, and 2.1 percent,
respectively, and settlements increased by an average of 0.7 percent.
Planned Improvements
A 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.
Table 6-9: 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.
Land Use, Land-Use Change, and Forestry 6-23

-------
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
emissions from coastal wetlands have been developed for the Inventory using the NOAA C-CAP land cover product.
The NOAA C-CAP product is currently not used directly in the land representation analysis, however, so a planned
improvement for the next (i.e., 1990 through 2019) 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 improved 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.
Lastly, additional land use data from NRI, which currently provides land use information through 2015, and NLCD,
which currently provides land use information through 2016, will be incorporated and used to recalculate the end
of the time series for land use and land use change associated with the conterminous United States, Alaska and
Hawaii. There are also other databases that may need to be integrated into the analysis, particularly for
Settlements.
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 the litter, fumic, and humic layers, and all non-living biomass with a diameter less
than 7.5 centimeters (cm) at transect intersection, lying on the ground.
•	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
6-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CO2 in the case of decomposition and as CO2, Cm, N2O, CO, and NOxwhen the wood product combusts. The rate
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. 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 6.1 Representation of the U.S. Land Base. All annual NFI plots included in the public FIA
database as of May 2019 (which includes data collected through 2018) were used in this Inventory. The NFIs from
each of the conterminous 48 states (CONUS; USDA Forest Service 2018a, 2018b) 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. 2014) and the forest land area estimates included in this report, which are based on the
annual NFI data through 2018 for all states (USDA Forest Service 2018b). 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. (2014). 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
Land Use, Land-Use Change, and Forestry 6-25

-------
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 6.1 Representation of the 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 et al. 2005).
An estimated 77 percent (211 million hectares) of U.S. forests in southeast and southcentral coastal Alaska 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 southeast and southcentral coastal Alaska
forest land and 80 percent of forest land in the conterminous United States are classified as timberland. Of the
remaining non-timberland, 30 million hectares are reserved forest lands (withdrawn by law from management for
production of wood products) and 69 million hectares are lower productivity forest lands (Oswalt et al. 2014).
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 southeast and southcentral coastal Alaska and the conterminous
United States has increased by about 14 million hectares (Oswalt et al. 2014) with the southern region of the
United States containing the most forest land (Figure 6-3). 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 6.1
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 to the net C flux from forest land
across the 1990 to 2018 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. 2014). 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-233 for annual differences between the forest area reported in Section 6.1 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 6.1
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 2018d).
6-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-3: Changes in Forest Area by Region for Forest Land Remaining Forest Land in the
conterminous United States and Alaska (1990-2018, Million Hectares)
100n
or 90H
 South
¦ North
Pacific
Coast
, Rocky
Mountain
_ i i i i | i i i i | i i i i | i i i i | i i i i | i i i
1990 1995 2000 2005 2010 2015
Year
Mountain
Pacific
North
South
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 2018. 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
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 663.2 MMT CO2 Eq. (180.9 MMT C) in 2018 (Table 6-10 and Table 6-11),
The estimated net uptake of C in the Forest Ecosystem was 564,5 MMT CO2 Eq. (153.9 MMT C) in 2018 (Table 6-10
Land Use, Land-Use Change, and Forestry 6-27

-------
and Table 6-11). The majority of this uptake in 2018, 385.2 MMT CO2 Eq. (105.1 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 192 MT C ha 1 from 1990 to 2018. This was calculated by
dividing the Forest Land area estimates by Forest Ecosystem C Stock estimates for every year (see Table 6-12) and
then calculating the mean across the entire time series, i.e., 1990 through 2018. 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 6-4). 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 6-4).
The estimated net uptake of C in HWP was 98.8 MMT CO2 Eq. (26.9 MMT C) in 2018 (Table 6-10 and Table 6-11).
The majority of this uptake, 67.2 MMT CO2 Eq. (18.3 MMT C), was from wood and paper in SWDS. Products in use
were an estimated 31.5 MMT CO2 Eq. (8.6 MMT C) in 2018.
Table 6-10: Net CO2 Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land
and Harvested Wood Pools (MMT CO2 Eq.)
Carbon Pool
1990
2005
2014
2015
2016
2017
2018
Forest Ecosystem
(610.1)
(572.6)
(532.8)
(587.4)
(565.5)
(552.0)
(564.5)
Aboveground Biomass
(425.1)
(391.3)
(390.8)
(404.6)
(397.0)
(381.2)
(385.2)
Belowground Biomass
(98.6)
(90.8)
(88.9)
(92.9)
(91.1)
(87.6)
(88.6)
Dead Wood
(81.9)
(84.1)
(80.3)
(88.4)
(87.6)
(83.1)
(86.4)
Litter
(5.0)
(5.2)
30.2
(3.1)
(0.9)
(3.5)
(3.1)
Soil (Mineral)
0.3
(1.8)
(2.7)
(0.6)
8.2
1.4
(3.3)
Soil (Organic)
(0.6)
(0.1)
(1.0)
1.4
2.3
1.4
1.4
Drained Organic Soil3
0.8
0.8
0.8
0.8
0.8
0.8
0.8
Harvested Wood
(123.8)
(106.0)
(86.0)
(88.7)
(92.4)
(95.7)
(98.8)
Products in Use
(54.8)
(42.6)
(22.3)
(24.6)
(27.8)
(30.3)
(31.5)
SWDS
(69.0)
(63.4)
(63.7)
(64.1)
(64.6)
(65.5)
(67.2)
Total Net Flux
(733.9)
(678.6)
(618.8)
(676.1)
(657.9)
(647.7)
(663.2)
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 6-22 and Table 6-23 for non-C02 emissions from drainage of organic soils from both Forest Land
Remaining Forest Land and Land Converted to Forest Land.
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 6.1 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 6.1
6-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Representation of the U.S. Land Base so there are small differences in the forest land area estimates in this
Section and Section 6.1. See Annex 3.13, Table A-231 for annual differences between the forest area
reported in Section 6.1 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.
Table 6-11: Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land
and Harvested Wood Pools (MMT C)
Carbon Pool
1990
2005
2014
2015
2016
2017
2018
Forest Ecosystem
(166.4)
(156.2)
(145.3)
(160.2)
(154.2)
(150.5)
(153.9)
Aboveground Biomass
(115.9)
(106.7)
(106.6)
(110.4)
(108.3)
(104.0)
(105.1)
Belowground Biomass
(26.9)
(24.8)
(24.2)
(25.3)
(24.9)
(23.9)
(24.2)
Dead Wood
(22.3)
(22.9)
(21.9)
(24.1)
(23.9)
(22.7)
(23.6)
Litter
(1.4)
(1.4)
8.2
(0.8)
(0.3)
(1.0)
(0.8)
Soil (Mineral)
0.1
(0.5)
(0.7)
(0.2)
2.2
0.4
(0.9)
Soil (Organic)
(0.2)
(0.0)
(0.3)
0.4
0.6
0.4
0.4
Drained Organic Soil3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Harvested Wood
(33.8)
(28.9)
(23.4)
(24.2)
(25.2)
(26.1)
(26.9)
Products in Use
(14.9)
(11.6)
(6.1)
(6.7)
(7.6)
(8.3)
(8.6)
SWDS
(18.8)
(17.3)
(17.4)
(17.5)
(17.6)
(17.9)
(18.3)
Total Net Flux
(200.2)
(185.1)
(168.8)
(184.4)
(179.4)
(176.7)
(180.9)
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 6-22 and Table 6-23 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.
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 6.1 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 6.1 Representation
of the U.S. Land Base so there are small differences in the forest land area estimates in this Section and
Section 6.1. See Annex 3.13, Table A-231 for annual differences between the forest area reported in Section
6.1 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.
Land Use, Land-Use Change, and Forestry 6-29

-------
Stock estimates for forest ecosystem and harvested wood C storage pools are presented in Table 6-12. 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 6-12, but these do not precisely match those in
Section 6.1 Representation of the U.S. Land Base for Forest Land Remaining Forest Land. This is because the forest
land area estimates in Table 6-12 only include managed forest land in the conterminous 48 states and Alaska while
the area estimates in Section 6.1 include all managed forest land in Hawaii. Differences also exist because forest
land area estimates are based on the latest NFI data through 2018 and woodland areas previously included as
forest land have been separated and included in the Grassland categories in this Inventory.32
Table 6-12: Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C)

1990
2005
2015
2016
2017
2018
2019
Forest Area (1,000 ha)
279,748
279,749
280,041
280,041
279,893
279,787
279,682
Carbon Pools (MMT C)







Forest Ecosystem
51,527
53,886
55,431
55,592
55,746
55,897
56,051
Aboveground Biomass
11,833
13,484
14,561
14,672
14,780
14,884
14,989
Belowground Biomass
2,350
2,734
2,982
3,008
3,033
3,056
3,081
Dead Wood
2,120
2,454
2,683
2,707
2,731
2,753
2,777
Litter
3,662
3,647
3,638
3,639
3,639
3,640
3,641
Soil (Mineral)
25,636
25,639
25,640
25,640
25,637
25,637
25,638
Soil (Organic)
5,927
5,929
5,927
5,927
5,926
5,926
5,926
Harvested Wood
1,895
2,353
2,567
2,591
2,616
2,642
2,669
Products in Use
1,249
1,447
1,490
1,497
1,505
1,513
1,521
SWDS
646
906
1,076
1,094
1,112
1,129
1,148
Total C Stock
53,423
56,239
57,998
58,183
58,362
58,539
58,720
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 6.1 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 6.1 Representation of the U.S. Land Base so
there are small differences in the forest land area estimates in this Section and Section 6.1. See Annex 3.13, Table A-
231 for annual differences between the forest area reported in Section 6.1 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 2018 requires estimates of C stocks for 2018 and 2019.
32 See Annex 3.13, Table A-233 for annual differences between the forest area reported in Section 6.1 Representation of the
U.S. Land Base and Section 6.2 Forest Land Remaining Forest Land.
6-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-4: Estimated Net Annual Changes in C Stocks for All C Pools in Forest Land
Remaining Forest Land in the Conterminous U.S. and Alaska (1990-2018, MMT C per Year)
20-1
2 L
.r ro
1 Ł
¦S o
^ I-
o>
,c
¦Ł
a)
CD
C
o-
-20-
^0-
-60-
-80-
g « -lOO-
ts
•S to
^5 §
-120-
-140-
€ -160-
o
-180-
-200-
i I i I | i I I I | l I i I | I i I I | I i i I | I i i I
1995 2000 2005 2010 2015
Year
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: CCh 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 land already includes CCh 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
CO2 emissions from fire disturbance, these separate estimates are highlighted here. Note that these CCh
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 CO2 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, the most-recent fire data available were for 2017. That is, fire data for
2018 were not available so estimates from 2017 were used. The 2018 estimates will be updated in subsequent
reports as fire data become available. Estimated CO2 emissions for wildfires in the conterminous 48 states and
Land Use, Land-Use Change, and Forestry 6-31

-------
in Alaska as well as prescribed fires in 2018 were 151 MMT CO2 per year (Table 6-13). This estimate is an
embedded component of the net annual forest C stock change estimates provided previously (i.e., Table 6-11),
but this separate approach to estimate CO2 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 6-13 include only
managed forest land within the state, which is consistent with C stock change estimates provided above.
Table 6-13: 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 Alaska
Prescribed Fires
Total C02 emitted
Year
States (MMTyr1)
(MMTyr1)
(MMTyr1)
(MMTyr1)
1990
6.2
5.3
0.2
11.7

2005
20.5
44.1
1.5
66.2

2014
60.3
3.5
10.4
74.2
2015
115.8
41.2
6.1
163.1
2016
34.0
1.7
9.7
45.4
2017
141.1
1.5
8.6
151.1
2018b
141.1
1.5
8.6
151.1
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 The data for 2018 were unavailable when these estimates were summarized; therefore 2017, the most recent
available estimate, is applied to 2018.
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 2018b) 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 6.1 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. To implement the stock-difference approach, forest Land
conditions in the CONUS were observed on NFI plots at time to and at a subsequent time ti=to+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 to
was then projected from ti to 2018. 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
6-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018. 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 2018. 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 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 2018a, 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
2018d, 2018a). 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
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 2018d). Forest C estimates are organized according to these state surveys, and the frequency of surveys
varies by state.
Land Use, Land-Use Change, and Forestry 6-33

-------
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 2018b, 2018c). 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
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
6-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 6-10 and Table 6-11 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-CC>2 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
(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.
Land Use, Land-Use Change, and Forestry 6-35

-------
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 CO2 flux using IPCC
Approach 1 (Table 6-14). 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 2018 net annual change for forest C stocks was estimated to be between -
846.3 and -480.6 MMT CO2 Eq. around a central estimate of -663.2 MMT CO2 Eq. at a 95 percent confidence level.
This includes a range of-745.5 to -383.4 MMT CO2 Eq. around a central estimate of-564.5 MMT CO2 Eq. for forest
ecosystems and -125.9 to -74.7 MMT CO2 Eq. around a central estimate of -98.8 MMT CO2 Eq. for HWP.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
Table 6-14: Quantitative Uncertainty Estimates for Net CO2 Flux from Forest Land
Remaining Forest Land: Changes in Forest C Stocks (MMT CO2 Eq. and Percent)
2018 Flux Estimate Uncertainty Range Relative to Flux Estimate
(MMT C02 Eq.)	(MMT CP2 Eq.)	(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Forest Ecosystem C Pools3
C02
(564.5)
(745.5)
(383.4)
-32.1%
32.1%
Harvested Wood Products'5
C02
(98.8)
(125.9)
(74.7)
-27.4%
24.4%
Total Forest
C02
(663.2)
(846.3)
(480.6)
-27.6%
27.5%
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.
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 2018d).
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
standard inventory summaries such as the forest resource statistics of Oswalt et al. (2014) or selected population
estimates generated from the FIA database, which are available at an FIA internet site (USDA Forest Service
2018b). 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
6-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 CFU
emissions from landfills based on EPA (2006) data are reasonable in comparison to CFU 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 2018 are consistent with those used in the 1990 through 2017 Inventory.
New NFI data contributed to increases in forest land area and stock changes, particularly in the Intermountain
West region (Table 6-15). 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 6-15) This resulted in a structural
change in the soil organic carbon estimates for mineral and organic soils across the entire time series (Table 6-10).
Updated HWPs data from 2003 through 2017 led to changes in Products in Use and SWDS between the previous
Inventory and the current Inventory (Table 6-16).
Table 6-15: Recalculations of Forest Area (1,000 ha) and C Stocks in Forest Land Remaining
Forest Land and Harvested Wood Pools (MMT C)

Previous Estimate
Current Estimate
Current Estimate

Year 2018,
Year 2018,
Year 2019,

2019 Inventory
2020 Inventory
2020 Inventory
Forest Area (1000 ha)
273,791
279,787
279,682
Carbon Pools (MMT C)



Forest
57,687
55,897
56,051
Aboveground Biomass
14,664
14,884
14,989
Belowground Biomass
3,042
3,056
3,081
Dead Wood
2,744
2,753
2,777
Litter
3,639
3,640
3,641
Soil (Mineral)
27,816
25,637
25,638
Soil (Organic)
5,781
5,926
5,926
Harvested Wood
2,640
2,642
2,669
Products in Use
1,510
1,513
1,521
SWDS
1,130
1,129
1,148
Total Stock
60,328
58,539
58,720
Note: Totals may not sum due to independent rounding.


Table 6-16: 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 2017,
Year 2017,
Year 2018,
Carbon Pool (MMT C)
2019 Inventory
2020 Inventory
2020 Inventory
Forest
(141.2)
(150.5)
(153.9)
Aboveground Biomass
(97.4)
(104.0)
(105.1)
Belowground Biomass
(22.9)
(23.9)
(24.2)
Dead Wood
(21.1)
(22.7)
(23.6)
Litter
(1.0)
(1.0)
(0.8)
Soil (Mineral)
0.6
0.4
(0.9)
Soil (Organic)
0.4
0.4
0.4
Land Use, Land-Use Change, and Forestry 6-37

-------
Drained organic soil	0.2	0.2	0.2
Harvested Wood	(28.2)	(26.1)	(26.9)
Products in Use	(9.7)	(8.3)	(8.6)
SWDS	(18.4)	(17.9)	(18.3)
Total Net Flux	(169.4)	(176.7)	(180.9)
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., wild fire, 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 2018b). 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.
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 2018b). 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
6-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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-CC>2 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 2018, emissions from this source were estimated to be
11.3 MMT CO2 Eq. of CH4 and 7.5 MMT CO2 Eq. of N2O (Table 6-17; kt units provided in Table 6-18). The estimates
of non-CC>2 emissions from forest fires include wildfires and prescribed fires in the conterminous 48 states and all
managed forest land in Alaska.
Table 6-17: N011-CO2 Emissions from Forest Fires (MMT CO2 Eq.)a
Gas
1990
2005
2014
2015
2016
2017
2018b
ch4
0.9
5.0
5.6
12.2
3.4
11.3
11.3
n2o
0.6
3.3
3.7
8.1
2.2
7.5
7.5
Total
1.5
8.2
9.2
20.3
5.6
18.8
18.8
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.
bThe data for 2018 were unavailable when these estimates were developed, therefore
2017, the most recent available estimate, is applied to 2018.
Table 6-18: N011-CO2 Emissions from Forest Fires (kt)a
Gas
1990
2005
2014
2015
2016
2017
2018b
ch4
35
198
222
489
136
452
452
n2o
2
11
12
27
8
25
25
CO
801
4,507
5,055
11,125
3,092
10,314
10,314
NOx
22
127
142
312
87
289
289
a These estimates include Non-C02 Emissions from Forest Fires on Forest Land Remaining
Forest Land and Land Converted to Forest Land.
bThe data for 2018 were unavailable when these estimates were summarized, therefore
2017, the most recent available estimate, is applied to 2018.
Methodology
Non-CC>2 emissions from forest fires—primarily CFU and N2O 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
Land Use, Land-Use Change, and Forestry 6-39

-------
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-CC>2 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 6-19.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
Table 6-19: Quantitative Uncertainty Estimates of Non-C02 Emissions from Forest Fires
(MMT CO2 Eq. and Percent)3
Source
Gas
2018 Emission Estimate
(MMT C02 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
11.3
9.8
13.0
-13%
15%
Non-C02 Emissions from
Forest Fires
n2o
7.5
6.7
8.3
-11%
12%
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.
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 2018) Inventory to compile estimates of non-C02 emissions from
forest fires are consistent with those used in the previous 1990 through 2017 Inventory. Forest within the MTBS
defined fire perimeters (MTBS Data Summaries 2018) are estimated according to NLCD spatial datasets (Homer et
al. 2015) rather than Ruefenacht et al. (2008) as in past reports. Most of the differences in annual forest area
burned (and thus associated emissions) is due to improperly adjusting the proportion of forest land within a fire to
account for no-data values in an MTBS raster image rather than a similar modified NLCD raster image that
conformed to the spatial extent of the fire. This calculation error only affected some fires; specifically those where
the Landsat images included masked areas (such as for cloud cover). The greater the masked area, the greater the
error in estimated forest land within the fire bounds. These area changes are reflected in the emissions estimates,
which are also revised. See Annex 3.13 for additional information on these changes. Fuel estimates are based on
the distribution of stand-level carbon pools (USDA Forest Service 2017) classified according to ecological
subregions defined in the forest inventory data. Combustion estimates are partly a function of the MTBS severity
classifications and thus can vary within a fire. Most of the differences in annual forest area burned (and thus
associated emissions) as seen in Table A-234 relative to the same table in the previous inventory is due to
improperly adjusting the proportion of forest land within a fire to account for no-data values in an MTBS raster
6-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
image rather than a similar modified NLCD raster image that conformed to the spatial extent of the fire. This
calculation error only affected some fires; specifically those where the Landsat images included masked areas
(such as for cloud cover). The greater the masked area, the greater the error in estimated forest land within the
fire bounds.
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 N2O 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 [NO3], and later converted into N2O 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 N2O emissions from Forest Land Remaining Forest Land and Land Converted to Forest Land33 in 2018
were 0.3 MMT CO2 Eq. (1 kt), and the indirect emissions were 0.1 MMT CO2 Eq. (0.4 kt). Total emissions for 2018
were 0.5 MMT CO2 Eq. (2 kt) and have increased by 455 percent from 1990 to 2018. Total forest soil N2O emissions
are summarized in Table 6-20.
Table 6-20: 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
Direct N20 Fluxes from Soils







MMTCO2 Eq.
0.1
0.3
0.3
0.3
0.3
0.3
0.3
kt N20
+
1
1
1
1
1
1
Indirect N20 Fluxes from Soils







MMTC02 Eq.
0.0
0.1
0.1
0.1
0.1
0.1
0.1
kt N20
+
+
+
+
+
+
+
Total







MMT C02 Eq.
0.1
0.5
0.5
0.5
0.5
0.5
0.5
kt N20
+
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.
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.
Land Use, Land-Use Change, and Forestry 6-41

-------
+ Does not exceed 0.05 MMT C02 Eq. or 0.5 kt.
Methodology
The IPCC Tier 1 approach is used to estimate N2O 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 N2O 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
2018, 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 and the portion of fertilized area are multiplied to obtain annual area estimates of fertilized Douglas-fir
stands. Similar to the Southeast, data are not available for 2005 through 2018, 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 N2O 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 for the portion of volatilized N that is converted to N2O off-site. 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 N2O off-site. The resulting estimates are summed to obtain total indirect emissions.
Uncertainty and Time-Series Consistency
The amount of N2O 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 N2O
flux 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 estimated areas of forested 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).
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 and emission factor input variables are directly applied to the 2018 emission estimates. IPCC (2006)
provided estimates for the uncertainty associated with direct and indirect N2O emission factor for synthetic N
fertilizer application to soils.
34 Uncertainty is unknown for the fertilization rates so a conservative value of ±50 percent is used in the analysis.
6-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Uncertainty is quantified using simple error propagation methods (IPCC 2006). The results of the quantitative
uncertainty analysis are summarized in Table 6-21. Direct N2O fluxes from soils in 2018 are estimated to be
between 0.1 and 1.1 MMT CO2 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 CO2 Eq. for 2018. Indirect N2O emissions in 2018 are 0.1
MMT CO2 Eq. and have a range are between 0.02 and 0.4 MMT CO2 Eq., which is 86 percent below to 238 percent
above the emission estimate for 2018.
Table 6-21: Quantitative Uncertainty Estimates of N2O Fluxes from Soils in Forest Land
Remaining Forest Land and Land Con verted to Forest Land (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT C02 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 2018. 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 N2O and uncertainty ranges are
checked and verified based on the sources of these data.
Recalculations Discussion
No recalculations were performed for the 1990 to 2017 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 CO2 and N2O 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
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 2018).
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 2018) coincident with mapped organic soil locations
35 Estimates of C and C02 emissions from drained organic soils are described in this section but reported in Table 6-10 and Table
6-11 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

-------
(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 2018) 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-CC>2 emissions on forest land
with drained organic soils in 2018 are estimated as 0.1 MMT CO2 Eq. per year (Table 6-22).
The Tier 1 methodology provides methods to estimate C emission as CO2 from three pathways: direct emissions
primarily from mineralization; indirect, or off-site, emissions associated with dissolved organic carbon releasing
CO2 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-CC>2 emissions provided here include CH4 and N2O. 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 N2O can be significant from these drained organic soils
in contrast to the very low emissions from wet organic soils.
Table 6-22: N011-CO2 Emissions from Drained Organic Forest Soilsa'b (MMT CO2 Eq.)
Source
1990
2005
2014
2015
2016
2017
2018
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 6-10 and Table 6-11 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 6-23: Non-C02 Emissions from Drained Organic Forest Soilsa'b (kt)
Source
1990
2005
2014
2015
2016
2017
2018
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 6-10 and Table 6-11 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.
Methodology
The Tier 1 methods for estimating CO2, CFU and N2O 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
6-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
-------
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-CC>2 emissions in 2018 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 CO2 Eq. around a central estimate
of 0.106 MMT CO2 Eq. at a 95 percent confidence level.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018. Details on the emission trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
Table 6-25: Quantitative Uncertainty Estimates for Non-C02 Emissions on Drained Organic
Forest Soils (MMT CO2 Eq. and Percent)3
2018 Emission
Source Estimate Uncertainty Range Relative to Emission Estimate
	(MMT C02 Eq.)	(MMT CP2 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.
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-CC>2 emissions may be included in
either the Wetlands or sections on N2O 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 2017 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.
6-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 1.1 million ha
year"1.
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 6-26). The net flux of C from all forest pool stock changes in 2018 was -110.6
MMT C02 Eq. (-30.2 MMT C) (Table 6-26 and Table 6-27).
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.
Table 6-26: Net CO2 Flux from Forest C Pools in Land Converted to ForestLandby Land Use
Change Category (MMT CO2 Eq.)
Land Use/Carbon Pool
1990
2005
2014
2015
2016
2017
2018
Cropland Converted to Forest Land
(45.9)
(46.1)
(46.3)
(46.3)
(46.3)
(46.3)
(46.3)
Aboveground Biomass
(26.1)
(26.3)
(26.4)
(26.4)
(26.4)
(26.4)
(26.4)
Belowground Biomass
(5.1)
(5.1)
(5.1)
(5.2)
(5.2)
(5.2)
(5.2)
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. See Annex 3.13, Table A-233 for annual differences between the forest area
reported in Section 6.1 Representation of the U.S. Land Base and Section 6.3 Land Converted to Forest Land.
Land Use, Land-Use Change, and Forestry 6-47

-------
Dead Wood
(5.9)
(6.0)
(6.0)
(6.0)
(6.0)
(6.0)
(6.0)
Litter
(8.4)
(8.5)
(8.5)
(8.5)
(8.5)
(8.5)
(8.5)
Mineral Soil
(0.3)
(0.3)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Grassland Converted to Forest Land
(9.8)
(9.6)
(9.6)
(9.6)
(9.7)
(9.7)
(9.7)
Aboveground Biomass
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
Belowground Biomass
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
Dead Wood
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
Litter
(3.8)
(3.8)
(3.8)
(3.8)
(3.8)
(3.8)
(3.8)
Mineral Soil
0.2
0.3
0.3
0.3
0.3
0.3
0.3
Other Land Converted to Forest Land
(14.3)
(14.8)
(14.9)
(14.9)
(14.9)
(14.9)
(14.9)
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
(2.0)
(2.0)
(2.0)
(2.0)
(2.0)
(2.0)
(2.0)
Litter
(4.1)
(4.2)
(4.2)
(4.2)
(4.2)
(4.2)
(4.2)
Mineral Soil
(0.6)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Settlements Converted to Forest Land
(38.6)
(38.7)
(38.8)
(38.9)
(38.9)
(38.9)
(38.9)
Aboveground Biomass
(23.2)
(23.3)
(23.4)
(23.4)
(23.4)
(23.4)
(23.4)
Belowground Biomass
(4.4)
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
(4.5)
Dead Wood
(4.6)
(4.6)
(4.6)
(4.6)
(4.6)
(4.6)
(4.6)
Litter
(6.3)
(6.4)
(6.4)
(6.4)
(6.4)
(6.4)
(6.4)
Mineral Soil
+
+
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Wetlands Converted to Forest Land
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
Aboveground Biomass
(0.4)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
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
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Mineral Soil
+
+
+
+
+
+
+
Total Aboveground Biomass Flux
(60.6)
(60.9)
(61.0)
(61.0)
(61.0)
(61.0)
(61.0)
Total Belowground Biomass Flux
(11.8)
(11.9)
(11.9)
(11.9)
(11.9)
(11.9)
(11.9)
Total Dead Wood Flux
(13.3)
(13.4)
(13.4)
(13.4)
(13.4)
(13.4)
(13.4)
Total Litter Flux
(22.9)
(23.0)
(23.1)
(23.1)
(23.1)
(23.1)
(23.1)
Total Mineral Soil Flux
(0.8)
(1.1)
(1.0)
(1.1)
(1.1)
(1.1)
(1.1)
Total Flux
(109.4)
(110.2)
(110.5)
(110.6)
(110.6)
(110.6)
(110.6)
+ Absolute value does not exceed 0.05 MMT C02 Eq.
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 6.1
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-233 for annual differences
between the forest area reported in Section 6.1 Representation of the U.S. Land Base and Section 6.3 Land Converted to
Forest Land. 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 6-10 and
Table 6-11 of the Forest Land Remaining Forest Land section of the Inventory.
Table 6-27: NetCFIuxfrom Forest C Pools in Land Converted to Forest Land by Land Use
Change Category (MMT C)
Land Use/Carbon Pool
1990
2005
2014
2015
2016
2017
2018
Cropland Converted to Forest Land
(12.5)
(12.6)
(12.6)
(12.6)
(12.6)
(12.6)
(12.6)
Aboveground Biomass
(7.1)
(7.2)
(7.2)
(7.2)
(7.2)
(7.2)
(7.2)
Belowground Biomass
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
Dead Wood
(1.6)
(1.6)
(1.6)
(1.6)
(1.6)
(1.6)
(1.6)
Litter
(2.3)
(2.3)
(2.3)
(2.3)
(2.3)
(2.3)
(2.3)
Mineral Soil
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Grassland Converted to Forest Land
(2.7)
(2.6)
(2.6)
(2.6)
(2.6)
(2.6)
(2.6)
6-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Aboveground Biomass
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
Belowground Biomass
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Dead Wood
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Litter
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
Mineral Soil
+
0.1
0.1
0.1
0.1
0.1
0.1
Other Land Converted to Forest Land
(3.9)
(4.0)
(4.1)
(4.1)
(4.1)
(4.1)
(4.1)
Aboveground Biomass
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
Belowground Biomass
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Dead Wood
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
Litter
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Mineral Soil
(0.2)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Settlements Converted to Forest Land
(10.5)
(10.6)
(10.6)
(10.6)
(10.6)
(10.6)
(10.6)
Aboveground Biomass
(6.3)
(6.4)
(6.4)
(6.4)
(6.4)
(6.4)
(6.4)
Belowground Biomass
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
Dead Wood
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
(1.2)
Litter
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
Mineral Soil
+
+
+
+
+
+
+
Wetlands Converted to Forest Land
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Aboveground Biomass
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Belowground Biomass
+
+
+
+
+
+
+
Dead Wood
+
+
+
+
+
+
+
Litter
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Mineral Soil
+
+
+
+
+
+
+
Total Aboveground Biomass Flux
(16.5)
(16.6)
(16.6)
(16.6)
(16.6)
(16.6)
(16.6)
Total Belowground Biomass Flux
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
(3.2)
Total Dead Wood Flux
(3.6)
(3.7)
(3.7)
(3.7)
(3.7)
(3.7)
(3.7)
Total Litter Flux
(6.3)
(6.3)
(6.3)
(6.3)
(6.3)
(6.3)
(6.3)
Total Mineral Soil Flux
(0.2)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Total Flux
(29.8)
(30.1)
(30.1)
(30.2)
(30.2)
(30.2)
(30.2)
+ Absolute value does not exceed 0.05 MMT C.
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 6.1 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-233 for annual differences between the forest area reported in Section 6.1 Representation of
the U.S. Land Base and Section 6.3 Land Converted to Forest Land. 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 6-10 and Table 6-11 of the Forest Land Remaining
Forest Land section of the Inventory.
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 2018b, 2018c). 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).
Land Use, Land-Use Change, and Forestry 6-49

-------
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 to and at a subsequent time ti=to+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 to was then
projected from ti to 2018. 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
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
6-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 N2O Emissions, CFU
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.
Uncertainty and Time-Seri insistency
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 CO2 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 6-28 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 10 percent below to 10 percent above the 2018 C stock
change estimate of-110.6 MMT CO2 Eq.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
Table 6-28: Quantitative Uncertainty Estimates for Forest C Pool Stock Changes (MMT CO2
Eq. per Year) in 2018 from Land Converted to Forest Land by Land Use Change
zuib mux	Uncertainty Range Relative to Flux Range3
Land Use/Carbon Pool	Estimate
	(MMT C02 Eq.)	(MMT CP2 Eq.)	(%)


Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Cropland Converted to Forest Land
(46.3)
(55.1)
(37.5)
-19%
19%
Aboveground Biomass
(26.4)
(35.0)
(17.8)
-33%
32%
Belowground Biomass
(5.2)
(6.2)
(4.1)
-21%
21%
Dead Wood
(6.0)
(7.2)
(4.8)
-20%
20%
Litter
(8.5)
(9.6)
(7.4)
-12%
13%
Mineral Soils
(0.2)
(0.5)
0.1
-133%
133%
Grassland Converted to Forest Land
(9.7)
(12.1)
(7.2)
25%
25%
Aboveground Biomass
(4.5)
(5.9)
(3.1)
-32%
32%
Belowground Biomass
(0.9)
(1.2)
(0.6)
-31%
31%
Dead Wood
(0.7)
(0.9)
(0.6)
-21%
21%
Land Use, Land-Use Change, and Forestry 6-51

-------
Litter
(3.8)
(4.4)
(3.3)
-14%
14%
Mineral Soils
0.3
(0.1)
0.6
-134%
134%
Other Lands Converted to Forest Land
(14.9)
(17.3)
(12.6)
-16%
16%
Aboveground Biomass
(6.3)
(8.4)
(4.2)
-33%
33%
Belowground Biomass
(1.2)
(1.7)
(0.8)
-35%
35%
Dead Wood
(2.0)
(2.6)
(1.5)
-28%
28%
Litter
(4.2)
(4.8)
(3.5)
-15%
15%
Mineral Soils
(1.1)
(1.9)
(0.4)
-62%
62%
Settlements Converted to Forest Land
(38.9)
(45.3)
(32.4)
-17%
17%
Aboveground Biomass
(23.4)
(29.6)
(17.2)
-26%
26%
Belowground Biomass
(4.5)
(5.8)
(3.2)
-29%
29%
Dead Wood
(4.6)
(5.7)
(3.4)
-25%
25%
Litter
(6.4)
(7.3)
(5.5)
-14%
14%
Mineral Soils
(0.1)
(0.1)
+
-37%
37%
Wetlands Converted to Forest Land
(0.9)
(1.1)
(0.7)
-18%
18%
Aboveground Biomass
(0.5)
(0.6)
(0.3)
-31%
31%
Belowground Biomass
(0.1)
(0.1)
(0.1)
-35%
35%
Dead Wood
(0.1)
(0.2)
(0.1)
-40%
40%
Litter
(0.2)
(0.3)
(0.2)
-26%
26%
Mineral Soils
+
+
+
NA
NA
Total: Aboveground Biomass
(61.0)
(71.9)
(50.2)
-18%
18%
Total: Belowground Biomass
(11.9)
(13.7)
(10.1)
-15%
15%
Total: Dead Wood
(13.4)
(15.2)
(11.7)
-13%
13%
Total: Litter
(23.1)
(24.7)
(21.5)
-7%
7%
Total: Mineral Soils
(1.1)
(1.7)
(0.6)
-48%
48%
Total: Lands Converted to Forest Lands
(110.6)
(121.9)
(99.3)
-10%
10%
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 6-10 and Table 6-11 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
QA/QC and Verification
See QA/QC and Verification sections under Forest Land Remaining Forest Land and for mineral soil estimates
Cropland Remaining Cropland.
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 May 2019 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 8 percent in 2018 between the previous Inventory and the current Inventory (Table 6-29).
6-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
This decrease is directly attributed to the incorporation of annual NFI data into the compilation system and new
data and methods used to compile estimates of C in mineral soils. 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 6-29).
Table 6-29: 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
2017 Estimate,
2017 Estimate,
2018 Estimate,
and Carbon pool (MMT C)
Previous Inventory
Current Inventory
Current Inventory
Cropland Converted to Forest Land
(13.1)
(12.6)
(12.6)
Aboveground Biomass
(7.4)
(7.2)
(7.2)
Belowground Biomass
(1.5)
(1.4)
(1.4)
Dead Wood
(1.7)
(1.6)
(1.6)
Litter
(2.5)
(2.3)
(2.3)
Mineral soil
+
(0.1)
(0.1)
Grassland Converted to Forest Land
(3.0)
(2.6)
(2.6)
Aboveground Biomass
(1.5)
(1.2)
(1.2)
Belowground Biomass
(0.3)
(0.3)
(0.3)
Dead Wood
(0.2)
(0.2)
(0.2)
Litter
(1.1)
(1.0)
(1.0)
Mineral soil
0.1
0.1
0.1
Other Land Converted to Forest Land
(5.0)
(4.1)
(4.1)
Aboveground Biomass
(2.5)
(1.7)
(1.7)
Belowground Biomass
(0.5)
(0.3)
(0.3)
Dead Wood
(0.6)
(0.5)
(0.5)
Litter
(1.4)
(1.1)
(1.1)
Mineral soil
+
(0.3)
(0.3)
Settlements Converted to Forest Land
(11.4)
(10.6)
(10.6)
Aboveground Biomass
(6.8)
(6.4)
(6.4)
Belowground Biomass
(1.3)
(1.2)
(1.2)
Dead Wood
(1.3)
(1.2)
(1.2)
Litter
(1.8)
(1.7)
(1.7)
Mineral soil
+
+
+
Wetlands Converted to Forest Land
(0.4)
(0.2)
(0.2)
Aboveground Biomass
(0.2)
(0.1)
(0.1)
Belowground Biomass
+
+
+
Dead Wood
+
+
+
Litter
(0.1)
(0.1)
(0.1)
Mineral soil
+
+
+
Total Aboveground Biomass Flux
(18.5)
(16.6)
(16.6)
Total Belowground Biomass Flux
(3.6)
(3.2)
(3.2)
Total Dead Wood Flux
(3.9)
(3.7)
(3.7)
Total Litter Flux
(6.9)
(6.3)
(6.3)
Total SOC (mineral) Flux
+
(0.3)
(0.3)
Total Flux
(32.9)
(30.2)
(30.2)
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
Land Use, Land-Use Change, and Forestry 6-53

-------
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 may 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 (SOC) is the main source and sink for atmospheric
CO2 in most soils. IPCC (2006) recommends reporting changes in SOC stocks due to agricultural land-use and
management activities on both 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 SOC 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
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 CO2 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
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.
39	N20 emissions from drained organic soils are included in the Agricultural Soil Management section of the Agriculture chapter
of the Inventory.
6-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2015 United States Department of Agriculture
(USDA) National Resources Inventory (NRI) land-use survey 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 6-33 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 6-30 and Table 6-31). In 2018, mineral soils are estimated to
sequester 49.4 MMT CO2 Eq. from the atmosphere (13.5 MMT C). This rate of C storage in mineral soils represents
about a 15 percent decrease in the rate since the initial reporting year of 1990. Carbon dioxide emissions from
organic soils are 32.8 MMT CO2 Eq. (8.9 MMT C) in 2018, which is a 6 percent decrease compared to 1990. In total,
United States agricultural soils in Cropland Remaining Cropland sequestered approximately 16.6 MMT CO2 Eq. (4.5
MMT C) in 2018.
Table 6-30: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT
COz Eq.)
Soil Type
1990
2005
2014
2015
2016
2017
2018
Mineral Soils
(58.2)
(62.4)
(44.7)
(44.9)
(54.3)
(55.1)
(49.4)
Organic Soils
35.0
33.4
32.5
32.1
31.6
32.8
32.8
Total Net Flux
(23.2)
(29.0)
(12.2)
(12.8)
(22.7)
(22.3)
(16.6)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net

sequestration.







rable 6-31: Net CO2 Flux from Soil C Stock Changes in CroplandRemaim
:)
Soil Type
1990
2005
2014
2015
2016
2017
2018
Mineral Soils
(15.9)
(17.0)
(12.2)
(12.3)
(14.8)
(15.0)
(13.5)
Organic Soils
9.5
9.1
8.9
8.8
8.6
8.9
8.9
Total Net Flux
(6.3)
(7.9)
(3.3)
(3.5)
(6.2)
(6.1)
(4.5)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net
sequestration.
Soil C stocks increase in Cropland Remaining Cropland largely due to sequestration in lands enrolled in CRP (i.e.,
set-aside cropland), as well as from conversion of land into hay production, adoption of conservation tillage (i.e.,
reduced- and no-till practices), and intensification of crop production by limiting the use of bare-summer fallow in
semi-arid regions, and growing a cover crop. However, there is a decline in the net amount of C sequestration (i.e.,
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.
Land Use, Land-Use Change, and Forestry 6-55

-------
2018 is 15 percent less than 1990), and this decline is largely due to lower sequestration rates and less annual
cropland enrolled in the CRP that was initiated in 1985. Soil C losses from drainage of organic soils are relatively
stable across the time series with a small decline associated with the land base declining by 6 percent (based on
2015 estimates) for Cropland Remaining Cropland on organic soils since 1990.
The spatial variability in the 2015 annual soil C stock changes41 are displayed in Figure 6-5 and Figure 6-6 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 and cover crop management, 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 aiong the Pacific Coast
(particularly California), which coincides with the largest concentrations of organic soils in the United States that
are used for agricultural production.
Figure 6-5: 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 C stock changes are estimated for 2016 to 2018 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 C stocks, and positive values represent a net decrease
in soil C stocks.
41 Only national-scale emissions are estimated for 2016 to 2018 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.
6-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-6: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural
Management within States, 2015, Cropland Remaining Cropland
r—
L - /y5
r—i
—r
—Li-
mit C02 ha"1 yr"1
~	<10
~	10 to 20
¦	20 to 30
¦	30 to 40
¦ >40
Note: Only national-scale soil C stock changes are estimated for 2016 to 2018 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 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 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 Methodology
section for a list of crops in the Tier 2 and 3 methods) (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 wili be recalculated in a future Inventory report using
the Tier 2 and 3 methods when data become available.
Soil 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 (i.e., each
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

-------
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)
were collected for each NRI point on a 5-year cycle beginning from 1982 through 1997. For cropland, data had
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 had 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 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 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 C stock changes, soil nitrous oxide (N2O) 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 N2O) 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, tobacco, 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 C stock changes
on federal croplands. Mineral SOC 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 C stock changes from 2016 to 2018 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 2018 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 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 (ARMA) errors
(Brockwell and Davis 2016) is used to estimate the relationship between the surrogate data and the modeled
43 See .
6-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018.
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 SOC 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 from 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).
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

-------
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 C stock changes on the majority of agricultural land on
mineral soils. This approach results in a more complete and accurate accounting of soil C stock 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 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 (i.e., daily time-step version of the Century model)
simulates soil C dynamics (and CO2 emissions and uptake) on a daily time step based on C emissions and
removals from plant production and decomposition processes. These changes in soil 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. 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
6-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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). 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 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, 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 was 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 6.1, 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 C stock changes from 2016 to 2018 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 2018 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 soil C stock change factors to estimate soil 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 at these locations in the survey program (i.e., NRI is restricted
to data collection on non-federal lands). Therefore, land-use patterns at 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.
Land Use, Land-Use Change, and Forestry 6-61

-------
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. U.S.-specific C stock change
factors are derived from published literature to determine the impact of management practices on SOC storage
(Ogle et al. 2003, 2006). The factors include changes in tillage, cropping rotations, intensification, and land-use
change between cultivated and uncultivated conditions. U.S. 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 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 U.S.-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 C stock changes from 2016 to 2018 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 U.S.-specific C loss rates (Ogle et al. 2003) rather than default IPCC rates. The
final estimates include a measure of uncertainty as determined from the Monte Carlo Stochastic Simulation with
1,000 iterations. Emissions are based on the annual data for drained organic soils from 1990 to 2015 for Cropland
Remaining Cropland areas 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 is used to estimate annual C emissions from organic soils from 2016 to 2018 as described
in Box 6-4 of this section. Estimates for 2016 to 2018 will be recalculated in future Inventories when new NRI data
are available.
Uncertainty and Time-Series Consistency
Uncertainty associated with the Cropland Remaining Cropland land-use category is addressed for changes in
agricultural soil C stocks (including both mineral and organic soils). Uncertainty estimates are presented in Table
6-32 for each subsource (mineral soil C stocks 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 2018, additional uncertainty is propagated
through the Monte Carlo Analysis that is associated with the surrogate data method. Soil 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 C stocks in Cropland Remaining Cropland ranges from 497 percent below to 497
percent above the 2018 stock change estimate of -16.6 MMT CO2 Eq. The large relative uncertainty around the
2018 stock change estimate is mostly due to variation in soil C stock changes that is not explained by the surrogate
data method, leading to high prediction error with this splicing method.
6-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-32: Approach 2 Quantitative Uncertainty Estimates for Soil C Stock Changes
occurring within Cropland Remaining Cropland (MMT CO2 Eq. and Percent)
Source
2018 Flux Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Flux Estimate3
(MMTCOzEq.) (%)


Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Mineral Soil C Stocks: Cropland Remaining
Cropland, Tier 3 Inventory Methodology
Mineral Soil C Stocks: Cropland Remaining
Cropland, Tier 2 Inventory Methodology
Organic Soil C Stocks: Cropland Remaining
Cropland, Tier 2 Inventory Methodology
(43.5)
(5.9)
32.8
(123.6)
(12.3)
13.8
36.6
(0.5)
51.8
-184%
-109%
-58%
184%
109%
58%
Combined Uncertainty for Flux associated
with Agricultural Soil Carbon Stock Change in
Cropland Remaining Cropland
(16.6)
(99.2)
66.0
-497%
497%
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. The IPCC (2006) does not recommend reporting of annual crop biomass in Cropland Remaining Cropland
because all of the biomass senesces each year and so there is no long-term storage of C in this pool. For woody
plants, biomass C stock changes are likely minor in perennial crops, such as orchards and nut plantations. There
will be some 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 be significantly
changing over the Inventory time series, 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.
However, 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 2018. 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. The comparisons include 72 long-term experiment
sites and 142 NRI soil monitoring network sites, with 948 observations across all of the sites (see Ogle et al. 2007
and Annex 3.12 for more information). The original statistical model developed from the comparisons to
experimental data did not separate croplands and grasslands, and it was discovered through additional testing that
the DayCent model had less bias in predicting soil C stock changes for croplands than grasslands. Therefore,
corrective actions were taken to include a grassland/cropland indicator variable in the statistical model to address
differences in the DayCent model prediction capability.
Land Use, Land-Use Change, and Forestry 6-63

-------
Recalculations Discussion
Several major improvements have been implemented in this Inventory leading to the need for recalculations,
including (1) development of a more detailed time series of management activity data by combining information in
an imputation analysis from USDA-NRCS CEAP survey, USDA-ERS ARMS data, CTIC data and USDA Census of
Agriculture Data; (2) incorporating new land use and crop histories from the NRI survey; (3) incorporating new land
use data from the NLCD; (4) modeling SOC stock changes to 30 cm depth with the Tier 3 approach (previously
modeled to 20 cm depth); (5) modeling the N cycle with freeze-thaw effects on soil N2O emissions; (6) addressing
the effect of cover crops on greenhouse gas emissions and removals; and (7) incorporating measurements of soil
organic C stocks from NRI survey locations for evaluating uncertainty in DayCent model estimates. Other
improvements include better resolving the timing of tillage, planting, fertilization and harvesting based on the
USDA-NRCS CEAP survey and state level information on planting and harvest dates; improving the timing of
irrigation; and crop senescence using growing degree relationships; and estimating soil C stock changes on federal
lands in the conterminous United States. The surrogate data method was also applied to re-estimate stock changes
from 2016 to 2017. These changes resulted in an average increase in soil C sequestration of 2.5 MMT CO2 Eq., 36
percent, from 1990 to 2018 relative to 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).
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.
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 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 6-33 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 time line may be extended if there are insufficient resources to fund all
or part of these planned improvements.
6-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-33: Area of Managed Land in Cropland Remaining Cropland that is not included in
the current Inventory (Thousand Hectares)
Area (Thousand Hectares)


Not Included in
Year
Managed Land
Inventory
Inventory
1990
162,163
162,163
<1
1991
161,721
161,721
<1
1992
161,252
161,252
<1
1993
159,449
159,449
<1
1994
157,732
157,732
<1
1995
157,054
157,054
<1
1996
156,409
156,409
<1
1997
155,767
155,767
<1
1998
152,016
152,016
<1
1999
151,135
151,135
<1
2000
150,981
150,981
<1
2001
150,471
150,471
<1
2002
150,175
150,175
<1
2003
150,843
150,843
<1
2004
150,645
150,645
<1
2005
150,304
150,304
<1
2006
149,791
149,791
<1
2007
150,032
150,032
<1
2008
149,723
149,723
<1
2009
149,743
149,743
<1
2010
149,343
149,343
<1
2011
148,844
148,844
<1
2012
148,524
148,524
<1
2013
149,018
149,018
<1
2014
149,492
149,492
<1
2015
148,880
148,880
<1
2016
ND
ND
ND
2017
ND
ND
ND
2018
ND
ND
ND
Note: NRI data are not available after 2015, and these years are designated as ND (No data).
6.5 Land Converted to Cropland fCRF 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 Use, Land-Use Change, and Forestry 6-65

-------
Land Converted to Cropland (see Section 6.1 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 6-37 in Planned Improvement 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 2006IPCC Guidelines recommend reporting changes in biomass, dead organic matter and soil organic carbon
(SOC) stocks with land use change. All SOC 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 2018, 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 16
percent of the total emissions (Table 6-34 and Table 6-35).
The net change in total C stocks for 2018 led to CO2 emissions to the atmosphere of 55.3 MMT CO2 Eq. (15.1 MMT
C), including 28.5 MMT CO2 Eq. (7.8 MMT C) from aboveground biomass C losses, 5.6 MMT CO2 Eq. (1.5 MMT C)
from belowground biomass C losses, 5.9 MMT CO2 Eq. (1.6 MMT C) from dead wood C losses, 8.5 MMT CO2 Eq.
(2.3 MMT C) from litter C losses, 3.1 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 2018 are 2 percent higher than emissions in the
initial reporting year, i.e., 1990.
Table 6-34: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Land Converted to Cropland by Land Use Change Category (MMT CO2 Eq.)

1990
2005
2014
2015
2016
2017
2018
Grassland Converted to Cropland
6.9
7.5
9.7
10.2
8.5
8.7
8.5
Mineral Soils
4.1
4.0
6.2
6.9
5.2
5.4
5.1
Organic Soils
2.7
3.5
3.4
3.3
3.3
3.3
3.3
Forest Land Converted to Cropland
48.6
48.4
48.6
48.7
48.7
48.7
48.7
Aboveground Live Biomass
28.4
28.4
28.4
28.5
28.5
28.5
28.5
Belowground Live Biomass
5.6
5.6
5.6
5.6
5.6
5.6
5.6
Dead Wood
5.8
5.8
5.9
5.9
5.9
5.9
5.9
Litter
8.3
8.4
8.5
8.5
8.5
8.5
8.5
Mineral Soils
0.4
0.2
0.1
0.1
0.1
0.1
0.1
Organic Soils
0.1
0.1
+
+
+
+
+
Other Lands Converted to Cropland
(2.2)
(2.9)
(2.0)
(2.0)
(2.1)
(2.2)
(2.2)
Mineral Soils
(2.3)
(2.9)
(2.0)
(2.0)
(2.1)
(2.2)
(2.2)
Organic Soils
0.2
0.1
+
+
+
+
+
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
+
+
+
+
+
+
+
Wetlands Converted to Cropland
0.8
0.9
0.5
0.5
0.5
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
0.3
0.4
Aboveground Live Biomass
28.4
28.4
28.4
28.5
28.5
28.5
28.5
Belowground Live Biomass
5.6
5.6
5.6
5.6
5.6
5.6
5.6
46 Changes in biomass C stocks are not currently reported for other land use conversions (other than forest land) 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 (i.e., other than forest land) to cropland.
6-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Dead Wood
5.8
5.8
5.9
5.9
5.9
5.9
5.9
Litter
8.3
8.4
8.5
8.5
8.5
8.5
8.5
Total Mineral Soil Flux
2.3
1.3
4.4
5.0
3.3
3.4
3.1
Total Organic Soil Flux
3.7
4.3
3.8
3.7
3.7
3.7
3.7
Total Net Flux
54.1
53.8
56.7
57.2
55.5
55.6
55.3
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.05 MMT C02 Eq.
Table 6-35: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Land Converted to Cropland (MMT C)

1990
2005
2014
2015
2016
2017
2018
Grassland Converted to Cropland
1.9
2.0
2.6
2.8
2.3
2.4
2.3
Mineral Soils
1.1
1.1
1.7
1.9
1.4
1.5
1.4
Organic Soils
0.7
1.0
0.9
0.9
0.9
0.9
0.9
Forest Land Converted to Cropland
13.3
13.2
13.3
13.3
13.3
13.3
13.3
Aboveground Live Biomass
7.8
7.7
7.8
7.8
7.8
7.8
7.8
Belowground Live Biomass
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Dead Wood
1.6
1.6
1.6
1.6
1.6
1.6
1.6
Litter
2.3
2.3
2.3
2.3
2.3
2.3
2.3
Mineral Soils
0.1
+
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Other Lands Converted to Cropland
(0.6)
(0.8)
(0.5)
(0.6)
(0.6)
(0.6)
(0.6)
Mineral Soils
(0.6)
(0.8)
(0.5)
(0.6)
(0.6)
(0.6)
(0.6)
Organic Soils
+
+
+
+
+
+
+
Settlements Converted to Cropland
+
+
+
+
+
+
+
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted to Cropland
0.2
0.3
0.1
0.1
0.1
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
7.8
7.7
7.8
7.8
7.8
7.8
7.8
Belowground Live Biomass
1.5
1.5
1.5
1.5
1.5
1.5
1.5
Dead Wood
1.6
1.6
1.6
1.6
1.6
1.6
1.6
Litter
2.3
2.3
2.3
2.3
2.3
2.3
2.3
Total Mineral Soil Flux
0.6
0.4
1.2
1.4
0.9
0.9
0.8
Total Organic Soil Flux
1.0
1.2
1.0
1.0
1.0
1.0
1.0
Total Net Flux
14.8
14.7
15.5
15.6
15.1
15.2
15.1
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 organic soil 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 2018)
however there is no country-specific data for cropland biomass, so only a default biomass estimate (IPCC 2006) for
croplands was used to estimate carbon stock changes (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
Land Use, Land-Use Change, and Forestry 6-67

-------
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
SOC 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 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. SOC 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 2018, a surrogate data method is used to estimate soil 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
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).
6-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018 will be recalculated in
future inventories when new NRI data are available.
Tier 3 Approach. For the Tier 3 method, mineral SOC 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 C stock changes from 2016 to 2018 are 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 Planned Improvements section in
Cropland Remaining Cropland).
Tier 2 Approach. For the mineral soils not included in the Tier 3 analysis, SOC 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 6-4 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 Cropland are estimated using the Tier 2
method provided in IPCC (2006), with U.S.-specific C loss rates (Ogle et al. 2003) as described in the Cropland
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 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 2018, there is additional uncertainty propagated through the Monte Carlo Analysis associated
with a surrogate data method, which is also described in Cropland Remaining Cropland.
48 See .
Land Use, Land-Use Change, and Forestry 6-69

-------
Uncertainty estimates are presented in Table 6-36 for each subsource (i.e., biomass C stocks, dead wood C stocks,
litter C stocks, mineral soil C stocks and organic soil C stocks) 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 98
percent below to 98 percent above the 2018 stock change estimate of 55.3 MMT CO2 Eq. The large relative
uncertainty in the 2018 estimate is mostly due to variation in soil C stock changes that is not explained by the
surrogate data method, leading to high prediction error with this splicing method.
Table 6-36: 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)
Source
2018 Flux Estimate
Uncertainty Range Relative to Flux Estimate3
(MMT C02 Eq.)
(MMT C02
Eq.)
(%)



Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Grassland Converted to Cropland
8.5
(29.3)
46.2
-446%
446%
Mineral Soil C Stocks: Tier 3
0.9
(36.7)
38.4
-4302%
4302%
Mineral Soil C Stocks: Tier 2
4.3
1.3
7.2
-69%
69%
Organic Soil C Stocks: Tier 2
3.3
0.9
5.8
-74%
74%
Forest Land Converted to Cropland
48.7
9.5
87.8
-80%
81%
Aboveground Live Biomass
28.5
(7.7)
64.7
-127%
127%
Belowground Live Biomass
5.6
(1.5)
12.8
-127%
127%
Dead Wood
5.9
(1.6)
13.3
-127%
127%
Litter
8.5
(2.3)
19.4
-127%
127%
Mineral Soil C Stocks: Tier 2
0.1
+
0.3
-122%
122%
Organic Soil C Stocks: Tier 2
+
(0.1)
0.1
-994%
994%
Other Lands Converted to Cropland
(2.2)
(3.5)
(1.0)
-57%
57%
Mineral Soil C Stocks: Tier 2
(2.2)
(3.5)
(1.0)
-57%
57%
Organic Soil C Stocks: Tier 2
+
+
+
+
+
Settlements Converted to Cropland
(0.1)
(0.3)
+
-109%
109%
Mineral Soil C Stocks: Tier 2
(0.2)
(0.3)
+
-85%
85%
Organic Soil C Stocks: Tier 2
+
+
0.1
-84%
84%
Wetlands Converted to Croplands
0.6
+
1.1
-92%
92%
Mineral Soil C Stocks: Tier 2
0.2
+
0.5
-101%
101%
Organic Soil C Stocks: Tier 2
0.4
(0.1)
0.9
-138%
138%
Total: Land Converted to Cropland
55.3
0.9
109.8
-98%
98%
Aboveground Live Biomass
28.5
(7.7)
64.7
-127%
127%
Belowground Live Biomass
5.6
(1.5)
12.8
-127%
127%
Dead Wood
5.9
(1.6)
13.3
-127%
127%
Litter
8.5
(2.3)
19.4
-127%
127%
Mineral Soil C Stocks: Tier 3
0.9
(36.7)
38.4
-4302%
4302%
Mineral Soil C Stocks: Tier 2
2.2
(1.0)
5.4
-145%
145%
Organic Soil C Stocks: Tier 2
3.7
1.2
6.2
-67%
67%
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 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
significant 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
6-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
See the QA/QC and Verification section in Cropland Remaining Cropland for information on QA/QC steps.
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. Recalculations for the soil C stock changes are associated with
several improvements to both the Tier 2 and 3 approaches that are discussed in the Recalculations section of
Cropland Remaining Cropland. As a result of these improvements to the Inventory, Land Converted to Cropland has
a smaller reported loss of C compared to the previous Inventory, estimated at an average of 13.4 MMT CO2 Eq.
over the time series. This represents a 19 percent decline in losses of C for Land Converted to Cropland compared
to the previous Inventory, and is largely driven by the methodological changes for estimating the soil C stock
changes.
Soil C stock changes with Forest Land Converted to Cropland are undergoing further evaluation to ensure
consistency in the time series. Different methods are used to estimate soil C stock changes in forest land and
croplands, and while the areas have been reconciled between these land uses, there has been limited evaluation
of the consistency in C stock changes with conversion from forest land to cropland.
There is also an improvement 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 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 6-37 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.
Table 6-37: Area of Managed Land in Land Converted to Cropland that is not included in the
current Inventory (Thousand Hectares)
QA/QC and Verification
Recalculations Discussion
Planned Improvements
Area (Thousand Hectares)
Year
Managed Land
Inventory
Not Included in
Inventory
1990
1991
1992
1993
1994
1995
1996
1997
1998
12,308
12,654
12,943
14,218
15,400
15,581
15,888
16,073
17,440
12,308
12,654
12,943
14,218
15,400
15,581
16,073
17,440
15,888
<1
<1
<1
<1
<1
<1
<1
<1
<1
Land Use, Land-Use Change, and Forestry 6-71

-------
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
Note: NRI data are not available after 2015, and 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 (SOC) 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 SOC, the 2006 IPCC Guidelines (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 privately-owned and federal 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 6-41 in Planned
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: 1990-2018

-------
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 soil C stocks between 1990 and 2018.
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 2018 led to net CO2 emissions to the atmosphere
of 11.2 MMT CO2 Eq. (3.1 MMT C), including 1.4 MMT CO2 Eq. (0.4 MMT C) from net losses of aboveground
biomass C, 0.1 MMT CO2 Eq. (<0.05 MMT C) from net losses in belowground biomass C, 2.6 MMT CO2 Eq. (0.7
MMT C) from net losses in dead wood C, 0.1 MMT CO2 Eq. (<0.05 MMT C) from net gains in litter C, 1.8 MMT CO2
Eq. (0.5 MMT C) from net losses in mineral soil C, and 5.4 MMT CO2 Eq. (1.5 MMT C) from losses of C due to
drainage and cultivation of organic soils (Table 6-38 and Table 6-39). Losses of carbon are 23 percent higher in
2018 compared to 1990, but as noted previously, stock changes are highly variable from 1990 to 2018, with an
average annual change of 9.0 MMT CO2 Eq. (2.5 MMT C).
Table 6-38: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Grassland Remaining Grassland (MMT CO2 Eq.)
Soil Type
1990
2005
2014
2015
2016
2017
2018
Aboveground Live Biomass
1.6
1.5
1.5
1.5
1.5
1.4
1.4
Belowground Live Biomass
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Dead Wood
3.4
3.1
2.7
2.7
2.6
2.6
2.6
Litter
+
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Mineral Soils
(2.2)
0.8
10.0
4.0
0.1
1.5
1.8
Organic Soils
6.3
5.2
5.5
5.4
5.4
5.4
5.4
Total Net Flux
9.1
10.7
19.7
13.6
9.6
10.9
11.2
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
+ Does not exceed 0.05 MMT C02 Eq.
Table 6-39: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Grassland Remaining Grassland (MMT C)
Soil Type
1990
2005
2014
2015
2016
2017
2018
Aboveground Live Biomass
0.4
0.4
0.4
0.4
0.4
0.4
0.4
Belowground Live Biomass
+
+
+
+
+
+
+
Dead Wood
0.9
0.8
0.7
0.7
0.7
0.7
0.7
Litter
+
+
+
+
+
+
+
Mineral Soils
(0.6)
0.2
2.7
1.1
+
0.4
0.5
Organic Soils
1.7
1.4
1.5
1.5
1.5
1.5
1.5
Total Net Flux
2.5
2.9
5.4
3.7
2.6
3.0
3.1
+ Does not exceed 0.05 MMT C Eq.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
The spatial variability in the 2015 annual soil C stock changes50 associated with mineral soils is displayed in Figure
6-7 and organic soils in Figure 6-8. 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.
50 Only national-scale emissions are estimated for 2016 to 2018 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 6-7: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural
Management within States, 2015, Grassland Remaining Grassland
Note: Only national-scale soil C stock changes are estimated for 2016 to 2018 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 C stocks, and positive values represent a net decrease
in soil C stocks.
6-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Figure 6-8: 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 carbon stock changes are estimated for 2016 to 2018 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) the impact from all management on mineral and organic soil C stocks.
Biomass, Dead Wood and Litter Carbon Stock Changes
The methodology described herein is consistent with IPCC (2006). Woodlands are lands that do not meet the
definition of forest land or agroforestry (see Section 6.1 Representation of the U.S. Land Base) but include woody
vegetation and thus may include the five C storage pools (IPCC 2006) 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 between the years to obtain the stock change. The methods for
estimating carbon stocks and stock changes on 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 2019) 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. In some cases, particularly in the Central Plains and Southwest United States, woodlands,
which do not meet the definition forest land, have been measured. 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.
Land Use, Land-Use Change, and Forestry 6-75

-------
Soil Carbon Stock Changes
The following section includes a brief description of the methodology used to estimate changes in soil 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 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) and additional stock changes associated with biosolids (i.e., treated sewage sludge) amendments. SOC
stock changes on the remaining soils are estimated with the IPCC Tier 2 method (Ogle et al. 2003), including land
on very gravelly, cobbly, or shaley soils (greater than 35 percent by volume) and land transferred to private
ownership from federal ownership.51
A surrogate data method is used to estimate soil C stock changes from 2016 to 2018 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
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 surrogate data method.
Stock change estimates for 2016 to 2018 will be recalculated in future inventories when new NRI data are
available.
Tier 3 Approach. Mineral SOC 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, and the remainder is
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.
6-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
deposited on federal lands (i.e., the amount that is not included in DayCent simulations is assumed to be applied
on federal grasslands). Carbon stocks and 95 percent confidence 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 C stock changes from 2016 to 2018 are 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 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 land use and management data that are
used in the Inventory for federal grasslands. 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 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 2018 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 2018 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 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 provided in IPCC (2006), which utilizes U.S.-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 is used to estimate annual C emissions from organic soils from 2016 to 2018 as described
in Box 6-4 of the Methodology section in Cropland Remaining Cropland. Estimates for 2016 to 2018 will be updated
in future Inventories when new NRI data are available.
Land Use, Land-Use Change, and Forestry 6-77

-------
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 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 2018, there is additional uncertainty propagated through the Monte Carlo Analysis associated with the
surrogate data method.
Uncertainty estimates are presented in Table 6-40 for each subsource (i.e., mineral soil C stocks and organic soil C
stocks) 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 C stocks in Grassland Remaining Grassland ranges from more than 1,296 percent
below and above the 2018 stock change estimate of 11.2 MMT CO2 Eq. The large relative uncertainty is mostly due
to variation in soil C stock changes that is not explained by the surrogate data method, leading to high prediction
error with this splicing method.
Table 6-40: Approach 2 Quantitative Uncertainty Estimates for C Stock Changes Occurring
Within Grassland Remaining Grassland (MMT CO2 Eq. and Percent)
2018 Flux Estimate Uncertainty Range Relative to Flux Estimate3
(MMT C02 Eq.)	(MMT CP2 Eq.)	(%)	


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Woodland Biomass:





Aboveground live biomass
1.4
1.0
1.9
-31%
31%
Belowground live biomass
0.1
0.1
0.1
-16%
16%
Dead wood
2.6
2.0
3.1
-22%
22%
Litter
(0.1)
(0.1)
+
-105%
105%
Mineral Soil C Stocks Grassland Remaining
Grassland, Tier 3 Methodology
2.9
(142.3)
148.0
-5054%
5054%
Mineral Soil C Stocks: Grassland Remaining
Grassland, Tier 2 Methodology
(0.9)
(9.8)
8.0
-998%
998%
Mineral Soil C Stocks: Grassland Remaining





Grassland, Tier 2 Methodology (Change in Soil
C due to Biosolids [i.e., Treated Sewage
(0.2)
(0.3)
(0.1)
-50%
50%
Sludge] Amendments)





Organic Soil C Stocks: Grassland Remaining
5.4
1.3
9.5
-77%
77%
Grassland, Tier 2 Methodology
Combined Uncertainty for Flux Associated





with Carbon Stock Changes Occurring in
11.2
(134.3)
156.7
-1,296%
1,296%
Grassland Remaining Grassland





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.
6-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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, on an annual basis, 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 2018. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
See the QA/QC and Verification section in Cropland Remaining Cropland. In addition, quality control uncovered an
error in the DayCent simulations associated with no grazing on pastures and rangelands during the recent
historical period from 1980 to 2015. In the initial simulations, this led to a large increase in soil C stocks.
Corrective actions were taken to ensure grazing was simulated on those lands, which reduced C input to soils and
the amount of C stock change.
This Inventory is the first reporting of biomass, dead wood and litter C stock changes for woodlands. Recalculations
for the soil C stock changes are associated with several improvements to both the Tier 2 and 3 approaches that are
discussed in the Cropland Remaining Cropland section. As a result of these improvements to the Inventory, C
stocks decline on average across the time series for Grassland Remaining Grassland, compared to an average
increase in C stocks in the previous Inventory. The average reduction in C stock change is 14.0 MMT CO2 Eq. over
the time series, which is a 738 percent decrease in C stock changes compared to the previous Inventory. This is
largely driven by the methodological changes associated with estimating soil C stock changes and to a lesser extent
by the inclusion of biomass, dead wood and litter C stock changes for woodlands.
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 6-41 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 6-41: Area of Managed Land in Grassland Remaining Grassland's Alaska that is not
included in the current Inventory (Thousand Hectares)
QA/QC and Verification
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
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	272,8,13	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
Land Use, Land-Use Change, and Forestry 6-79

-------
2000
316,242
266,202
50,040
2001
315,689
265,649
50,040
2002
315,232
265,192
50,040
2003
315,442
265,403
50,039
2004
315,459
265,421
50,038
2005
315,161
265,123
50,038
2006
314,841
264,804
50,037
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
Note: NRI data are not available after 2015, and these years are designated as ND (No data).
Non-C02 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-CC>2
greenhouse gases such as Cm and N2O, 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-CC>2
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 2018, Cm and N2O emissions from biomass burning in grasslands were 0.6 MMT CO2 Eq. (12 kt) and
0.3 MMT CO2 Eq. (1 kt), respectively. Annual emissions from 1990 to 2018 have averaged approximately 0.3 MMT
CO2 Eq. (12 kt) of CH4 and 0.3 MMT C02 Eq. (1 kt) of l\l20 (see Table 6-42 and Table 6-43).
Table 6-42: ChU and N2O Emissions from Biomass Burning in Grassland (MMT CO2 Eq.)

1990
2005
2014
2015
2016
2017
2018
ch4
0.1
0.3
0.4
0.3
0.3
0.3
0.3
n2o
0.1
0.3
0.4
0.3
0.3
0.3
0.3
Total Net Flux
0.2
0.7
0.8
0.7
0.6
0.6
0.6
Note: Totals may not sum due to independent rounding.
53 A planned improvement is underway to incorporate woodland tree biomass into the Inventory.
6-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-43: ChU, N2O, CO, and NOx Emissions from Biomass Burning in Grassland (kt)

1990
2005
2014
2015
2016
2017
2018
ch4
3
13
16
13
12
12
12
n2o
+
1
1
1
1
1
1
CO
84
358
442
356
325
345
331
NOx
5
22
27
21
20
21
20
+ Does not exceed 0.5 kt.
Methodology
The following section includes a description of the methodology used to estimate non-CC>2 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-CC>2
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 2018, which is discussed later in this section.
The land area designated as managed grassland is based primarily on the 2012 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. These survey locations 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 6.1 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 6-44).
Table 6-44: 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
Notes: Burned area are not
estimated (NE) for 2015 to 2018
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
54 See .
Land Use, Land-Use Change, and Forestry 6-81

-------
emission factors for Cm (2.3 g Cm per kg dry matter), N2O (0.21 g Cm per kg dry matter), CO (65 g CH4 per kg dry
matter) and NOx (3.9 g Cm 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 2018 because
new activity data have not been compiled for the current Inventory. 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 2014, and the trend is used to approximate the 2015 to 2018 emissions. The Tier
1 method described previously will be applied to recalculate the 2015 to 2018 emissions in a future Inventory.
Uncertainty and Time-Series Consistency
Emissions are estimated using a linear regression model with ARMA errors for 2015 to 2018. The linear regression
ARMA model produced estimates of the upper and lower bounds of the emission estimate and the results are
summarized in Table 6-45. Methane emissions from Biomass Burning in Grassland for 2018 are estimated to be
between approximately 0.0 and 0.7 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 100
percent below and 146 percent above the 2018 emission estimate of 0.3 MMT CO2 Eq. Nitrous oxide emissions are
estimated to be between approximately 0.0 and 0.8 MMT CO2 Eq., or approximately 100 percent below and 146
percent above the 2018 emission estimate of 0.3 MMT CO2 Eq.
Table 6-45: Uncertainty Estimates for Non-C02 Greenhouse Gas Emissions from Biomass
Burning in Grassland (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)

Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Grassland Burning
Grassland Burning
ch4
n2o
0.3
0.3
+
+
0.7
0.8
-100% 146%
-100% 146%
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.
55 See .
6-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 and recalculate the emissions.
Two other planned improvements have been identified for this source category, including a) 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 lead to more accurate estimates of non-CC>2 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 2018).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 6-49 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 (SOC) stocks due
to land use change. All soil 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
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).
Land Use, Land-Use Change, and Forestry 6-83

-------
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 6-46 and Table 6-47). These three
pools led to net emissions in 2018 of 9.4, 2.4, and 4.9 MMT CO2 Eq. (2.6, 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 C stocks, estimated
at 42.2 MMT CO2 Eq. (11.5 MMT C) in 2018. The gains are primarily associated with conversion of Other Land,
which have relatively low soil 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 CO2 emissions to the atmosphere of 1.9 MMT
CO2 Eq. (0.5 MMT C). The total net C stock change in 2018 for Land Converted to Grassland is estimated as a gain of
24.6 MMT CO2 Eq. (6.7 MMT C), which represents an increase in C stock changes of 268 percent compared to the
initial reporting year of 1990.
Table 6-46: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Grassland (MMT CO2 Eq.)

1990
2005
2014
2015
2016
2017
2018
Cropland Converted to Grassland
(18.3)
(23.5)
(14.5)
(15.5)
(17.8)
(18.0)
(18.0)
Mineral Soils
(18.9)
(25.0)
(15.9)
(16.9)
(19.1)
(19.4)
(19.3)
Organic Soils
0.6
1.5
1.3
1.4
1.4
1.4
1.3
Forest Land Converted to







Grassland
15.9
16.0
15.9
15.9
15.9
15.9
15.9
Aboveground Live Biomass
9.8
9.7
9.5
9.4
9.4
9.4
9.4
Belowground Live Biomass
2.5
2.5
2.4
2.4
2.4
2.4
2.4
Dead Wood
(1.2)
(1.0)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
Litter
4.8
4.8
4.9
4.9
4.9
4.9
4.9
Mineral Soils
(0.1)
(0.1)
+
(0.1)
(0.1)
+
+
Organic Soils
+
0.2
0.2
0.2
0.2
0.2
0.2
Other Lands Converted Grassland
(4.2)
(31.7)
(25.5)
(22.8)
(22.2)
(22.1)
(21.9)
Mineral Soils
(4.2)
(31.7)
(25.6)
(22.9)
(22.3)
(22.2)
(21.9)
Organic Soils
+
+
0.1
0.1
0.1
0.1
0.1
Settlements Converted Grassland
(0.2)
(1.4)
(1.1)
(1.0)
(0.9)
(1.0)
(0.9)
Mineral Soils
(0.2)
(1.4)
(1.1)
(1.0)
(0.9)
(1.0)
(0.9)
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted Grassland
0.1
0.2
0.3
0.3
0.3
0.3
0.3
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
0.1
0.2
0.3
0.3
0.3
0.2
0.2
Aboveground Live Biomass
9.8
9.7
9.5
9.4
9.4
9.4
9.4
Belowground Live Biomass
2.5
2.5
2.4
2.4
2.4
2.4
2.4
Dead Wood
(1.2)
(1.0)
(0.9)
(0.9)
(0.9)
(0.9)
(0.9)
Litter
4.8
4.8
4.9
4.9
4.9
4.9
4.9
Total Mineral Soil Flux
(23.4)
(58.2)
(42.5)
(40.8)
(42.4)
(42.5)
(42.2)
Total Organic Soil Flux
0.8
1.9
1.9
1.9
1.9
1.9
1.9
Total Net Flux
(6.7)
(40.3)
(24.9)
(23.2)
(24.8)
(24.9)
(24.6)
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 6-47: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Grassland (MMT C)







1990
2005
2014
2015
2016
2017
2018
Cropland Converted to Grassland
(5.0)
(6.4)
(4.0)
(4.2)
(4.8)
(4.9)
(4.9)
Mineral Soils
(5.2)
(6.8)
(4.3)
(4.6)
(5.2)
(5.3)
(5.3)
Organic Soils
0.2
0.4
0.4
0.4
0.4
0.4
0.4
Forest Land Converted to Grassland
4.3
4.4
4.3
4.3
4.3
4.3
4.3
Aboveground Live Biomass
2.7
2.6
2.6
2.6
2.6
2.6
2.6
Belowground Live Biomass
0.7
0.7
0.6
0.6
0.6
0.6
0.6
6-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Dead Wood
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Litter
1.3
1.3
1.3
1.3
1.3
1.3
1.3
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
+
+
0.1
0.1
0.1
0.1
0.1
Other Lands Converted Grassland
(3.8)
(8.6)
(6.9)
(6.2)
(6.1)
(6.0)
(6.0)
Mineral Soils
(1.2)
(8.6)
(7.0)
(6.3)
(6.1)
(6.1)
(6.0)
Organic Soils
+
+
+
+
+
+
+
Settlements Converted Grassland
+
(0.4)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Mineral Soils
+
(0.4)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted 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
2.6
2.6
2.6
2.6
2.6
Belowground Live Biomass
0.7
0.7
0.6
0.6
0.6
0.6
0.6
Dead Wood
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Litter
1.3
1.3
1.3
1.3
1.3
1.3
1.3
Total Mineral Soil Flux
(6.4)
(15.9)
(11.6)
(11.1)
(11.6)
(11.6)
(11.5)
Total Organic Soil Flux
0.2
0.5
0.5
0.5
0.5
0.5
0.5
Total Net Flux
(1.8)
(11.0)
(6.8)
(6.3)
(6.8)
(6.8)
(6.7)
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 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)
and in the eastern US, IPCC (2006) defaults for biomass in grasslands.
There are limited data on grassland carbon stocks so default biomass estimates (IPCC 2006) for grasslands were
used to estimate carbon stock changes (litter and dead wood carbon stocks were assumed to be zero since no
reference C density estimates exist for croplands) in the eastern US. 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 carbon 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
carbon in woody biomass during abrupt or gradual land use change. To estimate this transfer of carbon in woody
biomass, state-specific carbon densities for woody biomass remaining on these former forest lands following
conversion to grasslands were developed and included in the estimation of carbon 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 carbon density was lost during conversion from
Forest Land to Grasslands. This estimate was used to develop state-specific carbon density estimates for biomass,
Land Use, Land-Use Change, and Forestry 6-85

-------
dead wood, and litter for Grasslands in the West and Great Plains states and these state-specific carbon densities
were applied in the compilation system to estimate the carbon 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,
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.
Soil Carbon Stock Changes
Soil 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 for Land Converted
to Grassland on most mineral soils that are classified in this land use change category. C stock changes on the
remaining soils 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 C stock changes from 2016 to 2018 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
6-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
regression models include 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 2018 will be recalculated in future inventories when new NRI data are
available.
Tier 3 Approach. Mineral SOC 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). C 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 C stock changes from 2016 to 2018 are 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 Planned Improvements section in
Cropland Remaining Cropland).
Tier 2 Approach. For the mineral soils not included in the Tier 3 analysis, SOC 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 U.S.-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 2018 as described in Box 6-4 of the Methodology section in Cropland
Remaining Cropland. Estimates for 2016 to 2018 will be recalculated in future Inventories when new NRI data are
available.
Uncertainty and Time-Seri insistency
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 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
2018, 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 6-48 for each subsource (i.e., biomass C stocks, mineral soil C stocks
and organic soil C stocks) 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
58 Biogeochemical cycles are the flow of chemical elements and compounds between living organisms and the physical
environment.
Land Use, Land-Use Change, and Forestry 6-87

-------
the IPCC (2006), as discussed in the previous paragraph. The combined uncertainty for total C stocks in Land
Converted to Grassland ranges from 138 percent below to 138 percent above the 2018 stock change estimate of
24.6 MMT CO2 Eq. The large relative uncertainty around the 2018 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 C stock changes that is not explained by the surrogate data
method, leading to high prediction error with this splicing method.
Table 6-48: 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
2018 Flux Estimate3
Uncertainty Range Relative to Flux Estimate3
(MMT C02 Eq.)
(MMT C02
Eq.)
(%)



Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Cropland Converted to Grassland
(18.0)
(47.7)
11.8
-166%
166%
Mineral Soil C Stocks: Tier 3
(15.6)
(45.2)
14.0
-189%
189%
Mineral Soil C Stocks: Tier 2
(3.7)
(6.6)
(0.7)
-81%
81%
Organic Soil C Stocks: Tier 2
1.3
+
2.7
-99%
99%
Forest Land Converted to Grassland
15.9
4.5
27.3
-72%
72%
Aboveground Live Biomass
9.4
(0.4)
19.3
-104%
104%
Belowground Live Biomass
2.4
(0.1)
4.8
-105%
104%
Dead Wood
(0.9)
(1.9)
+
-106%
104%
Litter
4.9
(0.2)
10.0
-105%
104%
Mineral Soil C Stocks: Tier 2
+
(0.2)
0.1
-264%
264%
Organic Soil C Stocks: Tier 2
0.2
+
0.4
-104%
104%
Other Lands Converted to Grassland
(21.9)
(33.6)
(10.1)
-54%
54%
Mineral Soil C Stocks: Tier 2
(21.9)
(33.7)
(10.2)
-54%
54%
Organic Soil C Stocks: Tier 2
0.1
+
0.2
-136%
136%
Settlements Converted to Grassland
(0.9)
(1.5)
(0.4)
-58%
58%
Mineral Soil C Stocks: Tier 2
(0.9)
(1.5)
(0.4)
-58%
58%
Organic Soil C Stocks: Tier 2
+
+
+
-289%
289%
Wetlands Converted to Grasslands
0.3
+
0.5
-104%
104%
Mineral Soil C Stocks: Tier 2
+
(0.1)
0.1
-569%
569%
Organic Soil C Stocks: Tier 2
0.2
+
0.5
-105%
105%
Total: Land Converted to Grassland
(24.6)
(58.6)
9.4
-138%
138%
Aboveground Live Biomass
9.4
(0.4)
19.3
-104%
104%
Belowground Live Biomass
2.4
(0.1)
4.8
-105%
104%
Dead Wood
(0.9)
(1.9)
+
-106%
104%
Litter
4.9
(0.2)
10.0
-105%
104%
Mineral Soil C Stocks: Tier 3
(15.6)
(45.2)
14.0
-189%
189%
Mineral Soil C Stocks: Tier 2
(26.6)
(38.7)
(14.5)
-46%
46%
Organic Soil C Stocks: Tier 2
1.9
0.5
3.2
-74%
74%
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 2018. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
6-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
Recalculations Discussion
Differences in biomass, dead wood and litter C stock changes in Forest Land Converted to Grassland can be
attributed to incorporation of the latest FIA data. Recalculations for the soil C stock changes are associated with
several improvements to both the Tier 2 and 3 approaches that are discussed in the Cropland Remaining Cropland
section. As a result of these improvements to the Inventory, Land Converted to Grassland has a larger reported
gain in C compared to the previous Inventory, estimated at 35.2 MMT CO2 Eq. on average over the time series. This
represents a 610 percent increase in C stock changes for Land Converted to Grassland compared to the previous
Inventory, and is largely driven by the methodological changes for estimating the soil C stock changes.
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.
Soil C stock changes with land use conversion from forest land to grassland are undergoing further evaluation to
ensure consistency in the time series. Different methods are used to estimate soil C stock changes in forest land
and grasslands, and while the areas have been reconciled between these land uses, there has been limited
evaluation of the consistency in C stock changes with conversion from forest land to grassland. 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 6-49 provides information on the amount of managed area in Alaska
that is Land Converted to Grassland, which can reach as high as 54 thousand hectares per year.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 are not available until 2002. For information about other improvements, see the Planned
Improvements section in Cropland Remaining Cropland and Grassland Remaining Grassland.
59 All of the Land Converted to Grassland based on the land representation is included in the inventory for 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.
Land Use, Land-Use Change, and Forestry 6-89

-------
Table 6-49: Area of Managed Land in Land Converted to Grass/andm 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
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
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.
6-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 Cm and N2O. The natural
production of Cm 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 Cm 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
site fertility (i.e., concentration of mineral N). 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. This Inventory estimates CO2, N2O, and Cm emissions from
peatlands managed for peat extraction in accordance with IPCC (2006 and 2013) guidelines.
CO2, N2O, and CH4 Emissions from Peatlands Remaining Peatlands
IPCC (2013) recommends reporting CO2, N2O, and Cm 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 CO2 emissions
from Peatlands Remaining Peatlands using the Tier 1 approach. The IPCC methodologies estimate only on-site N2O
and Cm emissions, since off-site N2O 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 CO2 is emitted from the oxidation of the
peat. Since N2O emissions from saturated ecosystems tend to be low unless there is an exogenous source of
nitrogen, N2O 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 N2O, 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).
Land Use, Land-Use Change, and Forestry 6-91

-------
Off-site CO2 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 CO2, 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 CO2 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 0.7 MMT CO2 Eq. in 2018 (see Table
6-50) comprising 0.7 MMT CO2 Eq. (696 kt) of CO2, 0.001 MMT CO2 Eq. (0.0001 kt) of l\l20, and 0.004 MMT CO2 Eq.
(0.0001 kt) of Cm. Total emissions in 2018 were about 5 percent less than total emissions in 2017.
Total emissions from Peatlands Remaining Peatlands have fluctuated between 0.7 and 1.3 MMT CO2 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 2018. Carbon dioxide emissions
from Peatlands Remaining Peatlands have fluctuated between 0.7 and 1.3 MMT CO2 across the time series, and
these emissions drive the trends in total emissions. Methane and N2O 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 2018. 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 2018.
Table 6-50: Emissions from Peatlands Remaining Peatlands (MMT CO2 Eq.)
Gas
1990
2005
2014
2015
2016
2017
2018
co2
1.1
1.1
0.8
0.8
0.7
0.7
0.7
Off-site
1.0
1.0
0.7
0.7
0.7
0.7
0.7
On-site
0.1
0.1
0.1
+
+
+
+
N20 (On-site)
+
+
+
+
+
+
+
CH4 (On-site)
+
+
+
+
+
+
+
Total
1.1
1.1
0.8
0.8
0.7
0.7
0.7
+ Does not exceed 0.05 MMT C02 Eq.
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.
Table 6-51: Emissions from Peatlands Remaining Peatlands (kt)
Gas
1990
2005
2014
2015
2016
2017
2018
co2
1,055
1,101
775
755
733
734
696
Off-site
985
1,030
725
706
686
687
652
On-site
70
71
50
49
47
47
44
N20 (On-site)
+
+
+
+
+
+
+
CH4 (On-site)
+
+
+
+
+
+
+
+ Does not exceed 0.5 kt.
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.
6-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Methodology
The following methodology sections first describes the steps taken to calculate emissions estimates for the years
1990 through 2017, followed by the basic methodology used to update 2018 values.
1990-2017 Off-Site CO2 Emissions
Carbon dioxide emissions from domestic peat production were estimated using a Tier 1 methodology consistent
with IPCC (2006). Off-site CO2 emissions from Peatlands Remaining Peatlands were calculated by apportioning the
annual weight of peat produced in the United States (Table 6-52) 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 2015; USGS 2016). 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).
The Alaska estimates rely on reported peat production from the Alaska Department of Natural Resources, Division
of Geological & Geophysical Surveys (DGGS) annual Alaska's Mineral Industry reports (DGGS 1993 through 2012).
Similar to the U.S. Geological Survey, 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 6-53). However, volume production data
were used to calculate off-site CO2 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 2018 (reported in cubic yards) was assumed to be equal to the 2012 value.
Consistent with IPCC (2013) guidelines, off-site CO2 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 CO2 Emissions section below. CO2 emissions from dissolved organic C were
estimated by multiplying the area of peatlands by the default emissions 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
CO2 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; from 2011 to 2014, imports of
sphagnum moss (nutrient-poor) peat from Canada represented 97 percent of total U.S. peat imports (USGS 2016).
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 CO2 emissions from apparent consumption would involve
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).
Land Use, Land-Use Change, and Forestry 6-93

-------
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 6-52: Peat Production of Lower 48 States (kt)
Type of Deposit
1990
2005
2013
2014
2015
2016
2017
Nutrient-Rich
595.1
657.6
418.5
416.5
405.0
388.1
374.0
Nutrient-Poor
55.4
27.4
46.5
51.5
50.1
52.9
66.0
Total Production
692.0
685.0
465.0
468.0
455.0
441.0
440.0
Note: Totals may not sum due to independent rounding.
Sources: United States Geological Survey (USGS) (1991-2015) Minerals Yearbook: Peat (1994-2014); United States
Geological Survey (USGS) (2016) Mineral Commodity Summaries: Peat (2016).
Table 6-53: Peat Production of Alaska (Thousand Cubic Meters)

1990
2005
2013
2014
2015
2016
2017
Total Production
49.7
47.8
93.1
93.1
93.1
93.1
93.1
Note: Totals may not sum due to independent rounding.
Sources: Division of Geological & Geophysical Surveys (DGGS), Alaska Department of Natural Resources (1997-2015)
Alaska's Mineral Industry Report (1997-2014).
1990-2017 On-site CO2 Emissions
IPCC (2006) recommends 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 6-54. The
annual land area estimates were then multiplied by the IPCC (2013) default emission factor in order to calculate
on-site CO2 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 same assumption that a
single hectare yields 100 metric tons, see Table 6-55. 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).
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).
6-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-54: Peat Production Area of Lower 48 States (hectares)

1990*
2005
2013
2014
2015
2016
2017
Nutrient-Rich
5,951
6,576
4,185
4,165
4,050
3,881
3,740
Nutrient-Poor
554
274
465
515
501
529
660
Total Production
6,920
6,850
4,650
4,680
4,550
4,410
4,400
Note: Totals may not sum due to independent rounding.
*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 6-52, an assumed yield of 100 metric tons per hectare
per year.
Table 6-55: Peat Production Area of Alaska (hectares)

1990
2005
2013
2014
2015
2016
2017
Nutrient-Rich
0
0
0
0
0
0
0
Nutrient-Poor
286
104
210
204
209
201
201
Total Production
286
104
210
204
209
201
201
Note: Totals may not sum due to independent rounding.
Sources: Calculated using peat production values in Table 6-53, an assumed yield of 100 metric tons per hectare per year.
1900-2017 On-site N2O Emissions
IPCC (2006) suggests basing the calculation of on-site N2O 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 CO2 emissions methodology above details the calculation of area data from production data. In order to
estimate N2O emissions, the area of nutrient rich Peatlands Remaining Peatlands was multiplied by the
appropriate default emission factor taken from IPCC (2013).
1990-2017 On-site CH4 Emissions
IPCC (2013) also suggests basing the calculation of on-site Cm 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 CO2 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 Cm 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 6-56 for the calculated area of ditches and drained land.
Table 6-56: Peat Production (hectares)

1990
2005
2013
2014
2015
2016
2017
Lower 48 States
Area of Drained Land
6,574
6,508
4,418
4,446
4,323
4,190
4,180
Area of Ditches
346
343
233
234
228
221
220
Total Production
6,920
6,850
4,650
4,680
4,550
4,410
4,400
Alaska
Area of Drained Land
272
99
200
194
198
191
191
Area of Ditches
14
5
11
10
10
10
10
Total Production
286
104
210
204
209
201
201
Note: Totals may not sum due to independent rounding.
Sources: Calculated using peat production values in Table 6-46, an assumed yield of 100 metric tons per hectare per year,
and an assumed value of 5 percent ditch area.
Land Use, Land-Use Change, and Forestry 6-95

-------
2018 Emissions
A basic inventory update was performed for estimating the 2018 inventory year emissions using values from the
previous 1990 to 2017 Inventory. Estimates of emissions from peatlands remaining peatlands were forecasted for
2018 and peat production values were set equal to 2017. Excel's FORECAST.ETS function was used to predict a
2018 value using historical data via an algorithm called "Exponential Triple Smoothing." This method determined
the overall trend and provided an appropriate estimate for 2018.
Uncertainty and Time-Series Consistency
A Monte Carlo (Approach 2) uncertainty analysis that was run on the 1990 to 2017 Inventory was applied to
estimate the uncertainty of CO2, Cm, and N2O emissions from Peatlands Remaining Peatlands for 2018, 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).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 6-57. Carbon dioxide
emissions from Peatlands Remaining Peatlands in 2018 were estimated to be between 0.6 and 0.8 MMT CO2 Eq. at
the 95 percent confidence level. This indicates a range of 15 percent below to 15 percent above the emission
estimate of 0.7 MMT CO2 Eq. Methane emissions from Peatlands Remaining Peatlands in 2018 were estimated to
be between 0.002 and 0.007 MMT CO2 Eq. This indicates a range of 55 percent below to 88 percent above the
emission estimate of 0.004 MMT CO2 Eq. Nitrous oxide emissions from Peatlands Remaining Peatlands in 2018
were estimated to be between 0.0002 and 0.0008 MMT CO2 Eq. at the 95 percent confidence level. This indicates a
range of 50 percent below to 62 percent above the emission estimate of 0.0005 MMT CO2 Eq.
Details on the emission/removal trends and methodologies through time are described in more detail in the
Introduction and Methodology sections.
6-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-57: Approach 2 Quantitative Uncertainty Estimates for CO2, Cm, and N2O Emissions
from Peat lands Remaining Peat/a nds (MMT CO2 Eq. and Percent)


2018 Emission




Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


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

(%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Peatlands Remaining Peatlands
C02
0.7
0.6
0.8
-15%
15%
Peatlands Remaining Peatlands
ch4
+
+
+
-55%
88%
Peatlands Remaining Peatlands
n2o
+
+
+
-50%
62%
+ 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.
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
No recalculations were performed for the 1990 through 2017 portion of the time series.
Planned Improvements
In order to further improve estimates of CO2, N2O, and Cm 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 find a new source for Alaska peat production. The current source has not been reliably
updated since 2012 and future publication of these data may discontinue.
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.
Additional guidance on quantifying greenhouse gas emissions and removals on Coastal Wetlands is provided in the
2013 Supplement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands (Wetlands
Supplement), which recognizes the particular importance of vascular plants in sequestering CO2 from the
atmosphere within biomass, dead organic material (DOM; including litter and dead wood stocks) and building soil
carbon stocks. Thus, the Wetlands Supplement provides specific guidance on quantifying emissions 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 stock change and presently does not calculate 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.
62 See .
Land Use, Land-Use Change, and Forestry 6-97

-------
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 CFU emissions on Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands,
2)	Carbon 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.
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 (i.e., when managed Vegetated Coastal Wetlands are lost due to subsidence), 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 2013IPCC Wetlands
Supplement methodologies for CFU emissions, coastal wetlands in salinity conditions less than half that of sea
water are sources of Cm 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 CFU
emissions. The Wetlands Supplement provides methodologies to estimate N2O emissions on coastal wetlands that
occur due to aquaculture. While N2O 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 N2O
emissions with the Agricultural Soils Management, Forest Land and Settlements categories. The N2O 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 N2O.
The Wetlands Supplement provides procedures for estimating C stock changes and CFU emissions from mangroves,
tidal marshes and seagrasses. Depending upon their height and area, stock changes from managed 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 use and level of regulatory oversight, all coastal wetlands within the conterminous United
States are included within the managed land area described in Section 6.1, and as such all estimates of C stock
changes, emissions of CH4, and emissions of N2O 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
6-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
while work continues to harmonize data from NOAA's Coastal Change Analysis Program63 with National Resources
Inventory (NRI) 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.
Emissions and Removals from Vegetated Coastal Wetlands
Remaining Vegetated Coastal Wetlands
The conterminous United States hosts 2.9 million hectares of intertidal Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands comprised of tidally influenced palustrine emergent marsh (603,445 ha), palustrine
scrub shrub (142,034 ha) and estuarine emergent marsh (1,837,618 ha), estuarine scrub shrub (97,383 ha) and
estuarine forest (192,151 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 (52,403 ha), warm temperate (901,671 ha),
subtropical (1,862,402 ha) and Mediterranean (56,155 ha) climate zones.
Soils are the largest C pool in Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands, reflecting long-
term removal of atmospheric CO2 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 aboveground biomass C stocks along with soils. Currently, insufficient data exist on C stock
changes in belowground biomass. 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 6-58 through Table 6-60 below summarize nationally aggregated aboveground biomass and soil C stock
changes and Cm emissions on Vegetated Coastal Wetlands. Intact Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands hold a relatively small aboveground biomass C stock (9 MMT C); however, wetlands
maintain a large C stock within the top 1 meter of soil (estimated to be 870 MMT C) to which C accumulated at a
rate of 9.9 MMT CO2 Eq. in 2018. Methane emissions of 3.6 of MMT CO2 Eq. in 2018 offset C removals resulting in
an annual net C removal rate of 6.3 MMT CO2 Eq. in 2018. Dead organic matter stock changes are not calculated in
Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands since this stock is considered to be in steady
state (IPCC 2014). Due to federal regulatory protection, loss of Vegetated Coastal Wetlands slowed considerably in
the 1970s and the current rates of C stock change and Cm emissions are relatively constant over time. Losses of
Vegetated Coastal Wetlands to Unvegetated Open Water Coastal Wetlands (described later in this chapter) and to
other land uses do occur, which, because of the depth to which soil C stocks are impacted, have a significant
impact on the net stock changes in Coastal Wetlands.
Table 6-58: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands (MMT CO2 Eq.)
Year
1990
2005
2014
2015
2016
2017
2018
Soil Flux
(9.9)
(10.0)
(9.9)
(9.9)
(9.9)
(9.9)
(9.9)
Aboveground Biomass Flux
(0.02)
0.04
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
Total C Stock Change
(9.9)
(9.9)
(9.9)
(9.9)
(9.9)
(9.9)
(9.9)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
63 See https://coast.noaa.gov/digitalcoast/tools/lca.html; accessed October 2019.
Land Use, Land-Use Change, and Forestry 6-99

-------
Table 6-59: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands (MMT C)
Year
1990
2005
2014
2015
2016
2017
2018
Soil Flux
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
Aboveground Biomass Flux
(0.01)
0.01
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
Total C Stock Change
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
(2.7)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
Table 6-60: CH4 Emissions from Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands {Vmi CCh Eq. and kt CH4)
Year	1990 2005 2014 2015 2016 2017 2018
Methane Emissions (MMT C02 Eq.) 3.4 3.5	3.6 3.6 3.6	3.6 3.6
Methane Emissions (kt CH4)	137	140 143 143 144 144 144
Methodology
The following section includes a description of the methodology used to estimate changes in aboveground biomass
C stocks, soil C stocks and emissions of Cm 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 2013).
Soil Carbon Stock Changes
Soil C stock changes are estimated for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands for
both mineral and organic soils on wetlands 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, 2005, and 2010
NOAA C-CAP surveys.64 Federal and non-federal lands are represented. Trends in land cover change are
extrapolated to 1990 and 2017 from these datasets. Based upon NOAA C-CAP, coastal wetlands are subdivided
into freshwater (palustrine) and saline (estuarine) classes and further subdivided into emergent marsh, scrub shrub
and forest classes.65 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 between organic and
mineral soils.
Tier 2 level estimates of soil C removal associated with annual soil C accumulation from 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. A single soil emission
factor was used based on Holmquist et al. (2018). The authors found no statistical support to disaggregate soil C
removal factors by climate region, vegetation type, or salinity range (estuarine or palustrine).
64	See https://coast.noaa.gov/digitalcoast/tools/lca.html; accessed October 2019.
65	See https://coast.noaa.gov/digitalcoast/tools/lca.html; accessed October 2019.
6-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Aboveground Biomass Carbon Stock Changes
Aboveground biomass C Stocks for Palustrine and Estuarine marshes are estimated for Vegetated Coastal
Wetlands Remaining Vegetated Coastal Wetlands. Biomass is not sensitive to soil organic content but is
differentiated based on climate zone. 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). Trends in land cover
change are derived from the NOAA C-CAP dataset and extrapolated to cover the entire 1990 to 2018 time series.
Aboveground biomass 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. Currently, a nationwide dataset for belowground
biomass has not been assembled.
Soil Methane Emissions
Tier 1 estimates of Cm 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 Cm emission factors provided in Table 4.14 of the Wetlands Supplement.
The methodology follows Eq. 4.9, Chapter 4 of the Wetlands Supplement, and is applied to the area of Vegetated
Coastal Wetlands Remaining Vegetated Coastal Wetlands on an annual basis.
Uncertainty and Time-Series Consistency
Underlying uncertainties in estimates of soil and aboveground biomass C stock changes and Cm include
uncertainties associated with Tier 2 literature values of soil C stocks, aboveground biomass C stocks and Cm 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 Cm
flux applied. Soil C stocks and Cm fluxes applied are determined from vegetation community classes across the
coastal zone and identified by NOAA C-CAP. Community classes are further subcategorized by climate zones and
growth form (forest, shrub-scrub, marsh). Aboveground biomass classes were subcategorized by climate zones.
Uncertainties for soil and 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. Because mean soil and aboveground 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 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). Uncertainties for Cm 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 apply Cm flux emission factors (delineation of an 18 ppt
boundary) 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 Methodology sections.
Table 6-61: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes and CH4
Emissions occurring within Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands [WW CO2 Eq. and Percent)
Source

2018 Estimate
Uncertainty Range Relative to Estimate
Gas
(MMT CO? Eq.)
(MMT C02 Eq.) (%)



Lower Upper Lower Upper



Bound Bound Bound Bound
Soil C Stock Change
C02
(9.9)
(11.7) (8.1) -29.5% 29.5%
Land Use, Land-Use Change, and Forestry 6-101

-------
Aboveground Biomass C Stock Change C02	(0.02)	(0.03) (0.02) -16.5%	16.5%
CH4 emissions	QU		2j>	4.7 -29.8%	29.8%
Total Flux	(6;3)	(8.8) (3.9) -38.5%	38.5%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
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. Aboveground 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. A team of two evaluated and verified there were no
computational errors within the calculation worksheets. Soil and aboveground biomass C stock change data are
based upon peer-reviewed literature and Cm emission factors derived from the IPCC Wetlands Supplement.
Recalculations Discussion
There were no recalculations for the 1990 through 2017 portion of the time series.
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 aboveground biomass
for coastal wetlands.66 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 for estuarine emergent wetlands. The C-CAP dataset for 2015 is
currently under development with a planned release in 2020. Additional data products for years 2003, 2008 and
2013 are also planned for release. Once complete, land use change for 1990 through 2018 will be recalculated and
extended to 2019 with this updated dataset. Work is currently underway to examine the feasibility of
incorporating seagrass soil and biomass C stocks into the coastal wetland inventory.
Emissions from Vegetated Coastal Wetlands Converted to
On vegetated Open Water Coastal Wetlands
Conversion of intact Vegetated Coastal Wetlands to Unvegetated Open Water Coastal Wetlands is a source of
emissions from soil, biomass, and DOM C stocks. It is estimated that 4,827 ha of Vegetated Coastal Wetlands were
converted to Unvegetated Open Water Coastal Wetlands in 2018. The Mississippi Delta represents more than 40
percent of the total coastal wetland of the United States, and over 90 percent of the conversion of Vegetated
Coastal Wetlands 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 2005 and 2010 C-CAP surveys
coincides with two such events, hurricanes Katrina and Rita (both making landfall in the late summer of 2005), that
66 See https://serc.si.edu/coastalcarbon; accessed October 2019.
6-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
occurred between these C-CAP survey dates. The dataset, consisting of a time series of four time intervals, each
five years in length, creates a challenge in utilizing it to represent the annual rate of wetland loss and for
extrapolation between 1990 and 2018. Future updates to the C-CAP surveys will include a new survey for 2008 in
addition to other years, which will improve the time series of coastal wetland area change.
Shallow nearshore open water within the U.S. Land Representation is recognized as falling under the Wetlands
category within the Inventory. While high resolution mapping of coastal wetlands provides data to support Tier 2
approaches for tracking land cover change, the depth 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 (Crooks et al. 2009; Couvillion et al. 2011;
Delaune and White 2012; IPCC 2013). ATier 1 assumption is also adopted 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.
Table 6-62: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands (MMT CO2 Eq.)
Year
1990
2005
2014
2015
2016
2017
2018
Soil Flux
4.8
3.1
4.8
4.8
4.8
4.8
4.8
Aboveground Biomass Flux
0.04
0.03
0.04
0.04
0.04
0.04
0.04
Dead Organic Matter Flux
0.001
0.0004
0.001
0.001
0.001
0.001
0.001
Total C Stock Change
4.8
3.1
4.8
4.8
4.8
4.8
4.8
Note: Totals may not sum due to independent rounding.
Table 6-63: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands (MMT C)
Year
1990
2005
2014
2015
2016
2017
2018
Soil Flux
1.3
0.8
1.3
1.3
1.3
1.3
1.3
Aboveground Biomass Flux
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Dead Organic Matter Flux
+
+
+
+
+
+
+
Total C Stock Change
1.3
0.9
1.3
1.3
1.3
1.3
1.3
Note: Totals may not sum due to independent rounding.
+ Absolute values does not exceed 0.0005 MMT C.
Methodology
The following section includes a brief description of the methodology used to estimate changes in soil,
aboveground biomass and DOM C stocks for Vegetated Coastal Wetlands Converted to Unvegetated Open Water
Coastal Wetlands.
Soil Carbon Stock Changes
Soil C stock changes 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, 2005 and 2010 NOAA C-CAP surveys. Publicly-owned and privately-
owned lands are represented. Trends in land cover change are extrapolated to 1990 and 2018 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. Country-specific soil C stocks were
Land Use, Land-Use Change, and Forestry 6-103

-------
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, nor wetland classes. Following the Tier 1 approach for
estimating CO2 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 land area for managed coastal wetlands. The methodology follows Eq. 4.6 in the
Wetlands Supplement.
Aboveground Biomass Carbon Stock Changes
Aboveground biomass C stocks for palustrine and estuarine marshes are estimated for Vegetated Coastal
Wetlands Converted to Unvegetated Open Water Coastal Wetlands. Biomass C stock is not sensitive to soil organic
content but is differentiated based on climate zone. 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). Trends in land cover change are derived from the NOAA C-CAP dataset and extrapolated
to cover the entire 1990 to 2018 time series. Conversion to open water results in emissions of all aboveground
biomass C stocks during the year of conversion; therefore, emissions are calculated by multiplying the C-CAP
derived area lost that year in each climate zone by its mean aboveground biomass. Currently, a nationwide dataset
for belowground biomass has not been assembled.
Dead Organic Matter
Dead organic matter (DOM) carbon stocks, which include litter and dead wood stocks for subtropical estuarine
forested wetlands as an emission for Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal
Wetlands across all years. Data 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 mangrove DOM were used (IPCC 2013). Trends in land cover change are derived from the NOAA
C-CAP dataset and extrapolated to cover the entire 1990 to 2018 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 lost that year in by its Tier 1 DOM C stock.
Soil Methane Emissions
A Tier 1 assumption has been applied that salinity conditions are unchanged and hence Cm 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 aboveground biomass C stock changes are associated with
country-specific (Tier 2) literature values of these stocks, and 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.
Uncertainty specific to coastal wetlands include differentiation of palustrine and estuarine community classes,
which determines the soil C stock applied. Soil C stocks applied are determined from vegetation community classes
across the coastal zone and identified by NOAA C-CAP. Community classes are further subcategorized by climate
zones and growth form (forest, shrub-scrub, marsh). Soil and 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 soil and aboveground biomass C stock to a disaggregation of a community class.
Because mean soil and aboveground 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
6-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 subtropical estuarine forested
wetland DOM stocks were derived from those listed for the Tier 1 estimates (IPCC 2013). 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).
Table 6-64: Approach 1 Quantitative Uncertainty Estimates for CO2 Flux Occurring within
Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal Wetlands (MMT
CO2 Eq. and Percent)
Source
2018 Flux Estimate
Uncertainty Range Relative to Flux Estimate
(MMT CO? Eq.)
(MMT CO
2 Eq.)

(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Soil C Stock
4.8
4.1
5.5
-41.7%
+41.7%
Aboveground Biomass C Stock
0.04
0.03
0.05
-16.5%
+16.5%
Dead Organic Matter C Stock
0.001
0.001
0.002
-25.8%
+25.8%
Total Flux
4.8
3.0
6.7
-32.1%
+32.1%
Note: Totals may not sum due to independent rounding.
The C-CAP dataset, consisting of a time series of four time intervals, each five years in length, and two major
hurricanes striking the Mississippi Delta in the most recent time interval (2006 to 2010), creates a challenge in
utilizing it to represent the annual rate of wetland loss and for extrapolation to 1990 and 2018. Uncertainty in the
defining the long-term trend will be improved with release of the 2015 survey, expected in 2020.
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), which could provide a more refined regional
Approach 2-3 for assessing wetland loss and support the national-scale assessment provided by C-CAP.
Based upon the IPCC Tier 1 methodological guidance in the Wetlands Supplement for estimating emissions with
excavation in coastal wetlands, it has been assumed that a 1-meter column of soil has been remobilized with
erosion and the C released immediately to the atmosphere as CO2. This depth of disturbance is a simplifying
assumption that is commonly applied in the scientific literature to gain a first-order estimate of scale of emissions
(e.g., Delaune and White 2012). It is also a simplifying assumption that all that C is released back to the
atmosphere immediately and future development of the country-specific estimate may refine the emissions both
in terms of scale and rate. Given that erosion has been ongoing for multiple decades the assumption that the C
eroded is released to the atmosphere the year of erosion is a reasonable simplification, but one that could be
further refined.
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
against primary scientific literature. Aboveground 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. Dead organic matter 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
were verified by a second QA team. A team of two evaluated and verified there were no computational errors
within the calculation worksheets. Two biogeochemists at the USGS, in addition to members of the NASA Carbon
Land Use, Land-Use Change, and Forestry 6-105

-------
Monitoring System Science Team, corroborated the assumption that where salinities are unchanged Cm emissions
are constant with conversion of Unvegetated Open Water Coastal Wetlands to Vegetated Coastal Wetlands.
Recalculations Discussion
There were no recalculations for the 1990 through 2017 portion of the time series.
Planned Improvements
A refined uncertainty analysis and efforts to improve times series consistency are planned for the 1990 through
2019 Inventory (i.e., 2021 submission to the UNFCCC). An approach for calculating the fraction of remobilized
coastal wetland soil C returned to the atmosphere as CO2 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.
The C-CAP dataset for 2015 is currently under development with a planned release in 2020. Additional data
products for years 2003, 2008, and 2013 are also planned for release. Once complete, land use change for 1990
through 2018 will be recalculated and extended to 2019 with this updated dataset. 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 Land Representation, is recognized as Coastal Wetlands
within the 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 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 a number of
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.
During wetland restoration, Unvegetated Open Water Coastal Wetland is a common intermediary phase bridging
land use transitions from Cropland or Grassland to Vegetated Coastal Wetlands. The period of open water may last
from five to 20 years depending upon management. The conversion of these other land uses to Unvegetated Open
Water Coastal Wetland will result in reestablishment of wetland biomass and soil C sequestration and may result
in cessation of emissions from drained organic soil. Only changes in soil, DOM and aboveground biomass C stocks
are reported in the Inventory at this time, but improvements are being evaluated to include belowground biomass
C stock changes.
Table 6-65: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands
Converted to Vegetated Coastal Wetlands (MMT CO2 Eq.)
Year	1990	2005	2014 2015 2016 2017 2018
Soil C Flux	(0.004) (0.002) (0.004) (0.004) (0.004) (0.004) (0.004)
Aboveground Biomass C Flux (0.01) (0.004)	(0.01) (0.01) (0.01) (0.01) (0.01)
6-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Dead Organic Matter C Flux	(+)
0	(+) (+) (+) (+) (+)
Total C Stock Change	(0.02)	(0.01)	(0.02) (0.02) (0.02) (0.02) (0.02)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.0005 MMT C02 Eq.
Table 6-66: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands
Converted to Vegetated Coastal Wetlands (MMT C)
Year
1990
2005
2014
2015
2016
2017
2018
Soil C Flux
Aboveground Biomass C Flux
Dead Organic Matter C Flux
(0.001)
(0.003)
(+)
(0.001)
(0.001)
0
(0.001)
(0.003)
(+)
(0.001)
(0.003)
(+)
(0.001)
(0.003)
(+)
(0.001)
(0.003)
(+)
(0.001)
(0.003)
(+)
Total C Stock Change
(0.005)
(0.002)
(0.005)
(0.005)
(0.005)
(0.005)
(0.005)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.0005 MMT C.
Methodology
The following section includes a brief description of the methodology used to estimate changes in soil,
aboveground biomass and dead organic matter C stocks, and CFU emissions for Unvegetated Open Water Coastal
Wetlands Converted to Vegetated Coastal Wetlands.
Soil Carbon Stock Change
Soil C stock changes are estimated for Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal
Wetlands on lands 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 and 2010 NOAA C-CAP surveys. Privately-owned and publicly-owned lands are represented. Trends in
land cover change are extrapolated to 1990 and 2018 from these datasets. C-CAP provides peer reviewed country-
level mapping of coastal wetland distribution, including conversion to and from open water. Country-specific soil C
stock change 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).
Tier 2 level estimates of C stock changes associated with annual soil C accumulation in managed 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. Emission 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.
Aboveground Biomass Carbon Stock Changes
Quantification of regional coastal wetland aboveground biomass C stock changes for palustrine and estuarine
marsh vegetation are presented for Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal
Wetlands. Biomass C stock is not sensitive to soil organic content but differentiated based on climate zone. Data
Land Use, Land-Use Change, and Forestry 6-107

-------
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). Trends in land cover change are derived from the NOAA C-CAP
dataset and extrapolated to cover the entire 1990 through 2018 time series. 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
aboveground biomass. Currently, a nationwide dataset for belowground biomass has not been assembled.
Dead Organic Matter
Dead organic matter (DOM) carbon stocks, which include litter and dead wood stocks, are added for subtropical
estuarine forested wetlands for Vegetated Coastal Wetlands Converted to Unvegetated Open Water Coastal
Wetlands across all years. Tier 1 or 2 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 mangrove DOM were used (IPCC 2013). Trends in land cover change
are derived from the NOAA C-CAP dataset and extrapolated to cover the entire 1990 through 2018 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.
Soil Methane Emissions
A Tier 1 assumption has been applied that salinity conditions are unchanged and hence Cm 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 aboveground 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. Soil C stocks applied are determined from vegetation community classes across the coastal
zone and identified by NOAA C-CAP. Community classes are further subcategorized by climate zones and growth
form (forest, shrub-scrub, marsh). Soil and 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
soil C stock to a disaggregation of a community class. Because mean soil and aboveground 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 subtropical estuarine forested
wetland DOM stocks were derived from those listed for the Tier 1 estimates (IPCC 2013). 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). Details on the emission/removal trends and methodologies through time are described in
more detail in the Introduction and Methodology sections.
6-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-67: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes Occurring
within Unvegetated Open Water Coastal Wetlands Converted to Vegetated Coastal Wetlands
(MMT CO2 Eq. and Percent)
Source
2018 Flux Estimate
(MMT CO? Eq.)
Uncertainty Range
(MMT CO? Eq.)
Relative to Flux Estimate
(%)


Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Soil C Stock Flux
Aboveground Biomass C Stock Flux
Dead Organic Matter C Stock Flux
(0.004)
(0.01)
(+)
(0.005)
(0.01)
(+)
(0.004)
(0.01)
(+)
-29.5%
-16.5%
-25.8%
29.5%
16.5%
25.8%
Total Flux
(0.02)
(0.02)
(0.01)
-32.1%
32.1%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
+ Absolute value does not exceed 0.0005 MMT C02 Eq.
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 produced
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. Dead organic matter 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 Cm emissions are constant with conversion of Unvegetated Open Water Coastal
Wetlands to Vegetated Coastal Wetlands.
Recalculations Discussion
There were no recalculations for the 1990 through 2017 portion of the time series.
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 aboveground biomass in coastal wetlands. Reference values for soil and aboveground biomass C stocks will be
updated as new data emerge. Refined error analysis combining land cover change and soil and aboveground
biomass C stock estimates will be updated at those times.
The C-CAP dataset for 2015 is currently under development with a planned release in 2020. Additional data
products for years 2003, 2008, and 2013 are also planned for release. Once complete, land use change for 1990
through 2018 will be recalculated and extended to 2019 with this updated dataset. C-CAP data harmonization with
the NLCD is an ongoing process and will occur in future iterations 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 N2O. Nitrous
oxide is generated and emitted as a byproduct of the conversion of ammonia (contained in fish urea) to nitrate
Land Use, Land-Use Change, and Forestry 6-109

-------
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 2013 Wetlands Supplement).
Aquaculture production in the United States has fluctuated slightly from year to year, with resulting N2O emissions
increasing from 0.1 in 1990 to upwards of 0.2 MMT CO2 Eq. between 1992 and 2010. Levels have essentially
remained consistent since 2011. Aquaculture production data were updated through 2016; however, data through
2018 are not yet available and in this analysis are held constant with 2016 emissions of 0.1 MMT CO2 Eq.
Table 6-68: N2O Emissions from Aquaculture in Coastal Wetlands (MMT CO2 Eq. and kt N2O)
Year
1990
2005
2014
2015
2016
2017
2018
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 N2O emissions from Aquaculture in Coastal Wetlands follows guidance in the 2013
IPCC 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, 2018), from which activity data for this analysis is derived.67 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, 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, 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 N2O 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 likely 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 N2O emissions from aquaculture is not sensitive to salinity using IPCC approaches
and as such the location of aquaculture ponds on the landscape 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 N2O 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 N2O-N per kg offish produced is applied to the
activity data to calculate total N2O emissions.
Uncertainty and Time-Series Consistency
Uncertainty estimates are based upon the Tier 1 default 95 percent confidence interval provided within the
Wetlands Supplement for N2O emissions. Uncertainties in N2O emissions from aquaculture are also based on
expert judgment of the NOAA Fisheries of the United States fisheries production data (± 100 percent) multiplied by
the default uncertainty level for N2O emissions found in Table 4.15, chapter 4 of the Wetlands Supplement. 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.
67 See https://www.fisheries.noaa.gov/resource/document/fisheries-united-states-2017-report; accessed October 2019.
6-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-69: Approach 1 Quantitative Uncertainty Estimates for N2O Emissions for
Aquaculture Production in Coastal Wetlands (MMT CO2 Eq. and Percent)

2018 Emissions




Estimate
Uncertainty Range Relative to Emissions Estimate3
Source
(MMT C02 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 0.31
-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 2018. 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 2013 IPCC Wetlands Supplement to
assess which fisheries production data to include in estimating emissions from aquaculture. It was concluded that
N2O emissions estimates should be applied to any fish production to which food supplement is supplied be they
pond or open water and that salinity conditions were not a determining factor in production of N2O emissions.
Recalculations Discussion
A NOAA report was released in 2018 that contained updated fisheries data for 2016 (National Marine Fisheries
Service 2018). This new value was applied for 2016 and also applied in 2017 and 2018 until more recent data are
released. This resulted in a decrease in N2O emissions by 0.01 MMT CO2 Eq. (0.04 kt N2O) for 2016 and 2017
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 2018 the rate of annual transition for Land
Converted to Vegetated Coastal Wetlands ranged from 2,619 ha/year to 5,316 ha/year.68 Conversion rates were
higher during the period 2010 through 2018 than during the earlier part of the time series.
68 Data from C-CAP; see https://coast.noaa.gov/digitalcoast/tools/lca.html; accessed October 2019.
Land Use, Land-Use Change, and Forestry 6-111

-------
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)69 with NRI, FIA and NLDC data used to compile the Land Representation.
Following conversion to Vegetated Coastal Wetlands, there are increases in plant biomass and soil C storage.
Additionally, at salinities less than half that of seawater, the transition from upland dry soils to wetland soils results
in Cm emissions. In this Inventory analysis, soil and aboveground biomass C stock changes as well as CH4 emissions
are quantified. Estimates of emissions and removals are based on emission factor data that have been applied to
assess changes in soil and aboveground biomass C stocks and CH4 emissions for Land Converted to Vegetated
Coastal Wetlands. The United States calculates emissions and removals based upon stock change and presently
does not calculate 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.
Table 6-70: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal
Wetlands (MMT COz Eq.)
Year
1990
2005
2014
2015
2016
2017
2018
Soil Flux
Aboveground Biomass Flux
(0.01)
(0.03)
(0.01)
(0.02)
(0.01)
(0.03)
(0.01)
(0.03)
(0.01)
(0.03)
(0.01)
(0.03)
(0.01)
(0.03)
Total C Stock Change
(0.04)
(0.03)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
Note: Totals may not sum due to independent rounding.





Table 6-71: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal
Wetlands (MMT C)
Year
1990
2005
2014
2015
2016
2017
2018
Soil Flux
Aboveground Biomass Flux
(0.004)
(0.01)
(0.002)
(0.01)
(0.004)
(0.01)
(0.004)
(0.01)
(0.004)
(0.01)
(0.004)
(0.01)
(0.004)
(0.01)
Total C Stock Change
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
Note: Totals may not sum due to independent rounding.
Table 6-72: ChU Emissions from Land Converted to Vegetated Coastal Wetlands (MMT CO2
Eq. and kt CH4)
Year
1990
2005
2014
2015
2016
2017
2018
Methane Emissions (MMT C02 Eq.)
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Methane Emissions (kt CH4)
0.6
0.5
0.6
0.6
0.6
0.6
0.6
Methodology
The following section includes a description of the methodology used to estimate changes in soil and aboveground
biomass C stocks and CH4 emissions for Land Converted to Vegetated Coastal Wetlands.
Soil Carbon Stock Changes
Soil C removals are estimated for Land Converted to Vegetated Coastal Wetlands 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, and 2010 NOAA C-CAP surveys.70 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. Delineating Vegetated Coastal Wetlands from
69	See https://coast.noaa.gov/digitalcoast/tools/lca.html; accessed October 2019.
70	See https://coast.noaa.gov/digitalcoast/tools/lca.html; accessed October 2019.
6-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. Federal and non-federal lands are represented. Trends in land cover
change are extrapolated to 1990 and 2018 from these datasets. 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. 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).
Tier 2 level estimates of soil C removal 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 that given year. Currently, data are not
available to account for C stock changes for the 20 years prior to conversion to coastal wetlands as per IPCC
convention. 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. Emission factors were developed from
literature references that provided soil C removal factors disaggregated by climate region, vegetation type by
salinity range (estuarine or palustrine) as identified using NOAA C-CAP as described above.
Aboveground Biomass Carbon Stock Changes
Aboveground biomass C stocks for palustrine and estuarine marshes are estimated for Lands Converted to
Vegetated Coastal Wetlands. Biomass is not sensitive to soil organic content but rather is differentiated based on
climate zone. 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). Trends in land cover change are derived from the
NOAA C-CAP dataset and extrapolated to cover the entire 1990 through 2018 time series. Stock changes that occur
by converting lands to vegetated wetlands are calculated by multiplying the C-CAP derived area gained that year in
each climate zone by its mean aboveground biomass. A nationwide dataset for belowground biomass has not been
assembled to date. Currently, data are not available to account for C stock changes for the 20 years prior to
conversion to coastal wetlands as per IPCC convention.
Soil Methane Emissions
Tier 1 estimates of Cm emissions for Land Converted to 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 CFU 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, and is applied to the total area of Land
Converted to Vegetated Coastal Wetlands on an annual basis. Currently, data are not available to account for C
stock changes for the 20 years prior to conversion to coastal wetlands as per IPCC convention.
Uncertainty and Time-Series Consistency
Underlying uncertainties in estimates of soil C removal factors, aboveground biomass change, and CFU emissions
include error in uncertainties associated with Tier 2 literature values of soil C removal estimates, aboveground
biomass stocks, and IPCC default CFU 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 the soil C removal and Cm flux applied. Soil C removal and CFU fluxes applied are determined
from vegetation community classes across the coastal zone and identified by NOAA C-CAP. Community classes are
further subcategorized by climate zones and growth form (forest, shrub-scrub, marsh). Aboveground biomass
Land Use, Land-Use Change, and Forestry 6-113

-------
classes were subcategorized by climate zones. Soil and 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. Because mean soil and aboveground 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 Cm flux are the Tier 1
default values reported in the 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
estimate the Cm flux (e.g., delineation of an 18 ppt boundary), which will need significant improvement to reduce
uncertainties.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2018. Details on the emission/removal trends and methodologies through time are described in more
detail in the Introduction and Methodology sections.
Table 6-73: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes occurring
within Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent)
2018 Estimate Uncertainty Range Relative to Estimate3
(MMT C02 Eq.)	(MMT CP2 Eq.)	(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Soil C Stock Change
(0.01)
(0.01)
(0.01)
-29.5%
29.5%
Aboveground Biomass C Stock Change
(0.03)
(0.03)
(0.03)
-16.5%
16.5%
Methane Emissions
0.01
0.01
0.02
-29.8%
29.8%
Total Uncertainty
(0.03)
(0.04)
(0.02)
-38.5%
38.5%
Note: Totals may not sum due to independent rounding.
a Range of flux estimates based on error propagation at 95 percent confidence interval.
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. Aboveground 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. 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 Cm emission
factors derived from the IPCC Wetlands Supplement.
Recalculations Discussion
An error was found in the calculation for soil carbon removal for subtropical estuarine scrub/shrub wetlands for
the 1990 to 2017 time series. There currently is no soil C accumulation rate calculated from field data for
subtropical estuarine scrub/shrub wetlands so the rate from the most applicable wetland type is used as a proxy.
This rate was erroneously entered as 0.45 t C ha 1 yr1, which is the value calculated for subtropical palustrine
emergent wetlands, and was changed to be 1.091 C ha 1 yr1, which is the value calculated for subtropical estuarine
emergent wetlands and the more applicable rate to this wetland type. This rate is also already used for the
subtropical estuarine scrub/shrub soil C accumulation rate for Wetlands Remaining Wetlands calculations. The
resulting changes in total C removals is below detection at the scale of MMT CO2 yr"1.
6-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 aboveground biomass
for coastal wetlands.71 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 2021 Inventory submission.
The C-CAP dataset for 2015 is currently under development with a planned release in early 2020. Additional data
products for years 2003, 2008, and 2013 are also planned for release. Once complete, land use change for 1996
through 2018 will be recalculated and extended to 2019 with this updated dataset. Currently, biomass from lands
converted to wetlands are only tracked for one year due to lack of available data. In 2020, data harmonization of
C-CAP with the National Land Cover dataset (NLCD) will occur that will enable 20-year tracking of biomass as per
IPCC guidance.
Once harmonization happens for the land cover data, analyses will occur to address the loss of biomass and dead
organic matter (litter and standing dead wood C stocks) that occurs when lands (e.g., forest lands, grasslands) are
converted to vegetated coastal wetlands.
6.10 Settlements Remaining Settlements
(CRF Category 4E1)
Soil Carbon Stock Changes (CRF Category 4E1)
Soil C stock changes for Settlements Remaining Settlements occur in both mineral and organic soils. The United
States does not, however, 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 2006 IPCC Guidelines
(IPCC 2006) that inputs equal outputs, and therefore the soil carbon 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 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 CO2 emissions.72 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 20 18)73 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
71	See https://serc.si.edu/coastalcarbon; accessed October 2019.
72	N20 emissions from soils are included in the N20 Emissions from Settlement Soils section.
73	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.
Land Use, Land-Use Change, and Forestry 6-115

-------
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 6.1
Representation of the U.S. Land Base) and the settlements area included in the Inventory analysis. There is a
planned improvement to include CO2 emissions from drainage of organic soils in settlements of Alaska and federal
lands as part of a future Inventory.
CO2 emissions from drained organic soils in settlements are 15.9 MMT CO2 Eq. (4.3 MMT C) in 2018. 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 6-74: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements
(MMT COz Eq.)
Soil Type
1990
2005
2014
2015
2016
2017
2018
Organic Soils
11.3
12.2 .P;
15.1
15.7
16.0
16.0
15.9
Table 6-75: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements
(MMT C)
Soil Type
1990
2005
2014
2015
2016
2017
2018
Organic Soils
3.1
3.3
4.1
4.3
4.4
4.4
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 6-76 (See Section 6.1,
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 6-76). The area of land on organic soils in Settlements
Remaining Settlements has increased from 2 thousand hectares in 1990 to over 36 thousand hectares in 2015. The
area of land on organic soils are not currently available from NRI for Settlements Remaining Settlements after
2015.
Table 6-76: Thousands of Hectares of Drained Organic Soils in Settlements Remaining
Settlements
Year
Area
(Thousand Hectares)
1990
220
2005
235
2013
284
2014
291
2015
303
2016
ND
2017
ND
2018
ND
Note: No NRI data are available after 2015,
designated as ND (No data)
6-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
To estimate CO2 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 2018
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 2018 emissions. The Tier 2 method described previously
will be applied in future inventories to recalculate the estimates beyond 2015 as activity data become available.
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 2018 based on the linear time series model. The results
of the Approach 2 Monte Carlo uncertainty analysis are summarized in Table 6-77. Soil C losses from drained
organic soils in Settlements Remaining Settlements for 2018 are estimated to be between 7.6 and 24.2 MMT CO2
Eq. at a 95 percent confidence level. This indicates a range of 52 percent below and 52 percent above the 2018
emission estimate of 15.9 MMT CO2 Eq.
Table 6-77: Uncertainty Estimates for CO2 Emissions from Drained Organic Soils in
Settlements Remaining Settlements (MMT CO2 Eq. and Percent)


2018 Emission


Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


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



Lower Upper
Lower Upper



Bound Bound
Bound Bound
Organic Soils
C02
15.9
7.6 24.2
-52% 52%
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 2018. 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. These checks uncovered a few errors in the spreadsheets that were corrected.
There was also an error in handling of activity data for this source category in which settlement areas were only
included if they had been in agriculture during the past. This led to a significant under-estimation in the area of
drained organic soils in settlements that has been corrected in this Inventory (see Recalculations Discussion
below).
Recalculations Discussion
The entire time series was recalculated based on updates to the land representation data with the release of the
2018 NRI (USDA-NRCS 2018) and additional information from the NLCD (Yang et al. 2018; Fry et al. 2011; Homer et
al. 2007, 2015). In addition, the data splicing method has been used to re-estimate CO2 emissions for 2016 to 2017
in the previous Inventory. However, the major change was the correction of a quality control problem that led to
Land Use, Land-Use Change, and Forestry 6-117

-------
an under-estimation of drained organic soils in settlements. The recalculations led to an increase in emissions of
11.9 MMT CO2 Eq., or >6,500 percent, on average across the entire time series.
Planned Improvements
This source will be updated to include CO2 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 6-78 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 6-78: Area of Managed Land in Settlements Remaining Settlements that is not
included in the current Inventory (Thousand Hectares)
Area (Thousand Hectares)
SRS Area SRS Area Not
SRS Managed Land Included in Included in
Year
Area (Section 6.1)
Inventory
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
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
6-118 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
2018	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 115.4
MMT CO2 Eq. (31.5 MMT C) over the period from 1990 through 2018. Net C sequestration from settlement trees in
2018 is estimated to be 129.8 MMT CO2 Eq. (35.4 MMT C) (Table 6-79). 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), which has been trending downward recently and decreasing net sequestration. 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 2018 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 6-79: Net Flux from Settlement Trees in Settlements Remaining Settlements (MMT
COz Eq. and MMT C)a
Year
MMTCOz Eq.
MMT C
1990
(96.4)
(26.3)
2005
(117.4)
(32.0)
2014
(129.4)
(35.3)
2015
(130.4)
(35.6)
2016
(129.8)
(35.4)
2017
(129.8)
(35.4)
2018
(129.8)
(35.4)
Note: Parentheses indicate net
sequestration.
aThese estimates include net C02 and C flux
from Settlement Trees on Settlements
Remaining Settlements and Land Converted
to Settlements.
Methodology
To estimate net carbon sequestration in settlement areas, three types of data are required by state:
1.	Settlement area
2.	Percent tree cover in settlement areas
3.	Carbon sequestration density per unit of tree cover
Land Use, Land-Use Change, and Forestry 6-119

-------
Settlement Area
Settlements area is defined in Section 6.1 Representation of the U.S. Land Base as a land-use category representing
developed areas. The data used to estimate settlement area within Section 6.1 comes from the NRI as updated
through 2015. Annual estimates of CO2 flux (Table 6-79) were developed based on estimates of annual settlement
area and tree cover derived from developed land. 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. Developed land is likely a better proxy for tree cover in settlement areas
than urban areas as urban land areas were about 36 percent smaller than settlement areas in 2011.
Percent Tree Cover in Settlement Areas
Percent tree cover in settlement area 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
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 2018: used 2011 NLCD tree cover adjusted with 2016 photo-interpreted values
6-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 database74 and Forest Service urban forest inventory data
(e.g., Nowak et al. 2016, 2017) (Table 6-80). 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, 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-
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 6-80: 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
74 See .
Land Use, Land-Use Change, and Forestry 6-121

-------
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
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 Dakota15
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
6-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 6-81)
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
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 6-81). 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 6-81. 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 6-81: 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 (2018)
Gross Annual	Net Annual Tree Gross Annual Net Annual Net: Gross
State	Sequestration	Sequestration Cover Sequestration Sequestration	Annual
Land Use, Land-Use Change, and Forestry 6-123

-------




per Area of
Tree Cover
per Area of
Tree Cover
Sequestration
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
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




Note: Totals may not sum due to independent rounding.
6-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 6-82). 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 2018. The results of this quantitative uncertainty analysis are summarized in Table 6-82.
The change in C stocks in Settlement Trees in 2018 was estimated to be between -195.4 and -62.2 MMT CO2 Eq. at
a 95 percent confidence level. This analysis indicates a range of 51 percent more sequestration to 52 percent less
sequestration than the 2018 flux estimate of -129.8 MMT CO2 Eq.
Table 6-82: Approach 2 Quantitative Uncertainty Estimates for Net CO2 Flux from Changes
in C Stocks in Settlement Trees (MMT CO2 Eq. and Percent)
Source Gas
2018 Flux Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Flux Estimate
(MMT C02 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 2018. 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
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
In this 2018 assessment, the settlement area estimates have been updated with the latest NRI data through 2015
(projected to 2018). Due to this update, settlement area in 2017 increased from 43,118,102 ha (2017 report
estimate) to 44,799,282 ha (+ 3.9 percent). This area increase led to a 4.8 percent overall increase in the net
carbon sequestration estimate in 2017 (from 123.9 MMT CO2 Eq. to 129.8 MMT CO2 Eq.).
Land Use, Land-Use Change, and Forestry 6-125

-------
Planned Improvements
A consistent representation of the managed land base in the United States is discussed in Section 6.1
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 interpretation of settlement tree cover will be updated every few years to update tree cover
estimates and trends
b)	Areas for photo interpretation of settlement area tree cover will be updated as new NLCD developed land
information becomes available
c)	Overlap between forest and NLCD developed land (settlement area proxy) will be estimated based on
Forest Service Forest Inventory plot data
N20 Emissions from Settlement Soils (CRF Source Category
4E1|
Of the synthetic N fertilizers applied to soils in the United States, approximately 1.5 percent are currently applied
to lawns, golf courses, and other landscaping within settlement areas, and contributes to soil N2O 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 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 N2O.
N additions to soils result in direct and indirect N2O 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 N2O (i.e., ammonia [NH3] and nitrogen oxide [NOx] volatilization, nitrate
[NO3 ] leaching and runoff), and later converted into N2O at the off-site location. The indirect emissions are
assigned to settlements because the management activity leading to the emissions occurred in settlements.
Total N2O emissions from soils in Settlements Remaining Settlements75 are 2.4 MMT CO2 Eq. (8.1 kt of N2O) in
2018. There is an overall increase of 20 percent from 1990 to 2018 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 6-83.
Table 6-83: N2O Emissions from Soils in Settlements Remaining Settlements (MMT CO2 Eq.
and kt N2O)
75 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.
6-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
1990 2005 2014 2015 2016 2017 2018
MMTC02 Eq.
Direct N20 Emissions from Soils
1.6
2.5
1.9
1.8
1.9
2.0
2.0
Synthetic Fertilizers
0.8
1.6
0.9
0.8
0.9
1.0
1.0
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
0.4
0.3
0.3
0.4
0.4
Total
2.0
3.1
2.2
2.2
2.2
2.3
2.4
kt N20







Direct N20 Emissions from Soils
6
9
6
6
6
7
7
Synthetic Fertilizers
3
6
3
3
3
3
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
8
7
8
8
8
Note: Totals may not sum due to independent rounding.
Methodology
For settlement soils, the IPCC Tier 1 approach is used to estimate soil N2O emissions from synthetic N fertilizer,
biosolids additions, and drained organic soils. Estimates of direct N2O 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 2018 are based on 2012 values adjusted for annual total N fertilizer sales
in the United States because there is 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
applied to settlements is multiplied by the IPCC default emission factor (1 percent) to estimate direct N2O
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 N2O emissions (IPCC
2006) for 1990 to 2018.
The IPCC (2006) Tier 1 method is also used to estimate direct N2O 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 N2O 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 N2O 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 N2O off-site. The
resulting estimates are summed to obtain total indirect emissions from 1990 to 2015 for fertilizer and from 1990
to 2018 for biosolids.
A linear extrapolation of the trend in the time series is applied to estimate the direct and indirect N2O emissions
for fertilizer and drainage of organic soils from 2016 to 2018 because N fertilizer inputs and area data for these
Land Use, Land-Use Change, and Forestry 6-127

-------
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 2018 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 2018.
Uncertainty and Time-Series Consistency
The amount of N2O 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
N2O emissions is complex and highly uncertain. The IPCC default methodology does not explicitly incorporate any
of these variables, except variations in the total amount of fertilizer N and biosolids applications, 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.76 Uncertainty in the area of
drained organic soils is based on the estimated variance from the NRI survey (USDA-NRCS 2018). For 2016 to 2018,
there is also additional uncertainty associated with the fit of the linear regression ARMA 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
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 N2O emissions from fertilizer N and drainage of organic soils
based on a Monte Carlo analysis. The results are combined with the uncertainty in N2O emissions from the
biosolids application using simple error propagation methods (IPCC 2006). The results are summarized in Table
6-84. Direct N2O emissions from soils in Settlements Remaining Settlements in 2018 are estimated to be between
1.4 and 2.8 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 30 percent below to 38 percent
above the 2018 emission estimate of 2.0 MMT CO2 Eq. Indirect N2O emissions in 2018 are between 0.2 and 0.5
MMT CO2 Eq., ranging from 39 percent below to 39 percent above the estimate of 0.4 MMT CO2 Eq.
Table 6-84: Quantitative Uncertainty Estimates of N2O Emissions from Soils in Settlements
Remaining Settlements (MMT CO2 Eq. and Percent)
Source
Gas
2018 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
2.8
-30% 38%
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
76 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.
6-128 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
through 2018. 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
The spreadsheet containing fertilizer, drainage of organic soils, and biosolids applied to settlements and
calculations for N2O and uncertainty ranges have been checked. An error was found in the uncertainty calculation
that was corrected.
Recalculations Discussion
The entire time series was recalculated based on updates to the land representation data with the release of the
2018 NRI (USDA-NRCS 2018) and additional information from the NLCD (Yang et al. 2018; Fry et al. 2011; Homer et
al. 2007, 2015). The amount of fertilizer applied to settlements was also revised based on the USGS data product
with information about off-farm fertilizer application (Brakebill and Gronberg 2017). In addition, the data splicing
method has been used to re-estimate N2O emissions for 2016 and 2017 from the previous Inventory. These
recalculations led to a decrease in emissions of 0.27 MMT CO2 Eq., or 15 percent, on average across the time
series.
Planned Improvements
This source will be extended to include soil N2O 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 2018 in a future
Inventory.
Changes in Yard Trimmings and Food Scrap Carbon Stocks in
Landfills (CRF Category 4E1)
In the United States, yard trimmings (i.e., grass clippings, leaves, and branches) and food scraps (food waste from
residential, commercial, and institutional sources) 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.
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 and because food scraps are generated by settlements. While the majority of food scraps
originate from cropland and grassland, this Inventory has chosen to report these 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 therefore reporting these C stock
changes that occur entirely within landfills fits most appropriately within the Settlements Remaining Settlements
section.
Both the amount of yard trimmings collected annually and the fraction that is landfilled have declined over the last
decade. In 1990, over 58 million metric tons (wet weight) of yard trimmings and food scraps were generated (i.e.,
put at the curb for collection to be taken to disposal sites or to composting facilities) (EPA 2016). Since then,
programs banning or discouraging yard trimmings disposal in landfills have led to an increase in backyard
composting and the use of mulching mowers, and a consequent 1.4 percent decrease 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 31 percent in 2017 (EPA 2018). The net effect of the reduction in
Land Use, Land-Use Change, and Forestry 6-129

-------
generation and the increase in composting is a 57 percent decrease in the quantity of yard trimmings disposed of
in landfills since 1990.77
Food scrap generation has grown by 61 percent since 1990, and while the proportion of total food scraps
generated that are eventually discarded in landfills has decreased slightly, from 82 percent in 1990 to 76 percent in
2017, the tonnage disposed of in landfills has increased considerably (by 50 percent) due to the increase in food
scrap generation.78 Although the total tonnage of food scraps disposed of in landfills has increased from 1990 to
2017, the difference in the amount of food scraps added from one year to the next has generally decreased, and
consequently the annual carbon stock net changes from food scraps have generally decreased as well (as shown in
Table 6-85 and Table 6-86). As described in the Methodology section, the carbon stocks are modeled using data on
the amount of yard trimmings and food scraps landfilled since 1960. These materials decompose over time,
producing Cm and CO2. 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 yard trimmings and food scraps in a
single more recent year; and (2) is much larger than the carbon introduced from yard trimmings and 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 in landfill C storage from 24.5
MMT CO2 Eq. (6.7 MMT C) in 1990 to 12.0 MMT C02 Eq. (3.3 MMT C) in 2018 (Table 6-85 and Table 6-86).
Table 6-85: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills
(MMT COz Eq.)
Carbon Pool
1990
2005
2014
2015
2016
2017
2018
Yard Trimmings
(20.1)
(7.5)
(8.3)
(8.4)
(8.4)
(8.5)
(8.5)
Grass
(1.7)
(0.6)
(0.8)
(0.8)
(0.8)
(0.7)
(0.7)
Leaves
(8.7)
(3.4)
(3.8)
(3.9)
(3.9)
(4.0)
(4.0)
Branches
(9.8)
(3.4)
(3.7)
(3.7)
(3.7)
(3.8)
(3.8)
Food Scraps
(4.4)
(3.9)
(3.9)
(3.7)
(3.5)
(3.6)
(3.5)
Total Net Flux
(24.5)
(11.4)
(12.3)
(12.1)
(11.9)
(12.1)
(12.0)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
Table 6-86: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills
(MMT C)
Carbon Pool
1990
2005
2014
2015
2016
2017
2018
Yard Trimmings
(5.5)
(2.0)
(2.3)
(2.3)
(2.3)
(2.3)
(2.3)
Grass
(0.5)
(0.2)
) (0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Leaves
(2.4)
(0.9)
! (1-0)
(1.1)
(1.1)
(1.1)
(1.1)
Branches
(2.7)
(0.9)
! (1.0)
(1.0)
(1.0)
(1.0)
(1.0)
Food Scraps
(1.2)
(1.1)
§ (1-1)
(1.0)
(1.0)
(1.0)
(1.0)
Total Net Flux
(6.7)
(3.1)
(3.3)
(3.3)
(3.2)
(3.3)
(3.3)
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
77	Landfilled yard trimming amounts were not estimated for 2018; the values are estimated from 1990-2017.
78	Food scrap generation was not estimated for 2018; the values are estimated from 1990-2017.
6-130 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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, 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.
To determine the total landfilled C stocks for a given year, the following were estimated: (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 2015 (EPA 2018), which provides data for 1960,1970,1980,1990, 2000, 2005, 2010, 2014, and 2015.
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 2018), 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.
Due to the limited update this inventory year, the amount of yard trimming and food scraps for 2018 were not
estimated (2018 emissions were projected, as described later in this chapter). It is assumed that the proportion of
each individual material (food scraps, grass, leaves, branches) that is landfilled is the same as the proportion across
the overall waste stream, although the EPA (2018) report and historical data tables (EPA 2016) do not subdivide
the discards (i.e., total generated minus composted) of individual materials into amounts landfilled and combusted
(it provides a mass of overall waste stream discards managed in landfills79 and combustors with energy recovery).
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 (the EPA reports provide wet
weight data), 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 6-87).
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
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 6-87).
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
79 EPA (2018 and 2016) reports discards in two categories: "combustion with energy recovery" and "landfill, other disposal,"
which includes combustion without energy recovery. For years in which there is data from previous EPA reports on combustion
without energy recovery, EPA assumes these estimates are still applicable. For 2000 to present, EPA assumes that any
combustion of MSW that occurs includes energy recovery, so all discards to "landfill, other disposal" are assumed to go to
landfills.
Land Use, Land-Use Change, and Forestry 6-131

-------
time, resulting in emissions of CFU and CO2 (CFU 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 6-87.
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 CFU 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 2010 U.S. Census. Population data were broken into three
categories: less than 20 inches of rain per year, 20 to 40 inches of rain per year, and greater than 40 inches of rain
per year. To calculate the weighted national average for component-specific decay rates, the percentage of the
population within each precipitation category was multiplied by the component-specific decay rate for that
category, and then summed. The component-specific decay rates are shown in Table 6-87.
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 ICCix {[CSi x ICG[ + [(1 - (C.Sx ICG)) x e-^-")]}
n
where,
k
n
t
LFO,t
Win
MO
CSi
ICO
e
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),
Year in which the waste was disposed of (year, where 1960 
-------
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 Equation 2 as the change in
stock compared to the preceding year:
Ft= TLFCt- TLFCt- d
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 (518,000 metric tons) decomposes,
leaving a total of 617,000 metric tons (the persistent portion, plus the remainder of the degradable portion).
Continuing the example, by 2017, the total food scraps C originally disposed of in 1960 had declined to 178,900
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 2017), the total landfill C from food
scraps in 2017 was 45.3 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 2017, yielding a value of 275.5 million metric tons (as shown in
Table 6-88).80 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 6-86) is the difference in the landfill C
stock for that year and the stock in the next year. For example, the net change in 2017 shown in Table 6-86 (3.3
MMT C) is equal to the stock in 2017 (275.5 MMT C) minus the stock in 2018 (278.8 MMT C). The C stocks used in
the net change calculation are shown in Table 6-88.
Table 6-87: Moisture Contents, C Storage Factors (Proportions of Initial C Sequestered),
Initial C Contents, and Decay Rates for Yard Trimmings and Food Scraps in Landfills
Variable

Yard Trimmings

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 (year1)
0.313
0.179
0.015
0.151
Table 6-88: C Stocks in Yard Trimmings and Food Scraps in Landfills (MMT C)
Carbon Pool
1990
2005
2014
2015
2016
2017
2018
2019
Yard Trimmings
156.0
203.1
223.4
225.7
228.0
230.3
232.6
234.9
Grass
14.6
18.1
20.0
20.2
20.4
20.6
20.8
21.0
Leaves
66.7
87.3
96.6
97.7
98.7
99.8
100.9
102.0
Branches
74.7
97.7
106.8
107.8
108.9
109.9
110.9
111.9
Food Scraps
17.9
33.2
42.2
43.3
44.3
45.3
46.3
47.2
Total Carbon Stocks
173.9
236.3
265.7
269.0
272.3
275.5
278.8
282.1
Note: Totals may not sum due to independent rounding.
To develop the 2018 and 2019 C stock estimates, estimates of yard trimming and food scrap carbon stocks were
forecasted for 2018 and 2019, based on data from the 1990 through 2007 inventory. These forecasted values were
used to calculate net changes in carbon stocks for the previous year. Excel's FORECAST.ETS function was used to
predict a 2018 and 2019 value using historical data via an algorithm called "Exponential Triple Smoothing". This
method determined the overall trend and provided appropriate carbon stock estimates for 2018 and 2019.
80 Carbon stock mass and decomposition was not estimated for 2018; the values are only estimated from 1990 to 2017.
Land Use, Land-Use Change, and Forestry 6-133

-------
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 estimates of C storage in landfills 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 that was run on the 1990 to 2017 inventory was applied to
estimate the overall uncertainty of the C storage estimate for 2018. The results of the Approach 2 quantitative
uncertainty analysis are summarized in Table 6-89. Total yard trimmings and food scraps CO2 flux in 2018 was
estimated to be between -18.9 and -4.9 MMT CO2 Eq. at a 95 percent confidence level (or 19 of 20 Monte Carlo
stochastic simulations). This indicates a range of 57 percent below to 59 percent above the 2018 flux estimate of -
12.0 MMT CO2 Eq.
Table 6-89: Approach 2 Quantitative Uncertainty Estimates for CO2 Flux from Yard
Trimmings and Food Scraps in Landfills (MMT CO2 Eq. and Percent)

2018 Flux


Source Gas
Estimate
Uncertainty Range Relative to Flux Estimate3

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


Lower Upper
Lower Upper


Bound Bound
Bound Bound
Yard Trimmings and Food
CO2
Scraps
(12.0)
(18.9) (4.9)
-57% 59%
Note: Parentheses indicate negative values or net C storage.
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 2018. 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
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. No errors were
found.
Recalculations Discussion
A recent review of the total net flux methodology determined that the net flux was calculated incorrectly for this
category in the 1990 to 2017 Inventory. The net change for a specific year was calculated by subtracting the C
stock in the previous year from the C stock in the specific year. This calculation has been corrected, to calculate
the net change by subtracting the C stock in the next year from C stock in the specific year. The corrections
resulted in slight changes across the time series. The methodological approach now used is consistent with the
calculation of net C flux for forest ecosystems and harvested wood products in Chapter 6.2 of this Inventory.
6-134 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 Cm emissions described in the Waste chapter. For example, the Waste chapter does
not distinguish landfill Cm emissions from yard trimmings and food scraps separately from landfill Cm 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. However, because there are
no plans to separate out yard trimmings and food scraps when estimating landfill emissions in the Waste chapter
(Section 7.1) this evaluation may not be possible. In part, this is because the estimates in Section 7.1 are developed
using data from EPA's Greenhouse Gas Reporting Program for which only very few facilities break out these types
of waste (for more details on the landfills methodology see Section 7.1).
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 to take into account the fact that these items are relative to
population. 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.
EPA will also 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. In addition, based on comments received during the Public Review phase EPA will further
evaluate Equation 1 to determine if adjustments are needed to either the presentation of the equation and/or the
use of the equation in the inventory calculations.
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).81 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
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
81 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 6-135

-------
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 (SOC) stocks due
to land use change. All soil 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 2018, 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 2018 are 36.9, 7.2,
6.7, and 9.9 MMT CO2 Eq. (10.1, 2.0,1.8, and 2.7 MMT C). Mineral and organic soils also lost 16.2 and 2.4 MMT
CO2 Eq. in 2018 (4.4 and 0.6 MMT C). The total net flux is 79.3 MMT C02 Eq. in 2018 (21.6 MMT C), which is a 26
percent increase in CO2 emissions compared to the emissions in the initial reporting year of 1990. 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 6-90: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Settlements (MMT CO2 Eq.)

1990
2005
2014
2015
2016
2017
2018
Cropland Converted to







Settlements
3.4
9.8
6.7
6.2
6.0
6.0
5.9
Mineral Soils
2.8
8.4
5.8
5.3
5.2
5.2
5.2
Organic Soils
0.6
1.3
0.9
0.8
0.8
0.8
0.8
Forest Land Converted to







Settlements
54.6
59.9
62.9
63.0
62.9
62.9
62.9
Aboveground Live Biomass
32.5
35.1
36.8
36.9
36.9
36.9
36.9
Belowground Live Biomass
6.3
6.8
7.1
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
9.8
9.9
9.9
9.9
9.9
Mineral Soils
1.1
2.0
2.1
2.0
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
12.7
11.9
11.3
11.3
11.3
Mineral Soils
4.6
14.9
11.7
11.0
10.4
10.4
10.4
Organic Soils
0.6
1.4
1.0
0.9
0.9
0.9
0.9
Other Lands Converted to







Settlements
(0.4)
(1.4)
(1.3)
(1.2)
(1.2)
(1.2)
(1.2)
Mineral Soils
(0.4)
(1.6)
(1.5)
(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
36.8
36.9
36.9
36.9
36.9
Total Belowground Biomass Flux
6.3
6.8
7.1
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
9.8
9.9
9.9
9.9
9.9
Total Mineral Soil Flux
8.1
23.8
18.2
17.0
16.3
16.2
16.2
Total Organic Soil Flux
1.4
3.6
2.7
2.5
2.4
2.4
2.4
Total Net Flux
62.9
85.0
81.4
80.1
79.4
79.3
79.3
+ Does not exceed 0.05 MMT C02 Eq.
Note: Totals may not sum due to independent rounding.
6-136 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 6-91: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Settlements (MMT C)

1990
2005
2014
2015
2016
2017
2018
Cropland Converted to







Settlements
0.9
2.7
1.8
1.7
1.6
1.6
1.6
Mineral Soils
0.8
2.3
1.6
1.5
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
17.1
17.2
17.1
17.1
17.1
Aboveground Live Biomass
8.9
9.6
10.0
10.1
10.1
10.1
10.1
Belowground Live Biomass
1.7
1.9
1.9
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
0.6
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
3.5
3.2
3.1
3.1
3.1
Mineral Soils
1.3
4.1
3.2
3.0
2.8
2.8
2.8
Organic Soils
0.2
0.4
0.3
0.3
0.2
0.2
0.2
Other Lands Converted to







Settlements
(0.1)
(0.4)
(0.4)
(0.3)
(0.3)
(0.3)
(0.3)
Mineral Soils
(0.1)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted to







Settlements
+
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
Total Aboveground Biomass Flux
8.9
9.6
10.0
10.1
10.1
10.1
10.1
Total Belowground Biomass Flux
1.7
1.9
1.9
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
5.0
4.6
4.4
4.4
4.4
Total Organic Soil Flux
0.4
1.0
0.7
0.7
0.7
0.7
0.6
Total Net Flux
17.1
23.2
22.2
21.9
21.6
21.6
21.6
+ Does not exceed 0.05 MMT C.
Note: Totals may not sum due to independent rounding.
Methodology
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 mineral and organic soil 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
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 2018),
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
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
Land Use, Land-Use Change, and Forestry 6-137

-------
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 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 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. U.S.-specific C stock change factors are derived from
published literature to determine the impact of management practices on SOC 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 C with conversion to settlements under the assumption that there are
additional soil C losses with land clearing, excavation and other activities associated with 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 C stock changes from 2016 to 2018
because NRI activity data are not available for these years. Specifically, a linear regression model with
6-138 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
2018. The Tier 2 method described previously will be applied to recalculate the 2016 to 2018 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 CO2
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 C
stocks changes, a linear extrapolation of the trend in the time series is applied to estimate the emissions from 2016
to 2018 because NRI activity data are not available for these years to determine the area of Land Converted to
Settlements.
Uncertainty and Time-Seri insistency
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 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 6-92 for each subsource (i.e., biomass C, dead wood, litter, mineral
soil C and organic soil C) 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 C stock changes from 2016
to 2018. The combined uncertainty for total C stocks in Land Converted to Settlements ranges from 33 percent
below to 33 percent above the 2018 stock change estimate of 79.3 MMT CO2 Eq.
Table 6-92: 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)
2018 Flux Estimate Uncertainty Range Relative to Flux Estimate3
Source	(MMT CP2 Eq.)	(MMT CP2 Eq.)	(%)


Lower
Upper
Lower
Upper


Bound
Bound
Bound
Bound
Cropland Converted to Settlements
5.9
2.6
9.3
-56%
56%
Mineral Soil C Stocks
5.2
1.9
8.4
-63%
63%
Organic Soil C Stocks
0.8
0.2
1.4
-76%
76%
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%
Litter
9.9
3.7
16.0
-62%
62%
Mineral Soil C Stocks
1.9
1.4
2.4
-27%
27%
Organic Soil C Stocks
0.3
0.1
0.5
-68%
68%
Grassland Converted to Settlements
11.3
7.2
15.3
-36%
36%
Mineral Soil C Stocks
10.4
6.4
14.4
-38%
38%
Land Use, Land-Use Change, and Forestry 6-139

-------
Organic Soil C Stocks
0.9
0.2
1.6
-80%
80%
Other Lands Converted to Settlements
(1.2)
(1.8)
(0.5)
-56%
56%
Mineral Soil C Stocks
(1.3)
(1.9)
(0.7)
-49%
49%
Organic Soil C Stocks
0.1
0.1
0.3
-152%
152%
Wetlands Converted to Settlements
0.4
0.1
0.8
-83%
133%
Mineral Soil C Stocks
0.1
+
0.1
-87%
87%
Organic Soil C Stocks
0.3
+
0.8
100%
161%
Total: Land Converted to Settlements
79.3
53.0
105.7
-33%
33%
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
11.0
21.4
-32%
16%
Organic Soil C Stocks
2.4
(6.0)
10.7
-351%
352%
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 2018. 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. These checks uncovered errors in the calculation of uncertainty for mineral soils
that were corrected. There was also an error in handling of activity data for this source category in which
settlement areas were only included if they had been in agriculture during the past. This led to an under-
estimation of drained organic soils in settlements that has been corrected in this Inventory.
Recalculations Discussion
The entire time series for mineral and organic soils was recalculated based on updates to the land representation
data with the release of the 2018 NRI (USDA-NRCS 2018) and additional information from the NLCD (Yang et al.
2018; Fry et al. 2011; Homer et al. 2007, 2015), as well as the data splicing method that was applied to re-estimate
CO2 emissions from mineral and organic soils for 2016 to 2017. In addition, the entire time series was updated with
recalculated biomass and dead organic matter losses for Forest Land Converted to Settlements. The time series was
also corrected based on the quality control problem that led to an under-estimation of drained organic soils in
settlements. The recalculations led to a decrease in emissions of 1.8 MMT CO2 Eq., or 1.8 percent, on average
across the time series.
Planned Improvements
A planned improvement for the Land Converted to Settlements category is to develop an inventory of mineral soil
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 6-93 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 with Forest
Land Converted to Settlements (i.e., currently assume that all biomass is removed during conversion). There are
also plans to extend the Inventory to include C losses associated with drained organic soils in settlements occurring
6-140 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
Land Use, Land-Use Change, and Forestry 6-141

-------
Table 6-93: Area of Managed Land in Settlements Remaining Settlements that is not
included in the current Inventory (Thousand Hectares)

Area (Thousand Hectares)

Year
LCS Managed Land
Area (Section 6.1)
LCS Area
Included in
Inventory
LCS Area Not
Included in
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
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
6-142 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Remaining Other Land is known (see Table 6-7), 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 CO2, Cm or N2O fluxes on Other Land Remaining Other Land at this time.
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 6-7), 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 CO2, Cm or N2O fluxes on Land Converted to
Other Land from fluxes on Other Land Remaining Other Land at this time.
Land Use, Land-Use Change, and Forestry 6-143

-------
7. Waste
Waste management and treatment activities are sources of greenhouse gas emissions (see Figure 7-1). Landfills
accounted for approximately 17.4 percent of total U.S. anthropogenic methane (Cm) emissions in 2018, the third
largest contribution of any Cm source in the United States. Additionally, wastewater treatment and composting of
organic waste accounted for approximately 2.2 percent and 0.4 percent of U.S. Cm emissions, respectively. Nitrous
oxide (N2O) emissions from the discharge of wastewater treatment effluents into aquatic environments were
estimated, as were N2O emissions from the treatment process itself. Nitrous oxide emissions from composting
were also estimated. Together, these waste activities account for 1.7 percent of total U.S. N2O emissions. Nitrogen
oxides (NOx), carbon monoxide (CO), and non-Cm 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: 2018 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
Landfills
Wastewater
Treatment
Composting
waste as a Portion of All Emissions
111
Energy
Agriculture
IPPU
Waste
50 60 70
MMT CO2 Eq.
110
Overall, in 2018, waste activities generated emissions of 134.4 MMT CO2 Eq., or 2.0 percent of total U.S.
greenhouse gas emissions.1
Emissions reported in the Waste chapter for landfills and wastewater treatment include those from all 50 states, including
Hawaii and Alaska, as well as from U.S. Territories to the extent those waste management activities are occurring. 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.
Waste 7-1

-------
Table 7-1: Emissions from Waste (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
2018
ch4
195.3
148.6
129.0
128.0
124.7
124.3
127.2
Landfills
179.6
131.3
112.6
111.3
108.0
107.7
110.6
Wastewater Treatment
15.3
15.4
14.3
14.6
14.4
14.1
14.2
Composting
0.4
1.9
2.1
2.1
2.3
2.4
2.5
n2o
3.7
6.1
6.6
6.7
6.9
7.2
7.2
Wastewater Treatment
3.4
4.4
4.8
4.8
4.9
5.0
5.0
Composting
0.3
1.7
1.9
1.9
2.0
2.2
2.2
Total
199.0
154.7
135.6
134.7
131.6
131.4
134.4
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
7,811
5,945
5,160
5,120
4,988
4,971
5,089
Landfills
7,182
5,253
4,503
4,452
4,322
4,308
4,422
Wastewater Treatment
614
618
573
583
575
566
569
Composting
15
75
84
85
91
98
98
n2o
12
20
22
22
23
24
24
Wastewater Treatment
11
15
16
16
16
17
17
Composting
1
6
6
6
7
7
7
Note: Totals may not sum due to independent rounding.
Carbon dioxide (CO2), Cm, and N2O 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 2018 resulted in 11.4 MMT CO2 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 7.4.
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 2017) to ensure that the trend
is accurate. Revisions to Wastewater Treatment included updated population data, revised pulp and paper
wastewater generation data, and methodology updates for estimating ethanol production resulting in 0.25 percent
increase from the previous inventory. For more information on specific methodological updates, please see the
Recalculations for each category, in this chapter.
7-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 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 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 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 CO2 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 CO2 Eq. per year.
Waste Data from EPA's Greenhouse Gas Reporting Program
EPA's Greenhouse Gas Reporting Program (GHGRP)2 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 the Landfills category facility level data to compile the national estimate of
emissions from Municipal Solid Waste (MSW) landfills (see section 7.1). EPA uses directly reported GHGRP data
for net CH4 emissions from MSW landfills for the years 2010 to 2018 of the Inventory. MSW landfills subject to
the GHGRP began collecting data in 2010. This data is also used to recalculate emissions from MSW landfills for
the years 2005 to 2009 to ensure time series consistency. See Annex 9 for more information on use of EPA's
GHGRP in the Inventory.
2 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).
Waste 7-3

-------
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
commonly 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 2016).
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 (Cm)
producing anaerobic bacteria convert the fermentation products into stabilized organic materials and biogas
consisting of approximately 50 percent biogenic carbon dioxide (CO2) and 50 percent CFU, 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
Modern, managed landfills are well-engineered facilities that are located, designed, operated, and monitored to
ensure compliance with federal, state, and tribal regulations. 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
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. Additionally, state and
7-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
tribal requirements may exist.3
Methane and CO2 are the primary constituents of landfill gas generation and emissions. However, the 2006IPCC
Guidelines set an international convention to not report biogenic CO2 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 (N2O) 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 N2O emissions. Furthermore, the 2006
IPCC Guidelines did not include a methodology for estimating N2O 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
Cm 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 commonly 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 2018, landfill CH4 emissions were approximately 110.6 MMT CO2 Eq. (4,422 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 95 percent of total landfill emissions (95.6 MMT CO2 Eq.), while
industrial waste landfills accounted for the remainder (15.0 CO2 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
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). With regard to 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
3 For more information regarding federal MSW landfill regulations, see
.
Waste 7-5

-------
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 8.8 percent to 212 MMT in
2018 (see Annex 3.14, Table A-236). 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 Cm 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 10.1 MMT in 2018 (see Annex
3.14, Table A-236). Cm emissions from industrial waste landfills have also remained at similar levels recently,
ranging from 14.3 MMT CO2 Eq. in 2005 to 15.0 MMT CO2 Eq. in 2018 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 2019, LMOP identified 22 new landfill gas-to-
energy (LFGE) projects (EPA 2019a) 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 Cm generation from the amount of organic MSW landfilled as the U.S. population grows (EPA 2019b).
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
2014
2015
2016
2017
2018
MSW CH4 Generation
205.3
-
-
-
-
-
-
Industrial CH4 Generation
12.1
15.9
16.6
16.6
16.6
16.6
16.7
MSW CH4 Recovered
(17.9)
-
-
-
-
-
-
MSW CH4 Oxidized
(18.7)
-
-
-
-
-
-
Industrial CH4 Oxidized
(1.2)
(1.6)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
MSW net CH4 Emissions







(GHGRP)
-
117.0
97.7
96.4
93.1
92.7
95.6
Industrial CH4 Emissions3
10.9
14.3
14.9
14.9
14.9
15.0
15.0
Total
179.6
131.3
112.6
111.3
108.0
107.7
110.6
a 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 2018 (EPA 2019b).
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 2018, 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.
7-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Table 7-4: ChU Emissions from Landfills (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
MSW CH4 Generation
8,214
-
-
-
-
-
-
Industrial CH4 Generation
484
636
662
663
664
665
666
MSW CH4 Recovered
(718)
-
-
-
-
-
-
MSW CH4 Oxidized
(750)
-
-
-
-
-
-
Industrial CH4 Oxidized
(48)
(64)
(66)
(66)
(66)
(67)
(67)
MSW net CH4 Emissions







(GHGRP)
-
4,681
3,907
3,855
3,724
3,709
3,823
Industrial CH4 Emissions3
436
572
596
597
598
599
599
Total
7,182
5,253
4,503
4,452
4,322
4,308
4,422
a 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 2018 (EPA 2019b).
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 2018, 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
Methane emissions from landfills can be estimated using two primary methods. The first method uses the first
order decay (FOD) model as described by the 2006IPCC 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 CO2 as it passes
through the landfill cover (e.g., soil, clay, geomembrane). This method is presented below and is similar to
Equation HH-5 in 40 CFR Part 98.343 for MSW landfills, and Equation TT-6 in 40 CFR Part 98.463 for industrial
waste landfills.
CH4,Solid Waste = [CH4.MSW + CH4,Ind — R] — Ox
where,
CH4 ,solid waste	— Net CH4 emissions from solid waste
CH4,msw	= CH4 generation from MSW landfills
CH4 ,ind	— CH4 generation from industrial waste landfills
R	= CH4 recovered and combusted (only for MSW landfills)
Ox	= CH4 oxidized from MSW and industrial waste landfills before release to the atmosphere
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 CH4in 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
Waste 7-7

-------
not captured by the collection system; this amount is then adjusted for oxidation. The second portion of the
equation adjusts the portion of CI-I4 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 = [(-	-	R) X (1 ~ OX) + R X (l - (DE X fDest))\
\CE X f REC '
where,
CH4 ,soiid 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)
fkEc	= 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 7-2 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,
Figure 7-2: Methodologies Used Across the Time Series to Compile the U.S. Inventory of
Emission Estimates for MSW Landfills
-o
O
0)

-------
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 Cm generation, estimates for those years
were included in the FOD model for completeness in accounting for Cm 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
is disposed in managed, anaerobic landfills. The FOD method was applied to estimate annual Cm
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.
•	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 CFUgeneration. Landfill-specific CFU recovery amounts were then
subtracted from CFU generation and the result was adjusted with a 10 percent oxidation factor to derive
the net emissions estimates.
•	2005 through 2009: Emissions for these years are estimated using net Cm 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 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 C^flux rate. The average oxidation
factor from the GHGRP facilities is 19.5 percent (from reporting years 2011 to 2017).
•	2010 through 2018: 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 discussion of the data sources and methodology used to calculate CH4 generation and recovery is
provided below. Supporting information, including details on the technique used to ensure time-series consistency
Waste 7-9

-------
by incorporating the directly reported GHGRP emissions is presented in the Time-Series Consistency section of this
chapter and in Annex 3.14.
Methodology Applied for Industrial Waste Landfills
Emissions from industrial waste landfills are estimated from industrial production data (ERG 2019), waste disposal
factors, and the FOD method. There are currently no 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 169 facilities that reported to Subpart TT of the
GHGRP in 2018, 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. To validate this assumption, EPA recently conducted
an analysis of data reported to Subpart TT of the GHGRP in the 2016 reporting year. Waste streams of facilities
reporting to Subpart TT were designated as either relating to food and beverage, pulp and paper, or other based
on their primary NAICS code. The total waste disposed by facilities under each primary NAICS reported in 2016
were calculated in order to determine that 93 percent of the total organic waste quantity reported under Subpart
TT is originating from either the pulp and paper or food and beverage sector (RTI 2018b). Although this memo
concluded that Subpart TT data reported to the GHGRP are able to confirm the Inventory methodological
assumption that most organic waste placed in industrial waste landfills is from pulp and paper or food processing
facilities, EPA is currently unable to use these net emissions directly reported to the GHGRP for industrial landfills.
While Subpart TT waste disposal information for pulp and paper facilities correlates well with the production data
currently used to estimate Inventory emissions, the same cannot be said for food and beverage facilities. Waste
disposal data prior to 1990 does not correlate well between the two data sources, and no waste disposal data are
reported for these facilities through Subpart TT of the GHGRP prior to 1960. GHGRP data for food and beverage
facilities in the 1960s are an order of magnitude smaller than production data currently used to estimate emissions
for this sector in the Inventory. Because of these discrepancies, EPA is maintaining its current approach to
estimating emissions from industrial landfills using production data from the pulp and paper and food and
beverage sectors.
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 169 facilities, or 1 percent of facilities, have active
gas collection systems (EPA 2019b). 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.
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.
7-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Box 7-3: Nationwide Municipal Solid Waste Data Sources
Municipal solid waste generated in the United States can be managed through landfilling, recycling, composting,
and combustion with energy recovery. There have been three main sources for nationwide solid waste
management data in the United States:
•	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, now 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. 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 is not available, the survey asked for total tons landfilled. The data are adjusted for
imports and exports across state lines so that the principles of mass balance are adhered to, whereby the
amount of waste managed does 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, 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 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, composted, or combusted is assumed to be landfilled. The data
presented in the report are nationwide totals.
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. 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.
Uncertainty and Time-Series Consistency
Several types of uncertainty are associated with the estimates of CFU 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 Cm generation
potential (L0) and the rate of decay that produces Cmfrom MSW, as determined from several studies of Cm
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 Cm
generation at some landfills if used at the facility-level, the result is expected to balance out because it is being
applied nationwide.
Waste 7-11

-------
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 Cm 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 factor applied for years 2005 to 2009 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 still exist and assumptions were made for
many landfills in order to estimate the scale-up factor. Additionally, a simple methodology was used to back-cast
emissions for 2005 to 2009 using the GHGRP-reported emissions from 2010 to 2018. 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 emissions for 2005 to 2009.
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. Therefore, EPA decided to back-cast
the GHGRP emissions from 2009 to 2005 only, in order 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
7-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 Cm 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 Cm 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 Cm 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 Cm gas recovery; however, the number of unique landfills between
the four databases does differ.
The 2006IPCC 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 Cm 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
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 2006 IPCC 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)


2018 Emission




Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3


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




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Total Landfills
ch4
110.6
85.0
135.0
-23%
+22%
MSW
ch4
95.6
71.8
119.6
-25%
+25%
Industrial
ch4
15.0
10.3
18.8
-31%
+25%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Waste 7-13

-------
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 2006IPCC 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 Cm 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
are accurate, complete, and consistent.4 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 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.
4 See .
7-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 Cm
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
annual GHG reports resubmitted for 2010 to 2017 slightly increased or decreased total Subpart HH reported net
emissions by +/-0.1 percent or less in the years the Subpart HH data are applied (i.e., 2005 to 2017). These changes
resulted in changes to the net Inventory emissions by +/-0.1 percent. An increase in net Subpart HH reported
emissions resulted in an increase in the Inventory emissions for that year, and vice versa. For example, in 2017, the
changes in net Subpart HH reported emissions decreased by 0.04 MMT CO2 Eq. from the previous Inventory,
which resulted in a net decrease in landfill emissions in this year's Inventory by -0.04 percent.
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 CO2
equivalent for 3 consecutive years or less than 25,000 metric tons of CO2 equivalent for 5 consecutive years). If
warranted, EPA will revise the scale-up factor to reflect newly acquired information to ensure completeness of the
Inventory.
EPA has received comments from industry stakeholders requesting that the default oxidation factor of 10 percent
applied in the 1990 to 2004 time series be updated to a higher value to correspond with findings in recent
literature and facility-specific methane flux-derived oxidation factors from the GHGRP. Upon consideration of
available data, EPA has decided not to revise the oxidation factor applied in the 1990 to 2004 time series on the
basis that emissions estimates from the earlier part of the time series are not being used to inform policy. EPA has
increased the oxidation factor applied in the latter half of the time series by incorporating the GHGRP data and will
focus available resources on planned improvements that directly impact and improve the accuracy,
comprehensiveness, and completeness of net emissions from 2005 and later. In the next (1990 to 2019) Inventory
cycle, EPA will also begin investigating the prevalence of food-related waste deposited into industrial waste
landfills. EPA will record the findings from this exercise in a memorandum and if any changes to the methodology
or assumptions for industrial waste landfills are warranted, EPA will implement the changes during the following
Inventory cycle.
Additionally, with the recent publication of the 2019 Refinement to the 2006IPCC Guidelines for National
Greenhouse Gas Inventories (2019 Refinement), EPA will begin to review and update applicable emission factors,
Waste 7-15

-------
methodologies, and assumptions underlying emission estimates for landfills and make any applicable changes
during the next (1990 to 2019) Inventory cycle per the 2019 Refinement.
Box 7-4: Overview of U.S. Solid Waste Management Trends
As shown in Figure 7-3 and Figure 7-4, 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 7-3: Management of Municipal Solid Waste in the United States, 2017
Recycled
25.1%
Landfilled
52.1%
Composted
10.1%
MSW to
WTE
12.7%
Note: 2017 is the latest year of available data.
Source: EPA (2019c)
Figure 7-4: MSW Management Trends from 1990 to 2017
160
140
120
100
JCsl
Note: 2017 is the latest year of available data.
Source: EPA (2019c).
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
7-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 in 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 Cm 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 CFU generation potential and Cm 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 7-5). As shown in Figure 7-5, 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 7-5 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 Discarded3 in the Municipal Waste Stream by Waste Type from 1990
to 2017 (Percent)b
Waste Type
1990

2005

2014
2015
2016
2017
Paper and








Paperboard
30.0%

24.7%

14.3%
13.3%
12.7%
13.1%
Glass
6.0%

5.8%

5.2%
5.0%
4.9%
4.9%
Metals
7.2%

7.9%

9.5%
9.5%
9.8%
9.9%
Plastics
9.5%

16.4%

18.5%
18.9%
18.9%
19.2%
Rubber and Leather
3.2%

2.9%

3.0%
3.3%
3.4%
3.5%
Textiles
2.9%

5.3%

7.3%
7.7%
8.0%
8.0%
Wood
6.9%

7.5%

8.1%
8.0%
8.8%
8.7%
Otherc
1.4%

1.8%

2.2%
2.2%
2.2%
2.2%
Food Scraps
13.6%

18.5%

21.7%
22.0%
22.1%
22.0%
Yard Trimmings
17.6%

7.0%

7.9%
7.8%
6.9%
6.2%
Miscellaneous








Inorganic Wastes
1.7%

2.2%

2.3%
2.3%
2.3%
2.3%
a Discards after materials and compost recovery. In this table, discards include combustion with energy
recovery. Does not include construction & demolition debris, industrial process wastes, or certain
other wastes.
b Data for all years are from the EPA's Advancing Sustainable Materials Management: Facts and Figures
2016 and 2017 Tables and Figures report (Table 4) published in November 2019 (EPA 2019c).
c Includes electrolytes in batteries and fluff pulp, feces, and urine in disposable diapers. Details may not
add to totals due to rounding.
Note: 2017 is the latest year of available data.
Waste 7-17

-------
Figure 7-5: Percent of Degradable Materials Diverted from Landfills from 1990 to 2017
(Percent)

90%
80%
Paper and Paperboard
— — —Food Scraps


70%
¦Yard Trimmings /


60%



50%
>	


40%



30%
/


20%



10% -
no/
_____


U/o
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

Source: (EPA 2019c). Note: 2017 is the latest year of available data.

7.2 Wastewater Treatment (CRF Source
Category 5D)
Wastewater treatment processes can produce anthropogenic methane (Cm) and nitrous oxide (N2O) emissions.
Wastewater from domestic and industrial sources is treated to remove soluble organic matter, suspended solids,
pathogenic organisms, and chemical contaminants.5 Treatment may either occur on site, most commonly through
septic systems or package plants, or off site at centralized treatment systems. In the United States, approximately
19 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). Centralized wastewater treatment systems may include a variety
of processes, ranging from physical separation of material that readily settles out, to treatment operations that use
biological processes to convert and remove contaminants, to advanced treatment for removal of targeted
pollutants, such as nutrients. Some wastewater may also be treated through the use of constructed (or semi-
natural) wetland systems, though it is much less common in the United States (ERG 2016). Constructed wetlands
may be used as the primary method of wastewater treatment, or as a later 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).
Soluble organic matter is generally removed using biological processes in which microorganisms consume the
organic matter for maintenance and growth. The resulting biomass (sludge) is removed from the effluent prior to
discharge to the receiving stream. Microorganisms can biodegrade soluble organic material in wastewater under
aerobic or anaerobic conditions, where the latter condition produces CH4. During collection and treatment,
wastewater may be accidentally or deliberately managed under anaerobic conditions. In addition, the sludge may
be further biodegraded under aerobic or anaerobic conditions. The generation of N2O may also result from the
5 Throughout the Inventory, emissions from domestic wastewater also include any commercial and industrial wastewater
collected and co-treated with domestic wastewater.
7-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
treatment of domestic wastewater during both nitrification and denitrification of the nitrogen (N) present, usually
in the form of urea, ammonia, and proteins. These compounds are converted to nitrate (NO3) through the aerobic
process of nitrification. Denitrification occurs under anoxic conditions (without free oxygen) and involves the
biological conversion of nitrate into dinitrogen gas (N2). Nitrous oxide can be an intermediate product of both
processes but has typically been associated with denitrification. More recent research suggests that higher
emissions of N2O may in fact originate from nitrification (Ahn et al. 2010), while other research suggests that N2O
may also result from other types of wastewater treatment operations (Chandran 2012).
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 Cm 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 BODs. Because BOD is an aerobic parameter, it is preferable to use COD to
estimate Cm production, since CH4 is produced only in anaerobic conditions. The principal factor in determining
the N2O 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 N2O generation potential.
In 2018, Cm emissions from domestic wastewater treatment were 8.4 MMT CO2 Eq. (334 kt CH4). Emissions
remained fairly steady from 1990 through 1999 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 2004; U.S. Census Bureau 2017). In 2018, CH4
emissions from industrial wastewater treatment were estimated to be 5.9 MMT CO2 Eq. (235 kt CH4). Industrial
emission sources have generally increased across the time series through 1999 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. Table 7-7 and Table 7-8 provide CH4 emission estimates from domestic and industrial
wastewater treatment.
With respect to N2O, the United States identifies two distinct sources for N2O emissions from domestic
wastewater: emissions from centralized wastewater treatment processes, and emissions from effluent from
centralized treatment systems that has been discharged into aquatic environments. The 2018 emissions of N2O
from centralized wastewater treatment processes and from effluent were estimated to be 0.4 MMT CO2 Eq. (1.2 kt
N2O) and 4.6 MMT CO2 Eq. (15.6 kt N2O), respectively. Total N2O emissions from domestic wastewater were
estimated to be 5.0 MMT CO2 Eq. (16.8 kt N2O). Nitrous oxide emissions from wastewater treatment processes
gradually increased across the time series as a result of increasing U.S. population and protein consumption.
Nitrous oxide emissions are not estimated from industrial wastewater treatment because there is no IPCC
methodology provided or industrial wastewater emission factors available. Table 7-7 and Table 7-8 provide N2O
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
2014
2015
2016
2017
2018
ch4
15.3
15.4
14.3
14.6
14.4
14.1
14.2
Domestic
10.4
10.0
8.9
9.0
8.7
8.3
8.4
Industrial3
4.9
5.5
5.4
5.5
5.7
5.8
5.9
n2o
3.4
4.4
4.8
4.8
4.9
5.0
5.0
Centralized WWTP
0.2
0.3
0.3
0.3
0.4
0.4
0.4
Waste 7-19

-------
Domestic Effluent 3.2	4.1 ; 4.4 4.4 4.5 4.6 4.6
Total	18^7	19J5	19.1 19.3 19.2 19.1 19.2
Note: Totals may not sum due to independent rounding.
a Industrial activity includes the pulp and paper manufacturing, meat and poultry processing,
fruit and vegetable processing, starch-based ethanol production, petroleum refining, and
breweries industries.
Table 7-8: ChU and N2O Emissions from Domestic and Industrial Wastewater Treatment (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
ch4
614
618
573
583
575
566
569
Domestic
417
398
356
361
348
334
334
Industrial3
197
219
217
221
227
232
235
n2o
11
15
16
16
16
17
17
Centralized WWTP
1
1
1
1
1
1
1
Domestic Effluent
11
14
15
15
15
15
16
Note: Totals may not sum due to independent rounding.
a Industrial activity includes the pulp and paper manufacturing, meat and poultry processing, fruit
and vegetable processing, starch-based ethanol production, petroleum refining, and breweries
industries.
Methodology
Domestic Wastewater CH4 Emission Estimates
Domestic wastewater Cm emissions originate from both septic systems and from centralized treatment systems,
such as publicly owned treatment works (POTWs). Within these centralized systems, CH4 emissions can arise from
aerobic systems that are not well managed 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 anaerobic digesters when the captured biogas is not completely combusted. The methodological equations
are:
Emissions from Septic Systems = A
= USpop X (% onsite) X (EFseptic) X 1/109 X 365.25
Emissions from Centrally Treated Aerobic Systems (other than Constructed Wetlands) + Emissions from
Centrally Treated Aerobic Systems (Constructed Wetlands Only) + Emissions from Centrally Treated Aerobic
Systems (Constructed Wetlands used as Tertiary Treatment) = B
where,
Emissions from Centrally Treated Aerobic Systems (other than Constructed Wetlands)
= {[(% collected) x (total BODs produced) x (% aerobicoTcw) x (% aerobic w/out primary)] + [(%
collected) x (total BODs produced) x (% aerobicoTcw) x (% aerobic w/primary) x (1-% BOD removed in
prim, treat.)]} x (% operations not well managed) x (B0) x (MCF-aerobic_not_well_man)
Emissions from Centrally Treated Aerobic Systems (Constructed Wetlands Only)
= [(% collected) x (total BODs produced) x (%aerobiccw)] x (B0) x (MCF-constructed wetlands)
Emissions from Centrally Treated Aerobic Systems (Constructed Wetlands used as Tertiary Treatment)
= [(POTW_flow_CW) x (BODcwjnf) x 3.79 x (B0) x (MCF-constructed wetlands)] x 1/106 x 365.25
Emissions from Centrally Treated Anaerobic Systems = C
7-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
= {[(% collected) x (total BODs produced) x (% anaerobic) x (% anaerobic w/out primary)] + [(%
collected) x (total BODs produced) x (% anaerobic) x (% anaerobic w/primary) x (1-% BOD removed in
prim, treat.)]} x (B0) x (MCF-anaerobic)
Emissions from Anaerobic Digesters = D
= [(POTW_flow_AD) x (digester gas)/(100)] x 0.0283 x (FRAC_CH4) x 365.25 x (662) x (1-DE) x 1/109
Total Domestic CH4 Emissions from Wastewater (kt) = A+ B + C+ D
where,
USpop
% onsite
% collected
% aerobicoTcw
% aerobiccw
% anaerobic
% aerobic w/out primary
% aerobic w/primary
% BOD removed in prim, treat.
% operations not well managed
% anaerobic w/out primary
% anaerobic w/primary
EFseptic
Total BODs produced
BODcw.inf
Bo
1/106
365.25
3.79
MCF-aerobic_not_well_man.
MCF-anaerobic
MCF-constructed wetlands
DE
P OTW_f I o w_C W
POTW_flow_AD
digester gas
100
0.0283
FRAC_CH4
662
1/109
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) and an emission factor (10.7 g CHVcapita/day) (Leverenz
et al. 2010), 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 2019) and include the populations of the United States,
= U.S. population
= Flow to septic systems / total flow
= Flow to POTWs / total flow
= Flow to aerobic systems, other than wetlands only / total flow to
POTWs
= Flow to aerobic systems, constructed wetlands used as sole treatment
/ total flow to POTWs
= Flow to anaerobic systems / total flow to POTWs
= Percent of aerobic systems that do not employ primary treatment
= Percent of aerobic systems that employ primary treatment
= Percent of BOD removed in primary treatment
= Percent of aerobic systems that are not well managed and in which
some anaerobic degradation occurs
= Percent of anaerobic systems that do not employ primary treatment
= Percent of anaerobic systems that employ primary treatment
= Methane emission factor - septic systems
= kg BOD/capita/day x U.S. population x 365.25 days/yr
= BOD concentration in wastewater entering the constructed wetland
= Maximum CFU-producing capacity for domestic wastewater
= Conversion factor, kg to kt
= Days in a year
= Conversion factor, gallons to liters
= Cm correction factor for aerobic systems that are not well managed
= CH4 correction factor for anaerobic systems
= CH4 correction factor for surface flow constructed wetlands
= CH4 destruction efficiency from flaring or burning in engine
= Wastewater flow to POTWs that use constructed wetlands as tertiary
treatment (MGD)
= Wastewater influent flow to POTWs that have anaerobic digesters
(MGD)
= Cubic feet of digester gas produced per person per day
= Wastewater flow to POTW (gallons/person/day)
= Conversion factor, ft3 to m3
= Proportion of CH4 in biogas
= Density of CH4 (g CFU/m3 CH4)
= Conversion factor, g to kt
Waste 7-21

-------
American Samoa, Guam, Northern Mariana Islands, Puerto Rico, and the U.S. Virgin Islands. Table 7-9 presents U.S.
population for 1990 through 2018.
Emissions from Centrally Treated Aerobic and Anaerobic Systems:
Methane emissions from POTWs were estimated by multiplying the total BODs produced in the United States by
the percent of wastewater treated centrally, or percent collected (about 82 percent) (U.S. Census Bureau 2017),
the relative percentage of wastewater treated by aerobic and anaerobic systems (other than constructed
wetlands), the relative percentage of aerobic systems at wastewater facilities with and without primary treatment
(EPA 1992,1996, 2000, and 2004), the relative percentage of anaerobic systems at wastewater facilities with and
without primary treatment (EPA 1992,1996, 2000, and 2004), the percentage of BODs treated after primary
treatment (67.5 percent, 32.5 percent removed in primary treatment) (Metcalf & Eddy 2014), the maximum CFU-
producing capacity of domestic wastewater (B0, 0.6 kg Cm/kg BOD) (IPCC 2006), and the relative methane
correction factors (MCF) for not well-managed aerobic (0.3) (IPCC 2006), and anaerobic (0.8) (IPCC 2006) systems.
All aerobic systems are assumed to be well-managed as there are currently no data available to quantify the
number of systems that are not well-managed.
Table 7-9 presents total BODs produced for 1990 through 2018. The proportions of domestic wastewater treated
onsite versus at centralized treatment plants were based on data from the 1989,1991,1993,1995,1997,1999,
2001, 2003, 2005, 2007, 2009, 2011, 2013, 2015 and 2017 American Housing Surveys conducted by the U.S. Census
Bureau (U.S. Census Bureau 2017), with data for intervening years obtained by linear interpolation and 2018
forecasted using 1990 to 2017 data. The BODs production rate was determined using BOD generation rates per
capita both with and without kitchen scraps (Metcalf & Eddy 2003; Metcalf & Eddy 2014) as well as an estimated
percent of housing units that utilize kitchen garbage disposals (ERG 2018a). The percent BODs removed by primary
treatment for domestic wastewater was obtained from Metcalf & Eddy (2014).The percent of wastewater flow to
aerobic and anaerobic systems, the percent of aerobic and anaerobic systems that do and do not employ primary
treatment, and the wastewater flow to POTWs that have anaerobic digesters were obtained from the 1992,1996,
2000, and 2004 Clean Watersheds Needs Survey (CWNS) (EPA 1992,1996, 2000, and 2004). Data for intervening
years were obtained by linear interpolation and the years 2005 through 2018 were forecasted from the rest of the
time series. The percent of wastewater flow to aerobic systems that use only constructed wetlands and
wastewater flow to POTWs that use constructed wetlands as tertiary treatment were obtained from the 1992,
1996, 2000, 2004, 2008, and 2012 CWNS (EPA 1992,1996, 2000, 2004, 2008b, and 2012). Data for intervening
years were obtained by linear interpolation and the years 2013 through 2018 were forecasted from the rest of the
time series.
Table 7-9: U.S. Population (Millions) and Domestic Wastewater BODs Produced (kt)
Year	Population BOD5
1990	253	8,131
2005	300	9,624
2014	323	9,657
2015	325	9,743
2016	327	9,828
2017	329	9,911
2018	333	10,032
Sources: U.S. Census Bureau (2019); ERG
(2019a).
For constructed wetlands, an MCF of 0.4 was used, which is the IPCC suggested MCF for surface flow wetlands.
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 BODs concentration of 30 mg/L was used for wastewater entering
7-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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).
In addition, methane emissions were calculated for systems that treat wastewater with constructed wetlands and
systems that use constructed wetlands as tertiary treatment; however, constructed wetlands are a relatively small
portion of wastewater treated centrally (<0.1 percent).
Emissions from Anaerobic Digesters:
Total Cm emissions from anaerobic digesters were estimated by multiplying the wastewater influent flow to
POTWs with anaerobic digesters, the cubic feet of digester gas generated per person per day divided by the flow to
POTWs, the fraction of CH4 in biogas (0.65), the density of CH4 (662 g Cm/m3 CH4) (EPA 1993a), one minus the
destruction efficiency from burning the biogas in an energy/thermal device (0.99 for enclosed flares) and then
converting the results to kt/year.
The CH4 destruction efficiency for CH4 recovered from sludge digestion operations, 99 percent, was selected based
on the range of efficiencies (98 to 100 percent) recommended for flares in AP-42 Compilation of Air Pollutant
Emission Factors, Chapter 2.4 (EPA 1998), along with data from CAR (2011), Sullivan (2007), Sullivan (2010), and
UNFCCC (2012). The cubic feet of digester gas produced per person per day (1.0 ft3/person/day) and the
proportion of CH4 in biogas (0.65) come from Metcalf & Eddy (2014). The wastewater flow to a POTW (100
gal/person/day) was taken from the Great Lakes-Upper Mississippi River Board of State and Provincial Public
Health and Environmental Managers, "Recommended Standards for Wastewater Facilities (Ten-State Standards)"
(2004).
Table 7-10 presents domestic wastewater CH4 emissions for both septic and centralized systems, including
anaerobic digesters, in 2018.
Table 7-10: Domestic Wastewater ChU Emissions from Septic and Centralized Systems
(2018, MMT CO2 Eq. and Percent)

CH4 Emissions (MMT C02 Eq.)
% of Domestic Wastewater CH4
Septic Systems
5.9
70.4%
Centrally-Treated Aerobic Systems
0.03
0.4%
Centrally-Treated Anaerobic Systems
2.2
26.8%
Anaerobic Digesters
0.2
2.4%
Total
8.4
100%
Note: Totals may not sum due to independent rounding.
Industrial Wastewater CH4 Emission Estimates
Methane emission estimates from industrial wastewater were developed according to the methodology described
in the 2006IPCC Guidelines. 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 these
industries 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 emissions
for these sectors for 2018 are displayed in Table 7-11 below. Table 7-12 contains production data for these
industries.
Waste 7-23

-------
Table 7-11: Industrial Wastewater ChU Emissions by Sector (2018, MMT CO2 Eq. and
Percent)

CH4 Emissions (MMT
CO2 Eq.)
% of Industrial
Wastewater CH4
Meat & Poultry
4.8
81.3%
Pulp & Paper
0.6
9.8%
Fruit &



0.2
3.0%
Vegetables
Petroleum

2.6%

0.2
Refineries
Ethanol

2.4%

0.1
Refineries
Breweries
0.05
1%
Total	5.9	100%
Note: Totals may not sum due to independent rounding.
Table 7-12: 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


Petroleum
Year
Pulp and Paper3
Killed)
Killed) Juices
Ethanol
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
23.2
818.6
2014
80.9
32.2
26.9 45.3
42.8
22.5
903.9
2015
80.9
32.8
27.7 44.6
44.2
22.4
914.5
2016
79.9
34.2
28.3 43.2
45.8
22.3
926.0
2017
80.0
35.4
28.9 42.7
47.2
21.8
933.5
2018
75.7
36.4
29.4 42.1
48.0
21.5
951.4
a Pulp and paper production is the sum of market pulp production plus paper and paperboard production.
Sources: FAO (2019a) and FAO (2019b); USDA (2019a); Cooper (2018) and RFA (2019a and 2019b); Beer Institute
(2011) and TTB (2019); EIA (2019).
Methane emissions from these categories were estimated by multiplying the annual product output by the
average outflow, the organics loading (in COD) in the outflow, the maximum Cm producing potential of industrial
wastewater (B0), and the percentage of organic loading assumed to degrade anaerobically in a given treatment
system (MCF). Ratios of BOD:COD in various industrial wastewaters were obtained from EPA (1997a) and used to
estimate COD loadings. The B0 value used for all industries is the IPCC default value of 0.25 kg Cm/kg COD (IPCC
2006).
For each industry, the percent of plants in the industry that treat wastewater on site, the percent of plants that
have a primary treatment step prior to biological treatment, and the percent of plants that treat wastewater
anaerobically were defined. The percent of wastewater treated anaerobically onsite (TA) was estimated for both
primary treatment (%TAP) and secondary treatment (%TAS). For plants that have primary treatment in place, an
estimate of COD that is removed prior to wastewater treatment in the anaerobic treatment units was
incorporated. The values used in the %TA calculations are presented in Table 7-13 below.
The methodological equations are:
CH4 (industrial wastewater) = [P x W x COD x %TAP xB0x MCF] + [P x W x COD x %TAS xB0x MCF]
7-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
where,
o/0TAp = [%Plants0 x %WWa,P x %CODP]
o/0TAs = [%Plantsa x %WWa,s x %CODs] + [%Plantst x %WWa,t x %CODs]
CH4 (industrial wastewater) = Total CH4 emissions from industrial wastewater (kg/year)
P
W
COD
%TAP
%TAS
%Plants0
%WWa,p
%CODP
%Plantsa
%Plantst
%WWa,s
%WWa,t
%CODs
Bo
MCF
= Industry output (metric tons/year)
= Wastewater generated (m3/metric ton of product)
= Organics loading in wastewater (kg/m3)
= Percent of wastewater treated anaerobically on site in primary treatment
= Percent of wastewater treated anaerobically on site in secondary treatment
= Percent of plants with onsite treatment
= Percent of wastewater treated anaerobically in primary treatment
= Percent of COD entering primary treatment
= Percent of plants with anaerobic secondary treatment
= Percent of plants with other secondary treatment
= Percent of wastewater treated anaerobically in anaerobic secondary
treatment
= Percent of wastewater treated anaerobically in other secondary treatment
= Percent of COD entering secondary treatment
= Maximum Cm producing potential of industrial wastewater (kg Cm/kg COD)
= CH4 correction factor, indicating the extent to which the organic content
(measured as COD) degrades anaerobically
Alternate methodological equations for calculating %TA were used for secondary treatment in the pulp and paper
industry to account for aerobic systems with anaerobic portions. These equations are:
%TAa = [%Plantsa x %WWa,s x %CODs] + [%Plantsa,t x %WWa,t x CODs]
%TAa,t = [%Plantsa,t x %WWa,s x %CODs]
where,
%TAa
%TAa,t
%Plantsa
%Plantsa,t
%WWa,s
%WWa,t
%CODs
Percent of wastewater treated anaerobically on site in secondary treatment
Percent of wastewater treated in aerobic systems with anaerobic portions
on site in secondary treatment
Percent of plants with anaerobic secondary treatment
Percent of plants with partially anaerobic secondary treatment
Percent of wastewater treated anaerobically in anaerobic secondary
treatment
Percent of wastewater treated anaerobically in other secondary treatment
Percent of COD entering secondary treatment
As described below, the values presented in Table 7-13: were used in the emission calculations and are described
in detail in ERG (2008), ERG (2013a), and ERG (2013b).
Table 7-13: Variables Used to Calculate Percent Wastewater Treated Anaerobically by
Industry (Percent)
Industry
Variable
Pulp


Fruit/
Ethanol
Ethanol


Breweries
and
Meat
Poultry
Vegetable
Production
Production
Petroleum
Breweries
- Non-

Paper
Processing
Processing
Processing
- Wet Mill
- Dry Mill
Refining
-Craft
Craft
%TAP
0
0
0
0
0
0
0
0
0
%TAS
0
33
25
4.2
33.3
75
23.6
0
0
%TAa
2.2
0
0
0
0
0
0
0
0
<
1—
0s-
11.8
0
0
0
0
0
0
0
0
Waste 7-25

-------
%Plants0
60
100
100
11
100
100
100
100
1
%Plantsa
5
33
25
5.5
33.3
75
23.6
0
0
%Plantsa,t
28
0
0
0
0
0
0
0
0
%PlantSt
35
67
75
5.5
66.7
25
0
0
0
%WWa,p
0
0
0
0
0
0
0
0
0
%WWa,s
100
100
100
100
100
100
100
0
0
%WWa,t
0
0
0
0
0
0
0
0
0
%CODp
100
100
100
100
100
100
100
0
0
%CODs
42
100
100
77
100
100
100
0
0
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 (2008); ERG (2013a); and ERG (2013b).
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. In determining the
percent that degrades anaerobically, both primary and secondary treatment were considered. 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). However, because the vast majority of primary treatment operations at U.S.
pulp and paper mills use mechanical clarifiers, and less than 10 percent of pulp and paper wastewater is managed
in primary settling ponds that are not expected to have anaerobic conditions, negligible emissions are assumed to
occur during primary treatment.
Approximately 42 percent of the BOD passes on to secondary treatment, which consists of activated sludge,
aerated stabilization basins, or non-aerated stabilization basins. Based on EPA's OAQPS Pulp and Paper Sector
Survey, 5.3 percent of pulp and paper mills reported using anaerobic secondary treatment for wastewater and/or
pulp condensates (ERG 2013a). Twenty-eight percent of mills also reported the use of quiescent settling ponds.
Using engineering judgment, these systems were determined to be aerobic with possible anaerobic portions. For
the truly anaerobic systems, an MCF of 0.8 is used, as these are typically deep stabilization basins. For the partially
anaerobic systems, an MCF of 0.2 is used, which is the 2006IPCC Guidelines-suggested MCF for shallow lagoons.
A time series of CFU emissions for 1990 through 2018 was developed based on paper and paperboard production
data from the Food and Agricultural Organization of the United Nations (FAO) database FAOSTAT. (FAO 2019a) and
market pulp production data from FAO Pulp and Paper Capacities Reports (FAO 2019b). Market pulp production
values were available directly for 1998, 2000 through 2003, and 2010 through 2017. 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 2019a). The percent of woodpulp that is market pulp for 1990 to 1997
was assumed to be the same as 1998,1999 was interpolated between values for 1998 and 2000, 2000 through
2009 were interpolated between values for 2003 and 2010, and 2018 was forecasted from the rest of the time
series. A time series of the overall wastewater outflow 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 2015, 2017 and 2018
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
2018 (EPA 1997b; EPA 1993b; World Bank 1999; Malmberg 2018). Data for intervening years were obtained by
linear interpolation. The COD:BOD ratio used to convert the organic loading to COD for pulp and paper mills was
2.5 for the entire time series (Malmberg 2018).
7-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. 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 B0 of 0.25 kg Cm/kg COD and default MCF
of 0.8 for anaerobic lagoons were used to estimate the Cm produced from these on-site treatment systems.
Production data on carcass weight and live weight killed for the meat and poultry industry were obtained from the
USDA Agricultural Statistics Database and the Agricultural Statistics Annual Reports (USDA 2019a). Data collected
by EPA's Office of Water provided estimates for wastewater flows into anaerobic lagoons: 5.3 and 12.5 m3/metric
ton for meat and poultry production (live weight killed), respectively (EPA 2002). The loadings are 2.8 and 1.5 g
BOD/liter for meat and poultry, respectively (EPA 2002). The COD:BOD ratio used to convert the organic loading to
COD for both meat and poultry facilities was 3 (EPA 1997a).
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. Effluent is suitable for discharge to POTWs. This
industry is likely to 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).
Consequently, 4.2 percent of these wastewater organics are assumed to degrade anaerobically (ERG 2008). The
IPCC default B0 of 0.25 kg Cm/kg COD and default MCF of 0.8 for anaerobic treatment were used to estimate the
Cm produced from these on-site treatment systems. The USDA National Agricultural Statistics Service (USDA
2019a, 2019c) provided production data for potatoes, other vegetables, citrus fruit, non-citrus fruit, and grapes
processed for wine. Outflow and BOD data, presented in Table 7-14 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. The COD:BOD
ratio used to convert the organic loading to COD for all fruit, vegetable, and juice facilities was 1.5 (EPA 1997a).
Table 7-14: Wastewater Flow (m3/ton) and BOD Production (g/L) for U.S. Vegetables,
Fruits, and Juices Production
Commodity	Wastewater Outflow (mB/ton)	BOD (g/L)
Vegetables
Potatoes	10.27	1.765
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 only about 2 percent of ethanol production and is only in an experimental stage 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 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
Waste 7-27

-------
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 digesters is commonly collected and either flared
or used as fuel in the ethanol production process (ERG 2006).
Available information was compiled from the industry on wastewater generation rates, which ranged from 1.25
gallons per gallon ethanol produced (for dry milling) to 10 gallons per gallon ethanol produced (for wet milling)
(Ruocco 2006a; Ruocco 2006b; Merrick 1998; Donovan 1996; NRBP 2001). COD concentrations were found to be
about 3 g/L (Ruocco 2006a; Merrick 1998; White and Johnson 2003). One hundred percent of plants were
estimated to have on-site wastewater treatment, and the variables used to calculate percent wastewater treated
anaerobically are presented in Table 7-13. A default MCF of 0.8 for anaerobic treatment was used to estimate the
Cm 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 2006). Methane emissions for dry milling and wet milling processes were then estimated as
follows:
Methane = [Production x Flow x COD x 3.785 x ([%Plants0 x %WWa,P x %C0DP] + [%Plantsa x %WWa,s x
%C0Ds] + [%Plantst x %WWa,t x %C0Ds]) xB0x MCF x % Not Recovered] + [Production x Flow x 3.785 x
COD x ([%PlantSo x %WWa,P x %C0DP] + [%Plantsa x %WWa,s x %C0Ds] + [%Plantst x %WWa,t x %C0Ds])
x Bo x MCF x (% Recovered) x (1-DE)] x 1/109
where,
Production
= Gallons ethanol produced (wet milling or dry milling)
Flow
= Gallons wastewater generated per gallon ethanol produced
COD
= COD concentration in influent (g/l)
3.785
= Conversion factor, gallons to liters
%Plants0
= Percent of plants with onsite treatment
%WWa,p
= Percent of wastewater treated anaerobically in primary treatment
%CODP
= Percent of COD entering primary treatment
%Plantsa
= Percent of plants with anaerobic secondary treatment
%Plantst
= Percent of plants with other secondary treatment
%WWa,s
= Percent of wastewater treated anaerobically in anaerobic secondary treatment
%WWa,t
= Percent of wastewater treated anaerobically in other secondary treatment
%CODs
= Percent of COD entering secondary treatment
Bo
= Maximum methane producing capacity (g CFU/g COD)
MCF
= Methane correction factor
% Recovered
= Percent of wastewater treated in system with emission recovery
% Not Recovered = 1 - percent of wastewater treated in system with emission recovery
DE	= Destruction efficiency of recovery system
1/109	= Conversion factor, g to kt
A time series of CFU emissions for 1990 through 2017 was developed based on dry and wet milling production data
from the Renewable Fuels Association (RFA) (Cooper 2018). In 2018, production for dry and wet milling was based
on total production data and the average monthly grain-use for dry and wet milling (RFA 2019a; RFA 2019b).
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
7-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Collection Request (ICR) for petroleum refineries in 2011.6 Of the responding facilities, 23.6 percent reported
using non-aerated surface impoundments or other biological treatment units, both of which have the potential to
lead to anaerobic conditions (ERG 2013b). In addition, the wastewater generation rate was determined to be 26.4
gallons per barrel of finished product (ERG 2013b). An average COD value in the wastewater was estimated at 0.45
kg/m3 (Benyahia et al. 2006). A default MCF of 0.3 was used for partially aerobic systems.
The equation used to calculate Cm generation at petroleum refining wastewater treatment systems is presented
below:
A time series of Cm emissions for 1990 through 2018 was developed based on production data from the EIA 2019.
Breweries. Since 2010, the number of breweries has increased from less than 2,000 to more than 7,000 (Brewers
Association 2019). 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 Cm emissions from anaerobic wastewater
treatment. However, because many breweries recover their Cm, 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 2019). 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 2018.
The amount of water usage by craft breweries was estimated using the Brewers Association's 2015 Sustainability
Benchmarking Report (Brewers Association 2016a) and the 2016 Benchmarking Update (Brewers Association 2017;
ERG 2018b). Non-craft brewery water usage values were from the Beverage Industry Environmental Roundtable
(BIER) benchmarking study (BIER 2017).
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. On average, full-strength
wastewater is about 10,600 mg/L BOD, with a typical BOD:COD ratio of 0.6 (Brewers Association 2016b). 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 B0 of 0.25 kg CHVkg COD and default MCFs of 0.8 for anaerobic
treatment and 0 for aerobic treatment were used to estimate the CH4 produced from these on-site treatment
6 Available online at .
Methane = Flow x COD x %TA xB0x MCF
where,
Flow
COD
%TA
Bo
MCF
Annual flow treated through anaerobic treatment system (m3/year)
COD loading in wastewater entering anaerobic treatment system (kg/m3)
Percent of wastewater treated anaerobically on site
Maximum methane producing potential of industrial wastewater (kg CHVkg COD)
Methane correction factor
Waste 7-29

-------
systems (IPCC 2006). The amount of CFU 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.
The assumed distribution of wastewater treatment for craft and non-craft breweries are shown in Table 7-15.
Table 7-15: Wastewater Treatment Distribution for Breweries
Operation Type
Treatment Type
Non-Craft
Craft
Discharge to POTW with no pretreatment
0%
99%
Discharge to POTW following side streaming
0%
0.5%
Pretreatment with aerobic biological


treatment
1%
0%
Pretreatment with anaerobic reactor
99%
0.5%
Source: Stier, J. (2018)
Methane emissions were then estimated for non-craft breweries and for craft breweries as follows:
Methane = [(Production x Water Usage x WW:W x 31)/264.172) x COD x ([%Plantspotw x MCFpotw] +
[%Plantsss x MCFpotw] + [%Plantsaer x MCFaer] + [%Plantsa x MCFa]) x Bo x % Not Recovered] +
[(Production x Water Usage x WW:W x 31)/264.172) x COD x ([%Plantspotw x MCFpotw] + [%Plantsss x
MCFpotw] + [%Plantsaer x MCFaer] + [%Plantsa x MCFa]) x Bo x (% Recovered) x (1-DE)] x 1/106
where,
Production
= Barrels beer produced (non-craft breweries or craft breweries)
Water Usage
= Barrels water utilized per barrels beer produced
WW:W
= Ratio, barrels of wastewater generated per barrels of water utilized
COD
= COD concentration in influent (kg/m3)
31
= Conversion factor, gallons to barrels beer
264.172
= Conversion factor, gallons to m3
%PlantsPotw
= Percent of plants that discharge to POTW without pretreatment
MCFpotw
= Methane correction factor, discharge to POTW
%PlantSss
= Percent of plants with sidestreaming prior to POTW discharge
%PlantSaer
= Percent of plants with primary aerobic treatment
MCFaer
= Methane correction factor, aerobic systems
%Plantsa
= Percent of plants with anaerobic treatment
MCFa
= Methane correction factor, anaerobic systems
Bo
= Maximum methane producing capacity (g CFU/g COD)
% Recovered
= Percent of wastewater treated in system with emission recovery
% Not Recovered = 1 - percent of wastewater treated in system with emission recovery
DE	= Destruction efficiency of recovery system
1/106	= Conversion factor, kg to Gg
Domestic Wastewater N2O Emission Estimates
Nitrous oxide emissions from domestic wastewater (wastewater treatment) were estimated using the IPCC (2006)
methodology and supplemented with IPCC (2014) methodology to include constructed wetland emissions,
including calculations that take into account N removal with biosolids, non-consumption and
industrial/commercial wastewater N, and emissions from advanced and constructed wetlands at centralized
wastewater treatment plants:
7-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
In the United States, a certain amount of N is removed with biosolids, which is applied to land, incinerated, or
landfilled (Nsludge). The value for N discharged into aquatic environments as effluent is reduced to account for the
biosolids application.
The 2006IPCC Guidelines use annual, per capita protein consumption (kg protein/person-year). For this Inventory,
the amount of protein available to be consumed is estimated based on 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.
Small amounts of gaseous nitrogen oxides are formed as byproducts in the conversion of nitrate to N gas in anoxic
biological treatment systems. Approximately 7 g N2O is generated per capita per year if wastewater treatment
includes intentional nitrification and denitrification (Scheehle and Doom 2001). Analysis of the use of treatment
systems in the United States that include denitrification has shown a significant increase in the time period
between 2004 and 2012, from serving populations totaling 2.4 million people to 21.3 million people (EPA 2004 and
EPA 2012). This is consistent with efforts throughout the United States to improve nutrient removal at centralized
treatment systems in response to specific water quality concerns. Based on an emission factor of 7 g per capita per
year, and data from CWNS 2004, 2008, and 2012, approximately 21.2 metric tons of additional N2O may have been
emitted via denitrification in 2004, while about 186 metric tons may have been emitted via denitrification in both
2008 and 2012. Similar analyses were completed for each year in the Inventory using data from CWNS on the
amount of wastewater in centralized systems treated in denitrification units. Plants without intentional
nitrification or denitrification are assumed to generate 3.2 g N2O per capita per year.
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. The emission factor of 0.0013 kg INhO-N/kg N produced for constructed wetlands is
from IPCC (2014).
N2O emissions from wastewater treatment plants are estimated, and as such, the N associated with these
emissions is subtracted from the amount of N estimated to be discharged into aquatic environments as effluent,
consistent with the 2006 IPCC Guidelines.
Nitrous oxide emissions from domestic wastewater were estimated using the following methodology:
N2OPLANT = N2ONIT/DENIT + N2OWOUTNIT/DENIT+ N2OCWONLY + N2OCW TERTIARY
N20nit/denit= [(USpopnd) X EF2 X Find-com] X 1/109
N20woutnit/denit = {[(USpop X WWTP) - USpopnd - USpopcw] X 106 X Find-com X EFi} X 1/109
N20cwonly = {[(USpopcw X 106 X Protein X Fnpr X Fnon-con X Find-com) X EF4] X 44/28} X 1/106
N2O CW TERTIARY — {[(New,inf x POTW_flow_CW x 3.79 x 365.25) x EF4] x 44/28} x 1/106
N20effluent = [(USpop X WWTP X Protein X Fnpr X Fnon-con X Find-com) - Nsludge - (N20plant X 106 X 28/44)] X
EFs X 44/28 X 1/106
N20total = N20plant + N20effluent
where,
N20cwonly
N20total
N20plant
N20nit/denit
N2OWOUT NIT/DENIT
Annual emissions of N2O (kt)
N2O emissions from centralized wastewater treatment plants (kt)
N2O emissions from centralized wastewater treatment plants with
nitrification/denitrification (kt)
N2O emissions from centralized wastewater treatment plants without
nitrification/denitrification (kt)
N2O emissions from centralized wastewater treatment plants with constructed
wetlands only (kt)
Waste 7-31

-------
N20cwtertiary	= N2O emissions from centralized wastewater treatment plants with constructed
wetlands used as tertiary treatment (kt)
NzOeffluent	= N2O emissions from wastewater effluent discharged to aquatic environments (kt)
USpop	= U.S. population
USpopnd	= U.S. population that is served by biological denitrification
USpopcw	= U.S. population that is served by only constructed wetland systems
WWTP	= Fraction of population using WWTP (as opposed to septic systems)
POTW_flow_CW	= Wastewater flow to POTWs that use constructed wetlands as tertiary treatment
(MGD)
EFi	= Emission factor - plants without intentional denitrification
EF2	= Emission factor - plant with intentional nitrification or denitrification
Protein	= Annual per capita protein consumption (kg/person/year)
New,inf	= Influent nitrogen concentration to constructed wetlands used as tertiary
treatment (mg/L)
Fnpr	= Fraction of N in protein (kg N/kg protein)
Fnon-con	= Factor for non-consumed protein added to wastewater
Find-com	= Factor for industrial and commercial co-discharged protein into the sewer
Nsludge	= N removed with sludge, kg N/year
EF3	= Emission factor (kg N2O -N/kg sewage-N produced) - from effluent
EF4	= Emission factor (kg N2O -N/kg N produced) - constructed wetlands
3.79	= Conversion factor, gallons to liters
44/28	= Molecular weight ratio of N2O to N2
28/44	= Molecular weight ratio of N2 to N2O
1/106	= Conversion factor, kg to Gg
1/109	= Conversion factor, g to Gg
U.S. population data were taken from the U.S. Census Bureau International Database (U.S. Census Bureau 2019)
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 wastewater treatment plants is based on data
from the 1989, 1991, 1993, 1995, 1997, 1999, 2001, 2003, 2005, 2007, 2009, 2011, 2013, 2015 and 2017 American
Housing Survey (U.S. Census Bureau 2017). Data for intervening years were obtained by linear interpolation and
2018 was forecasted using 1990 to 2017 data. The emission factor (EFi) used to estimate emissions from
wastewater treatment for plants without intentional nitrification or denitrification was taken from IPCC (2006),
while the emission factor (EF2) used to estimate emissions from wastewater treatment for plants with intentional
nitrification or denitrification was taken from Scheehle and Doom (2001). The emission factor (EF4) used to
estimate emissions from surface flow constructed wetlands (0.0013 kg N2O -N/kg N produced) was taken from
IPCC (2014). Data on annual per capita protein intake were provided by the U.S. Department of Agriculture
Economic Research Service (USDA 2019b) and FAO (2019c). Protein consumption data was used directly from
USDA for 1990 to 2010 and 2011 through 2013 was calculated using FAO data and a scaling factor. 2014 through
2018 were forecasted from data for 1990 through 2013. An emission factor to estimate emissions from effluent
(EF3) has not been specifically estimated for the United States, thus the default IPCC value (0.005 kg N20-N/kg
sewage-N produced) was applied (IPCC 2006). The fraction of N in protein (0.16 kg N/kg protein) was also obtained
from IPCC (2006). The factor for non-consumed protein (1.2) and the factor for industrial and commercial co-
discharged protein (1.25) were obtained from IPCC (2006). The amount of nitrogen removed by denitrification
systems was taken from EPA (2008a), while the population served by denitrification systems was estimated from
Clean Watersheds Needs Survey (EPA 1992,1996, 2000, 2004, 2008b, and 2012). Sludge generation was obtained
from EPA (1999) for 1988,1996, and 1998 and from Beecher et al. (2007) for 2004. Intervening years were
interpolated and estimates for 2005 through 2018 were forecasted from the rest of the time series. The influent
nitrogen concentration to constructed wetlands used as tertiary treatment (25 mg/L) was obtained from Metcalf &
Eddy (2014). An estimate for the N removed as sludge (Nsludge) was obtained by determining the amount of sludge
disposed by incineration, by land application (agriculture or other), through surface disposal, in landfills, or through
ocean dumping (EPA 1993b; Beecher et al. 2007; McFarland 2001; EPA 1999). In 2018, 301 kt N was removed with
7-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
sludge. Table 7-16 presents the data for U.S. population, population served by biological denitrification, population
served by wastewater treatment plants, available protein, protein consumed, and nitrogen removed with sludge.
Table 7-16: U.S. Population (Millions), Population Served by Biological Denitrification
(Millions), Fraction of Population Served by Wastewater Treatment (percent), Available
Protein (kg/person-year), Protein Consumed (kg/person-year), and Nitrogen Removed with
Sludge (kt-N/year)






N Removed
Year
Population
PopulatioriND
WWTP Population
Available Protein
Protein Consumed
with Sludge
1990
253
2.0
75.6
43.1
33.2
214.2
2005
300
7.1
78.8
44.9
34.7
261.1
2014
323
20.8
80.8
44.3
34.1
288.7
2015
325
21.8
80.1
44.3
34.1
291.8
2016
327
22.8
81.1
44.3
34.1
294.8
2017
329
23.8
82.1
44.3
34.1
297.9
2018
333
24.8
81.9
44.3
34.1
300.9
Sources: Population: U.S. Census Bureau (2019); Population^: EPA (1992), EPA (1996), EPA (2000), EPA (2004), EPA (2008b),
EPA (2012); WWTP Population: U.S. Census Bureau (2017); Available Protein: USDA (2019b); N Removed with sludge: Beecher
et al. (2007), McFarland (2001), EPA (1999), EPA (1993c).
Uncertainty and Time-Series Consistency
The overall uncertainty associated with both the 2018 Cm and N2O emission estimates from wastewater
treatment and discharge was calculated using the 2006IPCC 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 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 N2O emissions include that
of biosolids disposal, total U.S. population, average protein consumed per person, fraction of N in protein, non-
consumption nitrogen factor, emission factors per capita and per mass of sewage-N, and for the percentage of
total population using centralized wastewater treatment plants. Uncertainty associated with 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 7-17. Methane emissions
from wastewater treatment were estimated to be between 10.2 and 17.4 MMT CO2 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 23 percent above the 2018 emissions estimate of 14.2 MMT CO2 Eq. Nitrous oxide emissions
from wastewater treatment were estimated to be between 1.3 and 10.5 MMT CO2 Eq., which indicates a range of
approximately 74 percent below to 109 percent above the 2018 emissions estimate of 5.0 MMT CO2 Eq.
Table 7-17: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from
Wastewater Treatment (MMT CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
(MMT CO? Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT C02 Eq.) (%)



Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Wastewater T reatment
ch4
14.2
10.2
17.4
-28%
+23%
Domestic
ch4
8.4
6.0
10.2
-28%
+22%
Waste 7-33

-------
Industrial	CH4	5.9	3.0	8.8	-48%	+50%
Wastewater Treatment N20	5.0	1.3	10.5	-74%	+109%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
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 emissions estimates; and
•	Compared estimates to previous estimates to identify significant changes.
All transcription errors identified were corrected and documented. The QA/QC analysis did not reveal any systemic
inaccuracies or incorrect input values.
Recalculations Discussion
Population data were updated to reflect revised U.S. Census Bureau datasets which resulted in changes to 2010
through 2017 values (U.S. Census Bureau 2019). American Housing Survey data were updated for percent of
wastewater treated centrally which affected 2016 and 2017 (U.S. Census Bureau 2017). EPA also updated the
percent calculation for centrally treated aerobic systems without primary sedimentation which affected the entire
time series.
EPA evaluated pulp and paper wastewater generation data and updated values for 2005 and 2016 which affected
emissions calculations for 2005 and 2015 through 2017 (AF&PA 2018). Market pulp production values were
updated to include "pulp of other fiber and paper and paperboard" and "dissolving pulp, wood and other raw
materials" after confirmation with NCASI that these values were appropriate to include in the market pulp
production (Malmberg 2019). This update affected emissions calculations for 1998 and 2000 through 2003.
EPA investigated updated sources for fruits, vegetables, and juices wastewater characteristics and outflow. EPA
evaluated a source that includes updated BOD and wastewater outflow information for some fruits and vegetables
included in the Inventory and determined updates to activity data were appropriate (CAST 1995). This update
affected industrial emissions calculations for the entire time series.
EPA updated the methodology used to estimate ethanol production for wet and dry milling as the source used in
previous Inventories is no longer readily available. EPA conferred with RFA and determined publicly available
production data used in conjunction with monthly grain-use data are an appropriate surrogate for calculating the
ethanol production at wet and dry mills (Lewis 2019; RFA 2019a; RFA 2019b).
The cumulative effect of these recalculations had minimal impact on the overall wastewater treatment emissions
estimates. Over the time series, the average total emissions increased by 0.25 percent from the previous Inventory
cycle. The changes ranged from the largest decrease, 0.19 percent (0.05 MMT CO2 Eq.), in 2017, to the largest
increase, 0.93 percent (0.16 MMT CO2 Eq), in 2016.
7-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Planned Improvements
IPCC recently announced the availability of the 2019 Refinement to the 2006 Guidelines for National Greenhouse
Gas Inventories. EPA is planning to incorporate the following improvements to the Inventory based on the 2019
Refinement:
•	Restructure the activity data on treatment systems in use at domestic and industrial treatment plants to
mirror the types of systems provided in the 2019 Refinement and incorporate updated emission factors,
including incorporating nitrous oxide emission estimates for septic systems.
•	Although there are insufficient data to capture emissions from collection systems, EPA plans to update
emission factors for centralized aerobic treatment based on the 2019 Refinement. The revised emission
factors account for incoming dissolved methane that is formed in the collection system and liberated
during aerobic treatment.
•	Develop the activity data to estimate methane and nitrous oxide emissions associated with wastewater
discharge using the new IPCC emission factors and updated U.S. activity data on BOD and N discharged
from domestic and industrial wastewater treatment plants.
•	Review and update the estimate of total organics in the wastewater, total organics and N removed during
treatment, and sludge produced, using updated default factors where necessary.
•	Identify key industries that have potential to generate nitrous oxide emissions for inclusion in the
Inventory. EPA expects that this improvement may take more than one cycle to fully incorporate into the
Inventory.
EPA is continuing to monitor the following potential sources for updating inventory data, including:
•	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;
•	Reports based on international research and other countries' inventory submissions to inform potential
updates to the Inventory's emission factors, methodologies, or included industries; and
•	Additional data sources for improving the uncertainty of the estimate of N entering municipal treatment
systems.
EPA also investigated data collected under the EPA's Greenhouse Gas Reporting Program (GHGRP) Subpart II,
Industrial Wastewater Treatment for use in improving the emission estimates for the industrial wastewater
category and for identifying whether anaerobic sludge digesters are in use. Because reporting data from the
GHGRP are not available for all inventory years and because only a few industrial facilities are required to report,
GHGRP data are not able to be used to improve estimates in the Inventory.
The inclusion of wastewater treatment emissions from dairy products processing into inventory estimates was
investigated. To date, there are insufficient data to determine if this industry constitutes a key source for the
United States. EPA will continue focusing on collecting wastewater treatment system data and wastewater
characteristics data. Anecdotal information obtained during previous investigations into the dairy products
processing industry noted that wastewater is often discharged to the sewer. EPA therefore reviewed the factor
used to reflect the contribution of nitrogen to domestic wastewater treatment systems from industrial and
commercial wastewater (Find-com = 1.25) to determine if it is appropriate for U.S. emissions estimates (and thereby
captures the vast majority of dairy products processing wastewater). EPA reviewed available industrial and
commercial flow contributions to POTWs using the CWNS data. After evaluating CWNS flow data for all available
years (1992,1996, 2000, 2004, 2008, and 2012), EPA determined the default IPCC factor of 1.25 appropriately
reflects the contributions of industrial and commercial wastewater flow to POTWs across the time series.
EPA will continue to look for methods to improve the transparency of the fate of sludge produced in wastewater
treatment.
Waste 7-35

-------
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 (CO2) 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 Cm is then oxidized to a large extent in the aerobic sections of the compost.
The estimated Cm 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 (N2O)
emissions can also be produced. The formation of N2O 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 N2O than, for
example, yard waste, however data are limited.
Even though CO2 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 2018, the amount of waste composted in the United States increased from 3,810 kt to 24,594 kt.
There was some fluctuation in the amount of waste composted between 2006 to 2009. A peak of 20,049 kt
composted was observed in 2008, followed by a steep drop the following year to 18,824 kt composted,
presumably driven by the economic crisis of 2009. Since then, the amount of waste composted has gradually
increased, and when comparing 2010 to 2018, a 34 percent increase in waste composted is observed. Emissions of
Cm and N2O from composting from 2010 to 2018 have increased by the same percentage. In 2018, CH4 emissions
from composting (see Table 7-18 and Table 7-19) were 2.5 MMT CO2 Eq. (98 kt), and N2O emissions from
composting were 2.2 MMT CO2 Eq. (7 kt), representing consistent emissions trends when compared to 2017. 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 N2O 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.
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. By 2010, 25 states, representing about 50 percent of the nation's
7-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
population, had enacted such legislation (ILSR 2014; BioCycle 2010). There are many more initiatives at the metro
and municipal level across the United States. More than 3,280 composting facilities exist in the United States with
most (71 percent) composting yard trimmings only (ISLR 2014).
In more recent years, bans and diversions have become more common for food wastes as well. As of September
2018, five states (California, Connecticut, Massachusetts, Rhode Island, Vermont) and six municipalities (Austin, TX;
Boulder, CO; New York City, NY; San Francisco, CA; Seattle, WA) had implemented organic waste bans or
mandatory recycling laws, most having taken effect after 2013 (BioCycle 2018a). In 2017, BioCycle released a
report in which 27 of 43 states that responded to their organics recycling survey noted that food waste (collected
residential, commercial, institutional, and industrial food waste) was recycled via anaerobic digestion and/or
composting. These 27 states reported an estimated total of 1.8 million tons of food waste diverted from landfills in
2016 (BioCycle 2018b). 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.
Table 7-18: ChU and N2O Emissions from Composting (MMT CO2 Eq.)
Activity
1990
2005
2014
2015
2016
2017
2018
ch4
0.4
1.9
2.1
2.1
2.3
2.4
2.5
n2o
0.3
1.7
1.9
1.9
2.0
2.2
2.2
Total
0.7
3.5
4.0
4.0
4.3
4.6
4.7
Table 7-19: ChU and N2O Emissions from Composting (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
ch4
15
75
84
85
91
98
98
n2o
1
6
6
6
7
7
7
Methodology
Methane and N2O 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 emissions shown in Table 7-18 and Table 7-19 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):
Et =MxEFt
where,
Ei = Cm or N2O emissions from composting, kt CH4 or N2O,
M = mass of organic waste composted in kt,
EFi = emission factor for composting, 41 Cl-U/kt of waste treated (wet basis) and
0.31 N20/kt of waste treated (wet basis) (IPCC 2006), and
i	= designates either CH4or N2O.
Per IPCC Tier 1 methodology defaults, the emission factors for CH4 and N2O 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
and Cornell 1996).
Estimates of the quantity of waste composted (M, wet weight as generated) are presented in Table 7-20 for select
years. Estimates of the quantity composted for 1990, 2005, 2010, and 2014 to 2015 were taken from EPA's
Advancing Sustainable Materials Management: Facts and Figures 2015 (EPA 2018); the estimates of the quantities
Waste 7-37

-------
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 extrapolated using the
2017 quantity composted and a ratio of the U.S. population growth between 2017 to 2018 (U.S. Census Bureau
2019).
Table 7-20: U.S. Waste Composted (kt)
Activity
1990
2005
2014
2015
2016
2017
2018
Waste Composted
3,810
18,643
20,884
21,219
22,780
24,485
24,594
Uncertainty and Time-Series Consistency
The estimated uncertainty from the 2006IPCC Guidelines is ±50 percent for the Tier 1 methodology.
Emissions from composting in 2018 were estimated to be between 2.3 and 7.0 MMT CO2 Eq., which indicates a
range of 50 percent below to 50 percent above the 2018 emission estimate of each gas (see Table 7-21).
Table 7-21: Tier 1 Quantitative Uncertainty Estimates for Emissions from Composting (MMT
CO2 Eq. and Percent)
Source
Gas
2018 Emission Estimate
Uncertainty Range Relative to Emission Estimate
(MMT CO? Eq.)
(MMT C02
Eq.)
(%)




Lower
Upper
Lower
Upper



Bound
Bound
Bound
Bound
Composting
ch4
2.5
1.2
3.7
-50%
+50%







n2o
2.2
1.1
3.3
-50%
+50%
ition
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
Composting estimates for 2016 and 2017 were revised with the November 2019 publication of EPA's Advancing
Sustainable Materials Management: 2016 and 2017 Tables and Figures report. These revisions resulted in changes
to the quantity of waste composted and the estimated emissions. The quantity of waste composted increased
from 23.7 million tons in the previous Inventory report to 27.0 million tons (or 14 percent) in the current Inventory
report for 2017; and increased from 23.5 million tons in the previous inventory report to 25.1 million tons (or 7
percent) in the current inventory report for 2016. This change increased total emissions by 28 percent or 0.6 MMT
CO2 Eq. for 2017, and by 13 percent or 0.3 MMT CO2 Eq. for 2016.
Planned Improvements
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. This information will be used to determine whether the emission factors
used in the current methodology can be revised or expanded to account for geographical differences and/or
7-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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. Additionally,
composting systems that primarily compost food waste may generate Cm at different rates than those that
compost yard trimmings because the food waste may have a higher moisture content and more readily degradable
material. This information will also be used to reassess the variance in emissions and associated uncertainty factors
applied to each greenhouse gas (Cm and N2O).
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. Additional efforts are being made
to collect information on the year the identified facilities began operating, an estimate of the quantity of waste
composted, and approximate land area or population (or households) the facilities serve. This data may be
incorporated into the current or future Inventories as a methodological improvement.
Additionally, EPA is actively collecting information on stand-alone anaerobic digesters in the United States so that
this source may be included in future Inventory estimates. In 2018, EPA conducted a review of publicly available
information on anaerobic digestion in the United States. While many primary sources were evaluated, EPA
determined that a report by the Environmental Research and Education Foundation (EREF) and data from an
information collection request (ICR) by EPA Region 5 provided the most relevant data; however, the data provided
by each report were not detailed enough to allow for the creation of a time series of waste sent to anaerobic
digesters in the United States for purposes of including this source in future Inventory emissions estimates. EPA is
aware of a new ICR report which is expected to be published in Fall 2019 which could potentially be used to
construct an emissions time series for this source. Once this ICR is published, EPA will determine if a time series for
emissions from stand-alone anaerobic digesters can indeed be created for Inventory purposes, and if so, will
incorporate this emission source within the next two Inventory cycles.
7.4 Waste Incineration (CRF Source Category
5C1)	
As stated earlier in this chapter, carbon dioxide (CO2), nitrous oxide (N2O), 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 2018 resulted in 11.4 MMT CO2 Eq. of emissions,
over half of which (6.4 MMT CO2 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
Waste 7-39

-------
500 kt CO2 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 UNFCCC7 request that information be provided on
precursor greenhouse gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-Cm volatile organic
compounds (NMVOCs), and sulfur dioxide (SO2). 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 SO2, 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. Total emissions of NOx, CO, and NMVOCs from waste sources for the years 1990 through 2018 are provided
in Table 7-22. Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and Annex 6.3.
Table 7-22: Emissions of NOx, CO, and NMVOC from Waste (kt)
Gas/Source	1990	2005	2014 2015 2016 2017 2018
NOx
+
2
2
2
2
2
2
Landfills
+
2
2
2
2
2
2
Wastewater Treatment
+
0
0
0
0
0
0
Miscellaneous3
+
0
0
0
0
0
0
CO
1
7
8
8
8
8
8
Landfills
1
6
8
8
8
8
8
Wastewater Treatment
+
+
1
1
1
1
1
Miscellaneous3
+
0
0
0
0
0
0
NMVOCs
673
114
68
68
68
68
68
Wastewater Treatment
57
49
29
29
29
29
29
Miscellaneous3
557
43
26
26
26
26
26
Landfills
58
22
13
13
13
13
13
Note: Totals may not sum due to independent rounding.
+ Does not exceed 0.5 kt.
3 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 2018 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2019) 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.
7 See .
7-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 2018. Details on the
emission trends through time are described in more detail in the Methodology section, above.
Waste 7-41

-------
The United States does not report any greenhouse gas emissions under the Intergovernmental Panel on Climate
Change (IPCC) "Other" sector.
Other 8-1

-------
9. Recalculations and 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, including the transparency, completeness, consistency and
overall usefulness of the report. In this effort, 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; 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 2017) 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 Table 9-1 and Table 9-2. Table 9-1 summarizes the quantitative effect of all changes on U.S. greenhouse gas
emissions in 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 tables present results relative to the previously published Inventory (i.e., the 1990 to 2017
report) in units of million metric tons of carbon dioxide equivalent (MMT CO2 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 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.
• Agricultural Soil Management (N2O). Several major improvements have been implemented in this
Inventory leading to the need for recalculations, including additional information from the United States
Department of Agriculture-Natural Resource Conservation Service's Conservation Effects Assessment
Project (USDA-NRCS CEAP) survey, United States Department of Agriculture- Economic Research Service's
Agricultural Resource Management Survey (USDA-ERS ARMS) data, Conservation Technology Information
Center (CTIC) data and USDA Census of Agriculture data, Natural Resource Inventory (NRI) survey,
(National Land Cover Database) NLCD data, modeling soil organic carbon stock changes to 30 cm with the
Tier 3 approach (previously modeled to 20 cm depth), modeling the N cycle with freeze-thaw effects on
soil N2O emission, and addressing the effect of cover crops on greenhouse gas emissions and removals.
Other improvements include better resolving the timing of tillage, planting, fertilization and harvesting
based on the USDA-NRCS CEAP survey and state-level information on planting and harvest dates;
improving the timing of irrigation; and crop senescence using growing degree relationships. The surrogate
Recalculations and Improvements 9-1

-------
data method was also applied to re-estimate N2O emissions from 2016 to 2017. These changes resulted in
an average increase in emissions of 57.3 MMT CO2 Eq. (22 percent) from 1990 to 2017 relative to the
previous Inventory.
•	Forest Land Remaining Forest Land: Changes in Forest Carbon Stocks (CO2). New national forest inventory
(NFI) data contributed to increases in forest land area and stock changes, particularly in the Intermountain
West region. 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. This resulted in a
structural change in the soil organic carbon estimates for mineral and organic soils across the entire time
series. Updated harvested wood products (HWPs) data from 2003 through 2017 led to changes in
Products in Use and Solid Waste Disposal Sites (SWDS) between the previous Inventory and the current
Inventory. The recalculations resulted in an average annual increase in C stock change losses of 46.4 MMT
CO2 Eq. (7 percent), across the 1990 through 2017 time series, relative to the previous Inventory.
•	Land Converted to Grassland: Changes in all Ecosystem Carbon Stocks (CO2). Differences in biomass, dead
wood and litter C stock changes in Forest Land Converted to Grassland can be attributed to incorporation
of the latest Forest Inventory and Analysis National Program (FIA) data. Recalculations for the soil C stock
changes are associated with several improvements to both the Tier 2 and 3 approaches that are discussed
in the Cropland Remaining Cropland section. As a result of these improvements to the Inventory, Land
Converted to Grassland has a larger reported gain in C compared to the previous Inventory, estimated at
an average of 35.2 MMT CO2 Eq. over the time series. This represents greater than 610 percent increase
of C for Land Converted to Grassland compared to the previous Inventory and is largely driven by the
methodological changes for estimating the soil C stock changes.
•	Natural Gas Systems (CHa). EPA thoroughly evaluated relevant information available and made several
updates to the Inventory, including: using EPA's Greenhouse Gas Reporting Program (GHGRP), Bureau of
Ocean Energy Management (BOEM), and other data to calculate emissions from offshore production; and
using GHGRP and Zimmerle et al. study data to calculate gathering and boosting station emissions. In
addition, certain sources did not undergo methodological updates, but CH4 and/or CO2 emissions changed
by greater than 0.05 MMT CO2 Eq., comparing the previous estimate for 2017 to the current
(recalculated) estimate for 2017 (the emissions changes were mostly due to GHGRP data submission
revisions). These sources include: hydraulically fractured (HF) gas well completions; production segment
pneumatic controllers; liquids unloading; production segment storage tanks; HF and non-HF gas well
workovers; and acid gas removal (AGR) vents, flares, reciprocating compressors, and blowdowns at gas
processing plants. The recalculations resulted in an average decrease in CH4 emission estimates across the
1990 through 2017 time series, compared to the previous Inventory, of 14.2 MMT CO2 Eq., or 8 percent.
•	Grassland Remaining Grassland: Changes in Mineral and Organic Carbon Stocks (CC>2). The current
Inventory is the first reporting of biomass, dead wood and litter C stock changes for woodlands.
Recalculations for the soil C stock changes are associated with several improvements to both the Tier 2
and 3 approaches that are discussed in the Cropland Remaining Cropland section. As a result of these
improvements to the Inventory, C stocks decline on average across the time series for Grassland
Remaining Grassland, compared to an average increase in C stocks in the previous Inventory. The average
reduction in C stock change is 14.0 MMT CO2 Eq. over the time series, which is a 738 percent decrease in C
stock changes compared to the previous Inventory. This is largely driven by the methodological changes
associated with estimating soil C stock changes and to a lesser extent by the inclusion of biomass, dead
wood and litter C stock changes for woodlands.
•	Land Converted to Cropland: Changes in all Ecosystem Carbon Stocks (CO2). 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. Recalculations for the soil C stock changes are associated with several
improvements to both the Tier 2 and 3 approaches that are discussed in the Recalculations section of
Cropland Remaining Cropland. As a result of these improvements to the Inventory, Land Converted to
Cropland has a smaller reported loss of C compared to the previous Inventory, estimated at an average of
9-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
13.4 MMT CO2 Eq. over the time series. This represents a 19 percent decline in losses of C for Land
Converted to Cropland compared to the previous Inventory and is largely driven by the methodological
changes for estimating the soil C stock changes.
•	Settlements Remaining Settlements: Changes in Organic Soil Carbon Stocks (CO2). The entire time series
was recalculated based on updates to the land representation data with the release of the 2018 NRI
(USDA-NRCS 2018) and additional information from the National Land Cover Database (Yang et al. 2018;
Fry et al. 2011; Homer et al. 2007, 2015). In addition, the data splicing method has been used to re-
estimate CO2 emissions for 2016 to 2017 in the previous Inventory. However, the major change was the
correction of a quality control problem that led to an under-estimation of drained organic soils in
settlements. The recalculations led to an increase in emissions of 12.0 MMT CO2 Eq., or more than 6,500
percent, on average across the entire time series.
•	Land Converted to Forest Land: Changes in Carbon Stocks (CC>2). The Land Converted to Forest Land
estimates in this Inventory are based on the land use change information in the annual NFI. 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 8 percent in 2018 between the previous Inventory and the current
Inventory. This decrease is directly attributed to the incorporation of annual NFI data into the compilation
system and new data and methods used to compile estimates of C in mineral soils. 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. These changes resulted in an average annual increase in C stock of 9.8 MMT CO2 Eq. (8
percent) relative to the previous Inventory.
•	Fossil Fuel Combustion (CO2). The Energy Information Administration (EIA 2019) updated energy
consumption statistics across the time series relative to the previous Inventory. As a result of updated
liquid petroleum gas (LPG) heat contents, EIA updated LPG consumption in the residential, commercial,
industrial, and transportation sectors across the time series. EIA also revised sector allocations for
propane and total hydrocarbon gas liquids for 2010 through 2017, and for distillate fuel oil in 2017, which
impacted petroleum consumption by sector for those years. EIA also revised 2017 natural gas
consumption in all sectors. EIA revised assumptions for the percentage of fossil fuels consumed for non-
combustion use which impacted non-energy use sequestration statistics, particularly for petroleum coke
and residual fuel across the time series relative to the previous Inventory. These changes resulted in an
average annual decrease of 6.6 MMT CO2 Eq. (0.1 percent) in CO2 emissions from fossil fuel combustion
for the period 1990 through 2017, relative to the previous Inventory.
•	Substitution of Ozone Depleting Substances (HFCs). For the current Inventory, updates to the Vintaging
Model included renaming the non-metered dose inhaler (non-MDI) aerosol end-use to consumer aerosol
and updating stock and emission estimates to align with a recent national market characterization. In
addition, a technical aerosol end-use was added to the aerosols sector, in order to capture a portion of
the market that was not adequately encompassed by the former non-MDI aerosol end-use (EPA 2019b).
Within the Fire Protection sector, a correction was made to the lifetime for streaming agents, which was
changed from 18 years to 24 years. The polyurethane rigid spray foam end-use was divided into two end-
uses representing high pressure and low pressure two-component spray foam. Market size,and foam
blowing agent transition assumptions were adjusted to align with stakeholder input and market research.
Together, these updates increased greenhouse gas emissions an average of 3.3 percent across the
timeseries, relative to the previous Inventory.
Recalculations and Improvements 9-3

-------
Finally, in addition to the more significant methodological updates noted above, the Inventory includes new
categories not included in the previous Inventory that improve completeness of the national estimates.
Specifically, the current report includes fluorinated greenhouse gas emissions (HFCs, PFCs, SF6, and NF3) from the
Electronics Industry from manufacturing micro-electronic mechanical systems (MEMS) and photovoltaics (PV).401
Table 9-1: Revisions to U.S. Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Source
1990
2005
2014
2015
2016
2017
Average
Annual
Change
CO?
7.1
1.3
(10.4)
(10.5)
(14.4)
(17.1)
(1.2)
Fossil Fuel Combustion
1.2
(4.1)
(14.6)
(15.3)
(19.5)
(19.7)
(6.6)
Electric Power Sector
NC
NC
NC
NC
NC
+
+
Transportation
+
(0.9)
(7.9)
(8.7)
(13.7)
(13.3)
(2.5)
Industrial
(0.5)
(3.4)
(6.7)
(6.6)
(6.2)
(5.6)
(4.3)
Residential
+
+
+
+
0.3
(0.6)
+
Commercial
1.7
0.1
+
+
0.2
(0.1)
0.2
U.S. Territories
NC
+
NC
NC
NC
+
+
Non-Energy Use of Fuels
+
0.1
0.1
0.1
(0.1)
(0.1)
+
Natural Gas Systems
2.1
2.7
4.1
4.3
4.4
4.0
2.9
Cement Production
NC
NC
NC
NC
NC
NC
NC
Lime Production
NC
NC
NC
NC
(0.3)
(0.3)
+
Other Process Uses of Carbonates
NC
NC
NC
NC
(0.5)
(0.2)
+
Glass Production
NC
NC
NC
NC
+
+
+
Soda Ash Production
NC
NC
NC
NC
NC
NC
NC
Carbon Dioxide Consumption
NC
NC
NC
NC
NC
NC
NC
Incineration of Waste
+
+
+
+
0.2
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
3.1
1.9
(0.2)
0.1
1.3
(1.2)
1.8
Ferroalloy Production
NC
NC
NC
NC
NC
NC
NC
Ammonia Production
NC
NC
NC
NC
NC
NC
NC
Urea Consumption for Non-Agricultural Purposes
NC
NC
NC
NC
NC
(1.2)
+
Phosphoric Acid Production
NC
NC
+
NC
NC
+
+
Petrochemical Production
0.4
0.6
(0.2)
NC
0.2
0.7
0.4
Carbide Production and Consumption
NC
NC
NC
NC
NC
NC
NC
Lead Production
NC
NC
NC
NC
+
0.1
+
Zinc Production
NC
NC
NC
NC
NC
NC
NC
Petroleum Systems
0.7
0.6
0.9
1.0
0.8
1.1
0.8
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Magnesium Production and Processing
NC
NC
NC
NC
NC
NC
NC
Liming
NC
NC
NC
NC
(0.1)
(0.1)
+
Urea Fertilization
(0.4)
(0.4)
(0.6)
(0.6)
(0.8)
(0.5)
(0.5)
International Bunker Fuelsb
NC
NC
NC
NC
NC
NC
NC
Wood Biomass, Ethanol, and Biodiesel







Consumptiona
NC
NC
NC
NC
+
+
+
CH4c
(5.4)
(11.9)
(23.0)
(22.9)
(30.6)
(26.0)
(10.9)
Stationary Combustion
+
+
+
+
+
+
+
Mobile Combustion
+
+
0.1
0.1
0.1
0.1
+
401 This completeness improvement was phased so while these emissions are currently reported as an "Unspecified Mix of
HFCs, NFs, PFCs, and SF6," EPA anticipates being able to report the specific gases in future submissions.
9-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Coal Mining
Abandoned Underground Coal Mines
Natural Gas Systems
Petroleum Systems
Abandoned Oil and Gas Wells
Petrochemical Production
Carbide Production and Consumption
Iron and Steel Production & Metallurgical Coke
Production
Ferroalloy Production
Enteric Fermentation
Manure Management
Rice Cultivation
Field Burning of Agricultural Residues
Landfills
Wastewater T reatment
Composting
Incineration of Waste
International Bunker Fuelsb
N2Oc
Stationary Combustion
Mobile Combustion
AdipicAcid Production
Nitric Acid Production
Manure Management
Agricultural Soil Management
Field Burning of Agricultural Residues
Wastewater T reatment
N20 from Product Uses
Caprolactam, Glyoxal, and Glyoxylic Acid
Production
Incineration of Waste
Composting
Electronics Industry
Natural Gas Systems
Petroleum Systems
International Bunker Fuelsb
HFCs, PFCs, SF6and NF3
HFCs
Substitution of Ozone Depleting Substancesd
HCFC-22 Production
Electronics Industry
Magnesium Production and Processing
PFCs
Aluminum Production
Electronics Industry
Substitution of Ozone Depleting Substancesd
SF6
Electrical Transmission and Distribution
Electronics Industry
Magnesium Production and Processing
NF3
Electronics Industry
Unspecified Mix of HFCs, NF3, PFCs and SF6
Electronics Industry
NC
NC
NC
NC
NC
(0.9)
+
NC
NC
NC
NC
NC
NC
NC
(9.8)
(13.3)
(24.0)
(25.3)
(29.9)
(26.3)
(14.2)
4.0
2.1
1.4
1.1
0.8
1.0
3.5
+
+
+
+
+
0.1
+
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
+
+
+
+
+
+
(2.2)
(3.5)
(3.0)
(1.9)
(1.8)
(1.7)
+
1.3
2.7
3.9
(0.2)
1.4
1.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
NC
(0.1)
+
0.1
0.1
+
+
0.1
+
+
0.1
0.2
(0.1)
+
NC
NC
NC
NC
0.1
0.3
+
NC
NC
NC
NC
NC
NC
NC
NC
NC
+
+
+
+
+
64.3
56.9
86.5
69.7
61.6
60.7
57.1
+
+
+
+
+
+
+
+
(1.7)
(0.5)
(0.5)
(0.5)
(0.6)
(0.3)
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
(0.1)
(0.1)
(0.1)
(0.1)
+
(0.1)
64.2
58.5
86.9
70.3
62.2
61.0
57.3
0.1
0.1
0.1
0.1
0.1
0.1
0.1
+
+
+
+
+
+
+
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
(0.2)
(0.3)
0.1
+
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
0.1
0.3
+
NC
NC
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
NC
NC
+
+
+
+
+
+
6.4
15.9
16.3
15.3
13.9
6.6
(0.1)
6.3
15.7
16.6
15.6
14.2
6.3
(0.1)
6.3
15.7
16.6
15.6
14.2
6.3
NC
NC
+
NC
NC
+
+
NC
+
+
+
+
+
+
NC
NC
NC
NC
NC
NC
NC
NC
+
+
+
+
(0.1)
+
NC
NC
NC
NC
+
(0.1)
+
NC
+
+
+
+
(0.1)
+
NC
NC
+
+
+
+
+
0.1
+
0.2
(0.3)
(0.3)
(0.2)
+
0.1
+
0.2
(0.3)
(0.3)
(0.2)
+
NC
+
+
+
+
+
+
NC
NC
NC
NC
+
+
+
NC
+
+
+
+
+
+
NC
+
+
+
+
+
+
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
NC*
Recalculations and Improvements 9-5

-------
Net Emissions (Sources and Sinks)
Percentage change
19.6 (21.9)	16.0 (11.9) (34.5) (18.3) (8.7)
0.4% -0.3%	0.3% -0.2% -0.6% -0.3% -0.1%
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 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 from Peatlands Remaining Peatlands; CH4 and N20 emissions reported for Non-C02 Emissions from Forest
Fires, Non-C02 Emissions from 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.
e 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.
' The LULUCF Sector Net Total is the net sum of all CH4 and N20 emissions to the atmosphere plus net carbon stock changes.
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
2014
2015
2016
2017
Average
Annual
Change
Forest Land Remaining Forest Land
(63.3)
(39.6)
(50.9)
(31.5)
(31.4)
(15.9)
(48.2)
Changes in Forest Carbon Stocks3
(62.3)
(39.2)
(50.0)
(30.9)
(29.0)
(26.7)
(46.4)
Non-C02 Emissions from Forest Firesb
(1.0)
(0.4)
(0.8)
(0.6)
(2.4)
10.7
(1.8)
N20 Emissions from Forest Soilsc
NC
NC
NC
NC
NC
NC
NC
Non-C02 Emissions from Drained Organic







Soilsd
+
+
+
+
+
+
+
Land Converted to Forest Land
9.6
9.7
10.0
10.0
10.1
10.0
9.8
Changes in Forest Carbon Stocks6
9.6
9.7
10.0
10.0
10.1
10.0
9.8
Cropland Remaining Cropland
17.8
(2.5)
(0.2)
(6.5)
(12.8)
(12.0)
2.5
Changes in Mineral and Organic Soil







Carbon Stocks
17.8
(2.5)
(0.2)
(6.5)
(12.8)
(12.0)
2.5
Land Converted to Cropland
(21.5)
(12.8)
(10.1)
(9.5)
(11.9)
(11.2)
(13.4)
Changes in all Ecosystem Carbon Stocks'
(21.5)
(12.8)
(10.1)
(9.5)
(11.9)
(11.2)
(13.4)
Grassland Remaining Grassland
13.3
5.2
27.3
4.0
11.2
11.0
14.0
Changes in Mineral and Organic Soil







Carbon Stocks
13.3
5.2
27.3
4.0
11.2
11.0
14.0
Non-C02 Emissions from Grassland Fires8
NC
NC
NC
NC
NC
NC
NC
Land Converted to Grassland
(15.4)
(45.4)
(32.8)
(32.9)
(33.3)
(33.3)
(35.2)
Changes in all Ecosystem Carbon Stocks'
(15.4)
(45.4)
(32.8)
(32.9)
(33.3)
(33.3)
(35.2)
Wetlands Remaining Wetlands
+
+
+
+
+
+
+
Changes in Organic Soil Carbon Stocks in







Peatlands
NC
NC
NC
NC
NC
NC
NC
Changes in Aboveground and Soil Carbon







Stocks in Coastal Wetlands
+
+
+
+
+
+
+
CH4 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
NC
NC
NC
NC
NC
NC
NC
N20 Emissions from Coastal Wetlands







Remaining Coastal Wetlands
NC
NC
NC
NC
+
+
+
Non-C02 Emissions from Peatlands
NC
NC
NC
NC
NC
NC
NC
9-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Remaining Peatlands
Land Converted to Wetlands
Changes in Aboveground and Soil Carbon
Stocks
CH4 Emissions from Land Converted to
Coastal Wetlands
Settlements Remaining Settlements
Changes in Organic Soil Carbon Stocks
Changes in Settlement Tree Carbon
Stocks
Changes in Yard Trimming and Food Scrap
NC
13.1
11.2
(0.1)
NC
11.7
11.7
(0.6)
NC
8.9
13.8
NC
8.3
14.4
NC
8.7
14.7
NC
8.5
14.7
NC
11.6
12.0
(4.4) (6.0) (5.9) (5.9) (1.2)
Carbon Stocks in Landfills
1.5
+
(0.2)
0.2
0.2
(0.2)
0.5
N20 Emissions from Settlement Soilsh
0.6
0.6
(0.4)
(0.4)
(0.2)
(0.1)
0.2
Land Converted to Settlements
(0.1)
(1.0)
(5.2)
(6.3)
(7.0)
(6.9)
(1.8)
Changes in all Ecosystem Carbon Stocks'
(0.1)
(1.0)
(5.2)
(6.3)
(7.0)
(6.9)
(1.8)
LULUCF Emissions'
(0.4)
0.2
(1.2)
(0.9)
(2.7)
10.6
(1.6)
LULUCF Total Net Flux'
(46.0)
(74.9)
(51.8)
(63.6)
(63.7)
(60.5)
(59.1)
LULUCF Sector Totalk
(46.4)
(74.6)
(53.0)
(64.5)
(66.3)
(49.8)
(60.7)
Percent Change
-5.7%
-10.1%
-7.9%
-9.1%
-9.2%
-7.0%
-8.4%
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 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 emissions from drained organic soils on both Forest Land Remaining Forest Land and Land Converted to
Forest Land.
e 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.
g Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grass/and.
h Estimates include 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.
i 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 CH4 and N20 emissions to the atmosphere plus net carbon stock
changes.
Recalculations and Improvements 9-7

-------
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 (2019a) Electricity Generation. Monthly Energy Review, November 2019. Energy Information Administration,
U.S. Department of Energy, Washington, D.C. DOE/EIA-0035(2019/11).
EIA (2019b) Electricity in the United States. Electricity Explained. Energy Information Administration, U.S.
Department of Energy, Washington, D.C. Available online at:
.
EIA (2018) International Energy Statistics 1980-2018. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EPA (2019a) Acid Rain Program Dataset 1996-2018. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2019b) Greenhouse Gas Reporting Program (GHGRP). 2019 Envirofacts. Subpart HH: Municipal Solid Waste
Landfills and Subpart TT: Industrial Waste Landfills. Available online at:
.
EPA (2019c) "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 (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 2018) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Available online at:
.
IEA (2019) CO2 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 (2019a) Trends in Atmospheric Carbon Dioxide. Available online at:
. 19 December 2019.
NOAA/ESRL (2019b) Trends in Atmospheric Methane. Available online at:
. 19 December 2019.
NOAA/ESRL (2019c) Nitrous Oxide (N2O) 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:
. 19 December 2019.
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: 1990-2018

-------
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 (2019a) Trends in Atmospheric Carbon Dioxide. Available online at:
. October 7, 2019.
NOAA/ESRL (2019b) Trends in Atmospheric Methane. Available online at:
. October 7, 2019.
NOAA/ESRL (2019c) Nitrous Oxide (N2O) 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:
. October 7, 2019.
NOAA/ESRL (2019d) Sulfur Hexafluoride (SFe) 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:
. October 7, 2019.
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 (2020) Fuel Oil and Kerosene Sales. Energy Information Administration, U.S. Department of Energy,
Washington, D.C. February 2020.
EIA (2019) Monthly Energy Review, November 2019. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2018) "In 2017, U.S. electricity sales fell by the greatest amount since the recession" 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 (2019b) 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: .
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 (2019) Monthly Energy Review, November 2019, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/11).
IEA (2019) CO2 Emissions from Fossil Fuel Combustion - Overview. International Energy Agency. Available online
at: .
Carbon Dioxide Emissions from Fossil Fuel Combustion
AAR (2008 through 2018) Railroad Facts. Policy and Economics Department, Association of American Railroads,
Washington, D.C. Obtained from Clyde Crimmel at AAR.
AISI (2004 through 2018) Annual Statistical Report, American Iron and Steel Institute, Washington, D.C.
10-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2020) Table 1.1.6. Real Gross Domestic Product, Chained 2012 Dollars. Bureau of Economic Analysis (BEA),
U.S. Department of Commerce, Washington, D.C. February 2020. 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 (2019) Updated On-highway Cm and N2O 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 CHa and N2O Emission Factors for U.S. GHG Inventory. Technical
Memo, November 2018.
Browning, L (2017) Updated Methodology for Estimating CHa and N2O 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) CO2 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 (2019) 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 2017) 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 2018) 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) Quarterly Coal Report: July - September 2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0121.
EIA (2020b) Form EIA-923 detailed data with previous form data (EIA-906/920), Energy Information Administration,
U.S. Department of Energy. Washington, DC. DOE/EIA. March 2020.
EIA (2019a) Monthly Energy Review, November 2019, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/11).
EIA (2019b) "Natural gas prices, production, consumption, and exports increased in 2018." Today in Energy.
Available online at: < https://www.eia.gov/todayinenergy/detail.php?id=37892>.
References 10-5

-------
EIA (2019c) Electric Power Annual 2018. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. DOE/EIA-O348(17).
EIA (2019d) Natural Gas Annual 2018. Energy Information Administration, U.S. Department of Energy. Washington,
D.C. DOE/EIA-O131(17).
EIA (2019e) Annual Coal Report 2018. Energy Information Administration, U.S. Department of Energy. Washington,
D.C. DOE/EIA-0584.
EIA (2019f) 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 (2017) International Energy Statistics 1980-2016. 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 2006. Energy Information Administration, U.S. Department of
Energy. 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 (2020a) Acid Rain Program Dataset 1996-2018. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2020b) -The 2019 EPA Automotive Trends Report: Greenhouse Gas Emissions, Fuel Economy and Technology
since 1975. Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Available online at: <
https://www.epa.gov/automotive-trends>.
EPA (2018) MOtor Vehicle Emissions Simulator (MOVES) 2014b. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency, Washington, 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 CO2 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.
10-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
FAA (2019) Personal Communication between FAA and John Steller, Mausami Desai, and Vincent Camobreco for
aviation emissions estimates from the Aviation Environmental Design Tool (AEDT). December 2019.
FHWA (1996 through 2018) 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 2018) 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) (2020) 2019 Mineral Commodity Summary: Titanium and Titanium Dioxide.
U.S. Geological Survey, Reston, VA.
USGS (2019) 2019 Mineral Commodity Summaries: Aluminum. U.S. Geological Survey, Reston, VA.
USGS (2014 through 2019a) Mineral Industry Surveys: Silicon. U.S. Geological Survey, Reston, VA.
USGS (2014 through 2019b) Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA.
USGS (2014 through 2018) Minerals Yearbook: Nitrogen [Advance Release], Available online at:
.
USGS (1991 through 2017) Minerals Yearbook - Iron and Steel Scrap. U.S. Geological Survey, Reston, VA.
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.
References 10-7

-------
USGS (1995,1998, 2000, 2001, 2002, 2007) Minerals Yearbook: Aluminum Annual Report. U.S. Geological Survey,
Reston, VA.
Stationary Combustion (excluding C02)
EIA (2019) Monthly Energy Review, November 2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2019/11).
EIA (2017) International Energy Statistics 1980-2016. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EPA (2020) Acid Rain Program Dataset 1996-2018. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2018). 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 2018) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. 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.
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 2018) Railroad Facts. Policy and Economics Department, Association of American Railroads,
Washington, D.C. Obtained from Clyde Crimmel at AAR.
ANL (2018) The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model (GREET2018).
Argonne National Laboratory. Available online at: .
ANL (2006) Argonne National Laboratory (2006) GREET model Version 1.7. June 2006.
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.
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 (2019) Updated On-highway CH4 and N2O 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 Memo, October 2018.
10-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Browning, L (2018b) "Updated Non-Highway Cm and N2O Emission Factors for U.S. GHG Inventory." Technical
Memo, November 2018.
Browning, L (2017) "Updated Methodology for Estimating Cm and N2O Emissions from Highway Vehicle
Alternative Fuel Vehicles." Technical Memo, 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 (2019) 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 2017) Transportation Energy Data Book. Office of Transportation Technologies, Center for
Transportation Analysis, Energy Division, Oak Ridge National Laboratory. ORNL-6978.
DOT (1991 through 2018) Airline Fuel Cost and Consumption. U.S. Department of Transportation, Bureau of
Transportation Statistics, Washington, D.C. DAI-10. Available online at: .
EIA (2019a) Monthly Energy Review, November 2019, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/11).
EIA (2019f) Natural Gas Annual 2018. Energy Information Administration, U.S. Department of Energy, Washington,
D.C. DOE/EIA-0131(11).
EIA (1991 through 2018) 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:
.
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 (2019b) 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 (2019c) Confidential Engine Family Sales Data Submitted to EPA by Manufacturers. Office of Transportation
and Air Quality, U.S. Environmental Protection Agency.
EPA (2019d) Annual Certification Test Results Report. Office of Transportation and Air Quality, U.S. Environmental
Protection Agency. Available online at: .
References 10-9

-------
EPA (2018aJ Motor Vehicle Emissions Simulator (MOVES) 2014b. 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 (2019) Personal Communication between FAA and John Steller, Mausami Desai and Vincent Camobreco for
aviation emission estimates from the Aviation Environmental Design Tool (AEDT). January 2019.
FHWA (1996 through 2018) 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:
.
Gaffney, J. (2007) Email Communication. John Gaffney, American PublicTransportation 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 EFs for 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.
10-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
ICF (2017b) Updated Non-Highway Cm and N2O 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 2018) Raillnc Short line and Regional Traffic Index. Carloads Originated Year-to-Date.
December 2019. Available online at: < https://www.railinc.com/rportal/railinc-indexes>.
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 (2019a) "Guide to the Business of Chemistry, 2019," American Chemistry Council.
ACC (2019b) "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."
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 (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: <
https://www.bankofcanada.ca/rates/exchange/annual-average-exchange-rates/>.
References 10-11

-------
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 (2019) Monthly Energy Review, November 2019. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035 (2019/11).
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.
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 (2019a) "Criteria pollutants National Tier 1 for 1970 - 2018." National Emissions Inventory (NEI) Air Pollutant
Emissions Trends Data. Office of Air Quality Planning and Standards, May 2019. Available online at:
.
EPA (2019b) 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:
.
10-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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: .
References 10-13

-------
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:
.
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 por Subsector, 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.
10-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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:
http://www.rma.org/scrap_tires/scrap_tire_markets/scrap_tire_characteristics/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 (1990-2018) "Interactive Tariff and Trade DataWeb: Quick Query." Available
online at: . Accessed September 2019.
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 (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 (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:
.
References 10-15

-------
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.
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 (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: < https://www.ustires.org/sites/default/files/MAR_027_USTMA.pdf >.
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.
10-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2019) DGO Data Information System. Department of Mines, Minerals and Energy of Virginia. Available
online at .
EIA (2019) Annual Coal Report 2018. Table 1. Energy Information Administration, U.S. Department of Energy.
El Paso (2009) Shoal Creek Mine Plan, El Paso Exploration & Production.
EPA (2019) Greenhouse Gas Reporting Program (GHGRP) 2018 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 (2019). Correspondence between ERG and Buchanan Mine.
Geological Survey of Alabama State Oil and Gas Board (GSA) (2019) Well Records Database. Available online at
.
IEA (2019) 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 (2019) Marshall County VAM Abatement Project Offset Verification Statement submitted to
California Air Resources Board, July 2019.
MSHA (2019) 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) (2019) Oil & Gas Production Data. Available online at
.
Abandoned Underground Coal Ivinies
EPA (2004) Methane Emissions Estimates & Methodology for Abandoned Coal Mines in the U.S. Draft Final Report.
Washington, D.C. April 2004.
References 10-17

-------
MSHA (2019) 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.
Enverus Drillinglnfo (2019) March 2019 Download. Dl Desktop" Enverus Drillinglnfo, Inc.
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 (2019) Greenhouse Gas Reporting Program. U.S. Environmental Protection Agency. Data reported as of August
4, 2019.
EPA (2020) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2018: Updates for Offshore Production
Emissions (Offshore Production memo). U.S. Environmental Protection Agency. April 2020. Available at:
https://www.epa.gov/ghgemissions/natural-gas-and-petroleum-systems.
EPA/GRI (1996) Methane Emissions from the Natural Gas Industry. Prepared by Radian. U.S. Environmental
Protection Agency. April 1996.
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.
Natural Gas Systems
Enverus Drillinglnfo (2019) March 2019 Download. Dl Desktop" Enverus Drillinglnfo, Inc.
EPA (2019) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2018: Updates Under Consideration for
Natural Gas Gathering & Boosting Station Emissions (G&B Station memo). U.S. Environmental Protection Agency.
September 2019. Available at: https://www.epa.gov/ghgemissions/natural-gas-and-petroleum-systems.
E PA (2019) Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2018: Updates Under Consideration for
Offshore Production Emissions (Offshore Production memo). U.S. Environmental Protection Agency. September
2019. Available at: https://www.epa.gov/ghgemissions/natural-gas-and-petroleum-systems.
EPA (2019) Greenhouse Gas Reporting Program- Subpart W-Petroleum and Natural Gas Systems. Environmental
Protection Agency. Data reported as of August 4, 2019.
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.
GTI (2001) Gas Resource Database: Unconventional Natural Gas and Gas Composition Databases. Second Edition.
GRI-01/0136.
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.
10-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
PHMSA (2019) Gas Distribution Annual Data. Pipeline and Hazardous Materials Safety Administration, U.S.
Department of Transportation, Washington, DC. Available online at: .
Zimmerle et al. (2019) "Characterization of Methane Emissions from Gathering Compressor Stations." October
2019. Available at https://mountainscholar.org/handle/10217/195489.
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.
Abandoned 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." http://www.geology.ar.gov/pdf/IC-10%20SUPPLEMENT_v.pdf.
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 Drillinglnfo (2019) March 2019 Download. Dl Desktop" Enverus Drillinglnfo, 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:

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: .
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:
.
References 10-19

-------
Energy Sources of 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), NASA Technical 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 (2019) 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.
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 (2019) Monthly Energy Review, November 2019, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/11).
FAA (2019) Personal Communication between FAA and Vince Camobreco for aviation emission estimates from the
Aviation Environmental Design Tool (AEDT). December 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.
USAF (1998) Fuel Logistics Planning. U.S. Air Force pamphlet AFPAM23-221, May 1,1998.
Wood Biomass and Biofuel Consumption
EIA (2019a) Monthly Energy Review, November 2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2019/11).
EIA (2019b) Monthly Biodiesel Production Report. October 2019. Energy Information Administration, U.S.
Department of Energy. Washington, D.C.
EPA (2020) Acid Rain Program Dataset 1996-2018. 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.
10-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2018) Aggregation of Reported Facility Level Data under Subpart H -
National Level Clinker Production from Cement Production for Calendar Years 2014, 2015, 2016, and 2017. 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) (2019) Mineral Commodity Summaries: Cement 2019. U.S. Geological
Survey, Reston, VA. February 2019. Available at: .
USGS (1995 through 2014) Minerals Yearbook - Cement. U.S. Geological Survey, Reston, VA.
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 (2018) Greenhouse Gas Reporting Program (GHGRP). Aggregation of Reported Facility Level Data under
Subpart S-National Lime Production for Calendar Years 2010 through 2017. 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.
References 10-21

-------
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.
United States Geological Survey (USGS 2020) 2020 Mineral Commodities Summary: Lime. U.S. Geological Survey,
Reston, VA (February 2020).
USGS (2019) 2019 Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (February 2019).
USGS (2018) (1992 through 2016) Minerals Yearbook: Lime. U.S. Geological Survey, Reston, VA (August 2019).
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
.
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 (2019) Mineral Industry Surveys: Soda Ash in December 2018. U.S. Geological Survey, Reston, VA. Accessed
September 24, 2019.
USGS (2018) Mineral Industry Surveys: Soda Ash in February 2018. U.S. Geological Survey, Reston, VA. Accessed
September 2018.
USGS (1995 through 2015a) 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 (2016a) Minerals Yearbook: Crushed Stone Annual Report: Advance Data Release of the 2016 Annual Tables.
U.S. Geological Survey, Reston, VA. November 2018.
Willett (2019a) Personal communication, Jason Willett, U.S. Geological Survey and John Steller, U.S. Environmental
Protection Agency. September 5, 2019.
Willett (2018a) Personal communication, Jason Christopher Willett, U.S. Geological Survey and John Steller, U.S.
Environmental Protection Agency. January 4, 2018.
Willett (2018b) Personal communication, Jason Christopher Willett, U.S. Geological Survey and John Steller, U.S.
Environmental Protection Agency. December 4, 2018.
10-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Other Process Uses of Carbonates
AISI (2018 through 2019) 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) (2013) Magnesium Metal Mineral Commodity Summary for 2013. U.S.
Geological Survey, Reston, VA.
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 (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 2012) Minerals Yearbook: Magnesium Annual Report. U.S. Geological Survey, Reston, VA.
Willett (2017a) Personal communication, Jason Christopher Willett, U.S. Geological Survey and Mausami Desai and
John Steller, U.S. Environmental Protection Agency. March 9, 2017.
Willett (2018a) Personal communication, Jason Christopher Willett, U.S. Geological Survey and John Steller, U.S.
Environmental Protection Agency. January 4, 2018.
Willett (2018b) Personal communication, Jason Christopher Willett, U.S. Geological Survey and John Steller, U.S.
Environmental Protection Agency. December 4, 2018.
Willett (2019) Personal communication, Jason Christopher Willett, U.S. Geological Survey and John Steller, U.S.
Environmental Protection Agency. September 5, 2019.
Ammonia Production
ACC (2019) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
Bark (2004) CoffeyvilleNitrogen 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:
.
References 10-23

-------
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:
.
CVR (2016)
CVR
Energy,
Inc.
2016 Annual Report.
Available
online
at:
.
CVR (2017)
CVR
Energy,
Inc.
2017 Annual Report.
Available
online
at:
.
CVR (2018)
CVR
Energy,
Inc.
2018 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.
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: 2010 Summary.
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.
10-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
United States Geological Survey (USGS) (2019) 2019 Mineral Commodity Summaries: Nitrogen (Fixed) - Ammonia.
February 2018. Available online at: .
USGS (2018a) 2016 Minerals Yearbook: Nitrogen [Advance Release], August 2018. Available online at:
.
USGS (2018b) Minerals Commodity Summaries: Nitrogen (Fixed)-Ammonia. Available online at:
.
USGS (2017) 2015 Minerals Yearbook: Nitrogen [Advance Release], August 2017. Available online at:
.
USGS (2016) 2014 Minerals Yearbook: Nitrogen [Advance Release], October 2016. Available online at:
.
USGS (2015) 2013 Minerals Yearbook: Nitrogen [Advance Release], August 2015. Available online at:
.
USGS (2014) 2012 Minerals Yearbook: Nitrogen [Advance Release], September 2014. 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.
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.
TFI (2002) U.S. Nitrogen Imports/Exports Table. The Fertilizer Institute. Available online at:
. August 2002.
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.
USGS (1994 through 2019a) Minerals Yearbook: Nitrogen. Available online at:
.
USGS (2019b) Minerals Commodity Summaries: Nitrogen (Fixed)-Ammonia. Available online at:
.
References 10-25

-------
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 (2018) Greenhouse Gas Reporting Program (GHGRP). Aggregation of Reported Facility Level Data under
Subpart V -National Nitric Acid Production for Calendar Years 2010 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
https://www.epa.gov/sites/production/files/2015-07/documents/ghgrp_verification_factsheet.pdf>.
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: .
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 (2019) 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.
10-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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) 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
https://www.epa.gov/sites/production/files/2015-07/documents/ghgrp_verification_factsheet.pdf>.
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 N2O 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 ai,
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.
Caprolactam, Glyoxal and Glyoxylic Acid Production
ACC (2019) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
AdvanSix (2018). AdvanSix Hopewell Virginia Information Sheet. Retrieved from:
https://www.advan6.com/hopewell/ on February 5, 2020.
BASF (2018). BASF: Freeport, Texas Fact Sheet. Retrieved from https://www.basf.com/documents/corp/en/about-
us/strategy-and-organization/verbund/BASF_Freeport.pdf on February 5, 2020.
Fibrant (2018). Fibrant LLC Contact Page. Retrieved from: http://www.fibrant52.com/en/contact on February 5,
2020.
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.
References 10-27

-------
TechSci n.d. (2017). Fibrant B.V. to Discontinue Caprolactam Plant in the United States. Retrieved from:
https://www.techsciresearch.com/news/1356-fibrant-b-v-to-discontinue-caprolactam-plant-in-the-united-
states.html.
Carbide Production and Consumption
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 (2005 through 2019) USITC Trade DataWeb. Available online at: .
United States Geological Survey (2018a) 2016 Minerals Yearbook: Abrasives, Manufactured [Advance Release], 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: https://www.usgs.gov/centers/nmic/manufactured-abrasives-
statistics-and-information.
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:
https://prd-wret.s3-us-west-2.amazonaws.com/assets/palladium/production/s3fs-public/atoms/files/mcs-2019-
abras.pdf.
USGS (1991a through 2015) 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.
IPCC (2006) 2006 IPCC 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 (2020) Mineral Commodity Summary: Titanium and Titanium Dioxide. U.S.
Geological Survey, Reston, Va. January 2020. Available online at: <
https://pubs.usgs.gov/periodicals/mcs2020/mcs2020-titanium.pdf>.
USGS (1991 through 2015) Minerals Yearbook: Titanium. U.S. Geological Survey, Reston, VA.
Soda Ash 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) (2019) Mineral Industry Surveys: Soda Ash in April 2019. U.S. Geological
Survey, Reston, VA. Accessed August 2019.
USGS (2018a) Mineral Commodity Summary: Soda Ash. U.S. Geological Survey, Reston, VA. Accessed August 2019.
10-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
USGS (2018b) 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, 2018c) 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 (2019) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
ACC (2014a) U.S. Chemical Industry Statistical Handbook. American Chemistry Council, Arlington, VA.
ACC (2014b) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
ACC (2002, 2003, 2005 through 2011) Guide to the Business of Chemistry. American Chemistry Council, Arlington,
VA.
AN (2014) About Acrylonitrile: Production. AN Group, Washington, D.C. Available online at:
.
EPA Greenhouse Gas Reporting Program (2019) Aggregation of Reported Facility Level Data under Subpart X -
National Petrochemical Production for Calendar Years 2014 through 2018. Office of Air and Radiation, Office of
Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
EPA Greenhouse Gas Reporting Program (2017) Aggregation of Reported Facility Level Data under Subpart X -
National Petrochemical Production for Calendar Years 2010 through 2013. 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
https://www.epa.gov/sites/production/files/2015-07/documents/ghgrp_verification_factsheet.pdf>.
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.
References 10-29

-------
HCFC-22 Production
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
https://www.epa.gov/sites/production/files/2015-07/documents/ghgrp_verification_factsheet.pdf>.
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.
10-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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: .
Carbc oxide Consumption
ARI (1990 through 2010) CO2 Use in Enhanced Oil Recovery. Deliverable to ICF International under Task Order 102,
July 15, 2011.
ARI (2007) CO2-EOR: 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) CO2-EOR: 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 Pestrusak, ICF International. September 5, 2003.
COGCC (2014) Monthly CO2 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 (2019). Aggregation of Reported Facility Level Data under Subpart PP -
National Level CO2 Transferred for Food & Beverage Applications for Calendar Years 2010 through 2018. 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) 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.
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 CO2. Available online at:
.
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.
References 10-31

-------
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 March 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 2018) Annual Statistical Report, American Iron and Steel
Institute, Washington, D.C.
AISI (2006 through 2017) Personal communication, Mausami Desai, U.S. EPA, and American Iron and Steel
Institute, December 2017.
AISI (2008) Personal communication, Mausami Desai, U.S. EPA, and Bruce Steiner, Technical Consultant with the
American Iron and Steel Institute, October 2008.
Carroll (2016) Personal communication, Mausami Desai, U.S. EPA, 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. EPA, 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 2018) Quarterly Coal Report: October-December, Energy Information Administration, U.S.
Department of Energy. Washington, D.C. DOE/EIA-0121.
EIA (2016b) Natural Gas Annual 2016. Energy Information Administration, U.S. Department of Energy. Washington,
D.C. DOE/EIA-0131(06).
EIA (2017c) Monthly Energy Review, December 2017, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2015/12).
10-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
EIA (2016c) Monthly Energy Review, December 2016, 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.
Fenton (2015 through 2018) Personal communication. Michael Fenton, Commodity Specialist, U.S. Geological
Survey and Marty Wolf, Eastern Research Group. September 16, 2015.
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.
USGS (2019) 2019 USGS Minerals Yearbook - Iron and Steel. U.S. Geological Survey, Reston, VA.
USGS (2018) 2018 USGS Minerals Yearbook - Iron and Steel. U.S. Geological Survey, Reston, VA.
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) (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.
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.
References 10-33

-------
Aluminum Production
EPA (2019) Greenhouse Gas Reporting Program (GHGRP). Envirofacts, Subpart: F Aluminum Production. Available
online at: .
EPA (2015). Greenhouse Gas Reporting Program Report Verification. 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.
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.
USGS (2019a) 2017 Mineral Yearbook: Aluminum. U.S. Geological Survey, Reston, VA.
USGS (2019b) 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 casting." Journal of Cleaner Production, 15: 979-987, March.
10-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
EPA (2019) Envirofacts. Greenhouse Gas Reporting Program (GHGRP), Subpart T: Magnesium Production and
Processing. Available online at: . Accessed on October 2018.
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 (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) CO2 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.
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.
References 10-35

-------
Zinc Production
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 (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:
http://www.nyrstar.eom/~/media/Files/N/Nyrstar/operations/melting/fact-sheet-clarksville-en.pdf>. 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.
Steel Dust Recycling (SDR) (2017) Personal communication. Jeremy Whitten, EHS Manager, Steel Dust Recycling
LLC and John Steller, U.S. Environmental Protection Agency. January 26, 2017.
10-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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) CO2 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.
USGS (1995 through 2014) Minerals Yearbook: Zinc Annual Report. U.S. Geological Survey, Reston, VA.
Viklund-White (2000) The use ofLCAforthe 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.
Doering, R. and Nishi, Y (2000) "Handbook of Semiconductor Manufacturing Technology", Marcel Dekker, New
York, USA, 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.
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. < https://fas.org/sgp/crs/misc/R42509.pdf>
SEMI - Semiconductor Equipment and Materials Industry (2018) World Fab Forecast, June 2018 Edition.
References 10-37

-------
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:
.
United States Census Bureau (USCB) (2011, 2012, 2015, 2016, 2017, 2018) Historical Data: Quarterly Survey of
Plant Capacity Utilization. Available online at: < https://www.census.gov/programs-surveys/qpc.html>.
U.S. 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.
U.S. EPA Greenhouse Gas Reporting Program (GHGRP) Envirofacts. Subpart I: Electronics Manufacture. 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 (2020). Proposed Updates to the Spray Foam 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. March 19, 2020.
EPA (2019a) Suppliers of Industrial GHGs and Products Containing GHGs. Greenhouse Gas Reporting Program.
Available online at: .
EPA (2019b) Proposed Updates to the Non -MDI Aerosol 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. October 3, 2019.
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.
10-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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)
SFs 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 SFe 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,
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.
Ottinger (2014) Personal communication. Deborah Ottinger (CCD, U.S. EPA) and Mausami Desai (U.S. EPA). Email
received on January 29, 2014.
Tupman, M. (2003) Personal communication. Martin Tupman, Airgas Nitrous Oxide and Daniel Lieberman, ICF
International. August 8, 2003.
Industrial Processes and Product Use Soi	cursor
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) 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.
References 10-39

-------
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
ic 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.
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.
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.
10-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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: .
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
.
References 10-41

-------
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
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.
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.
10-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (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.
USDA (2019a) Quick Stats: Agricultural Statistics Database. National Agriculture Statistics Service, U.S. Department
of Agriculture. Washington, D.C. Available online at: .
References 10-43

-------
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: <
https://www.nass.usda.gov/AgCensus/index.php>. 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:
.
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:
.
10-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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:
.
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:
.
References 10-45

-------
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
.
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:
.
tivation
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 N2O 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.
10-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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, FL and 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." Soil 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.
References 10-47

-------
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) Cm 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.
TAMU (2015) Texas Rice Crop Survey. Texas A&M AgriLIFE 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.
10-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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(1): 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.
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:
.
References 10-49

-------
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 N2O 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 N2O Emissions from U.S.
Cropland Soils." Global Biogeochemical Cycles, 24, GB1009, doi:10.1029/2009GB003544.
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 N2O-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.
10-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Firestone, M. K., and E.A. Davidson, Ed. (1989) Microbiological basis of NO and N2O 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 2006IPCC 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., L.A. 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.
NEBRA (2007) A National Biosolids Regulation, Quality, End Use & Disposal Survey. North East Biosolids and
Residuals Association, July 21, 2007.
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: .
References 10-51

-------
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 www.nass.usda.gov/AgCensus.
USDA-NASS (2012) 2012 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at www.nass.usda.gov/AgCensus.
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,
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/stelprdbl042093.pdf
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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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.
USFS (2019) Forest Inventory and Analysis Program. United States Department of Agriculture, U.S. Forest Service,
https://www.fia.fs.fed.us/tools-data/default.asp.
Van Buuren, S. (2012) "Flexible imputation of missing data." Chapman & Hall/CRC, Boca Raton, FL.
Wagner-Riddle, C., Congreves, K. A., Abalos, D., Berg, A. A., Brown, S. E., Ambadan, J. T., Gao, X. & Tenuta, M.
(2017) "Globally important nitrous oxide emissions from croplands induced by freeze-thaw cycles." Nature
Geosciences 10(4): 279-283.
Wisconsin Department of Natural Resources (1993) Wisconsin Greenhouse Gas Emissions: Estimates for 1990.
Bureau of Air Management, Wisconsin Department of Natural Resources, Madison, Wl.
10-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
Liming
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.
Tepordei, V.V. (1997 through 2015) "Crushed Stone," In Minerals Yearbook. U.S. Department of the Interior/U.S.
Geological Survey. Washington, D.C. Available online at: .
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.
USGS (2019) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2019, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2018) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2018, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2017) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2017, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2016) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2016, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2015) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2015, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2014) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2014, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2013) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2013, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2012) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2012, U.S.
Geological Survey, Reston, VA. Available online at:
.
References 10-53

-------
USGS (2011) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2011, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2010) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2010, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2009) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2009, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2008) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2008, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2007) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of 2007. 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. (2016) "Crushed Stone," In Minerals Yearbook. U.S. Department of the Interior/U.S. Geological Survey.
Washington, D.C. Available online at: . Accessed: 30 August 2017.
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. (2018) Personal communication. Jason Willett. Preliminary data tables from "Crushed Stone," In 2016
Minerals Yearbook. U.S. Department of the Interior/U.S. Geological Survey. Washington, D.C. November 16, 2018.
Willett, J.C. (2017) Personal communication. Jason Willett. Preliminary data tables from "Crushed Stone," In 2015
Minerals Yearbook. U.S. Department of the Interior/U.S. Geological Survey. Washington, D.C. August 31, 2017.
Willett, J.C. and Thompson, D.V. (2017) Crushed stone and sand and gravel in the second quarter 2015: U.S.
Geological Survey Mineral Industry Surveys, . Accessed: 30 August 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.
10-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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 ofGuelph.
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.
References 10-55

-------
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.
De Pinheiro Henriques, A.R., and Marcelis, L.F.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.
EPA (1994) International Anthropogenic Methane Emissions: Estimates for 1990, Report to Congress. EPA 230-R-93-
010. Office of Policy Planning and Evaluation, U.S. Environmental Protection Agency, Washington, D.C.
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,
http://ag.arizona.edu/pubs/crops/azll70/.
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.
10-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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, LB., 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.
References 10-57

-------
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." / Plant Nutr. 20:531-548.
Pettigrew, W.T., Meredith, W.R., Jr., and Young, LD. (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.
Sadras, V.O., and Wilson, LJ. (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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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.
10-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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.
References 10-59

-------
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.
U.S. Census Bureau (2010) Topological^ 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. http://srsfi
a2.fs.fed.us/php/tpo_2009/tpo_rpa_intl.php. 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.
10-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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., 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.
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,
References 10-61

-------
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.
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.; Pugh, S.A. (2014) Forest Resources of the United States, 2012. Gen. Tech.
Rep. WO-91. Washington, D.C. U.S. Department of Agriculture, Forest Service, Washington Office. 218 p.
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.
10-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2018a) Forest Inventory and Analysis National Program: Program Features. U.S. Department
of Agriculture Forest Service. Washington, D.C. Available online at: .
Accessed 1 November 2018.
USDA Forest Service. (2018b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 1 November 2018.
USDA Forest Service. (2018c) 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 1 November 2018.
USDA Forest Service (2018d) Forest Inventory and Analysis National Program, FIA library: Database
Documentation. U.S. Department of Agriculture, Forest Service, Washington Office. Available online at:
. Accessed on 1 November 2018.
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., LS. 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:
.
References 10-63

-------
Forest Land Remaining Forest Land: Non-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.
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: http://mtbs.gov/direct-download
[06Aug2018],
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 (2018b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 1 November 2018.
USDA Forest Service (2018d) Forest Inventory and Analysis National Program, FIA library: Database
Documentation. U.S. Department of Agriculture, Forest Service, Washington Office. Available online at:
. Accessed on 1 November 2018.
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.
10-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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 (2018) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, DC; 2015. Available online at . Accessed 1 November 2018.
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.
References 10-65

-------
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 (2018b) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, D.C. Available online at: . Accessed on 1 November 2018.
USDA Forest Service (2018c) 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 1 November 2018.
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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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., LS. 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.
10-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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
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.
References 10-67

-------
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., L.A. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURYSoil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSRTech. 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.
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 CO2 emissions." Soil Use and Management 13:
230-244.
10-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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 www.nass.usda.gov/AgCensus.
USDA-NASS (2012) 2012 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at www.nass.usda.gov/AgCensus.
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,
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/stelprdbl042093.pdf.
References 10-69

-------
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
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.
10-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 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 CO2 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., L.A. 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.
References 10-71

-------
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., LS. 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.
10-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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., L.A. 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.
NEBRA (2007) A National Biosolids Regulation, Quality, End Use & Disposal Survey. North East Biosolids and
Residuals Association. July 21, 2007.
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 (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.
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 10-73

-------
Grassland Remaining Grassland: Non-C02 Emissions from
Grassland 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.
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.
10-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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 CO2 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.
References 10-75

-------
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., L.A. 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., 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.
.
10-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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., LS. 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.
Wetlands Remaining Wetlands: C02, CH4, and N20 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 correspondence. 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 (1991-2015) Minerals Yearbook: Peat (1994-2014). United States Geological Survey, Reston, VA. Available
online at < http://minerals.usgs.gOv/minerals/pubs/commodity/peat/index.html#myb >.
USGS (2016) Mineral Commodity Summaries: Peat (2016). United States Geological Survey, Reston, VA. Available
online at .
Wetlands Remaining Coastal Wetlands: Emissions and
Remova	stal 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
References 10-77

-------
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, https://doi.org/10.5066/F77943K8.
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.
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) N2O 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,
10-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2013) 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 (2018) Fisheries of the United States, 2017. U.S. Department of Commerce,
NOAA Current Fishery Statistics No. 2017.
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.
M., Parker, V. T.,... and Castaneda-Moya, E. (2017) Biomass/Remote Sensing dataset: 30m resolution tidal marsh
References 10-79

-------
biomass samples and remote sensing data for six regions in the conterminous United States: U.S. Geological Survey
data release, https://doi.org/10.5066/F77943K8.
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.
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 (2003) Good Practice Guidance for Land Use, Land-Use Change and Forestry. LUCF Sector Good Practice
Guidance, Chapter 3. Jim Penman, Michael Gytarsky, Taka Hiraishi, Thelma Krug, Dina Kruger, Riitta Pipatti,
Leandro Buendia, Kyoko Miwa, Todd Ngara, Kiyoto Tanabe and Fabian 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.
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.
10-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Dissertation, University of Maryland, College Park, MD, 342pp.
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-
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.
References 10-81

-------
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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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.
Frelich, L.E. (1992) Predicting Dimensional Relationships for Twin Cities Shade Trees. University of Minnesota,
Department of Forest Resources, St. Paul, MN, p. 33.
Fleming, L.E. (1988) Growth Estimation of Street Trees in Central New Jersey. M.S. thesis, Rutgers University, New
Brunswick, NJ.
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.
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.
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.
10-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
Nowak, D.J., R.E. Hoehn, D.E. Crane, J.C. Stevens, J.T. Walton, and J. Bond (2008) Aground-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: N20 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,
https://doi.org/10.5066/F7H41PKX.
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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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.
References 10-83

-------
EPA (2018) Advancing Sustainable Materials Management: Facts and Figures 2015. 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) 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 (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.
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.
10-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2018) Forest Inventory and Analysis National Program: FIA Data Mart. U.S. Department of
Agriculture Forest Service. Washington, DC; 2015. Available online at . Accessed 1 November 2018.
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.
https://www.nrcs.usda.gov/lnternet/FSE_DOCUMENTS/nrcseprd 1422028.pdf.
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. (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.
References 10-85

-------
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:
.
Czepiel, P., B. Mosher, P. Crill, and R. Harriss (1996) "Quantifying the Effect of Oxidation on Landfill Methane
Emissions." Journal of Geophysical Research, 101(D11):16721-16730.
EIA (2007) Voluntary Greenhouse Gas Reports for EIA Form 1605B (Reporting Year 2006). Available online at:
.
EPA (2019a) Landfill Methane Outreach Program (LMOP). 2019 Landfill and Project Level Data. September 2019.
Available online at: < https://www.epa.gov/lmop/landfill-gas-energy-project-data>.
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 (2016) Industrial and Construction and Demolition Landfills. Available online at:
https://www.epa.gov/landfills/industrial-and-construction-and-demolition-cd-landfills.
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 (2019) Draft Production Data Supplied by ERG for 1990-2018 for Pulp and Paper, Fruits and Vegetables, and
Meat. August.
10-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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.
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.
U.S. Census Bureau (2019) Annual Estimates of the Resident Population: April 1, 2010 to July 1, 2018. Available
online at
.
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: <
https://www.wto.org/english/news_e/newsl7_e/impl_03octl7_e.htm>.
Wastewater Treatment
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.
Ahn et al. (2010) N20 Emissions from Activated Sludge Processes, 2008-2009: Results of a National Monitoring
Survey in the United States. Environ. Sci. Technol. 44: 4505-4511.
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.
References 10-87

-------
BIER (2017) Beverage Industry Environmental Roundtable. 2016 Trends and Observations. Available online at:
. Accessed April 2018.
Brewers Association (2019) Statistics: Number of Breweries. Available online at:
. Accessed July 2019.
Brewers Association (2017). 2016 Sustainability Benchmarking Update. Available online at:
. Accessed
April 2018.
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.
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: < http://www.cast-
science.org/download.cfm?PublicationlD=2889&File=70E92280D92EC9AlEED60A5AA8D2734E.cfusion>.
Climate Action Reserve (CAR) (2011) Landfill Project Protocol V4.0, June 2011. Available online at:
.
Chandran, K. (2012) Greenhouse Nitrogen Emissions from Wastewater Treatment Operation Phase I: Molecular
Level Through Whole Reactor Level Characterization. WERF Report U4R07.
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 (2019) Energy Information Administration. U.S. Refinery and Blender Net Production of Crude Oil and
Petroleum Products (Thousand Barrels). Available online at:
. Accessed September 2019.
EPA (2020) U.S. Environmental Protection Agency. Frequent Questions on Biosolids. February 2020. Available
online at: . Accessed March 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 (2008a) U.S. Environmental Protection Agency. Municipal Nutrient Removal Technologies Reference
Document: Volume 2 - Appendices. U.S. Environmental Protection Agency, Office of Wastewater Management.
Washington, D.C.
EPA (2008b) 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:
10-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
. Accessed December
2015.
EPA (2004) 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 (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 (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 (2018a) Updates to Domestic Wastewater BOD Generation per Capita. August 2018.
ERG (2018b) Inclusion of Wastewater Treatment Emissions from Breweries. July 2018.
References 10-89

-------
ERG (2016) Revised Memorandum: Recommended Improvements to the 1990-2015 Wastewater Greenhouse Gas
Inventory. November 2016.
ERG (2013a) Revisions to Pulp and Paper Wastewater Inventory. October 2013.
ERG (2013b) Revisions to the Petroleum Refinery Wastewater Inventory. October 2013.
ERG (2008) Planned Revisions of the Industrial Wastewater Inventory Emission Estimates for the 1990-2007
Inventory. August 10, 2008.
ERG (2006) Memorandum: Assessment of Greenhouse Gas Emissions from Wastewater Treatment of U.S. Ethanol
Production Wastewaters. Prepared for Melissa Weitz, EPA. 10 October 2006.
FAO (2019a) FAOSTAT-Forestry Database. Available online at:
. Accessed May 2019.
FAO (2019b) "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:< http://www.fao.org/forestry/statistics/81757/en/> Accessed June
2019.
FAO (2019c) FAOSTAT-Food Balance Sheets. Available online at:
. Accessed June 2019.
Great Lakes-Upper Mississippi River Board of State and Provincial Public Health and Environmental Managers.
(2004) Recommended Standards for Wastewater Facilities (Ten-State Standards).
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.
Leverenz, H.L., G. Tchobanoglous, and J.L. Darby (2010) "Evaluation of Greenhouse Gas Emissions from Septic
Systems." Water Environment Research Foundation. Alexandria, VA.
Lewis, A. (2019). Email correspondence. Ann Lewis, RFA to Kara Edquist, ERG. "Wet Mill vs Dry Mill Ethanol
Production." August 20, 2019.
Malmberg, B. (2019) Email correspondence. Barry Malmberg, NCASI to Kara Edquist, ERG. "Question on
Wastewater Inventory Pulp and Paper production data." July 1, 2019.
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.
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.
10-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
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 (2019a). Renewable Fuels Association. Annual U.S. Fuel Ethanol Production. Available online at:
. Accessed May 2019.
RFA (2019b). Renewable Fuels Association. Monthly Grain Use for U.S. Ethanol Production Report. Available online
at: . Accessed May 2019.
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.
Scheehle, E.A., and Doom, M.R. (2001) "Improvements to the U.S. Wastewater Methane and Nitrous Oxide
Emissions Estimate." July 2001.
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 (2019) Alcohol and Tobacco Tax and Trade Bureau. Beer Statistics. Available online at:
. Accessed May 2019.
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 (2019) International Database. Available online at:
. Accessed June 2019.
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 June 2019.
USDA (2019a) U.S. Department of Agriculture. National Agricultural Statistics Service. Washington, D.C. Available
online at:  and
. Accessed May 2019.
USDA (2019b) U.S. Department of Agriculture. Economic Research Service. Nutrient Availability. Washington D.C.
Available online at:. Accessed May 2019.
USDA (2019c) U.S. Department of Agriculture. National Agricultural Statistics Service. Vegetables 2018 Summary.
Available online at: . Accessed June
2019.
U.S. Poultry (2006) Email correspondence. John Starkey, USPOULTRY to D. Bartram, ERG. 30 August 2006.
References 10-91

-------
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.
Composting
BioCycle (2018a) Organic Waste Bans And Recycling Laws to Tackle Food Waste. Prepared by E. Broad Lieb, K.
Sandson, L. Macaluso, and C. Mansell. Available online at: .
BioCycle (2018b). State Food Waste Recycling Data Collection, Reporting Analysis. Prepared by Nora Goldstein.
Available online at: .
BioCycle (2010) The State of Garbage in America. Prepared by Rob van Haaren, Nickolas Themelis and 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 (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
.
EPA (2016) Advancing Sustainable Materials Management: Facts and Figures 2014. Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at
.
EPA (2014) 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
.
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 .
Institute for Local Self-Reliance (ISLR) (2014). State of Composting in the US: What, Why, Where & How. Available
at .
University of Maine (2016). Compost Report Interpretation Guide. Soil Testing Lab. Available online at:
.
U.S. Census Bureau (2019) Population Estimates: Vintage 2018 Annual Estimates of the Resident Population for the
United States, Regions, States, and Puerto Rico, April 1, 2010 to July 1, 2018. Available online at <
https://www2.census.gov/programs-surveys/popest/tables/2010-2018/state/totals/nst-est2018-01.xlsx>.
10-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------
U.S. Composting Council (2010) Yard Trimmings Bans: Impact and Support. Prepared by Stuart Buckner, Executive
Director, U.S., Composting Council. Available online at .
Waste Incineration
RTI (2009) Updated Hospital/Medical/lnfectious Waste Incinerator (HMIWI) Inventory Database. Memo dated July
6, 2009. Available online at: .
Waste Sources of Precursor Greenhouse Gas Emissions
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) 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
BOEM (2011) OCS Platform Activity. Bureau of Ocean Energy Management, U.S. Department of Interior.
CTIC (2004) National Crop Residue Management Survey: 1989-2004. Conservation Technology Information Center,
Purdue University, Available online at: .
EIA (2019) Monthly Energy Review, November 2019, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/11).
EPA (2019) Greenhouse Gas Reporting Program- Subpart W-Petroleum and Natural Gas Systems. Environmental
Protection Agency. Data reported as of August 4, 2019.
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.
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-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing
81(5):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.
MRLC (2013) National Land Cover Database 2001 (NLCD 2001). Available online at:
. Accessed August 2013.
References 10-93

-------
USDA-ERS (2018) Agricultural Resource Management Survey (ARMS) Farm Financial and Crop Production Practices:
Tailored Reports. Available online at: .
USDA-NASS (2017) 2017 Census of Agriculture. USDA National Agricultural Statistics Service, Complete data
available at www.nass.usda.gov/AgCensus.
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/porta l/nrcs/detail/national/technical/nra/ceap/na/?cid=nrcsl43_014163
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.
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.
10-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2018

-------